The “Wojewoda Mazowiecki” ruling: the recognition of same-sex couples to have and to hold, to love and to cherish, until death unites them in a gold two-handled urn*

 

Chiara De Capitani (PhD), a linguist agent at the European Commission and member of Amnesty International Belgique francophone’s Sexual Orientation and Gender Identity (OSIG) coordination. The views, thoughts and opinions expressed in this paper are solely that of the author and do not necessarily reflect the views, policy or opinion of the European Commission or of Amnesty International.

Photo credit: O. Ortelpa, via Wikimedia Commons

 

Introduction

In the Wojewoda Mazowiecki case (hereafter “the present case”), the Court of Justice of the European Union (hereafter “the Court”) was called upon to rule on whether to extend its jurisprudence concerning the civil registration of a surname or gender identity change obtained in another Member State to marriage certificates. In the present case, the applicants' State of origin does not recognize same-sex marriages and refused to transcribe a marriage certificate concluded in another EU Member State (hereafter “host Member State”).

In the present case, the Grand Chamber of the Court examined several crucial aspects of the recognition of same-sex marriages and found, for the first time, the existence of a general principle prohibiting any form of discrimination based on sexual orientation.

Indeed, according to the Court, although the rules relating to marriage and civil status fall under the jurisdiction of the Member States, they are obliged to recognize - for the purpose of exercising the rights conferred by EU law - the marital status acquired legally in another Member State. Furthermore, although Member States may establish the procedures for such marriages' recognition, these procedures must not render impossible or excessively difficult the implementation of rights conferred by European law.

Additionally, the present case highlights the necessity to ensure the effective enjoyment of rights that a Union citizen derives from Article 21 TFEU, notably that the family life that this citizen has created or strengthened in a Member State can be continued upon their return to their Member State of nationality.

The dispute in the main proceedings

Mr Trojan and Mr Cupriak-Trojan are a Polish and German-Polish same-sex couple who got married in Germany in 2018. Wishing to reside in Poland as a married couple, they filed a request for the transcription of the German marriage certificate in the Polish civil status register with the head of the Warsaw civil registry office in 2019. However, this request was denied on the grounds that Polish legislation does not allow same-sex marriages. After a series of failed appeals, the case was finally brought by referring Court to the Court of Justice.

Scope of the case: right to freedom of movement or EU citizenship?

At first glance, the facts resemble those of Coman in 2018 - a case concerning the application of the right of free movement to same-sex marriages - but the context is different here because both applicants are European citizens. Therefore, their right to move and reside freely within the Union does not depend on the recognition of their marriage for the purposes of family reunification under the provisions of the freedom of movement directive, but instead relies on the rights conferred to them by European citizenship according to Articles 20 and 21 of the TFEU.

Firstly, the Court recalls that, pursuant to these articles and its established case law, citizenship of the Union confers on each Union citizen a right to move and reside freely within the territory of the Member States and to rely on the rights pertaining to said EU citizenship (paras 41 and 42). These rights include the ability to lead a “normal family life” both in the host Member State and in the Member State of which the citizen is a national, benefiting from the presence of family members upon their return.

Additionally, the present case allows the Court to address one of the several unresolved issues from Coman: the obligation to recognize the marriage between the applicants extends beyond the specific context of family reunification, “extending the obligation to all areas of life in which a married couple may need to invoke their status as a couple” (Tryfonidou).

Practical effects of the non-recognition of the applicants' marriage

In light of these considerations, the Court specifies that the refusal by Polish authorities to recognize the applicants' marriage would cause them significant prejudice, resulting in serious administrative, professional, and private inconveniences (para 51). Indeed, the lack of recognition of their “certain” legal status, which is also enforceable against third parties, would hinder the everyday organization of the applicants' family life in their country of nationality - both in public and private spheres (paras 52-54).

The applicants, in their observations submitted to the Court, clarified that they had already experienced a series of constraints due to the non-recognition of their marital status: for example one of the applicants did not benefit from public health insurance coverage when he was unemployed, and his request to update his surname in the land register was rejected for one of his properties (para 50).

Furthermore, the Court has previously clarified in the Coman ruling that the notion of “public policy” as justification for a derogation from a fundamental freedom must be interpreted strictly and may be relied on if only there is a genuine and sufficiently serious threat to a fundamental interest of society (see Coman, para 44). In the present case, the Court notes that the obligation for the State of origin to recognize a marriage between same-sex Union citizens, contracted in the host Member State, does not undermine the institution of marriage in the State of origin, which remains defined by national law and falls under the competence of the Member States.

