Thursday 5 December 2019

Pilate washing his hands. The CJEU on pre-trial detention

By Adriano Martufi (Assistant Professor, Leiden University) and Christina Peristeridou (Assistant Professor, Maastricht University)


The very recent ruling of the CJEU in DK (C-653/19 PPU, 28 November 2019) came to verify two quite depressing suspicions about the current status of European criminal law. First, Directive 2016/343 on the presumption of innocence remains an instrument with staggeringly limited applicability especially in the field of pre-trial detention. Second, pre-trial detention stands as a political and legal hot potato: neither the CJEU nor the EU legislator are eager to provide common standards on pre-trial detention, even if the lack of these standards is partly to blame for problems of mutual trust between judicial authorities in the Member States.

The facts

In DK, the Bulgarian Specialised Criminal Court requested a preliminary ruling for the compatibility of the national code of criminal procedure with Article 6 of the Directive on the presumption of innocence, and Articles 6 (right to liberty) and 47 (fair trial and effective remedy) of the EU Charter of Fundamental Rights. In Bulgarian criminal procedure, when the case of an already detained suspect reaches trial, the trial court is responsible to deal with the detention, next to the merits of the case. Once the trial court finds the detention lawful, it becomes indefinite and can only be reviewed on application of the defendant; in such application the defendant must convince the court of changed circumstances that would justify release.

The Bulgarian court wonders whether this national rule shifts the onus from the prosecution to the defendant in providing evidence for release, thereby conflicting with Article 6 of the Directive on the burden of proof, which states that the burden of proof in proving guilt is for the prosecution. DK has been in custody since 11 June 2016 and has filed various applications for release to the trial courts (first and second instances). All of these were rejected since he failed to provide evidence of changed circumstances pointing towards release.

The ruling

The Court, ruling under the urgent procedure, delivered a short and to the point judgement where it found that Article 6 of the Directive on the presumption of innocence and Articles 6 and 47 of the Charter do not apply to the situation at hand (para 42). The syllogism is based on a collection of arguments from the wording of the Directive together with previous jurisprudence. First, the Directive only provides for minimum harmonisation and it does not lay down rules on pre-trial detention exhaustively. Indeed, in previous case law the Court verified that the examination of reasonable suspicion, the evidence used and the judicial reasoning in ordering pre-trial detention remain subject to national law (Milev). Second, a grammatical reading of Articles 4 and 6, and recitals 16 and 22 of the Directive reveals, according to the Court, an implicit distinction between judicial decisions on guilt and other procedural acts such as remand procedures. Following AG Pitruzzella’s Opinion, the argument is that, in contrast with Article 4 (public reference of guilt) which applies only to decisions of pre-trial nature, Article 6 (burden of proof) applies solely to decisions on guilt. Thus, the burden of proof must be borne by the prosecution only for judicial decisions pertaining to a finding of guilt, and not for other decisions of pre-trial nature. Third, since this case falls outside the realm of the Directive – and thus EU law – the Charter is not applicable (Article 51). Articles 6 and 47 of the Charter cannot be invoked ad hoc, as there is no application of EU law.


The Court in DK confirms the European Union legislator’s view on the presumption of innocence being a thin, limited concept. The Directive on the presumption of innocence had little ambition, putting forward a minimalistic presumption. To this extent, DK has no transformative effect: there is no brave expansion of EU law with some sort of judicial activism. The Court upholds the status quo. Yet this mundane result does not sit well with the increasing discomfort of having judicial cooperation without common pre-trial detention standards. Given that the overuse of detention on remand has become increasingly problematic in numerous EU legal systems – even leading to exceptions to the once unyielding mutual trust in Aranyosi – one could hope that this Directive would form a stepping stone to regulate certain aspects of pre-trial detention.

