Franked Investment Income Group Litigation, The Test Claimants In v Revenue And Customs

GENERAL INTRODUCTION

The present appeal and cross-appeal (with two associated applications for permission to appeal) are the most recent stage in the long-running Franked Investment Income (“FII”) group litigation. The litigation arises out of the way in which, under the regime in force until 5 April 1999, advance corporation tax (“ACT”) and corporation tax under Schedule D Case V were charged on dividends received by UK-resident companies from non-resident subsidiaries. The original test Claimants in the litigation are all UK-resident companies in the British American Tobacco (“BAT”) group; but they have since been joined, in respect of specific issues, by companies in the Ford and GKN groups. We will refer to them simply as “the Claimants”. The Respondents are the Commissioners for Her Majesty’s Revenue and Customs (“HMRC”).

The Claimants’ case in the litigation is, in bare outline:

All those points have been authoritatively decided in the Claimants’ favour at the previous stages of this litigation. The result is that they are in principle entitled to repayment and associated relief in relation to payments made by them as far back as the commencement of the ACT regime in 1973. The issues before us are concerned partly with certain specific defences which HMRC wish to invoke in relation to all or part of the claims and partly with their quantification.

The procedural history is complicated and has already involved two decisions of this Court, a decision of the Supreme Court, and no fewer than three decisions of the CJEU. We give a bare summary below, but the story is further complicated by the existence of other litigation – the so-called Portfolio Dividends group litigation and a claim brought by the Littlewoods group of companies – which raises overlapping issues. In all three cases the assigned judge has most recently been Henderson J, and it is against a judgment of his that this appeal lies.

The amounts at stake in the litigation are very large. As appears below, the outcome of the test claims is that HMRC have been ordered to pay the BAT Claimants over £1.184 billion. It is not yet known what amounts will be payable, subject to this appeal, in the other FII cases; but at an earlier stage in the litigation Henderson J recorded that the total amount potentially payable was some £5 billion.

HMRC were represented by Mr David Ewart QC, Mr Rupert Baldry QC, Mr Andrew Burrows QC and Ms Barbara Belgrano. The Claimants were represented by Mr Graham Aaronson QC, Mr Tom Beazley QC, and Mr Jonathan Bremner.

THE PROCEDURAL HISTORY

The Test Claims

The procedural history has been elaborately explained in earlier decisions. For present purposes all we need do is set out the bare minimum necessary to explain how the issues which we have to decide arise.

BAT commenced proceedings on 18 June 2003. A group litigation order was made for the FII cases on 8 October 2003.

On 11 October 2004 Park J made an order for the reference of certain questions raised by the FII litigation to the CJEU (“the first CJEU reference”). The Grand Chamber gave its judgment on 12 December 2006 (case C-446/04), [2006] ECR I-11753, (“FII CJEU1”). It followed from its judgment that the ACT regime did indeed contravene EU law in some respects and that the Claimants were entitled to a remedy in accordance with the principles established in the San Giorgio decision. But the judgment left a number of points affecting both liability and remedy for resolution by the domestic court.

On 5 July 2007 Rimer J directed a trial of “all … issues raised by the test claims, including liability for restitution, save in so far as those issues concern causation or quantification of the Claimants’ claims”, on the basis that the excluded issues would be determined so far as necessary at a second hearing. That split has been labelled by way of shorthand as being between “liability” and “quantification”; but that is not quite accurate in as much as the “liability trial” would consider issues of principle affecting remedy.

The liability trial took place before Henderson J in July 2008. His judgment was handed down on 27 November 2008 – [2008] EWHC 2893 (Ch), [2009] STC 254 (“FII HC1”). We need not attempt to summarise his decision here. Broadly speaking, he decided most, but not all, of the various issues in favour of the Claimants. On two issues he held that a further reference to the CJEU was necessary.

