A typology of reverse discrimination in EU citizenship law

Abstract

This Article sheds new light on one of the longest-running debates in the European Union (EU) citizenship literature: the concept of ‘reverse discrimination’ and the question of whether it is justified. Reverse discrimination has divided EU lawyers into roughly two distinct groups. One group believes that it constitutes an unjustified violation of the principle of equality; a second that it is inevitable in a Union governed by the constitutional principle of divided powers. This Article questions this by offering a typology of reverse discrimination. While most scholars assume that reverse discrimination is a singular phenomenon that demands a singular response, this Article shows that it is a variegated phenomenon that demands a variegated response. It distinguishes three types of reverse discrimination and explains that the proper response depends on the type we are considering. Type I is caused by the application of the principle of mutual recognition; Type II by an interaction between domestic federalism and internal discrimination; and Type III by the CJEU’s confusion over the aim of the right to free movement and residence. Through this typology, the Article shows that reverse discrimination is never a corollary of the principle of divided powers, nor is it always incompatible with the principle of equality. Finally, the Article shows that to the extent that reverse discrimination violates the principle of equality, the solution is not to equalise rights upwards but downwards to the lower (national or regional) level of government. This shows that the principle of equality and the principle of divided powers need not collide.

1. Introduction

This Article sheds new light on one of the longest-running debates in the EU citizenship literature: the concept of ‘reverse discrimination’ and the question of whether it is justified. Reverse discrimination is the concept used to describe situations where static EU citizens are worse off than mobile EU citizens. The position of static citizens (eg, German nationals living in Germany) is regulated exclusively by national law.¹ This is so because EU free movement law does not apply to so-called ‘purely internal situations’,² that is, situations whose facts are ‘confined in all respects within a single Member State’.³ In contrast, the position of mobile citizens – citizens who exercise their right to move and reside freely under EU law – is regulated not only by national law but also by EU free movement law. This can result in reverse discrimination, namely when mobile citizens enjoy more favourable rights under EU free movement law than static citizens can claim under national law. Reverse discrimination is a widespread phenomenon in EU citizenship law, occurring in areas as diverse as family reunification law, social security law, and the recognition of personal statuses. It is also a controversial phenomenon that has been subject to extensive debate. However, by providing a typology of reverse discrimination, this Article shows that the concept remains poorly understood. The conceptual clarification that this typology provides should allow us to better evaluate reverse discrimination.

Reverse discrimination has divided EU lawyers into roughly two distinct groups.⁴ According to the first group, reverse discrimination constitutes a violation of the principle of non-discrimination on grounds of nationality and, more generally, the ideal of equality – an ideal embodied by EU citizenship. Not long ago, Advocate General Wahl, in his Opinion in Austria v Germany, argued that requiring German taxpayers to fund the German motorway network on their own ‘would be an unreasonable (and possibly even perverse) interpretation of Article 18 TFEU⁵: rather than avoiding discrimination against non-nationals, it would de facto impose reverse discrimination against nationals’.⁶ This view echoes that of other Advocates General (AGs),⁷ as well as a large number of EU citizenship scholars who have taken the position that there is no better ‘argument than that built on Union citizenship’ for prohibiting reverse discrimination.⁸

According to the second group, reverse discrimination is inevitable in a Union governed by the constitutional principle of divided powers. Scholars who take this view may agree that there is a tension between reverse discrimination and the ideal of equality, but they disagree with the solution favoured by many in the first group: that the Court of Justice of the European Union (CJEU or Court) must revisit the purely internal rule, extending the privileges enjoyed by mobile citizens to static citizens.⁹ Scholars in the second group think this solution is incompatible with the EU’s division of powers. For them, reverse discrimination is ‘the result of application of the purely internal rule’,¹⁰ which is ‘a suitable instrument to meet the constitutional necessity of respecting the division of powers between the Union and its Member States’.¹¹ It is, as Iglesias Sánchez put it, ‘the flip side coin of purely internal situations and an intrinsic by-product of the principle of attributed competences’.¹²

The disagreement running through the reverse discrimination debate thus seems to be a result of ‘colliding constitutional principles’, to concern the question of whether the principle of equality or the principle of divided powers must carry more weight.¹³ This Article questions this narrative by offering a typology of reverse discrimination. While most scholars assume that reverse discrimination is a singular phenomenon that demands a singular response, this Article shows that it is a variegated phenomenon that demands a variegated response.¹⁴ Looking at the underlying causes, this Article distinguishes three types of reverse discrimination and shows why and how this distinction matters for the justification of reverse discrimination, ie, its compatibility with the above-mentioned constitutional principles of equality and divided powers. In other words, this Article shows that a befitting response to practices of reverse discrimination requires a more adequate understanding of its underlying causes.

By developing this typology, the Article improves (i) our understanding of the causes of reverse discrimination, (ii) our ability to evaluate reverse discrimination, and (iii) our understanding of the correct remedy to problematic forms of reverse discrimination. First, the Article shows that reverse discrimination has three distinct causes. Type I is caused by the application of the principle of mutual recognition (section 2). Type II is caused by domestic federalism and internal discrimination (section 3). Type III is caused by confusion over the aim of free movement (section 4). Second, as for the evaluation of reverse discrimination, the Article shows that neither of the two dominant views has got it right. First, reverse discrimination is not the flip side of the EU’s division of powers; it will be shown that all three types can be abolished without compromising this constitutional principle (whether that is desirable is another matter). Second, however, neither the principle of equality nor the concept of EU citizenship offers a compelling justification for the abolition of the purely internal rule. Of the three types, only the third type of reverse discrimination is truly problematic. However, third, as for the appropriate solution, it will be shown that equality between static and mobile citizens should be ensured not by abolishing but by respecting the purely internal rule. This solution shows that the principle of equality and the principle of divided powers need not collide.

Before proceeding, a clarification about a recurring theme of the Article is in order. Federalism is the theme that runs like a thread through the argument; the question that emerges in the discussion of each of the three types of reverse discrimination, and perhaps even ties the argument together, is when the EU as a federal-like entity should intervene in, and how it should position itself vis-à-vis, the (federal) constitutions of Member States. Type I reverse discrimination is the direct result of the EU’s construction of a federal market; Type II relates to how Member States can shape their domestic (federal) structures within the contours of this market; Type III is an example of federal overreach into the national (federal) orders. Nevertheless, the Article has no ambition to enter the long-standing debate of whether the EU has become a federal union of states in all but name.¹⁵ However, regardless of our view on that question, the Article shows that federalism can offer a promising angle from which to analyse and understand specific features of the EU’s constitutional order, its interaction with the domestic legal orders, and the challenges this interaction poses.