Obligation to establish a legal framework providing the recognition of same-sex couples

Polish legislation does not currently recognize the possibility for same-sex individuals to marry or conclude a civil union, since marriage is defined by the Polish Constitution as a union between a man and a woman. This notion was recently criticized by the European Court of Human Rights (hereafter “ECtHR”). In the Przybyszewska and others v. Poland ruling, the ECtHR found a violation of Poland’s positive obligation to protect private and family life arising from Article 8 of the European Convention for the Protection of Human Rights and Fundamental Freedoms (hereafter “ECHR”), due to the absence of a national legislative framework offering recognition and legal protection to same-sex couples.

The rights guaranteed by Article 7 of the EU Charter of Fundamental Rights (hereafter “The Charter”) have the same meaning and scope as those guaranteed by Article 8 of the ECHR, with the latter provision constituting a minimum protection threshold. Therefore, the Court confirms, failure to recognise the marriage concluded by two Union citizens of the same-sex having exercised their freedom to move and reside is contrary to Article 7 of the Charter (para 67).

It is interesting to note that the Court appears to foresee the possibility of such a Charter violation even in a situation where people have stayed in another Member State to marry “on the grounds that” the law of the Member State of which they are nationals does not allow same-sex marriage and, therefore, only for the purpose of circumventing such a lack of authorization. This possibility appears to be implied throughout the text of the present case due to the recognition by the Court of the right of same-sex individuals to continue a family life that would only have been “strengthened” in the host Member State – without, I believe, the need for them to “genuinely reside” there.

It follows from Article 7 of the Charter and Article 8 ECHR that a Member State that does not authorize same-sex marriages must establish adequate procedures for recognizing such marriages when legally concluded in a host Member State by two Union citizens exercising their freedom of movement and residence rights.

On this point, several authors (Luku, Tryfonidou, Acconciamessa) question the risk of “downgrading” due to the divergence of rights recognized by Member States who do recognise unions contracted by same-sex individuals. Indeed, the rights recognized to same-sex individuals vary depending on the type of union that the Member State allows them to contract - civil union or marriage - as well as the content of the rights (notably regarding the recognition of parenthood). Moreover, the applicability of the protection recognised by the Court in the present case to same-sex marriages contracted outside the European Union remains uncertain (Tryfonidou, Bribosia and Rorive).

Choice of procedures for recognising marriages concluded in a host Member State

As a preliminary point, the Court notes that Member States retain the power to choose the procedures for recognizing marriages concluded in a host Member State and that the procedure of transcribing marriage certificates in the civil status register constitutes only one modality among others (paras 69-70).

However, to avoid diverging results between Member States and their national procedures, the Court sets up a series of requirements Member States have to respect. Firstly, whichever types of procedures they pick, they must not render “impossible or excessively difficult” the implementation of rights conferred by Article 21 TFEU. Secondly, the Court assesses the specific situation of Member States - such as Poland - where the transcription of the marriage certificate in the civil status register is the only way possible to legally recognize a marriage concluded in a host Member State. In these circumstances, given the absence of a recognition procedure equivalent to that granted to opposite-sex couples, the Court recognises the existence of discrimination based on sexual orientation, prohibited by Article 21, paragraph 1, of the Charter.

It follows from the above that when a Member State provides a single procedure for recognizing marriages concluded in another Member State, it must apply this modality indistinctly to marriages concluded between same-sex couples and those concluded between opposite-sex couples.

This approach, both protective and pragmatic, aligns - I believe, with the Court's recent case-law regarding the recognition of a person's gender identity. Similarly to the commented case, in the Deldits ruling, the Court prohibited unnecessary or disproportionate administrative practices for rectifying personal data relating to a person's gender identity in a public register, further adding that “a medical certificate - including a psychiatric diagnosis” might suffice (Deldits ruling, paras 47-49).

Prohibition of discrimination based on sexual orientation: a general principle of EU law with direct effect

What truly distinguishes the commented decision is the Court's interpretation of Article 21 of the Charter—dealing with the prohibition of any discrimination based notably on sexual orientation—to which the Court recognizes an imperative character as a general principle of EU law (para 70). Indeed, the existence of such a general principle has only been expressly recognized until now for discriminations on grounds of age (Mangold, Association de médiation sociale, Dansk Industri rulings) and religion (Egenberger, Cresco Investigation rulings).