After all, the ECtHR has repeatedly linked the presumption of innocence and pre-trial detention, as the latter may negatively affect the former. In the absence of legislative harmonisation, it would arguably fall to the CJEU to fill the ‘gaps’ of protection left by the EU legislator. DK could offer such opportunity: very recent ECtHR case law has taken aim at practices of shifting the burden of proof within remand proceedings similar to those seen in DK. It would have been quite easy for the CJEU to fall back to ECHR standards, which seemed tailored to the problem raised in DK. Yet the Court did not engage into a carpe diem moment. What is more, the arguments developed by the Luxembourg judges are far from flawless and deserve closer scrutiny.

Interpretation of the Directive

To begin with, the Court held that the situation in DK escapes the scope of the Directive in a simplified and incomplete manner. The Court draws the strength of its arguments from the grammatical reading of the Directive as provided by AG Pitruzzella. According to the latter, while some provisions of the Directive cover trial and pre-trial measures alike, others target exclusively trial measures. This is extracted from references made in Article 4 (public references to guilt) to all ‘judicial decisions’ other the one on guilt. An all-encompassing notion is surely not repeated by Article 6 (burden of proof) and the elucidations on this article provided by the preamble. This grammatical analysis, according to the AG and the Court, points to the conclusion that the two provisions are meant to apply to different stages of criminal process. It is indeed the case that Article 6 speaks of the ‘burden of proof for establishing the guilt’ of the suspect, and not of any burden of proof. Having said that, one could argue that if the legislator had made such distinction between pre-trial and trial stage pertinent for the applicability of these articles, it would have made that more explicit in the text.

In light of the above, it is perplexing that in the previous case of RH the Court dealt with the issue quite differently. There it provided a different reading of Article 6 of the Directive by stating that this provision broadly refers to ‘any obligation on the judge or the competent court to seek both elements of inculpatory and exculpatory evidence’ (para 56 in RH). In doing so, it established a conceptual connection between Articles 4 and 6 of the Directive, implying that Article 6 would be relevant for pre-trial detention. By contrast, the Court in DK tried to nuance this link, in an effort to rule out the relevance of Article 6 for pre-trial detention proceedings. This is however surprising as, in reality, the presumption of innocence does indeed apply to those proceedings under ECHR standards and under most, if not all, national legal systems.

Conclusively, even if the Court’s reading of the Directive is not prime facie implausible, the lack of any further argumentation against extending Article 6 to pre-trial detention leaves much to be desired. All in all, the AG’s Opinion offers a more complete and holistic account to support such a thin understanding of the presumption of innocence. The travaux preparatoires showed that the Commission had insisted on excluding pre-trial detention from this Directive all together (point 33 of Opinion). The Court's reading of the Directive would have appeared less legalistic and more convincing, had it explained its change of heart from RH by referring to the telos of the Directive as intended by the drafters.

The Charter

But the Court was also asked to interpret Articles 6 (right to liberty) and – surprisingly – 47 (effective remedy and fair trial) of the Charter. Based on the facts of the case, one would have expected the Bulgarian court to request the interpretation of Article 48 (presumption of innocence). It is not apparent why the right to an effective remedy (or generally to a fair trial) is relevant here, although one may argue that the shift of the burden of proof described above challenges the effectiveness of the remedy: being in custody, DK could not easily provide new circumstances to challenge his detention. Having said that, an (additional) argument including the presumption of innocence (Article 48 Charter) would have been stronger. This is confirmed by the Opinion, as AG Pitruzzella groups Articles 47 and 48 together although no reference to Article 48 was made by the referring court.

In spite of these inconsistencies, the Court’s way of dealing with the interpretation of the Charter is bizarrely introvert. With a few laconic sentences (para 40-41), it dismisses the national court's requests stating that the situation at hand did not fall under EU law (the Directive) and therefore the Charter's safeguards could not be consistently triggered. Relying on Article 51, the Court aseptically reiterates that the Charter comes into play only when national authorities are ‘implementing EU law’.

