Hamoudi v Frontex: Advocate General Norkus’ Opinion - Reversing the Burden of Proof and the Presumption of Frontex’s Privileged Access to Evidence

 

Antje Kunst*

*Antje Kunst is an international lawyer and 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 related to fundamental rights within the CFSP and other fields. She has appeared in numerous cases before both the Court of Justice and the General Court, within the Court of Justice of the European Union.  

Photo credit: Francesco Placco, via Wikimedia Commons

 

Introduction

Following the hearing  on 4 February 2025 by the Grand Chamber on the appeal of Hamoudi v Frontex in Case C-136/24 against the General Court (GC)’s Order of 13 December 2023 the Advocate-General rendered his opinion on 10 April 2025 (‘Opinion’).  

As outlined  here, the case of Hamoudi v. Frontex involves a Syrian asylum seeker, Mr. Alaa Hamoudi, who, claims that on 28 and 29 April 2020 he was a victim of collective expulsion in the Aegean Sea. In this context, Mr. Hamoudi asserts that while a group of 22 individuals including himself were at sea, a private surveillance aircraft, equipped with a camera and operated by the European Border and Coast Guard Agency (Frontex), flew over the scene twice. A highly damaging report by the EU Anti-Fraud Office (‘OLAF report’) on Frontex was made public through an NGO in October 2022 and covered precisely the period in question — including April 2020 — and confirmed that incidents of pushbacks occurred under Frontex's watch, involving Frontex surveillance aircraft in the Aegean Sea.

In an action for damages before the EU General Court, Mr. Hamoudi requested that he be compensated by Frontex for the non-material damage suffered due to the collective expulsion. In his arguments before the General Court Mr. Hamoudi inter alia asserted that because Frontex surveillance aircraft flew over the collective expulsion it had specific knowledge of that incident but failed to report it and take appropriate action as per its mandate (see para. 11 of the Opinion)

The General Court dismissed the action as manifestly lacking any foundation in law. It found that Mr. Hamoudi had failed to prove the actual damage he claimed to have suffered. The evidence adduced by the appellant ‘was manifestly insufficient to demonstrate conclusively that he was present at and involved in the alleged incident of 28 and 29 April 2020’ (see paras. 39 and 62 of the Court Order).  

Advocate General's limited Analysis: Reversal of the Burden of Proof

As requested by the Court of Justice of the European Union (‘Court of Justice’ or ‘CJEU’) the Advocate General centred his analysis on a possible reversal of the burden of proof in relation to the existence of damage in collective expulsion cases. (para.22 of the Opinion).  In doing so, he looked in detail at the case-law of the Court of Justice in various areas including discrimination (paras. 30 to 37 of the Opinion) and the case law of the European Court of Human Rights (ECHR) in expulsion cases (paras. 40 to 50).

In his analysis of the case-law of the ECHR in expulsion cases he considered the recent successful collective expulsion case of A.R.E. v Greece in particular ‘instructive’.

Importantly, the Advocate General referred to the fact that in A.R.E., the ECtHR applied its case law on the burden of proof related to secret detention cases. Once the applicant presents prima facie evidence, the Court may draw strong adverse inferences from the respondent government’s failure to disclose essential documents to establish the facts, or failure to offer a satisfactory and convincing explanation of the alleged events. (paras. 41 to 44 of the Opinion)

In line with the jurisprudence of the CJEU and ECtHR, the Advocate General rightly emphasizes that the burden of proof cannot be placed on the claimant to establish facts or produce evidence that is likely to be exclusively in the possession or control of the respondent, here Frontex. If any concrete evidence of collective expulsions exists, it is far more likely to be held by the alleged perpetrator rather than the victim. (para. 51 of the Opinion)

The Advocate General’s three conditions

In the Advocate General’s view, based on the CJEU’s and ECtHR’s case law, the burden of proof may be reversed in cases like Mr. Hamoudi’s case when three conditions are met.

The FIRST CONDITION: (see para. 57 of the Opinion) is that the claimant must present prima facie evidence in support of his or her claim. If the claimant’s account is inconsistent, incoherent, or if the claimant lacks credibility, this initial burden is not met, and the case should be dismissed. Consequently, the reversal of the burden of proof becomes relevant only once a prima facie case has been established. In this context, the Advocate General relies in particular on the case law of the ECtHR, for example as set out in the recent case of G.R.J. v Greece (see § 179).

As stated here, from the questions of the judges at the hearing it appears that the Court of Justice is considering to hold that prima facie evidence had been presented by Mr. Hamoudi. In any event, the Advocate General correctly opines that the General Court may have placed the ‘evidentiary bar’ too high in the first place (para. 64 of the Opinion). In this context, though, the Advocate General omits that the General Court in its assessment of the evidence failed to consider the various contradictory public statements about the events by Frontex outside the proceedings, in particular when the OLAF report was made public and crucial general contextual evidence. Such general contextual evidence was considered in the recent G.R.J. and A.R.E. cases by the ECtHR. This evidence led the ECtHR to conclude that there is a systemic practice of pushbacks in the Aegean See and that the Greek Government failed to refute the evidence by providing a satisfactory and convincing alternative explanation (see § 190 in G.R.J. and § 229 in A.R.E. and footnote 80 of the Opinion).

The SECOND CONDITION: (para. 59 of the Opinion) for the burden of proof to shift is the existence of a clear or structural imbalance in access to evidence—specifically, where the claimant faces significant obstacles in presenting evidence, while the respondent, in this case Frontex, is in a better or more privileged position to refute the allegations. Mr. Hamoudi, a vulnerable asylum seeker with his mobile phone confiscated, lacks the evidence to prove Frontex’s involvement in the events, whereas the respondent is best placed, in a privileged position to prove or disprove the allegations (see in more detail under: ‘Presumption of privileged access must be applicable to Frontex’). There can be no doubt that the second condition is met.

The THIRD CONDITION: (para. 60 of the Opinion) is that the failure to shift the burden of proof would render ineffective the claimant's (fundamental) rights protected under EU law while a shift would not undermine the respondent’s (fundamental) rights under EU law. In Mr. Hamoudi’s case the failure to shift the burden of proof would render ineffective his fundamental rights inter alia under Article 19 of the EU Charter of Fundamental Rights and undermine his right to an effective remedy under Article 47 of the Charter. The shift would not undermine any (fundamental) right of Frontex under EU law. On the contrary, by providing evidence to prove or disprove the allegation, Frontex assists the Court in reaching an accurate outcome (regarding a public authority’s duty pursuant to Article 24 of the Court’s Statute to assist the Court in reaching the correct result see more here).

Presumption of privileged access to evidence not applicable?

Somewhat surprisingly towards the end of the Advocate General’s Opinion, he expresses the view that the presumption established in the jurisprudence of the ECtHR on expulsion cases—namely, that the claimant is at a disadvantage in presenting evidence while the respondent is in a stronger or more privileged position to rebut the allegations – may not be applicable to Frontex. The ‘automatic’ shift of the burden of proof, once prima facie evidence has been presented, allegedly could only be applied when the authorities of a Member State are involved (paras. 61 and 62 of the Opinion). This despite the fact that the AG opined earlier in his Opinion (para. 29) that the differences in the procedures before Strasbourg and the CJEU are more apparent than they are real.

According to the Advocate General, actors like Frontex possess more limited powers compared to the authorities of a Member State, and it was unclear whether—and to what extent—their actions contribute to the difficulties claimants face in adducing evidence of their involvement in the events at issue. Due to these limited powers, it was not clear that they would be in a better or more privileged position to rebut the claimant’s allegations (para. 62 of the Opinion).

