OPINION OF ADVOCATE GENERAL
CAMPOS SÁNCHEZ-BORDONA
delivered on 25 October 2017 (1)
Joined Cases C‑398/16 and C‑399/16
X BV (C‑398/16),
X NV (C‑399/16)
v
Staatssecretaris van Financiën
(Request for a preliminary ruling
from the Hoge Raad der Nederlanden (Supreme Court of the Netherlands))
(Preliminary-ruling proceedings — Corporation tax — Freedom of establishment — Deduction by a resident parent company of interest on a loan for the purchase of shares in a non-resident subsidiary — Deduction by a resident parent company of the reduction in the value of shares in a non-resident subsidiary due to fluctuations in the exchange rate — Integrated group)
1. The Court has had the opportunity on many occasions to rule on the legislation of the Member States relating to tax levied on the profits of companies in the case of groups composed of a parent company and its subsidiaries. (2)
2. In particular, the Netherlands tax provisions on integrated groups of companies have been examined in at least two previous judgments. (3) In accordance with those provisions, a group can benefit from the tax integration scheme only if all the companies in the group are resident in the Netherlands, which means that non-resident subsidiaries must be excluded.
3. In the judgment in X Holding, (4) the Court recognised, in principle, the compatibility of that Netherlands legislation with EU law (to be precise, the provisions of the FEU Treaty relating to freedom of establishment). The Court agreed, then, that the exclusion of non-resident companies from that tax integration scheme was justified in view of the need to safeguard the allocation of the power to impose taxes between the Member States.
4. However, the scope of the judgment in X Holding was later qualified in the judgment of 2 September 2015, Groupe Steria, (5) in which the Court observed that it could not be inferred from the former judgment that ‘any difference in treatment between companies belonging to a tax-integrated group, on the one hand, and companies not belonging to such a group, on the other, is compatible with Article 49 TFEU.’ (6) The Court added that the justification accepted in the judgment in X Holding related only to the provisions of the Netherlands scheme which allowed losses to be transferred within the tax-integrated group. (7)
5. By the questions it has submitted in the two references for a preliminary ruling in the present proceedings, the national court in reality seeks clarification of the case-law laid down in that area. The national court needs that clarification in order to adjudicate on the compatibility with EU law of the Netherlands tax integration scheme, pursuant to which certain items of expenditure may be deducted in the parent company’s accounts if a subsidiary is resident but not if a subsidiary is non-resident.
6. Netherlands law provides that costs (interest) incurred by a company in obtaining finance from another entity in the group are not deductible unless a tax integration scheme has been concluded between the parent company and the subsidiary, which is open solely to resident companies. Case C‑398/16 is concerned with that national rule.
7. That same criterion applies to capital gains and losses (including those derived from losses due to fluctuations in the foreign currency exchange rate), which are not taken into account when calculating the profit. Therefore, a currency loss arising from a shareholding which a parent company has in its subsidiary will not be deductible unless, I repeat, both companies are members of an integrated group, which is open solely to resident companies. That is precisely the rule applied in Case C‑399/16.
I. Legal framework
A. EU law
8. Article 49 TFEU provides:
‘Within the framework of the provisions set out below, restrictions on the freedom of establishment of nationals of a Member State in the territory of another Member State shall be prohibited. Such prohibition shall also apply to restrictions on the setting-up of agencies, branches or subsidiaries by nationals of any Member State established in the territory of any Member State.
Freedom of establishment shall include the right to take up and pursue activities as self-employed persons and to set up and manage undertakings, in particular companies or firms within the meaning of the second paragraph of Article 54, under the conditions laid down for its own nationals by the law of the country where such establishment is effected, subject to the provisions of the Chapter relating to capital.’
9. In accordance with Article 54 TFEU:
‘Companies or firms formed in accordance with the law of a Member State and having their registered office, central administration or principal place of business within the Union shall, for the purposes of this Chapter, be treated in the same way as natural persons who are nationals of Member States.
“Companies or firms” means companies or firms constituted under civil or commercial law, including cooperative societies, and other legal persons governed by public or private law, save for those which are non-profit-making.’
B. Netherlands law
Wet op de vennootschapsbelasting 1969 (Law on corporation tax, 1969; ‘the 1969 Law’)
10. Pursuant to Article 10a(2):
‘When profits are being calculated, interest may not be deducted if it relates to debts owed to a related entity, to the extent that those debts relate to the acquisition of … shares in a related entity, except in so far as a change is made to the ultimate ownership or the ultimate control of that entity.’
11. Under Article 10a(3), paragraph 2 will not be applicable if the taxpayer demonstrates conclusively that the loan and the associated legal transaction are based on economic reasons.
12. In accordance with Article 13(1), benefits derived from a holding and costs incurred in the purchase or transfer of that holding are not taken into account when calculating profits (‘the participation exemption’).
13. Article 15 provides:
‘1. Where a taxable person (the parent company) holds, legally and economically, at least 95% of the shares in the nominal paid-up capital of another taxable person (the subsidiary) and where both taxable persons so request, tax shall be levied on them as if they were a single taxable person, that is, as if the activities and assets of the subsidiary formed part of the activities and assets of the parent company. The parent company shall be liable to pay the tax. Both taxable persons shall together be regarded as a tax entity. More than one subsidiary may form part of the same tax entity.
