Establishing a Common European Asylum System by Leaving European Human Rights Standards Behind: Is this the Way Forward?

February 4, 2015

Francesco Maiani 
RefLaw, Editorial Advisory Panel
Swiss Graduate School of Public Administration of the University of Lausanne, Associate Professor

Things have definitely come to a head between the two European Courts, and for the time being, reconciliation is nowhere in sight. The “two European Courts” are the European Court of Justice (ECJ), the supreme judicature of the European Union (EU), and the European Court of Human Rights (ECtHR), the ultimate interpreter of the European Convention of Human Rights (ECHR). The point of contention is whether the pan-European standards of the ECHR apply with full force to inter-State cooperation in the EU “Area of Freedom, Security and Justice” (AFSJ).

As a matter of ECHR Law, there is no discussion: the ECHR applies to any act or omission of the Contracting Parties regardless of whether they are Members of the EU and regardless of whether they are implementing EU Law. Out of self-restraint, the ECtHR has developed its “Bosphorus” doctrine, whereby it will not normally review national measures that are strictly mandated by EU Law, as this would mean indirectly reviewing EU measures. But this doctrine does not affect the applicability of the substantive standards of the ECHR, and is conditional inter alia on the EU affording human rights protection equivalent to or greater than the ECHR (on all these aspects, see Bosphorus judgment, paras. 149-158).

As a matter of EU Law, it was also undisputed until recently whether Member States must fully abide by ECHR standards when implementing EU Law, particularly because such standards are part of EU general principles [see Art. 6(3) of the Treaty on the European Union(TEU)] and are incorporated in the Union’s own Charter of Fundamental Rights (CFR: see Art. 52(3)). And yet, in December 2014, the Full Court of the ECJ stated the following in paras. 191-194 of Opinion 2/13 on EU accession to the ECHR:

[The] principle [of mutual trust between the Member States] requires, particularly with regard to the [AFSJ], each of those States, save in exceptional circumstances, to consider all the other Member States to be complying with EU law and particularly with the fundamental rights recognised by EU law […]. Thus, when implementing EU Law, the Member States may, under EU law, be required to presume that fundamental rights have been observed by the other Member States, so that […], save in exceptional cases, they may not check whether that other Member State has actually, in a specific case, observed the fundamental rights guaranteed by the EU […]. In so far as the ECHR would […] require a Member State to check that another Member State has observed fundamental rights, even though EU law imposes an obligation of mutual trust between those Member States, [EU] accession [to the ECHR] is liable to upset the underlying balance of the EU and undermine the autonomy of EU law.

To understand this passage one must have in mind the typical “AFSJ measures,” e.g. the State-to-State surrender of a convicted person pursuant to a European Arrest Warrant or the State-to-State transfer of an “applicant for international protection” pursuant to the so-called Dublin system. What the ECJ implies here is that – as the law stands – it is up to the EU (i.e. largely to the ECJ itself) to define in what “exceptional cases” the executing or transferring State may ascertain the human rights consequences of such measures, regardless of what the ECHR prescribes.

Before commenting on this position, it is useful to consider its roots and its ramifications. Opinion 2/13 has in fact likely been shaped by a protracted exchange between the two Courts concerning the limits set to Dublin transfers by the prohibition of inhuman or degrading treatment (Art. 3 ECHR = Art. 4 CFR), and it has far-reaching implications in this area (see Opinion 2/13).

Human Rights in the Dublin System: the Soering test vs. the NS test

In 2009-2010, approximately 2,200 persons were sent to Greece under the Dublin Regulation. 14,000 more were scheduled for removal. The Member States were unmoved by the fact that – owing to the notorious dysfunctions of the Greek asylum system – this amounted to denying the “transferees” fair status determination and to exposing them to degrading detention conditions or abject destitution.

