Constitutional Blind Spots in the EU Hydrogen Economy: How Atypical Legal Instruments in the Renewable Energy Directive Challenge Legal Certainty

 

This post was originally published on EU Law Live.

Simon Vanhove*

Photo credit: DLR, CC-BY 3.0

The European Union’s ambition to decarbonise its energy system has elevated renewable hydrogen to a strategic priority. For sectors that cannot easily electrify—such as steelmaking, chemicals, and heavy transport—renewable hydrogen and its derivatives (ammonia, methanol) are widely viewed as indispensable. Yet the Union’s regulatory architecture is characterised by a striking tension: strict supply‑side criteria for renewable fuels of non‑biological origin (RFNBOs) coexist with ambitious demand‑side quotas in the Renewable Energy Directive (RED III). This mismatch is not merely technical.  As the European Court of Auditors’ Special Report 11/2024 noted, the framework is mostly in place. Still, complexity, delays, and weak investment signals threaten the ramp‑up required to meet the targets by the regulatory deadlines. The legal instruments used are not helping to provide legal certainty, either.

Policy Backdrop

Hydrogen entered the Union’s canon with the 2020 Hydrogen Strategy (COM/2020/301). That Communication anchored hydrogen in the European Green Deal’s decarbonisation pathway, prioritising renewable production and mapping a phased scale‑up of electrolysers (6 GW by 2024; 40 GW by 2030). In 2022, Russia’s aggression against Ukraine super‑charged hydrogen’s political salience. The REPowerEU Plan (COM/2022/230) endorsed the now‑familiar twin target of 10 million tonnes of domestic renewable hydrogen and 10 million tonnes of imports by 2030, knitting energy security narratives to decarbonisation. The legal linchpin for hydrogen sits in RED III: binding RFNBO uptake in industry and transport, plus a framework of technical rules that define what qualifies as “renewable” hydrogen.

From Delegated Acts to Q&A Documents

Two 2023 delegated acts do the heavy lifting. Delegated Regulation (EU) 2023/1184 elaborates the ‘additionality’ and spatio‑temporal correlation requirements linking hydrogen electrolysis to renewable electricity. Delegated Regulation (EU) 2023/1185 sets the greenhouse‑gas (GHG) accounting methodology and the 70% savings threshold. To further operationalise intricate RFNBO criteria, the Commission published web‑based Q&A documents, first in September 2023, and again in March 2024. They are explicitly styled as “living tools,” disclaiming binding effect and reserving freedom to argue otherwise before the Court of Justice. Crucially, they are not adopted in a formal procedure, not published in the Official Journal, and earlier versions are not systematically archived for comparison. Matching them with the legal acts listed in Article 288 TFEU is difficult, especially since these documents are clearly not Commission communications, such as the related Guidance on targets for the consumption of RFNBOs of September 2024.

Still, these Q&As influence behaviour, because market actors treat them as signals of how authorities will read compliance. That mix—de facto steering, de jure disclaimers—sits uneasily with transparency and legal certainty. It also tangles with the doctrine of legitimate expectations. The Court protects expectations only where assurances are ‘precise, unconditional and consistent,’ and even then expects recipients to stay ‘prudent and alert,’ as the Court indicated in Falqui v. Parliament (C‑391/21 P, paras 105-106).

Recitals as Implicit Carve-Outs

Legal uncertainty does no stop there. Two recitals of RED III are used as stopgaps in the EU’s hydrogen policy.

‘(62) The Union’s hydrogen strategy, […] recognises the role of existing hydrogen production plants retrofitted to reduce their greenhouse gas emissions in achieving the increased 2030 climate ambition. In light of that strategy, and within the framework of the call for projects organised under the Union’s Innovation Fund […], early movers have taken investment decisions with a view to retrofitting pre-existing hydrogen production facilities based on steam methane reforming technology with the aim of decarbonising hydrogen production. For the purpose of calculating the denominator in the contribution of renewable fuels of non-biological origin used for final energy and non-energy purposes in industry, hydrogen produced in retrofitted production facilities based on steam methane reforming technology for which a Commission decision with a view to the award of a grant under the Innovation Fund has been published before the date of entry into force of this Directive and that achieve an average greenhouse gas reduction of 70 % on an annual basis, should not be taken into account.’

Recital 62 acknowledges early movers: fossil hydrogen plants retrofitted with carbon capture, often with Innovation Fund support. It suggests that hydrogen produced there (meeting a 70% GHG reduction) should be excluded from the denominator when computing industrial RFNBO uptake, significantly relieving pressure on governments to achieve those targets. Thus, pre-existing investments (and subsidies) may be salvaged.

Recital 63, then, recognises the specific challenges in integrated ammonia facilities.