…
3. Paragraph 1 shall apply only if:
…
b. both taxable persons are subject to the same legislation for the purposes of calculating the profits;
c. both taxable persons are established in the Netherlands ...’
II. Facts of the disputes and the questions referred for a preliminary ruling
A. Case C‑398/16
14. The Netherlands company (8) X BV is part of a Swedish group, which also includes an Italian company. In order to purchase shares in the latter, which was controlled by third parties, X BV set up another company in Italy, to which it contributed capital in the amount of EUR 237 312 000. That contribution was financed by means of a loan (plus interest) to X BV from a Swedish company in the same group.
15. In 2004, as a result of the loan, X BV owed the Swedish lending company the sum of EUR 6 503 261 in respect of interest. In its 2004 corporation tax return, X BV deducted that amount as a cost deductible from its revenue. However, the Netherlands tax authority did not allow that deduction, pursuant to Article 10a(2)(b) of the 1969 Law, and it issued X BV with the notice of assessment disputed in the proceedings.
16. In the action contesting that assessment, X BV argued that it could have deducted interest on the loan if it had been permitted to create a single tax entity with its subsidiary. Since Netherlands law reserves that right to resident companies, X BV claims that its freedom of establishment has been limited contrary to Articles 49 and 54 TFEU.
17. The Hoge Raad der Nederlanden (Supreme Court of the Netherlands), which is seised of the case on appeal, has referred the following question to the Court for a preliminary ruling:
‘Must Articles 43 EC and 48 EC (now Articles 49 and 54 TFEU) be interpreted as precluding national legislation on the basis of which a parent company established in a Member State is not allowed to deduct interest in respect of a loan associated with a capital contribution made to a subsidiary established in another Member State, whereas that deduction could have been availed of if that subsidiary had been included with that parent company in a single tax entity — with characteristics such as those of a Netherlands single tax entity — in view of the fact that, in that case, by reason of consolidation, there would be no obvious association with such a capital contribution?’
B. Case C‑399/16
18. The Netherlands company X NV belongs to a group of companies which comprises, inter alia, and as a single tax entity, the subsidiary A Holdings BV. The latter, for its part, holds all the shares in the United Kingdom company A Holdings UK.
19. On 11 November 2008, A Holdings BV transferred its shares in A Holdings UK to its United Kingdom subsidiary C. (9)
20. In its corporation tax returns for 2008 and 2009, X NV sought to deduct as an expense the loss on its shareholdings resulting from fluctuations in the exchange rate. The Netherlands authorities did not allow that deduction, in accordance with Article 13(1) of the 1969 Law. (10)
21. X NV brought an action against the administrative decision, arguing that, if it had been permitted to form an integrated group with its United Kingdom subsidiary, it would have been able to deduct the currency loss incurred. Since Netherlands law reserves that right to resident companies alone, X NV claims that it has been limited in the exercise of its freedom of establishment.
22. The Hoge Raad der Nederlanden (Supreme Court of the Netherlands), which is seised of the case on appeal, has referred the following questions to the Court for a preliminary ruling:
‘(1) Must Articles 43 EC and 48 EC (now Articles 49 and 54 TFEU) be interpreted as precluding national legislation on the basis of which a parent company established in a Member State cannot take into account a currency loss in connection with the amount which it has invested in a subsidiary established in another Member State, whereas it would be able to do so if that subsidiary were to be included in a single tax entity — with characteristics such as those of the Netherlands single tax entity — with that parent company established in the first-mentioned Member State, as a result of consolidation within the single tax entity?
(2) If the answer to Question 1 is in the affirmative: can or must the point of departure for determining the currency loss to be taken into account be that (one or more of) the direct and indirect subsidiaries indirectly held by the parent company concerned, through the subsidiary in question, and established in the European Union, should also be included in the single tax entity?
(3) If the answer to Question 1 is in the affirmative: should account be taken only of currency losses that would have been reflected on the parent company’s inclusion in the single tax entity in the years to which the dispute relates, or should the currency exchange results that would have been reflected in earlier years also be taken into account?’
III. Summary of the arguments of the parties
A. Case C‑398/16
23. X BV maintains that it is entitled to deduct the interest on a loan received from a Swedish subsidiary which is part of its group of companies. It would have been allowed to do so if the subsidiary were resident for tax purposes in the Netherlands and X BV had formed an integrated group with it, something which is not possible under Netherlands law. The difference in treatment which that law creates means that investment to establish subsidiaries in other EU Member States is less attractive than investment in the Netherlands.
24. X BV further maintains that that difference can be justified only by an overriding reason in the public interest, provided that it does not go beyond what is necessary to protect that interest.
25. Citing the judgment in X Holding, X BV argues that refusal to allow the deduction does not satisfy the objective of safeguarding the balanced allocation of the power to impose taxes between the Member States, since the deduction does not lead to the transfer of the tax base from one Member State to another.
