OPINION OF ADVOCATE GENERAL
SHARPSTON
delivered on 16 February 2012 (1)
Case C‑542/09
European Commission
v
Kingdom of the Netherlands
(Access to education — Funding for higher education abroad — Residence requirement — ‘Three out of six years rule’)
1. Erasmus of Rotterdam was an early beneficiary of funding to study abroad. The then bishop of Cambray, Henry of Bergen (for whom Erasmus had started to work as secretary), gave him both leave and a stipend in 1495 to go and study at the University of Paris. Erasmus never looked back; and, in a career that spanned Paris, Leuven, Cambridge and Basel, he became arguably the outstanding scholar of his generation: the ‘Prince of the Humanists’. It is tolerably safe to say that he put the funding for his university studies abroad to excellent use (2) — and, indeed, the current exchange programmes between EU universities bear his name.
2. Modern day compatriots of Erasmus enjoy similar good fortune. Under the provisions of the Wet Studiefinanciering (Law on the Financing of Studies — ‘the WSF’), they can often obtain funding for higher education pursued outside the Netherlands. However, do the detailed rules governing the grant of such funding — in particular, the rule under which an applicant must, in addition to being eligible for funding to study in the Netherlands, also have resided lawfully in the Netherlands during at least three out of the last six years (the ‘three out of six years rule’) — fall foul of Article 45 TFEU (formerly Article 39 EC) (3) and Article 7(2) of Regulation (EEC) No 1612/68 (4) inasmuch as they discriminate indirectly and without justification against migrant workers and their dependent family members?
Legal background
Treaty provisions
3. Article 45 TFEU states:
‘1. Freedom of movement for workers shall be secured within the Union.
2. Such freedom of movement shall entail the abolition of any discrimination based on nationality between workers of the Member States as regards employment, remuneration and other conditions of work and employment.
…’
4. Pursuant to Article 165(1) TFEU (formerly Article 149(1) EC), Member States are responsible ‘for the content of teaching and the organisation of education systems’. Article 165(1) states that ‘[t]he Union shall contribute to the development of quality education by encouraging cooperation between Member States and, if necessary, by supporting and supplementing their action’. Union action is also to be aimed at ‘encouraging mobility of students’. (5)
Regulation No 1612/68
5. Regulation No 1612/68 aimed to secure the freedom of nationals of one Member State to work in another Member State and thereby implement the Treaty provisions on freedom of movement for workers. The first recital in the preamble to that regulation described its overall objective as being to achieve ‘the abolition of any discrimination based on nationality between workers of the Member States as regards employment, remuneration and other conditions of work and employment, as well as the right of such workers to move freely within the [Union] in order to pursue activities as employed persons subject to any limitations justified on grounds of public policy, public security or public health’.
6. The third and fourth recitals, respectively, stated that ‘freedom of movement constitutes a fundamental right of workers and their families’ and that that right was to be enjoyed ‘by permanent, seasonal and frontier workers and by those who pursue their activities for the purpose of providing services’.
7. According to the fifth recital, the exercise of this fundamental freedom, ‘by objective standards, in freedom and dignity, require[d] that equality of treatment shall be ensured in fact and in law in respect of all matters relating to the actual pursuit of activities as employed persons and to eligibility for housing, and also that obstacles to the mobility of workers shall be eliminated, in particular as regards the worker’s right to be joined by his family and the conditions for the integration of that family into the host country’.
8. Article 7(2) of Regulation No 1612/68 provided that a worker, who is a national of a Member State, in the territory of another Member State ‘shall enjoy the same social and tax advantages as national workers’.
9. Article 12 of Regulation No 1612/68 read:
‘The children of a national of a Member State who is or has been employed in the territory of another Member State shall be admitted to that State’s general educational, apprenticeship and vocational training courses under the same conditions as the nationals of that State, if such children are residing in its territory.
...’
Directive 2004/38
10. Article 7 of Directive 2004/38/EC (6) governs the conditions under which EU citizens can reside more than three months in another Member State. It states:
‘1. All Union citizens shall have the right of residence on the territory of another Member State for a period of longer than three months if they:
(a) are workers or self-employed persons in the host Member State; or
(b) have sufficient resources for themselves and their family members not to become a burden on the social assistance system of the host Member State during their period of residence and have comprehensive sickness insurance cover in the host Member State; or
...’
11. Article 24 of that directive provides:
‘1. Subject to such specific provisions as are expressly provided for in the Treaty and secondary law, all Union citizens residing on the basis of this Directive in the territory of the host Member State shall enjoy equal treatment with the nationals of that Member State within the scope of the Treaty. The benefit of this right shall be extended to family members who are not nationals of a Member State and who have the right of residence or permanent residence.
2. By way of derogation from paragraph 1, the host Member State shall not be obliged … prior to acquisition of the right of permanent residence, to grant maintenance aid for studies, including vocational training, consisting in student grants or student loans to persons other than workers, self-employed persons, persons who retain such status and members of their families.’
National law
12. The WSF defines who can receive funding to study in the Netherlands and abroad. Funding to study abroad is called ‘meeneembare studie financiering’ (‘MNSF’), that is to say, ‘portable’ funding for studies.
13. For higher education in the Netherlands, funding for studies is available to students who are between 18 and 29 years old, study at a designated or approved educational establishment and satisfy a nationality condition. (7) Article 2(2) defines the nationality condition. Those eligible are: (i) Netherlands nationals, (ii) non-Netherlands nationals who are treated, in the area of funding for studies, as Netherlands nationals based on a treaty or a decision of an international organisation and (iii) non-Netherlands nationals who live in the Netherlands and belong to a category of persons who are treated, in the area of funding for studies, as Netherlands nationals on the basis of a general administrative measure.
14. The second category includes EU citizens who are economically active in the Netherlands and their family members. They need not have resided in the Netherlands to qualify for this type of funding. Thus, cross-border workers and their family members are covered. The third category includes EU citizens who are not economically active in the Netherlands. They qualify for funding after five years of lawful residence in the Netherlands.
