Oh Lord, don’t follow the Opinion in Mercedes-Benz: Advocate General Rantos proposes a retroactive interpretation of European type approval law which makes up to 200 Mio type approved motor vehicles in the EU illegal

 

 

Dr. Benedikt Wolfers, M.A. & Sebastian Lutz-Bachmann, LL.M.*

 

* Partners of the Law Firm Posser Spieth Wolfers & Partners in Berlin.

 

Photo credit: Reinhold Möller, via Wikimedia Commons

 

In recent years, the Court of Justice decided a series of cases in relation to vehicle type approval law. On 21 November 2024, Advocate General Rantos published his Opinion in two further cases (joint cases C-251/23, C-308/23 Mercedes-Benz). A regulatory key question in these cases is whether EU5 and EU6 vehicles registered in the EU that comply with the emission limits on the legally required chassis dynamometer test in the New European Driving Cycle (NEDC) also have to comply with these emission limits in real driving conditions outside the NEDC. For NEDC vehicles the more recent Real Driving Emissions (RDE) test, that has been introduced as of 1.9.2017, is not applicable.

 

In his Opinion, Advocate General Rantos nevertheless considers NEDC emission limits to be applicable in real driving conditions: In his view, NEDC vehicles “must simultaneously comply with all of the limits values set out in Annex I to Regulation No 715/2007, both in the NEDC test and in real driving conditions” (para. 61). He states “that compliance with the NOx emission limit values must be verified not only in the context of the NEDC test but also under real driving conditions, in normal use” (para. 47). This general, far reaching statement on Art. 4(2) and Art. 5(1) of Regulation (EC) No 715/2007 (on type approval of motor vehicles) is the basis to answer the initial question put the court: whether a NEDC vehicle has to meet the emission limit values in the NEDC test not only after a cold start (as Annex I to Regulation (EC) No. 715/2007 requires) but also after a warm start (as Annex I does not require). Advocate General Rantos confirms this and justifies his position with his general statement, “in so far as driving with a warm engine constitutes normal use […] in real driving conditions, as they are usually present in the territory of the European Union” (paras. 47, 92).

 

Since the Advocate General interprets the scope of Art. 4(2) and Art. 5(1) of Regulation (EC) No 715/2007 generally his view applies to all EU5 and EU6 NEDC light passenger and commercial vehicles (be extended to EU4 ), with a diesel or petrol engine, and is relevant for all courts and EU type approval and market surveillance authorities who prove compliance of NEDC vehicles with the EU emissions regulation - far beyond the individual case at hand.

 

Should Advocate General Rantos’ view prevail, holders of NEDC vehicles would have to meet retroactively the additional requirement to meet the emission limits in real driving conditions which did not exist at the time of the type approval and marketing of the vehicles. Holders will not able to not do so - because NEDC vehicles are no RDE vehicles and can typically not meet the NEDC limits in real driving conditions. This means that up to 200 Mio of EU citizens, tradespeople, service providers, etc. in the EU could be exposed to the risk that their NEDC vehicles, which did comply with the statutory test when approved, would today, retroactively, be considered illegal. This could lead to massive damage claims and recall of vehicles, imposing an unbelievably high burden on customers and the EU industry.

 

Advocate General Rantos’ opinion is not only very dangerous but also, legally, erroneous. It infringes the legal and factual status of the last 20 years as well as the fundamental legal principle that measures should have no retroactive effects. For more than 20 years, EU law has provided that the limit values for NEDC vehicles are tested in the NEDC test only, not in real driving conditions. For more than 20 years, the EU Commission, all EU member states, all vehicle approval authorities, scientists, courts, manufacturers, importers and consumers have applied the limit values for NEDC vehicles exclusively to the NEDC on the test bench, but not to normal use in real driving conditions. To comply with the law and to avoid massive risks for consumers and the industry in the EU the Court of Justice should not follow Advocate General Rantos’ proposal.

