Top of page ↑


30 June 2020

2020/19 Freedom of religion: a tale of two cities

By: Filip Dorssemont, Professor of Labour Law at Université catholique de Louvain and Guest Professor at Free University of Brussels.

Why divergences can be problematic

The judgment issued by the Court of Justice of the European Union (CJEU) in the case Achbita v G4S Secure Solutions[i] has prompted a debate about the convergence of the approach adopted on religious apparel at work in an enterprise which claimed to be neutral with the allegedly more rigorous stance of the European Court of Human Rights (ECtHR) in a case concerning a crucifix.[ii]

Formally, both Courts examined the same issue through a different lens. Whereas the CJEU needed to assess primarily whether the person fell victim to discrimination based upon religion or belief, whether it is direct or indirect and can be justified, the ECtHR needed to assess whether a restriction of the freedom of religion can be justified or not. The latter presupposes that such a restriction would need to be prescribed by law, is justified by an aim recognized as legitimate under Article 9(2) of the European Convention on Human Rights (ECHR) and that such restriction is necessary in a democratic society. Any discrimination in the field of religion or belief will inevitably constitute a restriction of the freedom of religion. Article 14 ECHR implies that the enjoyment of freedoms of religion shall be secured without discrimination on any ground such as sex, race, colour, language, religion, political or other opinion, national or social origin, association with a national minority, property, birth or other status. States need to treat monotheistic and other religions on an equal footing. Some people tend to forget that all monotheistic religions have in fact emerged outside the European continent.

For a number of legal and practical reasons, it is important that there is a minimum of convergence between the case law of the Supreme Court (CJEU) which assesses the justification of an alleged discrimination based upon religion or belief and the case law of the Supreme Court (ECtHR) which assesses the justification of the restriction of the freedom of religion that such a discriminatory situation entails. In case of divergence, a number of problems arise. Member States of the European Union will in practice be bound by both EU Directive 2000/78/EC as well as by Article 9 ECHR. They will need to abide by a double standard. The mere fact that the CJEU would accept a discriminatory situation will not free the Member States from the obligation to demonstrate that the restriction of the freedom of religion which this discrimination entails can be justified as well. Preventing the discrimination will then be the best way to overcome the problem, if the discrimination would prove to be a violation of the ECHR.

In case of such a divergence, there is also a problem of constitutional legitimacy for the CJEU. The CJEU is bound by the ECHR in its interpretation of the freedom of religion. Article 52(3) of the Charter of Fundamental Rights of the European Union (the Charter) provides that:

Insofar as this Charter contains rights which correspond to rights guaranteed by the Convention for the Protection of Human Rights and Fundamental Freedoms, 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.

Hence, the Charter puts forward a canon of intertextual interpretation. If the CJEU would interpret the notion of discrimination based upon religion or belief in a way which is incompatible with the Strasbourg case law on freedom of religion, it is difficult to see how it would have done justice to the freedom of religion as enshrined in the Charter.

At present the CJEU has only had to deal with four landmark cases related to discrimination based upon religion or belief since the adoption of Directive 2000/78.[iii] This case law can be divided in two sets of two cases. The oldest strand of case law only dates back to 2017. It dealt with the issue of female workers wishing to express their religious convictions by wearing a headscarf in an enterprise which claimed in an explicit way to be neutral or in an enterprise which tried to hide behind the lack of religious tolerance stemming from its clients. In a subsequent strand of two cases, the Court dealt with the opposite situation of a candidate for a job in an NGO based upon Protestant religious convictions and a doctor working in a Catholic hospital. The candidate was not recruited because she was not affiliated to a Protestant church neither was she considered to be a socialized Christian. The doctor was fired when he concluded a second civil marriage although his religious marriage consecrated before the church was neither annulled by an ecclesiastical court nor dissolved by the death of his previous wife, who was indeed still alive and kicking. These cases have all been analysed at length in this journal.[iv]

I want to focus on the question whether the outcome the CJEU produced in these four cases was essentially different from the outcome of similar cases dealing with restrictions on the freedom of religion ruled by the ECtHR. Insofar as a different outcome could be explained by the fact that the cases were just not comparable, this will be highlighted as well. Such a comparison necessitates a small caveat. The procedures before the CJEU and before the Strasbourg Court are essentially different. The CJEU in preliminary procedures interprets EU law, but is not supposed to apply it to an individual case. It is up to the referring judge to do that. The Strasbourg Court does not assess in abstracto conformity of national law with the provisions of the ECHR, but it examines whether the way in which a national judge has applied their national law in a given situation is in conformity with the ECHR. If this is not the case, there will not be a role for any judge to be played, but the State will be required to pay a compensation to the applicant.

