September 08, 2016
By Lucy Vickers, Oxford Brookes University
In this post, I focus on two issues of note regarding the divergent reasoning of the Advocates General. The first is the question of whether or not religion is immutable, and whether the answer to that question is helpful in determining the extent to which religion should be protected at work. The second is the use of ‘margin of appreciation’ reasoning, drawn from human rights case law on freedom of religion and belief, in the context of CJEU equality law.
The role of choice in determining religious rights
The question of whether religion is a chosen characteristic often arises when considering the proper scope of religious equality provisions. Those who argue that it is chosen usually use this to suggest that religion should either not be protected at all (because religious individuals chose their religion and so can chose not to be disadvantaged by their religion too), or that religion should receive less protection than other, immutable, characteristics such as sex, race or sexual orientation.
This debate becomes relevant when assessing whether the different equality grounds should be treated in the same way in law. For example, should it be as difficult to justify religious discrimination as it is to justify race or gender discrimination? If religion is not chosen, it should, presumably, be treated the same as other grounds; if it is chosen, this may justify affording it lesser protection.
In Achbita AG Kokott’s decision that the wearing of a headscarf by an employee could lawfully be restricted by the employer was based in part on her approach to the question of choice. In her view, religion is different to other characteristics because
the practice of religion is not so much an unalterable fact as an aspect of an individual’s private life, and one, moreover, over which the employees concerned can choose to exert an influence… an employee may be expected to moderate the exercise of his religion in the workplace…
Although religious observances can involve an element of choice in practice, nonetheless, the question of whether religion is a chosen characteristic can be contested. First, some aspects of religious identity are clearly not chosen. Unfavourable treatment on grounds of religion can be based on others’ perceptions of religion, such as because of a religiously-associated name, or because of assumptions made by the discriminator, linked to ethnicity, matters which are not chosen by the ‘victim’ of the discrimination. In such cases it would be unfair to restrict protection on the basis that religion is not immutable.
Even in cases of religious observance such as the wearing of religious symbols that was in issue in these cases, the choice argument can also be questioned. For example, few adherents experience their religion as chosen; and even where this is the case, the cost to the individual of renouncing key aspects of identity and culture is high. Moreover, the assumption that religion is different because it is chosen suggests that other grounds of equality are not chosen. Yet such a claim is perhaps not so simple: at times, other characteristics can involve an element of choice, such as some cases of pregnancy. In these cases, choice is not accepted as a reason to deny protection, because any such choice is a ‘fundamental choice’, and closely related to an individual’s concept of identity and self-respect. It would seem that the same should apply to religion.
Indeed, this latter view is taken by AG Sharpston in Bougnaoui, who notes that
to someone who is an observant member of a faith, religious identity is an integral part of that person’s very being… it would be entirely wrong to suppose that, whereas one’s sex and skin colour accompany one everywhere, somehow one’s religion does not. 
Moreover, AG Sharpston follows the reasoning of AG Maduro in Coleman who firmly bases the protection for equality on all grounds, including religion and belief, on autonomy and self-fulfilment. In endorsing Maduro’s reasoning, AG Sharpston confirms that the protection for religion and belief should be based on notions of equality and autonomy, and not limited by the question of choice or immutability.
It is noteworthy that AG Sharpston does not suggest that religion is necessarily immutable, merely that it is an integral part of identity. This allows space for the view that religion is in some senses chosen. This is important, because despite the criticism above, a recognition of the element of choice in religious practice may be needed if we are to explain why religion is a protected characteristic in the first place. One of the strongest reasons for the protection of religion in human rights law is that of autonomy and the freedom to choose one’s own conception of the good life. If religion is understood as unrelated to choice, then some of the underlying reasons for its protection are weakened.
The role of choice in the protection of religion is thus somewhat paradoxical in that religion acquires its value from the notion of autonomy which is based on choice; but equally for many individuals religion is not experienced as freely chosen. Indeed perhaps this paradox explains such different approaches to the role of choice and immutability in the opinions of the AG’s in the two cases.
In my opinion the reasoning of AG Sharpston is to be preferred, with her recognition that for religious observers, religious identity is integral to a person’s very being. Thus religion’s role in maintaining identity and autonomy is accepted, even if at some deeper level religion represents the exercise of free choice.