HMRC appealed to the Court of Appeal. The Claimants cross-appealed on the issues on which it had lost. The appeal was heard by Arden, Stanley Burnton and Etherton LJJ in October 2009. 23 issues were identified – ten relating to liability and thirteen to remedy. The Court’s judgment was handed down on 23 February 2010 – [2010] EWCA Civ 103, [2010] STC 1251 (Footnote: ²) (“FII CA1”). Again, it is unnecessary to attempt a summary of its decision. It dismissed most of the challenges to Henderson J’s decision, but it allowed HMRC’s appeal on four points. In particular, it held that the Claimants could not advance a mistake-based claim in addition to a Woolwich claim: this had the result that the ordinary six-year limitation period would apply. On one further point relating to liability it held that a reference to the CJEU was required.

Both parties appealed to the Supreme Court. On 8 November 2010 the Court refused permission to appeal against the Court of Appeal’s decision as regards the reference to the CJEU, and it identified a further liability issue on which a reference was required. It extended time for appealing as regards the remainder of the liability issues until after the Court of Appeal’s decision following the reference. The Claimants were given permission to appeal on some of the remedy issues. Those issues included an issue as to the effect of legislation – specifically, section 320 of the Finance Act 2004 and section 107 of the Finance Act 2007 – which was designed to limit or preclude recovery on the basis of mistake in claims such as these.

On 20 December 2010 a reference was made by Henderson J to the CJEU (“the second CJEU reference”) incorporating the issues identified by Henderson J and the further issues identified by the Court of Appeal and the Supreme Court.

The appeal to the Supreme Court on the remedy issues was heard in February 2012. Judgment was handed down on 23 May 2012 – [2012] UKSC 19, [2012] 2 AC 337 (“FII SC”). The Court allowed the appeal in part but held that a further reference to the CJEU was required on, in effect, the question whether the cut-off provisions in section 320 Finance Act 2004 were compatible with EU law. That reference (“the third CJEU reference”) was made on 25 July 2012.

As regards the second CJEU reference (case C-35/11), the Opinion of Advocate General Jääskinen was delivered on 19 July 2012. The decision of the Grand Chamber was given on 13 November 2012 – [2013] Ch 431 (“FII CJEU2”). It resolved all but one of the referred questions in the Claimants’ favour.

As for the third CJEU reference (Case C-362/12), the Opinion of Advocate General Wathelet was delivered on 5 September 2013. The decision of the Court was given on 12 December 2013 – [2014] STC 638 (“FII CJEU3”). Its effect was that HMRC were not entitled to rely on the statutory cut-off provisions to defeat the Claimants’ mistake-based claims. A further hearing in the Supreme Court was not required; a formal order allowing the Claimants’ appeal was made on 16 April 2014.

In the meantime Henderson J gave directions for the trial of the remaining issues in the summer of 2014, in the (correct) anticipation that the decision of the CJEU in the third reference would by then be available. HMRC applied for permission to re-amend their defence to raise a new point relating to the special regime for foreign income dividends introduced with effect from 1 July 1994 (the “FID regime”). Henderson J refused permission ([2013] EWHC 3757 (Ch)). HMRC appealed. On 27 March 2014 the Court of Appeal (Moore-Bick, McFarlane and Gloster LJJ) upheld Henderson J’s decision, giving its reasons in a judgment dated 2 September 2014 – [2014] EWCA Civ 1214 (“FII CA2”).

The trial of the remaining issues took place before Henderson J over 17 days in May and June 2014. 29 issues were identified for resolution, several of them involving sub-issues. Judgment was handed down on 18 December 2014 – [2014] EWHC 4302 (Ch), [2015] STC 1471 (“FII HC2”). We will not attempt to summarise Henderson J’s decision at this stage. He decided all the issues raised (though issues 13, 26(b) and 27 were agreed) save for issue 29, relating to detailed quantification, which was adjourned and subsequently agreed. A formal order embodying his conclusions was made on 30 January 2015.

