The Council’s position on proposed EU law on migrant smuggling: cynical political theatre?

 

Professor Steve Peers, Royal Holloway University of London

Photo credit: Ggia, via Wikimedia Commons

Introduction

Member States’ ministers (the EU Council) are likely to agree a negotiating position this week on replacement of existing EU law on smuggling of migrants – raising questions in particular about the compatibility of this law with human rights, including as regards humanitarian assistance for migrants. The following analysis first sets out the current law, comparing it to the Commission’s proposal for replacing that law last year, then comparing that proposal in turn to the likely text of the Council position, concluding with comments on the latest text. The proposal will still have to be negotiated with the European Parliament, which does not yet have a negotiation position of its own.

Current law

Because it was adopted before the entry into force of the Treaty of Lisbon, the current law is divided between two measures – a Directive setting out the prohibition on facilitating illegal entry or stay, and a Framework Decision providing for criminal offences based on the prohibition. (The proposed law is a Directive that will combine and replace the two).

According to the 2002 Directive, Member States must impose ‘appropriate sanctions’ on anyone who ‘intentionally assists a person who is not a national of a Member State to enter, or transit across, the territory of a Member State in breach of the laws of the State concerned on the entry or transit of aliens’ (illegal entry), and anyone ‘who, for financial gain, intentionally assists a person who is not a national of a Member State to reside within the territory of a Member State in breach of the laws of the State concerned on the residence of aliens’ (illegal stay). Note that the prohibition on facilitating ‘illegal entry’ is not subject to a ‘financial gain’ motive; but the Directive goes on to say that any Member State ‘may decide not to impose sanctions’ in such cases, ‘by applying its national law and practice for cases where the aim of the behaviour is to provide humanitarian assistance to the person concerned’. Member States must also apply sanctions where people are instigators of or accomplices in, or attempt, such actions.

The Framework Decision then specifies that the prohibitions in the Directive must be enforced by ‘effective, proportionate and dissuasive criminal penalties which may entail extradition’, possibly accompanied by ‘confiscation of the means of transport used to commit the offence’, ‘a prohibition on practising directly or through an intermediary the occupational activity in the exercise of which the offence was committed’, or ‘deportation’. In the case of illegal entry for financial gain or instigation, the Framework Decision is more precise about sentencing: there should be a possible maximum penalty of at least eight years where the offence was committed as part of an organised crime group (as defined by a separate EU law), or ‘while endangering the lives of the persons who are the subject of the offence’. But that could be lowered to six years if that was ‘imperative to preserve the coherence of the national penalty system…provided that it is among the most severe maximum sentences available for crimes of comparable gravity.’

After standard provisions on liability of, and sanctions for, legal persons, the Framework Decision requires Member States to establish jurisdiction over an offence committed ‘in whole or part upon its territory’, by its nationals, or for the benefit of a legal person established there. But Member States may waive either of the latter two rules, or apply them only in ‘specific cases or circumstances’. There is also a specific rule on extradition and prosecution (no longer relevant in light of the European Arrest Warrant), and rules on communication between Member States, where one of them is informed about breaches of the law of another Member State, or requests another Member State to prosecute as regards breaches of its own law.

Finally, the Framework Decision includes a safeguard for asylum:

This framework Decision shall apply without prejudice to the protection afforded refugees and asylum seekers in accordance with international law on refugees or other international instruments relating to human rights, in particular Member States' compliance with their international obligations pursuant to Articles 31 and 33 of the 1951 Convention relating to the status of refugees, as amended by the Protocol of New York of 1967.

There is limited case law on the current law, but the Advocate-General’s opinion in the pending case of Kinsa addresses its application to a family member being prosecuted for assisting her child to enter the territory. According to the opinion, this is covered by the obligation to criminalise assistance in the Directive (unless Member States opt out of that aspect), and the Directive is valid; but Member States are obliged to impose lower penalties in such cases.

 

Proposal

The Commission’s proposal starts with a merged criminal offence of assisting illegal entry or stay, either subject to a financial benefit test or a serious harm test:

a) the person who carries out the conduct requests, receives or accepts, directly or indirectly, a financial or material benefit, or a promise thereof, or carries out the conduct in order to obtain such a benefit; or

b) there is a high likelihood of causing serious harm to a person.

The preamble states that these elements of the offence:

…will usually not be fulfilled when it comes to assistance among family members or the provision of humanitarian assistance or the support of basic human needs. Third-country nationals should not become criminally liable for having been the subject to such criminal offences. Moreover, it is not the purpose of this Directive to criminalise, on the one hand, assistance provided to family members and, on the other hand, humanitarian assistance or the support of basic human needs provided to third-country nationals in compliance with legal obligations.

Although the asylum safeguard would be dropped from the main text, a version of it would still appear in the preamble:

Member States should apply this Directive in accordance and in full compliance with the 1951 Convention Relating to the Status of Refugees as amended by the Protocol of New York of 1967, obligations related to access to international protection, in particular the principle of non-refoulement, and fundamental rights.

The proposal also suggests a fleshed-out offence of ‘public instigation’ of illegal entry or stay, although the preamble states that providing ‘objective information’ about immigration and asylum law is not to be covered by this, and that the instigation would amount to ‘illegal content’ covered by the social media content regulation rules in the Digital Services Act.

The offence must be treated as ‘aggravated’ where it was committed within the framework of organised crime, ‘deliberately or by gross negligence caused serious harm to, or endangered the life of, the third-country nationals who were subject to the criminal offence’, was ‘committed by use of serious violence’, where those subject to the offence ‘were particularly vulnerable, including unaccompanied minors’, or where it caused the death of the migrant. Member States must also criminalise incitement, aiding and abetting and attempt – except as regards the ‘public instigation’ offence.

There would be more details on penalties: a maximum penalty possible of at least three years in general for the main offences; ten years (up from eight, or maybe six) for the (longer) list of aggravated cases, also now applying to illegal stay assistance, not only assistance for illegal entry; and fifteen years where a death was caused. (There would be no specific sentencing rule as regards the inchoate offences of incitement, aiding and abetting, or attempt).  

An expanded list of other penalties would include ‘withdrawal of permits or authorisations to pursue activities which have resulted in committing the criminal offence, or prohibition on practising directly or through an intermediary the occupational activity in the exercise of which the criminal offence was committed’ (the latter part reflecting the current law); return to a non-EU country either following sentence in the Member State concerned, or serving some or all of the sentence in the non-EU country, without prejudice to more favourable EU or national rules (building upon the current ‘deportation’); an entry ban for an appropriate period of a maximum ten years, again without prejudice to more favourable EU or national rules; exclusions from public funding etc; fines; and freezing and confiscation of proceeds and instrumentalities (building on the current confiscation of means of transport).

The current provisions on liability of legal persons would be supplemented by more detailed rules on levels of fines, resembling more the penalties in EU competition law (or the Digital Services Act): 3% of worldwide turnover as the general rule, 5% in the case of aggravated offences, 6% where a death is caused. There would also be a list of aggravated circumstances: an offence committed by a public official; a link to illegal employment offences; a prior conviction for the same offence; a link to ‘the exploitation or instrumentalisation of’ the migrant; dispossessing migrants of their ‘identity or travel documents’; or committing the offence while carrying a firearm. Conversely, helping the authorities to ‘find evidence’ or ‘identify or bring to justice other offenders’ would be a mitigating circumstance. But the consequences of either would not be further specified by the Directive.

Like several recent EU criminal law measures, there would be rules on minimum limitation periods: at least seven years after the commission of the offence as a general rule; ten years for aggravated offences; and fifteen years for deaths (including attempts). The same limitation periods would apply as regards enforcement of the sentence, from the date of conviction. As a derogation, these periods could be reduced to five, eight and ten years, ‘provided that the period may be interrupted or suspended in the event of specified acts’.