The Advocate General’s suggestion that Frontex’s actions must have contributed to the difficulties experienced by the claimants in adducing evidence of their involvement in the events at issue is misplaced. This is not a requirement under the case law of the Court of Justice, nor under that of the ECtHR. According to this case law, it is sufficient to establish the evidentiary difficulties faced by the applicant and the respondent’s capacity to provide evidence in rebuttal.

It is wholly irrelevant whether a Member State may have more authority over certain types of evidence.  The issue is not whether Frontex differs from a Member State in terms of powers over evidence but rather who holds relevant evidence, and who does not. This must be assessed in casu,  e.g. here in relation to Hamoudi and Frontex. The Agency is not being asked to produce evidence concerning Member States’ actions, but rather evidence regarding its own actions or inactions, and knowledge, particularly in light of its established presence and two active joint operations in the region.

At the hearing, Frontex claimed its hands were tied because it did not have access to evidence held by the Member State. This is irrelevant, what matters is that it has evidence in its possession and control that may reveal what the agency did, failed to do, or knew regarding the alleged collective pushback action on 28 and 29 April 2020 in the Aegean Sea.

Undesirable Consequences of having to demonstrate Privileged Access to Evidence

If, as the Advocate General proposes, the burden of proof only shifts when it is first shown that Frontex is better placed to refute the allegations than a claimant, then in practice, the burden might never shift. The Court of Justice should firmly reject this approach.

Accepting that Frontex is - unlike a Member State - not presumed to have privileged access to evidence would undermine its positive obligations to protect fundamental rights of individuals in distress at sea and enable Frontex’s impunity and outright ignores a textual reading of article 7(4) in line with article 80(2) and 80(3) of the Frontex Regulation, which outlines the exclusive positive human rights responsibilities that Frontex has within the context of joint operations. It would risk making Frontex’s legal obligations unenforceable. This would also contradict the Advocate General’s THIRD CONDITION outlined in paragraph 60 of the Opinion.

Presumption of privileged access must be applicable to Frontex

Related to Frontex’s own actions, its own responsibility to comply with its own fundamental rights obligations as per the Frontex Regulation, and the harm resulting for claimants, Frontex must be presumed of being in a better or more privileged position in collective expulsion cases such as the present one.

The appellant is a Syrian refugee, pushed back at night, with his mobile phone confiscated and lacks access to the evidence which shows Frontex’s involvement. Frontex, on the other hand, operates its own aerial surveillance systems and joint operations logs and collects video and radar data from its flights. The likelihood that it possesses or has under its control relevant video surveillance footage and/or incidents reports, relevant logbooks on surveillance operations is extremely high which places Frontex in a particularly privileged position with regard to access to evidence in these types of cases. Moreover, it was established in the OLAF report and during the hearing that Frontex was fully aware of the practice of so-called ghost landings pursued by the Hellenic coast guard and from that flows the duty within the context of its joint operations to collaborate with Member State authorities to prevent human rights abuses.

While it is true, as the Advocate General points out, that the General Court did not examine Frontex’s involvement in or knowledge of the alleged events of 28 and 29 April 2020, this omission stems from the General Court’s erroneous exclusive focus on the question of damage to the appellant and whether he was present and affected by those events. However, this cannot mean that Frontex cannot be regarded as possessing or having under control the evidence needed to rebut Mr. Hamoudi’s allegations. (c.f. para. 63 of the Opinion). The very nature of ghost landings entails that the Hellenic Coastguard is no longer present at sea.

Accordingly, Frontex is in the exclusive position and best placed to confirm or deny the prima facie evidence provided by Mr. Hamoudi regarding the events that took place at sea, including whether it had a surveillance aircraft with a camera operating over the area during the collective expulsion of 28 and 29 April 2020 but failed to report it.

If there is an shift of the burden of proof based on the available contextual and personal prima facie evidence, Frontex would not face a probatio diabolica (an impossible proof). It is also not an unreasonable proof for Frontex to provide: recall that its access  led to the damning  report by OLAF which affirmed its presence on the night of 28-29 April. Frontex is not asked to prove facts that lie completely outside its sphere of influence and knowledge (see the case law of the CJEU cited at para. 53 of the Opinion). To the contrary: its being asked to provide evidence regarding events taking place within the very area of the joint operations in the region, falling entirely and exclusively within its mandate.

Conclusion

The Court of Justice in the present case should accept that the three conditions proposed by the Advocate General must be met in expulsion cases like this one for the burden of proof to shift. It should hold that this entails a reversal of the burden of proof for Frontex when the case concerns its own actions or inactions.

Contrary to the Advocate General’s assertion in his Conclusion, the state of the proceedings before the General Court permits the Court of Justice to assess that the appellant adduced prima facie evidence (FIRST CONDITION) which does allow for the burden of proof to shift, as Frontex is in a better and more privileged position than the applicant to prove or disprove its involvement in, and knowledge of, the alleged events (see Footnote 96 of the Opinion). This is not a question for the General Court to determine in proceedings following a referral back to it (contrary to what the Advocate General proposed in his Conclusion of his Opinion).

It is well established that Frontex possesses evidence directly linked to its core (and exclusive) responsibilities, including the gathering of surveillance data, the production of incident and operational reports, and the monitoring of compliance with fundamental rights as per Article 7(4) and 80(3) and 80(4) of the Frontex Regulation. This places the agency in a similarly ‘privileged’ evidentiary position as Member States are in relation to their own responsibilities in this case.

The Court of Justice should therefore set aside the order under appeal, hold that the three conditions for shifting the burden of proof as outlined by the Advocate General are met in Hamoudi’s case as it would have been the case if the respondent had been a Member State (see footnote 96 of the Advocate General’s Opinion). It should refer the case back to the General Court to reassess the matter, taking into account that the burden of proof has shifted to Frontex.

Between pragmatic and legal considerations: comment on the Advocate General's opinion in joined cases C‑758/24 [Alace] and C‑759/24 [Canpelli]

 

 

 

 

Matteo Zamboni (human rights lawyer working between Italy and the UK. He is a partner to the Immigration Law and Policy clinic of Goldsmiths, University of London)

 

Photo credit: Matteo Zamboni - The Grand Chamber of the CJEU in Luxembourg before the hearing of 25 February 2025

  

Introduction

 

On 10 April 2025, Jean Richard de la Tour, Advocate General (AG) at the Court of Justice of the European Union (CJEU), delivered his conclusions in joined cases C‑758/24 [Alace] and C‑759/24 [Canpelli], regarding the implementation of the Italy-Albania Protocol on asylum and return procedures, which was the subject of my previous blog.

 

In a nutshell, as stated in the press release from the Court, the AG argued that (i) ‘a Member State may designate safe countries of origin by a legislative act’; (ii) a Member State ‘must disclose, for the purpose of judicial review, the sources of information upon which that designation is based’; (iii) ‘[a] Member State may […] under certain conditions, grant a third country the status of safe country of origin, while identifying limited categories of persons likely to be at risk of persecution or serious harm in that country’.

 

This post tries to dig a little bit deeper in the opinion in order to highlight its nuances, strengths, and criticalities.

 

Context

 

Some background information helps put the opinion into context.

