Opinion of Mr Advocate General Geelhoed delivered on 6 April 2006.

A — Overview of context of dividend taxation

2. Prior to setting out the relevant provisions of the UK tax regime at issue, it is important to outline the broader framework for taxation of distributed company profits (dividends) within the EU, which forms the legal and economic backdrop to the case. In principle, two levels of taxation can arise when taxing the distribution of company profits. The first is at the company level, in the form of corporation tax on the company's profits. The levying of corporation tax at company level is common to all Member States. The second is at the shareholder level, which can take the form of either income taxation on the receipt of the dividends by the shareholder (a method used by most Member States), and/or withholding tax to be withheld by the company upon distribution.⁽³⁾

3. The existence of these two possible levels of taxation may lead, on the one hand, to economic double taxation (taxation of the same income twice, in the hands of two different taxpayers) and, on the other hand, juridical double taxation (taxation of the same income twice in the hands of the same taxpayer). Economic double taxation arises, when, for example, the same profits are taxed first in the hands of the company as corporation tax, and second in the hands of the shareholder as income tax. Juridical double taxation happens, when, for example, a shareholder suffers first withholding tax and then income tax, levied by different States, on the same profits.

4. The present case concerns the legality under Community law of a system set up by the UK with the principal aim and effect of providing a measure of relief for shareholders from economic double taxation.

5. In deciding whether and how to achieve such an aim, there are essentially four systems open to Member States, which may be termed the ‘classical’, ‘schedular’, ‘exemption’ and ‘imputation’ systems. States with a classical system of dividend taxation tax have chosen not to relieve economic double taxation: company profits are subjected to corporation tax, and distributed profit is taxed once again at the shareholder level as income tax. In contrast, schedular, exemption and imputation systems aim at fully or partially relieving economic double taxation.⁽⁴⁾ States with schedular systems (of which various forms exist) choose to subject company profits to corporation tax, but tax dividends as a separate category of income. Those with exemption systems choose to exempt dividend income from income taxation. Finally, under imputation systems, corporation tax at company level is fully or partially imputed onto the income tax due on the dividends at shareholder level, such that the corporation tax serves as a prepayment for (part of) this income tax. Thus, shareholders receive an imputation credit for all or part of the corporation tax attributable to the profits out of which the dividends were paid, which credit can be set against the income tax due on these dividends.

6. At the time relevant to the present case, the United Kingdom used a partial imputation system of dividend taxation.

B — Relevant UK legislation

7. From 1965 (when corporation tax was introduced in the United Kingdom) until 1973, the United Kingdom operated a classical system of dividend taxation which thus, as I described above, did not relieve economic double taxation. In 1973, the United Kingdom moved to a partial imputation system of dividend taxation, with the aim of removing discrimination against distributed profits.⁽⁵⁾ This system essentially functioned as follows.

C — Relevant Community legislation

23. The principal piece of secondary Community legislation of relevance to the present case is the Parent-Subsidiary Directive, which provides for a framework of tax rules regulating the relations between parent companies and subsidiaries of different Member States, with the aim of facilitating the grouping together of companies.⁽²⁷⁾ Article 4 of the Parent-Subsidiary Directive expressly allows both the exemption and the credit method of relieving cross-border double taxation, providing that:

‘1.

Where a parent company, by virtue of its association with its subsidiary, receives distributed profits, the State of the parent company shall, except when the latter is liquidated, either:

• refrain from taxing such profits, or

• tax such profits while authorising the parent company to deduct from the amount of tax due that fraction of the corporation tax paid by the subsidiary which relates to those profits and, if appropriate, the amount of the withholding tax levied by the Member State in which the subsidiary is resident, pursuant to the derogations provided for in Article 5, up to the limit of the amount of the corresponding domestic tax.

2.

However, each Member State shall retain the option of providing that any charges relating to the holding and any losses resulting from the distribution of the profits of the subsidiary may not be deducted from the taxable profits of the parent company. Where the management costs relating to the holding in such a case are fixed as a flat rate, the fixed amount may not exceed 5 % of the profits distributed by the subsidiary.

3.

Paragraph 1 shall apply until the date of effective entry into force of a common system of company taxation.

The Council shall at the appropriate time adopt the rules to apply after the date referred to in the first subparagraph.’

24. Article 6 of the Directive provides that the Member State of a parent company may not charge witholding tax on the profits that company receives from a subsidiary.

A — Applicability of Article 43 or 56 EC: Questions 1 to 4

30. As the national court has raised both Articles 43 and 56 EC in questions 1 to 4, it is necessary as a preliminary matter to consider which of these articles applies in the present case. Just as I observed in my Opinion in Test Claimants in Class IV of the ACT Group Litigation,⁽³⁰⁾ it is my view that the UK legislation at issue may in principle fall within the ambit of either Article 43 or 56 EC, depending on the quality of holding that a given parent company possesses in the relevant foreign subsidiary. The Court has consistently held that a company established in one Member State with a holding in the capital of a company established in another Member State which gives it ‘definite influence over the company's decisions’ and allows it to ‘determine its activities’ is exercising its right of establishment.⁽³¹⁾ As a result, in the case of UK-resident parent companies whose holdings in non-UK companies satisfy this criterion, therefore, it is the compatibility of the UK legislation with Article 43 EC that should be assessed. The application of this criterion in a given case is a matter for the national courts after analysis of the circumstances of the claimant company.