Jurisdiction would be expanded, to include not only acts committed on the territory in whole or part, but also by nationals and non-EU habitual residents, for the benefit of a legal person also where it does business in the territory, on board its ships or aircraft (registered or flying its flag), or resulting in the entry, transit or stay of the migrants concerned. Unlike the current law, Member States could not opt out of any of these jurisdiction rules. Where an offence was committed outside its territory, Member States would have to drop any requirement that the act was a criminal offence in the other country, or that prosecution required information from the other country first.  

There are some vague new provisions on prevention, resources, training, investigative tools, and data collection, and finally a requirement to apply the new law within one year – shorter than the normal two years for Directives.

 

Council position

First of all, the Council version simplifies the main criminal offence by dropping the provision about causing serious harm as well as the ‘public instigation’ offence. The offence remains assistance to illegal entry or stay, with a financial benefit element. A further safeguard clause has been added to the preamble:

In particular, nothing in this Directive should affect the rights, obligations and responsibilities of Member States and individuals to provide assistance to third-country nationals, for humanitarian reasons or aimed at meeting their basic human needs, under applicable international law - including international human rights law and in particular, where applicable, the 1951 Convention relating to the status of refugees as amended by the Protocol of New York of 1967, the United Nations Convention on the Law of the Sea, and in compliance with legal obligations. This assistance can also include legal, linguistic or social advice or support.

However, the preamble also expressly reserves the possibility for Member States to criminalise support not motivated by financial benefits (new text underlined):

Assistance of unauthorised entry, transit or stay in the Union should constitute a criminal offence at least when there is a link with an actual or a promised financial or material benefit. This is without prejudice to the way in which Member States deal in their national law with facilitation conducts for which an actual or a promised financial or material benefit is not a constituent element of the offence.

A wholly new recital in the preamble makes the point even more clearly:

This Directive is an instrument of minimum harmonisation. As a consequence, Member States are free to adopt or maintain legislation providing for a broader incrimination than what is set out in this Directive, in the interests of enhancing the effectiveness of the fight against migrant smugglers. Member States should ensure that intentionally assisting a third-country national to enter, or transit across, or stay within the territory of any Member State in breach of relevant Union law or the laws of the Member State concerned on the entry, transit and stay of third-country nationals constitutes a criminal offence at least where the person who carries out the conduct requests, receives or accepts, directly or indirectly, a financial or material benefit, or a promise thereof, or carries out the conduct in order to obtain such a benefit. However, since this Directive is an instrument of minimum harmonisation, Member States are free to criminalise such conduct when no financial or another material benefit has been provided.

And the Commission’s proposed recital stating that family members, and migrants subject to the offence, should not themselves be criminalised, is replaced by two recitals giving Member States leeway to do just that:

Nothing in this Directive should be understood as requiring the criminalisation, on the one hand, of assistance provided to close family members and, on the other hand, of humanitarian assistance or the support of basic human needs provided to third-country nationals in compliance with the applicable national and international legal framework.

Without prejudice to rules of national law relating to unauthorised entry in the territory of the Member State concerned, third-country nationals should not become criminally liable for the sole fact of having been the subject to the criminal offence of facilitating the unauthorised entry, transit across, or stay within the territory of any Member State.

In this context, it should be noted that the case law on the Returns Directive, beginning with El Dridi, places limits on Member States enforcing criminal law against migrants for irregular entry or stay by means of custodial sentences – although this is subject to exceptions, and the rationale for it is to speed up expulsion.  

The penalty would follow the Commission proposal for at least a three year possible maximum sentence for the offences. But it would revert to eight years (as at present, although without the option to lower that to six years) for (a longer list of) aggravated offences. It would also revert to an option to apply these sentences to illegal stay cases, and refer to serious endangerment of life – a higher threshold than the current law – as regards one of the grounds of aggravated offences (and the proposed extension to include serious harm would be dropped). The possible maximum sentence would be at least ten (not fifteen) years where a death is caused (subject to the added requirement that the death was caused by seriously endangering the migrant’s life; a new recital in the preamble refers to ‘intention’ being defined by national law, and the avoidance of strict liability). As for additional penalties, the Council would drop the reference to return (thus being less stringent on this point than the current law) as well as the specific reference to a 10-year entry ban (the length of the ban would not be specified).

The penalties for legal persons would be simplified by adding a possibility of a lump sum (€24 million for basic cases, €40 million for aggravated cases); deaths of migrants would be subsumed within the rules on other aggravated cases. A new provision would allow for different penalties than these, if necessary to be consistent with national penalty systems. Aggravating circumstances would no longer include a link to illegal employment, exploitation of migrants (although ‘instrumentalisation’ would now be defined in the preamble), or carrying a firearm, but would now include cases where ‘the criminal offence was committed involving cruel, inhuman, or degrading treatment of a third-country national who was subject to the criminal offence.’

Limitation periods for prosecution would be reduced compared to the proposal, matching the sentencing rules – ie a limitation period of three years where the maximum penalty must be at least three years, eight years for aggravated offences, and ten where death was caused via serious endangerment. A separate set of limitation periods would apply to enforcement after a sentence – three, five, eight or ten years from the conviction, depending on the length of the sentence. Member States could reduce either set of limitation periods to five years (where they were longer than that).

Next, Member States would only be obliged to assert jurisdiction where the offence was committed in whole or part on its territory, or by their nationals, or on board their ships or aircraft. Unlike the proposal, it would be optional to assert jurisdiction as regards non-EU habitual residents, where the act was for the benefit of a legal person established or doing business in the territory, or where the act aimed at the entry, transit or stay of the migrants concerned. Where an offence was committed outside its territory, Member States could retain a requirement that the act was a criminal offence in the other country. Finally, Member States would have two years to give effect to the Directive – probably bringing us to the second half of 2027.

 

Comments

If the European Parliament accepts the Council’s position – and we have no idea what the Parliament (now less liberal than it was before on immigration issues) will argue for – then the main changes from the current law would be:

-          criminalising assistance to illegal entry without financial benefit would no longer be the rule under EU law, with an opt out for Member States, but rather outside the scope of the law, leaving Member States free to criminalise it if they wish;  

-          a three year possible maximum criminal sentence for the main offences, and ten years where a death was caused by seriously endangering a migrant’s life;

-          returns would no longer be explicitly referred to as a penalty, although entry bans now would be;

-          there would be more specific possible penalties for legal persons;

-          there would be new rules on aggravating circumstances, mitigating circumstances, and limitation periods; and

-          jurisdiction would have to be asserted also as regards ships, aircraft and nationals.

The most controversial element here is likely to be exactly how to treat offences not committed for financial gain. (On this point, see also the Belgian Presidency redraft six months ago) Although nominally removing the default obligation under EU law to criminalise all cases of assisting illegal entry, unless Member States choose to opt out, Member States will still be left to criminalise such cases if they wished to, outside the scope of EU law.

In fact, if the Advocate-General’s opinion in the pending Kinsa case is followed, the position of anyone criminalised in this situation will arguably be worse: if EU law no longer applies to them, then even the obligation to reduce the possible penalties applicable to them, recommended by the Advocate-General, will no longer apply. Only a fully-fledged humanitarian defence to criminal liability would suffice. But on that point, it might be argued that since the EU can only set minimum rules on criminal offences (according to the Treaties), EU law is unable to provide fully for defences, because Member States are always free to insist upon greater criminal obligations than EU law provides for.

As for the other changes to the law, the provisions on sentencing, entry bans, limitation periods, special circumstances and jurisdiction may already reflect what many Member States do in practice. It’s hard to be sure because the Commission did not produce an impact assessment on its proposal, although the explanatory memorandum to that proposal did give some examples of potential maximum sentences of less than three years in some Member States. Taken as a whole, this exercise looks like another example of ‘migration theatre’ – an activity intending to signal some effort to control migration flows, but which does not change things much in practice, either as regards criminal liability for those profiting from smuggling, or exemption from liability for humanitarians and desperate family members who do not.