2. Type I: Reverse discrimination caused by mutual recognition

The oldest form of reverse discrimination finds its origins in the internal market and is caused by the application of the principle of mutual recognition. This principle is meant to foster the free movement of goods, services, and persons by obliging Member States to recognise goods, services, and personal statuses lawfully produced, provided, or acquired in other Member States, unless the former can provide a legitimate justification for non-recognition.¹⁶ For example, this principle allows goods produced in a Member State with lenient product standards (MS A) to be marketed in another Member State with stricter standards (MS B). MS B cannot require imported goods to meet its regulatory standards but must accept as lawful the sale of products manufactured in accordance with MS A’s standards. And because manufacturers in MS B cannot benefit from MS A’s more lenient product standards, they are being reversely discriminated against compared to manufacturers in MS A.¹⁷ Indeed, the application of the principle of mutual recognition may place companies that trade across borders, as well as EU citizens who move across borders, in a privileged position when compared to companies or citizens in the country of destination.¹⁸

As the case law on recognition of names shows,¹⁹ the principle of mutual recognition has found its way into EU citizenship law. Grunkin and Paul concerned Leonhard Matthias, the son of Dr. Paul and Mr. Grunkin. Leonhard Matthias was born in Denmark but possessed German nationality, as did his parents. His parents had registered the surname Grunkin-Paul in Denmark and requested the German authorities to register their son, who resided with the mother in Denmark but often stayed with the father in Germany, under the surname registered in Denmark. The German authorities refused, insisting that since nationality was the sole connecting factor, only one of the surnames, Grunkin or Paul, could be accepted.²⁰ The parents challenged this decision invoking EU citizenship law. The CJEU ruled in their favour, deciding, in line with previous case law,²¹ that national rules disadvantaging Member State nationals simply because they exercised the right to free movement constitutes a restriction of Article 21 TFEU. In other words, the Member State of nationality may not treat its own nationals less favourably because they have resided and acquired rights in another Member State. Such is the case for someone ‘having to use a surname, in the Member State of which the person concerned is a national, that is different from that conferred and registered in the Member State of birth and residence’.²² The CJEU further ruled that the justifications brought forward for using nationality as the sole connecting factor could not be accepted, since this would undermine the continuity and stability of the personal status in question. In other words, in the absence of a legitimate public interest justification, the Member State of destination must recognise the personal status acquired by the citizen in accordance with the law of the other Member State. It must thus also treat that citizen more favourably than its own nationals who have no connection with that other state and, hence, EU law (unless there were a change in its domestic laws). It must, in other words, accept reverse discrimination between static and mobile citizens.

Reverse discrimination through mutual recognition can also occur because EU citizens retain their nationality when they settle in another Member State.²³ As a result, they may possess membership of two Member States: the Member State of nationality and the Member State of residence by virtue of the right to non-discrimination. García Avello showed that this is not without consequences. Esmeralda and Diego, the children of Mrs. Weber and Mr. García Avello, were born in Belgium where they had lived all their lives. They had dual Spanish–Belgium nationality. The Belgian authorities entered the children in the national registers under the surname García Avello. Meanwhile, the children were registered in Spain as García Weber. The father’s request to change the surname to García Weber, in accordance with Spanish law, was rejected by the Belgian authorities. In response to the question of whether this decision was contrary to the provisions on Union citizenship,²⁴ the Court ruled that a link with EU law exists for persons ‘who are nationals of one Member State lawfully resident in the territory of another Member State’.²⁵ This being the case, the applicants could rely on the right to non-discrimination on grounds of nationality, enshrined in Article 18 TFEU. After concluding that Belgian citizens with dual Spanish nationality find themselves in a different position from those who only have Belgian nationality,²⁶ the Court decided, ‘[i]t is common ground that such a discrepancy in surnames is liable to cause serious inconvenience for those concerned at both professional and private levels’.²⁷ Belgium was obligated to recognise the Spanish surname.

While critics of reverse discrimination have challenged the purely internal rule (see introduction), it was ironically the watering down of this rule in García Avello that exacerbated reverse discrimination.²⁸ Traditionally, the principle of mutual recognition applied to ensure that barriers to free movement were removed.²⁹ However, Belgium was required to recognise the Spanish surname, and treat the Spanish–Belgium children more favourably than it would have treated children with only Belgian nationality, even though they had never exercised the right to free movement. Belgian nationals with divergent surnames, by reason of the different laws to which they are linked by their nationality, may plead difficulties specific to their situation that distinguish them from persons holding only Belgian nationality.³⁰ In the absence of clear rules, the ECJ has provided EU citizens with choice of law autonomy if two or more laws are equally applicable in a given case.³¹ Member States are subsequently obliged to recognise this choice.

In any event, the above two cases illustrate how the application of the principle of mutual recognition may cause reverse discrimination in EU citizenship law: EU citizens who exercised the right to free movement or with a dual nationality may claim certain names that static citizens may not. According to Tryfonidou, however, differences caused by mutual recognition are not actually reverse discrimination. For her, it is a side-effect of the desire ‘to compensate persons/traders taking part in the process of European economic integration for disadvantages they would have suffered as a result of exercising one of the fundamental freedoms’.³² And therefore, she prefers to classify such situations as a realisation of substantive equality rather than the creation of reverse discrimination.³³ However, while the principle of mutual recognition may indeed alleviate certain burdens that the exercise of the right to free movement may entail, and therefore contribute to the realisation of substantive equality between mobile and static citizens,³⁴ this distinction between reverse discrimination and substantive equality will seem strange to readers familiar with the use of the term reverse discrimination elsewhere. This term is traditionally used to describe the practice of favouring historically disadvantaged groups with the purpose of addressing past injustices by the dominant group. As such, reverse discrimination is associated with affirmative action policies that seek to strengthen the position of vulnerable groups, such as women, minority religious or racial groups, or LGBT people. Following this definition, differences in treatment are reversely discriminatory only if their aim is to attain substantive equality. It is therefore interesting that EU lawyers have always used the term reverse discrimination in a broader sense, to include differences of treatment that do not contribute to producing substantive equality (the discussion of Type II and Type III below will show this).³⁵ And following this definition, differential treatment between mobile and static citizens through the application of the principle of mutual recognition clearly amounts to reverse discrimination, whether or not it realises substantive equality.

How should we evaluate reverse discrimination caused by mutual recognition? As we saw in the introduction, one group of scholars has condemned reverse discrimination as contrary to principles of equal citizenship, while a second group sees it as an inevitable consequence of the EU’s constitutional division of powers. Type I reverse discrimination already shows why neither position is plausible. Consider the free movement of same-sex couples and the related issue of mutual recognition of marriages and registered partnerships. There are important similarities between the recognition of different forms of personal status such as names and marital status.³⁶ First, like the law on the recognition of family names, the issue of the recognition of the marital status of same-sex couples ‘epitomizes the cultural identities underlying the most pertinent European conflicts cases’.³⁷ Moreover, by requiring recognition of same-sex marriage,³⁸ the CJEU created reverse discrimination between mobile and static same-sex couples.³⁹ Member States opposed to same-sex marriage are now under an obligation to recognise the marital status of same-sex couples who celebrated their marriage in another EU Member State while genuinely resident there,⁴⁰ but they remain entitled to deny same-sex couples who are their nationals and resident within their territory the right to marry.