31. In the case of the test claimants in the present reference, it seems clear from the order for reference that these are UK-resident companies (all members of the BAT group) with wholly-owned non-UK-resident subsidiaries. As a result, the test case falls to be considered under Article 43 EC. As I observed in my Opinion in Test Claimants in Class IV of the ACT Group Litigation, although the exercise of freedom of establishment by these UK-resident companies will also inevitably involve the movement of capital out of the UK insofar as this is necessary to establish a subsidiary, this is a purely indirect consequence of the exercise of freedom of establishment. As a result, Article 43 EC takes priority of application for such companies.⁽³²⁾

32. In the case of UK-resident companies holding an investment in a non-UK-resident company which does not give them a ‘decisive influence’ over the latter's activities, or allow them to determine that company's activities, the UK legislation should be assessed for compatibility with Article 56 EC. I note in this regard that the UK legislation at issue clearly concerns what can be termed ‘movement of capital’.⁽³³⁾

33. In principle, therefore, due to the nature of the present case as a group action where the particular circumstances and nature of shareholding of each claimant have not been put before the Court, it is necessary to consider the compatibility of the UK legislation at issue with both Articles 43 and 56 EC.

34. I would add that, although the substantive principles for analysis of whether a breach has occurred are the same for both articles, the geographic and temporal scope of Article 56 EC differs from that of Article 43 EC: Article 43 EC applies only to restrictions on the exercise of freedom of establishment between Member States and entered into force as part of the Treaty of Rome, while Article 56 EC also prohibits restrictions on the movement of capital between Member States and third countries and entered into force on 1 January 1994 (although the principle of free movement of capital had already been established by Directive 88/361).⁽³⁴⁾ Moreover, Article 56 EC is subject to a ‘standstill’ provision — Article 57(1) EC — as regards third States.

35. As a result, as regards the substantive principles for assessment of compatibility, I will only expressly consider Article 43 EC, as the same principles apply to the assessment under Article 56 EC. I will deal separately with certain issues of temporal and geographic scope particular to Article 56 EC (raised in Question 5).

B — Question 1

36. By its first question, the national court asks whether it is contrary to Article 43 or 56 EC for a Member State to keep in force and apply measures which exempt from corporation tax dividends received by a company resident in that Member State from other resident companies and which subject dividends received by the resident company from companies resident in other Member States to corporation tax (after giving double taxation relief for any withholding tax payable on the dividend and, under certain conditions, for the underlying tax paid by the non-resident companies on their profits in their country of residence).

37. The Court has consistently held that, although direct taxation falls within their competence, Member States must none the less exercise that competence consistently with Community law.⁽³⁵⁾ This includes the obligation to comply with Article 43 EC, which prohibits restrictions on the setting up of agencies, branches or subsidiaries by nationals of any Member State established in the territory of any Member State.

38. As I observed in my Opinion in Test Claimants in the ACT Group Litigation,⁽³⁶⁾ Article 43 EC is infringed in a case where the different treatment applied by the relevant Member State to its tax subjects is not a direct and logical consequence of the fact that, in the present state of development of Community law, different tax obligations for subjects can apply for cross-border situations than for purely internal situations. In other words, Article 43 EC prohibits restrictions on free movement of establishment that go beyond those resulting inevitably from the fact that tax systems are national, unless these restrictions are justified and proportionate.⁽³⁷⁾

39. This means that, in order to fall under Article 43 EC, disadvantageous tax treatment should follow from direct or covert discrimination resulting from the rules of one jurisdiction, and not purely from disparities or the division of tax jurisdiction between two or more Member States' tax systems, or from the co-existence of national tax administrations.⁽³⁸⁾

40. In the case of a Member State exercising worldwide (home State) tax jurisdiction, as I noted in my Opinion in Test Claimants in the ACT Group Litigation, this principle means essentially that such a State must treat foreign-source income of its residents consistently with the way it has divided its tax base. Insofar as it has divided its tax base to include this foreign-source income — i.e., by treating it as taxable income — it must not discriminate between foreign-source and domestic income.⁽³⁹⁾ In particular, its legislation should not treat foreign-source income less favourably than domestic-source income.

41. The present question essentially asks whether it is compatible with Article 43 EC for a Member State exercising home State tax jurisdiction to relieve economic double taxation on dividends using an exemption method for domestic-source income, but using a credit method for foreign-source income which

1. in the case of UK companies holding less than 10 % of the voting power of the company paying the dividend (which I will term ‘portfolio holdings’), gave credit only for the withholding tax levied by the source State on the dividends, and

2. in the case of UK companies which directly or indirectly controlled, or were subsidiaries of companies which directly or indirectly controlled, not less than 10 % of the voting power of the company paying the dividend (which I will term ‘non-portfolio holdings’), gave credit for the underlying foreign corporation tax on the profits out of which the dividends were paid.

42. Dealing first with foreign non-portfolio holdings, as the UK and the Commission observe, the objective of eliminating double economic taxation of dividends is achieved by the UK rules in the case of both domestic-source income and foreign-source dividend income. However, this is achieved by different means: in the case of domestic-source income, by exemption from corporation tax on dividend income in the shareholder's hands; in the case of foreign-source income, by crediting the amount of foreign corporation tax paid on the profits constituted in the dividends.

43. In principle, the choice of whether and how to relieve economic double taxation of dividends lies purely with Member States; that is, whether to use a classical (no relief of economic double taxation), schedular, exemption or imputation system (full or partial relief of economic double taxation). If applied in the same way to foreign-source and domestic-source dividend income, each of these systems is perfectly compatible with Article 43 EC.⁽⁴⁰⁾

44. Thus, for example, it is in principle perfectly possible for a credit-based method of economic double taxation relief to apply in a manner that is consistent with Article 43 EC. An example is the solution of the Court in Manninen,⁽⁴¹⁾ which concerned Finnish legislation whereby Finland granted a full imputation tax credit to Finnish shareholders for Finnish corporation tax levied on profits distributed as dividends, but no tax credit for foreign corporation tax levied on foreign-source profits distributed as dividends. In holding that Article 56 EC obliged Finland to extend this tax credit to account for corporation tax levied on dividends incoming from another Member State (Sweden), the Court observed that under the Finnish system, where a Finnish taxpayer invested capital in a Swedish company, there was no way of escaping double taxation of the profits distributed by the company in which the investment was made.⁽⁴²⁾ In contrast, granting a tax credit for Swedish-source dividends would eliminate double taxation on the dividends in the same way as for domestic-source profits.⁽⁴³⁾

