Tuesday, 13 November 2018

CJEU case law on EU citizenship: normatively consistent? Unlikely! - A response to Davies’ ‘Has the Court changed, or have the cases?’

Alexander Hoogenboom, PhD, MSc. LL.M. Senior Policy Officer at the Dutch Healthcare Authority and associate researcher at the Institute for Transnational and Euregional cross border cooperation and Mobility, Faculty of law, Maastricht University. The position taken in this paper solely reflects the views of the author.


Recent case law of the Court of Justice on EU citizens’ access to benefits has been seen by some as a restrictive turn compared to prior case law, in response to a rise in populism. However, the article by Davies in a recent special issue of the Journal of European Public Policy is to be commended for its original take on this alleged ‘turn to restrictiveness’. The goal of his article is, as I see it, questioning whether the Court has indeed recently become stricter (in the sense of more State-friendly, less Union citizen-friendly) in response to the populist turn in the European political landscape. In that vein, Davies submits, contrary to what he sees is the main thrust in the scholarship, that the court has been ‘normatively consistent’ (see also this research paper he authored) and that the perceived difference in recent litigation outcomes from the golden years of Union citizenship are due to the litigants being less ‘deserving’ of access to benefits provided by the host Member State: ‘what goes in will provide an overwhelmingly plausible explanation of the outcomes on its own’.

This argument is supported by relying on a methodology for measuring the ‘deservedness’ of the litigants based on a set of indicators: the ‘good behaviour’ of the applicant, the possibility for exceptional harm should the benefit be denied, the possible cost to society resulting from the grant of a benefit (e.g. because it is long-term or applies to a large category of persons), whether granting the benefit would mandate a particular positive outcome and whether the State was somehow at fault for the specific conundrum that the litigant finds him or herself in.  The more ‘deserving’ (high contribution to society, low cost), the greater the likelihood of a litigant-positive result.

Applying these criteria to a selection of case law, Davies proceeds to show that indeed, the Dano’s (never worked, never integrated, lacked resources for self-support) of the world seem to be less deserving of benefits than the Sala’s (long-term legal resident, child benefit denied due to a technicality by an inconsistent state). (On the Dano case – the first judgment showing the perceived turn toward a stricter approach – see the discussion here).

There are, however, a few issues one could take with this approach. In part, it is questionable whether the facts as available to scholars of EU law decisions allow one to accurately apply the proposed test – especially given that the publication of the Reports for the Hearing was abolished since 2012 (see also para 119 of the AG opinion in Breyer). This is exacerbated, as Davies admits, by the fact that a negative outcome throws its shadow forwards: one then tends to present the facts in such a way as to support the later conclusion.

In part, it is questionable whether deservedness is objectively verifiable: after all, one could also argue that the rise of populism or say an economic crisis – the ‘times we live in’ – simply provide a different kind of lens through which to view the migratory Union citizen. The same fresh-faced youth enrolling in higher education in a Member State different from the one of his nationality can be cast as a self-improving, future productive member of that society (compare to Gravier, para 24) or as a welfare-abusing, locust-like creature eating his or her fill and subsequently returning from whence he or she came (see the main arguments by Belgium and Austria in the the judgment in Bressol).

However, the main challenge I would like to level is that the methodology suggested is applied to an incomplete ‘data set’ (the cases), in part due to selection bias (the focus in the article on the cases ‘most discussed’).

Grzelczyk by another name: The curious case of Förster

The case of Grzelczyk is among those used by Davies to support his thesis: his particular circumstances ‘paint a sympathetic picture’. A hard-working young Frenchman, resident in Belgium for some time and whose claim to financial support finish his last few months of study would seem altogether reasonable. Indeed, the Court, while leaving it to the national court to make the final decision, seemed to suggest he should be so entitled.

Nonetheless, the Court was not so generous in Förster. Jacqueline Förster was a German national who grew up in a town not far from the Dutch border. She moved to and resided in the Netherlands from March 2000 onwards – partly to be with her Dutch boyfriend. There, she enrolled into a teacher training programme and later in a course on educational theory at the Hogeschool van Amsterdam. In the period 2000-2002 she undertook various part-time jobs, until she engaged in full-time paid practical training at a Dutch school providing secondary education for children with special needs (October 2002 – June 2003). After her practical training she did not engage in gainful employment until July 2004. She graduated from her course that summer.

At stake was the intermezzo period: the Dutch Student Benefit Authority (then called IB-Groep) assessed in 2005 that she had not been eligible for study maintenance assistance in the second half of 2003 and ordered her to repay the amounts received.

It would seem to me that Ms. Förster ticked practically the same as mr. Grzelczyk. Measuring her ‘deservedness’ according to the matrix suggested by Davies we find:

-       Good behaviour: self-support initially (three years), enrolled in studies and contributing to Dutch society even during her studies in her practical training period. Advocate-General Mazàk moreover made the explicit point that she did not seem to have moved with the goal of claiming benefits.
-       Exceptional harm: Although the withdrawal of the benefit was retroactive, when assessing her right to access the benefit as matters stood in 2003 one could plausibly maintain that, as with Grzelczyk, non-access to the benefit would have made the completion of the degree much more difficult.
-       The support requested covered only a limited time: at issue was a six-month period only.
-       Limited cost of the support: As with Grzelczyk, this aspect is hard to estimate.
-       Positive outcomes: the completion of the degree would enable her to contribute to Dutch society and economy as a teacher.
-       State at fault: One could make the argument that given the fact that they had initially granted her the benefit and only much later (some one and a half years) took the final decision to recover the amount, should count against the Dutch state.

Finally, in addition to these elements from Davies’ deservedness matrix, the applicant had a relationship with a Dutch national and given her work, presumably, spoke fluent Dutch – both elements that the Court in other cases has found relevant when assessing eligibility to benefits (see Prinz and Seeberger, for instance).