It is not quite accurate to argue that reverse discrimination is a consequence of the EU’s division of powers. One can easily envision a polity with a division of powers between two or more levels of government without reverse discrimination. That would be a federal-type union with no principle of mutual recognition: one that applies the host rather than the home state principle, regulating citizens according to the laws of the host state. Consider the position of same-sex couples within the EU that are unmarried but have contracted a registered partnership. According to EU legislation, registered partnerships must be recognised only when ‘the legislation of the host Member State treats registered partnerships as equivalent to marriage and in accordance with the conditions laid down in the relevant legislation of the host Member State’.⁴¹ Imagine same-sex couples in a registered partnership that move to a Member State that does not know registered partnerships. The Member State will not, in accordance with the above provision, need to recognise the status, as its national law does not treat it as equivalent to marriage. Hence, the host state can apply its own standards and no reverse discrimination will arise. Or think about the United States, a federal union in which the central level of government has placed fewer constraints on the regulatory powers of the constituent states.⁴² Prior to Obergefell v Hodges,⁴³ the decision that legalised same-sex marriages in the US, the US Constitution was interpreted as leaving decisions on whether to recognise out-of-state same-sex marriages to the individual states.⁴⁴ Under those conditions, reverse discrimination created by recognition of individual status need not arise. The individual states can decide not to recognise marriages of same-sex couples celebrated under the laws of other states.

This is not to say this is desirable. Federal-type unions that aim to guarantee and facilitate the free movement of persons among the constituent states will want to ensure that the effects of movement on the personal statuses of citizens are, as far as possible, neutralised. In federal unions without a unified substantive law, demands for legal certainty and justice are therefore likely to result in the development of some form of an ‘interstate private law’.⁴⁵ However, as the two examples illustrate, reverse discrimination is not unavoidable in unions that divide powers between different levels of government. On the contrary, it can be avoided by not enforcing a principle of mutual recognition. Indeed, reverse discrimination occurs not due to a desire to safeguard the EU’s division of powers, but because of the intention to guarantee free movement under conditions of certainty and stability.

However, the EU cannot apply mutual recognition and combat reverse discrimination without violating its division of powers. To avoid Type I reverse discrimination, either mutual recognition or the division of powers must go. As we just saw, the EU can eliminate this type of reverse discrimination by not applying the principle of mutual recognition. But if it wants to apply this principle, for example to ensure that same-sex marriage is recognised throughout the EU, and also avoid reverse discrimination, it should abolish or ignore the purely internal rule to allow static same-sex couples to marry as well. This example illustrates the shortcomings of the view that reverse discrimination is an abomination in a citizens’ Europe, especially when it is argued that it is up to the CJEU to abolish the purely internal rule. It is beyond the CJEU’s powers to require Member States to extend marriage to same-sex couples and, more generally, to require them to accept the standards of the most lenient state any time a situation of reverse discrimination by mutual recognition arises. Indeed, the EU must accept reverse discrimination as a logical extension of the constitutional principle of divided of powers if it wants to enforce a principle of mutual recognition as part of free movement law.

As this shows, from an EU perspective, reverse discrimination caused by mutual recognition is not ‘logically unexplainable’ and ‘a big problem begging for resolution’.⁴⁶ On the contrary, it is the logical outcome of a wide interpretation of free movement rules that should ensure that free movement within a union with different regulatory regimes can happen under conditions of certainty and stability. If the EU is to ensure that certainty and stability, it must preserve the principle of mutual recognition. In that case, however, reverse discrimination must be accepted – not as an inevitable side-effect of the division of powers but as a consequence of the desire to guarantee that citizens can move freely across the EU. Of course, national actors may be able to resolve reverse discrimination by aligning their legislation with that of other Member States. For example, those that do not allow same-sex couples to enter into marriage could eliminate reverse discrimination by opening up marriage for same-sex couples. But such domestic solutions may not be available where national standards diverge more greatly, as in the case of the spelling of names. It thus seems that reverse discrimination Type I is there to stay, at least as long as the EU wishes to protect the free movement of its citizens by means of the principle of mutual recognition. It certainly seems, from the example of the recognition of same-sex marriage, that those who have argued for its abolition have not fully understood the extreme consequences of their arguments.

3. Type II: Reverse discrimination caused by domestic federalism and internal discrimination

The second type of reverse discrimination is reverse discrimination caused by the interaction of domestic federalism and internal discrimination. This type is most common in states organised along federal lines, but it can arise in any state with internal municipal subdivisions (ie, all Member States). By internal discrimination, I mean that regions within a Member State may treat national citizens from other regions within that same Member State differently from their own residents and from EU citizens from other Member States. As we will see, such internal discrimination can arise in two situations.

We find the first situation in the facts that later led to the Walloon case. While this case has been extensively discussed in the reverse discrimination debate, it has been overlooked is that it involved two different types of reverse discrimination. The academic debate has mostly focused on the reverse discrimination created by the CJEU in the Walloon judgement, but this is Type III reverse discrimination and will be discussed in the next section. The situation preceding the judgement involved reverse discrimination Type II, which arose after the Commission challenged a care insurance scheme established by the Flemish government.⁴⁷ Initially, the scheme covered only non-medical expenses incurred by persons residing in the Dutch-speaking region, who were obliged to join, or persons residing in the bilingual region of Brussels, who could join on a voluntary basis.⁴⁸ The Commission claimed that this violated Regulation 1408/71 EEC, according to which ‘a worker employed in the territory of one Member State shall be subject to the legislation of that State even if he resides in the territory of another Member State’.⁴⁹ On this basis, it argued that persons living in other EU Member States but working in the Dutch-speaking region or bilingual region of Brussels should be included in the scheme.⁵⁰ In accordance with EU law, the Flemish government extended the scheme’s personal scope to persons residing in other EU Member States but working in the Dutch-speaking region or bilingual region of Brussels.⁵¹ By contrast, residents of Wallonia (Belgian and non-Belgian EU citizens) who worked in Flanders or Brussels remained excluded from the scheme.

This inclusion of EU citizens living outside Belgium but working in the Dutch-speaking region or bilingual region of Brussels created reverse discrimination. Indeed, their inclusion put them in a more favourable position compared to residents of Wallonia working in these regions. This is the consequence of two contrasting principles that federally organised polities can use to condition the applicability of social security schemes: lex loci domicilii and lex loci laboris. Like many other federal regimes,⁵² social legislation in Belgium is conditioned by the principle of lex loci domicilii: those resident in the region providing benefits are entitled to inclusion.⁵³ EU legislation, on the other hand, applies the principle of lex loci laboris: those employed in the region providing benefits are eligible.⁵⁴ The lex loci laboris principle was meant to ensure equal working conditions for employees, but because the principle applies also to residence-based benefits such as the Flemish care insurance scheme,⁵⁵ reverse discrimination occurred.