‘(63) Moreover, it should be acknowledged that the replacement of hydrogen produced from the steam methane reforming process might pose specific challenges for certain existing integrated ammonia production facilities. It would necessitate the rebuilding of such production facilities, which would require a substantial effort by Member States depending on their specific national circumstances and the structure of their energy supply.’

In contrast to the preceding recital, this recital does not suggest an approach to accommodate those ‘specific challenges’, let alone any criteria or conditions for support.  A Commission statement recorded in Council minutes upon final voting of the RED III indicates that these facilities could be exempted, essentially on case-by-case discretion of the Commission. Amortisation and final investment decisions may be relevant considerations in that respect. The Commission thus pulls itself up by its own bootstraps, installing an ad hoc grandfathering regime based on a single recital, that, at best, acknowledges economic hardship.

As guardian of the Treaties, the Commission oversees implementation and may adopt delegated or implementing acts where empowered to do so. There is, however, no general executive prerogative to carve out obligations when not duly authorised by the co‑legislators. In the case of recital 62 of RED III, the conditions for exclusion appear only in the preamble – not the provisions of the Directive. A lawful derogation would require an explicit empowerment under Article 290 TFEU and respect for the ‘essential elements’ doctrine. Absent that mandate, guidance cannot rewrite targets by administrative interpretation. The same caution applies to recital 63, where the wording is even more nebulous. If transitional flexibilities are politically necessary, the correct path is to legislate them—be it by amending RED III or by a narrow implementing act where authorised. Anything less is convenient in the short run but comes at the expense of legal certainty and equal treatment.

The politics are intelligible: early investments and hard‑to‑retrofit facilities pose transition dilemmas. The legal technique is more problematic. Recitals have interpretative value; they do not, however, derogate from operative provisions. Nor can unilateral declarations in Council minutes alter the legal nature of a directive. The Court’s long‑standing guidance—famously in Antonissen (C‑292/89, para. 18)—makes clear that such declarations cannot drive interpretation where the text is silent. If the legislator wants exemptions, they belong in the articles. Where there is no legal mandate, it is difficult to argue for Commission discretion at all.

Could Q&A documents or Council statements generate protected expectations? Sometimes—if the assurance is precise, unconditional, consistent, and within the administration’s discretion. Where a Q&A simply explains how the Commission will conduct its own assessments, reliance may be reasonable. However, where it delivers an interpretation of EU law, even inadvertently, the disclaimer and the Court’s interpretive monopoly cut the other way. The question then remains what those Q&A documents bring to the table.

Why This Matters: Climate Governance by Law

Governing through Q&As, recitals, and unilateral declarations may buy flexibility in a fast‑moving technological field, but it blurs lines of accountability and invites uneven implementation across Member States. The better course is to anchor exemptions and obligations in law, with explicit delegations, consultation, impact assessment, and judicial review. That path is slower. It is also the one most likely to reassure investors, safeguard equal treatment, and maintain the rule of law as the green transition’s backbone rather than its afterthought.

The EU’s hydrogen economy will not be built by ambition alone. It requires a legally coherent framework that reconciles tight production rules with credible uptake trajectories, robust certification, bankable support schemes, and predictable enforcement. The toolkit already exists: RED III sets the targets and the 2023 delegated acts define what counts as renewable. Ultimately, legality depends on the separation of instruments: laws in the articles; context in the recitals; implementation via empowered acts; explanation in guidance; and politics in press lines.

*Dr. Simon Vanhove is a postdoctoral researcher in energy law at Tilburg University, where he is affiliated with both the Law School’s department on Technology, Law and Society (TILT) and the Academic Collaborative Centre on Energy and Climate. He acknowledges funding of the Belgian federal Ministry of economic affairs under the Energy Transition Fund’s project MuSe (Molecules at Sea) 2023-2025.   

 

Frontex before the EU Courts: Damage, Causation, and Liability under Article 340(2) TFEU

 

Cases: Hamoudi v Frontex   (C-136/24 P) and WS and Others v Frontex (Case C-679/23 P)

 

Antje Kunst*

Photo credit: Rock Cohen

Introduction

In two judgments delivered on 18 December 2025 - Hamoudi v Frontex (Case C-136/24 P) and WS and Others v Frontex (Case C-679/23 P) - the Court of Justice significantly clarified the scope of Frontex’s responsibility for fundamental rights violations.

By referring both cases back to the General Court for re-examination, the Court of Justice highlighted serious flaws in how the General Court approached burden of proof, causation and Frontex’s fundamental rights obligations in actions for damages against the Agency.

In both cases it emphasized that Frontex bears its own obligations to protect the fundamental rights of individuals affected during its operations. This is a positive development which was long overdue. Frontex, like any other EU Agency, must be held accountable for the consequences of its operations. EU law, in particular the Frontex Regulation in connection with the EU Charter of Fundamental Rights is unambiguous on this point, despite the Agency’s persistent attempts to argue otherwise.