26. In X BV’s submission, nor can the limitation applied be attributed to the need to ensure the coherence of the tax scheme for integrated groups. X BV maintains that the Court of Justice (11) does not allow that justification unless there is a direct link between the granting of the tax advantage concerned and the offsetting of that advantage by a particular tax. In the situation at issue, there is no direct link (in the sense that the advantage is offset against the tax) between the right to deduct the interest on the loan from the profits of the integrated group and the disadvantages referred to in the order for reference. (12)
27. X BV claims that the Court should reply in the affirmative to the question referred.
28. The Commission contends that the relationship between a Netherlands parent company and its subsidiary which is also a Netherlands company is subject to different treatment from that afforded to the relationship between that parent company and a non-resident subsidiary. Under Article 10a of the 1969 Law, interest on a loan from a company in the group may be deducted, by means of consolidation, only in the first situation where the amount of that loan is allocated to a capital contribution to another subsidiary.
29. The Commission states that that disparity is not derived directly from Article 10a of the 1969 Law, since that article refers to certain transactions between associated entities, for the purposes of preventing abuse, and it is applicable without distinction to domestic and cross-border situations. The difference between domestic and intra-Community relationships is a result of the tax regime for integrated groups: whereas, in a transnational situation, it is impossible to avoid the application of Article 10a of the 1969 Law, it is possible to do so in a purely national situation through the creation of an integrated group.
30. Based on the judgment in X Holding, the Commission contends that the difference in treatment must be justified by overriding reasons in the public interest, for the situation of a resident parent company wishing to form a single tax entity with a resident subsidiary and the situation of a resident parent company wishing to do so with a non-resident subsidiary are objectively comparable.
31. From the point of view of the allocation of the power to impose taxes between the Member States, the Commission submits that there is no justification, since this case involves only the Netherlands’ power to impose taxes.
32. As regards the possible effects on the coherence of the Netherlands tax system, the Commission submits that this cannot be invoked because there is no direct link between the tax advantage concerned and the offsetting of that advantage by a particular tax levy. (13) In the Commission’s submission, it is the Netherlands tax system itself which lacks coherence because, on the one hand, it considers it necessary to apply the anti-abuse provisions laid down in Article 10a of the 1969 Law to both national and cross-border situations, whilst, on the other hand, it permits purely national integrated groups to avoid the application of those rules.
33. The Netherlands Government contends that Article 10a(2) of the 1969 Law does not, in itself, conflict with the freedom of establishment. Any obstacles in this case are derived from the fact that a resident parent company cannot form a single tax entity with a non-resident subsidiary. Nevertheless, the Netherlands Government relies on overriding reasons in the public interest to justify that provision.
34. The deductibility of interest within an integrated group is derived from its very nature. As a result of consolidation, a capital contribution between a parent company and a subsidiary is not fiscally visible within the single tax entity because inter-group transactions are neutralised. Since in the integration scheme there is just a single pool of assets allocated to the parent company, a capital contribution will not be fiscally possible within the single tax entity. Therefore, in those circumstances, Article 10a of the 1969 Law is not applicable and the deduction of interest is linked directly and inextricably to consolidation within the single tax entity.
35. The Netherlands Government concludes that the so-called ‘per-element approach’ which followed the judgment in X Holding is not applicable to the disputed provision. However, should the Court adopt such an approach, the difference in treatment may be justified by an overriding reason in the public interest: the limitation of the right to deduct interest is aimed at precluding artificial arrangements not based on economic reasons. The right granted to taxpayers in Article 10a(3) of the 1969 Law, which enables them to demonstrate that they have not used an artificial arrangement, ensures the proportionality of the measure.
B. Case C‑399/16
1. The first question
36. X NV and the Commission proceed on the basis that, under Netherlands law, the currency loss incurred in relation to X NV’s shares in a United Kingdom subsidiary cannot be deducted in the corporation tax return. That loss would have been deductible within the tax integration scheme if the subsidiary had been established in the Netherlands, a situation objectively comparable to that of a Netherlands company with a Netherlands subsidiary which carries on its activities in the United Kingdom. The different treatment of comparable situations constitutes a barrier to freedom of establishment.
37. The only justifications for that difference in treatment are the preservation of the balanced allocation of the power to impose taxes between Member States and the safeguarding of the coherence of the Netherlands tax system.
38. As regards the first justification, X NV and the Commission submit that the Netherlands’ power to impose taxes is not undermined in this case. The currency loss incurred by the Netherlands company in connection with its holding in the United Kingdom subsidiary is not visible in the latter’s accounts which are drawn up in pounds sterling.
39. As regards the second (possible) justification, X NV and the Commission observe that, for the safeguarding of the coherence of the tax system to be acceptable as such, there has to be a direct link between the granting of the tax advantage, on the one hand, and the offsetting of that advantage by a particular tax levy, on the other. In this case, that link does not exist between the advantage (that is, deduction of the currency loss) and the disadvantages mentioned by the referring court. (14)
40. The Netherlands Government contends that X NV is seeking to avoid the negative effect of Article 13 of the 1969 Law (the fact that it is not possible to deduct the currency loss) and to avail itself, nevertheless, of the ‘participation exemption’. However, under Netherlands law, that limited application of the ‘participation exemption’ cannot be obtained by a parent company with a resident subsidiary.