It was the MSS judgment, delivered by the Grand Chamber of the ECtHR in January 2011, that put an end to this mass human rights violation. In the judgment, the Court reiterated that Dublin transfers are fully subject to the ECHR and to its own review (paras 338-340). Bosphorus was held inapplicable for the good reason that the Dublin Regulation never mandates transfers. To the contrary, under its “sovereignty clause” (Art. 17(1)) every Member State may examine any protection claim that is lodged with it and may therefore forgo any transfer that would be incompatible with the ECHR.

The ECtHR also reiterated that “mutual trust” and “safety presumptions” cannot be absolute: Member States must refrain from transfers when the conditions of the well-established “Soering test” are met, i.e. “where substantial grounds ha[ve] been shown for believing that the person in question, if expelled, would face a real risk of being subjected to [inhuman or degrading treatment] in the receiving country” (paras. 342 and 365). Relatedly, according to the ECtHR, any arguable claim that such a risk exists must be subjected to “close and rigorous scrutiny” by the authorities of the sending State before the removal can take place (para. 387).

Finally, the Court found that transfers to Greece at the material time did breach Art. 3 of the ECHR. Remarkably, it based its finding mainly on the general conditions prevailing in that country, as documented by numerous reports. To the objection that such reports did not disclose any individualized risks for the applicant, the Court replied: “The fact that a large number of asylum seekers in Greece find themselves in the same situation as the applicant does not make the risk concerned any less individual where it is sufficiently real and probable” (para. 359). In other words, the Court affirmed that systemic risks, which had up until then been wrongly dismissed as irrelevant in national litigation on Dublin transfers, may prohibit a removal under Art. 3 of the ECHR  just as well as individualized risks.

Eleven months later the Grand Chamber of the ECJ delivered the NS judgment on identical facts, and essentially confirmed the findings of MSS, however, its emphasis was placed differently. After recalling the principle of mutual trust underpinning the Dublin system, the ECJ conceded that the Common European Asylum System “may, in practice, experience major operational problems in a given Member State, meaning that there is a substantial risk that asylum seekers may, when transferred to that Member State, be treated in a manner incompatible with their fundamental rights,” (para. 81; emphasis added). It then added that not “any infringement of fundamental rights” in the responsible State will “affect the obligations (sic) of the other Member States to comply with the [Dublin Regulation],” (para. 82). The reason offered was that “[a]t issue here is the raison d’être of the European Union and the creation of an [AFSJ] and, in particular, the [CEAS], based on mutual confidence and a presumption of compliance, by other Member States, with EU Law and, in particular, fundamental rights,” (para. 83). The conclusion, a few paragraphs below, was that Member States are precluded from effecting a transfer “where they cannot be unaware that systemic deficiencies in the asylum procedure and in the reception conditions of asylum seekers in that Member State amount to substantial grounds for believing that the asylum seeker would face a real risk of being subjected to inhuman or degrading treatment within the meaning of that provision,” (para. 2 of the operational part, emphasis added).

This “NS test” is open to many interpretations because it is unclear what a “systemic deficiency” is. Opinions range from holding that a “systemic deficiency” is a synonym for “Greek conditions,” the complete failure of the asylum system in all its components, to holding that it is just a failure or lacuna making a risk of ill-treatment “predictable.” I am inclined to think, along with other authors, that any deficiency in the asylum system capable of putting whole categories of persons at risk qualifies as “systemic,” without having to reach “Greek” proportions, but that the ECJ intended to exclude risks that are dependent on individual circumstances. This reading is borne out by the evidentiary requirement that transfers are precluded when Member States “cannot be unaware” of the deficiency, based on dependable field reports enabling them to “assess the functioning of the asylum system” (paras. 90-92, emphasis added). As an Advocate General (AG) of the Court has later summarized: “the evidential standard is set out to the point where it has become notorious that asylum seekers cannot be transferred to the Member State concerned.”