In Hamoudi, the Court held that the General Court failed to adapt the burden of proof when assessing whether a pushback operation, in which Frontex was involved, had occurred and whether the applicant was present. In WS and Others, the Court underlined that causation must be assessed in concreto, taking into account the specific circumstances of each case, particularly the vulnerability of asylum seekers.

Factual background of the two cases

The Hamoudi case concerns a young Syrian who was allegedly pushed back in April 2020 by Greek authorities involving Frontex operations. As the Court noted at the time of the alleged incident two Frontex operational activities were ongoing in the geographical zone where Mr. Hamoudi claims it took place. (see factual background at para.10). Mr. Hamoudi requested compensation based on Article 340 (2) TFEU from Frontex for the non-material damage he suffered as a result of the pushback in which Frontex had been involved.

In WS and Others, a Syrian family of six, including children, was unlawfully returned from Greece to Turkey in a Frontex-coordinated joint return operation in 2016, from where they fled to Iraq. The family had expressed their desire to apply for asylum, they were nevertheless placed on the return flight without that there was a return decision by the Member State. The family requested to be compensated for the harm suffered on the same basis as in Hamoudi.

Three conditions under Article 340 (2) TFEU

To establish the non-contractual liability of Frontex under Article 340 (2) TFEU, applicants must satisfy three cumulative conditions: (1) unlawful conduct, (2) actual damage and (3) a causal link between the two.

In Hamoudi, the Court of Justice examined whether the General Court had committed an error of law in its assessment of actual  damage, the second condition. In doing so, the Court of Justice was confined in its review to alleged errors of law on appeal; it could not reassess the facts or evidence as established by the General Court. In WS and Others the Court of Justice focused on whether the General Court had erred in law in its analysis of the existence of a causal link between the alleged unlawful conduct and the damage (the third condition). Upon referral, the General Court will have to reassess the facts and evidence examining whether all three conditions of Article 340 (2) TFEU are met in light of the rulings of the Grand Chamber.

Difficult (factual) assessments will still need to be made by the General Court upon referral. The good news is, however, that the Court of Justice has set out clear parameters for the General Court including active investigation duties as the blog will explain below.

Responsibility for own conduct & What the principle of sincere cooperation really means for Frontex

While the Court of Justice rejects the joint and several liability claim in WS and Others as inadmissible, it does not rule out the possibility of concurrent liability of Frontex and Member States as envisaged by Advocate General Ćapeta in her Opinion (see WS and Others at paras. 85-88 and 127 and here).

In Hamoudi, the Court of Justice stressed that, under the Frontex Regulation, “Frontex is fully responsible and accountable for any decision it takes and for any activity for which it is solely responsible” (para. 66). Beyond that, the Court highlighted that the Regulation requires Frontex, in the conduct of those activities, to ensure respect for fundamental rights. (see, inter alia para. 127).

Similarly, in WS and Others, the Court of Justice made it clear that under the then-applicable Frontex Regulation and the EU Return Directive, Frontex has its own obligations to ensure and effectively monitor respect for fundamental rights during joint return operations. (see inter alia paras. 96 to 102, 130). Joint return operations should only target those persons who are subject to enforceable written return decisions and Frontex is obliged to check that they exist for any individual a Member State plans to include in joint return operations (Frontex’s so-called “verification obligation”) (see paras. 101 and 102) and here.

The Court firmly rejected Frontex’s claim that the principle of sincere cooperation under Art.4 (3) TEU could allow it to evade the specific obligations imposed by the  then-applicable Frontex Regulation, most importantly its verification obligation. Instead, the Court made clear, which is important, that the principle of sincere cooperation requires Frontex to help ensure that Member States comply with EU law, especially fundamental rights. (see paras. 107 and 108)

While the Court of Justice’ s statements on Frontex’s legal obligations in both  cases may seem obvious, their significance lies in the fact that the Court explicitly rejects Frontex’s attempts to shield itself from responsibility, even for its own decisions and activities (see in detail the criticism of this here).

Adapting the burden of proof (Hamoudi)

Relying on Article 47 of the Charter, the Court of Justice held that the General Court applied the rules on the burden of proof and the taking of evidence in a manner incompatible with the right to effective judicial protection in the context of an alleged pushback involving Frontex. Article 47 of the Charter required an “adaptation” of the burden of proof. Individuals who claim to be a victim of a pushback involving Frontex cannot be reasonably expected to produce conclusive proof. It is sufficient to present prima facie evidence that a pushback occurred and that they were present during the incident. (paras. 104 and 110)

The Court of Justice also rightly recognized that, at the time of the facts, victims of pushbacks are in a highly vulnerable position, making it very difficult - or even impossible (probatio diabolica) - for them to gather the necessary evidence.  (para. 88 and see here in detail).  Ignoring this reality would risk granting Frontex de facto immunity (para. 105).