45. It is of course true that the application of a credit-based system by the UK for relieving double economic taxation on foreign-source dividends, in the case where the underlying foreign corporation tax levied on company profits was at a higher rate than the UK corporation tax rate, would result in a higher tax burden on these foreign-source dividends than on UK-source dividends (as the UK gives credit only up to the UK corporation tax rate, and not for all foreign corporation tax paid). While, in one sense, this could be said to ‘restrict’ investment in foreign subsidiaries as against UK subsidiaries, this is a good example of a restriction flowing purely from disparities between national tax systems, with which Article 43 EC is not concerned.⁽⁴⁴⁾ Likewise, while the fact that taxpayers receiving foreign-source dividends may, in the case of a credit system, be under an obligation to complete additional formalities in order to prove the amount of foreign corporation tax paid in order to qualify for the credit, this is what I have termed a ‘quasi-restriction’ resulting inevitably from the fact that tax administrations are at present national.⁽⁴⁵⁾

46. In sum, there is in principle no problem under Article 43 EC with the application of a credit system of double economic taxation relief.

47. The present question, however, asks whether Article 43 EC permits a Member State to apply an exemption system for domestic-source dividends and a credit system for foreign-source dividends. The answer to this question depends on whether this distinction has the effect that the UK treats foreign-source dividends less favourably than domestic-source dividends.

48. In this regard, the UK and the Commission argue that, in a domestic context, the effect of exemption and credit systems of economic double taxation relief would be precisely the same. The adoption of a credit system for domestic-source income, however, would mean pointless extra administration costs, while an exemption system, which leads to the same results, is far simpler and less costly to run. Similarly, the effect of the regime for domestic-source dividends (exemption) and foreign-source dividends (credit) is the same: in each case, economic double taxation is relieved.

49. The Test Claimants dispute this conclusion. They argue that a difference exists between the exemption and credit systems in cases where the UK distributing subsidiary has, pursuant to particular UK corporation tax exemptions and benefits (e.g., for investment or Research & Development), in fact paid a lower net rate of corporation tax than the standard UK rate. Under an exemption system, this is ‘passed on’ to the recipient parent company - i.e., the dividends distributed will ultimately thus have borne a tax rate lower than the standard UK corporation tax rate. Under a credit system applied in a domestic context, however, in a case where a lower effective corporation tax rate had been originally borne by the profits pursuant to exemptions and allowances, this rate would always be ‘topped up’ to the standard UK corporation rate upon distribution to the parent company.⁽⁴⁶⁾ Similarly, in the case of foreign-source dividends, the effect of a credit system is that the UK in all cases tops up the effective foreign corporation tax paid to the standard UK rate, without taking account of underlying corporation tax allowances granted at subsidiary level.

50. It would seem, therefore, that the application of a credit system by the UK for the relief of double economic taxation on foreign-source dividends can in certain cases have less favourable effects than the pure exemption system applied to domestic-source dividends. While, under an exemption system, the benefits of underlying corporation tax exemptions and allowances may be passed on to the parent company receiving the dividends, under a credit system these benefits cannot be passed on as the tax borne by the dividends is topped up to the standard UK corporation tax rate. In such cases, the effect of this could be seen as the application by the UK of a different (lower) tax rate to domestic-source dividends than to foreign-source dividends.

51. A further question arises as to whether such discriminatory treatment can be justified. On this point, the UK argues in its submissions that any restriction can be justified on the grounds of fiscal cohesion. It argues, relying on Manninen, that the effect of the UK's system is to relieve economic double taxation for foreign-source and domestic-source dividends. Cohesion is maintained in cross-border situations because the recipient parent company receives credit for all of the foreign tax paid on the profits from which the dividend arose. While the UK's arguments certainly show that, as I observed above, in principle the application of a credit system can be perfectly in accordance with Article 43 EC, they do not go towards justification of the possible difference in treatment, discussed above, between foreign- and domestic-source income as regards the potential ability to pass on the benefit of underlying tax allowances to recipient parent companies.

52. In the absence of a mechanism enabling such tax allowances to be taken into account in a similar way for foreign-source dividends as for domestic-source dividends, therefore — the presence of which has not been contended in the present case — it is my view that the UK taxation rules for non-portfolio dividends infringe Article 43 EC.

53. As regards foreign portfolio holdings, for which credit was given only for the foreign withholding tax levied on the foreign dividends, the UK's provisions seem clearly discriminatory. While UK corporation tax was not chargeable on dividends received by UK companies by virtue of a portfolio holding in another UK company, UK corporation tax was chargeable on dividends received from such a holding in a company resident in another Member State, subject only to credit for foreign withholding tax (and not for underlying foreign corporation tax). Put otherwise, the UK had chosen, in the exercise of its competence, fully to relieve double economic taxation of dividends coming from a portfolio holding in a UK company, without doing so for dividends coming from a foreign portfolio holding. This evidently represents less favourable treatment of foreign-source income falling within its tax jurisdiction than of equivalent domestic-source income.

54. In its written and oral submissions, the UK attempted to justify this on the ground that it would be disproportionately expensive and complex to administer and supervise the grant of tax credits for foreign underlying tax in respect of smaller shareholdings, which complexity would result in delay and legal uncertainty for taxpayers.

55. I am not convinced by this argument. While it is true that affording tax credits for dividends for foreign-source portfolio holdings would place an extra administrative burden on the UK authorities, this burden is in my view not disproportionate to the benefit of relieving economic double taxation for relevant UK corporate shareholders. I would refer on this point to the Court's judgment in Manninen where, in holding that Article 56 EC obliged Finland to extend its imputation tax credit to account for corporation tax levied on dividends incoming from Sweden, the Court rejected arguments based on potential difficulties for the taxpayer or the tax administration to obtain the necessary information on the corporation tax paid in another Member State.⁽⁴⁷⁾ Although, as the Court observed, the calculation of a tax credit granted to a Finnish-resident shareholder receiving dividends from a company established in another Member State must take account of the tax actually paid by that company, and that such tax would arise from the general rules on calculating the basis of assessment and from the rate of corporation tax in that latter Member State, ‘possible difficulties in determining the tax actually paid cannot, in any event, justify an obstacle to the free movement of capital such as that which arises from the legislation at issue in the main proceedings’.⁽⁴⁸⁾ Precisely the same considerations apply in the present case. I note that the option of exempting such dividend income from UK corporation tax (as happens for UK-source dividend income from portfolio holdings) would in any event be available to the UK, should it prefer to avoid additional administrative burdens.