All in all, one would assume, from the methodology applied that this applicant should be successful, or that should have led to a ‘discretionary-result-with-a-hint-in-favour-of-the applicant’. Yet, she categorically lost her case, notwithstanding the suggestion by the Advocate-General to consider the circumstances of the case along the lines suggested above. Yes, the conclusion was (partly) mandated by the EU citizens’ Directive 2004/38, but the Court certainly did not ‘bend over backwards to find exceptions to [this] restriction’ as his matrix would predict.

The cases of Commission v Austria and Commission v the Netherlands are perhaps even more glaring. The issue concerned a travel benefit aimed at students attending higher education. In the Austrian case, decided in 2012, Austria argued that it could refuse access to the benefit where the Union citizen in question had not yet obtained a right to permanent residence (a five-year prior residence requirement in practice). In contrast, the Court found that all persons enrolled in higher education should have automatic access.

The case is somewhat hard to fit into the matrix, but intuitively one could argue that the Court was not swayed by the hypothetical good behaviour of potential beneficiaries (no prior residence requirements, no integration), nor was the benefit for a limited time (it could last the entire study period), all persons enrolled in higher education persons were eligible (no ‘limited costs’) and it is hard to imagine an ‘exceptional harm’ had the benefit been denied to the hypothetical applicant. Notwithstanding this apparent lack of deservedness, however, a positive outcome for the hypothetical beneficiaries.

In 2016, in Commission v the Netherlands, the Court reached the opposite conclusion. The benefit at issue was in all respects the same as at stake in Commission v Austria (Table taken from A. Hoogenboom, Balancing Student Mobility Rights and National Higher Education Autonomy in the European Union (BRILL, 2017)):

The Netherlands
Nature of the benefit
Fee reduction for the use of public transport (grant)
Free use of public transport on some days of the week; reduced fees on the other days (conditional grant)
Apparent purpose
To facilitate access to education
To facilitate access to education
Commercial public transport operators
Commercial public transport operators
The individual governments of several Länder
Central Dutch government
Enrolment higher education
Parents in receipt of Austrian family allowances
Enrolment in higher education.
Student in receipt of Dutch studiefinanciering

Here the Court decided that the Netherlands could restrict the benefit along the terms suggested by Austria in the earlier case. It stretched the limits of consistency by trying to distinguish the latter case with a bizarre reference to the difference in national classification of the benefit, despite it being settled case law that such classifications are irrelevant - a point the Court itself made in Commission v Austria.

So here we initially have a judgment against the State and in favour of the Union citizen, whereas the matrix would likely predict a judgment in favour of the State. Four years later the Court came, on largely the same facts, to the opposite conclusion. This should not be seen as a (late) vindication of the matrix however, since the Court formally distinguished the latter case from the former, meaning that both cases continue to be good law.


Whereas the line of inquiry presented in Davies’ paper is an interesting one, it cannot account adequately account for the existence of Förster and the Commission v Austria/Netherlands saga. Apparently deserving applicants get rejected, and the same benefit with a similar hypothetical user base gets a different legal treatment in the space of a few years. Two possible explanations suggest themselves:

-       The Court is not consistent. This would mean that Davies’ hypothesis is disproven.
-       The matrix presented, by which the court supposedly measures deservedness, is flawed. This goes to the methodology used and means that we simply haven’t found calculus that the Court uses.

It would seem to me that the existence of both Commission v Austria and Commission v the Netherlands as ‘good law’ seems to hint at the former. In any case, however, to argue that the Court of Justice is normatively consistent, on the basis of the analysis carried out by Davies, is a bridge too far.

Barnard & Peers: chapter 13
Photo credit: i newspaper

Monday, 12 November 2018

The effectiveness of the right to family reunion of refugees: An analysis of the K. & B. ruling of the CJEU

Mark Klaassen, Assistant professor at the Institute of Immigration Law at Leiden University


For most EU Member States (the UK, Ireland and Denmark have opted out), family reunion between non-EU citizens residing on their territory and their non-EU citizen family members living abroad is governed by Directive 2003/86. This Directive includes a more favourable regime for refugees seeking family reunion, which Member States may limit to family relationships which predate entry to the Member State. These exceptions include the waiver of a number of conditions for family reunion: the optional derogation for those over 12; the conditions relating to accommodation, sickness insurance and “stable and regular resources”; and the waiting period. The rules on evidence of family relationships are also relaxed, in the event that documentary evidence is unavailable. Finally, the optional “integration measures” requirement can only be applied after family reunion, whether the family relationship existed before entry or not. 

However, Member States may limit the application of this more favourable regime as regards the waiver of the conditions relating to accommodation, sickness insurance and stable and regular resources if the sponsor or family members have “special links” with a non-EU country where family reunion is possible, or if the application for family reunification was lodged within a period of three months after refugee status has been granted. This latter competence is laid down in the last sub-paragraph of Article 12(1) of the Directive.

But how must the Member States deal with applications for family reunification by beneficiaries of international protection which are lodged after this three months period? This is essentially the question in Case C-380/17 K. & B., that was submitted to the Court of Justice of the EU (the Court) for a preliminary ruling by the Council of State – the Dutch court of highest instance in immigration matters.

In this blog I will first sketch the implementation of Article 12(1) of the Directive in the Netherlands. Then I will discuss the cases that have led to the preliminary reference. After that I will summarise the ruling of the Court. Finally I will discuss the implications of the ruling for the family reunification of refugees. Note that the judgment also confirmed the Court’s jurisdiction to interpret the Directive for the family reunion of persons with subsidiary protection in the Netherlands, even though they are excluded from the scope of the Directive, because the Netherlands has chosen to apply the Directive to them in the same way as refugees. I will not elaborate on the jurisdiction issue further.