41. The Netherlands Government submits that in the judgments in X Holding and Groupe Steria, (15) the Court held that the offsetting of profits and individual losses of companies in a single tax entity, attributed to the parent company, and the neutralisation of intragroup transactions are inextricably linked to the formation of a single tax entity, within which a shareholding in another company in the group and the profits from that shareholding do not have any tax implications.
42. The option of forming an integrated group does not entail, a priori, a tax advantage in relation to the risk of currency loss, for, although Netherlands law prohibits the deduction of currency losses, it does not include currency gains in the tax base for corporation tax, either. Accordingly, the exclusion of currency losses relating to a shareholding in a non-resident subsidiary is not an impediment to freedom of establishment.
43. Moreover, currency losses cannot be taken into account in the taxation of an integrated group, even if application of the ‘participation exemption’ is not permitted. Since factors related to the holding of shares (such as the distribution of profits and fluctuations in the value of the shareholding) are not included in the final results of a single tax entity, they cannot give rise to any deductions. Therefore, as far as currency losses are concerned, there is no difference in treatment between a parent company with a non-resident subsidiary and a parent company with a resident subsidiary (where both are part of an integrated group).
2. The second and third questions
44. X NV and the Commission observe in relation to the second question that Netherlands law leaves a parent company free to decide whether or not to form an integrated group with its resident subsidiaries and also allows it to choose which subsidiaries to form such a group with. The treatment afforded to cross-border groups should not be less favourable than that afforded to entirely national groups, as regards certain aspects. In other words, since the basis for comparison is the deductibility of currency losses in relation to the taxation of national single tax entities, the treatment of a parent company with a foreign subsidiary cannot be less favourable.
45. As regards the third question, X NV and the Commission maintain that the applicant company should not be taxed more unfavourably than a national group comprising a parent company and its subsidiaries in a single tax entity.
46. The Netherlands Government submits that the answer to the second and the third questions should be the same. The parent company cannot be permitted to choose which companies and which tax years it includes in the return of the fictional integrated group. If the parent company did have that freedom, it would be exercised a posteriori, based on information already known, which would enable the parent company to take into account changes in the exchange rate and to cherry pick the best subsidiary and tax year to include, with the risk that the tax base might be eroded.
IV. Proceedings before the Court of Justice
47. The orders for reference were received at the Court Registry on 18 July 2016.
48. On 9 August 2016, it was decided to join Cases C‑398/16 and C‑399/16.
49. Written observations were lodged by X BV, X NV, the Netherlands Government and the European Commission. It was not considered necessary to hold a hearing.
V. Assessment
A. Preliminary considerations
50. The questions which the Court must address in these preliminary-ruling proceedings concern the taxation of groups of companies, which may be subject to a number of different models of tax regime. In accordance with one such regime, each entity included in the group acts as a separate taxpayer, that is, it pays tax on the total revenue received even though some of that revenue comes from transactions carried out with entities in the same group.
51. However, other legal provisions grant groups the option of paying tax in accordance with the special tax integration scheme, so that tax is levied just once on the group itself, as an economic unit (specifically, on the parent company). That model means that transactions between entities in the group are fiscally neutral; in other words, they are not taken into account for the purpose of establishing the tax base for the levy.
52. The common factor in the two cases referred for a preliminary ruling is that the parent company, which is resident in the Netherlands, states that it has incurred financial losses as a result of its relationships with its subsidiaries, which cannot be deducted in its corporation tax return because Netherlands law prohibits this. The parent company further states that it would be allowed to deduct those losses if it were able to form a single tax entity (integrated group) with those non-resident subsidiaries.
53. Netherlands law provides that groups of companies may benefit from a tax integration scheme only if they are composed of companies resident in the Netherlands. Non-resident companies are not eligible for such a scheme. As I have observed, (16) the Court agreed in the judgment in X Holding that losses incurred by a subsidiary not resident in the Netherlands could not be taken into account in order to reduce the tax base of the parent company because Netherlands law restricts consolidation to resident subsidiaries. The Court held that that difference in treatment was justified by the need to protect the balanced allocation of the power to impose taxes between Member States.
54. The Hoge Raad (Supreme Court) interpreted the judgment in X Holding as meaning that Articles 43 EC and 48 EC not only do not preclude ‘the consequence arising from the very essence of the single tax entity (reserved for resident companies) that losses can be offset within the single tax entity, but also do not preclude other differences in treatment which, when determining tax liability, result from consolidation’. By those judgments, ‘the Hoge Raad (Supreme Court) has not allowed taxable persons, by invoking freedom of establishment, to share at will in the benefits of individual elements directly associated with the essence of the single tax entity (consolidation).’ (17)
55. Therefore, in the view of the referring court, it is possible to infer from the judgment in X Holding an ‘all or nothing’ rule, in accordance with which it is not lawful to choose only certain effects of the single tax entity regime. Any advantage derived from the formation of a single tax entity with resident subsidiaries is justified because it is not possible to adopt a ‘per-element approach’ in relation to a single tax entity. That is why it was decided not to extend the effects of tax integration to non-resident entities.
