European Court of Justice Allows Bans on Religious Dress in the Workplace

Headshot - Brems.jpg

Eva Brems

Ghent University


For decades now, the Islamic headscarf has been a topic of societal debate and legal disputes in several European countries. Headscarf bans (often formulated as either bans on headgear or as general bans on religious symbols or dress) have been introduced by schools, public authorities, service providers (such as sports facilities and restaurants), and private employers. This phenomenon is most widespread in France and Belgium. In many cases, bans have passed domestic judicial review. Bans in public schools and by public employers have moreover been accepted by the European Court of Human Rights. The latter Court has not yet ruled on a headscarf ban in the workplace. It has however found that banning the wearing of a Christian cross in the workplace can be a violation of religious freedom.

In addition to religious freedom, the prohibition of discrimination on grounds of religion is an obvious basis for judicial review of headscarf bans. As the anti-discrimination legislation in EU member states is to a large extent an implementation of EU law, it was only a matter of time before the issue would come before the European Court of Justice through the procedure of preliminary review. On March 14 the ECJ ruled on two separate headscarf queries from domestic courts concerning the interpretation of the EU Equal Treatment Directive. The case of Achbita was referred by the Belgian Court of Cassation, and the case of Bougnaoui by the French Court of Cassation. Both cases concerned women working in jobs in which their company sent them out to work for client companies. Both women lost their jobs on account of their refusal to comply with their employer’s request that they take off their headscarf in the workplace. Ms Achbita started wearing a headscarf at work in her fourth year of employment, and was immediately met by her employer’s refusal, based on the company’s ‘neutrality’ policy, which was only later (shortly before she was dismissed) incorporated in the employee code of conduct. Ms Bougnaoui on the other hand was asked to remove her headscarf at work after the company received a complaint from a client that her ‘veil’ had ‘embarrassed’ a number of its employees.

Interestingly in light of the strong similarity of the cases, the opinion of Advocate General Kokott in Achbita and that of Advocate General Sharpston in Bougnaoui reached opposite conclusions, based on diverging interpretations of several aspects of the Directive. For Kokott, there was no discrimination, for Sharpston, there was. For an analysis of the differences in their interpretations, see herehere and here. Thus, while the succinct style of the ECJ judgments does not betray this, it is clear that the ECJ was aware that it stood before a principled choice, in a field characterized by strong differences of opinion, including among experts of EU law.

Summary of the Judgments

In the case of Bougnaoui, where the headscarf ban was not part of a general neutrality policy, and was based on client preference, the ECJ found (indirectly, through its answer to the question of the referring judge) a violation. It ruled that this was a case of direct discrimination on grounds of religion, which can only be justified in case of a ‘genuine and determine occupational requirement’ (art. 4(1) of the Directive). The ECJ ruled that

The willingness of an employer to take account of the wishes of a customer no longer to have the services of that employer provided by a worker wearing an Islamic headscarf cannot be considered a genuine and determining occupational requirement. (para. 41)

In the case of Achbita however, the ECJ did not find a violation. It held that there was no direct discrimination, because

the internal rule at issue … refers to the wearing of visible signs of political, philosophical or religious beliefs and therefore covers any manifestation of such beliefs without distinction. The rule must, therefore, be regarded as treating all workers of the undertaking in the same way by requiring them, in a general and undifferentiated way, inter alia, to dress neutrally, which precludes the wearing of such signs. (para. 30)

The ECJ then continued on to consider the matter of indirect discrimination, which under the Directive can be justified if the measure is objectively justified by a legitimate aim and if the means of achieving that aim are appropriate and necessary.

The ECJ accepts ‘the desire to display, in relations with both public and private sector customers, a policy of political, philosophical or religious neutrality’ as a legitimate aim, linked to the freedom to conduct a business (Charter of Fundamental Rights of the EU, art 16).

It further considered the ban appropriate for the purpose of pursuing a policy of neutrality ‘provided that that policy is genuinely pursued in a consistent and systematic manner’.

Finally, the ECJ ruled that the ban can be considered ‘strictly necessary’ if two conditions are met. One, it should cover only the workers who interact with customers. Two, dismissal is allowed only if it is not possible (taking into account ‘the inherent constraints to which the undertaking is subject’ and without taking on an ‘additional burden’) to offer the applicant a post not involving any visual contact with customers.


While the Bougnaoui judgment is in line with previous case law ruling that client preference does not legitimate discrimination, the Achbita judgment surprises – and can be criticized- on a number of points.