In future cases, the General Court will need to take the applicants’ vulnerability into account when assessing whether damage has been established under Article 340 (2) TFEU – and, as the blog explains also when evaluating the causal link between the unlawful conduct and the damage.

Testimony of a pushback victim can constitute prima facie evidence (Hamoudi)

In Hamoudi, the Court carefully assesses the probative value of the testimony of a pushback victim and concludes that the General Court ought to have concluded that Mr. Hamoudi’s witness statement was “sufficiently detailed, specific and consistent” to constitute prima facie evidence that he had been a victim of a pushback operation. (see paras. 119 and 122). The mere fact that a witness statement contains several statements which are insufficiently specific with regard to essential points of fact does not, in itself, justify the conclusion that it cannot amount to prima facie evidence (para. 120)

Similarly, the Court held that the fact that a person claiming to be a victim of a pushback cannot recall the exact date of their journey to Europe, or cannot clearly identify other victims as witnesses, is not sufficient to undermine the probative value of that person’s witness statement for determining whether it constitutes prima facie evidence (para. 122).

The ruling makes clear to the General Court that an applicant’s witness statement may, on its own, constitute prima facie evidence. This is welcome news for applicants who face serious practical difficulties in obtaining additional evidence (which the Court acknowledged in para. 109).

The Court of Justice also relied on a press article - the Bellingcat article – which it treated as merely corroborative, allowing the prima facie evidence provided by the applicant’s witness statement to stand on its own (see para. 125 and also this article).

Contextual evidence (Hamoudi)

Unlike the ECtHR the Court of Justice’s ruling unfortunately does not address contextual evidence such as reports by the UN or NGOs. In the recent G.R.J. and A.R.E. cases, the ECtHR relied on such material to conclude that there is a systemic practice of pushbacks in the Aegean See. The General Court should similarly consider this type of evidence in future cases, as it can be crucial in establishing the facts surrounding pushbacks and Frontex’s involvement.

Evidence in Frontex’s hands (Hamoudi)

The Court emphasised a crucial point regarding proving pushbacks: due to Frontex’s monitoring role, its operational data-collection tasks, and its obligation to ensure compliance with EU fundamental rights, the Agency “must have” relevant information“ at its disposal” (paras. 127 and 133) to establishing whether pushbacks occurred.  This is even more so, when incidents take place in areas and at times where Frontex is on the ground, like during the rapid border intervention in the Aegean Sea and Joint Operation Poseidon.  (paras. 96 and 97).

In this context, the Court made clear that the Agency cannot simply claim it had no knowledge about an alleged pushback without a proper explanation why this is so (see para. 127).

Evidencing Pushbacks: The General Court’s duty to investigate (Hamoudi)

The Court of Justice rightly criticised in Hamoudi that Frontex had failed to cooperate voluntarily (para. 148) whilst it “must have had at its disposal” the relevant information. In this regard, the Court drew the correct conclusion by making clear to the General Court that it must actively use its investigation powers instead of accepting evidentiary gaps caused by Frontex’s lack of cooperation. The General Court failed to use all procedural tools at its disposal to obtain relevant operational information from Frontex which is responsible for upholding fundamental rights during joint operations. (para. 133)

The General Court can request parties to produce all documents, provide information or appear in person, and even seek evidence on behalf of applicants who face difficulties accessing it (paras. 81 and 82).  Claims of pushbacks cannot simply be dismissed for insufficient evidence. The Court must order, in particular at the request of the applicant, measures to clarify the facts.

This has concrete implications for the proceedings now returning to the General Court. Mr. Hamoudi had requested a number of investigative measures (see paras. 135-141) but they were unlawfully rejected (see paras. 142, 150). On referral, the General Court must fully use its powers to obtain all information from Frontex and importantly should draw appropriate conclusions if Frontex continues to withhold evidence.

From Prima Facie Evidence to Proof (Hamoudi)

The Court of Justice in Hamoudi did not go so far as to say the burden of proof should be reversed (cf. Advocate General’s Norkus’ Opinion). However, it made a crucial clarification: when the General Court, for example on referral, concludes that prima facie evidence of a pushback with Frontex involvement has not been successfully rebutted, whether through the applicant’s testimony at a hearing, evidence from other parties, or the Court’s own investigation, then the fact in question must be treated as proven (para. 132).

In this context, it is appropriate for the General Court to draw inferences if Frontex continues to be non-cooperative and fails to provide requested information. The General Court may, for instance, attach special significance to withheld evidence or question the credibility of Frontex’s account of events. (see here regarding cases before the ECtHR). 