A slightly different form of Type II reverse discrimination may no longer be relevant after Brexit but is nonetheless instructive to discuss: the reverse discrimination created by the Scottish university fee policy. Scottish residents can study at a lower tariff than English, Welsh, and Northern-Irish British nationals.⁵⁶ Previously, Scotland could not apply the higher tariff to EU citizens that arrive from other Member States, as the Court decided in Gravier that the principle of non-discrimination applies to tuition fees.⁵⁷ This interaction between the discriminatory Scottish policy and the EU non-discrimination principle led to reverse discrimination against non-Scottish British nationals.⁵⁸ Note that there is a slight difference between this situation and that of the Flemish care insurance scheme. Flanders treated all residents equally, meaning that being resident in Flanders was sufficient to be included in the care insurance scheme. Scotland, on the other hand, may charge English, Welsh, and Northern-Irish British nationals higher fees, even after taking up residence in Scotland. In both situations, however, there is reverse discrimination due to an interplay between domestic principles of (quasi-)federalism and norms of EU free movement law.

That clarified, the point of this discussion is that also Type II reverse discrimination shows that neither of the dominant positions on reverse discrimination – either that it violates principles of equal citizenship or that it is the inevitable consequence of the EU’s division of powers – is tenable. First, by contrast to Type I reverse discrimination, Type II reverse discrimination can be resolved without undermining the right to free movement of EU citizens. For instance, Scotland could treat university students from England, Wales, and Northern-Ireland in the same way as Scottish residents and other EU citizens. In other words, a domestic policy change could remedy the reverse discrimination in question without thereby disadvantaging mobile EU citizens. The same is most likely true of the reverse discrimination caused by the Flemish care insurance scheme. Belgium could in theory follow the EU’s example and condition the applicability of its social security schemes by the principle of lex loci laboris rather than lex loci domicilii, thereby remedying the reversely discriminatory effects of the interaction between the principle of lex loci domicilii at national level and the principle of lex loci laboris at EU level. As this shows, Type II reverse discrimination is not inevitable in a Union governed by a principle of divided powers; it very much depends on the political choices of domestic policymakers.

But it is obviously incompatible with the principle of divided powers for the EU to take such choices and impose these on the Member States. Those who argue that EU citizenship provides a justification for the reversal of reverse discrimination have paid too little attention to the limits of the EU’s powers. As Nic Shuibhne said in response to such views: ‘it is simply erroneous to argue that all Member State issues should, in fact, become Community issues. It doesn’t wash in terms of logic, nor does it take account of the deep seated social, cultural, political and financial objections that would oppose such a radical move in the first place’.⁵⁹ The reverse discrimination in the Flemish care insurance scheme case could perhaps be remedied by the EU legislature through an amendment of EU social security coordination legislation, by opting for lex loci domicilii rather than lex loci laboris as the coordinating principle.⁶⁰ While the division of powers does not stand in the way of such legal reform, it would entail such complex decisions that only the EU legislature is capable of instituting. But not only is the CJEU not capable of taking such decisions, it could also not justify such reforms by reference to, say, the concept of EU citizenship or the principle of equality. As reverse discrimination Type II thus shows, those who think that EU citizenship justifies a ban on reverse discrimination have not only missed the complexity of reverse discrimination, but have also misjudged the powers of the EU, and in particular its judiciary, to address this phenomenon.

4. Type III: Reverse discrimination caused by confusion over the aim of free movement

The third type of reverse discrimination is reverse discrimination caused by confusion over the aim of the right to free movement. Of the three types, this one is the most controversial and most often targeted by critics of reverse discrimination. Yet it might also be the type least well understood. Two important (lines of) judgements have led to reverse discrimination of this type: the Walloon judgement and the family reunification case law. Following a detailed discussion of these cases, it is shown that also Type III reverse discrimination demonstrates the deficiencies of the two dominant perspectives on reverse discrimination.

Not only is this type not a manifestation of the EU’s constitutional division of powers, but it even violates this division of powers, in a way that also amounts to an unjustifiable violation of the principle of equality.

Yet, contrary to what the argument from equal EU citizenship advocates, the solution is not to equalise rights upwards to the EU level (or national level in Walloon). If anything, the logical solution would be the equalisation of rights downward to the national level (or state level in Walloon).

The reason for this admittedly controversial conclusion is that the reverse discrimination that occurred with these judgements was unnecessary and entirely due to confusion on the CJEU’s part about the aim of the right of EU citizens to move and reside freely. This aim, as I understand it, is functional: to tackle unjustified obstacles to the exercise of the right to free movement, whether discriminatory or non-discriminatory.⁶¹ While this leaves ample room for different interpretations and applications of this right, it should, I think, be undisputed that it normally comes into play only when there are restrictions on free movement.⁶² Thus, it is settled case law that the fundamental freedoms do not apply to national measures that treat foreign goods or persons in the same manner, in law and in fact, as national goods or persons in the same circumstances.⁶³ Because such measures apply equally to foreign and national actors and do not otherwise obstruct free movement,⁶⁴ they do not conflict with the aim of free movement ‘to facilitate the pursuit by Union nationals of occupational activities of all kinds throughout the EU’.⁶⁵ Logically, therefore, the Treaty provisions on free movement do not capture national measures whose effect on cross-border movement is too uncertain or indirect, or entirely hypothetical.⁶⁶ They certainly do not apply where no limitation on the right to free movement exists at all. The latter is the case, for instance, when the law of the host state applies to EU citizens and the occupational activities they carry out only after they have exercised the right to free movement and established themselves in the host state.⁶⁷ It would be inconsistent with the aim of the right to move and reside freely to prohibit national measures that do not hinder cross-border movement and establishment but are otherwise unwanted by the person exercising the right to move and reside freely. Therefore, this right does not guarantee that cross-border movement will be neutral in terms of the rights and duties one enjoys.⁶⁸ In relation to social security, for example, the CJEU ruled in Konrad Erzberger that ‘given the disparities between the Member States’ social security schemes and legislation, such a move may be more or less advantageous for the person concerned’.⁶⁹ The Treaty provisions on the right to free movement are not intended to challenge or rule out such regulatory differences.

Of course, the right to free movement may be valuable for other reasons than removing restrictions on cross-border mobility. It can enable EU citizens to escape moral or economic constraints in their home states or, some argue, have their interests represented in the host state’s political process.⁷⁰ But these would merely be valuable side-effects of the right to free movement, which remain secondary to the functional aim of tackling unjustified restrictions to free movement. The phenomenon discussed in this Article attests to this. As Tryfonidou notes, ‘discrimination against persons who have not exercised their free movement rights – known as ‘reverse discrimination’ – has always been held by the Court to fall outside the scope of the free movement provisions, because it appears incapable of impeding the achievement of their aims’.⁷¹ The judgements discussed below also bear witness to this. The CJEU adjudicated the Walloon and family reunification cases as it did because it deemed this necessary to remove obstacles to the right to move and reside freely – indeed to realise the functional aim of this right. As we will see, however, it was wrong to assume that this right was at stake, which is why it unjustly produced reverse discrimination in these cases.

Before explaining why, I should clarify that I consider it not only descriptively true, but also normatively desirable, to limit the application of the right to free movement to obstacles to the exercise of free movement. Admittedly, what counts as an obstacle is controversial, but I take it as uncontroversial that the aim of the right to free movement is not to make EU citizens’ place of residence as attractive as possible. We just saw, for example, that this right does not guarantee that cross-border movement is neutral as regards the rights EU citizens enjoy. It would then also not allow me, a Dutchman living in Germany, to bring a case against the German state when it appears that Germany is treating me less favourably than the Netherlands would in similar situations; nor should it allow me to (this example will be fleshed out below).