The first strand: positive freedom of religion in a ‘neutral environment’

In a previous issue of this journal, Hashemi already pointed out the controversial character of one of the two headscarf cases, heavily criticized for being incompatible with the stance adopted by the ECtHR in the Eweida case.[v] ,[vi]

The author made an interesting attempt to reduce the hermeneutic gap between both Courts, or in other words to reconcile both judgments. The different outcome, where Strasbourg seems to save Miss Eweida carrying the crucifix and the CJEU seems to provide a pathway to be followed by a national judge to save the employer is not seen to be in se et per se inconsistent. Hashemi argues that the facts of the cases were very different. British Airways had to some extent destroyed its own credibility by changing its policy rules on uniforms and by applying these rules in an inconsistent way. The author also argues that a small crucifix is something different than a headscarf. Personally, I think such a comparison depends upon the perspective adopted. It reduces the examination of the ‘ostentatious’ character to the rather superficial matter of size. A more qualitative comparison could amount to a different conclusion. A crucifix is at odds with the second commandment of the Decalogue which states: “You shall not make for yourself a carved image, or any likeness of anything that is in heaven above, or that is in the earth beneath, or that is in the water under the earth.” A crucifix is extremely ostentatious by representing a divinity as a semi-nude person attached to a cross. The Catholic Church has attempted to overcome this tension between religious practice and the second commandment by introducing a distinction between the prohibited adoration of religious idols and the recognized practice of the veneration of such sacred objects. In sum, an abstract piece of clothing which still allows any spectator to see the identity of his or her colleague at work is much less explicit, hence ‘ostentatious’ from an iconographical perspective in the circumstances outlined. Is it allowed to state that the Strasbourg Court’s understanding of ‘ostentatious’ is slightly influenced by Christian standards and traditions?

Contrary to Hashemi, I don’t think that a part of the legal doctrine created what she calls a storm in a teacup. It is true that the CJEU did not apply as such a proportionality test, but just explained it to the referring judge, whereas Strasbourg applied it. However, this will always be the case. It is in the nature of a preliminary procedure that the CJEU will not solve the legal dispute.

However, I am inclined to see insurmountable divergences between the approaches of both Courts. The CJEU is satisfied with consistently and systematically applied policies of neutrality, whereas the Strasbourg Court will have to insist that such a policy is based upon a piece of prescribed law. In Achbita, no such prescribed rule existed from a legal point of view. Prescribed comes from the latin prae-scribere (scripsi, scriptum). It suggests a written text put on paper, duly notified and made public to the workers. The contract of employment of Achbita did not contain a neutrality clause, neither was such a clause included in the written shop rules at the time the dispute arose. The shop rules were in fact amended after the dispute arose to include such a clause and these amendments had not entered into force when Achbita was dismissed. One might argue that the Strasbourg Court has in the past also recognized that general principles could be a source of restrictions prescribed by law. It has considered that employees need to give proof of loyalty, reserve and discretion mitigating the extent of their freedom of expression.[vii] However, in freedom of religion (at work) cases, the Court has always pointed to the existence of explicit contractual clauses restricting the freedom of religion in so-called Tendenzbetriebe.[viii]

It is entirely inconsistent for the CJEU to recognize that ‘religion’ also covers the expression of a religious conviction by the carrying of religious apparel on the one hand, and on the other hand to rule that a discrimination based upon the existence of such apparel is merely indirect. The mere fact that in the enterprise concerned the so-called policy would not differentiate among workers irrespective of the nature of their convictions (philosophical, ideological and religious) does not show in my modest opinion that there has not been direct discrimination. In a famous case, a school after having fired a worker who had undergone a correction or transformation of ‘his’ sex, argued that there was no direct discrimination based upon sex. It claimed that it would have fired anyone changing his or her sex, whether formerly female or male. The CJEU rejected that argument firmly and concluded that it was direct discrimination.[ix]

The Strasbourg Court clearly stated prior to measuring the size of the crucifix that a balancing operation needs to take into account the ‘scale’ of the interests at stake. It did consider that there is a difference between a fundamental right protected under the ECHR and the legitimate economic interest of a company to protect a ‘corporate image’. In sum, there is a problem of incommensurability of conflicting interests, which is such that the size of the religious apparel in my view cannot be the decisive factor, although it could and had to be taken into account.