Assessing causal link to damage (WS and Others)

In WS and Others the Court of Justice clarified the causal link between unlawful conduct and damage, namely the costs related to the family’s temporary residence in Turkey, flight to and residence in Iraq. In exceptional cases the connection between the conduct and the harm may remain “unbroken”, even if the affected person made a decision between the conduct and the harm suffered. The Court agreed with Advocate General Ćapeta  (see here an analysis of her Opinion), that this may be reasonable given the extraordinary circumstances faced by asylum seekers, including trauma,  displacement, and risk to safety (see paras. 156 -157 and in more detail here).

The Court emphasized that the assessment of causation cannot be considered in isolation from the factual context of the case, including vulnerability and the risk of refoulment. The General Court upon referral will need to take this factual context into account when it re-examines the causal link.

Costs of legal assistance to be linked to Frontex (WS and Others)

The Court of Justice clarified that the costs of legal assistance incurred in the context of complaints against Frontex may be linked to Frontex conduct (see 166 to 174). The complaints mechanism under the Frontex Regulation  is a tool available to those directly affected by alleged fundamental rights violations by Frontex. Here again, the Court of Justice refers to the particular vulnerability of complainants (para.171), stressing that the decision to seek legal assistance cannot be treated as merely a personal choice, it was a necessary and reasonable step. The General Court will have to take this into account when examining the causal link between the costs of legal representation relating to the complaints mechanism and any unlawful conduct on the part of Frontex.

Conclusion

All eyes are now on the General Court, which must take a fresh look at these two cases. Its room for manoeuvre is limited, given the detailed rulings by the Court of Justice in both cases. Prospects in WS and Others look relatively strong, while in Hamoudi a positive outcome is possible though meeting the causal link requirement may be challenging. What is already clear however, is that the era of Frontex’s non-accountability is over.

 A serious engagement by the General Court is now expected setting the stage for how similar claims will be treated in future litigation.

*Antje Kunst is barrister of Garden Court North Chambers, admitted to the Bar of England and Wales, and the Bar of Berlin, advising and representing individuals in a wide range of matters relating to their fundamental rights in the context of EU external action and other areas, as well as access to justice for individuals, including staff cases. She has appeared in numerous cases before both the Court of Justice and the General Court of the Court of Justice of the European Union, including litigation involving EU bodies, agencies and missions operating under the EU’s external action framework.

 

The Council of the EU’s Position on the Return Proposal: Trump-Inspired Approaches Dominate

 

Dr. Izabella Majcher, Independent Consultant

Photo credit: Alamy stock photo

 

On 11 March 2025, the European Commission published a proposal for a Return Regulation intended to replace the 2008 Return Directive. The stated objective of the proposal is to increase the number of people returned, with the Commission regretting that only around 20% of persons issued with a return decision actually leave. As the proposal explains, it seeks to streamline and harmonise return procedures, attributing the perceived ineffectiveness of EU return policy to divergent national approaches. In this spirit, the Commission proposes replacing the directive with a regulation that would establish “modern, simplified and common” procedures. EU institutions have repeatedly asserted that the current return framework is complex and outdated. However, this claim is difficult to verify, as is the assumption that the proposed regulation would effectively increase return rates. This is largely due to the Commission’s failure to conduct a prior impact assessment, despite this being required under its own Better Regulation Guidelines.

On the side of the European Parliament, the draft report on the Commission’s proposal was presented to the Committee on Civil Liberties, Justice and Home Affairs (LIBE) on 11 November 2025, and the Parliament is currently working towards adopting its final position. The other co-legislator, the Council of the European Union, adopted its “general approach” on 8 December 2025, during the Justice and Home Affairs (JHA) Council, alongside its position on the safe third country concept and on a common EU list of safe countries of origin. This JHA Council was the final meeting under Denmark’s Presidency. Denmark—widely known for its restrictive asylum and migration policies—sought to secure Member States’ agreement on these files before the end of its term (for instance, the Council removes the future mandatory character of the mutual recognition of return decision proposed by the Commission, on which the agreement among Member States proved difficult).

While the Commission’s proposal already displays an overall coercive approach and restricts human rights safeguards—as highlighted in a joint statement by over 200 civil society organisations—the Council’s position further erodes existing protections. This non-exhaustive analysis examines ten key measures introduced or reinforced by the Council that are particularly problematic from a human rights perspective.