56. The subsequent rulings of the Court in the judgments in Groupe Steria and Finanzamt Linz (18) have led to the referring court’s uncertainties. In those two cases, the differences in treatment ruled unlawful by the Court concerned precisely ‘specific elements’ (the costs and expenses related to a parent company’s shareholding, in the former case, and the depreciation of goodwill, in the latter) of the relationship between parent companies and resident and non-resident subsidiaries, in the context of the taxation of groups of companies.
57. Specifically, in Groupe Steria, the Court held that a separate assessment must be made of ‘tax advantages other than the transfer of losses within the tax-integrated group’. Only at the end of that assessment will it be possible to determine ‘whether a Member State may reserve those advantages to companies belonging to a tax-integrated group and consequently exclude them in cross-border situations.’ (19)
58. I shall examine the questions referred for a preliminary ruling in accordance with that premiss, following which it will be necessary to determine whether the Netherlands legislation applied in the present two cases is incompatible with Article 49 TFEU, which requires the abolition of restrictions on the freedom of establishment.
59. I shall follow the method that the Court frequently uses when tackling references for preliminary rulings in similar cases relating to the scope of direct taxation. The method is to proceed in steps or stages, seeking first of all to identify the relevant freedom and the possible restriction which might have occurred. The second step is for the Court to compare the situations at issue to see if they have been treated differently, which requires a detailed examination of the domestic legislation giving rise to this. Lastly, the Court will investigate any possible justifications based on overriding reasons in the public interest and the proportionality of the national legislation restricting the relevant freedom.
60. I should note at the outset that a difference in tax treatment between resident and non-resident subsidiaries in the State of the parent company may signify for the latter an impediment to its freedom of establishment by deterring it from setting up subsidiaries in other Member States. (20) The provision of the FEU Treaty in question is, therefore, Article 49 and that difference in treatment between resident and non-resident subsidiaries involves a restriction of the freedom enshrined therein.
61. In order for such a difference in treatment to be compatible with the provisions of the Treaty on the freedom of establishment, it must relate to situations which are not objectively comparable or be justified by an overriding reason in the general interest. (21) Even if the difference is justified, it is necessary that its application be appropriate to ensuring the attainment of the objective thus pursued and not go beyond what is necessary to attain it. (22)
B. Case C‑398/16
62. In summary, the factual starting point is the relationship involving three companies in the same group which are resident in a number of Member States. First, a Swedish company granted an interest-bearing loan to a Netherlands company. Secondly, that Netherlands company invested the loan capital in the purchase of shares in an Italian subsidiary.
63. The dispute arose because the Netherlands company sought to deduct in its corporation tax return the amount of interest owed to the company in the same group.
64. In general terms, Article 10a(2) of the 1969 Law prohibits the deduction of interest where a loan has been effected between undertakings in the same group (associated undertakings). However, it is possible to circumvent that limitation if those associated undertakings exercise the option of being taxed as an integrated group or single tax entity.
65. In Netherlands law, the formation of integrated groups is governed by the following principles:
– The companies which are going to form the integrated group have freedom of choice. (23)
– The right to form an integrated group is restricted to companies which are resident in the Netherlands. (24)
66. According to paragraph 2.8.3 of the order for reference, the difference in treatment is derived from the fact that if the subsidiary had been established in the Netherlands, it could have been included in the single tax entity with the Netherlands company. In those circumstances, Article 10a of the 1969 Law would not have been applicable and the interest on the loan would have been deductible.
67. Therefore, investment to purchase the entire capital of a resident subsidiary is more attractive than investment to purchase the capital of a non-resident subsidiary: the financial costs (interest) of the loan taken out to buy the shares may be deducted in the first case but not in the second.
68. Are the situations under consideration comparable? The Court has replied to that question in the affirmative, specifically in relation to the same Netherlands legislation (Article 15 of the 1969 Law) as that applicable to the present two references for a preliminary ruling.
69. In the judgment in X Holding, the Court observed that ‘the situation of a resident parent company wishing to form a single tax entity with a resident subsidiary and the situation of a resident parent company wishing to form a single tax entity with a non-resident subsidiary are objectively comparable with regard to the objective of a tax scheme such as that at issue in the main proceedings in so far as each seeks to benefit from the advantages of that scheme, which, in particular, allows the profits and losses of the companies constituting the single tax entity to be consolidated at the level of the parent company and the transactions carried out within the group to remain neutral for tax purposes’. (25)
70. The judgment in X Holding concerned whether the losses of a subsidiary could be deducted by the parent company, for which purpose the subsidiary’s profits in the tax year as a whole were taken into account. Although the deduction sought in the present case is not identical, (26) I believe that there are also two objectively comparable situations, since the case relates to the financial cost borne by the parent company which is linked to its holding in the subsidiary, irrespective of whether or not there is consolidation.