To make this more concrete: the right to free movement does not – and should not – allow citizens to enjoy the same (fundamental) rights in their host state as in their home state. In this respect, my argument may run counter to the idea of linking fundamental rights to EU citizenship,⁷² including by extending EU fundamental rights protection to mobile citizens.⁷³ There are several reasons for this. One objection to such proposals is that the bounded nature of citizenship is at odds with the universalist aspirations of fundamental rights. For that reason alone, fundamental rights should not be part of the substantive content of EU citizenship.⁷⁴ In addition, such proposals threaten the EU’s division of powers, especially the limitation enshrined in Article 51(1) of the Charter of Fundamental Rights that EU fundamental rights apply ‘to the Member States only when they are implementing Union law’. Indeed, as Iglesias Sánchez observed, linking fundamental rights and EU citizenship ‘would lead to a virtually complete generalisation of EU fundamental rights review, dismantling the structure of the current Charter system and imposing an unbearable burden on the Court’.⁷⁵

But perhaps the most fundamental objection to such proposals is a democratic one. The principle of non-discrimination on grounds of nationality not only imposes limitations on how Member States can treat EU citizens who do not have their nationality; it also limits the extent to which the EU can intrude in the democratically legitimated political choices of the Member States. In this respect, the principle performs an important democratic function. As Menéndez has argued,

As long as free movement of persons was considered as an operationalisation of the principle of non-discrimination on the basis of nationality, the constitutional standards being applied were still national ones, an outcome in full accordance with the key legal role played by the collective of national constitutions as the deep constitution of the European Union, and consequently as the key source of democratic legitimacy of the synthetic constitutional order.⁷⁶

Of course, the concept of non-discrimination is an elastic one; it can be used to tackle not only overt discrimination, but also other, more covert obstacles to free movement.⁷⁷ Indeed, as we have seen in section 2, there can be good reasons to construe the right to free movement more broadly and to extend it to non-discriminatory obstacles, for example, to ensure that EU citizens can move around the EU under conditions of certainty and stability. But the right to free movement should not be interpreted more broadly than that: national laws and policies that in no way impede cross-border mobility should not be caught by that right. Instead, the EU should in such cases afford respect to domestic constitutional structures, local practices of democratic self-determination, and the legitimate diversity that flows from such structures and practices. Construing the right to free movement with respect for the Member States’ democratic constitutions is also in accordance with the EU’s own federal structure. Following on from this, as Schütze argued, is ‘a conception of the common market that accepts collective differences stemming from distinct political communities: the Member States’.⁷⁸

If this is correct, the judgements in the Walloon and family reunification cases are not. As will be explained now, these judgements show confusion about the aim of the right to free movement. Moreover, because the resulting reverse discrimination was a result of the CJEU’s misinterpretation of this right, the correct solution is not to equalise rights upwards and interfere to an even greater extent in the Member State’s regulatory competences than already occurred with these judgements. On the contrary, if this type of reverse discrimination is to be reversed, the logical solution is to equalise rights downwards and respect both the principle of divided powers and the principle of equality. This will be demonstrated in what follows by highlighting the confused application of the right to free movement in Walloon (A) and the family reunification case law (B), before showing how the cases demonstrate the shortcomings of the two dominant perspectives on reverse discrimination (C).

A. Walloon

In section 3, we saw that the Commission’s objections to the Flemish case insurance scheme were overcome by including EU citizens working in the Dutch-speaking region or bilingual region of Brussels but living in another Member State within the scope of that scheme. However, residents of Wallonia – Belgian and non-Belgian EU citizens alike – were still excluded from its coverage.⁷⁹ Their exclusion was challenged by the governments of Wallonia and the French Community in Belgium. The CJEU partially agreed with the claims put forward by both governments: the exclusion of non-Belgian EU citizens living in Wallonia was deemed problematic, but not the exclusion of Belgian citizens living there. With respect to the latter group, it confirmed the purely internal rule, saying that EU citizenship is ‘not intended to extend the material scope of the Treaty to internal situations which have no link with Community law’.⁸⁰ Belgian citizens resident within Wallonia with no connection to another Member State thus fell outside of the reach of EU free movement law and did not have to be included under the care insurance scheme. In other words, their exclusion was compatible with EU citizenship law.

In contrast, with respect to non-Belgian EU citizens living in Wallonia, the CJEU ruled that they fell within the scope of EU law because they had exercised the right to free movement.⁸¹ Furthermore, it reasoned that their exclusion from the care insurance scheme amounted to an obstacle to the exercise of that right because those

pursuing or contemplating the pursuit of employment or self-employment in one of those two regions, might be dissuaded from making use of their freedom of movement and from leaving their Member State of origin to stay in Belgium, by reason of the fact that moving to certain parts of Belgium would cause them to lose the opportunity of eligibility for the benefits which they might otherwise have claimed.⁸²

The Flemish government invoked the federal division of powers within Belgium to justify the restricted personal scope of its provisions, but this argument was resolutely rejected. The CJEU retorted ‘that a Member State cannot plead provisions, practices or situations prevailing in its domestic legal order, including those resulting from the constitutional organisation of that State, to justify the failure to observe obligations arising under [EU] law’.⁸³

The outcome of this decision was reverse discrimination. Belgian nationals who lived in Wallonia and did not exercise the right to free movement were excluded from the insurance scheme, while those who had exercised that right had to be included under the insurance scheme. However, it is important to recognise that this type of reverse discrimination is of a different type from the discrimination resulting from the interaction between the lex loci domicilii and the lex loci laboris principles (discussed in section 3). After all, the principle of lex loci laboris in EU social security coordination law helps protect the right to free movement: its application guarantees that mobile citizens are not deprived of social protection. In contrast, the argument in Walloon that the exclusion of EU citizens living in Wallonia from Flemish health care constituted an impediment to their right to free movement was fanciful. Certainly, residence in Wallonia might be less attractive if citizens do not enjoy the same social benefits as in Flanders (or, for that matter, in other Member States). However, an inherent aspect of federalism is that the constituent states can set their own rules and standards within the confines of federal law. By definition, then, the choice to move from one federal region to another has implications for the applicable rules and standards, and thus for the kind of rights and benefits one is entitled to.⁸⁴ That is, indeed, ‘an unavoidable by-product of the choice not to harmonize social security schemes’.⁸⁵