 

European type approval law does not provide for the application of NEDC emission limits in real drive conditions

 

Advocate General Rantos argues that the application of NEDC emission limits in real drive conditions would result from the existing case law of the Court of Justice on diesel vehicles (esp. from the decision in Case C-134/20 Volkswagen). However, in these cases the Court of Justice only decided on the interpretation of “normal driving conditions” in the context of defeat devices and did not decide that NEDC emission limits have to be met in real drive conditions. This is not the same. The definition and prohibition of a defeat device (in Art. 3(10) and Art. 5(2) of Regulation (EC) No 715/2007) is strictly to be distinguished from the question whether emission limits specified in Annex I to Regulation (EC) No. 715/2007 have to be met in real driving situations.

 

In that regard, European type approval law, to which Advocate General Rantos refers in his Opinion, does not provide for the application of NEDC emission limits in real drive conditions. Rather, Regulation (EC) No. 715/2007 and Regulation (EC) No. 692/2008, which form together the core of emission-related type approval law for vehicles in the European Union, clearly establish that compliance with the relevant emission limits is only required during the specified NEDC test procedures:

 

Emission limits for EU5 and EU6 are specified in Annex I to Regulation (EC) No. 715/2007. For passenger and light duty diesel vehicles NOx limits are 180mg/km (EU5) and 80mg/km (EU6) respectively. Art. 4(1) sub 2 Regulation (EC) No. 715/2007 obliges manufacturers to comply with the emission limit values set out in Annex I. The reference to meet the limits is immediately followed by a reference to the implementing measures. According to these implementing measures, which are set out in Regulation (EC) No. 692/2008, the emission limits must only be met under test procedures specified in the detailed annexes to Regulation (EC) No. 692/2008 (Art. 3 para 6 Regulation (EC) No. 692/2008). These test procedures most notably include the NEDC test procedures. Until 2017 the test procedures set out in the annexes to Regulation (EC) No. 692/2008 did not, however, refer to any testing under real driving conditions. Therefore, until 2017 all type-approved vehicles had to comply with the emission limit values set out in Annex I to Regulation (EC) No. 715/2007 only in the prescribed test procedures on the dynamometer.

 

Only for vehicles that have been type-approved since 1.9.2017 new test procedures that include normal use under specified real driving conditions on the road have been introduced. However, these RDE tests are not retroactively applicable for vehicles type-approved according to the NEDC before 2017.

 

Tests for NEDC vehicles under real drive conditions would not have been possible

 

An interpretation of the law which requires the impossible demonstrates that it is incorrect. This applies to Advocate General Rantos’ view that NEDC vehicles should meet the emission limits in real driving conditions: No type approval authority, no technical service and no manufacturer could have implemented this – for two reasons:

 

First, until 2017 EU law did not provide for any procedure to test emission limits of EU4, EU5 and EU6 NEDC passenger cars and light duty vehicles in real driving conditions. Each emission limit, however, requires a description of the test procedure with which it is to be tested. This is a scientific-technical law and logically mandatory. Every material test and every technical limit requires a description of the procedure in which it is tested. However, a procedure for testing emissions from passenger cars and light duty vehicles in real driving conditions was only introduced in EU law for new vehicle types as of 1.9.2017 (introduction of RDE tests). Without a legally recognized test procedure, type approval authorities could not have tested NEDC vehicles in real driving conditions.

 

Second, until 2017 it would not have been technically possible to test the emissions of passenger cars and light duty vehicles in real driving conditions. So-called PEMS devices, which have been used to carry out the RDE test of passenger cars and light duty vehicles since 2017, were not available when EU4 and EU5 were introduced in 2005 and 2009 respectively. The first recognition of PEMS devices as a measurement technology under EU law was for heavy duty vehicles in 2011 which was not applicable for passenger cars and light duty vehicles (Art. 2 No. 12 and Annex II Appendix 1 of Commission Regulation 582/2011 of 25 May 2011 on Heavy Duty Vehicles Emissions). For NEDC passenger cars and light duty vehicles, PEMS measurement technology for real driving conditions was only created and recognised in EU law with the introduction of RDE tests for new vehicle types as of 1.9.2017. Without a legally recognized PEMS test equipment, type approval authorities could not have tested NEDC vehicles in real driving conditions.