 

‘Safe countries of origin’ in asylum law: the CJEU first interprets the concept

 

Professor Steve Peers, Royal Holloway University of London

Photo credit: viewsridge, via Wikimedia commons

 

Introduction

Along with judgments on Afghan women asylum seekers (discussed here) and the status of Turkey as a ‘safe third country’ (discussed here), the Court of Justice of the European Union recently gave its first judgment interpreting the substance of the concept of ‘safe countries of origin’ in EU law. The judgment addressed a number of important points, interpreting the law in ways that may have broad relevance to EU law on asylum procedures going forward.

EU legal framework

Previous rules

The concept of ‘safe countries of origin’ goes back, at EU-wide level, initially to ‘soft law’ adopted in the early 1990s (one of the ‘London Resolutions’ of 1992). Next, the principle first took on binding legal form at EU level in the first-phase 2005 asylum procedures Directive, which provided for an option for Member States to accelerate considering asylum applications (albeit in accordance with the usual procedural rules), inter alia where the applicant is from a ‘safe country of origin’. The latter concept was then further defined: it applied to countries (as described in Annex II to the Directive) where:  

on the basis of the legal situation, the application of the law within a democratic system and the general political circumstances, it can be shown that there is generally and consistently no persecution as defined in Article 9 of Directive 2004/83/EC [the first-phase qualification Directive], 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.

When ‘making this assessment’, Member States had to take account, ‘inter alia, of the extent to which protection is provided against persecution or mistreatment by’:

	- the relevant laws and regulations of the country and the manner in which they are applied;
	- observance of the rights and freedoms laid down in the European Convention for the Protection of Human Rights and Fundamental Freedoms and/or the International Covenant for Civil and Political Rights and/or the Convention against Torture, in particular the rights from which derogation cannot be made under Article 15(2) of the said European Convention;
	- respect of the non-refoulement principle according to the Geneva Convention;
	- provision for a system of effective remedies against violations of these rights and freedoms.

When applying these tests, Member States had to ‘have regard to the legal situation, the application of the law and the general political circumstances in the third country concerned’, and the ‘assessment of whether a country is a safe country of origin’ had to ‘be based on a range of sources of information, including in particular information from other Member States, the UNHCR, the Council of Europe and other relevant international organisations’.

Furthermore, the Directive expressly provided that Member States could designate ‘part of a country as safe where the conditions in Annex II are fulfilled in relation to that part’. It was also open to Member States to continue in place pre-existing rules that fell short of the standards in the Directive to some extent; these rules could also be applied to part of a country, or to designate a country (or part of it) ‘as safe for a specified group of persons’. There were safeguards for individuals: for the ‘safe country of origin’ concept to apply, they would have to be nationals of the country concerned (or stateless persons who were habitual residents of it), and must not have ‘submitted any serious grounds for considering the country not to be a safe country of origin in his/her particular circumstances and in terms of his/her qualification as a refugee in accordance with’ the qualification Directive. Also, Member States were obliged to ‘lay down in national legislation further rules and modalities for the application of the safe country of origin concept’ – although there was no indication of what those rules and modalities might concern.

Finally, there was a power for the EU to adopt common lists of ‘safe countries of origin’, by means of a special procedure. But this power was soon annulled by the CJEU, when the European Parliament successfully challenged the prospect of using such a special procedure.

Current rules

The current asylum procedures Directive, adopted in 2013, retains the ‘safe countries of origin’ concept as an option for Member States to apply an accelerated procedure to. There is no longer a reference to potentially treating only part of a country as safe, and the option for Member States to retain pre-existing lower standards on this issue (along with pre-existing rules on designating part of a country as ‘safe’, or as ‘safe’ for groups of people) was dropped. The notion of a common EU list was not revived in the Directive; although the Commission subsequently proposed such a list (consisting of Western Balkans states and Turkey) in 2015, that proposal did not receive enough support and was withdrawn in 2019. An express requirement for regular review of the supposedly ‘safe countries of origin’ by Member States has been added. Otherwise the rules on ‘safe country of origin’, including the core definition of the concept, were not amended.

However, the rules on appeals, elaborated compared to the previous Directive, now address the issue too. Although the 2013 Directive provides that in principle asylum-seekers can remain on the territory pending resolution of their appeal, there is an exception for (inter alia) most accelerated procedures, including the ‘safe country of origin’ rule. In such cases Member States, if they opt not to give an appeal suspensive effect, must at least allow a court to rule on whether or not the asylum-seeker can stay pending the appeal; and the asylum-seeker can at least stay until a court has ruled on this point.

Before the recent judgment, the CJEU has only ruled on these provisions once, and on a purely procedural point: according to the Court, Sweden could not apply a ‘safe country of origin’ rule until it legislated for one in national law. Sweden duly did so. In passing, the Court confirmed the usual understanding that the ‘safe country of origin’ rules set out a ‘rebuttable presumption’.  

Future rules

The 2024 asylum procedures Regulation has amended the ‘safe country of origin’ rules again. This entails: explaining the concept of ‘non-refoulement’; confirming that rebuttal of the presumption by an asylum-seeker must take place ‘in the framework of an individual assessment’; again expressly providing for exceptions for ‘specific parts’ of the non-EU country’s territory and (not only in the context of pre-existing law) for ‘clearly identifiable categories of persons’; and again allowing for the adoption of a common EU list (this time not via a special procedure, so not repeating the previous legal problem with the concept of such a list). There is no longer any reference to national rules providing for further ‘rules and modalities’ of the concept.

In the broader context, the ‘safe country of origin’ concept is still on the list of possible accelerated procedures, but the use of those procedures will now be mandatory for Member States. In practice the concept might cross over sometimes with the new ground of accelerated procedures where the country of origin has an international protection recognition rate below 20%. It will be expressly possible (although arguably it will only be optional) to apply this ground for accelerated procedures to unaccompanied minors. The implications of accelerated procedures are more explicitly harmonised: three months (instead of the usual six months) to decide on the application, and between five and ten days (instead of the usual two weeks to a month) to bring an appeal. The exception to the right to remain pending appeal (subject to a court ruling otherwise) in such cases becomes mandatory, with some harmonised rules added: a time limit of at least five days to request a court to remain on the territory, along with interpretation and legal aid. As before, accelerated procedures are linked to the rules on a border procedure, which have also been heavily amended.    

National practice

The EU asylum agency has produced a useful map of which countries have been designated as ‘safe countries of origin’ (and ‘safe third countries’) by Member States. Furthermore the agency reported on the national application of these concepts at the end of 2022. As regards ‘safe country of origin’, the report indicates that a large majority (although not quite all) Member States apply the concept. The exceptions are Poland, Latvia, Spain, Bulgaria, Lithuania, Portugal and Romania; Finland applies the concept, but without a list.

Several Member States apply geographic exceptions, namely for parts of Armenia, Bosnia and Herzegovina, Georgia, Moldova, the USA and India. Some Member States also apply exceptions for groups of people, in particular LGBT applicants, minorities, criminal cases, political activists, journalists, human rights defenders, victims of discrimination or forced marriage, and women and girls. There is some variation in which countries are listed as ‘safe’, with Western Balkans states being listed most frequently. The report details a number of judgments of national courts concerning both the listing of particular countries and their application to individual circumstances, with mixed results – some judgments uphold the listing or its application to an individual asylum-seeker, but some do not. The recent CJEU judgment is, as we shall see, relevant to the geographic exceptions; and a pending case asks the CJEU about the exceptions for groups of people.   

The judgment

The recent judgment concerned the Czech designation of part of Moldova (except the Russian-occupied Transnistria) as ‘safe’. Faced with an appeal by a Moldovan asylum seeker who was subjected to a ‘safe country of origin’ rule, the national court asked three questions: a) could a country be designated a ‘safe country of origin’ even if it had triggered the derogation from (some) ECHR rights set out in Article 15 ECHR (which Moldova had done after the Russian invasion of neighbouring Ukraine)?; b) could only part of a country be designated as ‘safe’; and c) could a national court raise the legality of the designation of its own motion?