71. There are also certain parallels with the Groupe Steria case. In that case, the issue to be determined was whether the so-called ‘proportion of costs and expenses’ (which represented the costs to be borne by the parent company by virtue of its holding in the subsidiary) were deductible within the tax integration scheme, from which non-resident subsidiaries were excluded. Now, the debate relates to a different cost (the interest on the loan) incurred by the parent company, with a view to determining whether the classification of that cost as non-deductible renders the exercise of freedom of establishment less attractive to the same extent.
72. Thus, from the perspective of the ‘per-element approach’ to which the national court refers, I believe that the situations are undeniably comparable and, therefore, that the treatment afforded to similar tax conduct is undeniably different.
73. Having established the difference in the treatment of objectively comparable situations, attention must now be focused on whether there is any overriding reason in the public interest which justifies that difference. The referring court cites, in that connection, the coherence of the Netherlands tax integration scheme.
74. The Court held in the judgment in Groupe Steria that ‘[f]or an argument based on such justification to succeed, a direct link has to be established between the tax advantage concerned and the offsetting of that advantage by a particular tax levy, the direct nature of that link falling to be examined in the light of the objective pursued by the rules in question (judgment [of 13 March 2014,] … Bouanich, C‑375/12, EU:C:2014:138, paragraph 69 and the case-law cited).’ (27)
75. I am unable to find in the order for reference or in the submissions of the Netherlands Government sufficient reasons to accept that justification. In reality, the arguments put forward by the Netherlands Government (28) are, rather, confined to proposing the ‘unfettered’ application of the judgment in X Holding, and the rejection of the ‘per-element approach’. Further, in citing the judgment in Groupe Steria, the Netherlands Government merely refers to the need to protect the balanced allocation of the power to impose taxes between Member States, an issue on which it does not expand (it fails to explain why there is an imbalance in this case).
76. As concerns Case C‑398/16, the Netherlands Government’s assertion that the tax integration scheme is part of a coherent package of advantages and disadvantages is too general. As I have just stated, the Netherlands Government fails to provide any conclusive evidence that, as far as the deduction of interest on a loan to a subsidiary is specifically concerned, the coherence of that system is undermined.
77. The Netherlands Government’s observations concerning the fight against tax evasion, as an overriding reason in the public interest (which, however, the referring court does not mention in paragraph 2.8.6 of its order) are more explicit. The Netherlands Government explains that the aim of Article 10a of the 1969 Law is to preclude the creation of artificial arrangements which are not based on genuine economic reasons and are instead devised solely for the purpose of avoiding payment of the tax due on profits made on Netherlands territory. (29)
78. In the same vein, the Netherlands Government contends that acceptance that a taxpayer is entitled to make a choice a posterioriwould be tantamount to giving precedence to the most favourable tax option. However, that argument is outlined very briefly as regards the difficulties relating to the right to deduct the interest on the loan (it is developed in a little more detail in connection with the losses at issue in Case C‑399/16).
79. A provision like that in Article 10a(2) of the 1969 Law, which is designed to combat tax evasion, may indeed justify certain restrictions of freedom of establishment. (30) As the Netherlands Government explains, whereas dividends received from a company in the group increase the tax base, interest on a loan between the same companies reduces it. There is, therefore, a certain risk to the integrity of the tax base and the aim is to prevent revenue from being ‘counteracted’ in the form of loans giving rise to interest which, in addition to not being included in the recipient’s tax base, also enable the tax base to be reduced in the amount of that interest.
80. Moreover, Article 10a of the 1969 Law applies to relationships between companies in the same group, regardless of whether or not they are resident in the Netherlands. Accordingly, the difficulties do not arise as a result of that provision because, in the context of the fight against tax evasion, it treats residents and non-residents equally.
81. That explanation is, however, undermined by the fact that that aim of tackling tax evasion does not affect resident companies which have exercised the option of being taxed under a tax integration scheme. The fight against tax evasion may explain the existence of the provision but it is difficult to understand why such unequal treatment is afforded to relationships between companies in the same group depending on whether or not they have availed themselves of the tax integration scheme. Whilst, as a general rule, interest on inter-group loans will not be deductible, irrespective of where the companies concerned are established, such interest will be deductible in the case of tax-integrated groups.
82. In the context of tax integration, interest on a loan from a Netherlands (parent) company to a Netherlands subsidiary may be deducted by the former. By contrast, where that same loan is made by the parent company to an Italian subsidiary, integration is not possible and the interest cannot be deducted. From the perspective of tax evasion, if the intention is to prevent the artificial reduction of the parent company’s tax base in the Netherlands, it is impossible to see why that is tolerated when the companies concerned are exclusively Netherlands companies and prohibited when a company from another Member State is involved: the likelihood of tax evasion is the same in both cases.