The Court further specifies that Articles 20 and 21 TFEU and Articles 7 and 21 of the Charter “are sufficient in themselves and do not need to be made more specific by provisions of EU or national law to confer on individuals rights which they may rely on as such”. Therefore “if the referring court were to find that it is not possible to interpret its national law in conformity with EU law, it would be required to [disapply] the national provisions concerned” (para 76).

While the recognition that the prohibition of any discrimination based on sexual orientation constitutes a general principle of EU law is revolutionary— its practical effects are currently difficult to appreciate.

Conclusion

The present case undoubtedly constitutes a turning point in the protection of LGBTIQ+ individuals. The Court, embracing its recent and consolidated jurisprudence (Mousse, Deldits, Mirin) concerning the rights of non-binary and transgender individuals in the EU, further enshrines the principle that “lives, names, and ties should not unravel when crossing internal borders” (Escach-Dubourg).

It is worth mentioning, in this regard, that the Court has recently received a request for a preliminary ruling by a Romanian court similar to the present one. This request would open the door for the Court to clarify a series of unsolved questions regarding the cross-border mobility of same-sex couples. The case concerns the non-recognition of an Italian civil union by Romania for the purpose of granting a residence permit to a spouse from a third country.

Firstly, the preliminary ruling request concerns the recognition of a civil union rather than a same-sex marriage, allowing the Court to clarify whether the protective considerations highlighted within the present case also apply to this type of union. Moreover, the case will enable the Court to confirm if a same-sex couple can “strengthen” their union in a Member State without actually residing in it.

Finally, the Court's reasoning is particularly striking: it is grounded in the search for practical solutions to real, current administrative obstacles encountered by same-sex couples in several Member States. The Court favours transparent, direct solutions to avoid the possibility that same-sex couples live in the shadows and face relentless administrative hurdles due to its ruling’s lack of clarity. Such an approach is reminiscent of the landmark Minister voor Immigratie en Asiel v X and Y and Z 2013 ruling dealing with the conditions for granting refugee status. Asked whether homosexual third country nationals could be expected to conceal their orientation from everyone in their country of origin in order to avoid persecution, the Court famously replied that: “requiring members of a social group sharing the same sexual orientation to conceal that orientation is incompatible with the recognition of a characteristic so fundamental to a person’s identity that the persons concerned cannot be required to renounce it” (Minister voor Immigratie en Asiel v X and Y and Z ruling, para 70).

* [Patroclus to Achilles]: “A last request—grant it, please. Never bury my bones apart from yours, Achilles, let them lie together . . . just as we grew up together in your house. […] So now let a single urn, the gold two-handled urn your noble mother gave you, hold our bones-together!”

The Iliad, Book 23, lines 99-110

National Courts, Judicial Independence, and the Risk of Getting It Wrong

 

Darren Harvey, Senior Lecturer in Law, Kings College London

Photo credit: Lukas Plewnia, via Wikimedia Commons – see also  www.polen-heute.de

Introduction

Much has been written in praise of the Court of Justice’s role in promulgating legally binding obligations pertaining to judicial independence on the basis of Article 19(1) TEU. To ensure effective legal protection in the fields covered by Union law, Member States must ensure that all national courts which may rule on the interpretation or application of Union law meet the requirements of independence, impartiality and previous establishment by law.

But might there be a trap door built into the rather impressive stage that the Court has been constructing in recent years? I happen to think so. Simply stated, my concern is this – what happens when national courts try to interpret and apply the manifold legal principles flowing from the CJEU’s judicial independence case law and get things wrong?

The stakes have recently been raised by the CJEU’s confirmation in AW “T” that provisions of national law may need to be declared null and void as a matter of domestic law whenever they infringe EU judicial independence standards. This includes lower-instance courts being obliged to find decisions of hierarchically superior national courts to be null and void.

Thus, in the name of protecting the rule of law, one now faces the very real prospect of national courts (either in good or in bad faith) expunging measures from their domestic legal systems on the basis of an erroneous application of EU judicial independence standards.