On the first point, the Court ruled that invoking Article 15 ECHR did not automatically prevent or cancel the designation of a country as a ‘safe country of origin’. This was because the ECHR set out guarantees for use of the derogation in Article 15, limiting its use to cases where it was necessary, requiring no conflict with other international law, allowing derogations only from certain rights and remaining subject to review by the European Court of Human Rights. (On the relevant case law, see that Court’s guide to Article 15). Also, it could not necessarily be assumed that invoking Article 15 meant that rights were actually derogated from; nor did use of the derogation as such determine the nature and extent of the derogation.

However, the CJEU pointed out that triggering an Article 15 ECHR derogation must mean that Member States had to review the ‘safe country of origin’ listing, on the basis of the obligation to review the designations in light of developments set out in the Directive. That review should be triggered by significant events likely to affect whether a non-EU country still meets the ‘safe country of origin’ criteria, which include the invocation of a derogation from the ECHR – even though the derogation would affect rights other than Article 3 ECHR (the ban on torture or other inhuman or degrading treatment), because Article 15 ECHR does not allow derogation from Article 3.

The Court did not address the arguments in the Advocate-General’s opinion, concerning the protocol on asylum for EU citizens by analogy. That protocol provides that EU Member States are ‘safe countries of origin’ for each other’s citizens, meaning (in the context of the protocol) that asylum applications from EU citizens cannot be considered at all – unless a Member State either: invokes a derogation under Article 15 ECHR; or the Article 7 TEU process of either sanctioning a Member State for a ‘serious and persistent breach’ of EU values, or warning a Member State that there is a ‘clear risk of a serious breach’ of such values, has been applied (or triggered, as regards a warning; on this point, see the Hungary v European Parliament judgment); or a Member State decides to do so unilaterally for an individual applicant. In the Advocate-General’s view, the protocol could not apply by analogy due to its different context: an expression of mutual trust applicable to EU citizens only.

On the second question – whether a non-EU country could be designated as a ‘safe country or origin’ only in part – the Court first looked at the wording of the Directive, which used the word ‘country’ without indicating that it may be referring to only part of the territory. Secondly, the context of the Directive included the definition of ‘safe country of origin’ in the Annex; that definition did not suggest either that only part of a country could be designated, and the words ‘generally and uniformly’ suggested the opposite.

Next, the ‘safe country of origin’ rule was an ‘exception’ to the normal process of considering asylum applications, and such derogations should have a ‘strict interpretation’ (referring to the case law on inadmissible applications), notably where an interpretation had no support from the text. Also, the Court looked at the historical context, noting that: the 2005 Directive expressly provided for parts of countries to be listed; the 2013 Directive does not; and the proposal for what became the 2013 Directive explicitly stated an intention to remove the possibility. Finally, the Court referred to the objectives of the asylum procedures Directive, taking the view that the EU legislature was trying to balance the objectives of a speedy versus a thorough assessment of asylum applications. This was a discretionary choice that was reversed by the 2024 Regulation; a change of mind was the prerogative of the EU legislature, provided that it complied with the Refugee Convention and the Charter. And the wording of the 2024 Regulation reinforced the point that designation of only part of a country was not provided for in the 2013 Directive.  

Again the Court’s judgment does not discuss a point analysed by the Advocate-General – as to whether the ‘effective territorial control’ principle, which limits the liability of countries (including Moldova, in this context) for breaches of the ECHR on their territory controlled by another State, applies in the context of EU asylum law to allow for the designation of only part of their territory as ‘safe’. In the Advocate-General’s view, the issue of state liability for breaches of international law is distinct from the issue of the listing of ‘safe country of origin’ under EU asylum law; and the EU legislature decided to provide for listing of whole countries only, even being aware of the ‘effective territorial control’ issue.

On the third point – consideration of the designation of ‘safe country of origin’ by a court on appeal, of its own motion – the CJEU ruled that, taking account of its prior case law giving a strong role to courts reviewing asylum decisions, including on procedural issues, such designations have to be reviewed by courts of their own motion even if the asylum applicant does not raise the point. In particular, the courts have to review the invocation of derogations from the ECHR and the territorial scope of the ‘safe country of origin’ designation. Again, the Court did not address an issue raised by the Advocate-General – the analysis, in the alternative, of how the EU law principle of equal and effective remedies would apply in a case like this one.

Comments

The Court’s judgment raises a number of points that could be relevant to EU asylum procedures law more generally. On the first point (the ECHR derogation), the ruling at first sight may only apply to ECHR parties; but logically it applies by analogy also to any derogations from other human rights treaties (including the ICCPR, which is expressly mentioned in the Directive), and/or from national human rights protection – taking into account any relevant differences from the ECHR (the prospects of effective judicial review of the derogation, and the particular rights being derogated from, which were key features of the Court’s judgment). After all, human rights abuses (triggering asylum claims) often take place during supposed states of emergency. 

The judgment also makes clear that reviews of ‘safety’ must be triggered by significant events – not only, as Member States might prefer, taking place at regular prescheduled intervals. It follows that a failure to conduct a review in light of a significant event is a procedural flaw that can be challenged. Moreover, this judgment should logically apply by analogy to the reviews of the EU common list provided for in the asylum pact Regulation – as well as to national and EU reviews of ‘safe third countries’.  

As for the dog that didn’t bark – the comparison with the protocol on asylum for EU citizens – the better argument than in the Advocate-General’s opinion is surely that in addition to the obvious difference in wording, the context of the Directive is hugely different too. Although the protocol uses the phrase ‘safe countries of origin’, unlike the Directive it does not create a rebuttable presumption, but rather a complete ban on considering applications (in principle) from EU citizens; so it makes sense that the conditions for applying the protocol are more stringent, since the key safeguard of the possibility of the individual right to attempt to rebut the presumption is absent.

On the second point – the inability to label part of a country as ‘safe’ – as we can see from the practice above, this would mean changes in a number of Member States. Of course, this is only short term, because the asylum pact legislation will soon allow such designations again. But the Court’s express reference to the new Regulation may give us a broader indication of how it will interpret the pact. It suggests that while the EU legislature has a discretion to choose between prioritising speedy or thorough consideration of applications, that choice must comply with the Charter and the Refugee Convention. Given the reduction in standards in the Regulation – and the further reductions in standards that the Commission and many Member States seem to be gagging to make as soon as possible – the Court’s assessment of whether EU asylum law indeed complies with the Charter and the Refugee Convention is likely to be of increasing importance in the future.

This part of the judgment also confirms that the concept of ‘safe country of origin’ is an exception which must be interpreted strictly, by analogy with the case law on inadmissibility which often finds that Member States’ interpretations of the grounds for inadmissibility have been too restrictive. This logically applies by analogy to all grounds for accelerated procedures, a fortiori after they become mandatory under the 2024 asylum pact.

Finally, the third point – judicial review of the ‘safe country of origin’ designation, on its own initiative – logically applies by analogy too, to other aspects of the principle besides those at issue in this case, and to listings of ‘safe third countries’ also. Although strengthening judicial review may have limited effect in such cases, given the absence of suspensive effect of appeals, confirming the wider power of courts to examine the validity of designating countries in the list on their own motion may make it easier to convince them to grant such suspensive effect, given the wider array of remedies which an applicant can therefore seek.

Latest Updates on The Legitimate Interest Ground for Processing Personal Data (Article 6(1)(f) of GDPR): the latest CJEU Case and EDPB New Guidelines

 

 

 

Aolan Li*

*The author is a third-year PhD candidate in Law at Queen Mary University of London. Her ongoing doctoral thesis research delves into the application of Article 6(1)(f) of GDPR from a comparative perspective. Email: aolan.li@qmul.ac.uk

Photo credit: TheDigitalArtist, via Wikipedia Commons

 

A positive spirit has spread among business-side stakeholders across the EU since the Court of Justice of the European Union (CJEU) published its preliminary ruling in the Koninklijke Nederlandse Lawn Tennisbond case (C-621/22) on 4 October 2024, where the court confirms a purely commercial interest could constitute a legitimate interest for processing personal data under Article 6(1)(f) of GDPR. Commentators go as far as to say - “what this means is that under the GDPR, your data can be used without your consent solely for a company’s commercial interests.”