 

Advocate General Rantos’ opinion contradicts with practice of European Commission, Member States and all type approval authorities for more than 20 years

 

The incorrectness of Advocate General Rantos’ view is confirmed by the practice of the European Commission, EU Member States, vehicle approval authorities, scientists and courts, which for more than 20 years, as numerous documents show, have applied the EU4, EU5 and EU6 emission limits for NEDC vehicles only in the NEDC and not under normal use in real driving conditions.

 

Because the NEDC was criticised for not being representative for real driving conditions it was replaced at EU level in 2017 for new vehicle types by a new test procedure which provides for a more demanding driving profile on the chassis dynamometer (WLTP) and additionally and for the first time by a test under normal use under specified real driving conditions on the road (RDE test). This new legislation and its historical background confirm that before the change of law in 2017 the emission limits of NEDC vehicles had to tested in the NEDC only. This is also the view of the European Commission which in 2019 explicitly stated that “the mandatory NOx emission limits for diesel cars in the EU have been progressively lowered from 500 mg/km at Euro 3 […] to 80 mg/km at Euro 6. However, there is evidence that despite these increasingly stringent NOx limits, that are verified under laboratory conditions and on a standardised test cycle, the actual NOx emissions under real driving conditions did not come down to the same extent (Weiss et al. 2011). In response to this, the Commission has developed the real driving emissions procedure (RDE). […] It first applied for monitoring purposes from early 2016 to complement the laboratory test and has been applicable to new vehicle types as from 1 September 2017” (see: Proposal by EU Commission for a Regulation amending Regulation 2007/715; COM(2019) 208 final).

 

Conclusion: Court of Justice should not follow Advocate General’s Opinion

 

Retroactively expanding the scope of emission limits for NEDC vehicles on real driving conditions would undermine the trust in the harmonized European type-approval law and in the internal market which for more than 20 years has provided that the emission limit values for NEDC vehicles are tested in the NEDC only, not in real driving conditions. Up to 200 Mio of EU citizens, tradespeople, service providers, etc. in the EU could be exposed to the risk that their NEDC vehicles, which did comply with the statutory test when approved, would today, retroactively, be considered illegal. Such an important decision falls outside of the competence of the Court of Justice and would rather have to be adopted by a legislative act (Art. 289(3) TFEU). However, such a legal act, that would retroactively impose additional duties (and potentially sanction), would not fulfil the requirements established by the Court of Justice under which such a retroactive application is exceptionally permissible (Case 98/78, ECR 1979, 69, para. 20 (Racke); Case 84/81, ECR 1982, 1763 para. 13 - 15 (Staple Dairy Products); Case 114/81, ECR 1982, 3189, para. 5-11 (Tunnel Refineries)). Therefore, the Court of Justice should not follow the arguments in Advocate General Rantos’ opinion.

 

As long as the system of remedies and the objectives are not undermined: The Court of Justice on GDPR enforcement (Case C-21/23, Lindenapotheke)

 

 

 

Alessandra Fratini and Giorgia Lo Tauro, Fratini Vergano European lawyers

Photo credit: via Wikimedia Commons

Introduction

On 4 October 2024, the Grand Chamber of the Court of Justice of the European Union issued its judgment in Lindenapotheke (Case C-21/23), a case concerning the online sale of pharmacy-only medicinal products and its implications as regards GDPR compliance. In its request for a preliminary ruling, the German Federal Court of Justice (Bundesgerichtshof) raised two questions on the interpretation of the GDPR. While acknowledging the importance of the second question on the meaning of ‘data concerning health’, this post focuses on the first one, concerning the compatibility of the system of remedies established in Chapter VIII GDPR with other remedies under national law. The paragraphs below, after a short overview of the facts of the case and the preliminary questions, review the main findings of the Advocate General and of the Court of Justice on the first question and conclude by placing the judgment within the rising trend of addressing the challenges of digital markets through a broader enforcement of EU digital regulation.