15. For funding for higher education pursued outside the Netherlands, students must be eligible for funding for higher education in the Netherlands and, pursuant to Article 2(14)(2)(c) of the WSF, must additionally have resided lawfully in the Netherlands during at least three out of the six years preceding enrolment at an educational establishment abroad. This requirement applies irrespective of students’ nationality.
16. As long as they satisfy the relevant conditions, students can apply sequentially for funding to study in the Netherlands and then for MNSF to study abroad.
17. Until 1 January 2014, the three out of six years rule does not apply to students, whatever their nationality, pursuing higher education in the ‘border areas’ of the Netherlands (Flanders and the Brussels-Capital Region in Belgium, and North-Rhine Westphalia, Lower Saxony and Bremen in Germany).
18. MNSF consists of four components: (i) a basic grant, which is a fixed amount paid per month and based on whether the student lives at home or independently, together with an allowance for travel costs and an additional allowance if the student has a partner or is a single parent, (ii) an additional grant, based on the income and contribution of the student’s parents and subject to a maximum limit, (iii) a basic loan, if applied for, subject to a maximum limit and (iv) a loan to cover fees, if applied for, limited in principle to the maximum fee chargeable by Netherlands educational institutions for an equivalent course.
19. The basic grant, the additional grant (except for the first year of studies) and the allowance for travel costs are given as loans. They become grants if the studies are completed within 10 years of their commencement.
20. The maximum limit for MNSF funding, excluding allowances, ranges from EUR 739.15 to EUR 929.69 per month, depending on whether the student lives at home or independently. The same limit applies to funding for studies in the Netherlands.
Procedure
21. Following a regular pre-litigation procedure, the Commission asks the Court to declare that, by requiring that migrant workers, including cross-border workers, and their dependent family members fulfil a residence requirement (that is, the three out of six years rule) to be eligible under the WSF for the funding of educational studies abroad, the Kingdom of the Netherlands indirectly discriminates against migrant workers and has failed to fulfil its obligations under Article 45 TFEU and Article 7(2) of Regulation No 1612/68, and to order the Kingdom of the Netherlands to pay the costs.
22. The Netherlands Government contends that the Court should dismiss the application and order the Commission to pay the costs.
23. The Governments of Belgium, Denmark, Germany and Sweden have intervened in support of the Netherlands.
24. The principal parties and all the interveners made oral submissions at the hearing on 10 November 2011.
Assessment
Preliminary remarks
25. The Commission has throughout limited its claim to Article 45 TFEU and Article 7(2) of Regulation No 1612/68. It argues that there is indirect discrimination against migrant workers working in the Netherlands and their dependent family members with respect to MNSF. It makes no complaint under Article 24 of Directive 2004/38, Article 21 TFEU or any other provisions of EU law governing citizenship rights.
26. Article 7(2) of Regulation No 1612/68 expresses the principle of equal treatment set out in Article 45 TFEU with regard to social and tax advantages and must be interpreted in the same manner. (8) Thus, if a measure regulating access to a social advantage infringes Article 7(2) because it treats migrant workers less favourably than national workers, it is also incompatible with Article 45 TFEU. However, even if a measure is compatible with Article 7(2), it may still infringe Article 45. (9) I shall therefore first consider the residence requirement in the light of Article 7(2) of Regulation No 1612/68. If it infringes Article 7(2), it is equally prohibited by Article 45 TFEU.
27. The Netherlands, supported by the intervening Member States, submits that Article 7(2) of Regulation No 1612/68 does not apply. In the alternative, the Netherlands argues that the residence requirement is not indirectly discriminatory against migrant workers.
28. In any event, the Netherlands and the intervening Member States contend that the residence requirement is justified for two reasons. First, the requirement serves to identify the desired target group of students: namely, those who, without MNSF, would study in the Netherlands and, if they study abroad, will return to the Netherlands. Second, the residence requirement prevents the scheme from becoming an unreasonable financial burden which could have consequences for the overall level of funding that is granted. That objective was endorsed by the Court in Bidar and confirmed in Förster. (10)
Does the residence requirement infringe Article 7(2) of Regulation No 1612/68 in principle?
The beneficiaries of equal treatment under Article 7(2) of Regulation No 1612/68
29. The Netherlands contends that Article 7(2) of Regulation No 1612/68 does not apply in principle to dependent family members of migrant workers, irrespective of their place of residence. It accepts that an exception exists in cases of direct discrimination against children of migrant workers. Generally, however, such persons are covered by Article 12 of Regulation No 1612/68, not by Article 7(2). This is because Article 12 is a specific expression of the equal treatment obligation as it applies to children and to access to general educational, apprenticeship and vocational training courses. Reading Article 7(2) as applying to children of migrant workers risks rendering the residence requirement in Article 12 meaningless.
30. The Commission contends that the Court’s case‑law confirms that Article 7(2) applies to all dependent family members of the migrant workers.
31. I agree with the Commission.
32. The direct beneficiaries of the equal treatment guaranteed by Article 7(2) are nationals of a Member State who work in another Member State. Cross-border workers, who reside by definition outside the host Member State, belong to this category. (11) Thus, workers are not required to reside where they work to enjoy protection under Article 7(2), nor does Article 7(2) make entitlement to equal treatment conditional on where the social advantage is actually enjoyed.
33. Dependent family members of a migrant worker are the indirect beneficiaries of the equal treatment obligation under Article 7(2) because discrimination against them with respect to a social advantage also discriminates against the migrant worker who then has to support the family member. The Court has already made clear that this group of indirect beneficiaries includes the workers’ dependent family members in the descending and ascending line and spouses. (12) They are not required to reside in the Member State where the migrant worker is employed to enjoy protection under Article 7(2). (13)
34. The term ‘social advantages’ in Article 7(2) includes funding for higher education studies pursued by migrant workers or their dependent family members. (14) In the present case, the dependent children of migrant workers working in the Netherlands may, in particular, wish to apply for MNSF to study elsewhere than in the Netherlands.
35. The Netherlands relies heavily on the fact that the cases in which the Court has held that Article 7(2) applies to children of migrant workers have all involved direct discrimination. Unlike the Netherlands, I see no logic in an interpretation which renders the personal scope of an equal treatment obligation dependent on the type of discrimination involved. I therefore consider that it is of no consequence whether the alleged discrimination is direct or indirect.