The preceding saying is a total misunderstanding. The positive spirit should have been dampened as, on 9 October 2024, the European Data Protection Board (EDPB) published its new guidelines on Article 6(1)(f) for public consultation (hereafter as the new EDPB guidelines).

Bearing in mind the optimistic bubbles in the market, this writing articulates the EDPB’s stringent stance on the application of Article 6(1)(f) of GDPR, focusing on what has changed compared to the Article 29 Data Protection Working Party’s opinion on the legitimate interest ground under Directive 95/46/EC (hereafter as the WP29 Opinion).

General remark

The newly published EDPB guidelines align with the WP29 Opinion in some basic stances. First and foremost, the recognition of a legitimate interest is not itself sufficient to rely on Article 6(1)(f) of GDPR as a legal basis (this is why the saying is misleading) as there are three cumulative conditions for its application. Secondly, Article 6(1)(f) of GDPR should not be used “by default” nor as a “last resort”. The open-ended nature of Article 6(1)(f) has a unique role in the EU data protection law.

Not surprisingly, the new EDPB guidelines also substantially update the WP29 Opinion.

The update is partially attributed to judgments of CJEU issued after the adoption of the WP29 Opinion, including Rīgas (Case C-13/16), Fashion ID (C-40/17), TK (C-708/18), MICM (C-597/19), Meta v Bundeskartellamt (C-252/21), SCHUFA Holding (Joined Cases C-26/22 and C-64/22), and the latest Koninklijke Nederlandse Lawn Tennisbond (C-621/22). Many practical examples in the new guidelines mirror scenarios disputed in the abovementioned cases. For example, example 4 is analogous to Rīgas.

Building upon more detailed case law, the new EDPB guidelines are more logical and clearly articulated. Unlike the WP29 Opinion, the new guidelines make effects to draw a clearer line between the six grounds for legitimising data processing under Article 6(1) of GDPR. Also, the new guidelines follow the now well-accepted three-step approach to applying Article 6(1)(f), which was established by the CJEU in its judgment in Rīgas.

The update also corresponds to the evolvement of the law itself (GDPR vs Data Protection Directive). GDPR has strengthened data subject rights. It is worth noting the improvement of the right to object - a specific right for the processing based on Article 6(1)(e) and (f) of GDPR - as the burden of proof has been reversed on the controller. Also, GDPR and CJEU case law have escalated the reasonable expectation of data subjects to a more significant position in determining the application of Article 6(1)(f) of GDPR. Therefore, the new guidelines are observed to enhance the position of data subjects accordingly.

Besides being consistent with legislative developments and the CJEU’s case law, the EDPB is observed to add its unique understanding to narrow down the scope of Article 6(1)(f) of GDPR; here’s why I said the EDPB takes a stringent stance. The next part provides more discussions.

Overall, the new guidelines have been compiled from rich and up-to-date sources and provide much more nuanced interpretations of Article 6(1)(f) of GDPR. However, one might lament that Part IV of the new guidelines hesitates to touch on the application of Article 6(1)(f) of GDPR in more complicated and controversial contexts. For example, its application in the credit scoring industry seems like a real-world need, as demonstrated in SCHUFA Holding. Let alone its silence on applying Article 6(1)(f) of GDPR in AI-related scenarios.

The writing below touches on the substantial content of Article 6(1)(f) of GDPR. However, it does not intend to sketch the 37-page guidelines reductively. Instead, it aims to highlight the stringent stance of the new guidelines, read together with the Koninklijke Nederlandse Lawn Tennisbond case.

The Three Steps Approach

As mentioned above, three cumulative conditions must be fulfilled to rely on Article 6(1)(f) of GDPR as a legal basis, called the three steps approach, which are 1) the pursuit of a legitimate interest by the controller or by a third party; 2) the need to process personal data for the purposes of the legitimate interest(s) pursued; 3) the interests or fundamental freedoms and rights of the concerned data subjects do not take precedence over the legitimate interest(s) of the controller or of a third party (the new EDPB guidelines, p 2).

For the first step, the new guidelines narrow down the scope of interests with respect to the controller’s own interests and disentangle the third party’s interests from wider public interests.

As the information circulated, the qualifier “legitimate” is interpreted broadly, covering any interests that are not contrary to the law (Koninklijke Nederlandse Lawn Tennisbond, para 49).

However, sourced to the CJEU judgment in Meta v. Bundeskartellamt, the new guidelines confine that “as a general rule, the interest pursued by the controller should be related to the actual activities of the controller.” (the new EDPB guidelines, para 19) It means that, within the meaning of Article 6(1)(f) of GDPR, a controller whose activity is economic and commercial in nature is only allowed to pursue economic and commercial-related interests.

Other legitimate but non-economic/commercial interests might fall within the scope of interest(s) pursued by a third party. The new guidelines clarify that the controller needs to demonstrate the legitimate interest(s) are pursued by one or more specific third parties (para 20-25) and should not be confused with broader public interests despite the fact they can overlap, as seen in SCHUFA Holding.

Remarkably, the new EDPB guidelines indicate that relying on the interest(s) pursued by a third party in the first step is generally more challenging to pass the latter two steps (the necessity and balance test) than relying on the controller's own interests. (para 30)

For the second step, the processing involved should be necessary for the purposes of that interest identified in the first step, called the necessity test. The concept of necessity has its own free-standing meaning in EU law. The controller must demonstrate that there are no other reasonable, just as effective, but less intrusive alternatives to achieve the pursued legitimate interests.

Despite no given example in the new EDPB guidelines, the CJEU judgment in Koninklijke Nederlandse Lawn Tennisbond provides a least intrusive scenario in the direct marketing context. To be brief, without asking for consent, a Netherlands sports federation (KNLTB) sold its members’ personal data to its sponsors for the latter’s marketing purposes. The court considers it possible for KNLTB “to inform its members beforehand and to ask them whether they want their data to transmitted to those third parties for advertising or marketing purposes.” (para 51) The court deems a procedure as such may involve the least intrusion of data subjects’ rights and compliance with data minimisation principles. As will be explained below, the proposed approach resonates with the right to object and controllers’ notification obligations.

For the third step, the balance test entails a balancing of the controller side's rights and interests against those of the data subject side. The controller needs to ascertain, on a case-by-case basis, that the processing at issue would not disproportionately impact the data subject’s rights and interests.

One can observe the improved position of data subjects directly from the structures of exercising the balance test in the new EDPB guidelines and the WP29 Opinion. (See table below)

Methodology for the balancing test under new EDPB guidelines	Methodology for the balancing test under the WP29 opinion
The data subjects’ interests, fundamental rights and freedoms.	Assessing the controller’s legitimate interest. -      Exercise of a fundamental right; -          Public interests/the interests of the wider community; -          Other legitimate interests; -          Legal and cultural/societal recognition of the legitimacy of the interests.
The impact of the processing on data subjects, including -          The nature of the data to be processed; -          The context of the processing; -          Any further consequences of the processing.	Impact on the data subjects -          Assessment of impact; -          Nature of the data; -          The way data are being processed; -          Reasonable expectations of the data subject; -          Status of the data controller and data subject.
The reasonable expectations of the data subject.	Provisional balance.
The final balancing of opposing rights and interests, including the possibility of further mitigating measures.	Additional safeguards applied by the controller to prevent any undue impact on the data subjects.

Despite most of the content continuing to work, some remarkable points exist.