Family reunification of refugees in the Netherlands

In the Netherlands there is a one-status system, meaning that no distinction is made between refugee status and subsidiary protection status. This has implications for the implementation of the Directive. Beneficiaries of subsidiary protection are excluded from the Directive. However, as no distinction between refugees and beneficiaries of subsidiary protection is made in Dutch asylum law, all beneficiaries of international protection are eligible for family reunification pursuant to the implementation of the Directive. From the moment that international protection is granted, the sponsor has three months to make the application for family reunification of family members that were left behind in the country of origin. In case the application is made within this three month period, no substantive requirements for family reunification are imposed, as is required by Article 12(1) Directive. Also, no administrative fee is charged for making the application.

When the application is made after the three months period has passed, the application is rejected. Only in very exceptional circumstances is an application lodged outside of the three months period accepted – even though there is no legal basis for that in Dutch law. In making the determination whether such special circumstances apply, the Dutch authorities neither take the best interests of the child (Article 5(5) Directive) nor the specific individual circumstances of the case (Article 17 Directive) into account. A sponsor with international protection status is eligible to submit a “regular” application for family reunification outside the three month period. In regular applications, the substantive requirements – which are based on Article 7(1) of the Directive – are fully applicable.

The applicants in K. and B.

Applicant K. is the Afghan mother of an unaccompanied minor in the Netherlands. The minor sponsor has been granted international protection in the Netherlands on 14 February 2012. As an unaccompanied minor, the legal guardian was assigned to the sponsor. The sponsor applied to be reunited with his mother on 27 November 2013. From the order of reference, it appears that his legal guardian did not inform the applicant of the possibility to apply for family reunion. The application was rejected because it was not made within the three month period after being granted international protection.

An appeal was lodged with the District Court of Zwolle, which ruled in favour of the applicant. Referring to the best interests of the child as laid down in Article 3(1) of the UN Convention on the Rights of the Child and the case law of the European Court of Human Rights on Article 8 ECHR (the right to private and family life), the District Court held that the Secretary was bound to make an individual assessment of the competing interests involved in the case. The Secretary appealed the decision of the District Court to the Council of State, which is the referring court in this case. In the order of reference to the Court, the Council of State blames the sponsor – who was fifteen years old on the moment the application was lodged – for not informing himself of the three months’ period in which the application for family reunification must be made. The Council of State deems that it was reasonable to attribute the lack of information provided by the legal guardian to the minor sponsor.

Applicant B. is the baby daughter of an Eritrean national who was granted international protection in the Netherlands on 23 September 2014. The sponsor applied to be reunited with his wife and baby daughter on 22 January 2015. The application was lodged one month after the three month period had elapsed. The applicant appealed the rejection of his application to the District Court of Amsterdam. His appeal was rejected on 24 June 2016, upon which he lodged a subsequent appeal to the Council of State. The reason for the late submission of the application was a misunderstanding between the sponsor and the Dutch Refugee Council, which has a formalised role in assisting beneficiaries of international protection in making an application for family reunification. He misinterpreted a conversation with the Dutch Refugee Council. In the order for reference to the Court, the Council of State attributes the misunderstanding to the applicant.

These two cases are no exceptions. In 2014 the Advisory Committee on Migrations Affairs – an independent Committee that advises the Dutch Government and Parliament on immigration law and policy – expressed the view that the absolute character of the three months’ period is problematic. The Committee held that it is important that beneficiaries of international protection are swiftly reunited with their family members, also in view of the integration in Dutch society. The Committee advised the Dutch government to proactively ask each beneficiary of international protection whether they want to apply for family reunion and to introduce a hardship clause for situations in which an application is made outside the three months’ period.

The ruling of the Court

The Court rules that the EU legislature has explicitly allowed the Member States to apply the conditions from Article 7(1) Directive to refugees who apply for family reunification after the time limit from Article 12(1) has elapsed (para 46). Based on that, the Member States are allowed not to process applications made after the three months’ period has elapsed under the more favourable rules for refugees but under the general rules instead (para. 47).

According to the Court, the EU legislature did not regulate on the procedure to be followed regarding out of time applications (para. 55). Therefore, the principle of procedural autonomy applies, which is limited by the general principles of equivalence and effectiveness (para. 56). The latter principle requires that the national procedure may not render the rights conferred by the Directive impossible in practice or excessively difficult. Concerning the application of the three month period, the Court finds that rejecting an application for not being made within the three month period does not per se render the exercise of the right to family reunification impossible in practice or excessively difficult when a fresh application can be lodged in which the conditions of Article 7(1) are applicable (para. 59). In this manner, the right to family reunification can still be safeguarded (para. 60). The Court notes that this is not the case in situations in which the late submission of the application is objectively excusable (para. 62).

In this regard, the Member States must fully inform the applicant of the consequences of rejecting the application for non-compliance with the three month period. Besides that, the Member State must inform the applicant how the right to family reunification can nevertheless be effectively realised (para. 63). The Court concludes by remarking that even though the Member States are allowed to impose the requirements of Article 7(1) Directive in case an application is lodged after the three months’ period has elapsed, the other provisions of the Chapter on family reunification of refugees still apply (para. 64-65).

Analysis: A pyrrhic victory for the Dutch government?

The ruling of the Court means that the distinction that is made in Dutch immigration law between the family reunification of beneficiaries of international protection and “regular” family reunification can be maintained. However, in the application of the “regular” family reunification regime, the fact that a sponsor has been granted international protection must be taken into account (para. 53). In the refugee family reunification procedure, there must be a codification of the requirement to not apply the three months’ period in cases in which it was objectively excusable that the application was lodged out of time.

This is the second ruling of the Court on family reunification of refugees in the Netherlands. In the A. & S. judgment (discussed here), the Court held that an unaccompanied minor who is considered a minor on the moment that he lodged the application for international protection has the right to family reunification with family members that were left behind in the country of origin. Essentially that case also considered the procedural autonomy of the Member States to regulate where the Directive does not provide for procedural rules. The Court held that to make the right to family reunification dependent on the moment on which the domestic authorities grant international protection would call the effectiveness of the protection provided for by the Directive into question (A. & S., para 55).