It is against that order that the present appeal and cross-appeal are brought. Permission was sought by one or other party, or sometimes both, in relation to his decisions on all but three of the issues which had remained in dispute (being issues 3(c), 14 and 16). Henderson J himself gave permission in respect of all save issue 10, which concerned ACT paid on foreign income dividends under the FID regime: he refused permission in relation to that issue on the basis that it had already been determined by this Court in FII CA1. Patten LJ refused permission on the papers on the same basis. The application was renewed and was directed to be heard at the hearing of the appeal on the other issues.

The consequences of Henderson J’s decision on the various issues, in terms of the sums payable, required detailed computation. The computation was agreed between the parties and resulted in orders for the payment by HMRC to the various BAT Claimants of sums totalling £1,184,074,056.02.

The Other FII claims: the Evonik Degussa application

Following the outcome of the test claims, other claimants in the litigation have made applications for summary judgment and/or interim payments. One order has been made by consent.

One such application was made by a number of claimants, of which the first-named was Evonik Degussa Holdings Ltd., in respect of ACT paid by them under the FID regime. By a judgment handed down on 22 January 2016 ([2016] EWHC 86 (Ch)) Henderson J held that summary judgment should be given in relation to those payments (save in two immaterial respects): the total amount was approximately £160 million. HMRC sought permission to appeal. Vos LJ adjourned the application to be heard with the present appeal, with the substantive appeal to follow if permission were granted.

Portfolio Dividends and Littlewoods

We have already mentioned the Portfolio Dividends and Littlewoods cases. Portfolio Dividends is concerned with the treatment for the purposes ACT and corporation tax of dividends paid on shares in foreign companies held as investments, and Littlewoods with the overpayment of VAT; and both raise some of the same issues as the FII litigation. At this stage all that we need do is to identify the core decisions in each case.

Portfolio Dividends. The lead claimant is Prudential Assurance Co. Ltd. Henderson J gave the principal judgment on liability on 24 October 2013 – Prudential Assurance Co. Ltd v Commissioners for Her Majesty’s Revenue and Customs [2013] EWHC 3249 (Ch), [2014] STC 1236 (“Portfolio Dividends HC1”). He gave a consequential judgment on relief on 26 January 2015 – [2015] EWHC 118 (Ch), [2015] STC 1119 (“Portfolio Dividends HC2”). An appeal against both decisions was decided by this Court (Lewison, Christopher Clarke and Sales LJJ) on 19 April 2016 – [2016] EWCA Civ 376 (“Portfolio Dividends CA”).

Littlewoods. Most of the liability issues were decided by Vos J in a judgment handed down on 19 May 2010 – [2010] EWHC 1071 (Ch), [2010] STC 2072 (“Littlewoods HC1”). Following a reference to the CJEU (Littlewoods Retail Ltd v Revenue and Customs Commissioners (Case C-591/10) [2012] STC 1714 – “Littlewoods CJEU”), the remaining liability issues and the quantum issues were decided by Henderson J in a judgment handed down on 28 March 2014 – [2014] EWHC 868 (Ch), [2014] STC 1761 (“Littlewoods HC2”). Appeals against both decisions were decided by this Court (Arden, Patten and Floyd LJJ) on 21 May 2015 – [2015] EWCA Civ 515, [2016] Ch 373 (“Littlewoods CA”). Permission to appeal to the Supreme Court has been granted: the appeal is listed for next summer.

It will be observed that neither of the decisions of this Court in Portfolio Dividends and Littlewoods was available at the time of Henderson J’s decision in this case; but, as will appear, they effectively dispose of some of the issues in this appeal.

THE SHAPE OF THIS JUDGMENT

For ease of cross-reference we adopt the numbering of the issues before Henderson J, though this means that there are a few missing numbers where the issue in question is not live before us: a list of those issues is attached as an annex.