 

The opinion was published on the same day that the Italian military vessel Libra (already infamous for not taking action when called to the rescue of the victims of the massive shipwreck of 11 October 2013) was sailing toward the centres in Albania to bring 40 individuals whose claims for international protection had been rejected by the Italian authorities. According to the information available, these people ‘will be held in [the] Italian-run detention centres until they are repatriated to their home countries’. In fact, it was reported that, ‘facing criticism from the opposition over the legal confusion, Prime Minister Giorgia Meloni's conservative coalition […] decided to use [one of the two detention centres built in Albania, the one in Gjader] as a staging post for people whose asylum bids have already been turned down’. More precisely, by means of decree-law no 37 of 28 March 2025 the Italian government established that one of the two centres built in Albania  will temporarily be turned into a repatriation hub – in Italian, ‘Centri di permanenza per i rimpatri’, abbreviated CPR; that is to say, centres to detain undocumented migrants whose claims to international protection had been rejected with a final decision and who are awaiting deportation to their countries of origin.

 

Interestingly, the amendment enacted by the Italian government mirrors the proposal of the EU Commission for a Common European System for Returns, which, amid criticism from human rights organisations, ‘endorsed the idea of “return hubs” located outside the European Union’. Yet, according to sources in the press, the plan is only temporary. In fact, ‘the government still hopes to revert to its original plan and is waiting on a ruling from the European Union's Court of Justice which could compel Italian judges to process new asylum seekers dispatched to Albania’.

 

Moreover, on 16 April 2025 the EU Commission published a proposal for amendments to regulation 2024/1348 (the asylum procedures Regulation, forming part of the asylum pact) touching precisely on the matter of the designation of safe countries of origin (SCOs). Notably, the Commission highlighted that ‘Regulation 2024/1348 […] for the first time provides for the possibility to designate safe countries of origin at Union level’ and informed that ‘the EU Agency for Asylum (EUAA) had been asked to accelerate its analysis […] with a view to drawing up an EU list [of SOCs]’.

 

The proposal is of interest for the case under review for at least three main reasons: 

 

(i)                from a general perspective, the European Commission made it clear that the proposed amendments have an eminently practical purpose: ‘[to] help Member States manage asylum applications more efficiently’ and with ‘greater flexibility’;   

(ii)              the Commission took note of the fact that Article 61(2) of regulation 2024/1348 ‘allows for designation of […] safe countries of origin with exceptions’, including ‘by excluding specific regions or clearly identifiable categories of individuals’, and proposed to ‘bring[…] forward [the] application [of these exceptions]’; the Alace and Canpelli case is partly about the designation of categories of asylum seekers who may be at risk even though their country of origin is designated as ‘safe’;

(iii)           the provisional list of SCOs proposed to be designated at EU level includes Bangladesh; i.e., the country of origin of Alace and Canpelli, the fictitious names of the asylum seekers concerned by the present case – although the Commission does not propose any exceptions for categories of people in that country.

 

Summary of the AG opinion

 

The context briefly discussed above, to which it may be added the fact that pictures of the 40 asylum seekers handcuffed at the time of disembarkation in Albania made the front pages on Italian newspapers, could explain, at least in part, the reason why the AG opinion is full of reference to non-legal considerations.

 

Indeed, the opinion cites, for 4 times in total, the ‘high migratory pressure’ against some Member States as one of the factors to take into due consideration in the effort to advance a balanced interpretation of the relevant provisions of directive 2013/32 (the current asylum procedures Directive, which applies until the 2024 Regulation takes effect).

 

From a legal point of view, the balancing exercise transpires from the juxtaposition of two concepts: one the one hand, that of the margin of discretion (and/or appreciation) enjoyed by Member States when giving effect to the directive at hand, cited 10 times in total; on the other hand, the doctrines of “effet utile” and of the primacy of EU law, referred to in the opinion 6 times in total.

 

Generally speaking, the careful balancing of these competing legal concepts, coupled by practical considerations regarding the State response to mass migration, resulted in the nuanced replies given by the AG.

 

The first question

 

To begin with, the AG noted that the first question, asking whether EU law precludes a national legislature from designating a third country as a SCO by a legislative act of primary law, ‘does not raise, in itself, any particular problem’. The AG noted that directive 2013/32 does neither specify which national authorities are competent for such designation, nor requires that the designation is made with a specific instrument. To the contrary, the expression ‘legislation’, contained in Article 37(1) of the directive, ‘must be understood in its broadest sense, as including acts of a legislative, regulatory or administrative nature’. As a consequence, the AG acknowledged that, under the principle of ‘institutional and procedural autonomy’, Member States enjoy ‘a wide margin of discretion’ as regards the means and the procedures to be used to proceed to the designation of SCOs. Such discretion encompasses the possibility to proceed to the designation by means of primary legislation; i.e., acts of Parliament. At the same time, the AG clarified that ‘the act by which a Member State designates third countries as [SCOs] must not affect […] the basic principles and fundamental guarantees set out in […] directive 2013/32, and in particular […] the right to an effective judicial remedy’ (see AG opinion § 35-39, all translations from the original in Italian and French into English are mine).

 

Thus, the AG replied to the first question stating that:

 

‘Articles 36 and 37 of directive 2013/32 must be interpreted as not precluding Member States from designating SCOs by means of legislative acts’. However, this is possible only insofar as (i) the primacy of EU law is guaranteed; and (ii) the obligations and the objectives of the directive are fully implemented’ (see § 39).

 

The second and the third question

 

A similar tension between the doctrine of the margin of discretion and the overarching objective to ensure the full and consistent application of the acquis communautaire throughout the European Union legal space permeates the answer given to the second and the third question, regarding the need to make publicly available the sources used to justify the designation of a country as a SCO (second question) and the possibility, for the national judge, to assess the designation by making use of information drawn independently from the sources referred to in Article 37 and Annex I of directive 2013/32 (third question). (The Annex defines the criteria to designate a country as a ‘safe country of origin’)

 

At the outset, with a clarification that was much needed after some confusion in the arguments made by the parties (and, above all, by the Italian government and the intervening Member States) during the hearing of 25 February, the AG stated that ‘these questions do not concern the review to be carried out by [the] judicial authority with regard to Article 36(1) of directive 2013/32, which requires the rebuttal of the presumption of safety of a given country [designated as a SCO] in a particular case where, as a result of a specific and detailed assessment […], it appears that that country is not safe due to the individual circumstances of an applicant’ but rather the different issue of challenging the general designation of a country as a SCO (see § 44).

That said, the answers of the AG follow the reasoning applied with regard to the first question. Notably, the AG reiterated that the designation of SCOs by means of primary legislation cannot be construed as excluding the possibility to subject the designation to proper judicial review, as required by Article 46 of the directive. Indeed, when they proceed to the designation of SCOs under Article 37 of the directive, Member States are implementing EU law. Thus, quite regardless from the means employed, they must ‘ensure the respect of the substantive and procedural guarantees established […] by EU law’. It is, therefore, ‘essential’ that ‘sufficient and adequate publicity’ is given to the sources used by the national authorities to proceed to the designation of a county as a SCO (see § 48-49).

 

The AG acknowledged that the publication of the sources is not expressly required by the letter of directive 2013/32. However, he argued that such obligation can be inferred from a ‘systemic reading’ of the directive that takes into due account its overall ‘objectives’. Indeed, ‘the rebuttable nature of the presumption that a country is safe’ implies that  the asylum seekers concerned as well as the national (judicial or administrative) authorities are allowed to know the grounds relied on by Member States to proceed to the designation of a country as a SCO with a view to guaranteeing their right to, respectively, challenge and assess the legitimacy of such designation (see § 50-51 and 54-55).