56. For these reasons, the response to the first question should be that it is contrary to Articles 43 and 56 EC for a Member State to keep in force and apply measures such as those at issue in the present case, which exempt from corporation tax dividends received by a company resident in that Member State from other resident companies and which subject dividends received by the resident company from companies resident in other Member States to corporation tax, after giving double taxation relief for any withholding tax payable on the dividend and, under certain conditions, for the underlying tax paid by the non-resident companies on their profits in their country of residence.

C — Questions 2 and 3

57. By its second question, the national court asks whether, where a Member State has a system which in certain circumstances imposes ACT on the payment of dividends by a resident company to its shareholders and grants a tax credit to shareholders resident in that Member State in respect of those dividends, it is contrary to Article 43 or 56 EC or Article 4(1) or 6 of Directive 90/435 for the Member State to keep in force and apply measures which provide for the resident company to pay dividends to its shareholders without being liable to pay ACT to the extent that it has received dividends from companies resident in that Member State (either directly or indirectly through other companies resident in that Member State) and which do not provide for the resident company to pay dividends to its shareholders without being liable to pay ACT to the extent that it has received dividends from non-resident companies.

58. By its third question, the national court asks whether it is contrary to these provisions of Community law for the Member State to keep in force and apply measures which provide for the ACT liability to be set against the liability of the dividend-paying company, and that of other companies in the group resident in that Member State, to corporation tax in that Member State upon their profits, but which

1. do not provide for any form of set off of the ACT liability or some equivalent relief (such as the refund of ACT) in respect of profits earned, whether in that State or in other Member States, by companies in the group which are not residents in that Member State; and/or

2. provide that any double tax relief which a company resident in that Member State enjoys reduces the liability to corporation tax against which the ACT liability can be set.

59. The second question thus concerns the feature of the UK system at issue whereby

1. UK corporate shareholders receiving dividends from UK companies, which had paid ACT upon the distribution of these dividends, received a tax credit equal to the ACT paid by the distributing company, meaning that ACT was paid only once in respect of dividends passed up through UK-resident members of groups of companies; and

2. UK corporate shareholders receiving dividends from non-UK companies did not receive such a tax credit, and were as a result liable to pay ACT on the full amount of profit distributions made. The third question concerns the feature that foreign corporation tax paid on incoming dividends could not be set off against ACT, but only against UK MCT. As ACT paid could in turn only be set off against UK MCT, however, this meant that companies with substantial foreign-source income could have unrelieved ACT (i.e., ACT paid which could not be set off against a company's MCT liability in that accounting period: so-called ‘surplus ACT’). While there were possibilities ultimately to relieve such surplus ACT (e.g., by carrying it back or forward for set-off against MCT of other periods, or by surrendering it to UK-resident subsidiaries), not all companies could avail of such provisions.

60. As these two questions deal with what, in the domestic context, are complementary features of the UK system, the full effect of the system can, in my view, best be appreciated by considering them together.

D — Question 4

93. By its fourth question, the national court asks whether, where the UK measures provide in certain circumstances for resident companies, if they so elect, to recover the ACT paid on distributions to their shareholders to the extent that distributions are received by the resident companies from non-resident companies (including for this purpose companies resident in third countries), it is contrary to Article 43 or 56 EC or Article 4(1) or 6 of Directive 90/435 for those measures

1. to oblige the resident companies to pay ACT and to reclaim it subsequently; and

2. not to provide for the shareholders of the resident companies to receive a tax credit which they would have received on a dividend from a resident company which had not itself received dividends from non-resident companies.

94. This question concerns the compatibility with the abovementioned Community law provisions of the so-called Foreign Income Dividend (‘FID’) Scheme introduced in the UK with effect from 1 July 1994. As I explained above, under this system a UK-resident company could, prior to payment of a cash dividend to its shareholders, elect that this dividend was a foreign income dividend. ACT payment was payable on this FID but, if the company could match the FID with foreign profits, a claim for repayment could be made for surplus ACT arising in respect of the FID. Such surplus ACT was repaid at the same time as MCT became payable, i.e., 9 months after the end of the accounting period, and after set off against any MCT liability for the period. The shareholder receiving the FID was not entitled to a tax credit under section 231(1) TA, but an individual receiving a FID was treated as receiving income which had borne tax at the lower rate for the year of assessment. However, no repayment was made to such shareholders of income tax treated as having been paid, nor could a tax-exempt shareholder reclaim a tax credit similar to that which would have been payable on a non-FID qualifying distribution.

95. The obligation imposed on the UK by Articles 43 and 56 in implementing the FID was precisely the same as I described in the context of Questions 2 and 3 above: namely, an obligation to ensure equivalent full relief of double economic taxation for foreign-source dividends as for domestic-source dividends.

96. As regards the feature of the FID system raised in the first part of Question 4, therefore — the obligation for resident companies receiving foreign dividends to pay ACT on re-distribution and to reclaim this subsequently — this obligation infringed Articles 43 and 56 EC insofar as such ACT resulted in double economic taxation of these foreign-source profits. The fact that resident companies could subsequently reclaim the ACT paid is clearly no defence here: by analogy with the Court's observations in Metallgesellschaft,⁽⁶¹⁾ the cash-flow disadvantage for such companies in the interim period prior to reclaim qualifies, for the purposes of the non-discrimination principle, as less favourable treatment.⁽⁶²⁾

97. The second part of Question 4 concerns the feature of the FID system whereby, to the extent that UK-resident companies had received dividends from non-resident companies, (downstream) shareholders of these UK companies did not receive the tax credit which they would have received on a dividend from a UK-resident company which had not itself received dividends from nonresident companies.