83. In summary, I do not believe that the fight against tax evasion is an overriding reason in the public interest capable of legitimising unequal treatment, for it is precisely the design of the tax integration scheme in the Netherlands which provides an opportunity for lawful advantage on the part of groups of resident companies while excluding that opportunity for groups including non-resident companies. (31)
84. Furthermore, just as it would be possible to determine whether the grant of a loan by a parent company to a resident subsidiary conceals an artificial arrangement for which there is no genuine economic explanation and which is aimed at reducing, without justification, the former’s tax burden, (32) I see no reason why the same checks could not be carried out with regard to a relationship with a non-resident subsidiary. That possibility of verifying in each case the underlying economic reality is categorically excluded a priori by legislation which simply does not allow it in relation to non-resident companies, which are prohibited from involvement in the Netherlands tax integration scheme.
85. Lastly, the Court has already held, inter alia in the judgment in Euro Park Service, (33) that ‘the imposition of a general rule automatically excluding certain categories of operations from the tax advantage, without account being taken of whether or not there is actually tax evasion or avoidance, would go further than is necessary for preventing such tax evasion or avoidance ...’
C. Case C‑399/16
86. Before I examine the questions referred for a preliminary ruling in this case, I believe it is useful to make two observations. The first concerns the precise extent of the currency losses affecting the parent company’s shareholding in its non-resident subsidiary where both use different currencies.
87. On this point, I believe that the Netherlands Government is correct to make a qualification by distinguishing between ‘currency losses of the subsidiary’ and ‘currency losses on the subsidiary’. The former refer to losses derived from foreign-currency investments made by the subsidiary which are reflected in the subsidiary’s profits. However, ‘currency losses on the subsidiary’, at least from an accounting perspective, reduce the value of the parent company’s investment in its (foreign currency) shareholding in the subsidiary and have an effect on the parent company’s profits.
88. In the present case, the currency losses are directly linked to the value of the shares and not to the profit from investments made by the subsidiary. The case therefore concerns the second of the two scenarios referred to.
89. The second observation relates to the increase or decrease in value of the shares which the parent company holds in the subsidiary’s capital, which may be affected by fluctuations in the exchange rate. That factor may be considered from two angles: (a) changes in value while the shares were included in the parent company’s assets and (b) the difference in value occurring at the time when the shares were transferred, that is, the difference between the purchase value and the transfer value.
90. In my view, the order for reference is not particularly clear on this point: the first question appears to concern the situation where the shares changed hands, whereas the third concerns the period during which those shares were held by the parent company without being transferred.
91. Accordingly, in order to delimit the subject-matter of the dispute properly, the reply to the first question must relate to the currency loss when the shares were transferred; on the other hand, the reply to the third question must relate to the fall in value of the shares while they were included in the parent company’s assets, that is, to the depreciation merely in the book value of the shares.
1. The first question
92. In the light of the parties’ observations, in particular those of the applicant in the main proceedings, the dispute arose because, in the applicant’s view, the currency loss, which came to light at the time of the transfer of shares by A Holdings UK to its subsidiary C, would have been deductible if X NV had been able to include the United Kingdom subsidiary in its integrated group. (34)
93. The reason lies in the fact that Article 13 of the 1969 Law lays down the so-called ‘participation exemption’: when calculating a company’s profits, the advantages derived from a shareholding and the costs incurred when that shareholding is purchased or transferred are not to be taken into account. However, that rule does not apply to tax integration schemes.
94. The referring court identifies that difference in treatment, stating that a parent company established in the Netherlands ‘cannot take into account a currency loss in connection with the amount which it has invested in a subsidiary established in another Member State, whereas it would be able to do so if that subsidiary were to be included in a single tax entity — with characteristics such as those of the Netherlands single tax entity — with that parent company established in the [Netherlands], as a result of consolidation within the single tax entity’.
95. At first sight, therefore, a difference in treatment can be identified which is liable to limit freedom of establishment. That difference will be compatible with the provisions of the FEU Treaty governing that freedom only if it concerns situations which are not objectively comparable or if it is justified by an overriding reason in the public interest.
96. The Hoge Raad (Supreme Court) refers to three approaches for resolving the difficulty that has arisen: (a) the tax regime applicable to permanent establishments abroad should be applied, (b) the resident subsidiaries should operate using a functional currency other than the euro and (c) in addition to currency losses, any currency gains should also be disregarded when calculating the tax base.
97. I shall focus on the last of those three approaches, without it being necessary to address the first two. The referring court rightly refers in that regard to the case-law of the Court in Deutsche Shell (35) and X. (36) Relying on that case-law, the referring court suggests that the non-deductibility of a currency loss may be justified by the fact that any currency gains are not taken into account either. That position (37) is supported by paragraphs 38, 40 and 41 of the judgment in X and by the Opinion of Advocate General Kokott in that case. (38)
98. The dispute on which a preliminary ruling was given in the judgment in X also concerned the deduction of capital losses, in the value of shares, resulting from a currency loss. While the Swedish legislation admittedly prohibited the deduction of currency losses where a parent company sold, with a capital loss, its shares in a non-resident subsidiary, it did not levy tax on capital gains obtained in the same way from such shares either.