The Risk of National Courts (incorrectly) Policing Judicial Independence

The recently decided Daka case offers a good illustration of the concerns at play here. The case concerned decisions taken by the President of a national court to re-assign judges from one chamber to another chamber of that court. Those re-assignment decisions were taken without the consent of the re-assigned judges, did not contain any reasons and did not seem to be open to judicial challenge. Furthermore, the appointment of the court President issuing the relevant decisions was itself incompatible with Article 19(1) TEU. These factors prompted a national court, composed of re-assigned judges, to refer questions on the compatibility of the above with Article 19(1) TEU. In response, the CJEU surprised commentators (here and here) by finding that the re-assignment decisions of the non-independent court President were compatible with judicial independence standards flowing from Article 19(1) TEU.

But what if, instead of seeking guidance from the CJEU, the national court in Daka had concluded, not unreasonably in my view, that the re-assignment decisions were contrary to EU standards of judicial independence? In light of the abovementioned development in AW “T.” , a national court in a position similar to that in Daka might plausibly conclude that that national measure must be declared null and void.

The problem, however, is that we know that in Daka itself such a finding would have been incorrect as a matter of EU law. And it is here, I submit, that a possible trap door becomes partially visible. The problem stems from the role that EU law requires national courts to play in defending judicial independence on the ground in the Member States. According to settled case law, “the guarantees of access to an independent and impartial tribunal previously established by law, and in particular those which determine what constitutes a tribunal and how it is composed, represent the cornerstone of the right to a fair trial. That right means that every court is obliged to check whether, as composed, it constitutes such a tribunal where a serious doubt arises on that point…such a check is an essential procedural requirement, compliance with which is a matter of public policy and must be verified of the court’s own motion,.” There is also an “inextricable link” between the requirement of access to a court or tribunal previously established by law and the guarantees of independence and impartiality of judges. Finally, the requirements of judicial independence flowing from Article 19(1) TEU are capable of direct effect and must, therefore, be dutifully applied by national courts. It follows that the main responsibility for ensuring compliance EU judicial independence standards lies with national courts, whether in relation to their own independence, or the independence of other courts.

Consider in this regard the recent judgment on whether Article 19(1) TEU precluded national rules on setting national judicial salaries. In that case, the CJEU set down several very detailed principles that must henceforth be considered when national authorities enact rules for setting national judicial salaries. These principles also form the basis of any subsequent judicial review. They are intended to be applied by national courts to national rules and practices. Accordingly, what would prevent a national judge from incorrectly applying those principles in a future case and deciding, as a result, that national rules on judicial salaries must be disapplied or even declared null and void? And what if that national judge erroneously finds a violation of Article 19(1) TEU in such a case, without ever having referred the matter to Luxembourg? Crucially, the wide-ranging scope of application of Article 19(1) TEU means that virtually all national laws and practices touching upon the organization of justice within the State could now suffer a similar fate. Thus, far greater disruption can be caused to national legal orders by a misapplication of EU judicial independence principles than in other areas of EU law. There is a world of difference between a national court erring in its application of EU law governing free movement of workers, say, and the same national court erring in its conclusion that the national supreme court lacks independence under Article 19(1) TEU standards, with all the consequences that that may now entail.

It is worth dwelling on what those consequences might now be for national justice systems. As Dougan points out, disapplication is an individual remedy provided by a national court to resolve a particular dispute. Once disapplied, the national law in question remains in existence and operation for all other purposes. In contrast, declaring a provision of national law to be null and void is a remedy of general effect and application. The relevant national provision, once nullified, ceases to exist, including for all other purposes. Thus, Dougan notes that nullity is “a very different, more robust and more far-reaching, response to incompatibility than disapplication.” The former “carries potentially widespread and more penetrating implications within and across the national legal system, for example as regards the impact on other relationships, and for other powers or actions, as well as for third parties.” Admittedly, the circumstances in which this novel remedy of nullity will be required as a matter of Union law remain uncertain. Nonetheless, there would appear to be more at stake for the functioning of national justice systems in a world where national courts may declare measures of national law to be null and void.

One must also be attentive to an altogether more worrying prospect that, I believe, is now in play here. In the name of EU judicial independence standards, lower-instance courts may utilise these developments to disapply or invalidate judgments of hierarchically superior national courts. It is perfectly foreseeable that, in some national legal orders, first instance courts and tribunals will be far more susceptible to having their independence eviscerated by the ruling executive than apex courts in that state. In circumstances where the constitutional or supreme court remains independent, but the lower courts do not, Article 19(1) TEU, coupled with the primacy principle, can now be used (in bad faith) by those captured lower courts to effectively neutralize decisions of their (independent) superior courts that would otherwise be binding upon them. Again, this can be done (disingenuously of course) in the name of upholding EU standards of independence.