 

The conclusion is confirmed in the light of the general principle of the duty of cooperation (set out, with specific regard to refugee law, in Article 4(1) directive 2011/95) and the right to an effective remedy (guaranteed by Article 47 of the Charter), which impose that asylum seekers and national courts are given access to all relevant materials, including those allowing them to assess the possible ‘violation of the substantive conditions of [the] designation [of a country as a SCO]’ (see § 56-59).

 

Yet, such a strong petition of principle was not translated into an equally forceful practical answer. As a matter of fact, in answering the third question, the AG gave different options to Member States, arguing that they should include the sources as annexes to the legislative act adopted to designate SCOs, or, alternatively, communicate them at the request of the asylum seeker concerned and/or of the national administrative or judicial authorities. Moreover, in answering the fourth question, the AG held that ‘if the sources of information are not disclosed, then the competent judicial authority may review the legality of [the] designation [of a country as a SCO] in the light of the conditions set out in Annex I to the directive on the basis of the sources that were gathered by the judicial authorities among those listed in Article 37(3) of the directive’ (see § 62, 64, 65).

 

Admittedly, these answers may be said to fail to curb all the uncertainties, as they leave several options open to Member States who are unwilling to publish the sources on the basis of which they proceeded to the designation of SCOs. Moreover, the answer to the fourth question (and in particular the opening caveat ‘if the sources of information are not disclosed’) is capable to undermine the answer to the third question (which states that, anyway, the sources must be published).

 

Be this as it may, this does not seem problematic in the specific case of Alace and Canpelli, as the Italian government had already back-trailed on their position and, with a decision of 28 March 2025, had anticipated the judgment of the CJEU by providing for the publication of the country reports relied on for the purpose of the designation of SCOs.

 

The fourth question

 

The last question, regarding the possibility to designate a country as a SCO notwithstanding the existence of categories of persons for whom it does not meet the substantive conditions for such a designation, was rightly identified as the most complicated one, and has indeed prompted the AG to draft a very nuanced reply.

 

Namely, the AG came up with two possible solutions. The first, based on a ‘restrictive’ interpretation of Annex I to directive 2013/32, relied on the assumption that, to be considered as such under EU law, a safe country must guarantee ‘all nationals and stateless persons living there sufficient protection against the risk of persecution or serious harm, regardless of the portion of the territory in which they find themselves and [independent from] their race, nationality, political [opinions] or religious beliefs, or […] their belonging to a specific social group’. On the other hand, the second solution is meant to ‘allow [Member States] to designate a third country as a SCO even though one or more categories […] of individuals at risk […] have been identified’ (see § 68 and 70).

 

Between the two, the AG concluded that the second option would seem the correct one.

 

The reasons adduced to discard the first option seem more practical than legal. Indeed, the AG borrowed the expression used by the Italian government in their oral arguments and stated that such a restrictive interpretation of the concept of SCO under EU law would be ‘idealistic’ and would undermine the practical effect of directive 2013/32 insofar as it would prompt ‘Member States to treat all the requests advanced by citizens [coming] from those countries under the ordinary procedure, even though the overwhelming majority of those asylum seekers do not have any real need of international protection’. According to the AG, this would result in a ‘procedural congestion’ all the more intolerable in ‘a context characterised by a strong migratory pressure’ (see § 70).

 

In essence, this seems to be the reason prompting the AG to favour the second solution, albeit, as the AG himself did not fail to point out, this may be seen as being at odds with the findings of the CJEU in case C-406/22 (last year’s judgment in which the CJEU interpreted the currently applicable Directive to mean that Member States could not designate part of a country of origin as ‘safe’).

 

To overcome the contradiction, the opinion argued that the second option is grounded in (i) the letter of Annex I to directive 2013/32; (ii) the systematic reading of the relevant EU legislation; and (iii) its purpose.

 

Before addressing this three-pronged argument, however, the AG resorted, once again, to ‘pragmatic’ considerations, stating that ‘in the event that the asylum system of a Member State is under strong migratory pressure and [faces] a high proportion of manifestly ill-founded applications lodged by nationals from [SCOs]’, the second option ‘would [represent] a balanced solution which would make it possible, on the one hand, to […] expedite the examination of those applications [and, on the other hand,] to ensure [that] all applications [receive] appropriate treatment in accordance with the provisions of directive 2013/32’ (see § 71-72).

 

Having said that, the opinion focusses on the letter of Annex I, and specifically on the interpretation of the adverb ‘generally’ (included in the formula ‘a country is considered as a safe country of origin where […] it can be shown that there is generally and consistently no persecution […], no torture or inhuman or degrading treatment or punishment and no threat by reason of indiscriminate violence in situations of international or internal armed conflict’). According to the AG, the adverb ‘generally’ refers to ‘an event, a fact or any other circumstance which manifest itself in the majority of cases or […] which is applicable to a really wide number of persons, without taking into account particular cases’. It, then, follows that, ‘from a literal point of view, a third country may be designated as a SCO if it is demonstrated […] that it protects, not each of its nationals but, the majority of them’. In this scenario, as it had been argued by the Italian government, the second option would also be confirmed by the text of Whereas 42 of directive 2013/32 (see § 78-79).

 

This argument is reinforced by a systematic argument regarding the dichotomy between general and specific considerations underpinning directive 2013/32. In the opinion of the AG,

 

‘if [EU law] requires the competent national authorities to depart (ex post) from the presumption of the safety of a country whenever they determine, as a result of an individual examination […], that the person concerned may, by reason of his individual circumstances, be exposed to a risk of persecution or serious harm in his or her country of origin, then [there is] no valid reason why a Member State should not decide, as a result of the general assessment of that country, to exclude (ex ante) from the scope of that presumption the category or categories of persons whom it has […] identified as being at risk’ (see § 81).

 

Finally, the second solution is said to fit the purpose of the directive, which is to allow Member States to ‘speed up the procedure […] every time that a request for international protection may be ill-founded’. From a wider perspective, this is also confirmed by the consideration of the ‘margin of appreciation’ enjoyed by EU Member States when enacting the directive. Moreover, in reaching this conclusion the opinion also relies on the new provisions of regulation 2024/1348, which, at Article 61(2), explicitly allows for the possibility to designate a country as a SCO even though there exist categories of individuals at-risk. As a matter of fact, and even though the new regulation will become applicable only as of 12 June 2026 (unless the Commission’s recent proposal to bring forward parts of it is adopted beforehand), the AG noted that it would be ‘paradoxical’ to prevent Member States from using this possibility at a moment in which they are called to ‘adequately prepare to implement’ the new regulation, including Article 61(2) (see § 83, 85, 94). 

 

At the same time, the AG seemed to be conscious of possible abuses, and indeed strived to place clear limitations to the implementation of the second option by requiring that these categories are ‘limited’ and ‘clearly identifiable’. In general terms, the need for a qualified application of the personal exceptions to the designation of SCOs is predicated on the assumption that the margin of appreciation, or discretion, of Member States is limited by EU law and by the principle of proportionality. As a result, the use of discretion can never ‘impair the general objectives of directive 2013/32’ (see § 85-87 and 93). 