36. Article 12 of Regulation No 1612/68 gives a separate, distinct entitlement to children of migrant workers in their own right.
37. Pursuant to that provision, the host Member State must allow children of migrant workers access to its general educational, apprenticeship and vocational training courses. Article 12 applies also to children who pursue education outside the host Member State. (15)
38. Article 12 specifically applies to ‘[t]he children of a national of a Member State who is or has been employed in the territory of another Member State’ and who ‘are residing in its territory’. The Court has held that Article 12 grants children who have established their residence in a Member State during their parent’s exercise of the right of residence as a migrant worker in that Member State an independent right of residence in order to attend general educational courses there. (16) The child enjoys that right whether or not the parent retains the status of migrant worker in the host Member State. (17)
39. Furthermore, a child does not have to demonstrate dependence on the migrant worker to rely on Article 12. If the parent is no longer a migrant worker benefiting from equal treatment under Article 7(2) or providing for the maintenance of the child, the child may none the less claim in his own right access to the types of social advantage defined in Article 12 and under the same conditions as nationals, provided that the child resides in the host Member State. (18)
40. Unlike the Netherlands, I do not consider that because Article 12 expressly governs a defined, limited group of family members as direct beneficiaries, it necessarily follows that the personal scope of Article 7(2) should be read as excluding that group as indirect beneficiaries. The Netherlands relies on a series of cases in support of its position. None of these cases resolves the issue as to whether Article 7(2) protects dependent family members of a migrant worker seeking financial support for higher education.
41. In Brown, the claimant was denied protection under Article 7(2) because he acquired the status of migrant worker exclusively as a result of being accepted to undertake studies in the host Member State. (19) He could not seek protection under Article 12 (nor, on my reasoning, as an indirect beneficiary under Article 7(2)) because neither parent had the status of migrant worker after his birth. (20)Lair and Matteucci, on the other hand, concerned the application of Article 7(2) to claimants who were themselves migrant workers. (21)
42. In Casagrande, the Court interpreted Article 12 in a dispute involving the child of a migrant worker residing where the parent was employed, and held that that provision also covered general measures intended to facilitate educational attendance. (22) Similarly, di Leo (23) concerned the application of Article 12 to the child of a migrant worker leaving the host Member State to study abroad.
43. I conclude that dependent family members, including children, benefit from the migrant worker’s right to equal treatment under Article 7(2) of Regulation No 1612/68. That conclusion applies irrespective of where they or the migrant worker reside and whether the alleged discrimination is direct or indirect.
Does an objective difference exist between workers residing in the Netherlands and those residing outside the Netherlands?
44. The Commission claims that migrant workers (including cross-border workers) working in the Netherlands and their dependent family members are treated less favourably than Netherlands workers and their dependent family members.
45. The Netherlands argues that an objective difference exists between workers residing in the Netherlands and those residing outside the Netherlands because the latter do not require incentives to study abroad. That argument implies that migrant workers working in the Netherlands and residing in another Member State are not in a comparable situation to Netherlands workers (and migrant workers, for that matter) working and residing in the Netherlands.
46. I disagree with the Netherlands.
47. Discrimination under Article 7(2) exists when migrant workers are treated less favourably than national workers in a comparable situation. To decide whether that is the case, it is necessary to determine who benefits from equal treatment and in connection with what particular benefit. In that regard, the object of the rules establishing the difference in treatment is relevant to assessing whether an objective difference exists between the relevant categories of people. (24) I add that, in my view, the alleged objective difference must generally reflect a distinction made in law or in fact other than that made by the very legal rule that is at issue.
48. In the present case, the benefit is the grant of funding for studies anywhere outside the Netherlands. In the context of Article 7(2), migrant workers in the Netherlands benefit from equal treatment.
49. There is relatively little difficulty about accepting that the following two categories contain workers who may properly be compared with each other. First, migrant workers residing and working in the Netherlands are clearly comparable to, and are to be treated equally with, Netherlands nationals residing and working in the Netherlands. Second, migrant workers working in the Netherlands but residing elsewhere are clearly comparable to, and are to be treated equally with, Netherlands nationals working in the Netherlands but residing elsewhere.
50. The Netherlands uses the fact that these two identifiable categories exist to argue that no comparison can be made between those categories — that is, it claims that those residing in the Netherlands are objectively different from those residing outside the Netherlands. At one level, this is self-evidently true. Living in Amsterdam is not the same as living in Paris. But is this a relevant difference such as, objectively, to justify different treatment? (25)
51. I do not think so.
52. The Netherlands accepts (rightly) that children of migrant workers who wish to study in the Netherlands should have access to funding for such studies on exactly the same terms as Netherlands nationals, irrespective of whether those migrant workers (and their dependent children) reside in the Netherlands or elsewhere.
53. In so doing, it has implicitly accepted that at least some children of migrant workers may — like the children of Netherlands workers — be pre-disposed to study in the Netherlands (whether or not they are residing there) and that they should have access to funding to do so. But the necessary corollary of that — it seems to me — is that the Netherlands can no longer legitimately assert that the place of residence will, in a quasi-automatic manner, determine where the migrant worker or his dependent children will study. And, if that is right, then it is not legitimate to use place of residence as an allegedly ‘objective’ criterion for different treatment. On the contrary: a migrant worker employed in the Netherlands but residing in another Member State can properly be compared with a Netherlands worker residing and working in the Netherlands.
Does the residence requirement result in indirect discrimination?
54. It is settled case‑law that, in infringement proceedings, the Commission must prove the existence of the alleged infringement and provide the Court with the evidence necessary for it to establish that an obligation has not been fulfilled. In so doing, the Commission may not rely on a presumption. (26)
55. In this case, the Commission must demonstrate that migrant workers and Netherlands workers are treated differently with results similar to those that would follow from applying a condition of nationality.