98. In this regard, to the extent that the UK chose to relieve double economic taxation by grant of an imputation tax credit for domestic-source dividends, it is obliged by Articles 43 and 56 EC to relieve double taxation in an equivalent way for foreign-source dividends.⁽⁶³⁾

99. As regards the UK Government's argument that, under its FID system, shareholders of UK companies receiving FIDs could in fact avoid double taxation because they were treated as if they had received income which had borne tax at the lower rate for the year of assessment, I would observe that it is for the national court to assess, in a particular case, whether such treatment actually had the effect that double economic taxation was relieved in an equivalent manner as for domestic-source dividends.

100. Insofar as the FID system did not result in equivalent double economic taxation relief for UK companies' shareholders receiving FIDs as for those receiving domestic-source profits, it breaches Articles 43 and 56 EC unless justified.

101. In its defence, the UK Government argues first of all that the fact that the FID system was a wholly elective regime means that it could not amount to a restriction on any party's freedom of establishment or freedom to move capital: the underlying UK tax provisions (dealt with in Questions 2 and 3 above) remained in place at all material times. To the extent that these underlying tax provisions were also discriminatory and in breach of Articles 43 and 56, however, this argument is clearly fallacious. It was in neither case possible for UK companies with foreign shareholders to receive non-discriminatory treatment of their foreign-source income in comparison with domestic-source income.

102. Second, the UK Government argues that a non-resident subsidiary that has not been obliged to pay ACT upon distribution is in any event in a position to pay a larger dividend to its UK parent company than a resident subsidiary, which must pay ACT upon its distributions, would be able to pay. Again, this argument fails to recognise that, while the non-resident subsidiary has not been obliged to pay ACT, it has none the less been subject to foreign corporation tax which, as I observed above, places parent companies receiving foreign- and domestic-source dividends in a comparable position.

103. Finally, the UK Government raises the argument that the FID system is justified by the need to preserve the fiscal cohesion of the UK tax system, as it argued as regards Question 2 above, and to ensure the effectiveness of fiscal supervision, in particular as concerns third countries. As regards intra-Community restrictions, to the extent that these arguments are substantiated, they repeat those dealt with under Question 2 and should be dismissed for the same reasons. The question whether such arguments may carry greater force concerning ‘third country’ restrictions has been raised under Question 5 and I will deal with it briefly there.

104. A separate issue is whether the only potential claimants of loss caused by such a breach should be the shareholders themselves, and not the UK distributing company. On this point, the Test Claimants contend that parents paying distributions from foreign-source income were caused by the FID system to enhance dividends for shareholders to distribute an equivalent amount as parents paying distributions from domestic-source income. I deal with this point under Question 6, which concerns appropriate remedies for breach.

105. The national court also raises the compatibility of the two features it identified in the FID system — liability of UK companies re-distributing foreign dividends to pay ACT (Question 4(i)) and failure to give imputation tax credits to their downstream shareholders (Question 4(ii)) — with Articles 4(1) and 6 of the Parent-Subsidiary Directive. As regards Article 4(1), as I explained above in relation to Questions 2 and 3, this imposes an obligation fully to relieve double economic taxation at the level of the receiving corporate shareholder on distributions falling within its material and temporal scope. As such, the analysis under Question 4(i) (liability to ACT) is the same as for Articles 43 and 56. In the case of Question 4(ii) (grant of imputation tax credit), however, this concerns a difference in treatment at the level not of the receiving company itself, but of that company's shareholders. For this reason, such a restriction would not seem to me to fall within the scope of Article 4(1) of the Parent-Subsidiary Directive. Likewise, for analogous reasons as I explained above in the context of Questions 2 and 3, neither of the features raised in Question 4 constitutes, in my view, a breach of Article 6 of the Parent-Subsidiary Directive.

106. The answer to the fourth question should therefore be that, where the UK measures provide in certain circumstances for resident companies, if they so elect, to recover the ACT paid on distributions to their shareholders to the extent that distributions are received by the resident companies from non-resident companies (including for this purpose companies resident in third countries), (1) it is contrary to Articles 43 and 56 EC, as well as Article 4(1) of the Parent-Subsidiary Directive, to oblige the resident companies to pay ACT and to reclaim it subsequently, to the extent that this does not ensure equivalent full relief from double economic taxation to that provided for domestic-source dividends; and (2) it is contrary to Articles 43 and 56 EC not to provide for the shareholders of the resident companies to receive equivalent relief from double economic taxation which they would have received on a dividend from a resident company which had not itself received dividends from non-resident companies.

E — Question 5

107. By its fifth question, the national court asks whether where, prior to 31 December 1993, a Member State adopted the measures outlined in Questions 1 and 2, and after that date it adopted the further measures outlined in Question 4, and if the latter measures constitute a restriction prohibited by Article 56 of the EC Treaty, that restriction is to be taken to be a new restriction not already existing on 31 December 1993.

108. The national court raises this question in the context of Article 57(1) EC, which provides that the Article 56 EC prohibition of restrictions of free movement of capital shall be ‘without prejudice to the application to third countries of any restrictions which exist on 31 December 1993 under national or Community law adopted in respect of the movement of capital to or from third countries involving direct investment — including in real estate — establishment, the provision of financial services or the admission of securities to capital markets’. The question essentially asks therefore whether, to the extent that the measures outlined in Question 4 fall within the Article 56 EC prohibition, this prohibition includes restrictions on free movement of capital between Member States, and third countries.⁽⁶⁴⁾

109. The first question is whether the FID rules, which took effect from 1 July 1994, can be considered to form part of restrictions which ‘existed’ on 31 December 1993.