Recognition that religion is integral to identity does not lead inevitably to religion prevailing over other interests; the proportionality assessment may still result in religion being restricted. However, recognition that religion is not simply a chosen behaviour does allow more appropriate weight to be given to religious interests in any proportionality balancing exercise. In contrast, a decision that religion can be left ‘at the door’, or that religious staff should moderate the exercise of religion in the workplace (Achbita ), is likely to lead to insufficient protection.
The use of ‘margin of appreciation’ reasoning
In the context of human rights, the ECtHR routinely relies on the ‘margin of appreciation’ in deciding religious freedom cases to reflect different national contexts. This accounts for the fact that the refusal of an employee to be allowed to wear religious symbols at work can be both acceptable (Ebrahimian v France) and unacceptable (Eweida v UK) in human rights terms. In contrast, the CJEU case law on equality aims to set standards to eradicate inequality, rather than reflect different national traditions.
One of the questions arising in religion cases under the Directive is whether the CJEU will incorporate a similar notion to the margin of appreciation into its equality jurisprudence in religion cases. Of course, the margin of appreciation will not directly apply in the context of the Directive, but the process of reciprocal interpretation as between the CJEU and the ECtHR could mean that a similar concept could be used.
A particular concern arises regarding a possible reliance on ‘margin of appreciation’ style reasoning in the context of equality law, because equality law has usually been developed to eradicate entrenched inequality. Given that it would seem inconceivable that a court would allow a state to argue that national traditions should be allowed to justify sex or race discrimination in employment, it is questionable whether such reasoning should be accepted in the different context of religion.
The opinions in Achbita and Bougnaoui both take different approaches to the question of whether the national context should be taken into account in assessing the proportionality of any restrictions on religious practice. In Achbita, AG Kokott concludes that some discretion is needed for States when applying the Equality Directive, stating that one of the factors to which the court should have regard in assessing proportionality is ‘the national identities of Member States inherent in their fundamental structures’. This reasoning is entirely consistent with the approach of the ECtHR as it has consistently respected the individual States’ national identities in allowing restrictions on religious symbols at work in cases such as Sahin v Turkey, and Ebrahimian v France. However, the concerns noted above regarding the dangers of allowing established, but unequal, national practices remain following this approach. An additional concern is that such reasoning could then infiltrate the case law on other grounds, leading to the ‘levelling down’ of protection for other equality grounds.
In contrast, AG Sharpston in Bougnaoui is clear that different standards of protection should not be applied to different equality grounds. By implication then, if national identity would not be allowed to justify gender discrimination, nor should it justify religious discrimination. Moreover, her strong words at  regarding the danger of businesses relying on customer prejudice to justify discrimination, suggests that arguments based on established practice should not be accepted as justifying a refusal to accommodate religious manifestations at work. Instead, her assumption is that in the vast majority of cases it will be possible for employers and employees to reach an accommodation allowing the employer’s business needs to be met while providing for the manifestation of religion. The fact that the need to make a profit can prevail over an individual employee’s right to manifest religion only ‘in the last resort’, leads to the assumption that in most cases there will be no valid business reasons to restrict the wearing of religious symbols at work, particularly where these are based on general issues such as national identity.
My focus here has been on the implications of the very different approaches to the question of immutability and to the use of a ‘margin of appreciation’ in the two AG’s opinions. One factor that links these two issues is the question of whether the EU Court will countenance any formal hierarchy as between different grounds of equality.
In its earlier decisions, such as Chacón Navas  the CJEU has taken the view that there should be a common understanding of equality, with common standards introduced across the different grounds. This approach, also taken by AG Sharpston, guarantees significant protection for religious equality in the workplace.
However, in Achbita AG Kokott suggests grounds for treating religion and belief differently based on concepts of choice and immutability (at  and ). A benefit of accepting different approaches for different grounds of equality is that it allows any reasoning based on national identities and traditions to be limited to religion and belief only, without having to risk any levelling down of protection for other grounds.
Nonetheless, in my view the approach of AG Sharpston is to be preferred, because it confirms the centrality of the principle of equality to EU law: indeed at  she asserts that non-discrimination on grounds of religion or belief is a general principle of EU Law. Such an approach still allows some area of judgment to the Court, but if all equalities are to be treated equally, the focus will return to a question of proportionality as the decider of the issue of when religion and belief can be restricted at work. Such an approach allows for fact sensitive decision making, focussed on the interests of employer and employee, rather than more general, theoretical and at times paradoxical matters related to historical national context and the role of choice in religion.