Case C‑448/23, Commission v Poland (Polish Constitutional Tribunal) as a Possible Solution?

Now, the response here might be to fall back on the classic doctrines of EU law. The risk that national courts might err in their application of EU law has always been a structural feature of the EU legal system. There is nothing novel or surprising about EU law disrupting national judicial hierarchies. The above may be nothing more than an extension of Simmenthal to the rule of law context.  Moreover, tools such as Article 258 TFEU infringement proceedings and actions for state liability in damages serve to mitigate against these risks. It may also be possible to appeal against a judgment of a national court that allegedly misapplies Article 19(1) TEU, and for any appellate court to subsequently refer the matter to Luxembourg. However, all these corrective mechanisms have shortcomings that will be familiar to readers of this blog.

It is with this in mind that we can turn, in closing, to the recent CJEU judgment concerning the Polish Constitutional Tribunal. In my view, the judgment makes some important points about the role of national courts in ensuring compliance with Article 19 TEU standards, particularly in light of the problem I have identified above. The judgment hints in places (paras 102-120) at a solution to those problems that would centralize review in Luxembourg. The Court stresses that it has exclusive jurisdiction to give definitive and binding interpretations of EU law, including the requirements flowing from Article 19(1) TEU. Then, “in order to ensure the uniform application of EU law in all the Member States, as is required by Article 19(1) TEU, those requirements cannot depend on the interpretation of provisions of national law, or on the interpretation of provisions of EU law by a national court, which does not correspond to that of the Court (para 210, emphasis added). Thus, Article 267 TFEU sets up a system of judicial dialogue between national courts and the CJEU with the aim of securing the uniform interpretation of EU law. “This is consequently the path that a national court must take when it has doubts as to the compatibility of national law with a provision of EU law requiring interpretation of the latter.” (para 212, emphasis added). Moreover, the obligation upon national courts to disapply any national rule that is inconsistent with Article 19 TEU arises “in the event of an established infringement of the second subparagraph of Article 19(1).” (para 115, emphasis added). Finally, the obligation to declare provisions of national law to be null and void is said to arise when “a national court or tribunal which has referred a question to the Court for a preliminary ruling” and subsequently “concludes that it is apparent from all the conditions and circumstances” that national law is incompatible with Article 19(1) TEU (para 120).

To my mind, these passages hint in the direction of finding that national courts must refer questions to Luxembourg whenever they harbour doubts as to the compatibility of national law with Article 19 TEU standards. How else would a national court ensure that it did not arrive at an interpretation of Article 19(1) TEU that does not correspond to that of the CJEU? How else might a national court determine that there had been an “established infringement” of Article 19(1) TEU before disapplying or nullifying the infringing provision of national law? Such a centralized solution, if confirmed, would essentially extend the Foto-Frost obligation to refer to Luxembourg questions of validity of national law when they arise in relation to Article 19(1) TEU. Such a solution would certainly mitigate against the risks of national courts causing major disruption to the organization of national justice systems whenever they incorrectly apply EU judicial independence standards. However, it would sit uncomfortably with the structure and wording of Article 267 TFEU, which affords lower-instance national courts with a discretion to refer (or not refer) questions of interpretation to Luxembourg. It would also sit uncomfortably with the obligation that direct effect and primacy place on all national courts to give full and immediate effect to Article 19(1) TEU standards by disapplying inconsistent provisions of national law. The alternative reading is that determinations of whether national rules and practices comply with Article 19(1) TEU standards remains a task that is primarily for national courts, most of whom are not obliged to refer such matters to Luxembourg.

Conclusion

The above is an attempt to articulate what I take to be a difficulty in the law of judicial independence as it currently stands. There is always a risk in such endeavours that one is identifying a problem that does not really exist. Perhaps there is no trap door built into the stage. It is also possible that the problem identified does exist, but is not as significant or difficult to solve as the above analysis might suggest. To be corrected on either of these points would be very helpful. It would mean that the stage rests on a firmer foundation than first feared. Alternatively, the above analysis might well be on to something. If that proves to be true, it will at least offer the consolation of being persistently troubled about the right sort of thing.

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.