Firstly, the reasonable expectation of the data subject has been escalated to an independent element. It goes beyond the controller’s notification obligation and highlights the data subject’s genuine understanding; as the new EDPB guidelines put it, more than the mere fulfilment of Articles 12, 13, and 14 is needed to consider that the data subject can reasonably expect the said processing. (para 53)

Secondly, the mitigating measures, be it technical and organisational, within the meaning of Article 6(1)(f) of GDPR must go beyond existing principles and obligations set out in the GDPR. In this sense, the new EDPB guidelines encourage controllers who intend to rely on Article 6(1)(f) of GDPR to pursue a higher level of personal data protection than legal obligations.

Data subject rights

A comprehensive review of the enhanced data subject rights under the GDPR goes beyond the subject matter of this writing. Calling back to the least intrusive approach proposed in Koninklijke Nederlandse Lawn Tennisbond, this part of the writing articulates the significance of controllers’ notification obligations and data subjects’ right to object in the context of Article 6(1)(f) of GDPR.

The court considers that KNLTB can inform its members beforehand. KNLTB’s notification obligations are set out in Articles 13 and 14 of GDPR. Its members (data subjects) should be informed about, among other things, the legal basis of processing, the specific legitimate interests pursued by KNLTB or its sponsors, and data subject rights. According to Article 13(3), KNLTB should inform its member concerned prior to further processing.

The court also considers it good practice for KNLTB to ask members concerned whether they want their data transmitted to third parties for advertising or marketing purposes. One might feel at odds with the reintroduction of “consent” in assessing Article 6(1)(f) of GDPR. Actually, it is better to understand the “ask” as informing its members concerned about their right to object under Article 21 of GDPR.

The objection to direct marketing based on Article 6(1)(f) of GDPR is absolute. In other cases, however, the controller might have compelling legitimate grounds to disapprove the right. Here, it involves another balancing test to determine whether the controller has a compelling legitimate ground. Unlike Directive 95/46, the burden of proof is on the controller.

The new EDPB guidelines promote the idea that the controller’s compelling legitimate grounds can only be recognised in exceptional cases. The controller cannot circumvent the right to object by merely showing that the processing would be beneficial to the controller. Rather, the concept of compelling is understood as essential to the controller.

From the preceding standpoint, the right to object has been improved in favour of the data subject, and it is not much inferior to the right to withdraw consent.

Concluding remarks

In conclusion, while the CJEU’s preliminary ruling in the Koninklijke Nederlandse Lawn Tennisbond case initially sparked optimism among business stakeholders by holding that purely commercial interests could qualify as legitimate under Article 6(1)(f) of the GDPR, this enthusiasm is misplaced. The ongoing EDPB’s new guidelines underscore a more restrictive interpretation of the legitimate interest ground than the earlier WP29 Opinion, reinforcing the need for careful application and a balanced approach to personal data protection. This writing calls for a self-reassessment of GDPR compliance, in particular for controllers relying on legitimate interest as a main legal basis.

Pyrrhic victory for the Greek government: the CJEU rules on Turkey as a “safe third country”

Professor Steve Peers, Royal Holloway University of London

Photo credit: Mstyslav Chernov, via Wikimedia Commons

Hamster idiom credit: Adam Sharp

(Thanks to Zoe Gardner for discussion of the case; the following analysis is my own)

 

Introduction

Remember the ‘refugee crisis’ of 2015-16? One part of the EU’s response to it was to negotiate a ‘joint statement’ with Turkey (see comments and criticism here, here and here) that, among other things, led to Turkey agreeing to take back asylum-seekers from the Greek islands. Greece facilitated the application of this process by determining that Turkey was a ‘safe third country’ (ie that asylum-seekers could and should have applied there instead, given that this other country is ‘safe’) under EU asylum law, initially for Syrian asylum-seekers in the eastern Greek islands. Faced with a legal challenge to this determination, the Greek Council of State ruled in 2017 that the determination was legal, deciding by a 13-12 vote not to ask the CJEU to interpret the issue of EU law at stake. For its part the EU General Court (upheld on appeal by the CJEU) rejected legal challenges to the joint statement, on the frankly bizarre ground that the European Council had nothing to do with the EU/Turkey joint statement (which referred to EU law issues such as visa waivers).

Subsequently in 2020, Turkey decided to refuse all readmission of asylum-seekers from Greece, stating that (as quoted in note 10 of the Advocate-General’s opinion):

 Responding to repeated requests from the Greek authorities and the European Commission regarding the resumption of return operations, [the Republic of Türkiye] has stated that no return operation would take place unless the alleged pushbacks along the Turkish-Greek border stop and [the Hellenic Republic] revokes its decision to consider [the Republic of Türkiye] a Safe Third Country.

Greece nevertheless decided to list Turkey as a ‘safe third country’ more broadly in 2021, for asylum-seekers from Syria, Afghanistan, Pakistan, Bangladesh or Somalia; it added Albania and North Macedonia to its ‘safe third country’ list soon afterwards. In light of the EU law rules on ‘safe third country’ (discussed further below), two NGOs (the Greek Council for Refugees and Refugee Support in the Aegean) decided to challenge the Greek government decision to list Turkey, resulting this time in the view of the Council of State in February 2023 (by an 18-4 vote) that the listing of Turkey was invalid, on the grounds that EU law precluded designating a country which refused to readmit asylum-seekers as a ‘safe third country’. (Other EU law arguments against the designation of Turkey were rejected by the national court and not referred to the CJEU: see the analysis by Minos Mouzourakis).  But unlike in 2017, there was great deference to the view of the (much smaller) minority (two judges apiece argued that the issue of readmission was relevant instead when ruling applications inadmissible, or when enforcing that decision). So the national court decided to ask the CJEU questions about the interpretation of EU law on the readmission point before giving its final ruling. The CJEU delivered its judgment last Friday.

This is not only the first time that the CJEU has been asked about an aspect of the EU/Turkey asylum relationship. Remarkably, despite the huge number of asylum seekers that have come to Greece over the last decade and more, this is also the first time that the CJEU has been asked any questions about EU asylum law by any Greek court.

 

EU legal framework

Legislation

The definition of ‘safe third country’ for the EU is currently set out in Article 38 of the EU’s asylum procedures Directive. (This is about the ‘safety’ of non-EU countries, ie it is distinct from the EU Member States regarding each other as ‘safe’ countries, under the Dublin rules) First, Article 38(1) provides that the (optional) ‘safe third country’ concept can ‘only’ be applied if Member States are satisfied that certain ‘principles’ are ensured for asylum-seekers in that country: (a) ‘life and liberty are not threatened on account of race, religion, nationality, membership of a particular social group or political opinion’; (b) ‘there is no risk of serious harm as defined in’ the EU qualification Directive; (c) ‘the principle of non-refoulement in accordance with the Geneva [Refugee] Convention is respected’; (d)  ‘prohibition of removal, in violation of the right to freedom from torture and cruel, inhuman or degrading treatment as laid down in international law, is respected’; and (e) ‘the possibility exists to request refugee status and, if found to be a refugee, to receive protection in accordance with the Geneva Convention’.

Secondly, Article 38(2) requires Member States to subject the application of the ‘safe third country’ rule to national law rules, including: (a) ‘requiring a connection between the applicant and the third country concerned on the basis of which it would be reasonable for that person to go to that country’; (b) ‘rules on the methodology by which the competent authorities satisfy themselves that the safe third country concept may be applied to a particular country or to a particular applicant. Such methodology shall include case-by-case consideration of the safety of the country for a particular applicant and/or national designation of countries considered to be generally safe’; and (c) ‘rules in accordance with international law, allowing an individual examination of whether the third country concerned is safe for a particular applicant which, as a minimum, shall permit the applicant to challenge the application of the safe third country concept on the grounds that the third country is not safe in his or her particular circumstances’. Also, an asylum-seeker must ‘be allowed to challenge the existence of a connection between him or her and the third country’, as referred to in point (a).

Next, Article 38(3) requires Member States to inform the applicant that the principle is being applied to them, and also to give the asylum-seeker a document informing the supposedly ‘safe’ third country that the merits of their asylum claim have not been examined.