Yet, the Court has taken the view in the past that the Charter could be applicable not only when EU law is implemented, but also when a more tenuous connection exists. In our case, whereas pre-trial detention is not exhaustively regulated by the Directive, some parts are indeed affected (Article 4 prohibits references to guilt in pre-trial detention orders). Thus, a connection with EU law does exist, and there could be an argument in favour of using further the Charter following Åkeberg Fransson. Furthermore, the Charter has been used in the past to cover lacunas. In Aranyosi, while the EAW Framework Decision did not provide any ground for refusal on grounds of (potential) fundamental rights' violations, the Court used the Charter to fill up this lacuna and substantially created one.

In Aranyosi, the Court followed a more principled approach, even going against a literal or teleological interpretation of the legal instrument in question. Remarkably, this happened even though the Luxembourg judges were under huge pressure to uphold mutual trust and maintain the EAW procedure unaltered. Why didn’t the Court do the same in DK?

Notably, we are not the only ones to raise this question. AG Pitruzzella himself criticises heavily (and in a rather strong tone) the lack of common standards for pre-trial detention and urges the EU legislator to move forward with this subject (points 20-22). Quite interestingly, however, he concedes that he has ‘no choice’ but to conclude that the facts in DK fall outside Union law. Similar remarks were expressed by the Opinion in Milev, with AG Wathelet disagreeing strongly with the Commission’s view that the Directive did not include any substantial provisions for pre-trial detention (points 55-57).

Explanation of Court’s approach

So why didn’t the Court extend Article 6 application on the burden of proof to pre-trial detention, following an application of the Charter?

One plausible explanation is that the Charter is not automatically conducive to an expansion of the protective scope of this Directive. The Court has reiterated in Milev (para 47) and in DK that harmonisation is not exhaustive. In this sense, this Directive provides for a partial (and minimum) harmonisation as only certain aspects of the presumption of innocence are dealt with. Thus, in this case the Charter cannot expand the scope of the harmonisation more than the way the Directive already defines it. In constitutional terms, one could argue that the Court may have been wary to step beyond the red lines set by the principles of subsidiarity and conferral. Of course, the Luxembourg judges could easily shield themselves behind the inherent vagueness of the presumption of innocence. This is a notoriously elusive concept, poorly implemented in practice and theoretically divisive. Some national laws hardly go any further that a statement of principle and, despite sincere and sophisticated efforts, scholarship is deeply divided on its meaning. In this sense, the EU legislator was quite brave to consider touching upon it. In our view, the Directive would have enjoyed more success and less uncertainty, had the harmonisation of the presumption of innocence been complete.

A second explanation for not following a principled approach may derive from a general unwillingness to engage in judicial activism, especially in light of the delicate legal issue underlying the preliminary ruling. Pre-trial detention is a controversial topic, as the debate on its harmonisation at EU level clearly reveals. On multiple occasions, the EU has indicated that it may adopt minimum rules in these matters (Green Paper, EP Resolution), and yet so far the reaction by Member States has been lukewarm. Pre-trial detention is regarded as being incredibly diverse and therefore complicated to harmonise, so that EU legislator has chosen to focus on the less troubling alternative measures (with the so-called ESO Framework Decision).

Arguably, in the case at hand, the Court is reluctant to touch upon pre-trial detention given the strong opposition of both the Commission and the Council to include this within the scope of the Directive. But pre-trial detention does have a strong human component that is difficult to ignore: before excluding the situation at hand from Union law, AG Pitruzzella half-heartedly admits that the defendants in the domestic proceeding had as only option to bring their case before the ECtHR – a scenario which, as the AG himself recognises, may take years to materialise (point 21). This is a simple but powerful remark, shifting the perspective from the European courts, with their intricacies and conflicts, to the one of the defendant. The question – implicitly but ever so powerfully posed – is whether or not the time has come for the EU to address pre-trial detention as an EU matter.