 

1) Return decision and the risk of refoulement: the Council explicitly allows disregard of the principle of non-refoulement. It provides that authorities should not be precluded from issuing a return decision where there is a risk of refoulement (Art. 7(4a)). In such cases, the return decision should state that removal to the country concerned is postponed under Art. 14(1)(a). Given that a return decision may be enforceable, issuing it despite an identified risk of refoulement engages the State’s responsibility under Article 3 ECHR. Issuing a return decision in such circumstances—irrespective of a possible postponement of removal—may lead to refoulement in practice, for example if the postponement decision is not adopted in time. Moreover, as the suspensive effect of appeals is significantly curtailed (see below (2)), a person may be removed before such a decision is taken. This approach is also questionable in terms of effectiveness: postponement procedures would still need to be carried out, while the person would remain in an irregular situation, thereby undermining the stated objective of increasing return rates.

In practice, this provision appears to facilitate transfers to third countries acting as “return hubs” (see below (9)) during the postponement of removal to the country designated in the return decision. However, the prohibition of refoulement also covers indirect (chain) refoulement. Transfer to an intermediary country does not absolve the sending State of its responsibility under Article 3 ECHR to ensure that the “return hub” offers sufficient guarantees against onward removal to a risk of refoulement.

2) Procedural safeguards and the right to an effective remedy: the proposal further increases the risk of refoulement by weakening procedural safeguards to challenge return decisions. As introduced by the Commission and reinforced by the Council, a return decision need not specify the country of return (Art. 7(4)), while a removal decision may identify more than one country (Art. 4(4b)). This undermines legal certainty and clarity and places an excessive burden on individuals, who may be required to challenge removal in relation to several potential destinations. The proposal foresees a non-refoulement assessment prior to removal, which is particularly important given that a return decision may be issued despite an identified risk of refoulement (see above (1)). Accordingly, where there are indicators that removal would breach the principle of non-refoulement, the person should be referred to an appropriate procedure, including the asylum procedure, or the risk should be assessed in accordance with national law. However, this assessment may be carried out by the same authorities that issued or enforce the return decision, and may be dispensed with entirely where a prior assessment exists or where the relevant information is deemed insufficiently substantiated (Art. 12(3)–(3a)).

Remedies are also weakened. The Council deletes the Commission’s proposal to suspend enforcement of a return decision where there is a risk of refoulement. Automatic suspensive effect is no longer guaranteed; instead, Member States may grant it only upon request (Art. 28(1)). Given the irreversible nature of the harm, Art. 13 ECHR requires automatic suspensive effect where return is challenged on non-refoulement grounds.

3) Grounds for non-return: under the current Return Directive, Member States may grant an autonomous residence permit or other authorisation offering a right to stay for humanitarian, compassionate or other reasons. In such cases, no return decision shall be issued, or an existing decision must be withdrawn or suspended for the duration of the authorisation (Art. 6(4), current Directive). While the Commission retains this possibility (Art. 7(9), Commission proposal)—albeit outside the exceptions to the obligation to issue a return decision in Art. 8—the Council removes it entirely. Its position contains no human-rights-based exception to the obligation to issue a return decision. Yet such exceptions are essential to ensure respect for rights that may prevent removal, including non-refoulement, the right to private and family life, or other circumstances rendering return unlawful or inappropriate, such as serious health conditions. Eliminating the possibility to grant a residence permit in such situations risks violations of international and EU human rights obligations and conflicts with Member States’ own legal frameworks, many of which provide for regularisation mechanisms. Not every person in an irregular situation can be removed. Issuing return decisions in cases where removal would violate human rights is likely to generate litigation and leave individuals in legal limbo—ultimately undermining, rather than enhancing, the effectiveness of return policy.

4) Detention: the Commission’s proposal significantly expands the legal basis for detention. In addition to the two grounds provided under the Return Directive—risk of absconding and hampering return—it introduces three new grounds: “security risks” (see below (8)), determination or verification of identity or nationality, and non-compliance with restrictive measures (see below (5)) (Art. 29(3)). As discussed elsewhere, these grounds are broadly framed and raise serious concerns as to their necessity and justification. The Council further extends this list by allowing detention on “other relevant, necessary and proportionate grounds” to ensure an effective return procedure, as defined in national law. This open-ended clause undermines legal certainty and permits unchecked and potentially arbitrary detention, in violation of the right to liberty.

The Commission also introduces extensive criteria for assessing the risk of absconding, including three criteria establishing a rebuttable presumption and eight criteria for an overall assessment (Art. 30, Commission proposal). The Council expands both lists further, adding undefined criteria under national law (Art. 21a, Council version). This disproportionate approach risks categorising nearly every person in an irregular situation as a potential absconder, thereby transforming immigration detention from an exceptional measure of last resort into a standard tool of migration control, contrary to international human rights standards.