Crucially for this case, Article 38(4) then provides that:

Where the third country does not permit the applicant to enter its territory, Member States shall ensure that access to a procedure is given in accordance with the basic principles and guarantees described in Chapter II. [ie an ‘ordinary’ examination of the merits of an asylum claim]

Finally, Article 38(5) obliges Member States to inform the Commission of their use of the concept.

The practical relevance of applying the ‘safe third country’ notion is (among other things) that it is a ground of (optional) inadmissibility under Article 33 of the Directive. And if the case is inadmissible, the asylum application is not further considered on the merits and has failed in that Member State (subject to an appeal of the inadmissibility decision, or the limited prospect of a ‘repeat application’, discussed further below). The failed asylum-seeker can then deprived of support as an asylum seeker under the reception conditions Directive and detained and removed under the Returns Directive. Of course, the impact of this is in theory ameliorated in ‘safe third country’ cases because the supposedly ‘safe’ country will readmit the person concerned and consider their asylum application properly. In practice, as we have seen, Turkey was explicitly and unambiguously refusing to readmit asylum-seekers from Greece in the context of the ‘safe third country’ concept.

Case law

The CJEU had interpreted the ‘safe third country’ rule in some prior case law, all concerning Hungary. First of all, in Case C‑564/18 (paras 29-51 of the judgment), the CJEU confirmed that the grounds of inadmissibility listed in the Directive were exhaustive (‘only’), and ruled that a national law applying a ‘safe third country’ principle to non-EU States where the asylum-seeker ‘as not exposed to persecution or a risk of serious harm, or in which a sufficient degree of protection is guaranteed’ was not consistent with the requirements of Article 38(1) of the Directive, in particular because it omitted the ‘non-refoulement’ requirement. Moreover, mere transit through the non-EU country was not enough for there to be a ‘connection’ with that country as required by Article 38(2).

Next, the Court reiterated these points in its judgment in Joined Cases C-924/19 and C-925/19 PPU (paras 148-165); but questions about what happens when the non-EU State refuses to readmit the asylum-seekers – the issue since addressed in last week’s judgment – were ruled inadmissible (paras 166-174). However, the Advocate-General’s opinion in that case discusses the issue (paras 107-127), taking the view that the refusal of readmission is only relevant at the point of enforcement of a decision based on the principle.  

Finally, in Commission v Hungary (Case C‑821/19), along with ruling on the criminalisation of assistance to asylum seekers, the Court again reiterated that the Hungarian law breached Article 38(1) and (2) of the Directive. The existence of a (then) pending proposal to amend the Directive on this point (since adopted as part of the 2024 asylum pact – but without some of the changes originally proposed) was irrelevant in determining whether a Member State was in breach of EU law at the time of the alleged breach.

 

Greek practice

As noted above, Greece had listed Turkey as a ‘safe third country’ despite Turkey’s refusal to readmit anyone in that context. While the CJEU judgment says nothing about the practice of the authorities, and the Advocate-General’s opinion says little (paras 65-68 of the opinion), one can find rather more information at the bottom of the ivory tower, in the recent Pro-Asyl report on the ‘safe third country’ principle in Greece, and in the annual AIDA report on Greece (‘safe third country’ section).

The reports state that over 10,000 asylum claims have been found inadmissible on the basis of the Greek listing of Turkey as a ‘safe third country’, although around 12,000 asylum claims have been found admissible (in some cases because the asylum-seekers argued that Turkey had not allowed them to make an asylum claim, or due to past exposure to refoulement, ie return to an unsafe country, or because they lacked a sufficient ‘connection’ with Turkey, on the basis of the national law implementing Article 38(2) of the Directive). Also, the reports are highly critical of the conclusion that Turkey is ‘safe’ – although the recent CJEU judgment does not address this issue, because the national court did not ask the CJEU about it.

Like the Advocate-General’s opinion, the Pro-Asyl study reports that Greek authorities apply the readmission test at the point of enforcement of a ‘safe third country’ decision. When Turkey then refuses readmission, the Greek bodies refuse to withdraw their prior decisions, leaving asylum-seekers to make a ‘repeat application’. But there are delays in registering these applications, and the authorities often reject them because they do not believe that the refusal of readmission is a ‘new element or finding’ (as required for a repeat application to be admissible under the EU Directive). But at least some courts rule against detention in such cases, given that there is no reasonable prospect of return (as required by the Returns Directive).

 

The judgment

The CJEU’s judgment began by answering the first question referred by the national court, rephrasing it as asking ‘in essence, whether’ the ‘safe third country’ rule in the Directive, ‘read in the light of Article 18 of the Charter’ on the right to asylum, ‘must be interpreted as precluding legislation of a Member State classifying a third country as generally safe for certain categories of applicants for international protection where, despite the legal obligation to which it is subject, that third country has generally suspended the admission or readmission of those applicants to its territory and there is no foreseeable prospect of a change in that position’.

First, the Court reiterated its case law that the grounds for inadmissibility of asylum applications in the procedures Directive (including ‘safe third country’ as defined in the Directive) are ‘exhaustive’, ie the Member States cannot add more grounds of their own (such as a divergent national definition of ‘safe third country’). It pointed out that the EU law definition ‘is subject to compliance with the cumulative conditions laid down’ in Article 38 of the Directive, as summarised above. But as the national court had already ruled that Greek law complied with the basic principles in Article 38(1), the CJEU did not examine that point. But it summarised the requirements for national law set out in Article 38(2) and the information obligations in Article 38(3), then recalled the Article 38(4) rule that ‘where the third country concerned does not permit the applicant for international protection to enter its territory, the Member States are to ensure that access to a procedure is given in accordance with the basic principles and guarantees described in Chapter II of that directive’.

In the CJEU’s view, it followed from this first from the wording of the Directive that a Member State can designate a country as ‘generally safe’ by ‘an act of general application’. The references to national law in Article 38(2) referred to such a designation, as did the preamble to the Directive. Also, the Directive did not make the validity of such a general act ‘subject to the condition that it be proved that the applicants for international protection concerned will actually be admitted or readmitted to the territory of that third country’. This was not among the national law rules in Article 38(2); and the wording of Article 38(4) implied that the designation of the third country as ‘safe’ preceded the consideration of the issue of whether that country actually readmitted asylum seekers.

The CJEU also believed that this followed from the context of the Directive, in particular by comparison with the wording of the rules on ‘first country of asylum’ (ie where the asylum seeker already had protection, as distinct from where they could have sought it) on the readmission issue.

Finally, in the CJEU’s view, this interpretation also did not conflict with the objective of the Directive, as set out in the preamble, that in the interests of asylum-seekers in particular, asylum applications ‘should be made as soon as possible, without prejudice to an adequate and complete examination being carried out’. This objective was addressed by Article 38(4), requiring that a Member State ‘ensure’ access to an asylum decision on the merits, in accordance with the usual rules, if the ‘safe third country’ refused readmission. And then a key point:

54      It follows that, where it is established that the third country designated as generally safe by a Member State does not in fact admit or readmit the applicants for international protection concerned, that Member State cannot reject their applications for international protection as inadmissible on the basis of Article 33(2)(c) of Directive 2013/32. Furthermore, that Member State may not unjustifiably postpone the examination of those applications and must, inter alia, ensure that that examination is conducted on an individual basis, in accordance with Article 10(3)(a) of that directive and in compliance with the time limits set out in Article 31 thereof.

This interpretation, in the Court’s view, ‘is likewise not such as to deprive of any practical effect the right of an applicant for international protection, as enshrined in Article 18 of the Charter and given specific expression by that directive, to obtain the status of beneficiary of international protection, provided that the conditions required by EU law are met.

Having ruled, in effect, that the designation of Turkey was valid, the CJEU felt it was unnecessary to answer the second and third questions referred by the national court, on whether the refusal of readmission was a factor to be considered at the admissibility or the enforcement stage. But as we have seen, the Court had already de facto answered the second question: if it is ‘established’ that the ‘safe’ third country ‘does not in fact’ readmit the asylum seekers concerned, a Member State ‘cannot’ find their applications admissible, and must consider those applications on the merits (moreover, without ‘unjustifiable’ postponement, and subject to the deadlines in the Directive). And that renders the third question moot.