It is also in line with the interpretation of the Court with other provisions of the Directive. In Chakroun, the Court held that the competence of the Member State to require the sponsor to demonstrate stable, regular and sufficient resources may not be used in a manner which would undermine the effectiveness of the right to family reunification (Chakroun, para. 43). The Court confirmed this position in O., S. & L., in which the Court strongly emphasised the role of fundamental rights in determining the competence of the member states in imposing the requirement of stable, regular and sufficient resources (O., S. & L., para. 80). In that case, the Court held that “[i]t is for the competent national authorities, when implementing Directive 2003/86 and examining applications for family reunification, to make a balanced and reasonable assessment of all the interests in play, taking particular account of the interests of the children concerned. (para. 81)”  

The K. & B. judgment has demonstrated that even when an application was made outside the three month period laid down in Article 12(1) Directive, the effectiveness of the right to family reunification of a person with international protection may not be undermined by the application of the conditions of Article 7(1) Directive. This conclusion might sound obvious, but nevertheless has far-reaching implications.

To what extent can it be expected from the applicants in this case to comply with the stable and regular sufficient resources requirement? In order to answer this question, it is interesting to take a look at the drafting history of the Directive. In the initial proposal of the European Commission, refugees were always exempted from the substantive conditions, as “[r]efugees and persons enjoying subsidiary protection, given the overriding reasons why they have had to flee their country of origin and cannot lead a normal family life, cannot be subjected to the same additional conditions without their right lead a family life being imperilledCOM/1999/638 final. Even though the EU’s qualification Directive gives them a right to seek employment, beneficiaries of international protection start with an enormous backlog on the labour market, so requiring them to comply with the requirement of stable and regular sufficient resources could undermine the effectiveness of the right to family reunification. In this context, the reasoning of the Court in O., S. & L. that “a balanced and reasonable assessment of all the interests in play” must be conducted, is of utmost importance.

It must be noted that this assessment is different from the balancing of interests under the ECtHR case law on Article 8 ECHR, as the Directive grants a subjective right to family reunification. The question which must be assessed it whether imposing the stable and regular sufficient resources requirement would undermine the effectiveness of the Directive. The question is not (as under the ECtHR case law) whether a fair balance has been struck between the individual interest and the interests of the state to control immigration.

Lastly, as noted above, Article 12(1) allows the Member States to impose the conditions of Article 7(1) in case an application is lodged after the three month period has elapsed. This does not include the requirement to comply with integration measures from Article 7(2) Directive. This means that refugees can never be subjected to the requirement to pass a pre-entry integration exam in the country of origin, also in cases in which the application is lodged outside the three month period.

Barnard & Peers: chapter 26
JHA4: chapter I:6
Photo credit: AP/Boris Grdanoski

Sunday, 11 November 2018

You can teach a new court Mangold tricks – the horizontal effect of the Charter right to paid annual leave

Filippo Fontanelli, Senior Lecturer in International Economic Law, University of Edinburgh

On 6 November 2018, the Court of Justice of the European Union (the Court) delivered three judgments relating to paid annual leave. Some of its remarks transcend the specific topic, and touch on constitutional matters: the impact of the EU Charter of Fundamental Rights on private parties, the allocation of competences between the EU and the Member States, and the application of EU secondary law.

These cases teach a couple of lessons, and raise a wider point. The lessons are useful: first, alongside the right to non-discrimination and effective judicial protection, the right to annual paid leave in the Charter has direct effect not only in vertical disputes (ie disputes between the individual and the State), but also in horizontal disputes (ie disputes between individuals), even though EU Directives dealing with the same issue still do not in themselves have such horizontal direct effect.  Second, the Charter sometimes binds State acts in a decisive manner. This occurs rarely and, curiously, so far exclusively in disputes between individuals, when the domestic law transposing a directive is not in good order.

The wider point concerns the Charter’s application to domestic measures. In these cases, EU secondary law could not displace domestic measures, because Germany had not transposed a directive correctly. The directive’s mere existence, however, warranted the Charter’s application, and in turn enabled domestic courts to disregard German law and enforce the right to annual paid leave.

The now familiar combo “unimplemented directive plus Charter right” seems a Munchhausen trick. To justify the practice, one should go back to the Mangold case (in which the CJEU ruled that, prior to the Charter having binding force, the general principles of EU law meant that the right to non-discrimination could apply between private parties even if a Member State had not implemented a Directive), and perhaps look deeper into the difference between direct applicability and direct effect, or between a norm’s application and its scope of application. After looking back and looking in-depth, it is still difficult to see precisely how the Charter applied.

The judgments

In cases C-619/16 and C-684/16 Kreuziger and Max Planck, the facts were comparable. Messrs Kreuziger and Shimizu, respectively employed by the Land of Berlin and the Max Planck Institute, had failed to take the entire period of paid annual leave to which they were entitled. After their employment ended, the former employers denied their request to receive payment in lieu of leave. German law appeared to authorise the employers’ position that a failure to request paid annual leave automatically entailed its lapse upon termination.

The Court disposed of the Kreuziger case quickly, due to the dispute’s vertical nature (the employer was a German Land). Article 7 of Directive 2003/88 (the working time Directive) clearly confers the right to paid annual leave and, accordingly, payment in lieu for the leave not taken; the Court recently ruled on this issue in Bollacke. The Court dusted off the direct effect spiel. Since “provisions of a directive that are unconditional and sufficiently precise may be relied upon by individuals, in particular against a Member State and all the organs of its administration” [21], individuals can invoke the Directive and the judges must set aside domestic law if need be. The automatic lapse of this right upon termination of employment, without any safeguard to make sure that the employee could exercise it beforehand, violated EU law.

The reasoning of the Max Planck ruling extended further, since the main proceedings concerned a dispute between private parties. The reasoning of Kreuziger was copy-pasted: Article 7 of the Directive precludes an automatic lapse of the right to paid leave triggered by a mere failure to exercise it [40]. National courts must arrive, to the extent possible, at an interpretation of domestic law consistent with the Directive.