 

As a consequence, Member States must ‘confine [such] personal exceptions to a very limited number of persons’. Otherwise, the very operation of the presumption of safety would be questionable. In other words, if a given country is designated as a SCO notwithstanding the identification of numerous categories of people that may be exposed to the real risk of persecution of serious harm (as for example, all members of the LGBTQIA+ community), then the concept of safe country of origin would be tantamount to a ‘legal fiction’ (see § 70 and 91-92).

 

Based on this (quite balanced and nuanced) reasoning, the opinion concluded that

 

‘Articles 36 and 37(1) of, and Annex I to, Directive 2013/32 must be interpreted as not precluding a Member State from designating a third country as a safe country of origin for the purposes of examining applications for international protection, while at the same time identifying limited categories of persons as potentially exposed to a risk of persecution or serious harm in that country, provided that on the one hand, that the legal and political situation in that country characterises a democratic regime under which the population enjoys, in general, durable protection against that risk and, on the other, that Member State proceeds accordingly to expressly exclude those categories of persons from the application of the concept of safe country of origin and the presumption of safety attached to it’ (see § 95).

 

Final considerations

 

Few conclusive remarks further highlight the very nuanced nature of the AG opinion.

 

This is particularly evident as regards the first three questions, the answers to which really give the impression of a balancing exercise between the position of the Italian government (and the intervening Member States) and that of the defence. Indeed, the AG validated the practice to designate SCOs by means of primary legislation while at the same time making it very clear that this practice cannot be invoked in order to undermine the guarantees set out at EU level.

 

All in all, the AG found a workable compromise between the legal concepts of the margin of discretion/appreciation and the fundamental principle of the consistent application of the acquis communautaire and its supremacy over national law.

 

On the facts of the case, this approach resulted in a very careful stance toward what seems to be the crux of the questions. Indeed, the AG does not mention the possibility to file a challenge of constitutionality with the Italian constitutional court, which was forcefully put forward by the Italian government as the effective remedy provided by the domestic legal system to question the designation of a country as a SCO. In so doing, the opinion confirmed that the judicial review over the legislation enacting the provisions of directive 2013/32 must be performed by the judge of the case, including first-instance judges. At the same time, the AG opinion does not go as far as stating that the national judge is empowered to disapply national legislation in case of non-compliance with the provisions of the directive. Rather, the multiple references to the doctrine of “effet utile” seem to favour the interpretation that, as I had already argued, the case at hand concerns provisions of EU law lacking direct effect.

 

Still, the answers to the first three questions seem to find a reasonable balance between legal and non-legal considerations. Unfortunately, it seems that one could argue that the matter is different with regard to the answer to the fourth question, as it appears that the option to favour the solution allowing for personal exceptions to the designation of SCOs is grounded on ‘pragmatic’ considerations more than on sound legal reasoning.

 

In particular, this seems to be the case as regards the literal argument employed by the AG, which focusses on the term ‘generally’, but completely overlooks the other adverb in the hendiadys; that is, in the English version of the directive, ‘consistently’. According to the English dictionary, consistently means ‘in every case or on every occasion; invariably’. Therefore, it seems that the literal meaning of this term would rather support a more restrictive interpretation of Annex I; i.e., the interpretation excluding the possibility to designate as SCO a country which is not safe for specific categories of persons.

 

True, the opinion justified this approach by making reference to some inconsistencies in the different language versions of Annex I to directive 2013/32. However, the language versions that I have examined (i.e., the English version, that says ‘generally and consistently’; the French version, which states ‘d’une manière générale et uniformément’; and the Italian version, in which these terms are translated as ‘generalmente e costantemente’) seem to have the same meaning.

 

Also the reference to recital 42 in the preamble to directive 2013/32 may be said to cause raised eyebrows. In fact, when taking into due account the distinction (highlighted by the AG at § 44 of his opinion) between a (more specific) challenge to the application of the concept of SCO to a given case, and a (more general) challenge to the designation in itself, it seems that recital 42 refers to the latter question, while the matter addressed in the fourth question pertains to the second.

 

Finally, the argument regarding the future application of Article 61(2) of regulation 2024/1348 appears to be contradicted by the latest proposal from the EU Commission. Indeed, when the Commission proposes to ‘bring forward’ the application of the personal and territorial exceptions provided in the new legislation, it seems to imply that the EU legislative framework as it currently stands does not allow for such exceptions.

 

Hopefully, these aspects will be clarified by the CJEU in the judgment, which is expected before the summer. 

Budapest Pride: banned? - Banning Pride as a violation of EU free movement

 

Attila Szabó, LLM in European law, Head of Legal Aid Service, Hungarian Civil Liberties Union

The author would like to thank Steve Peers and Máté Szabó, HCLU's professional director, for their ideas for the text!

 

Photo credit: Tételes istentagadó, via Wikimedia Commons

 

On 5 April, Renáta Uitz published a pinpoint blog post on Verfassungblog about how and why Budapest Pride, which has been held every year for 30 years in Hungary for the equality of LGBTQ+ people, was banned. Uitz describes in detail why this is contrary to Article 2 of the TEU and how this government action violates the European Court of Human Rights case law. The CEU (and Royal Holloway) professor also explains how Hungary got to this point.

 

I would like to add two things to this reflection. One is just brief: Hungarian civil society organisations are committed to organising and hosting Pride and the Mayor of Budapest also supported the organizers, despite the categorical legal ban, on the very same fundamental rights basis that Uitz explains: assembly and thus standing up for LGBTQ+ rights is a fundamental right and can only be restricted if it restricts another fundamental right. The abstract and unscientific paedophile arguments that the Hungarian government is operating with are not. Consequently, the restriction of rights is also contrary to the Hungarian Fundamental Law and goes against the ECtHR case law cited by Uitz.

 

It is another matter that the Hungarian legislator allowed for fines to be imposed on the basis of the face recognition system at banned rallies, such as the Budapest Pride, which was planned to take place despite the ban. The legal issues involved would require a separate blog post. A very careful Hungarian-language piece concludes that "the use of facial recognition technology to detect and punish offences, and thus to restrict assembly and free expression, may be objectionable on a number of points. Although (...) the prohibition rules (Article 5 of the EU AI Regulation) do not directly preclude the non-real-time use of facial recognition technology, there are nevertheless a number of fundamental rights concerns which render the use of the system unlawful. The primary avenue of redress may therefore be to invoke these fundamental rights violations."

 

My argument, which may be news to many, is that the restriction runs counter to one of the EU's four freedoms, the right of free movement and residence. This is what I want to convince readers of.

 

Freedom of movement: limited?

 

In my view, if a non-Hungarian EU citizen can enjoy the right to participate in an LGBTQ+ rights march in every EU country then she can not freely choose this given country.

 

Participation in Pride is a fundamental right that has become part of Atlantic and therefore European culture. All EU citizens have the right to express their support for LGBTQ+ people in any EU Member State within the framework of Pride. If an EU citizen is not allowed to participate in Budapest Pride because it cannot take place, he or she is not free to choose Hungary as a place of residence, as this means that he or she cannot freely exercise in that Member State the fundamental rights that he or she could exercise in another Member State. So, an EU citizen who is committed to LGBTQ+ rights cannot come to Hungary to work or study, because he cannot stand up for the rights he could stand up for anywhere else in any city.

 

One might think that restrictions on freedom of movement are only violated if hard barriers are put in place, but in my view this is not the case. If a Member State introduces an ideological, ideological, theoretical or even religious restriction that alienates other EU citizens from exercising their right to move and reside freely, then that right is infringed. This is particularly so in cases where the exercise of the right in question is consensual in the EU or protected by human rights court decisions and other international law standards. Participation in Pride is such a case.