If the above is correct, the “NS test” is much less inclusive than the traditional “Soering test” of the ECtHR. And while in NS the ECJ refrained from saying that “only” risks stemming from systemic deficiencies could preclude a transfer, it took this step in the Abdullahi judgment, ironically delivered on Human Rights Day 2013 (see Abdullahi judgment, para. 60). On Nov. 4 2014 (the anniversary of the ECHR) the Grand Chamber of the ECtHR replied with the Tarakhel judgment. It held that real risks of ill-treatment in the destination State preclude removal whatever the “source of the risk” – systematic or individualized. For good measure, it added that “[the source of the risk] does not exempt [the sending] State from carrying out a thorough and individualised examination of the situation of the person concerned and from suspending enforcement of the removal order should the risk of inhuman or degrading treatment be established” (para. 104).

It is hard to imagine that the ECtHR will back down from its position in the future. The 28 EU Member States make up the majority of the ECHR Contracting parties, and the EU policies linked to the establishment of the AFSJ – immigration and asylum, cooperation in criminal matters, and judicial cooperation in civil matters including family matters – are by far its most human-rights-sensitive. Otherwise stated: the stakes for the integrity and relevance of the ECHR are immense. Opinion 2/13, for its part, suggests that the ECJ is determined to stick to its position even though it is now clear that this means opening a conflict with the ECHR. The question is whether it has good reasons to do so. Is its position sound from the standpoint of human rights, even though it conflicts with established ECtHR case law? Is it in line with the Union’s own constitution? Is it capable of commanding respect from the courts of the Member States? Does it reflect sound judicial policy on the part of the ECJ? I have serious doubts on all counts, and will in the following try to explain why.


Under human rights law, removal is prohibited as soon as it exposes a person to a sufficiently severe consequence (inhuman or degrading treatment) with sufficient likelihood (there has to be a “real risk”). Both “severity” and “likelihood” may depend on system-related circumstances and from individual circumstances in varying proportions. Consider the following “Dublin” situations:

A: Owing to the state of its reception system and detention facilities, asylum seekers sent to Greece were all exposed with near-certainty to life conditions so bad as to amount to inhuman or degrading treatment, regardless of their individual characteristics.

B: Italy has a well-known problem with reception capacities, raising a real possibility that reasonable accommodation will not be found for persons transferred; but that risk might not exist for a happy few having guaranteed access to dignified facilities, and even in the other cases it might depend on individual circumstances (think “healthy young adult” vs. “traumatized child”) whether the anticipated life conditions would actually amount to inhuman or degrading treatment.

C: In other cases still the risk itself may be wholly linked to individual circumstances – e.g. past abuse from a member of the family established in the destination State, or from an organization that had previously trafficked her there and cruelly exploited her. There may still be some “systemic” problems, but these do not necessarily relate to the asylum system as required by the NS test, and they may well escape notice until the individual case brings them to light.

The Soering test covers all such risks. The NS test would leave some out – possibly the “B” cases, quite certainly the “C” cases. But this distinction seems, with due respect, perfectly arbitrary: once it is established that a given course of action will expose a person with sufficient likelihood to degrading treatment, there is no inherent reason why we should protect her if the risk is “systemic,” and withhold protection in other cases. This is the precisely the point made by the ECtHR in Tarakhel (and also, quite vigorously, by the UK Supreme Court in the EM judgment).

For its part, the ECJ has not even tried to argue that the distinction it draws is inherently justified. Rather, it has based its reasoning entirely on public interest: the cases in which Dublin transfers may not be carried out must be “exceptional,” as said in Opinion 2/13; or as AG Jääskinen has put it in Puid, the whole point of the NS test is “establishing a high barrier against the setting aside of the principle of mutual trust.” All such arguments, however, are out of place. The prohibition of inhuman or degrading treatment is absolute. Once a relevant risk is established removals are outlawed and it is completely immaterial whether this is an exceptional occurrence or not, or whether public interest – be it the good functioning of the AFSJ or public security as in Chahal – requires otherwise.