Failing all attempts at consistent interpretation, Mr Shimizu could not rely on the Directive alone, since “Max Planck had to be considered an individual” [65], and directives normally lack horizontal direct effect [68]. The Court thus turned to Article 31(2) of the Charter on the right to annual paid leave, noting that it entailed a clear, enforceable right, withstanding only derogations in compliance with Article 52(1) of the Charter [73]. Article 31(2) of the Charter needing no implementing act to operate, it could warrant disapplication of contrary domestic law:

74           The right to a period of paid annual leave, affirmed for every worker by Article 31(2) of the Charter, is thus, as regards its very existence, both mandatory and unconditional in nature, the unconditional nature not needing to be given concrete expression by the provisions of EU or national law, which are only required to specify the exact duration of annual leave and, where appropriate, certain conditions for the exercise of that right. It follows that that provision is sufficient in itself to confer on workers a right that they may actually rely on in disputes between them and their employer in a field covered by EU law and therefore falling within the scope of the Charter.

75           Article 31(2) of the Charter therefore entails, in particular, as regards the situations falling within the scope thereof, that the national court must disapply national legislation negating the principle [that the right to paid leave and/or payment in lieu cannot lapse automatically].

The last hurdle for the direct invocability of Article 31(2) of the Charter in German courts was the dispute’s horizontal nature. The Court recycled from its own case-law the bold suggestion that Article 51(1) of the Charter, which sets out who is bound by the Charter, is ambiguous in this respect and does not preclude individuals from relying on the Charter against each other:

76           … although Article 51(1) of the Charter states that the provisions thereof are addressed to the institutions, bodies, offices and agencies of the European Union … and to the Member States only when they are implementing EU law, Article 51(1) does not, however, address the question whether those individuals may, where appropriate, be directly required to comply with certain provisions of the Charter and cannot, accordingly, be interpreted as meaning that it would systematically preclude such a possibility.

Notably, the Court distinguishes Article 27 of the Charter on workers’ consultation (at issue in AMS) from Article 31(2) on annual leave, because the former Charter Article refers to national and EU law limits and the latter Charter Article does not. If this is the test for whether Charter provisions can have horizontal direct effect, it should be noted that most of the Charter provisions on social rights refer to national and EU limits – but most of the other provisions of the Charter do not.

73      By providing, in mandatory terms, that ‘every worker’ has ‘the right’ ‘to an annual period of paid leave’ — like, for example, Article 27 of the Charter which led to the judgment of 15 January 2014, Association de médiation sociale (C‑176/12, EU:C:2014:2) — without referring in particular in that regard to the ‘cases’ and ‘conditions provided for by Union law and national laws and practices’, Article 31(2) of the Charter, reflects the essential principle of EU social law from which there may be derogations only in compliance with the strict conditions laid down in Article 52(1) of the Charter and, in particular, the fundamental right to paid annual leave.

The Joined Cases C-596/16 and C-570/16 (Wuppertal v Bauer; Willmeroth v Broßonn) largely replicated the reasoning of the Max Planck and Kreuziger cases. Only, in the underlying disputes, it was rather the employees’ heirs seeking payment in lieu, on behalf of the deceased workers. According to the referring judge, German law provided that the right to unpaid annual leave, necessarily turned into payment in lieu upon the workers’ death, would not become part of their estates.

Building on prior case law (discussed here), the Court, besides noting the importance of the right under Article 7 of the Directive, stated that it should accrue, after the death, to the worker’s estate:

48      … from a financial perspective, the right to paid annual leave acquired by a worker is purely pecuniary in nature and, as such, is therefore intended to become part of the relevant person’s assets, as a result of which the latter’s death cannot retrospectively deprive his estate and, accordingly, those to whom it is to be transferred by way of inheritance, from the effective enjoyment of the financial aspect of the right to paid annual leave.

With respect to horizontal disputes (like the Willmeroth v Broßonn controversy, whereas Bauer was a public employee), the reasoning of the Court was identical to that used in the Max Planck case [87-91]. Article 31(2) of the Charter can be invoked in disputes between individuals, possibly leading to the setting aside of domestic norms like those at issue in the main proceedings.

The Charter’s effect on domestic measures

The Charter has many functions. It guides the interpretation of EU law and serves as standard of legality of EU acts. It does not enlarge the competences of the EU at the expense of the member states, but binds their action when they act as agents of the EU. In this residual scenario, delimited by the sibylline “implement[ation of] EU law” notion of Article 51(1) of the Charter, the Charter should serve as standard of EU-legality of national measures. In other words, the Charter can preclude some national measures.

The Charter applying to state measures falling “within the scope of EU law,” (a formula sanctified in Fransson [21], and unsurpassably frustrating), the following circumstances can arise:

a)      EU law does not apply to the matter, so neither does the Charter;
b)      EU law applies to the matter, and precludes the domestic measure. The Charter applies too: it might also preclude the domestic measure (“double preclusion”) or not;
c)       EU law applies to the domestic measure, but does not preclude it. The Charter applies too, and likewise does not preclude it;
d)      EU law applies to the domestic measure, without precluding it. The Charter, which also applies, precludes the domestic measure.

In the scenarios a), b) and c), the Charter is irrelevant to the ultimate determination of EU-legality.

Only in scenario d) does the Charter show its teeth, doing its standard of review job fully. Exclusively in this scenario, an otherwise EU law-compliant measure can breach the Charter and, accordingly, might be set aside by domestic judges. To this day, scenario d) has never occurred in its garden variety. Never has been the case that, for instance, a national measure that justifiably restricts one fundamental freedom (and falls therefore under the scope of Treaty law, without being precluded thereby) was found to breach the Charter. This unlikely coincidence warrants a deeper analysis (but not here), because it suggests that the Court is discreetly keeping the Charter in a locked drawer, lest member states react like the German Constitutional Court reacted after Fransson. A low-profile use of the Charter – one that essentially emptied it of its post-Lisbon potential – is what the Court’s record shows in the last 9 full years.