Last but not least, there is a lack of convergence in the way in which the CJEU takes into account the willingness and the ability of the employer to offer a job in a back office. The CJEU accepts that such an offer insofar as it does not in any way provoke inconvenience for the employer is helpful to conclude that an indirect discrimination could be considered as proportionate. In Eweida, the ECtHR was confronted with such a pragmatic attitude stemming from British Airways, which indeed had offered Miss Eweida such a job in a back office. However, this very fact was immaterial for the judgment of the ECtHR that there had been a violation of Article 9 ECHR. The question in fact arises whether the CJEU was sufficiently aware of the fact that isolating workers and reducing their contacts is a text book example of harassment and that harassment is assimilated to discrimination. In sum, discrimination is in my view not a way to remedy or justify discrimination.

The second strand: negative freedom of religion in an environment based upon a religious ethos

The question whether the judgments in Egenberger and IR are compatible with the case law of the Strasbourg Court has drawn slightly less attention. Since both workers had been saved by the CJEU, less need might have been felt to engage in a comparison with the Strasbourg case law. However, an opposite exercise is useful. Thus, the question arises whether the Strasbourg case law has taken in the past a more employer friendly attitude towards the situation of employees in so-called Tendenzbetriebe, id est organisations based upon a religious ethos. The Strasbourg Court has committed itself to an intertextual interpretation of the provisions of the ECHR, which logically can only include the provisions of Directive 2000/78 as interpreted by the CJEU. Hence, a lack of convergence might be at odds with the outcome of a more intertextual approach and could be helpful to convince Strasbourg to reconsider its case law.

A factor which complicates the assessment of the Strasbourg case law is that freedom of religion works both individually and collectively. It tends to protect individuals with strong religious convictions as well as religious communities. In the cases on Tendenzbetriebe clashes occurred because workers did not share the religious convictions held by these institutions or were unwilling or unable to live up to the morals which were supposed to flow from them. The workers did not have religious convictions, they might have different ones or their private lives at some point proved to be at odds with the morality professed by these institutions.

The right of the citizens Egenberger and JQ not to be discriminated against on the basis of their religion or belief inevitably constituted a restriction of the freedom of religion in its collective dimension to which the Protestant Diaconate and the Catholic community were entitled. In both cases both employees and employers could invoke the freedom of religion. Citizens Egenberger and JQ invoked the freedom to hold a religious or metaphysical conviction other than that of their employer. The employer invoked the internal autonomy of the religious community to which they belonged. There is a conflict within Article 9 ECHR that has both a collective and an individual dimension. Mr JQ could also have invoked his right to respect for his private and family life. In my view, it is easy to imagine that pressure on employees not to marry which is accompanied by a ‘loss of livelihood’ affects Article 8 ECHR at its core.

In this context, it should be recalled that the Charter guarantees not only the principle of non-discrimination on the basis of religion or belief, but also freedom of religion. Moreover, freedom of religion must be interpreted in the light of the similar provision of the ECHR. The Court of Justice and, where appropriate, the national courts should give at least an equivalent degree of protection to the freedom of religion guaranteed by the Charter. It should be noted, however, that the Court has not succeeded in referring to the case law of the Strasbourg Court in either of its judgments Egenberger or IR .  Advocate General Tanchev did much better. He referred quite systematically to the Strasbourg case law. For example, he considered that the preliminary question on the criteria to be used to justify direct discrimination had to be assessed in the light of the Strasbourg case law on philosophical conflicts of loyalty.[x] He distinguished in this case law the following criteria: the nature of the position in question, the proximity of the activity in question to the proclamation task and the protection of the rights of others, as well as the balance between the competing rights and interests at stake.[xi]

There is, in my view, an interesting distinction between the two tests. It is inherent in the determination of direct discrimination based on religion and belief that it can only be justified on grounds of genuine, legitimate and justified occupational requirements having regard to the organisation’s ethos. The legitimate aims justifying a restriction of conventional rights are much broader. What in no way constitutes a distinction is the mere presence of  a proportionality test. Restrictions on freedom of religion must be assessed in the light of public security, the protection of public order, health or morals, or the protection of the rights and freedoms of others. The justification for unequal treatment and related harm must be proportionate to the justification put forward for direct discrimination, just as restrictions on freedom of religion must be proportionate to the legitimate aims pursued.