110. On this point, both the Test Claimants and the UK Government rightly refer to the Court's judgment in Konle.⁽⁶⁵⁾ That case concerned the interpretation of a derogating clause for Austria in the Act of Accession of the Republic of Austria, the Republic of Finland and the Kingdom of Sweden and the adjustments to the treaties on which the European Union is founded⁽⁶⁶⁾, which allows it to maintain for a certain period its existing legislation regarding (prior authorisation of) secondary residences. The Court first observed that, while it was for the Austrian courts to determine the content of the national legislation existing at the date of Austria's accession, it was for the Court to supply guidance on the Community concept of ‘existing legislation’.⁽⁶⁷⁾ The Court continued,

• ‘Any measure adopted after the date of accession is not, by that fact alone, automatically excluded from the derogation laid down in [the derogating article of the Act of Accession]. Thus, if it is, in substance, identical to the previous legislation or if it is limited to reducing or eliminating an obstacle to the exercise of Community rights and freedoms in the earlier legislation, it will be covered by the derogation.

• On the other hand, legislation based on an approach which differs from that of the previous law and establishes new procedures cannot be treated as legislation existing at the time of accession.’⁽⁶⁸⁾

111. As I observed in my Opinion in Ospelt, it also follows from Article 57(1) EC that Member States are empowered to adapt existing legislation without altering the existing legal situation.⁽⁶⁹⁾

112. In the context of the present case, on the basis of the description provided in the order for reference, it seems to me that the object and effect of the introduction of the FID regime was indeed to reduce (although not eliminate) the existing obstacle to exercise of the freedoms provided by Articles 43 and 56 EC — namely the failure fully to relieve double economic taxation on foreign-source dividends. Indeed, the existing ACT regime was not removed for foreign-source dividends upon the introduction of the FID changes, which remained optional for companies falling under their scope. I would add that this manner of interpreting Article 57(1) EC, to include legislation aimed at lessening existing restrictions, is wholly logical: in the absence of such an interpretation, Member States would have an incentive to maintain existing restrictions rather than attempt to reduce such restrictions partially or fully.

113. As a result, while final assessment of the content, purpose and effect of the FID system is for the national court, on the basis of the details provided it is my view that the system forms part of restrictions that existed on 31 December 1993 within the meaning of Article 57(1) EC.

114. The second question is whether the FID regime falls within the material scope of Article 57(1) EC.

115. The Test Claimants argue that the regime falls outwith the material scope of Article 57(1) EC, which is limited to movement of capital to or from third countries involving ‘direct investment — including in real estate — establishment, the provision of financial services or the admission of securities to capital markets’.

116. The Test Claimants argue that the term ‘investment’ should be interpreted strictly, and does not extend to payments flowing from such an investment. I do not agree. As the Court has held, as discrimination in the tax treatment of domestic- and foreign-source dividends may render investment in the shares established in other Member States less attractive, it should be considered to be a restriction on the free movement of capital.⁽⁷⁰⁾

117. The Test Claimants further argue that the phrase ‘direct investment’ in this regard should not extend to smaller participations (e.g., portfolio holdings) held by UK companies abroad. In this regard, it is true that, as an exception to Article 56 EC, Article 57(1) EC should be construed strictly.⁽⁷¹⁾ It is my view that the concept of direct investment should be interpreted in accordance with the guidance given in Annex I to Directive 88/361, which outlines the nomenclature for capital movements referred to in Article 1 of that directive. Heading 1 of the Annex deals with ‘Direct Investment’ and the most relevant inclusion for present purposes within this category is subheading 2: ‘Participation in new or existing undertaking with a view to establishing or maintaining lasting economic links.’ The Explanatory Memorandum to the directive sheds further light on what is meant by ‘direct investment’, namely, ‘Investments of all kinds by natural persons or commercial, industrial or financial undertakings, and which serve to establish or to maintain lasting and direct links between the person providing the capital and the entrepreneur to whom or the undertaking to which the capital is made available in order to carry on an economic activity. This concept must therefore be understood in its widest sense.’

118. The Explanatory Memorandum continues, ‘As regards those undertakings mentioned under I-2 of the Nomenclature which have the status of companies limited by shares, there is participation in the nature of direct investment where the block of shares held by a natural person of another undertaking or any other holder enables the shareholder, either pursuant to the provisions of national laws relating to companies limited by shares or otherwise, to participate effectively in the management of the company or in its control.’

119. It is for the national court to decide whether, in a given case, the investment that a UK company holds in a company resident in a third country serves to establish or maintain ‘lasting and direct links’ with the latter company, allowing the UK company to ‘participate effectively in the management of the company or in its control’. I would observe, however, that this is clearly a lower threshold than the ‘decisive influence’ test to which I referred in the context of distinguishing between the scope of application of Articles 43 and 56 EC under Title IV(A) above.

120. Only in the event that a UK company's holding in a third country company was not such as to allow effective participation, therefore, would the Article 56 EC prohibition apply.

121. This raises in turn the issue whether the same considerations apply to the analysis of Article 56 EC as regards intra-Community capital movements as opposed to capital movements between Member States and third countries. In this regard, it is evident from the wording of Article 56(1) EC that restrictions on free movement of capital between Member States and third countries are, in principle, prohibited. Nonetheless, it is my view that, in analysing whether such restrictions are justified (whether under Article 58(1) EC or under the discrimination analysis of Article 56 EC), different considerations may apply than is the case with purely intra-Community restrictions. As I have already observed in my Opinion inOspelt, the particular context of intra-Community freedom of capital is that it must be regarded as a constituent element of economic and monetary union.⁽⁷²⁾ As I stated there, the fact that monetary policy is, following completion of economic and monetary union, set by the European Central Bank, presupposes complete unity in terms of the movement of money and capital. This context does not obtain in the case of movement of capital between Member States and third countries, although the movement of capital has been to a large extent liberalised worldwide.⁽⁷³⁾ As a result, I will not exclude that a Member State may be able to prove that a restriction of capital movements with third countries is justified on a given ground, in circumstances where this ground would not amount to a valid justification of a restriction on purely intra-Community capital movements.