 

In my view, this conclusion is true in all cases where an EU citizen wishes to stay in Hungary on a long-term basis, but is discouraged from doing so by this kind of unlawful restriction of a fundamental right, and also in cases where an EU citizen wishes to come to Budapest specifically for Pride. The latter case is not unprecedented either, as Budapest Pride in recent years has hosted many EU citizens who would not otherwise have stayed in Hungary for a longer period. So banning Pride is not only an obstacle to the free movement of those who would like to attend such an event but cannot, but also to those who attend a formally illegal event and face sanctions for doing so.

 

Based on Art 6 of Directive 2004/38, free movement also extends to people who make short-term visits, ie to participate in an event or demonstration. The Cowan and Bickel and Franz judgments show that free movement rights in the context of short-term visits go further than entry onto the territory – ie. equal treatment as regards victims’ rights and fair trial rights respectively.

 

Legal bases and arguments

 

The right to freedom of movement is a cornerstone of EU law, enshrined in Article 21 of the Treaty on the Functioning of the European Union (TFEU). Furthermore, the EU Charter of Fundamental Rights reinforces this right, particularly Article 45, which explicitly states that every EU citizen has the right to move and reside freely within the territory of the Member States. Although Article 51 of the Charter says that the Charter itself should only be taken into account in the application of EU law, I am not arguing that the Charter should be taken into account in relation to the right of assembly in the Member States. I say that the Charter should be taken into account in relation to Article 21 TFEU. Indeed, the restriction on the right of assembly restricts this EU right, guaranteed in TFEU Article 21, as follows.

 

The Court of Justice of the European Union (CJEU) plays a vital role in shaping and expanding the interpretation of EU freedom of movement, particularly in light of the EU Charter of Fundamental Rights. The CJEU increasingly considers the Charter when assessing national measures that may restrict freedom of movement. This means that restrictions must not only be justified under EU law but also comply with fundamental rights.

 

Cases involving the rights of LGBTQ+ individuals and their families have highlighted the CJEU's willingness to use the Charter to protect fundamental rights within the context of freedom of movement. The Charter has strengthened the protection of freedom of movement by providing a clear and comprehensive list of fundamental rights. It has also given the CJEU a stronger basis for challenging national measures that are deemed to be incompatible with these rights.

 

The Coman and Hamilton (C-673/16) judgment helps us to understand the context. This case dealt with the refusal of a Romanian authority to recognize the marriage of a Romanian citizen with a US citizen, a same-sex couple, celebrated in Belgium. The CJEU ruled that the term "spouse" in the context of EU freedom of movement includes same-sex spouses. The Court emphasized the need to respect the fundamental rights of EU citizens, including the right to respect for private and family life (Article 7 of the Charter) and the principle of non-discrimination (Article 21 of the Charter). This case significantly expanded the rights of same-sex couples in the EU, ensuring that their family life is protected when exercising freedom of movement.

 

Steve Peers wrote about the case that “[t]he Court added that any measure restricting free movement rights also has to comply with human rights guaranteed by the EU Charter of Rights, which has to be interpreted consistently with the European Convention on Human Rights. According to the case law of the European Court of Human Rights, “the relationship of a homosexual couple may fall within the notion of ‘private life’ and that of ‘family life’ in the same way as the relationship of a heterosexual couple in the same situation.”

 

In another case (C-808/21), the CJEU determined that EU citizenship, established by the Maastricht Treaty and articulated in Article 20 TFEU, implies a right for EU citizens to join political parties in other Member States, despite the absence of explicit provisions. The Court reasoned that denying this right undermines the effective exercise of voting rights in municipal and European Parliament elections, which are expressly granted to EU citizens. The CJEU emphasized the principle of equal treatment under Article 22 TFEU, stating that Member States cannot impose conditions on non-national EU citizens that are not applied to their own nationals. While acknowledging national identity and the Member States' authority over national elections, the Court asserted that EU values, such as democracy and equal treatment, limit national discretion. The Court's rulings enhance EU citizenship's political rights, reinforcing its role as a fundamental status within the EU legal order. Finally, the court linked the right to join political parties to Article 12 of the Charter, which concerns freedom of association.

 

Peers wrote in the Verfassungblog that “remarkably, the Court links the implied political rights of EU citizens in municipal and European Parliament elections not only to the express rights of EU citizenship, but also to democracy and equal treatment, as represented in Article 10 TEU and the EU’s values – putting them at the same rank as national identity. So national identity cannot constitute an exception to democracy or the rule of law et al – but must be reconciled with those values.”

 

I think that the EU understandings, based on the ECtHR rulings, is that restrictions on certain political freedoms are not only a violation of EU law if they discriminate against EU citizens, but also if they restrict those rights without discrimination. For, as I argued above, a Member State that falls below the standards of European fundamental rights is in fact restricting the free movement of its citizens by alienating the citizens of other States from the freedom of movement and residence. Who would like to study in Budapest without being able to choose to participate in the celebration of equality according to sexual orientation? I think it can be argued that far fewer people would do so than if this right were not restricted.

 

The EU’s Fundamental Rights Agency has a more detailed collection of relevant cases.  These cases demonstrate the CJEU's commitment to protecting fundamental rights within the context of freedom of movement. The EU Charter of Fundamental Rights provides a vital framework for the CJEU's interpretation of EU law in this area.

 

Obviously, the facts of these cases are different from those of the Budapest Pride ban, but they follow a similar logical structure: that is, national measures become relevant under EU law when, in the exercise of free movement, they deprive an EU citizen of fundamental rights that would be guaranteed in another Member State.

 

Conclusion: no pasarán

 

I argue that banning Budapest Pride is not only a violation of the Hungarian Fundamental Law, not only a violation of the European Convention on Human Rights, not only a violation of Article 2 of the TEU, but also a violation of the right to freedom of movement and residence. There is the same human rights minimum without which an EU Member State cannot be considered so free, so grounded in European values, that a citizen of any other Member State is free to choose to travel and reside there. Member States must not violate the privacy prohibitions on which freedom of movement and residence are based, and they must not, in my opinion, exclude citizens of other Member States from joining political parties by excluding everyone equally. In the same way, Member States cannot exclude EU citizens from lawful assembly by European legal standards. If they do so, they restrict freedom of assembly.

 

Of course, many things can deter an EU citizen from moving to another Member State: from the climate, to the quality of social and health services, to tax policy. However, fundamental rights deterrence is different from other types of deterrence because of the Charter and the relevance of ECtHR jurisprudence. Denial of fundamental rights, on the other hand, carries a normative message: 'You cannot exercise certain fundamental rights here'. And it is no longer a matter of preference for the person, but of the exercise of his or her citizenship status. It is up to the Court of Justice of the European Union to work out the yardstick for this.

The New EU “Common System for Returns” under the Return Regulation: Evidence-Lacking Lawmaking and Human Rights Concerns

 

Dr. Izabella Majcher, Independent Consultant

Photo credit: Alamy Stock Photo 

On 11 March 2025, the European Commission published a proposal for a Return Regulation to replace the 2008 Return Directive. This new proposal also supersedes the Commission 2018 proposal to recast the Directive (discussed here and here). Negotiations on the 2018 recast proposal spanned several years, incurring significant costs, yet they stalled at the European Parliament level before even reaching interinstitutional negotiations between the Parliament and the Council of the EU. This raises questions about the necessity of launching fresh negotiations, their financial implications, and ultimately whether a comprehensive revision is needed at all. Why will this proposal pass through the negotiations if the previous one got stalled? In addition, opting for a regulation rather than a directive could make negotiations even more complex, as Member States recognize that a regulation’s provisions would be directly applicable in their domestic legal systems.