However, there is a hybrid category of cases, halfway between scenarios b) (double preclusion) and d) above. The paid annual leave rulings of 6 November 2018 belong in this atypical group, insofar as they relate to horizontal disputes. As far as their bearing on vertical disputes, they are squarely b)-type rulings: the measures are precluded by the Directive, and the breach of the Charter is just redundant.

The discussion below, instead, focuses only on the horizontal dimension of these rulings.

Did the Directive apply?

That the Directive applied must be assumed – otherwise the Charter would not have applied at all. Yet, the Directive could not apply to determine the outcome of the underlying dispute, for lack of horizontal direct effect. Can it be said that the Directive applied, and precluded the national measures, but was ineffective? The (ineffective) application of the Directive to the underlying scenario would then warrant the (very effective) application of the Charter. The Charter alone would warrant the disapplication of the German norms that breached both the ineffective Directive and the effective Charter.

The easier construction – that the Directive, lacking horizontal direct effect, could not apply – is untenable. Admitting that the Directive did not apply contradicts the notion that the German law fell within the “scope” of EU law, and would rule out the relevance of the Charter tout court. Some finer reasoning must support the Court’s decision. I suggest a couple of unnoticed distinctions that might help.

Direct application is not the same as direct effect: This is an all-time favourite for mid-term exams, so why not give this distinction a day in Court? Perhaps, the Directive did apply somehow (direct applicability), but could not be relied upon in domestic proceedings (lack of horizontal direct effect). This would explain the triggering of Article 51(1) of the Charter, but also the impossibility to use the Directive to solve the disputes between Mr Shimizu and Mrs Broßonn and their employers. Certainly the Directive applied, somehow. As soon as the transposition period expired, it deployed its legal obligatory effects and, even regarding horizontal disputes, triggered immediately Germany’s responsibility for failure to transpose correctly, which individuals can invoke to seek compensation without recourse to any implementing measure.

On the lack of direct effect, there is no real mystery: Article 7 of the Directive creates a precise and unconditional right, capable of invocation in domestic proceedings, so the direct effect checklist is in order. However, it can only work in vertical disputes. This distinction between direct applicability and direct effect could justify the Court’s truncated use of the Directive (sort-of applying to bring the Charter in under Article 51(1) of the Charter; not-really applying in its own right).

Scope of application is not the same as application: This is a contrived distinction, but it might operate in the background of these rulings. The Directive effectively cannot apply in domestic proceedings. However, the German law falls under the “scope” of the Directive, in a somewhat more abstract sense, as if applicable and applied were distinguishable. To accept that EU law reaches further than the circumstances to which it can actually apply is a head-scratcher. Yet, that is what the Court prescribed: domestic law must be set aside for intruding in an area that the Directive could not operate (the direct regulation of duties between private parties) but it nevertheless occupied – somehow. In the wake of Kücükdeveci (an earlier judgment on discrimination in employment), I noted the risk:

… it is necessary to evaluate the implications of Kücükdeveci: if the general principle of non-discrimination has a wider scope than the measures codifying it (the Directive), it follows that it can be invoked in a series of disputes that, despite concerning EU-regulated matters, fall outside the scope of the Directive.

Putting lipstick on Mangold

The distinctions above are nowhere mentioned in the rulings. So, if the Directive could not operate in domestic proceedings, how did the Court justify German law falling under the scope of EU law? This week’s rulings proffer only minimal reasons, but summon precedents to convey the idea that this apparent misalignment is nothing to worry about. From Bauer:

53           Since the national legislation at issue in the main proceedings is an implementation of Directive 2003/88, it follows that Article 31(2) of the Charter is intended to apply to the cases in the main proceedings (see, by analogy, judgment of 15 January 2014, Association de médiation sociale, C176/12, EU:C:2014:2, paragraph 43).

If one goes down the rabbit hole of cross-citations, however, Mangold awaits on the bottom. In terms of legal reasoning, it does not get any more controversial than Mangold. The judgment exasperated the former president of the German Constitutional Court and drafter of the Charter, who penned an op-ed wishing the Court of Justice to “Stop.” Ultimately, Mangold was criticised for encroaching into the Member States’ competence.

There are, in fact, a series of cases in which the Court prescribed the horizontal direct effect of a fundamental right, using a directive as a trampoline. Mangold was the first case, and Kücükdeveci followed. Dansk Industri (discussed here), also on discrimination on grounds of age, pulled the same trick, this time using the Charter rather than a general principle of law. AMS (discussed here) replicated the reasoning with respect to the workers’ right to consultation and information in the Charter, accepted its applicability but concluded that the Charter’s norm was not self-executing and stopped short of confirming its horizontal direct effect.

Egenberger (discussed here), very recently, asserted the horizontal effect of the right to non-discrimination on grounds of religion and the right of effective access to justice. In IR (discussed here), the Court consolidated Egenberger, and wisely recalled that non-discrimination is essentially a general principle: if the AMS explanation why the Charter applies to individuals were not compelling, the unwritten source might apply more liberally and come to the rescue just like in the good Mangold days:

69           Before the entry into force of the Treaty of Lisbon, which conferred on the Charter the same legal status as the treaties, that principle derived from the common constitutional traditions of the Member States. The prohibition of all discrimination on grounds of religion or belief, now enshrined in Article 21 of the Charter, is therefore a mandatory general principle of EU law and is sufficient in itself to confer on individuals a right that they may actually rely on in disputes between them in a field covered by EU law.


In all the cases mentioned, individuals could not rely on the directives in domestic proceedings. Yet, each directive somewhat marked the “scope” or “cover[age]” of EU law and let the fundamental right finish the work (unless some ingrained deficiency made it non self-executing, as in AMS). The simple annotation that the German law “is an implementation of [the] Directive” sanded over the uncomfortable truth: if the Directive was insufficient to set aside the norms of German law, perhaps these norms of German law lay outside the scope of EU law, and the Charter should not have found its way in the proceedings.