122. In the present case, however, as I observed above, the UK Government has not put forward any substantiated arguments as to why specific considerations apply to justify the restrictions comprised in the FID scheme in the case of third countries. Its arguments as to justification of the scheme are primarily based on fiscal cohesion, and it argues that the ‘drain of revenue’ out of the Community in the case of third country movements is of greater concern than in intra-Community situations. This abstract argument is, however, in my view not sufficient to prove that the restrictions of the FID scheme as regards third country-source dividends were, in that particular case, justified.

123. In any event, due my answer on the scope of Article 57(1) EC in this case, it is in my view not necessary to give a definite response on this issue.

124. The answer to the fifth question should therefore in my opinion be that, where prior to 31 December 1993 a Member State adopted the measures outlined in Questions 1 and 2, and after that date it adopted the further measures outlined in Question 4, and if the latter measures constitute a restriction prohibited by Article 56 EC, that restriction forms part of provisions that already existed on the 31 December 1993, within the meaning of Article 57(1) EC.

F — Questions 6 to 9

125. Questions 6 to 9 of the order for reference raise questions relating to the nature of remedies that should be available to affected UK-resident companies or other companies in the same group in the event of any of the measures set out in Questions 1 to 5 being in breach of any of the Community provisions referred to in those questions.

126. In this regard, the Court has consistently held that that the right to a refund of charges levied in a Member State in breach of rules of Community law is the consequence and complement of the rights conferred on individuals by Community provisions as interpreted by the Court.⁽⁷⁴⁾ The Member State is therefore required in principle to repay charges levied in breach of Community law.⁽⁷⁵⁾

127. In the absence of Community rules on the recovery of sums unduly paid, it is for the domestic legal system of each Member State to designate the courts and tribunals having jurisdiction and to lay down the detailed procedural rules governing actions for safeguarding rights which individuals derive from Community law, provided, first, that such rules are not less favourable than those governing similar domestic actions (principle of equivalence) and, second, that they do not render practically impossible or excessively difficult the exercise of rights conferred by Community law (principle of effectiveness).⁽⁷⁶⁾

128. The question raised in the present case is whether the plaintiffs' claims should be characterised as claims in restitution, damages or for an amount representing a benefit unduly denied.

129. On this point also, the Court's judgment in Metallgesellschaft is relevant. In that case, the second question posed by the national court concerned the appropriate remedy that should be available where a UK subsidiary and non-UK parent had been deprived of the benefit of the group exemption scheme contrary to Article 43 EC. In particular, the question was whether Article 43 EC entitled that subsidiary and/or its parent company to obtain a sum equal to the interest accrued on the advance payments made by the subsidiary from the date of those payments until the date on which the tax became chargeable, even when national law prohibited the payment of interest on a principal sum which was not due. The Court emphasised that it was not its task to assign a legal classification (under English law) to the actions brought by the plaintiffs before the national court, but was for the affected companies to specify the nature and basis of their actions — whether for restitution or for compensation in damages — subject to the supervision of the national court.⁽⁷⁷⁾

130. On this basis, the Court went on to consider issues arising on both of the hypotheses put forward by the national court: first, the hypothesis that the actions were to be treated as claims in restitution, and second, the hypothesis that they were to be treated as claims in damages.⁽⁷⁸⁾ It concluded that, in any event, Article 43 EC required that the plaintiffs should have an effective legal remedy in order to obtain reimbursement or reparation of the financial loss which they had sustained and from which the authorities of the Member State concerned had benefited as a result of the advance payment of tax.⁽⁷⁹⁾ The mere fact that the sole object of such an action would be the payment of interest did not constitute a ground for dismissing such an action.⁽⁸⁰⁾

131. I note that, in that case, no questions had been referred by the national court in respect of the interpretation of the Brasserie du Pêcheur general conditions for State liability for a breach of Community law, and the Court did not consider whether these conditions were made out.⁽⁸¹⁾ Advocate General Fennelly considered the issue briefly, although only in the alternative as his view was that it was ‘more correct and more logical to treat the plaintiffs’ claim as restitutionary rather than as a compensatory claim for damages'.⁽⁸²⁾

132. In the present case, it seems to me that, with one exception, the claims described in the national court's sixth question should be considered equivalent to claims for recovery of sums unduly paid, that is to say, claims for recovery of charges unlawfully levied within the meaning of the Court's case-law, which the UK is in principle obliged to repay. The underlying principle should be that the UK should not profit and companies (or groups of companies) which have been required to pay the unlawful charge must not suffer loss as a result of the imposition of the charge.⁽⁸³⁾ As such, in order that the remedy provided to the Test Claimants should be effective in obtaining reimbursement or reparation of the financial loss which they had sustained and from which the authorities of the Member State concerned had benefited, this relief should in my view extend to all direct consequences of the unlawful levying of tax. This includes to my mind:

1. repayment of unlawfully levied corporation tax (Questions 6(i), (iii) and (vii));

2. the restoration of any relief applied against such unlawfully levied corporation tax (Question 6(ii));

3. the restoration of reliefs foregone in order to set off unlawfully levied corporation tax (Question 6(v));

4. loss of use of money insofar as corporation tax was, due to the breach of Community law, paid earlier than it would otherwise have been (Questions 6(iv), (vi), and (viii)).⁽⁸⁴⁾ In each case, it would be for the national court to satisfy itself that the relief claimed was a direct consequence of the unlawful levy charged.

133. On this point, I am not convinced that the head of claim outlined in Question 6(ix) should qualify as equivalent to a claim for repayment of charges unlawfully levied. The Test Claimants essentially argue that the UK's discriminatory failure to grant equivalent imputation credits to shareholders of UK companies receiving FIDs caused those companies to enhance their distributions to compensate these shareholders. However, it does not seem to me that such actions on the part of the distributing company to increase the amount of distributions should be considered to be a direct consequence of the UK's unlawful failure to grant an equivalent credit to the shareholders. Rather, the direct consequence of this failure is simply the extra tax levied on those shareholders than would have been the case had the UK complied with its Community law obligations — which loss is suffered by the shareholders, and not the distributing companies. In contrast, any increase by these companies in the amount of dividend distributed to its shareholders does not seem to me to follow inevitably from the denial of tax credit, nor it is possible without more to conclude that the distribution of an increased dividend necessarily qualifies as a loss incurred for the distributing companies.