99. In those circumstances, the Court observed that ‘it cannot be inferred from the provisions of the FEU Treaty concerning freedom of establishment that that Member State would be required to exercise — asymmetrically, moreover — its taxation powers so as to permit the deduction of losses from operations whose results, if they were positive, would not in any event be taxed’. The Court concluded from this that ‘Article 49 of the TFEU must be interpreted as meaning that it does not preclude the tax legislation of a Member State which, in principle, exempts capital gains on holdings for business purposes from corporation tax and, by the same token, excludes the deduction of capital losses on such holdings, even where those capital losses are due to currency losses.’ (39)
100. Unless I have misconstrued the Netherlands legislation referred to in the order for reference in this case, as concerns the ‘participation exemption’ and the other provisions applicable to the taxation of capital gains and losses resulting from fluctuations in the exchange rate affecting shares in non-resident subsidiaries, I consider that the same solution as that in the judgment in X should be applied.
101. Although in Case C‑686/13, X, the dispute arose in the context of a relationship between a parent company and its subsidiary outside an integrated group, I see no reason why the judgment given in that case should not be applied to the present case. To my mind, the decisive point is that, in circumstances like those at issue, capital gains derived from fluctuations in the exchange rate are not included in the corporation tax base either. In other words, to quote the referring court, this is a situation in which ‘neither currency gains nor currency losses may be taken into account’. (40)
102. If that is the case — which the national court appears to suggest in its summary of domestic law, and the Netherlands Government confirms — the difference in the treatment of the currency loss in the value of the shares held by the parent company in the non-resident subsidiary, when the latter cannot form part of the integrated group, does not limit freedom of establishment, for the same reasons as the Court previously explained in the judgment in X.
103. The answer to this question precludes examination of the other two questions referred by the national court. However, I shall briefly examine those questions together (in line with the Netherlands Government’s submission) in case the Court should decide to answer the first question in the affirmative.
2. The second and third questions
104. If the parent company is entitled to count as losses capital losses on its shareholdings in non-resident subsidiaries derived from fluctuations in the exchange rate, the referring court asks: (a) whether that right extends to ‘(one or more of) the direct and indirect subsidiaries indirectly held by the parent company concerned, through the subsidiary in question, and established in the European Union’; and (b) whether only losses ‘reflected … in the years to which the dispute relates’ or also those in other earlier years should be taken into account.
105. It is apparent from the legislative information in the case-file that the tax integration scheme established under Netherlands law provides for the right to choose which companies form part of the group (created as a single tax entity) and which do not. (41) In principle, therefore, there are no reasons to refuse to allow non-resident subsidiaries something which resident subsidiaries are allowed.
106. The dispute in the main proceedings is concerned solely with the inclusion of currency losses on shares in a particular non-resident subsidiary. Therefore, the wording of the second question is somewhat hypothetical: the issue is whether it is possible to include the United Kingdom subsidiary, and not any other direct or indirect subsidiaries, in the integrated group. Expressed in those terms, the question is inadmissible.
107. Nor do I think that the Court would be able to provide a helpful response to the third question. In fact, none of the parties to the proceedings have proposed that it should do so directly, on the basis of EU law. Even the applicant company in the main proceedings (X NV) proposes that the solution to the problem is to be found in national law rather than EU law, for the concept of taxable annual profits is governed by domestic rules.
108. I should point out that, conceptually, a distinction may be drawn between the loss in value of the shares whilst they were held by A Holdings BV and the loss in their value as a result of the transfer. The question appears to concern the depreciation of the shares, as components of the company’s assets, by virtue of fluctuations in the exchange rate
109. Since the accounts are required to provide information about an undertaking’s financial position on a specific date, the depreciation of the shares may be encapsulated in the accounts using a number of criteria. Insofar as those shares are components of the investing company’s assets, the depreciation of those shares may be reflected in the company’s accounts by, for example, effecting the relevant adjustments to make an allowance for the impairment loss.
110. Where shares are denominated in a foreign currency which is not the currency used by the parent company, changes in the exchange rate may lead to more or less permanent economic fluctuations. For those fluctuations to have an impact on the tax base for the purposes of corporation tax, it will normally be necessary for them to constitute a real economic loss.
111. The Court may not examine whether the reflection in the accounts of the fall in value of the shareholdings has affected the determination of the tax base under national law. It is for the referring court to ascertain, in accordance with national taxation legislation, whether there has been a real economic loss affecting the parent company’s results, either only in the year in which the shares were transferred or in each of the previous years. (42)
VI. Conclusion
112. In the light of the foregoing considerations, I propose that the Court of Justice reply as follows to the Hoge Raad der Nederlanden (Supreme Court of the Netherlands):
Article 49 TFEU:
– Precludes national legislation on the basis of which a parent company established in a Member State cannot deduct interest in respect of a loan associated with a capital contribution made to a subsidiary established in another Member State, whereas that deduction could have been availed of if that subsidiary had been resident in the same State as the parent company.
– Does not preclude national legislation on the basis of which a parent company established in a Member State cannot deduct from its profits losses (capital losses) derived from fluctuations in the exchange rate, in connection with the value of its shares in a subsidiary established in another Member State, where the same legislation does not provide, symmetrically, for tax to be levied on gains (capital gains) derived from those fluctuations.