The underlying aim of the proposal is to increase the number of people returned, as it regrets that only around 20% of those issued a return decision actually leave. The proposal links this objective to the effectiveness of return. This objective is to be achieved through two main contributions of the proposal.

First, the proposal aims to streamline and harmonise return procedures, attributing the inefficiency of returns at the EU level to divergent national approaches. To address this, the Commission proposes shifting from a directive to a regulation and restricting the ability of Member States to provide stronger safeguards at the domestic level. Notably, the proposal criticises “significant room” the Directive leaves for national courts to interpret EU rules and removes the option for “more favourable provisions” option under the Directive (Art. 4). In the effort to create “modern, simplified and common” procedures, there is a risk that essential human rights safeguards will be eroded. It is crucial to remember that return procedures must comply not only with the principle of effectiveness but also with human rights standards. Second, the proposal upgrades “common standards and procedures” of return to a “common system for returns.” Besides the return standards and procedures, the “common system for returns” includes mutual recognition of return decisions (see discussion below), necessary resources and personnel to be ensured by the Member States, cooperation between Member States, support from EU bodies and agencies, and digital management systems. These measures are neither groundbreaking nor novel enough to justify an entirely new legislative text—especially considering the costs of adopting a new law.

The lack of an ex-ante impact assessment makes these questions even more pressing. Citing urgency and existing studies and consultations (most of which have not been made public), the Commission chose not to conduct an impact assessment for this proposal. None of these explanations is convincing, especially in light of the importance of evidence-based law-making. Under its own Better Regulation Guidelines, for every initiative that is likely to have significant economic, environmental or social impacts, the Commission should conduct an impact assessment to demonstrate the added value of the proposed measures and their coherence with key criteria—including impact on fundamental rights.

The proposal introduces some stronger human rights protections compared to the Directive, including the judicial form of appeal (Art. 26(1)), expanded monitoring of forced returns (Art. 15), basic needs provision during removal postponement (Art. 14(6)), and age assessment for children (Art. 19). However, these safeguards are overshadowed by the proposal’s overall coercive approach and several provisions that weaken rights. This analysis examines five key measures that are particularly problematic from a human rights perspective, questioning not only their human rights compliance but also their effectiveness.

1) Mutual recognition of return decisions: Building upon the Commission Recommendation from March 2023, the proposal enshrines the mutual recognition of return decisions. Under this mechanism, a Member State that apprehended a person who has already received a return decision in another Member State may (and in the future shall) enforce that decision rather than issuing its own (Art. 9). This is made possible through a “European Return Order” (a form listing main elements of the return decision) made available among the states through the Schengen Information System (Regulation 2018/1860) or other information exchange channels. In theory, mutual recognition enhances efficiency, as only one state would conduct the return procedure. It is therefore unsurprising that the Commission presents it as a key element of the common system for return. However, this measure was already provided in a 2001 Council Directive, yet it does not appear to be regularly used.

Recognizing another state’s return decision comes with practical and legal challenges. Since grounds for legal stay are not harmonised at the EU level, a person may be in an irregular situation in one state but not in another. Although the proposal aims to harmonise procedures, Member States will still be allowed to grant residence permits on humanitarian, compassionate, or other grounds (Art. 7(9)). Since these residence permits are not harmonised, situations may arise where a state is required to remove a person who would otherwise qualify for a residence permit under its domestic law. Additionally, despite formal harmonisation, there will remain the scope for domestic interpretation of grounds for a legal stay (currently demonstrated by discrepancies between asylum recognition rates for the same nationality across the EU). So a state enforcing another’s return decision may face litigation, even though appeals against the decision would have to be lodged against the issuing Member State, according to the proposal. The enforcing state would also bear the costs of removal and potentially detention. Although the proposal provides for Frontex funding or compensation by the issuing state, accessing this financing may be cumbersome to organise. Mutual recognition is thus generally in the interest of the transit countries while disadvantageous for the destination countries.

Crucially, mutual recognition of return decisions and entry bans raises proportionality concerns as it effectively spreads across the EU the most restrictive approaches. As demonstrated above, despite the proposal's aim to harmonise procedures, there will still be room for domestic non-harmonised statuses (e.g., under Art. 7(9) of the proposal) as well as variations in the interpretation of harmonised statuses (e.g., refugee or subsidiary protection statuses). The Commission’s approach lacks coherence on this issue. It glorifies the mutual recognition of return decisions as a key efficiency measure, yet it makes no effort to introduce the mutual recognition of protection statuses—exposing a clear double standard in EU migration policy.  

2) Expansion of detention: the proposal significantly expands the legal basis for detention. Under the Directive, states may only apply detention unless other sufficient but less coercive measures can be applied effectively in a specific case. The proposal removes this requirement (Art. 29(1)-(2)), making detention the first resort measure, which is at odds with the CJEU case-law and HRC jurisprudence. The proposal introduces three new grounds for detention (Art. 29(3)), which are also questionable under international human rights law. Among the new grounds for detention is the need to determine or verify the person’s identity or nationality, even where there is no risk of absconding. This ground could lead to unnecessary and prolonged detention of individuals whose nationality is disputed, not recognized, or who are stateless.

Detention would henceforth also be allowed if the person poses “security risks,” defined as a threat to public policy or public or national security, the existence of serious grounds for believing that the person has committed a serious offence, or a clear indication of their intention to commit one (Art. 16). The proposal further provides for derogations from the detention regime otherwise applicable. States will be able to detain those covered by this ground in prisons (albeit separated from “ordinary prisoners”) rather than in dedicated detention centres, and for a period longer than the maximum permissible length under the proposal—potentially indefinite. Detention based on “security risks” blurs the lines between (administrative) immigration detention and criminal detention. Because it offers fewer guarantees to detainees, immigration detention is an exceptional measure that should only be applied on narrowly defined grounds related to imminent removal. Immigration detention is by no means intended to address security risks; rather, criminal laws should apply equally to anyone under the state's jurisdiction. Criminal pre-trial and post-conviction detention provide stronger due process guarantees and clearer time-limits. The proposal thus intends to sideline the CJEU case-law, as the court was adamant in stressing that detention on public order or safety grounds cannot be based on the Directive.

The proposal extends the maximum permitted length of detention in ordinary cases from 18 months to 24 months (Art. 32(3)). While states are not required to maintain detention for the maximum period, in practice, they often do. As a result, the proposal would lead to longer detention periods across the EU. This stands in stark contrast to the Commission’s previous stance, which claimed that the Directive had a beneficial effect on detention lengths in Member States by generally shortening them. Additionally, the proposal clarifies that these limits apply within a given Member State. So, if a person is transferred between states (for instance, under the mutual recognition mechanism), the detention period may start anew. The proposal also facilitates extended and unnecessary detention by removing the principle that detention shall only be maintained as long as removal arrangements are in progress and executed with due diligence. However, this requirement stems from the ECtHR's well-established case-law, so the Member States remain bound by it in any case.

The expansion of detention thus conflicts with several safeguards flowing from the right to liberty. Additionally, as pointed out by the EPRS, detention is neither necessarily effective (since there is no evidence to suggest that more detention leads to higher return rates) nor efficient (due to the considerable costs involved).