 

Comments

Greece can have a dog; but it will be wagged by its Turkish tail. That is the effect of the Court’s judgment: in more technical legal language, Turkey can remain on Greece’s ‘safe third country’ list, but the practical impact of this is zero as long as Turkey refuses to readmit asylum-seekers from Greece, because asylum applications cannot be found inadmissible on this ground. The wheel of designation of Turkey as a ‘safe third country’ keeps spinning, but the hamster of inadmissibility of asylum applications is dead. Let’s unpack this further, examining first the reasoning of the judgment and then its potential consequences.  

Judicial reasoning

The CJEU begins by, with respect, answering a ‘straw man’ question that it had not been asked.  The national court did not ask the CJEU if it was possible to designate ‘safe third countries’ at all; rather it asked if those countries could be designated if they failed to readmit asylum-seekers. The references to common principles on ‘safe third countries’ in the preamble to the Directive do not answer the question as to what those common principles are; a rule that failure to readmit precludes listing a country would be a ‘common principle’ too. However, the Court’s points that Article 38(2) on national law and (especially) Article 38(4) on readmission presume that the country concerned is already on a list are more convincing.

Most importantly, the Court’s finding that applications cannot be inadmissible on ‘safe third country’ grounds where the country concerned refuses to readmit, requiring a Member State to consider the merits of such applications, is particularly convincing.  Given that between Greece and Turkey the readmission will not take place, and in light of the evidence that repeat applications are difficult to make and are often rejected, this is the best way to ensure that the obligations set out in Article 38(4), coupled with the principle of considering applications as soon as possible, are met.

Unlike some judgments, this is not so much a ‘liberal’ or ‘activist’ interpretation of EU asylum law by the CJEU, but rather the overdue termination of conservative activism by the Greek authorities and lower courts. In light of the clear obligation in Article 38(4) to consider the merits of an asylum application if the asylum seeker is not readmitted to a ‘safe third country’, in the context of the absolute refusal of Turkey to readmit, the failure of the authorities to consider the merits of many of those applications, and of the courts to require them to do so, is unjustifiable and frankly shocking.

Consequences in practice

First and foremost, what happens to the 10,000 plus asylum seekers whose applications were already rejected as inadmissible? As noted above, some had previously tried to make repeat applications, but faced the difficulty that they had to invoke ‘new elements or findings’ in order to do so, which was often difficult to do. But their ability to invoke the ‘repeat applications’ clause in the Directive has now, as a consequence purely of the Court of Justice’s recent judgment, undergone a Copernican revolution. This is because the CJEU has previously ruled (see Joined Cases C-924/19 and C-925/19 PPU, paras 175-203, confirmed in Case C-216/22) that its own judgments on EU asylum law are a new element or finding in filing a repeat application, where those judgments are relevant and add significantly to the likelihood of an asylum application being successful. For those asylum-seekers whose application was found inadmissible on ‘safe third country’ grounds despite Turkey’s refusal to readmit them, last week’s judgment is obviously highly relevant; and overturning the finding that their applications were inadmissible equally obviously significantly increases the odds that their application will be successful. (Moreover, the judgment in Joined Cases C-924/19 and C-925/19 PPU is especially relevant by analogy here, since the prior national judgments which conflicted with a later CJEU judgment in that case also concerned an unlawful application of the ‘safe third country’ rules).

Therefore, if the rejected asylum-seekers apply again, the Greek asylum system will now have to decide on the merits of thousands of asylum applications that it wrongly ruled were inadmissible in the first place. A pending CJEU case referred from Ireland may be relevant by analogy in establishing whether Greece has to pay damages too. Finally, if any Afghan women asylum-seekers were among those whose applications were considered inadmissible, they can also rely on another CJEU judgment from the same day (discussed here) as a ‘new element or finding’, supporting the merits and fast-tracking of their gender-based persecution claims.

As for current or future applicants, their applications cannot be rejected as inadmissible on the grounds that Turkey is a 'safe third country' at all. Nor can they be rejected on the grounds that Turkey is a ‘first country of asylum’ for them: without even looking at the substance of that concept, the CJEU judgment explicitly stated that this principle cannot even get off the ground if the country concerned will not readmit the asylum seeker.

What about applicants after the asylum pact takes effect? The Court of Justice judgment from last week only mentions the new asylum procedures Regulation (discussed here) in passing, but as the Advocate-General’s opinion pointed out (in para 64), the interpretation of the Directive ultimately followed by the Court is explicitly set out in the Regulation, now being built into the definition of inadmissibility – also as regards the ‘first country of asylum’ principle (these grounds of inadmissibility can be invoked for either principle ‘unless it is clear that the applicant will not be admitted or readmitted to that country’). (See also recital 53 in the preamble to the Regulation). Whether the planned review of the ‘safe third country’ rules in the Regulation in 2025 affects these issues remains to be seen.

Finally, the Court’s judgment understandably only dealt with situations where the refusal of readmission is certain at the time of considering the possible application of the ‘safe third country’ rule. It did not deal with the scenario where refusal was merely a possibility – as that was hypothetical on the facts. How should such cases be dealt with?

The starting point is that the current Directive (recital 44 in the preamble) implies the possibility of an inadmissibility decision in that scenario: Member States do not have to examine the merits of an application where ‘there are grounds for considering that the applicant will be admitted or readmitted to that [‘safe’ third] country’. The inadmissibility rule in the new Regulation points in the same direction: it can’t be applied if ‘it is clear that’ readmission is not going to happen (see also recital 43 in the preamble to the Regulation). But this implies that the inadmissibility rule can be applied if it is not clear that readmission is ruled out. On the other hand, the recent judgment points to the importance of deciding on applications as soon as possible, without unjustifiable postponements, while ensuring the application of the Charter right to asylum. How best to reconcile the possibility of the adoption of inadmissibility decisions with the need to take decisions speedily and afford an effective prospect of having an asylum application decided on the merits, if there is uncertainty about whether readmission will actually take place?

The 2024 procedures Regulation already expressly provides (in recital 53 in the preamble) that access to the Member State’s asylum system must be offered again if readmission does not take place:

…if the applicant is eventually not admitted or readmitted to the third country after the application has been rejected as inadmissible, the applicant should again have access to the procedure for international protection in accordance with this Regulation

How should that principle apply in concrete terms? The best approach is to infer an obligation upon the Member State to act with due diligence to attempt to secure the readmission, starting as soon as it takes the inadmissibility decision (after any appeals, or the prospect of appeals, against that decision are spent). From that point, the Member State should have a reasonable time frame to request readmission from the non-EU country concerned. By analogy with the procedures Regulation time limit for deciding on inadmissibility, two months should be sufficient. If the non-EU country refuses readmission, or does not reply within the time limit set out in the readmission arrangements in force, or accepts readmission in principle but does not carry it out in practice within the relevant time limit, then the Member State’s obligation to consider the application on the merits where readmission does not take place must be triggered. Procedurally either the inadmissibility decision must be considered as withdrawn from that point (it makes sense to consider it as only provisional until readmission is actually carried out), or a repeat application must be considered – the requirement of new developments being satisfied by the inability, subsequent to the inadmissibility decision, to ensure readmission in practice within a reasonable time frame. The inadmissibility decision was based on the presumption that readmission would be secured within a reasonable period; that presumption has since been rebutted.

The final word on the impact of last week’s judgment should go to the asylum NGOs who brought the case:

…the CJEU’s judgment is of major importance because it overturns an arbitrary and abusive practice of the Greek authorities which has been in force for many years, of wholesale rejecting asylum applications as inadmissible under the safe third country principle, and puts an end to the denial of the rights of thousands of applicants, by prohibiting their rejection when there is no possibility of their readmission to Türkiye and dictating their individual examination in accordance with the Directive.