Viewing such federal differences as an impediment to free movement is legally incorrect and politically problematic. First, such an interpretation of the right to free movement makes it almost impossible for Member States to organise their welfare policies along federal lines. Such federal divisions may dissuade EU citizens from moving to a region that offers less generous social benefits, which, according to Walloon, constitutes a restriction of the right to free movement that cannot be justified, as Member States cannot cite their federal organisation as a justification for such restrictions. The problem in the CJEU’s reasoning, however, was not only its condescending attitude toward national constitutional provisions regarding the organisation of the state. The bigger problem was that Walloon rests on an incorrect understanding of the aim of right to move and reside freely. The aim of this right, as explained above, has never been to make every potential place of residence as attractive as possible, nor should it be conceived as such. Otherwise, federal borders become quasi-automatic barriers to free movement, and federal states will continually violate EU law. After all, many rights within federal states are, as the CJEU argued in support of its decision, ‘dependent on the condition of residence in. .. a limited part of national territory’.⁸⁶ In fact, not just federally organised Member States, but even the most centralised among them will be in violation of EU law on a constant basis if Walloon were to become the default interpretation of the right to free movement. Even highly centralised Member States are divided in administrative jurisdictions, granting local centres of government some margin of discretion in deciding what rights and obligations those residing within their territory have.⁸⁷ To give but one example, persons working and living in Amsterdam are subject to a different taxation and benefits scheme than those working in Amsterdam but living in Rotterdam. According to Walloon, this may dissuade free movement and thus be impermissible, because, to cite the CJEU again, ‘moving to certain parts of [the country] would cause [EU citizens] to lose the opportunity of eligibility for the benefits which they might otherwise have claimed’.⁸⁸

So the problem with Walloon is that it is confused about the aim of the right to free movement, which is to combat obstacles to that right, not to make mobile citizens’ region of residence as attractive as possible by allowing them to claim benefits they can enjoy in other regions. Indeed, as Christensen and Malmstedt have argued, they are not ‘exempted from the consequences of existing differences’ between and within Member States.⁸⁹ As we saw earlier, this is also what the CJEU has consistently decided: the right to free movement does not guarantee that cross-border mobility is neutral in its affects. Indeed, Walloon is the outlier, misapplying the right to free movement and challenging domestic regulatory standards that do not impede free movement. On such a conception of the right to free movement, no national measure is safe anymore; any measure can be challenged if more favourable rights and benefits can be obtained by moving to another (region of a) Member State. That cannot be and has never been the aim of free movement.

B. Family reunification

If we accept that the right to free movement should not be used to challenge national regulatory standards that do not burden free movement – as I believe most EU lawyers would – then an uncomfortable conclusion follows: the family reunification case law rests on an incorrect interpretation of the right to free movement as well. This line of cases is complex, and I will not discuss them all in detail.⁹⁰ Instead, I will demonstrate the shortcomings of these cases through a detailed examination of the leading case on which much family reunification case law is based: Metock.⁹¹

But first consider the following hypothetical situation: A Dutch–German couple living in the Netherlands considers moving to Germany to build a new life. While they have always liked the idea of living in Berlin, the real reason they consider moving is that they would like to start a family in the coming years. They have learned that Germany has a much more generous statutory parental leave scheme than the Netherlands. Moreover, they know that childcare in the Netherlands is prohibitively expensive, while in Berlin their child can attend a good kindergarten for free. But not only are they considering their options in Berlin; they also decide to challenge the Dutch policies in court, arguing that these violate their right as EU citizens to move and reside freely. Their argument is that the unfavourable Dutch policies discourage them from staying in the Netherlands and encourage them to move to Germany. How should the Dutch Court decide such a case? Should it side with the couple and decide that EU free movement law requires the Dutch authorities to offer them the same parental leave and childcare to which they are entitled in Germany? Or should it commend the couple for its creativity but explain that it is up to each Member State to decide such issues, and that such policies cannot be challenged on the mere ground that other states have more favourable policies? Of course, as the case law discussed earlier this section also shows, the latter is the correct answer under EU free movement law. Their legal advisor could tell them that they should consider moving to Germany if that would be desirable. After all, EU citizens have the right to ‘vote with her feet’;⁹² they can move elsewhere to live the life they desire. But the fact that they can live a more attractive life elsewhere does not entitle them to the same standard of living in their state of residence. So much seems clear.

Now imagine a situation that differs from the original situation in only one respect: the right claimed is not parental leave or subsidised childcare, but the right to family reunification. Consider, more specifically, the following situation: an EU citizen who has moved from the Netherlands to Germany meets a third-country national (TCN) after some time in Germany. In time they marry, yet the German government decides that the TCN partner no longer meets the conditions for legal residence and orders him to leave Germany. The EU citizen objects, arguing that the decision of the German authorities violates her right under EU law to move and reside freely. She learned that she could have lived together with her partner if they had been in the Netherlands, which happens to have more favourable family reunification rules. She claims that this discrepancy between the Dutch and German rules, and especially the strictness of the latter, discourages her from staying in Germany and encourages her to return to the Netherlands. Therefore, she argues, the German authorities violate her right to reside freely in Germany. The German authorities reject this argument. They explain that they would have allowed family reunification if she had lived with her partner in the Netherlands before moving to Germany. They would have allowed this in the interest of her right to move to and reside freely in Germany. But her situation is different: she met her partner in Germany only after she had successfully exercised her right to move to Germany, which is why the rejection of her application does not constitute a violation of the right to free movement. In fact, the German authorities continue, they apply the German family reunification rules – equally in law and in fact – to her as they do to German nationals, thus giving effect to the prohibition of nationality discrimination. The rules may, of course, discourage her from staying in Germany, but EU law does not require Member States to make it as attractive for EU citizens to reside in their country as in other countries. Again, this situation differs from the original situation only in terms of the rights claimed: family reunification rather than parental leave and subsidised childcare. Is this a relevant difference that matters to how the CJEU should decide such cases?

The Metock judgement shows that the CJEU deemed this a relevant difference. It is unlikely (at the very least) that it would accept an argument such as that in the hypothetical scenario above, that being discouraged from staying in one and encouraged to move to another Member State because of more favourable childcare or parental leave policies, constitutes a violation of the right to move and reside freely. But it did accept an essentially similar argument with respect to family reunification. These were the facts of Metock: four applicants, all TCNs living in Ireland, had met and married EU citizens from Member States other than Ireland. This had happened only after the EU citizens had moved to Ireland. The applicants had applied for asylum and later the right to reside, but these applications were unsuccessful. Their application for family reunification and the right of residence under EU free movement law was rejected by the Irish government on the ground that

Member States have competence in relation to the admission into a Member State of nationals of non-member countries coming from outside Community territory, while the Community has competence to regulate the movement of Union citizens and their family members within the Union.⁹³

Ireland argued, in other words, that since the TCNs had not resided legally in another Member State and lived there together with their partner before moving to Ireland, the issue was not one of free movement governed by EU law. Instead, it concerned the admission into a Member State of nationals of a third country, which was within its national regulatory competences.

The CJEU rejected the Irish government’s claims and argued that the right to move and reside freely granted to EU citizens under the Citizenship Directive would be obstructed ‘if Union citizens were not allowed to lead a normal family life in the host Member State’.⁹⁴ It reasoned that the refusal to grant family members the right to reside is ‘such as to discourage that citizen from moving to or residing in that Member State, even if his family members are not already lawfully resident in the territory of another Member State’.⁹⁵ In other words, even if mobile EU citizens meet their partners only after arriving in the host state, that state must grant partners the right of residence under EU free movement law. Otherwise, EU citizens may be discouraged from staying, which amounts to a restriction of their right to move and reside freely.