Despite the requirement under international law that immigration detention be as short as possible, the Commission extends the maximum period of detention from 18 months (6 + 12) to 24 months (12 + 12) (Art. 32(3), Commission proposal). This detention period is exceptionally long and unnecessary, particularly given that immigration detention concerns persons who have not committed a criminal offence and who do not benefit from safeguards applicable under criminal law. The Council allows for repeated six-month extensions of the detention period where there is a risk of absconding and a reasonable prospect of removal arises from specific changes in circumstances (Article 32(3a), Council version). Such potentially indefinite detention amounts to arbitrary detention and violates the right to liberty. In addition, the Council clarifies that these detention limits apply per Member State, allowing a new period of detention to be imposed if the person moves to another EU country (Art. 32(3)).

Procedural safeguards are further weakened: the Council reduces the frequency of detention reviews from three to six months and removes the 15-day deadline for an initial judicial review (Art. 33).

5) Cooperation duties and restrictive measures: the Commission’s proposal imposes extensive cooperation obligations and restrictive measures on persons subject to return, which the Council substantially expands. The resulting web of detailed—and at times overlapping—obligations undermines legal clarity and proportionality. The obligation to cooperate (Art. 21) encompasses fifteen categories of obligations, including remaining available to authorities (such as appearing before third-country authorities for the issuance of travel documents or verification of identity or nationality), providing extensive information, and complying with third-country entry requirements, including health-related conditions. The list is open-ended, as it refers to national law. In practice, many of these obligations may be difficult or impossible for individuals to fulfil.

Non-compliance with cooperation duties triggers the mandatory imposition of restrictive measures, including residence requirements, reporting obligations, and other measures under national law (Art. 23(1)). It may also lead to further sanctions, such as refusal of allowances, extension of entry bans up to 20 years, criminal sanctions including imprisonment, or other penalties provided for under national law (Art. 23b(1)). Failure to cooperate also constitutes a criterion for establishing a risk of absconding (Art. 21a(2)(g)), which may in turn justify the imposition of so-called “alternatives to detention for preventing absconding”—including electronic monitoring and measures defined in national law (Art. 23c(1))—as well as detention itself (Art. 29(3)(a)). Moreover, failure to comply with specific cooperation duties is an independent ground for detention (Art. 29(3)(e)).

In practice, virtually any person subject to return may fail to comply with one or more of these extensive obligations and thus become exposed to a cascading set of restrictive measures and sanctions. The breadth of these measures fails to meet the principles of necessity and proportionality and leaves significant scope for arbitrary application.

6) “Investigative measures”: the Council introduces “investigative measures” (Art. 23a), allowing authorities—without the person’s consent—to search persons subject to return, their homes and other relevant premises, and to seize personal belongings, including electronic devices. Such measures may also be further defined under national law. While safeguards are foreseen, such as necessity, proportionality, respect for fundamental rights and access to remedies, the breadth of this provision raises serious concerns. In practice, it would permit police to raid homes, shelters and other premises that accommodate or support persons in an irregular situation, with significant implications for the right to privacy.

7) Entry ban: while the Commission already expands the use and duration of entry bans, the Council amplifies this approach further. The Commission introduces the possibility of imposing an entry ban without a prior return decision where an irregular stay is detected during exit border checks. The Council removes the Commission’s safeguards, namely the requirements that such a ban be justified by the individual circumstances of the case, respect the principle of proportionality, and safeguard the rights of defence. The Council’s position effectively eliminates the rights of defence by allowing an entry ban to be imposed and notified after the person has already left the territory (Art. 10(4)). It further allows the imposition of an entry ban without a prior return decision where a person departs before such a decision is issued (Art. 10(4a)). This raises serious concerns as to how authorities can assume, without any prior assessment, that a person in an irregular situation would not have a right to stay. Crucially, this deprives individuals of any effective remedy, in breach of Art. 47 of the Charter of Fundamental Rights of the EU.

As regards duration, in cases not involving a threat to public order or security, the Commission extends the maximum length of an entry ban from five to ten years, while the Council further increases it to twenty years (Art. 10(6)). Such durations may severely interfere with the right to private and family life under Art. 8 ECHR, particularly for individuals who have established strong personal and social ties in the EU prior to removal. Although the Council retains the possibility to withdraw, suspend or shorten an entry ban in justified individual cases (Art. 11(2)), it significantly restricts individuals’ ability to request such measures (Art. 11(3)).

8) Persons posing “security risks”: the Commission introduces a distinct category of persons subject to return to whom heightened coercive measures may apply: those deemed to pose “security risks” (Art. 16). The Council further expands this already vaguely defined category. In particular, it includes within the notion of a threat to public policy or security and national security persons subject to an alert in the Schengen Information System (SIS), which itself encompasses situations as minor as attempts to circumvent rules on entry or residence. The Council also defers to national law for defining “other security threats,” further diluting legal certainty. Persons falling within this category may be subjected to additional coercive measures, including detention—which may be carried out in prisons and potentially be of indefinite duration—and entry bans exceeding twenty years. This approach blurs the distinction between criminal and migration law, reinforces harmful stereotypes linking migration and security risks, and circumvents the procedural safeguards that apply in criminal proceedings. While States have a legitimate interest in protecting public order and national security, criminal law should apply equally to all persons under a State’s jurisdiction. Using migration law as a substitute for criminal justice undermines fundamental legal principles and the rule of law.

