Little EU Court of Justice (CJEU) case law has been as fiercely criticised as that relating to the right to be free from religious discrimination (including Cavalcanti, Pin, Henrard, and Howard in this Symposium). However, the CJEU recently found a sympathiser in Ronan McCrea. He argues that, given “the inevitable uncertainty produced by rapid and unprecedented religious change in Europe, it is the most prudent and politically sustainable approach for the time being”. In his view, the approach is one of “justifiable caution”.

Despite my disagreements with him, I believe that we must take McCrea’s position seriously. If we do so, we may be able to develop a more fine-grained view of when caution is warranted. McCrea rightly brings together two strands of case law that are all too often discussed disjointly. In Egenberger and IR v JQ, on the autonomy of churches and other religious employers, the CJEU construed the right of religious employers to discriminate fairly strictly. In contrast, in Achbita, Bougnaoui, I.X. v Wabe, L.F., and O.P.  v Commune d’Ans, concerning the right to wear religious dress in the workplace, the CJEU construed the right of public and private employers to discriminate broadly. Provided that a prohibition to wear the Islamic headscarf is part of a “neutrality policy” that bans any manifestations of religious, philosophical, and political beliefs at work, it is likely justified. Indeed, as McCrea notes, the cases on religious discrimination “bring together two areas of law in which the CJEU has taken markedly different approaches”.

While I have taken the view that the less cautious case law on the autonomy of religious employers can be justified (here) and that the more cautious case law on religious dress requirements cannot be (here and here), I believe that McCrea’s position is not without merit. At the same time, I still firmly believe that the case law on religious clothing reveals deeply troubling attitudes toward Muslim women that have no place under anti-discrimination law. In this contribution, I will attempt to reconcile these apparent opposites by offering a more nuanced understanding of when caution is justified.

Where should caution end, and where should it begin?

Any plea for judicial deference invariably raises the question of how much of it is justified. In other words, where should caution begin, and where should it end. In this respect, McCrea’s position raises difficult questions about EU fundamental rights and anti-discrimination law. He is right that, as regards religion’s role in society, “Member States have always taken different approaches … [and that] in many Member States, the situation is highly fluid with significant societal changes prompting intense political debate and significant legal and political changes”. However, it also seems right that religion is not unique. Europe has witnessed equally significant legal and political changes in relation to the position of sexual minorities.

As a result, his analysis raises the question if judicial caution must end with religion or be extended to sexual minorities and perhaps also the rest of EU anti-discrimination law. For example, what about the obligation to recognise same-sex marriages validly entered into in other Member States? Should the Court reverse its decisions in Coman and Trojan, and show greater respect for the national identities and traditions in Member States that have not legalised same-sex marriage? And what about discrimination on grounds of sexual orientation, which is covered by the same Directive as religious discrimination. Was the Court wrong to decide that anti-discrimination law is violated when a company refuses to hire or renew a contract with someone who is gay (here and here)? The CJEU does allow companies to engage in such practices vis-à-vis Muslim women, provided that they cloak their practices in “neutral” terms.

Having just published The End of the Gay Rights Revolution, McCrea likely has thoughtful answers to the questions I just posed. However, this does not alter the fact that the case law on the position of sexual minorities raises very similar questions to the case law on the permissibility of religious clothing bans at work. Moreover, given that the latter cases constitute an exception to the CJEU’s generally strict application of anti-discrimination norms in relation to other protected grounds (including sexual orientation and race and ethnicity), accepting McCrea’s arguments invariably poses questions about the structure of EU anti-discrimination law as such.

In my view, such questions do not have to be asked. While I will argue later that McCrea is right that caution may sometimes be justified, I will first show why it goes against the rationale of EU anti-discrimination law to grant it to the extent that it has been granted in the headscarf case law.

Back to the root of the problem

Once a year, I thank the CJEU for its verdicts. I do so on the day when I teach key concepts of EU anti-discrimination law, using the headscarf case law as example. Even the most dispassionate students become engaged, especially when we discuss Advocate General Kokott’s Opinion in the Achbita case. Her analysis typically provokes a great deal of headshaking, anger and frustration, and above all the question of how it is possible for one of Europe’s leading lawyers to take such an openly hostile stance towards Muslims.