 

Facts of the case and questions referred

The main proceedings involved two competitors operating pharmacies in Germany, ND and DR. ND, which operates a pharmacy under the trade name ‘Lindenapotheke’, has been selling pharmacy-only medicinal products via the ‘Amazon-Marketplace’ online platform since 2017.

DR brought an action before the German Regional Court seeking an order for ND to cease selling pharmacy-only medicinal products via the online marketplace on the basis that such marketing constituted an unfair commercial practice in so far as it was pursued in breach of Article 9 GDPR, which requires that the data subject’s prior explicit consent be obtained for the processing of data concerning health. According to the German law against unfair competition, in fact, “anyone who infringes a statutory provision intended, inter alia, to regulate market conduct in the interest of market players acts unfairly where that infringement is capable of having an appreciable adverse effect on consumers, other market players or competitors”; such an infringement constitutes a prohibited unfair commercial practice enabling any competitor to claim an injunctive relief (paras. 21-23 of the judgment). The Regional Court upheld the action and the subsequent appeal brought by ND was dismissed by the Higher Regional Court, which held that such an online marketing was contrary to the national law against unfair competition. ND lodged an appeal on a point of law before the German Federal Court of Justice, which raised a request for a preliminary ruling on the interpretation of Chapter VIII and Article 9(1) GDPR, but also Article 8(1) of Directive 95/46 (the previous data protection Directive) before the Court of Justice.

Question 1

With its first question, the referring court asked the Court of Justice whether a competitor, who is not a data subject within the meaning of Article 4(1) GDPR, has standing to bring an action before the civil courts against the alleged infringer of the GDPR, on the basis that the alleged infringement falls within the prohibition of unfair commercial practices. The referring court noted that the provisions of Chapter VIII GDPR do not mention, nor do they explicitly exclude, the possibility for competitors to bring an action against an undertaking, where the infringement of data protection law constitutes an unfair commercial practice (para. 35). The referring court underlined the uncertainty of the situation and highlighted both the risks of recognising such a possibility for competitors, in terms of potential encroaching on the powers of the supervisory authorities and ensuing divergences, and its potential benefits in terms of ‘effet utile’ to ensure the highest level of data protection (paras. 36-39).

Question 2

With its second question, the referring court asked the Court of Justice to clarify whether the data which customers must enter on the online sales platform when ordering medicinal products (such as name, delivery address and information required for individualising the medicinal products ordered) constitute ‘data concerning health’ within the meaning of Article 8(1) of Directive 95/46 and Article 9(1) GDPR. In particular, the doubts of the referring court concerned non-prescription medicinal products, since these may be intended not necessarily for the customers but for third parties, who may not be identifiable (para. 41).

In the opinion of the referring court, the questions of a competitor’s standing to bring proceedings (para. 39) and of the notion of ‘special categories of personal data’ (para. 43) had not been clarified by the case-law of the Court of Justice and warranted its request for a preliminary ruling.

 

The Opinion

In his Opinion, Advocate General Szpunar first changed the order of the proposed questions, as he considered that if the answer to the second one were to be negative, there would be no need to answer the first one (para. 31 of the Opinion). Addressing the second question at the outset, the AG suggested to answer that “the data of the customers of a pharmacist which are transmitted when an order is placed on an online sales platform for pharmacy-only but non-prescription medicines do not constitute ‘data concerning health’ within the meaning of Article 4(15) and Article 9 of the GDPR, in so far as only hypothetical or imprecise conclusions as to the health status of the person placing the online order may be drawn, which it is for the referring court to verify” (para. 54).