The Strasbourg Court applies a number of criteria closely linked to those used to justify direct discrimination.[xii]

These include the nature of the function (Obst), in particular the relationship between the function and the proclamation of faith (Schüth), and the non-conformity of an act in the light of the ethos of the religious community (Obst).

The ECtHR also weighs strongly the impact of the dismissal on the economic and professional situation of the employee. The dismissal of a young employee, with a minor seniority, is to the disadvantage of an employee. The monopoly position of the organisation in connection with the dismissal of a position on a specific labour market plays to the advantage of the employee (Schüth).

It is certainly important to balance the interests involved. The Court of Justice does not examine this aspect at all. The employability of the doctor JQ is not examined. Nor is a distinction made between non-recruitment and dismissal. Incidentally, the way in which the ECtHR carries out or checks this balancing of interests is sometimes open to criticism. In the Fernández Martínez case, it was taken into account that a dismissed teacher of religion who had been ordained as a priest could draw unemployment benefit. Neither his age as an older worker, nor the fact that he had not found an equivalent post afterwards were taken into account. His job as an attendant did not seem to me to be a textbook example of a similar job.

Another distinction between Luxembourg and Strasbourg seems to me to be the observation that the Strasbourg criteria do not play a role in the context of an examination of the lawfulness of a restriction of a fundamental right, but in the context of proportionality. In accordance with the case law of the Court of Justice, the examination of the post always forms part of an assessment of the legitimacy of discrimination. The relationship between religion and the nature of the function is also considerably more objective. In the case of Strasbourg case law, a dominant subjective criterion predominates: the so-called credibility of the institution invoking its ethos to justify dismissal. This subjective criterion, which is separate from an objective analysis of the relationship between religion and the function, does not actually appear in the case law of the Court of Justice. The question is whether this issue of credibility should not be mitigated to some extent. It seems to justify dismissals of people who are not exercising functions for which a religious conviction seems necessary, by invoking the idea that there mere presence would undermine the credibility of the institution in the eyes of the members of a congregation or in the eyes of clients, patients or parents. There is of course an argument to take into account the issue of credibility to some extent. Framework Directive 2000/78 already integrates these concerns, by stating that it does not prejudice the right of Tendenzbetriebe to require individuals working for them to act in good faith and with loyalty to the organisation’s ethos.[xiii]

This technical comparison between the architecture of the non-discrimination test and the test of the lawfulness of the restriction of certain conventional fundamental rights of workers should not make us forget that the ECtHR did not ‘save’ virtually all workers working in identity-based organisations. The only lucky person turned out to be the organist Schüth. The PR director Obst of the German Church of Mormons, Mrs Siebenhaar, who watched the little ones in a crèche, as well as the Spanish priest-religious teacher Fernández Martínez knocked at Strasbourg’s door in vain. In the first two cases, however, the same German doctrine of the Tendenzschutz was applied. This shows that the Court of Appeal never tests rules in abstracto against the ECHR, but only the application of those rules in a concrete case. Schüth’s rescue was based on two fundamental criticisms. According to Strasbourg the consequences of the dismissal actually amounted to a Berufsverbot (prohibition of occupation) in one of the few ‘sectors’ in which Schüth could exercise the profession of organist. Moreover, the Court had difficulty with the German judges’ refusal to examine the link between the organist’s activity and the proclamation of faith. The Court particularly criticised the fact that the German judges, without further investigation, relied on the Church’s view on this, which was held to be true. This actually put the finger on the German wound: the too marginal control by the German judges, who only looked at the Selbstverständnis (self-perception )of the Church. It seems to me that the connection between the activity and the religious ethos of Fernández Martínez and Obst was much more profound than in the case of Egenberger and IR . As far as Siebenhaar is concerned, caution seems advisable. However, I find it hard to imagine that the task of a child caretaker in a crèche had anything to do with ‘proclamation of faith’. Nor is there any trace of internal proselytism on the part of Mrs Siebenhaar in this case.