56. The Commission argues that the residence requirement infringes Article 7(2) of Regulation No 1612/68 because national workers are always likely to satisfy it more easily than migrant workers. It submits that Meeusen (27) and Meints (28) establish that a residence requirement is, by definition, indirectly discriminatory. In the present case, the residence requirement is indirectly discriminatory in any event to the extent that it necessarily excludes cross-border workers and their dependent family members. The Netherlands relies on Sotgiu and Kaba II to argue that a residence requirement is not discriminatory in all circumstances. (29)
57. I share neither reading of the Court’s case‑law.
58. In Meeusen, the Court found that ‘a Member State may not make the grant of a social advantage within the meaning of Article 7 … dependent on the condition that the beneficiaries be resident within its territory’. (30)Meeusen concerned a residence requirement that was directly discriminatory and therefore prohibited. The Court’s statement in Meeusen was based in turn on Meints. (31) In that case, the Court concluded that the residence requirement at issue was indirectly discriminatory only after examining whether that requirement was more easily met by national workers (and whether it could be justified). (32) Neither judgment therefore establishes that a residence requirement is always indirectly discriminatory.
59. However, nor are the Court’s rulings in Sotgiu and Kaba II authority for the contrary position, namely that it may be possible to impose a residence requirement on nationals and non-nationals who are in a comparable situation without that resulting in indirect discrimination. In Sotgiu, the workers concerned belonged to different categories based on whether they were obliged to move. The Court therefore considered that residence formed an objective criterion for different treatment of workers in objectively different situations. In Kaba II, the spouse of a migrant worker who was a national of a Member State other than the United Kingdom and the spouse of a person who was ‘present and settled’ in the United Kingdom were held not to be comparable due to a distinction made in a provision of national law other than that at issue. (33)
60. I agree none the less with the Commission that the residence requirement indirectly discriminates against migrant workers.
61. A requirement of past, present or future residence (especially if it stipulates residence for a particular duration) is intrinsically likely to affect national workers of a Member State less than migrant workers who are in a comparable situation. That is because such a condition always distinguishes between workers who do not need to move to satisfy it and workers who do need to move. The former are usually, although possibly not invariably, more likely to be nationals of the host Member State.
62. The three out of six years rule pertains to past residence of a certain duration. I consider that Netherlands workers are more likely to be able to satisfy that condition than migrant workers residing in the Netherlands.
63. It is conceivable that such a residence requirement may not discriminate against every cross-border worker. (34) Nevertheless, it is likely that a considerable number of cross-border workers and their dependent family members are excluded from MNSF because the family resides together in a border area and thus outside the Netherlands.
64. I therefore conclude that the residence requirement constitutes indirect discrimination prohibited in principle by Article 7(2) of Regulation No 1612/68.
Is the residence requirement none the less justified?
65. If the residence requirement constitutes indirect discrimination prohibited by Article 7(2) of Regulation No 1612/68, the Court must determine whether it is none the less justified. To that effect, the Netherlands must demonstrate that the residence requirement (i) pursues a legitimate aim which is justified by overriding reasons of public interest, (ii) is appropriate to achieve the legitimate objective pursued (appropriateness) and (iii) does not go beyond what is necessary to achieve the desired objective (proportionality). (35)
66. The Netherlands argues that the residence requirement is justified because it is appropriate and does not go beyond what it necessary (i) to avert an unreasonable financial burden resulting from making MNSF available to all students (the economic objective) and, at the same time, (ii) to ensure that MNSF is available solely to students who, without it, would pursue higher education in the Netherlands, and who are likely to return there if they study abroad (the social objective).
67. Before turning to the justification of the residence requirement on the basis of each objective, I should like to comment on the principles governing the burden of proof and the standard of proof. I do so because neither party in this case has applied those principles properly.
68. The Court has held that the defendant Member State must provide ‘reasons which may be invoked by a Member State by way of justification’ and ‘an analysis of the appropriateness and proportionality of the restrictive measure adopted by that State and specific evidence substantiating its arguments’. (36) It thus bears the onus of establishing a prima facie case that the measure is appropriate and does not go beyond what is necessary to achieve its objective(s).
69. However, the burden on the defendant Member State to demonstrate proportionality ‘cannot be so extensive as to require the Member State to prove, positively, that no other conceivable measure could enable that objective to be attained under the same conditions’. (37) Put another way, the Member State cannot be required to prove a negative.
70. If the defendant Member State establishes that the contested measure is prima facie proportionate, it is then for the Commission to rebut the Member State’s analysis by suggesting other less restrictive measures. The Commission cannot merely propose an alternative measure. It must also explain why and how that measure is appropriate to achieve the stated objective(s) and is, above all, less restrictive than the contested measure. Without such an explanation, the defendant Member State cannot know on what its rebuttal should focus.
Is the residence requirement justified on the basis of the economic objective?
– Is the economic objective a legitimate aim which is justified by overriding reasons in the public interest?
71. The Netherlands argues that the residence requirement is justified because it seeks to ensure that MNSF does not impose an excessive financial burden on society. In Bidar and Förster, the Court accepted that Member States may be legitimately concerned with the financial consequences of policies and therefore require a degree of integration before making funding for studies available. (38) The Netherlands estimates that eliminating the residence requirement would result in an additional financial burden of some EUR 175 million per year spent on providing MNSF for, in particular, children of migrant workers and Netherlands nationals who either live outside the Netherlands or have lived less than three out of the previous six years in the Netherlands.
72. The Commission argues that the reasoning in Bidar and Förster does not apply to migrant workers because EU law treats economically active EU citizens differently from economically inactive EU citizens. Article 24(2) of Directive 2004/38 confirms that distinction. Even if the Netherlands were allowed to require a degree of connection, the status of migrant worker itself demonstrates a sufficiently close connection with the Netherlands; and the Court in Bidar recognised that no residence requirement can be imposed in such circumstances. (39) Furthermore, mere concerns about budgetary implications cannot qualify as overriding reasons relating to the general interest.
73. I agree with the Commission.
74. The Court is being invited to apply the reasoning in Bidar and Förster as regards economically inactive EU citizens to migrant workers. But first: what precisely did the Court rule in Bidar and Förster?