Because of this broad interpretation of the right to freedom of movement and residence, static citizens are reversely discriminated against compared to mobile citizens with respect to the right to family reunification. Mobile citizens can evade the confines of national immigration law, while those who have never had genuine residence in another Member State must comply with the national rules on family reunification.⁹⁶ It is important, however, to realise that this type of reverse discrimination is different from reverse discrimination caused by the application of the principle of mutual recognition (Type I reverse discrimination). As we saw in section 2, the principle of mutual recognition ensures that EU citizens can move freely under conditions of certainty and stability. This principle can also be used to protect the family life of EU citizens, by requiring the host Member State to recognise the family status of EU citizens and their family members who reside in another Member State before they exercise the right to free movement. A good example is the requirement under EU law that the host Member State recognises same-sex marriages contracted elsewhere in the EU.

The type of reverse discrimination created by Metock is of a different kind. The ruling makes mobile citizens better off than static citizens, even though their privileged treatment was not needed to protect their right to free movement. After all, the refusal to grant partners the right to reside does not impede EU citizen’s right to free movement if their partners were not living with them in the territory of another Member State.⁹⁷ Indeed, the EU citizens in Metock had successfully exercised their right to move from another Member State to Ireland and met their partners after they had moved there – the right to free movement was therefore not at issue.

But if the rule that a right of residence must be granted by the host state to TCN family members who never enjoyed prior residence in the territory of another state cannot be justified as a way to facilitate the pursuit of the right to free movement, then how did the CJEU justify its decision? The Court must also have understood that its judgement could not be justified by reference to the right to free movement, which is presumably why it anchored its conclusions exclusively in the right to freedom of residence. It said that:

Where a Union citizen founds a family after becoming established in the host Member State, the refusal of that Member State to authorise his family members who are nationals of non-member countries to join him there would be such as to discourage him from continuing to reside there and encourage him to leave in order to be able to lead a family life in another Member State or in a non-member country.⁹⁸

According to the CJEU, therefore, making sure that EU citizens can lead a normal family life in the host state is necessary to guarantee their continued residence, since otherwise EU citizens may be discouraged from staying there and be encouraged to move to another Member State. It is irrelevant in this respect whether the family members were already residing in the territory of another Member State prior to the decision of the EU citizen to exercise the right to move and reside freely.

But while this may sound more plausible than justifying Metock by reference to the right to free movement, it s not; the CJEU’s interpretation of the right to freedom of residence cannot justify the judgement either. This is because this interpretation repeats the mistake of Walloon: it exempts mobile citizens from the consequences of regulatory differences between Member States. Citizens who decide to move elsewhere may, of course, as a consequence of their mobility, become subject to less favourable health case schemes, parental leave policies, and, indeed, family reunification rules than in their Member State of origin. And they may, as a result, be discouraged from staying and encouraged to move to another Member State with better health care schemes, parental leave policies, and family reunification rules. None of this is surprising or problematic. It so often happens that EU citizens are encouraged to move to another Member State because of the opportunities there, be they work-related, family-related, or out of economic and financial interest.

To my knowledge, no one believes that such situations violate the right of free residence, and the reason is simple: the right to free movement offers EU citizens an exit option, which they can use to move to a Member State that offers them the desired quality of life.⁹⁹ It allows them to be encouraged to build a new life elsewhere if they are discouraged from staying in their current Member State for whatever reason. Thus, being discouraged from remaining in a particular Member State and being encouraged to leave that Member State because of better opportunities elsewhere cannot logically be a violation of the right to freedom of movement of residence, since this right specifically allows citizens to act upon such feelings of discouragement and encouragement. Perhaps the CJEU did not mean to say that such feelings always constitute a restriction on the right to reside (although the reasoning in Metock supports such an interpretation), but then the question arises: why should family reunification be treated differently?

The following reason is often given to justify the distinct treatment of family reunification: reverse discrimination in family reunification is intended to compensate mobile citizens for the disadvantages they face relative to static citizens. As van der Mei has argued, mobile citizens face

the obstacle of having to integrate in another country, society and culture, whilst the [static citizen] enjoys the comfort of living in and already being fully integrated in the societal setting he knows and understands. … Can, or should, the right of mobile Union citizens to live in another Member State with their third country family members rather not be seen as a ‘privilege’ granted to compensate for actual inequalities?¹⁰⁰

Along similar lines, Barnard has argued that reverse discrimination in family reunification may be a justifiable way to address the political underrepresentation of mobile EU citizens: ‘migrants cannot necessarily gain access to the host state’s political processes so Union law intervenes on their behalf to correct laws which discriminate against them’.¹⁰¹

It is not clear, however, that these are compelling justifications. First, it is not the case that national family reunification laws discriminate against mobile citizens. In fact, the argument put forward by the Irish government in Metock was that the same rules that apply to static citizens should also apply to mobile citizens. This seems consistent with, and indeed the logical implication of, the prohibition of discrimination on grounds of nationality. It was the CJEU that created discrimination between mobile and static citizens through its interpretation of the right to move and reside freely in Metock. Second, regarding the argument that reverse discrimination in family reunification serves to compensate for the disadvantages faced by mobile citizens, it should first be noted that many of those who avail themselves of the right to free movement are among the privileged in our societies; it is, in this respect, unclear what mobile students, workers, or pensioners should be compensated for. Of course, plenty of mobile citizens are less privileged, but here the obvious question arises: if we believe that such citizens need to be compensated, why is family reunification a suitable form of compensation? Metock benefits only a very small portion of mobile citizens, namely those who met their partner in the host country after they exercised their right of free movement.¹⁰² If we are serious about compensating mobile citizens, why not compensate all mobile citizens? Why not, for example, exempt them from taxation? One reason is that mobile citizens must contribute their share, but another important reason not to compensate mobile citizens stems from the values underlying the right to non-discrimination on the basis of nationality. The justification for treating mobile and static citizens alike is grounded not only in the value of equality but also, as I explained above, in the value of democratic legitimacy. Indeed, as long as the free movement of citizens is materialised through the principle of non-discrimination on the grounds of nationality, the regulatory standards applied are national standards, legitimated through the national democratic process.¹⁰³ This requires that mobile citizens are treated equally with static citizens, both in terms of the rights they enjoy and the duties they owe, and in respect of the whole body of national laws and regulations, including of course the rules on family reunification.

It is from this perspective that we must evaluate Type III reverse discrimination. Unlike Type I, Type III cannot be justified as an evil necessary to remove obstacles to the right to free movement and residence. Nor can it be justified as a compensation for the burdens imposed on mobile citizens. So how should we evaluate Type III reverse discrimination then?