9) “Return hubs”: among the Commission’s so-called “innovative solutions” to increase return rate is the establishment of “return hubs.” The proposal allows Member States to remove individuals to a third country with which they have concluded an agreement or arrangement (Art. 17). The person need not have any link to that country, nor must the country qualify as a safe third country, although certain conditions are envisaged. The Council’s position clarifies that such agreements or arrangements may regulate the modalities of onward return and that detention may result from their implementation (Art. 17(2a)). It also relaxes notification obligations vis-à-vis the Commission and other Member States (Art. 17(3)), reducing transparency and oversight.

While the externalisation of returns may appear attractive to Member States, recent migration arrangements show that such models often entail significant legal challenges, operational difficulties and high financial costs for the outsourcing State. More fundamentally, return hubs raise serious human rights concerns, in particular regarding detention and onward return. States cannot evade their international obligations by outsourcing migration control. Depending on the degree of involvement, responsibility for violations may be attributed to the externalising State, the third State, or both, including through complicity in internationally wrongful acts.

10) Children: the Council reduces child-specific safeguards throughout the proposal. It weakens protections relating to age assessment (Art. 19), the appointment and role of a guardian for unaccompanied children (Art. 20(2)), and legal assistance and representation (Art. 25). The Council also removes the provision in the current Directive (Art. 10(1)) that requires assistance for an unaccompanied child by bodies other than the authorities enforcing return to be granted before issuing a return decision. This safeguard is crucial to ensure that the best interests of the child are properly assessed prior to any return decision. Under international law, children may be returned only where this is in their best interests. The Council further removes schooling from the explicit grounds for extending the period for voluntary departure (Art. 7(1c)). It also deletes the Commission’s proposed exemption of families with children from return to “return hubs” (Art. 17(4)), leaving only unaccompanied children excluded from this measure. Moreover, children are not exempt from newly introduced coercive measures, including cooperation obligations and sanctions for non-compliance (see above (5)), as well as investigative measures (see above (6)).

As regards detention, both the Commission and the Council disregard international standards according to which children should not be placed in immigration detention at all. The expanded detention regime (see above (4)) applies equally to children. The Council further removes the mandatory ex officio review of detention of unaccompanied children (Art. 33(2)) and relaxes requirements concerning conditions and services in detention, including access to education (Art. 35(2)–(3)). These changes are incompatible with the best interests of the child and with children’s rights to liberty and education under the UN Convention on the Rights of the Child.

 

Concluding thoughts

The Council: as this assessment demonstrates, the Council’s position lacks proportionality and opens the door to arbitrary measures. Human rights safeguards are systematically weakened. Even the fundamental, non-derogable prohibition of refoulement no longer prevents the issuance of a return decision. The accumulation of multiple—and at times overlapping—coercive measures, including far-reaching cooperation obligations and sanctions for non-compliance, runs counter to the stated objective of the reform, namely to streamline and simplify return procedures. Moreover, the Council’s frequent deferrals to national law—including on core issues such as grounds for detention, criteria for establishing a risk of absconding, and the definition of a security threat—undermine the very rationale for replacing a directive with a regulation and defeats the goal of harmonisation of return procedures.

Other players: the Council’s position was made possible by the Commission. The Commission has consistently relied on the narrative of a low return rate, while maintaining a taboo on non-return and regularisation, despite having funded multiple research projects that provided evidence on this matter. The Commission’s proposal itself already contained several measures raising serious concerns from a human rights perspective. By advancing this proposal in a context of political polarisation on migration—where centrist parties increasingly adopt right-wing and enforcement-driven approaches—the Commission has fallen short of its role as guardian of the Treaties and defender of the Union’s fundamental values. Attention now turns to the European Parliament, with the hope that its final report will offer a human-rights-compliant position to restore a measure of balance before inter-institutional negotiations begin.

First they came for migrants: the Council’s position reflects the view, advanced by some States, that the European Court of Human Rights—or human rights more broadly—has become excessively protective of migrants in an irregular situation and unduly restrictive of States’ prerogatives to expel them. To be clear, any attempt to narrow the scope or application of human rights conventions in relation to migrants constitutes an attack not only on migrants’ rights, but on the rights of all persons. Once human rights are weakened for one group deemed politically expendable, derogations and qualifications can be expected to follow for others. What is presented as an exceptional response to migration thus risks becoming a broader erosion of the universality and indivisibility of human rights protection.