For the time being, therefore, it seems very difficult to ascertain whether the application of the Strasbourg criteria will be as generous to workers as the application of the discrimination test. Prima facie, one has the impression that the discrimination test coupled with the strict interpretation of Article 4(2) of Framework Directive 2000/78 offers more opportunities to employees who feel discriminated against in an identity-related organisation. From a formal point of view, this does not seem problematic to me. It is not because the Strasbourg Court would be more lenient with regard to restricting the fundamental rights of employees in the workplace than the Court of Justice that a problem arises. The ECHR offers only a minimum level of protection. However, there is an important caveat. Neither the Court of Justice nor the identity-based organisations have examined in the proceedings the possibilities offered by Article 2(5) of Framework Directive 2000/78, which provides that:

the Directive shall be without prejudice to provisions of national law necessary in a democratic society for public security, for the maintenance of law and order and the prevention of criminal offences, for the protection of public health and for the protection of the rights and freedoms of third parties.

These freedoms of third parties include, of course, the freedom of religion of the denominations as well as the related internal autonomy.

The approach of the Court of Justice is, in my view, happily different from the Károly Nagy v Hungary[xiv] judgment of the Grand Chamber of the ECtHR. In this judgment, the Court ruled that a pastor of the Reformed Hungarian Church who wished to contest the financial compensation of his suspension and his resignation from his denomination before a civil court could not invoke Article 6 ECHR, despite the fact that these courts denied their jurisdiction. Indeed, the Hungarian courts had ruled that the pastor’s relationship was not of a civil law nature.

Now, the comparison between the two cases at the centre of this commentary and the Nagy case is not so obvious. Neither the Protestant NGO nor the hospital ever stated that the employment relationship was not an employment contract. Framework Directive 2000/78 does not even mention the civil law nature of the employment relationship. It applies, without further specification, to employment, self-employment and any other occupation.

One way to understand the fact that the Strasbourg Court has only saved a few people working for a Tendenzbetrieb is to compare the functions these people were actually carrying out. Contrary to Egenberger and JQ, the bulk of workers concerned were not exercising a merely technical function. Some of the workers concerned exercised the kind of important functions for which in the understanding of the CJEU religious convictions concerned could be considered to be necessary. Thus, Martínez was not just a teacher, but a teacher of religion. Obst was the face of the German Mormon Church. He was the head of the Public Relations department. In the case of Schüth, the Court was not at all convinced of the sufficient link between the job of the organist and the objective of the propagation of the faith. The only case which according to me might prove to have had a divergent outcome is the case of childcare assistant in a day nursery (Kindergarten). It is very difficult to assess the convergence issue, since the judgment is mute on the precise job description.

[i] CJEU, 14 March 2017, C-157/15 (Samira Achbita, Centrum voor gelijkheid van kanen en voor racismebestrijding).

[ii] See the case mentioned in footnote 1 above as well as: CJEU, 14 March 2017, C-188/15, (Asma Bougnaoui, Association de défense des droits de l’homme (ADDH) – v – Micropole SA).

[iii] See the cases referred in footnotes 1 and 2 above, as well as CJEU, 17 April 2018, C-414/16 (Vera Egenberger v. Evangelisches Werk für Diakonie und Entwicklung eV), CJEU, 11 September 2018, C-68/17 (IR v.JQ).

[iv] See M. Hashemi, ‘Eweida versus Achbita: a storm in a teacup?’, in EELC 2019 No. 3, pp. 174-177 and A. Swiatkowski, “The religious ethos and differences of treatment in employment on grounds of belief”, in EELC 2019 No. 3, pp. 179-187.

[v] M. Hashemi, ‘Eweida versus Achbita: a storm in a teacup?’, in EELC 2019 No. 3, pp. 174-177.

[vi] See ECtHR, Eweida v UK, 15 January 2013, nos. 48420/10, 59842/10, 51671/10 and 36516/10 (Eweida v UK).

[vii] ECtHR, 21 July 2011, nr. 28274/08 (Heinish – v – Germany).

[viii] See footnote xii below.

[ix] CJEU, 30 April 1996, C 13/94 (P – v – S and Cornwall).

[x] See the conclusions of Advocate General Tanchev in Egenberger, § 105.

[xi] See the conclusions of Advocate General Tanchev in Egenberger, § 111.

[xii] See especially the following cases: ECtHR, 23 September 2010, nr. 425/03 (Obst v Germany); ECtHR, 23 September 2010, nr. 1620/03 (Schüth v Germany); ECtHR, 3 February 2011, nr. 18136/02 (Siebenhaar v. Germany); ECtHR, 12 June 2014, nr. 56030/07, (Fernández Martínez v Spain).

[xiii] Article 4(2) of Directive 2000/78.

[xiv] ECtHR 14 September 2017, nr. 56665/09 (Károly Nagy v Hungary).