In the light of that proposed negative answer, the first question was dealt with in the Opinion only for the sake of completeness. Having acknowledged that the GDPR confers no rights on undertakings and their competitors, as that regulation grants rights only to data subjects (paras. 79-81), the AG assessed whether the GDPR system of remedies has to be seen as an exhaustive system, in the sense that it precludes undertakings from relying on a GDPR infringement in the context of other remedies provided for by national law (paras. 82-89).

First, he noted that the action at issue in the main proceedings was not based on a GDPR infringement, but took such an infringement into account in an incidental manner. The Court already accepted, in its judgment in Meta Platforms and others (2023), that data may be taken into account in an incidental manner and that an infringement of the GDPR may constitute an infringement of competition law (paras. 90-91), and the AG considered that was applicable to the present case (para. 91). Second, as regards the interaction between national actions in which the GDPR can be invoked incidentally and the GDPR system of remedies, the AG observed that the former should be accepted only on condition that they do not undermine the GDPR system of remedies or the attainment of its objectives (para. 95). In the present case, since an action brought by an undertaking against a competitor is not intended to ensure respect for the data subjects’ rights but pursues another objective, the actions made available to data subjects by the GDPR system of remedies are preserved and may still be exercised in those circumstances (paras. 100-101). Furthermore, in the AG’s view, the objectives pursued by the GDPR, such as the high level of protection of natural persons and the consistent and homogenous application of the data protection rules (recital 10), are not threatened (but, as for the high level of protection, actually strengthened) by the possibility afforded to an undertaking to bring an action for an injunction against a competitor based on the prohibition of acts of unfair competition, in reliance on a GDPR infringement by that competitor (paras. 103-104). Finally, the AG noted that, far from being undermined, the effectiveness of the GDPR would be reinforced by the fact that compliance with its provisions may also be enforced in judicial proceedings distinct from those within its system of remedies. Accordingly, he concluded that such national remedies may exist alongside the system established by the GDPR (paras. 105-108).

 

The Judgment

The Court of Justice considered the questions in the order they were raised by the referring court and departed from the Opinion with regard to the answer to the second question.

To address the first question, the Court interpreted the relevant provisions of Chapter VIII GDPR by relying on their wording, the context and the objectives pursued by the GDPR (para. 52 of the judgment). As to the wording, the Court noted that not only the provisions of Chapter VIII do not expressly rule out the possibility for additional national remedies, but the rights provided for by Article 77(1), Article 78(1) and Article 79(1) are ‘without prejudice’ to any other administrative, judicial or non-judicial remedy (para. 53). When it comes to the context, while it agreed with the AG that only data subjects are beneficiaries of the GDPR protection, the Court noted in addition that the infringement of its substantive provisions is also liable to adversely affect third parties (in this sense, it referred to the right to compensation provided for by Article 82(1); para. 55). The Court recalled that it had already held that the infringement of data protection rules may at the same time give rise to an infringement of rules on consumer protection or unfair commercial practices (judgment in Meta Platforms Ireland, 2022, para. 78) and may be “a vital clue” in the assessment of an abuse of a dominant position (judgment in Meta Platforms and others, 2023, para. 47) (para. 55). It also noted the importance of access to personal data and the ability to process such data, which “have become a significant parameter of competition between undertakings in the digital economy”, so that it may be necessary to consider rules on data protection when enforcing competition law and the rules on unfair commercial practices (para. 56).