75. In Bidar, the United Kingdom sought to justify a residence requirement of three years based on the need to ensure that (i) contributions made through taxation were sufficient to justify the grant of the funding and (ii) a genuine link existed between the student claiming the funding and the employment market of the host Member State. (40) In essence, the concern was that students from all over the European Union might arrive in the United Kingdom and forthwith apply for funding to study there.
76. In response to the first part of the United Kingdom’s argument, the Court accepted that ‘it is permissible for a Member State to ensure that the grant of assistance to cover the maintenance costs of students from other Member States does not become an unreasonable burden which could have consequences for the overall level of assistance which may be granted by that State’. (41) As a result, it was legitimate to grant funding ‘only to students who have demonstrated a certain degree of integration into the society of that State’. (42)
77. The Court did not accept the second part of the United Kingdom’s argument. A Member State was not entitled to make the grant of funding for studies dependent on a link between the student and the employment market. In essence, the Court found that an indirectly discriminatory residence requirement could not be justified based on the need to grant funding only to students who had already worked in the host Member State or would work there after their studies. Indeed, the Court found that education does not necessarily assign a student to a particular geographical employment market. (43) Unlike the Commission, I do not read this part of the judgment in Bidar as precluding any requirement that migrant workers demonstrate a degree of connection to the host Member State. The Court simply did not address that point. What it did was to reject the argument that linking the place of studies and the place of employment was an objective that could justify indirect discrimination.
78. The Court went on to accept that past residence for a certain time in the host Member State may establish the necessary degree of connection. (44) Limiting the group of recipients through a criterion expressing a degree of closeness to the financing Member State, such as past residence, was thus an appropriate measure to ensure that the grant of funding to students from other Member States did not become an unreasonable burden which could have consequences for the overall level of assistance which might be granted by that State.
79. The Netherlands appears to read the judgment in Förster as confirming that in Bidar.
80. I am not convinced by that reading of Förster.
81. In Förster, the Court first noted that, according to Bidar, it is legitimate for a Member State to ensure that a social advantage does not become an unreasonable burden which could have consequences for the overall level of assistance. (45) That was, indeed, the legitimate objective recognised in Bidar. (46)
82. Next, the Court stated that, also according to Bidar, it is legitimate to grant assistance to cover maintenance costs only to students demonstrating a certain degree of integration into the society of the Member State. (47) The Court referred to the passage in Bidar where it held that a student may be regarded as demonstrating a certain degree of integration in the host Member State if the student has resided there for a certain period of time. (48)
83. The Court next applied that reasoning to the facts in Förster. The Court needed to resolve whether the indirectly discriminatory residence requirement of five years could ‘be justified by the objective, for the host Member State, of ensuring that students who are nationals of other Member States have to a certain degree integrated into its society’. (49) The Court in Förster therefore examined the proportionality of the residence requirement in relation to the objective of ensuring integration of the student, and not that of avoiding the collapse of the existing scheme due to its financial cost. (50)
84. However, the Court in Bidar had not recognised that objective. In that judgment, evidence of a degree of integration was treated as a means to avert an unreasonable financial burden.
85. It would be unfortunate if a superficial reading of Förster were to lead to confusion between means and end. There is a risk that Förster might be read as indicating that Member States can set a residence requirement, irrespective of whether its purpose is to ensure that making available a social advantage does not adversely affect the stability of its public finances or the pursuit of any other legitimate objective justified by overriding reasons of public interest. On that basis, Member States might seek to justify less favourable treatment of (both economically active and inactive) EU citizens in terms of social policy (integration) by applying access criteria such as length of residence, marital and family status, language, diplomas, employment, and so forth, without ever explaining why the availability of a social benefit should be limited in that way.
86. Against the background of that reading of Bidar and Förster, I turn to examine whether averting an unreasonable financial burden which could have consequences for the overall level of funding for studies is an objective that can be transposed from the context of economically inactive EU citizens and invoked to justify indirect discrimination against migrant workers.
87. I consider it cannot.
88. I accept that the financial burden of making a social advantage widely available may compromise its existence and overall level. (51) In such circumstances, concerns about budgetary implications are intrinsically linked with the existence and objective of the social advantage itself and cannot therefore be wholly disregarded. Member States might otherwise forgo altogether providing particular forms of social advantage, to the detriment of the public interest.
89. I am nevertheless of the view that the Netherlands cannot invoke budgetary concerns to justify discriminatory treatment of migrant workers and their dependent family members. Any conditions attached to MNSF in order to keep expenditure within acceptable limits must be borne equally by migrant workers and Netherlands workers.
90. Migrant workers and their families enjoy the freedom to move to another Member State based on the consideration that ‘mobility of labour within the [Union] must be one of the means by which the worker is guaranteed the possibility of improving his living and working conditions and promoting his social advancement, while helping to satisfy the requirements of the economies of the Member States’. (52) Member States must therefore eliminate any obstacles to the exercise of the freedom of movement and related rights of migrant workers, including those affecting ‘the worker’s right to be joined by his family and the conditions for the integration of that family into the host country’. (53)
91. In my opinion, if Member States make a social advantage available to their own workers, irrespective of whether the benefit is tied to a person’s contributions or not, they must grant it on equal terms to migrant workers. Any limitation imposed for preserving financial integrity must be applied on equal terms to national workers and migrant workers. (54)
92. It is true that the Court has accepted that the objective of averting an unreasonable financial burden which can have consequences for the overall level of social assistance granted can justify discrimination against economically inactive EU citizens. In my opinion, the Court has done so because, as EU law stands, all EU citizens are not yet guaranteed full equal treatment with regard to social advantages.