A sceptical reader might wonder what “a field covered by EU law” means, and wonder whether the Charter spilled over from the EU-law scope comfort zone. An optimist one can celebrate the expansive force of social rights. Horizontal application of fundamental rights translates into the creation of EU-based fundamental duties, and lets us catch a glimpse of solidarity in the making.

Barnard & Peers: chapter 9, chapter 20, chapter 6
Photo credit: PureTravel

Friday, 9 November 2018

Harassment of Human Rights Defenders: Measuring Democracy, Bad Faith and Hidden Agendas - What Role for the European Court of Human Rights?

Dr Marco Antonio Simonelli, PhD, University of Siena, and Alast Najafi, LLM candidate, University of Leiden

A. Introduction

In its recent judgment of Aliyev v. Azerbaijan, the European Court of Human Rights (hereafter the Court) activated anew the long dormant Article 18 ECHR (which provides that States shall not abuse the possible limitations on human rights which the Convention allows)  to unanimously find a violation thereof. The judgment is the third case this year where the Court found an Article 18 violation in Azerbaijan, but still remains unprecedented, since it allowed the Court for the first time to find a violation of that provision in conjunction with Article 8 ECHR (the right to private and family life). The judgment also indicates the measures to be adopted by the Azeri Government to execute the judgement, while drawing at the same time a dramatic picture of the political situation in Azerbaijan.

The present piece briefly illustrates the facts of the case, to subsequently consider the application of Article 18 in the light of the test elaborated by the Grand Chamber in Merabishvili v. Georgia (1). It is claimed that whilst, on paper, the “predominant purpose” test may be flawed, its application in Aliyev unveils Article 18’s promising potential in countering rule of law backsliding (2). In a third step, the significance of the recommendations given by the Court under Article 46 ECHR will be examined, that prove themselves as being a useful tool for enhancing the European supervision of “rebel” States, while leaving the more political aspects of the Convention’s enforcement mechanisms to the Committee of Ministers (3). In the conclusions, it will be sketched out which role Strasbourg could play in safeguarding democracy across Europe (4). 

1. The facts

Similar to the facts in Rasul Jafarov v. Azerbaijan, the case concerned criminal proceedings brought against a civil-society activist, well-known human rights defender and chairman of a non-governmental organisation. Charged for illegal entrepreneurship, large-scale tax evasion and aggravated abuse of power, the applicant was detained, and several documents and objects were seized during the search of his association’s office and home. While still in detention, new charges were brought against the applicant that lead to a five-year suspended sentence and his release in March 2016. The Court held that there had been, inter alia, a violation of Article 18 taken in conjunction with Articles 5(1)(c) (concerning pre-trial detention) and 8(2) ECHR, since both his detention and the search of his office and home pursued the illegitimate aim to silence and punish the applicant for his activities in the field of human rights.

2. The considerations under Article 18

Article 18 ECHR states that restrictions on Convention rights shall not be applied for any purpose other than those for which they have been prescribed. For a long time, this Article has not been of any practical importance in the Court’s jurisprudence. Yet, a slow but steady rise can be observed since 2004, when the Court for the first time found a violation of that provision in Gusinskiy v. Russia, leading to eleven violations to the present day. This judicial chain culminated, at the end of 2017, in the Grand Chamber ruling in Merabishvili, which set out general principles with regards to the interpretation and application of Article 18. (See the Court’s guide to the application of Article 18).

Reiterating its findings in Merabishvili, the Court reaffirms the much criticized “predominant purpose” test. Indeed, this test faced criticism since the very moment of its genesis: the two concurring opinions attached to the judgment in Merabishvili attack the test for being vaguely defined and not providing objective criteria, and also scholars considered the test to be too narrowly constructed. Be that as it may, the test prescribes that a restriction of one of the substantive rights of the ECHR engages the responsibility of a State under Article 18 when a purpose which is prescribed by the Convention only serves to mask an illegitimate predominant and ulterior purpose. Contrary to what the Government tried to argue, the Court also confirms the main novelty of the Merasbishvili judgment, namely that States no longer enjoy a general presumption of good faith.

Applying these principles to the present case, the Court points out that the “combination of the relevant case-specific facts in the applicant’s case [were] similar to that of Rasul Jafarov (…) where proof of ulterior purpose derived from a juxtaposition of the lack of suspicion with contextual factors”. Unlike in Khodorkovskiy v. Russia, no “healthy core” was to be found in the charges against the applicant. Therefore, the Court, after delivering a comprehensive and alarming analysis of the political situation in Azerbaijan - the Court points out restrictive legislation towards non-governmental organisations and the fact that “[s]everal notable human-rights activists have been similarly arrested and charged with serious criminal offences” (Aliyev, §208-214) - has no problem in finding a violation of Article 18 in conjunction with Article 5 and 8 ECHR.

The manner in which the Court reached the conclusion that there has been a violation of Article 18 sheds a new light on the relevance of the “predominant purpose” test. In Merabishvili, the Court considered as non-decisive the evidence submitted by the applicant in an attempt to demonstrate that his detention had been politically motivated. By contrast, in the present case, the political situation appears to have played a decisive role in the Court’s assessment.

A paradigmatic example is provided by the different weight attached to the statements from government officials. While considering in Merabishvili that “[s]uch statements can only be seen as proof of ulterior purpose (…) if there is evidence that the courts were not sufficiently independent” (§ 324); the Court’s finding in Aliyev that “the applicant’s arrest was accompanied by stigmatising statements made by public officials” (§ 210) amounts to an element of proof to conclude that the proceedings brought against the applicant were politically motivated. This new approach to the application of Article 18 ECHR aligns better with what Judges Yudkivska, Tsotsoria and Vehabović had in mind when writing their separate opinion in Merabishvili: “when there is evident misuse of State machinery for improper political ends, the Court should treat it by default as the predominant purpose and thus find a violation of Article 18” (Joint Concurring Opinion of Judges Yudkivska, Tsotsoria and Vehabovic, § 38).