I find it impossible to disagree with my students. The Opinion is by quite some distance the worst Advocate General Opinion I have ever read. If I were asked to summarise it in one sentence, I would say that, willingly or not, it proposes to carve out a derogation for Muslim women from the protections offered by EU non-discrimination law. Since the Opinion lies at the root of the Achbita judgment, and thus of subsequent rulings on the right to wear the headscarf at work, re-examining its most problematic aspects may help illustrate why it amounts not to caution, but rather a capitulation to Islamophobia.

The problems begin when AG Kokott proposes to draw the scope of direct discrimination narrowly with regard to religion because this is a “mutable” personal characteristic (paras 44-46). Contrary to what she argues, this goes against established case law, which never regarded mutable characteristics as being less protection worthy (consider, for example, case law on discrimination against pregnant women), but it presents an important first step in her weakening of anti-discrimination norms for Muslim women.

The AG’s reasoning gets truly problematic when she considers whether a ban on religious clothing at work is a genuine and determining occupational requirement that justifies an interference with the right to non-discrimination under Article 4(1) of the Equal Treatment Directive. She reasons that businesses must be allowed to “take into careful account the preferences and wishes of its business partners” (para 90), but may not cater to demands from customers to “be served only by employees of a particular religion, ethnic origin, colour, sex, age or sexual orientation” (para 91). However, customers may demand “to be served without discrimination, courteously and to a basic standard of politeness” (para 92). So far, so good, but which of these does a company policy prohibiting religious clothing at work due to customer preferences amount to? Does it cater to discriminatory customer preferences or demand a basic level of courteousness and politeness from its (Muslim) employees? Astoundingly, she implies that demanding Muslim employees to remove their veil falls in the second category.

AG Kokott saves the worst for last. It is worth citing paragraphs 132 of her Opinion in full:

“The wearing by male or female employees of visible signs of their religious beliefs, such as, for example, the Islamic headscarf, in the workplace may be prejudicial to the rights and freedoms of others in two principal respects: on the one hand, it may have an impact on the freedoms not only of their colleagues but also of the undertaking’s customers (particularly from the point of view of the negative freedom of religion); on the other hand, the employer’s freedom to conduct a business may be adversely affected.”

It is when reading this paragraph that my students are rightly dismayed. According to the AG, employees wearing the Islamic headscarf may prejudice the rights and freedoms of their colleagues and customers. But who, other than persons who are blatantly Islamophobic, will consider their rights and freedoms to be prejudiced by having to interact with a Muslim woman wearing the headscarf? The answer is none. A Muslim woman wearing a headscarf is observing her religious beliefs; she is not proselytizing for them. In no way does the Islamic headscarf interfere with someone else’s rights and freedoms.

Unfortunately, not appreciating that observing one’s beliefs is different from proselytising for them, AG Kokott proposed that, within the framework of anti-discrimination law, those with discriminatory attitudes should receive more protection than the victims of such attitudes. This does not amount to justified caution, but a troubling misunderstanding of the very point of anti-discrimination law.

The CJEU’s case law

The CJEU has avoided some of the most troubling aspects of AG Kokott’s Opinion. Most importantly, in Bougnaoui, it ruled that an employer’s desire to accommodate the wishes of a customer to not be served by a person wearing an Islamic headscarf is not a genuine and determining occupational requirement (para 41). Moreover, perhaps listening to the vicious criticism of Achbita, it ruled in I.X. v Wabe that a company rule banning the manifestation of any political, philosophical, and religious belief is proportionate only when it is “strictly necessary in view of the adverse consequences that the employer is seeking to avoid by adopting that prohibition” (para 69).

Yet despite the stricter language, companies may still cater to the prejudiced views of customers who do not want to be served by Muslim employees if they can demonstrate that they would otherwise face adverse consequences. This is at odds with the objectives of anti-discrimination law. Essentially, the Court made one of two mistakes: either it wrongly decided that neutrality policies, adopted to accommodate customers, are not directly discriminatory; or, if such policies are indeed not directly discriminatory, it wrongly ruled that they serve a legitimate aim and are not indirectly discriminatory.

Regarding direct discrimination, the Court has tried to distinguish neutrality policies from the Feryn case. This case “concerned direct discrimination based on race or ethnic origin that allegedly arose from discriminatory requirements on the part of customers” (I.X. v Wabe, para 66). However, the clothing requirements were cloaked in neutral language, but, as their justification shows, they were put in place to cater to discriminatory customer requirements. Following Feryn, they should therefore have been found to discriminate directly on religious grounds.