93. Before the introduction of EU citizenship, several directives provided that nationals of Member States who were not exercising an economic right to free movement had the right to move to and reside in another Member State on condition that they and their family members were covered by sickness insurance and had ‘sufficient resources to avoid becoming a burden on the social assistance system of the host Member State during their period of residence’. (55) The condition was imposed because these nationals ‘must not become an unreasonable burden on the public finances of the host Member State’. (56) In particular, Directive 93/96 limited the right of students to reside in another Member State and did not establish any right to payment of maintenance grants by the host Member State. (57)
94. These nationals whatever their activity became EU citizens (58) following the entry into force of the Maastricht Treaty. Based on that status, they have the right to move and reside freely within the territory of the Member States, subject to limitations laid down in EU law. The Court has held that the host Member State must show a certain degree of financial solidarity to students who are nationals of other Member States and have exercised their right to move to and reside in the host Member State. (59)
95. Directive 2004/38 consolidated much of the earlier legislation and case‑law. It maintains the distinction between EU citizens who have exercised an economic right of free movement and other EU citizens and expressly preserves the right of Member States to discriminate for a certain time against the latter. Thus, Article 24(2) of Directive 2004/38 provides that, until students have acquired permanent residence in the Member State where they study, ‘[b]y way of derogation from’ the obligation to treat equally nationals and other EU citizens, the host Member State cannot be obliged to grant them maintenance assistance for studies, consisting out of grants or loans. Although the facts giving rise to Bidar preceded the adoption of Directive 2004/38, the reasoning in that case reflects Member States’ freedom to discriminate in those circumstances. The derogation does not, however, apply to ‘workers, self-employed persons, persons who retain such status and members of their families’. Such persons are, on the contrary, protected by the general rule of equal treatment.
96. I therefore conclude that the economic objective cannot be regarded as a legitimate aim which is justified by overriding reasons in the public interest. It follows that, unless the social objective can be upheld, the Netherlands’ defense must fail.
97. However, in case the Court should disagree with my conclusions on the economic objective, I shall briefly examine both the appropriateness of the residence requirement in relation to that objective and its proportionality.
– Is the residence requirement appropriate to achieve the economic objective?
98. The Netherlands argues that the residence requirement is an appropriate means to ensure that MNSF does not lead to an excessive, unreasonable financial burden. The Netherlands has submitted a study which it contends demonstrates that eliminating the requirement would result in an additional burden of some EUR 175 million per year.
99. The Commission laconically indicates that it has ‘doubts’ about the Netherlands’ position on the appropriateness of the measure.
100. Even if the Commission makes no convincing effort to refute the Netherlands’ argument and evidence, it is for the Netherlands to make a persuasive case that excluding students who have lived less than three out of six years in the Netherlands is correlated to the unreasonable financial burden it allegedly averts. That does not involve establishing that the residence requirement is the most appropriate measure to achieve the stated objective. (60)
101. I accept the Netherlands’ argument.
102. The residence requirement necessarily excludes a group of potential claimants and hence limits the cost of MNSF. The Netherlands appears to take the view that the additional burden of EUR 175 million per year would undermine the MNSF scheme as it presently stands.
103. I find no reason to question that position. After all, Member States remain free to decide at what point a particular level of funding for studies becomes an unreasonable financial burden with consequences for the overall level of assistance granted under the scheme. It is for the Member State, and not the Court, to determine where that threshold lies.
104. Since the Commission has made no effort to rebut the Netherlands’ position, I conclude that the Netherlands has established that the residence requirement is appropriate.
– Is the residence requirement proportionate in relation to the economic objective?
105. The parties’ arguments on proportionality became clearer at the hearing held at the Court’s initiative.
106. The parties disagree in essence about whether it is proportionate to require migrant workers, who are already connected to the Netherlands through their employment there, also to comply with the three out of six years rule.
107. The Commission contends that the status of migrant worker is sufficient by itself to demonstrate the required degree of connection and that the Netherlands cannot impose an additional residence requirement. It suggests coordination with other Member States as an alternative measure. The Netherlands argues that the status of migrant worker is insufficient and that no alternative measures are available. When deciding to impose the residence requirement, it also took into account that alternative sources of funding and types of financial support may be available, that other Member States make funding similar to MNSF conditional on past residence and that the residence requirement prevents certain risks of fraud.
108. I am not convinced that the residence requirement is proportionate.
109. Unlike the Netherlands, I find that the fact that the Court accepted a residence requirement of five years as proportionate in Förster does not mean that the three out of six years rule is proportionate here. In Förster, the Court relied on the text of Articles 16(1) and 24(2) of Directive 2004/38 to rule that a Member State was not required to grant maintenance assistance for studies to economically inactive EU citizens who had not resided legally in that Member State for a continuous period of five years. (61) Unlike the Advocate General, (62) the Court appeared not to be inclined to question the thesis that the required degree of connection could not be demonstrated through other means.
110. However, Article 24(2) makes clear that the five years’ residence condition in Directive 2004/38 cannot be imposed on migrant workers and their dependent families.
111. Can a Member State nevertheless impose a requirement of three out of six years residence on such persons?
112. I consider it cannot.
113. Unlike the Netherlands, I do not read Bidar as endorsing such a residence requirement. In that case, the Court did not need to examine proportionality because the effect of the residence requirement coupled with the rules on obtaining ‘settled status’ in the United Kingdom was that, whatever his actual degree of integration, Mr Bidar could never qualify for assistance to cover his maintenance costs.
114. The difficulty in assessing the proportionality of the residence requirement in this case is that the parties’ arguments are based on the understanding that the Netherlands can require a certain degree of connection without taking into account that that is a means to an end.
115. On my reading of Bidar, examining the proportionality of the residence requirement involves deciding whether the Netherlands has established that the three out of six years rule does not go beyond what is necessary to avoid an unreasonable financial burden.
116. The Netherlands has indeed submitted evidence to that effect.
117. The figure of EUR 175 million per year is based on a risk analysis that calculates the estimated additional cost of funding, in particular, children of migrant workers (group 1) and Netherlands nationals (group 2) who are currently excluded from MNSF. (63) Eliminating the residence requirement for children in group 2 would, it is said, result in an additional cost of EUR 132.1 million, which is almost three times as high as the cost of EUR 44.5 million resulting from eliminating the requirement for children in group 1.