3. The “recommendations” under Article 46: a new instrument of European supervision?

After having decided the case on the merits, the Court moves on to consider the application of Article 46 ECHR (§ 220-228), which concerns the legal force of Court judgments. Considering its earlier case-law, where the Court had already found Article 18 violations in Azerbaijan, the Court finds that the events “cannot be considered as isolated events” but reveal a “troubling pattern” of arbitrary arrest and detention of government critics and human rights activists and a “misuse of criminal law in defiance of the rule of law” (Aliyev, §223). Furthermore, the Court notices that applications raising similar issues have been communicated to the Azerbaijani Government or are pending before the Court.

It has to be underlined that the Court was not only faced with a practice incompatible with the Convention but with a State which openly disregards the rule of law, as the Court worryingly stressed itself. Indeed, the measures to be adopted when domestic courts limit themselves to an “automatic endorsement of the prosecution’s application” (Aliyev, §223) can hardly be identified.

Nonetheless, the Court decided to indicate general measures to be taken by Azerbaijan to comply with the judgment, i.e. the eradication of politically motivated prosecution against human rights activists. The practice to indicate measures of general application - without going the whole way of a pilot- judgment procedure – is becoming ever common in Strasbourg case-law (Sicilianos, 2017). However, notwithstanding their frequent use by the Court, the assessment of their legal value proves to be complex.

A prompt answer can be given by restricting the analysis to the operative part of the judgment: if the Court’s indications are not echoed in the operative part, they can be considered by the respondent State as no more than obiter dicta.

Such a conclusion proves unsatisfactory in the present case. On the one hand the judgment fails to address concrete measures towards Azerbaijan in its operative part. Yet, on the other hand, the formulations chosen by the Court – “the respondent State must focus on the protection of critics of the government”, “[t]he measures to be taken must ensure the eradication of retaliatory prosecutions”, “The individual measures to be taken by the respondent State [..] must be determined in the light of the terms of the Court’s judgment” (Aliyev, §226-227) – are of a strong prescriptive nature and cannot be classified as mere recommendations.

The most satisfying answer, at least in the present case, is to view these indications rather as a message directed at the Committee of Ministers (which has the task of following up the implementation of the Court’s judgments under Article 46 ECHR) than at the Azeri Government; in fact, there are two good reasons to think so.

First, according to Article 46(2) ECHR, the primary responsibility for the supervision of the execution of judgments lies with the Committee of Ministers; thus these “recommendations” can be considered as the backdrop against which the Committee of Ministers has to evaluate the implementation of the judgement by the Respondent State. (See Villiger, in Seibert and Villiger, eds). Second, the precise nature of the indications may be interpreted as an attempt by the Court to prevent an action of the Committee of Ministers under Article 46(4) ECHR, which allows the Committee of Ministers to refer to the Court the question whether a respondent state has faithfully executed a judgement. The procedure has been harshly criticized by scholars as being a game in which the Court has nothing to win but much to lose, i.e. its legitimacy (see De Londras and Dzethtsiarou, 2017).

The Committee of Ministers triggered Article 46(4) for the first time in December 2017 to inquire whether Azerbaijan had properly executed the Court’s Judgment in Ilgar Mammadov. Since the execution of the judgment only required the release of the applicant, who was still detained at the time the Committee of Ministers launched the proceeding, the question was purely rhetorical. This impasse, which put the Court in a “impossible position” (De Londras and Dzethtsiarou), was resolved by the eventual release of Mammadov in August 2018. It is very likely that the Court, seeking to avoid being again called upon to sanction the lack of effectiveness of its own judgments, issued precise instructions to address the problem of non-execution, a problem which is, indeed, political.

However, the relevance of these measures goes beyond a mere assistance to the Committee of Ministers in supervising the implementation of the judgments.

A closer look at the last judgments in which the Court indicated measures of general application under Article 46, reveals that, in the last four years, this practice only concerned States with compliance problems with the requirements of the rule of law, in particular Azerbaijan, Hungary, Poland, Russia, Turkey and Ukraine. This pattern suggests that Aliyev may very well be part of a judicial strategy of the Court aimed at enhancing the persuasive force of its rulings. In this reading, the issuing of general recommendations can be seen as a compromise between a judgment in which the Court only orders the State to pay just satisfaction under Article 41 ECHR, and a pilot or quasi-pilot judgment in which the Court prescribes specific measures to adopted by the Respondent State in its operative part. This strategy allows the Court to continue exerting its beneficial influence over countries characterized by a rule of law backsliding, without causing a direct clash with non-abiding States, as this may prove fatal for the very existence of the Council of Europe.

4. Conclusions

Witnessing the possible emergence of new legal approaches is an exciting thing but only the test of time and subsequent judgements will show whether the Court is actually developing a new test under Article 18. In this regard, the Navalny case, currently pending before the Grand Chamber, offers a good opportunity for the Court to further consolidate its case-law, and to revisit its predominant purpose test.

What is sure is that this provision is “the only measuring instrument for democracy” the Convention regime offers. Backing this provision up by taking on a more proactive stance under Article 46 in situations where similar violations recur, leaves the Court better equipped to challenge State practices that repeatedly violate the rule of law and close spaces for civil society. The rule of law backsliding is not intrinsic to the Convention system but emerged very prominently in the European Union, where the so-called Copenhagen dilemma – i.e. the question of how to find effective means to ensure continuous respect of the organisation’s values and sanction violations thereof, once a State has successfully acceded but fails to comply – remains unsolved.

However, the recent line of case-law from Strasbourg, of which Aliyev is the last piece, hints which role could be reserved for the European Court of Human Rights in cases where States abusively undermine democracy by targeting individuals.

Barnard & Peers: chapter 9
Photo credit: ISHR