C. Neither constitutional division of powers nor equal citizenship

It follows from the foregoing discussion that neither of the two main positions on reverse discrimination is entirely correct. First, as should be clear by now, Type III reverse discrimination is not inevitable under the EU’s constitutional principle of divided powers. On the contrary, this principle supports the arguments of the governments of Flanders and Ireland in Walloon and Metock, respectively. The Flemish government defended the personal scope of its care insurance scheme by reference to Belgium’s federal division of powers, and the Irish government specifically pointed to the EU’s division of competences to explain why it was for the Member States to decide on the admission of TCNs. Since there appears to be no justification for the outcomes in Walloon and Metock (as we saw, the right to move and reside freely cannot be invoked as justification), it would have been more appropriate, from the point of view of the principle of divided powers, to respect Belgium’s federal division of powers and Ireland’s admission policies. In other words, if domestic policy decisions and constitutional provisions do not impede the right to freedom of movement and residence, their existence should not be threatened by extravagant interpretations of this right.

There being no justification for Type III reverse discrimination, critics of reverse discrimination are correct to think that this type constitutes an unjustified violation of the principle of equality. But this does not answer the question of at what level of government equality is to be realised; it is in respect of this question, of the proper place of equal treatment, that critics have reached the wrong conclusion. They typically advocate to equalise rights upwards to the higher level of government. In her Opinion in Walloon, for example, AG Sharpston expressed her puzzlement that ‘although the last 50 years have been spent abolishing barriers to freedom of movement between Member States, decentralised authorities of Member States may nevertheless reintroduce barriers through the back door by establishing them within Member States’.¹⁰⁴ And she took the view that the case presented an opportunity to reconsider ‘the sustainability in its present form of the doctrine on purely internal situations’.¹⁰⁵ But these internal barriers were nothing less than Belgium’s federal constitutional boundaries, so that a revision of the purely internal rule would have challenged the essence of federalism within Belgium. It would have realised equality at the federal level, entitling everyone in Belgium to inclusion in a care insurance scheme established by one of the federal regions. And this while, as we saw in section 4.A, these internal constitutional boundaries did not truly impede the right to move and reside freely.

Similarly, in her Opinion in Ruiz Zambrano, AG Sharpston advocated the abolition of reverse discrimination when that would entail a fundamental rights violation, including most likely in cases of family reunification.¹⁰⁶ In other words, she advocated equal treatment of EU citizens with respect to family reunification at the EU level. But after what we learned in the previous section, namely that the reverse discrimination in the family reunification case law is the result of an incorrect interpretation of the right to freedom of movement of residence, this seems the wrong conclusion to reach. In Metock, the CJEU intruded into national competences for no good reason; and yet, critics of reverse discrimination argue that national rules in relation to the admission of TCNs should, by judicial fiat, be subsumed by common European standards. This is not a logical conclusion, nor the only way to respect the principle of equal treatment.

There is an alternative way of remedying Type III reverse discrimination that simultaneously respects the principle of divided powers and the principle of equality. These constitutional principles need not collide, at least as far as the resolution of this type of reverse discrimination is concerned. To see this, we must understand that the ideal of equal citizenship does not as a matter of principle or necessity require equal treatment at the higher level of government. Equality can just as well be realised at the lower level of government. In fact, it seems more logical to realise equality at the level of government that confers the benefits and rights in question, which in the cases discussed here would mean equalising rights downwards rather than upwards. Not just would this solution admit that the CJEU’s intrusion into national competences in Walloon and the family reunification case law was unjustified. More importantly, it is the way to guarantee equality between EU citizens while respecting the division of powers within the EU. The scope of EU free movement law should be such that only actual restrictions to the right to move and reside freely are captured thereby. For the rest, Member States must, under EU free movement law, be left free to determine their own regulatory standards and constitutional organisation. This would be more respectful from the point of view of the principle of divided powers, and at the same time ensure that the principle of equality is safeguarded at the correct level of government. This would allow the CJEU to treat national federal borders with more respect than it did in Walloon, while ensuring that EU citizens are treated equally with the nationals of the host Member State in the region where they reside, ie, treated equally with Belgian citizens living in Flanders if they are living there and with Belgian citizens in Wallonia if that is their place of residence. Likewise, equal citizenship does not as a matter of principle require equality at EU level; it can just as well imply equality at the national level if the applicable law has been enacted by the national government (in fact, this is what non-discrimination on grounds of nationality logically entails). In other words, decisions on the admission of third-country nationals can be left to the Member States and equality be realised at the national rather than EU level, so that mobile citizens are subject to the same family reunification standards as static citizens. This conclusion is hard to swallow, but it is consistent with the right to free movement and residence;¹⁰⁷ it is also consistent with both the division of powers and the principle of non-discrimination on the grounds of nationality.

To sum up, there is no justification for Type III reverse discrimination and there are good reasons to argue for its reversal to guarantee equality between EU citizens. However, this reversal need not and should not come at the expense of the principle of divided powers. Contrary to established wisdom, there is no inevitable collision between the principle of equality and the principle of divided powers. The current tension between both principles exists solely due to a confusion on the CJEU’s part over the aim of the right to free movement and residence. As we have seen, it is possible to interpret this right in a way that respects both principles, namely by restricting the application of this right to actual obstacles to the exercise of this right and by otherwise allowing Member States and their regions to apply their legal norms equally to mobile and static citizens. EU citizenship was never meant to be a universal and indivisible status guaranteeing equal rights to all irrespective of the Member State in which they reside.

5. Conclusion

Most scholars have assumed that reverse discrimination is a singular phenomenon that demands a singular response, though they have disagreed on what that response must be. One group has argued that reverse discrimination is incompatible with the ideal of equal citizenship and a second group that reverse discrimination is the logical consequence of the purely internal rule, which exists to protect the division of powers between the EU and the Member States. In this Article, I have shown that these positions overlook the complexity of reverse discrimination and that, taking this complexity into account, it cannot be said that reverse discrimination is a constitutional necessity in a Union that divides political powers with its Member States, nor that it always constitutes an unjustified violation of the principle of equality. Moreover, to the extent that it does, the logical remedy is not equalise rights upwards but downwards.

More specifically, I have shown that there are three different types of reverse discrimination and that each requires a different response. Type I is caused by the application of the principle of mutual recognition. If the EU wants to ensure that free movement takes place under conditions of certainty and stability, this type of reverse discrimination is inevitable. To avoid it, the EU would have to sacrifice either the principle of mutual recognition or the principle of divided powers. Neither option seems attractive. Type II reverse discrimination is caused by an interaction between domestic federal borders and internal discrimination. Unlike Type I, Type II reverse discrimination is not a necessity in a Union committed to free movement of citizens. But like Type I, it is not an inevitable by-product of the principle of divided powers, nor an unjustified violation of the principle of equality that demands intervention by the CJEU. This type could be abolished without violating the division of powers between the EU and its Member States, but only through (national) political action. Finally, Type III reverse discrimination is caused by confusion over the aim of the right to free movement and residence. This type is problematic and demands a judicial response, but a response that is different than critics of reverse discrimination have typically argued for. That is, the appropriate response is not for the CJEU to abolish the purely internal rule and equalise the position of static citizens with that enjoyed currently by mobile citizens. Rather, its duty is to protect both the principle of equality and of divided powers simultaneously, which it can do by equalising rights downwards to the level of government that is competent to act.