118. These estimates are based on a range of assumptions that appear, at best, questionable. For example, in calculating the number of children in group 1 residing outside the Netherlands, the authors of the study estimate that between 15% and 30% of Eastern European migrant workers in the Netherlands continue to reside with their families in their home Member State. These workers are therefore assumed to commute either on a daily or a less regular basis from, for example, Warsaw to the Netherlands. At the same time, the fact that these commuting migrant workers may spend more days a week in the Netherlands than in the home Member State is not taken into account in determining whether they are resident in the Netherlands. Another example is that the authors of the study assume that children of cross-border workers will study in the border area where they reside. They therefore do not appear to apply a correction for the children of migrant workers and Netherlands nationals residing abroad, whether or not in a border area, who are entitled to obtain MNSF to study in a border area.
119. Leaving aside these concerns about the methodology applied, children in groups 1 and 2 qualify for funding to study in the Netherlands despite the fact that they do not reside there. The Netherlands has voluntarily assumed the burden of financing such students up to certain maximum limits. The same limits apply to funding to study in the Netherlands and abroad. The Netherlands has not explained why the same financial burden is acceptable when assumed in connection with studies in the Netherlands, but unreasonable in the context of MNSF. (64)
120. If the Court should decide that the Netherlands can require a certain degree of connection independently of concerns about the financial cost of MNSF, I consider that it is nevertheless disproportionate to require a migrant worker and his dependent family members to satisfy the three out of six years rule.
121. The Court has accepted that a residence requirement may be disproportionate if it is too exclusive in nature because it ‘unduly favours an element which is not necessarily representative of the real and effective degree of connection … to the exclusion of all other representative elements’. (65) To be proportionate, the relevant connecting elements must also be known in advance and provision must be made for the possibility of a means of redress of a judicial nature. (66)
122. In my opinion, the Netherlands has not explained convincingly why either a more flexible residence requirement than the three out of six years rule or other elements expressing a comparable degree of connection, such as employment, would not achieve the same objective in a less restrictive manner. In particular, it has not explained why it accepts that an EU citizen residing in the Netherlands during three out of six years is always sufficiently connected to the Netherlands, irrespective of his participation in that society but rejects outright the possibility that a person’s status as a migrant worker might properly serve to demonstrate the requisite degree of connection with the Netherlands.
123. The Netherlands’ other arguments do not lead me to reconsider that conclusion.
124. Unlike the Netherlands, I consider that it is of no relevance that alternative sources of funding may be available to study outside the Netherlands or outside the home Member State for students excluded from MNSF, and that other Member States make funding for studies abroad conditional on a similar requirement. The fact that students may apply to the Netherlands to obtain funding to study in the Netherlands or that they may claim a generally available tax benefit and enjoy other benefits in connection with studies abroad cannot remedy the discriminatory treatment afforded to them in connection with MNSF. In any event, as the Commission submits in rebuttal, it would appear that these alternative benefits may not be as beneficial as MNSF; and their availability does not demonstrate that the residence requirement does not go beyond what is necessary to achieve the desired objective. Measures adopted by other Member States likewise cannot remedy the discriminatory treatment applied by the Netherlands. It is settled case‑law that a Member State cannot justify an unlawful measure based on the fact that other Member States have adopted the same measure and may thus be infringing EU law in the same manner. (67)
125. The Netherlands further contends that the residence requirement: (i) prevents students residing abroad from claiming that they live independently and are thus entitled to a higher grant when in fact they still live at home, and (ii) prevents people from acquiring the status of migrant worker in the Netherlands after a token period of employment, becoming entitled to MNSF and then studying outside the Netherlands (possibly, indeed, in their home Member State).
126. In my opinion, neither risk is peculiar to MNSF. Both also exist in connection with students’ applications to receive funding to study in the Netherlands. The Netherlands has presumably found other ways of addressing the same concerns adequately in relation to that funding since it is granted to Netherlands nationals and migrant workers alike irrespective of where they reside.
127. In any event, the Netherlands can verify a person’s status as a migrant worker (68) and take the measures to guard against abuse of rights and fraud, taking into the account the individual circumstances of the case and the distinction between taking advantage of a possibility conferred by law and an abuse of rights. (69)
128. I therefore conclude that the Netherlands has not demonstrated that the residence requirement is prima facie proportionate.
129. For the sake of completeness, I will consider none the less whether the Commission has put forward other less restrictive measures.
130. The Commission has proposed only one alternative. It suggests that the Netherlands should coordinate with other Member States. In so doing, it relies on a remark I made in Bressol that the host Member State and the home Member State share a responsibility actively to seek a negotiated solution for problems resulting from high volumes of student mobility. (70)
131. I agree with the Netherlands that EU law imposes no duty of coordination. Rather, coordination is a form of cooperation that requires the consent of at least one other Member State. If the Netherlands is entitled to invoke a legitimate aim to justify indirect discrimination, the means to achieve that aim cannot be made conditional upon other Member States’ consent and willingness to find a negotiated solution. Member States remain responsible for the organisation of their education systems. While coordination might resolve some of the difficulties facing Member States which, like the Netherlands, wish to promote student mobility through funding, requiring them to achieve coordination would run counter to the entire spirit of Article 165(1) TFEU. Coordination is not, therefore, an alternative measure.
132. In any event, the Commission has not explained how and why the possibility of coordination demonstrates that the residence requirement is not proportionate.
133. The Netherlands in its rejoinder appears to accept that the Commission put forward three possible measures: limiting where MNSF can be used, limiting the duration of MNSF and the obligation of coordination. However, the first and second options are canvassed in the section of the Commission’s reply where the Commission summarises the measures that the Netherlands itself put forward and discussed in its defence. I therefore do not consider that the Commission has put forward these suggestions. In any event, they are not, properly speaking, less restrictive alternatives. A Member State must be free to offer generous financial support for studies anywhere in the world, provided it respects its obligations under EU law (and, of course, assumes the financial responsibility for the cost of its generous scheme).
– Conclusion
134. I conclude that the indirect discrimination against migrant workers and their dependent family members resulting from the residence requirement cannot be justified on the basis of the economic objective recognised by the Court in Bidar. However, I must still examine whether the residence requirement can be justified on the basis of the social objective invoked by the Netherlands.
Is the residence requirement justified on the basis of the social objective?
– Is the social objective a legitimate aim which is justified by overriding reasons in the public interest?