Interestingly, while the above would have been sufficient to interpret Chapter VIII in the light of the context, the Court went further to consider the margin of discretion enjoyed by Member States in the implementation of the GDPR. In this respect, even though the GDPR “seeks to ensure the harmonisation of national legislation on the protection of personal data which is, in principle, full, the fact remains that several provisions of that regulation expressly make it possible for Member States to lay down additional, stricter or derogating national rules, which leave them a margin of discretion as to the manner in which those provisions may be implemented (‘opening clauses’)”(para. 57). After referring to its judgment in Meta Platforms Ireland (2022, para. 57), which concerned a provision of the GDPR (Article 80) expressly containing an opening clause, the Court added: “It is true that the provisions of Chapter VIII of the GDPR do not specifically provide for such an opening clause which would expressly allow Member States to make it possible for a competitor of an undertaking which allegedly infringes the substantive provisions of that regulation to bring an action in order to put an end to that infringement. However, it follows from the wording and context of the provisions of Chapter VIII (…) that, by adopting that regulation, the EU legislature did not intend to bring about an exhaustive harmonisation of the remedies available in respect of infringements of the provisions of the GDPR and, in particular, did not wish to rule out the availability of such remedies to competitors of the person allegedly responsible for an infringement of the laws protecting personal data, on the basis of national law relating to the prohibition of unfair commercial practices” (paras. 59-60, emphasis added).

In the Court’s view, that interpretation was corroborated by the GDPR objectives (i.e., ensuring a consistent and high level of protection of natural persons with regard to the processing of personal data and removing obstacles to the flow of such data within the EU; strengthening of the rights of data subjects and of the obligations of those who process and determine the processing of data, as well as equivalent powers for monitoring and ensuring compliance with the rules for the protection of personal data and equivalent sanctions for infringements in the Member States; providing natural persons in all Member States with the same level of legally enforceable rights and obligations and responsibilities for data controllers and processors, and ensuring consistent monitoring of the processing of personal data, and equivalent sanctions in all Member States) (para. 61). It found therefore that the possibility of national remedies like those at stake does not undermine those objectives but actually enhances the effectiveness of the GDPR provisions (para. 62). These national remedies are in addition to those of Chapter VIII and pursue an objective (fair competition) which is different from those pursued by the GDPR. In this context, as the German government observed, the uniform interpretation of the GDPR remains ensured by the preliminary ruling procedure under Article 267 TFEU (paras. 65-67). Furthermore, the Court held that national remedies aimed at ensuring fair competition undoubtedly contribute to compliance with the GDPR and, therefore, to strengthening the rights of data subjects: an application for injunctive relief filed by a competitor may also prove particularly effective in so far as it may prevent a large number of infringements of data subjects’ rights (paras. 69-70).

In the light of the above, the Court concluded that Chapter VIII does not preclude national legislation providing for such remedies to the benefit of competitors, while leaving to the referring court the assessment of whether the alleged infringement of the GDPR, in so far as it is established, also constitutes a breach of the prohibition of unfair commercial practices under the relevant national law (paras. 71-72).

As to the second question, suffice it to say that the Court, unlike the AG, found that the information which customers enter when ordering online pharmacy-only medicinal products, the sale of which does not require a prescription, does constitute ‘data concerning health’ even where it is “only with a certain degree of probability, and not with absolute certainty, that those medicinal products are intended for those customers” (para. 90). This, however, does not preclude it from being processed, in specific contexts, if the conditions for exemptions are met (para. 92), i.e. does not mean automatically that the processing is in breach of the GDPR.

 

Concluding remarks

The judgment in Lindenapotheke, as far as the first question is concerned, provides an interpretation of the GDPR system of remedies aimed at enhancing the effectiveness of data protection. The remarkable point of the reasoning is the emphasis placed on the margin of discretion recognised to Member States in implementing the GDPR, with a view to enhancing the protection afforded by it. While in Meta Platforms Ireland (2022) the Court could rely on the wording of the provision concerned (para. 59: “(…) Article 80(2) of the GDPR, which leaves the Member States a discretion with regard to its implementation. (…) Member States must make use of the option made available to them by that provision to provide in their national law for that mode of representation of data subjects”), in Lindenapotheke it admitted that Chapter VIII does not expressly provide for any opening clause allowing Member States to make available further remedies for actors other than data subjects invoking a GDPR infringement. However, by relying on the wording and context of Chapter VIII, as well as on the legislator’s intention and the GDPR objectives, it came to the conclusion that Member States can make available such remedies to competitors of the person allegedly responsible for an infringement of the laws protecting personal data, since such a possibility is not being ruled out by the GDPR system of remedies and its objectives (paras. 60-61). The Court’s interpretation actually seems to encourage Member States to make additional remedies available under national laws, insofar as they enhance the effectiveness of data protection (paras. 62 and 69).