This leads to a second, and closely related, criticism of NS and its progeny. Not only is this case law at odds with basic principles of international human rights law, but it also contradicts just about every foundational statement on human rights to be found in the EU Treaties. As already noted, Art. 6(3) TEU states that “[f]undamental rights, as guaranteed by the [ECHR] […]” constitute “general principles of the Union’s law” (emphasis added). The CFR, which is also part of the EU constitution (Art. 6(1) TEU) and which is binding on all EU institutions and on Member States whenever they implement EU Law, states in Art. 52(2): “In so far as this Charter contains rights which correspond to rights guaranteed by the [ECHR], the meaning and scope of those rights shall be the same as those laid down by the said Convention. This provision shall not prevent Union law providing more extensive protection.” (emphasis added). De lege lata, EU primary law is thus premised on the idea that the minimum standard set by ECHR law – necessarily including the case law of the ECtHR (see Art. 32 ECHR) – is to be respected and cannot be altered by EU authorities, without prejudice to their establishing more protective standards. Art. 78 TFEU adds that the common asylum policy must ensure respect for the principle of non-refoulement and be in compliance with the 1951 Geneva Convention and “other relevant treaties” – including the ECHR. De lege ferenda, Art. 6(2) TEU mandates the accession of the EU to the ECHR – nota bene to the ECHR, not to some watered-down version thereof.

The conclusion that the Treaties require inter-State cooperation untrammeled by the ECHR contradicts all these provisions, and the arguments deployed by the ECJ in support of it – essentially based on the “specific features of EU law” – do not seem particularly weighty or convincing. Such arguments are further weakened by the fact that the ECJ makes no argumentative effort to conciliate the various provisions and principles at stake. No attempt is made, e.g., to show that the EU system requires untrammeled cooperation based on mutual trust, but nonetheless guarantees respect for the ECHR standards through such or such other “special feature.” Rather, what we have is practically a naked claim that the ECHR may safely be disregarded – precisely in situations where the Treaties mandate respect for it.

With this “naked claim,” the ECJ may have effectively blocked accession to the ECHR – that was the main effect of Opinion 2/13. The question nonetheless remains: should the national authorities, i.e. the authorities that actually take and review measures such as Dublin transfers, apply the NS test, or should they stick to Soering? As EU law stands, the answer is relatively straightforward. As the ECJ itself has stressed in Halaf, the sovereignty clause of the Dublin Regulation leaves Member States entirely free to refrain from transfers (see Halaf, para. 36). In such conditions the ECJ could not – without doing extreme violence to the Regulation, and without blatantly disregarding the limits of its own authority – impose to Member States the execution of transfers conflicting with the ECHR (see also recital 32 of the preamble of the Dublin Regulation).

Should push come to shove, and national authorities be placed before a real dilemma between respecting EU Law or respecting ECHR Law, the ECJ might find that its judicial policy is not conducive to the authority and uniform application of EU Law. Already now, the UK Supreme Court has simply disregarded the Abdullahi judgment indirectly criticizing the “unacceptable artificiality” of its conclusions (see EM, paras. 48-50). It would not be surprising either if the statements included in  Opinion 2/13 led the ECtHR to reconsider its Bosphorus conclusion that the EU offers, in general, equivalent protection to human rights. In the future, if faced with a real conflict of obligations, the authorities of the Member States (save France) might remind themselves that their ECHR obligations predate accession to the EU, and that Art. 344 TFEU reserves such pre-existing obligations. While true, Art. 344 TFEU only reserves obligations owed by EU States to non-EU States, not obligations owed by EU States to other EU States (for a hint in this direction, see Opinion 2/13, para. 193), it is far from being clear that such distinctions hold in the case of human rights treaties.

The irony is that, in following its perilous course, the ECJ is striving to place the Union’s AFSJ and CEAS on solid foundations. In fact, it is undermining them, along with human rights protection in the EU and beyond. Consider this: should the NS test eventually supplant the Soering test in the operation of the Dublin system, the ECJ will effectively have built a “systemic deficiency” – a rule producing serial violations of the prohibition of inhuman or degrading treatment – in a CEAS whose supreme goal is to “ensur[e] compliance with the principle of non-refoulement,” (Art. 78 TFEU).

Suggested Citation: Francesco Maiani, Establishing a Common European Asylum System by Leaving European Human Rights Standards Behind: Is this the Way Forward?, RefLaw (Feb 4. 2015),


Submit A Comment

Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>