3) Generalized restriction on freedom of movement: the proposal introduces five freedom-restricting measures (including reporting obligations, residing in a specific place and electronic monitoring), labelling them as “alternatives to detention” (Art. 31). However unlike genuine alternatives to detention, these measures are not to be imposed instead of detention (i.e., when detention would otherwise be lawful) but in addition to detention – when detention is not or no longer justified (Art. 32). Labelling such measures as “alternatives to detention” is therefore misleading and creates confusion. Under the framework of detention and its so-called alternatives, individuals risk being either detained or subjected to these restrictive measures depending on the perceived level of the risk of absconding. Given the expansive definition of the risk of absconding (Art. 30), personal freedom would effectively become the exception rather than the rule. Although the proposal includes some human rights safeguards flowing from the right to freedom of movement within a State—such as individual assessment and proportionality requirements—these may not be sufficient to counterbalance the broad scope of restrictions.

Another set of freedom-restricting measures introduced in the proposal relates to the new obligation to cooperate and remain available for the return process (Art. 23). To ensure a swift, efficient, and effective return, the proposal subjects individuals to geographical restrictions (such as confinement to a designated area or residence at a specific address) and/or reporting duties for the duration of the return procedure. Notably, the proposal does not establish any additional conditions for imposing these measures. Subjecting every person in the return procedure to geographical restrictions is an indiscriminate and disproportionate measure. It fails to meet the necessity requirement under the right to freedom of movement, which demands that any restriction be necessary in the individual case for achieving the legitimate objective.

4) The downgrading of “voluntary” departure/return: The proposal rebrands "voluntary departure" (currently used in the Directive) as "voluntary return," which reduces clarity. A return that follows a return decision is not truly voluntary, as it is not based on the person’s informed and free consent. Even "voluntary departure" under the Directive is misleading, with "mandatory" or "accepted" return previously suggested as more accurate. The shift to "voluntary return" only deepens the confusion, likely aligning the language with assisted voluntary return and reintegration programs run by the International Organization for Migration.

Additionally, the proposal removes the priority currently given to "voluntary" departure/return, making removal the default option. The minimum 7-day departure period has also been eliminated, further downgrading the role of “voluntary” departure/return (Art. 12–13). This shift contradicts key legal principles and the Commission’s own position. The CJEU in Zh. and O. confirmed that “voluntary” departure/return is rooted in the principle of proportionality. Furthermore, this form of return is in states’ interests—it is cheaper and easier to organize. As the Commission acknowledged in its 2021 strategy, “voluntary” departure/return helps ensure more effective and sustainable outcomes. The explanatory memorandum applauds the increase in the uptake of “voluntary” departures/returns and, due to the absence of an impact assessment, it remains unclear why this form of return is now being restricted. While the proposal limits “voluntary” departures/returns, it simultaneously promotes them by introducing a new provision dedicated to return and reintegration assistance (Art. 46(3)). This inconsistency ultimately undermines both legal principles and practical considerations.

5) Externalisation of return: The proposal allows Member States to remove individuals—except unaccompanied children and families with children—to a third country with which they have an agreement or arrangement for return (Art. 17). This “return hub” would then be responsible for the further removal. In effect, the proposal establishes a legal basis in EU law for the externalisation of return. On paper, outsourcing returns may seem attractive to Member States. However, as a recent migration deal has shown, such measures often face legal challenges, operational hurdles, and high costs for the outsourcing state. The Commission itself considered “externally-located return centres” in 2018 but identified serious legal and practical challenges, including the risk of violating non-refoulement and conflicts with EU values.

Return hubs raise human rights concerns, particularly regarding detention and onward return. While the proposal requires third countries to uphold international human rights standards, including non-refoulement, it fails to specify how and by whom compliance would be assessed. Similarly, although independent monitoring is mandated, its scope and mandate remain undefined. Member States may not be able to escape responsibility for human rights violations in return hubs. Under the proposed framework, their involvement may be substantial enough to establish jurisdiction, as seen in recent UN Human Rights Committee decisions.

Concluding thoughts

As this analysis has shown, several measures in the Commission’s proposal risk violating fundamental human rights, including the prohibition of refoulement, arbitrary detention, and torture or ill-treatment. The generalised restrictions on freedom of movement and the downgrading of “voluntary” departure/return are not in line with the EU law principle of proportionality. At the same time, these measures are unlikely to improve the effectiveness of return, even when effectiveness is reduced to the sheer number of persons returned—let alone efficiency, which considers the human and financial resources required. In fact, detention does not necessarily lead to higher return rates, while being costly. Similarly, forced removal is more expensive and complex to organize than “voluntary” departure/return. The so-called “innovative solutions”—such as mutual recognition of return decisions and return hubs—lack clarity and may prove far more difficult to implement than envisioned on paper.

Moreover, the proposal fails to meet other key criteria for EU policymaking as outlined in the Commission’s Better Regulation Guidelines. The de-prioritisation of “voluntary” departure/return contradicts the criterion of sustainability. The criterion of coherence is also put into question. Internally, the downgrading of “voluntary” departure/return is inconsistent with the promotion of “voluntary” assisted return and reintegration. Externally, the use of “alternatives to detention” as standalone restrictive measures distorts the concept, which is meant for people who would otherwise be lawfully detained. Ultimately, the proposal disregards several core criteria of EU lawmaking—shortcomings that could have been identified through an ex-ante impact assessment.

The proposal fails to acknowledge that, despite its arsenal of measures to increase return numbers, not everyone in an irregular situation can or will leave EU territory. The obligation to issue a return decision for every person in an irregular situation—without a mandatory prior assessment of refoulement risks or other legal bars to removal (such as family and private life or health conditions)—inevitably creates a group of unreturnable individuals. This issue already existed under the Directive and is further exacerbated by the proposal. Currently, Member States may issue a residence permit for humanitarian or other reasons instead of a return decision, or withdraw/suspend an existing return decision for such reasons (Art. 6(4)). However, as the EPRS observed, these considerations are not automatically assessed within the return procedure.

Rather than addressing this gap, the proposal eliminates this possibility from the list of exceptions. Member States will only be able to withdraw or suspend a return decision (Article 7(9)); abstaining from issuing a return decision on such considerations will not be an option anymore. Impediments to return will be assessed only at the appeal stage of the return procedure, which undermines the effectiveness and efficiency of return policies let alone human rights compliance. Instead, an automatic assessment should be conducted before issuing a return decision, and a regular status should be granted to those who cannot be returned. Regular pathways for stay are widely recognized as a viable alternative to return and a normal feature of migration governance. A genuine "Common System for Returns" should at least recognize regularisation measures for unreturnable persons.

A legislative proposal based on coercion and restrictive measures—without any supporting evidence due to a missing impact assessment—would never see the light of day in the EU if its addressees were not people whose voices are typically unheard. People fall into irregular situations in many different ways and many have lived in and contributed to European societies for decades. They remain invisible, and those who defend their rights face increasing criminalisation. In this context, responding to the demands of some Member States, the Commission has enjoyed wide flexibility to propose unrestricted coercive measures—with little accountability for their impact on those affected. However, law-making without evidence sets a dangerous precedent. In the future, such an untransparent approach could extend beyond undocumented migrants, affecting asylum seekers, migrants in a regular situation, and ultimately EU citizens. Now, it is up to the European Parliament and the Member States in the Council to reaffirm the EU’s founding principles: democracy, the rule of law, and human rights.