A third explanation invites us to look at DK in the light of the prior case law concerning references issued by Bulgarian courts. Before DK, two more cases were raised by the same court, challenging the compatibility of domestic pre-trial detention regime with EU law. Both in Milev and in RH, the Bulgarian court requested clarification for the same national legislation and its relation to the Directive. As explained by AG Wathelet in Milev, the pre-trial detention regime had been amended due to ECtHR developments but the status quo divided the judges. In RH, the Bulgarian court even revealed some details of the internal judicial struggles amongst the Bulgarian Supreme Court and the lower courts: the former had ordered the latter not to wait for the response of the preliminary reference procedure, in order to rule on the pending status of the detention ‘within reasonable time’. Due to the defiance of the lower court, even disciplinary proceedings were launched. This occurrence was even put forward as a question to the CJEU regarding judicial independence (a rather fashionable topic these days), with regard to the power to prevent lower courts from waiting for the CJEU’s response in preliminary reference procedure. With this background in mind, we could see DK as a request for the CJEU to play the referee for a national debate. This may provide further leads to understand the Court's reluctance to rule on this topic.

But leaving aside the domestic quarrels between Bulgarian courts, the fact remains that a struggle is taking place in that country to uphold human rights standards and enhance their level of protection in criminal proceedings. It is only natural that the Directive on the presumption of innocence – especially in light of the wider interpretation given in RH – would be used by Bulgarian judges as a breeding ground to achieve that result. DK could in fact be the opportunity for the CJEU to exercise its role as ‘competence regulator’ – aiguilleur des compétences (point 21 Opinion). Referring to Vedel’s theory of constitutional control, AG Pitruzzella advised the Court to seize this opportunity and give an indication to the national authorities of the right path for reforming their national procedure on pre-trial detention. As has become apparent, the Court did not follow his advice.

DK might have the effect of discouraging national courts to follow up with more questions on the application of the Directive in relation to pre-trial detention. Yet it is a wonderful addition to previous case law, demonstrating that pre-trial detention is increasingly becoming a candidate for intervention by the EU; the question of harmonising pre-trial detention cannot be ignored much longer if national courts keep asking the CJEU to intervene.

Barnard & Peers: chapter 9, chapter 25
JHA4: chapter II:4
Art credit: Jan Lieven, via Wikicommons 


  1. I see you think Mr Johnson cancelling an outdoor speech because of four protesters is the height of hilarity when everyone outside the ever decreasing Corbynite bubble sees that he doesn't want the whole thing to degenerate into a playground 'na na na my Dad's bigger than yours' argument. Because that's why the protesters are doing it, like children having a tantrum they want to be noticed, not because they actually care.
    One more thing : reCAPCHA is hideously ablest.

    1. This comment has nothing to do with the blog post, which was written by other people - so I'll only do this once. If you object to something I tweeted, reply on Twitter. Your objection to peaceful protest and attempt to discredit freedom of assembly is disturbingly authoritarian.

  2. What does "DK" means? Its Denmark?

    1. No, it's the name of the case. The CJEU has been redacting the names of natural persons from its judgments for a few months now.

    2. Interesting aside there however - the way that supranationals use language.

      'Movement of Natural Persons,' 'Provide Services,' 'mode of service delivery' 'transposition.' It is, I suspect, a very careful avoidance of the language of workers, labour, employment, immigration, supremacy and the like.

      These supranationals have taken on lives of their own and developed a very careful language around the politics of their situation. It's a micro example of the constitutional deficit we see with supranationals.

      Indeed quite why the ECJ withholds names is anyone's guess.

    3. It's not anyonean guess; it was explained by the CJEU as its interpretation of data protection law, and was criticized on this blog.

      The first three bits of jargon come from the WTO, not EU law. The WTO is intergovernmental, not supranational. And again, this has nothing to do with the blog post. Quite embarrassing that people use it as a forum for random comments.

    4. Not *anyone's* guess, that is.

    5. Also, WTO services rules are not about long-term immigration. At least make some minimal effort to understand something before you pontificate about it.