134. In principle, it is for the national court to decide how the various claims brought should be characterised under national law. However, as I observed above, this is subject to the condition that the characterisation should allow the Test Claimants an effective remedy in order to obtain reimbursement or reparation of the financial loss which they had sustained and from which the authorities of the Member State concerned had benefited as a result of the advance payment of tax.⁽⁸⁵⁾ This obligation requires the national court, in characterising claims under national law, to take into account the fact that the conditions for damages as set out in Brasserie du Pêcheur may not be made out in a given case and, in such a situation, ensure that an effective remedy is nonetheless provided.

135. In the present case, for example, I am not convinced that the Brasserie du Pêcheur conditions would be satisfied for all the aspects of the UK system raised in the present reference that, in my opinion, breach Community law. Clearly, the first condition (breach of a rule of law intended to confer rights on individuals) is fulfilled, as each of the Community law provisions raised is directly effective. The same would seem true by and large for the third condition (existence of a direct causal link between the breach of an obligation resting on the State and damage sustained by injured parties), with the potential exception of the claim outlined in Question 6(ix) for the reasons I have mentioned above.

136. However, I have considerable doubts whether the second condition — the existence of a ‘sufficiently serious’ breach of Community law — is fulfilled for all aspects of the system which, in my view, breach Community law. As observed by the Court in Brasserie du Pêcheur,

• ‘… the decisive test for finding that a breach of Community law is sufficiently serious is whether the Member State or the Community institution concerned manifestly and gravely disregarded the limits on its discretion.

• The factors which the competent court may take into consideration include the clarity and precision of the rule breached, the measure of discretion left by that rule to the national or Community authorities, whether the infringement and the damage caused was intentional or involuntary, whether any error of law was excusable or inexcusable, the fact that the position taken by a Community institution may have contributed towards the omission, and the adoption or retention of national measures or practices contrary to Community law.

• On any view, a breach of Community law will clearly be sufficiently serious if it has persisted despite a judgment finding the infringement in question to be established, or a preliminary ruling or settled case-law of the Court on the matter from which it is clear that the conduct in question constituted an infringement.’⁽⁸⁶⁾

137. In Metallgesellschaft, the Court, as I observed above, did not consider this matter, nor was the question raised by the national court in that case. Advocate General Fennelly, who as I have noted was of the opinion that the plaintiffs remedy in that case was restitutionary in nature, none the less made some comments in the alternative on the question of whether the Brasserie du Pêcheur conditions were satisfied. He remarked that, ‘The issue is whether the clarity and precision of Article [43] of the EC Treaty were such that the breach may be regarded as sufficiently serious. This has to be viewed in the light of the widespread use of residence as a criterion for direct taxation purposes coupled with the state of development of the relevant case-law at the material time. This will concern the limits which affect the use by Member States of that criterion where it is detrimental to the interests of residents from other Member States. In short, was the refusal to allow the group income election, viewed objectively, excusable or inexcusable?’⁽⁸⁷⁾ He went on to opine that, as what was in issue was indirect discrimination, this,‘should, in general, be considered sufficiently serious … To classify a breach of Article 52 of the Treaty such as that involved in the present case as excusable, the national court must be satisfied not only that the United Kingdom authorities genuinely believed that refusing to extend the benefit of the group exemption in question to groups whose parent company was non-resident was strictly necessary, but also, viewed objectively in the light of Bachmann [⁽⁸⁸⁾ ] and the principle of strict interpretation of exceptions to fundamental Treaty rules like the freedom of establishment, that this belief was reasonable’.⁽⁸⁹⁾

138. I agree with Advocate General Fennelly that the crucial question in deciding whether a breach such as the UK's in that case is sufficiently serious is the question whether the error of law was, viewed objectively, excusable or inexcusable. I would also agree that, in most areas of Community law, indirect discrimination is likely to satisfy this test. However, as I have observed in my Opinion in Test Claimants in the ACT Group Litigation,⁽⁹⁰⁾ certain of the Court's case-law setting out the boundaries of the application of the Treaty free movement provisions in the field of direct taxation is extremely complex and, in parts, in the process of development. For example, it was not in my opinion wholly clear until the recent Verkooijen⁽⁹¹⁾ and Manninen⁽⁹²⁾ judgments that Member States acting in a home State capacity are obliged by Articles 43 and 56 EC to grant equivalent relief from double economic taxation to resident shareholders receiving foreign-source income as for resident shareholders receiving domestic-source income. Such areas may be contrasted, however, with obligations that clearly follow from secondary legislation such as the Parent-Subsidiary Directive, or that follow clearly from jurisprudence of the Court which existed at the time that the relevant measures were in force. In sum, it follows in my view that breaches occurring at what were at that time the boundaries of the development of the Court's case-law in this field should not be considered as a manifest and grave disregard of the limits or a Member State's discretion within the meaning of the Court's case-law. It is for the national court to make the final assessment of this issue on the facts of the present case.⁽⁹³⁾

139. The response to Questions 6 to 9 should therefore in my view be that, in the absence of Community rules on the recovery of taxes unduly paid, it is for the domestic legal system of each Member State to designate the courts and tribunals having jurisdiction and to lay down the detailed procedural rules governing actions for safeguarding rights which taxpayers derive from Community law, including the characterisation of actions brought by the claimants before the national court. However, in exercising such jurisdiction, national courts are obliged to ensure that the claimants have an effective legal remedy in order to obtain reimbursement or reparation of the financial loss which they have sustained as a direct result of tax levied in breach of Community law.