In the first place, it is astonishing that, in the eyes of the European Court, direct discrimination on grounds of religion or belief exists only when a measure targets a single religion or a selection of religions, but not when a measure targets all religions and beliefs. Generalized hostility toward religions is apparently a manifestation of neutrality. Not all individuals have a (strong) religion or belief that they would ever consider expressing in the work place. That is to say, even hostility against all religions and beliefs targets only part (arguably a small minority) of the employees. In that sense it is comparable to the discrimination ground ‘disability’ (as opposed to ‘sex’, ‘age’ or ‘ethnicity’, which are characteristics of all individuals). And yet it seems unthinkable that the ECJ would rule that a measure excluding without distinction persons with all kinds of disabilities would not constitute direct discrimination. What is different about religion? Has anti-religious sentiment become so normalized in Europe, that it has become invisible – the new normal?

Second, it is problematic that the ECJ accepts the expansion of ‘neutrality’ reasoning into the private sphere without the least degree of scrutiny. Neutrality has been accepted as a ground for rights restrictions when it is applied in the public sector, as it is not debated that neutrality is a duty for public authorities. But extending it to the private sector is a big leap. The link to the freedom to conduct a business does not solve the matter, since this freedom is by no means absolute, and would for example not include the freedom to conduct a business in a discriminatory manner – which is what is alleged in the case. Neutrality can be an easy cover-up for prejudice. When neutrality coincides with corporate image, an oxymoronic subjective neutrality is legitimized.

Third, it is hard to understand how the Court can come to the conclusion that the measure is ‘strictly proportionate’ to achieving the aim of neutrality, without even considering whether less restrictive measures could have realized the goal of corporate neutrality. In that respect, it is important to note that across Europe, the ‘exclusive’ view of neutrality (neutrality as rejecting manifestations of all religions on an equal footing) coexists with an ‘inclusive’ view of neutrality (neutrality as accepting manifestations of all religions on an equal footing). One could have expected the Court to at least consider the second scenario. In addition, the European Court of Human Rights, in its Eweida judgment accepts corporate image as a ‘legitimate aim’ for rights restrictions (which is referenced by the ECJ in Achbita, para. 39), yet grants it only little weight when put in the balance against the religious freedom of the applicant (Eweida, para. 94). The ECJ on the other hand does not even mention the applicant’s religious freedom or the importance of the headscarf for her. It is hard to accept that the ECJ has genuinely conducted a proportionality assessment if the judgment does not show any evidence of the weight that is attributed to the interests on one side of the balance. Moreover, the result of what the ECJ presents as its proportionality analysis, is that headscarf wearers are relegated to back-office jobs. In other words, they are closeted, made invisible. Set against the recent history of closeting of LGBT people in Europe, I consider it extremely cynical that the ECJ today presents the closet as an acceptable compromise for Muslim women.

Finally – and this is a point of critique of the ECJ’s style that transcends this case – it its bewildering to read a judgment of a supranational court ruling on fundamental rights that discusses the issue of  Islamic headscarf bans without any reference to either the Europe-wide context of Islamophobia, or the widespread existence of negative stereotypes about Muslim women, and in particular those who wear Islamic dress. Maybe the judges have a strong preference for French-style rulings, or maybe they could not agree on any language that would show the least degree of recognition of the plight of Muslim women in Europe. In any case, what some may perceive as remaining on ‘neutral technical ground’ is for others an expression of an ivory tower mentality. As illustrated by the following anecdote: when ECJ President Lenaerts read the judgment aloud in his native language Dutch, he read ‘Islamist headscarf’ instead of ‘Islamic headscarf’. In Dutch, this is only one letter extra (Islamitisch/Islamistisch), yet anyone who has some degree of familiarity with debates on Islam is aware of the crucial difference. Not the ECJ President, apparently.


The ECJ judgment imposes a uniform interpretation of the Equal Treatment Directive on this point. Across the EU, it will not be possible to use anti-discrimination law against bans on religious symbols in the private sector that comply with the conditions set out in Achbita. In that sense, the judgment can be read as a ‘how-to’ for employers wishing to discriminate against headscarf wearers: introduce a neutrality policy that applies to all types of religious dress; apply it consistently; apply it only to front-office employees; and if you want to dismiss a person, make sure to motivate why you cannot offer that person a back-office job.

Yet the judgment does not apply to challenges to headscarf bans that are based on the freedom to manifest one’s religion. So, the ‘how-to’ for persons who want to challenge a headscarf ban in the private sector is the following: argue a violation of religious freedom rather than a discrimination, and refer to Strasbourg (Eweida) rather than Luxembourg.

By Eva Brems, a Professor of human rights law at Ghent University (Belgium). She is a member of the Executive Committee of the IACL.