135. The aim of MNSF is to increase student mobility from the Netherlands to other Member States. It is not to promote mobility between two Member States other than the Netherlands, or from another Member State to the Netherlands, or to fund students residing outside the Netherlands who wish to study where they reside. MNSF is reserved for students who would otherwise study in the Netherlands, and who are — so the Netherlands argues — likely to return there if they study abroad. It thus seeks to target students who are likely to use their experience abroad to enrich Netherlands society and (possibly) the Netherlands employment market.
136. I accept that this is a legitimate aim. Nor does the Commission appear to contest it.
137. ‘Encouraging mobility of students’ is one of the EU’s objectives; and its importance has been stressed by the Parliament and the Council. (71) It is likewise a legitimate objective for Member States to pursue in the organisation of their educational and study finance systems. (72)
138. I also accept that encouraging student mobility serves the public interest. It promotes cultural and linguistic diversity and enhances professional development. In that way, it contributes to a pluralistic society in Member States and in the European Union as a whole.
139. In a fully-integrated European Union, it might not be acceptable to make access to funding conditional on the likely return of a student to the originating Member State, because that would impede freedom of movement for EU citizens. In the absence of harmonisation in this area, however, Member States retain considerable freedom to decide the conditions of entitlement to funding for studies, provided they do so in a manner consistent with EU law.
140. I therefore accept that the social objective is a legitimate aim which is justified by overriding reasons in the public interest.
– Is the residence requirement appropriate to achieve the social objective?
141. The Netherlands argues that the residence requirement is appropriate to ensure that MNSF goes only to the target group.
142. The Commission advances no argument in that regard. It merely states that it has ‘doubts’ about the Netherlands’ position.
143. Even if the Commission once again makes no convincing effort to refute the Netherlands’ argument, it is for the Netherlands to make a persuasive case that the residence requirement is appropriate to achieve the stated objective. (73)
144. I am not convinced that the Netherlands has done so.
145. I accept that where students reside prior to pursuing higher education may have some influence on where they study. It is true that the Netherlands has not submitted evidence substantiating that correlation. I do not consider that to be an obstacle. The actual or potential contribution of a measure to the stated objective can be established through quantitative or qualitative analysis. In the present case, I consider that qualitative analysis is sufficient, and that the argument is inherently plausible.
146. I also agree with the Netherlands that the residence requirement prevents students from using MNSF to study where they reside, since students residing outside the Netherlands are precluded from applying for MNSF.
147. However, I am not convinced that there is an obvious link between where students reside prior to pursuing higher education and the likelihood that they will return to that Member State after completing their studies abroad. I do not regard it as inherently likely that a majority of students who reside in the Netherlands and then study abroad will necessarily return to reside in the Netherlands. There may be ways of encouraging that to happen, (74) but it is not self-evident that past residence is a good way of predicting where students will reside and work in the future.
148. I conclude that the Netherlands has not established that the residence requirement is appropriate to identify the group of students to whom it wishes to give MNSF.
149. For the sake of completeness, I shall consider briefly whether the residence requirement is proportionate in relation to the social objective.
– Is the residence requirement proportionate in relation to the social objective?
150. It is for the Netherlands to show that the three out of six years rule does not go beyond what is necessary to identify the group of students who would otherwise study in the Netherlands and who are likely to return there if they study abroad. (75)
151. I consider that its arguments in that regard are insufficient.
152. I agree with the Netherlands that a requirement to know Dutch or to have a diploma from a Netherlands school would not be effective alternative measures.
153. Proficiency in Dutch is not necessarily a good indicator of whether students would study in the Netherlands without MNSF or whether they will return there after their studies abroad. A Dutch-speaking student may decide to study in Antwerp because he knows the language there. He might also opt to study in Paris to improve his French or in Warsaw to learn Polish.
154. The same reasoning applies to requiring the would-be student to hold a diploma from a Netherlands school. Assuming that a Netherlands school diploma is recognised in other Member States and that the Netherlands similarly accepts the equivalence of diplomas obtained abroad, it is difficult to see any necessary direct correlation between where a school diploma is obtained and whether a particular student would study in the Netherlands without MNSF and will return there after his studies abroad.
155. In any event, both those requirements appear indirectly discriminatory and likely to affect migrant workers in the same way as the residence requirement.
156. Is it sufficient for the Netherlands to advance two measures that are clearly not proportionate ways of achieving the objective (and that are, in any event, as (if not more) discriminatory as the residence requirement) in order to show that the residence requirement satisfies the proportionality test?
157. I consider it is not.
158. As the party bearing the burden of proof, the Netherlands needs at least to show why it favours residence of three out of six years to the exclusion of all other representative elements, such as (for example) residence of a shorter duration, or why the target group cannot be identified through other (possibly less restrictive) measures, such as (for example) a rule prescribing that MNSF cannot be used to study in the place of residence.
159. If the Court were none the less to take the view that the Netherlands has established that the residence requirement is in principle proportionate, I consider that the Commission has failed to show that other, less restrictive, measures exist that achieve the same result. It is quite unclear from the Commission’s written and oral observations whether it was putting forward any such alternatives. If its argument with regard to coordination is meant to apply in relation to the social objective, I consider that that argument should be rejected for the reasons already given. (76)
– Conclusion
160. I conclude that the indirect discrimination against migrant workers and dependent family members resulting from the residence requirement could in principle be justified on the basis of the social objective invoked by the Netherlands. However, I am not convinced that the Netherlands has shown that the residence requirement is an appropriate and proportionate means of attaining that objective. In my view, its defence must accordingly fail.
Conclusion
161. In the light of all the foregoing considerations, I am of the opinion that the Court should:
(1) declare that, by requiring that migrant workers and dependent family members fulfil a residence requirement to be eligible under the Wet Studiefinanciering for the funding of educational studies abroad, the Kingdom of the Netherlands has failed to fulfil its obligations under Article 45 TFEU and Article 7(2) of Regulation (EEC) No 1612/68 of the Council of 15 October 1968 on freedom of movement for workers within the Community;
(2) order the Kingdom of the Netherlands to pay the costs.