Moreover, if the CJEU was correct that a neutrality policy adopted to accommodate the wishes of customers does not amount to direct discrimination but possibly only indirect discrimination, it should still have ruled against them. Indirectly discriminatory policies are justified only if they serve a legitimate aim and if they are appropriate and necessary in view of this aim. Any neutrality policy adopted to meet the requirements of customers should never pass the first stage of the justification analysis. Allowing companies to cater to the prejudices of customers would be contrary to the very ideal of anti-discrimination law. Therefore, the problem is not primarily that the CJEU did not assess the necessity of the company’s neutrality policies sufficiently strictly, as it is sometimes argued, but rather that such policies should never be considered to serve a legitimate aim.

When caution would be justified

While the headscarf rulings represent a problematic departure from anti-discrimination law, in respect to one ruling, I can agree with McCrea that the CJEU’s approach may be one of justifiable caution. Four of the five cases involved private employers, but one, OP v Commune d’Ans, involved a public employer: the municipality of Ans in Belgium. The CJEU’s analysis of the municipality’s neutrality policy was lenient, and I can understand why (although it could have spelt out its reasons much more clearly).

The first reason is Article 4(2) TEU (national constitutional identity) combined with Article 17(1) TFEU, according to which the EU “respects and does not prejudice the status under national law of churches and religious associations or communities in the Member States”. According to McCrea, Article 17(1) TFEU offers sound legal reasons for caution in all cases pertaining to religion, but in my view, it does so only when the CJEU is asked to rule on a rule or practice that emanates from a national conception on the relationship between religion and the state. In other words, Article 17(1) TFEU may justify caution in relation to public employers, but not in cases where private employment practices are involved.

In this respect, it is worth reminding ourselves of the fact that even France and Belgium considered that the neutrality policy of the company in the Achbita case could not pass a proportionality assessment (AG Kokott Opinion, para 63). Applying the principles of laïcité and neutralité, both countries prohibit public servants to wear religious clothing at work, but these principles do not extend to the private sector (AG Sharpston Opinion in Bougnaoui, para 42). In other words, the CJEU decided to tolerate more “neutrality” than even France found acceptable. In the context of anti-discrimination law, this seems like a very poor choice.

Moreover, the fact that public sector neutrality policies stem from domestic conceptions on the appropriate relationship between religion and the state may also be relevant to their classification under anti-discrimination law. Private employers typically adopt such policies to accommodate the wishes of customers or colleagues. As I argued above, such policies should be considered to be directly discriminatory because they directly pander to discriminatory attitudes. Public employers applying constitutional principles of laïcité and neutralité do not necessarily do so (although these principles are often weaponized against Muslims nowadays). Therefore, it may be incorrect to classify neutrality policies of public employers as direct discrimination. While they may still discriminate indirectly, the more open-ended justification available for indirect discrimination, along with Article 17(1) TFEU, might indeed justify caution as regards public employers.

For similar reasons, McCrea may be right that the CJEU should have been more cautious in Egenberger and IR v JQ. Although I believe that both verdicts were legally defensible, it is certainly true that a different, more cautious, conclusion could have been justified under Article 17(1) TFEU and Article 4(2) of the Equal Treatment Directive. Space is too limited to delve into these cases (see here and the contributions of Vickers and Wendel and Geiger to this Symposium), but since EU law allowed, and maybe even push for, a large degree of caution, this might have been the politically wiser approach.

Political context supports this conclusion. The degree of autonomy enjoyed by religious organisations under German law meant that the German churches, with the backing of the German Constitutional Court, were able to discriminate on a large scale. Religious organisations are vital to the German welfare state, running hospitals, kindergartens, nursery homes, and the like. As a result, the two main churches are Germany’s second-largest employer after the state. They could demand that all their employees, including those not performing religious functions (such as doctors), join the church. However, in recent years, the churches have abolished the requirement of church membership as an absolute prerequisite for employment. Nowadays, membership is only required for roles involving proclamation, pastoral care, or education, or positions requiring the promotion of the church’s religious profile (Part I of the decision of the German Constitutional Court in Egenberger). As this greatly reduces the discriminatory practices of both churches, a degree of caution by the CJEU can indeed be justified.

By contrast, Muslims are facing an increasingly hostile environment in Europe, including in the workplace. In my view, for the EU’s highest court to stand by and, through its judgments and language, legitimise the exclusion of Europe’s Muslim population from important aspects of daily life is not a sign of caution, but of cowardice.