From this perspective, the Court’s conclusion is significantly relevant when placed in the context of the ongoing debate on the GDPR (under) enforcement (Gentile-Lynskey, 2022), the shortcomings of its composite enforcement system (Hofmann-Mustert, 2024) and the Commission’s Proposal for a Regulation laying down additional procedural rules relating to the enforcement of the GDPR (2023). When it comes to the handling of complaints and the role of complainants, it has been observed that these vary significantly among Member States, which in turn results in a limitation of individual procedural rights (Hofmann-Mustert, 2024). Against this background, some rightly fear, by comparing this judgment with previous case law, that its “implications have the potential to be more disruptive” as regards the consistent enforcement of the GDPR and introduce “greater potential risks of interference between administrative and judicial enforcement” (van den Poel, 2024).

However, the implications of the judgment are less daunting when considering the GDPR enforcement in the broader context of digital legislation. The Commission Second Report on the application of the GDPR, published on 25 July 2024, makes it clear that “the development of digital regulations raises the need for close cooperation across regulatory fields. Such cooperation is all the more necessary since data protection issues increasingly intersect with questions of, for example, competition law, consumer law, digital markets rules, electronic communications regulation and cybersecurity. (…) data protection authorities are taking steps to ensure their actions are complementary and coherent with other regulatory fields”. In its statement of 3 December 2024 on the Commission Second Report, the EDPB also recognised that it “would support a holistic methodological approach for the next evaluation of the GDPR that explores the interplay between the GDPR and other EU digital legislation”.

The judgment fits into this context of growing institutional awareness of the need for a holistic and coordinated approach for the effective protection of personal data, in line with the “more ‘collaborative approach’” proposed by scholars for the enforcement of data protection, competition law and unfair competition law (Vandendriessche, 2024). The Court insists on the likely enhanced effective enforcement of the GDPR via national remedies aiming at other objectives (Holtz, 2024), by proposing an interpretation where the GDPR as such calls upon the Member States for its effective enforcement (again, paras. 60-61). By stating that “such an application for injunctive relief brought by a competitor may prove, like that brought by a consumer protection association, to be particularly effective in ensuring such protection, in so far as it is capable of preventing a large number of infringements of the rights of data subjects by the processing of their personal data” (para. 70), the Court recognises the preventive effect of a potential “private enforcement” (Opinion, para. 93) through remedies allowed under national laws, which has been read as an ‘incentive’ for market players to contribute to GDPR compliance (Vandendriessche, 2024). In this sense, the judgment embraces an emerging approach in the EU regulation of the digital environment, which is aimed at involving in the enforcement multiple actors of society as a whole. This approach is evident when it comes to making the online world safer and fairer, namely with the DSA: for example, as far as institutional actors are concerned, in the cooperation required between the Commission and the Digital Services Coordinators with regard to systemic risk mitigation measures (Peukert, 2024); even more, as far as non-institutional actors are concerned, in the mechanisms required to allow any user - individual or entity - to notify illegal content online, or in the required cooperation with “trusted flaggers” (Articles 16, 22, 35 DSA) (in this sense, see also Commission’s dialogue with Civil Society Organisations for implementing the DSA).

It remains to be seen whether such an approach succeeds in becoming consolidated through greater coordination of EU institutions and national authorities and greater awareness of society at large, alongside the required adjustments for the effective implementation of the remedies the GDPR grants to data subjects.

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.