In broad terms, issues 1-4 before the judge concerned the calculation of unlawfully levied Schedule D Case V tax (“Case V tax”), and issues 5-13 concerned the calculation of unlawfully levied ACT. We refer to these issues together as the “taxation issues”. They are dealt with in Part II of this judgment.

Issues 17-29 concerned the extent of the recoverability of the unlawfully levied tax and were dealt with by Henderson J under the general heading “Remedies”. We deal with them in Part III of this judgment, under the general heading “Remedies”, save that we deal with issue 28, which concerns the “discoverability date” of the claims for limitation purposes, separately as Part IV, and that issue 29, which embraced the detailed quantification of the claims once the other issues were resolved, is not before us.

The judge grouped issues 14-16 together under the compendious heading “Other Issues of Principle”, but only issue 15 is live before us, and since the particular question of set-off which it raises is more closely related to the remedies issues, we deal with it also under Part III.

Although all members of the Court have contributed to each part of the judgment, Underhill LJ has been primarily responsible for drafting Parts I and IV, David Richards LJ for Part II and Vos LJ for Part III.

THE CORPORATION TAX REGIME

We start with a brief overview of the material features of the regime for corporation tax and ACT in force between 1973 and 1999. There have been a number of summaries in the many judgments given in the present and related litigation. In particular, reference may be made to FII CJEU1 at [6]-[22], FII HC1 at [12]-[22], and Portfolio Dividends CA at [30]-[32].

What follows is taken largely from the skeleton argument of counsel for HMRC in the present appeal. We do not understand it to be either controversial or, as a summary of the essential material features, incomplete.

The claims arise out of the different ways in which the UK tax system treated domestic dividends and cross-border dividends. For the purposes of the present case the two central differences concern (a) the charge to tax on foreign dividends and (b) the application of the ACT regime on the onward distribution of such dividends.

Under section 208 of the Income and Corporation Taxes Act 1988 (“ICTA”), a UK resident company (“a UK company”) was exempt from corporation tax on dividends received from another UK company. However, when a UK company received dividends from a company resident outside the UK (“a foreign company”), it was liable to Case V corporation tax on those dividends.

The Case V tax charge could be reduced by the grant of a credit for any withholding tax charged on the dividend in the distributing company’s state of residence. In addition, where the UK company owned 10% or more of the voting power of the foreign company, the UK company was entitled to relief for the underlying foreign corporation tax on the distributed profits, up to the amount due in the United Kingdom by way of corporation tax on the income concerned.

During the period relevant to this appeal, the UK operated a partial “imputation” system under which a UK company paid corporation tax on its profits but part of the corporation tax was imputed to non-corporate shareholders in the event of the profits being distributed to them, through the operation of the ACT system.

The system operated as follows. Where a UK company made a “qualifying distribution” (which included a dividend) it was liable to pay ACT on the distribution: section 14 (1) ICTA. The sum of the distribution and the ACT was called a franked payment: section 238 (1) ICTA.

ACT paid by a company could in principle be set against the company’s corporation tax liability on its profits for the relevant accounting period. This was referred to as “mainstream corporation tax” (“MCT”). ACT became “surplus” to the extent that a company’s MCT liability was insufficient to allow set-off. Surplus ACT could be carried forward or back by the company and could be surrendered to a company’s UK subsidiaries.

Individual shareholders were liable to income tax on dividends and other distributions received (under Schedule F, section 20 ICTA 1988). A UK-resident individual in receipt of a qualifying distribution from a UK company was entitled to a tax credit equal to such proportion of the amount or value of the distribution as corresponded to the rate of ACT: section 231 (1) ICTA. Income tax was chargeable on the total of the distribution and the tax credit: section 20 (2) ICTA. The tax credit extinguished all or part of the taxpayer’s liability. Lower-rate taxpayers and non-taxpayers, such as taxpayers whose income did not exceed the personal allowance, could recover some or all of the tax credit.

A UK company receiving a qualifying distribution from another UK company was also entitled to a tax credit: section 231 (1) ICTA. The total of the distribution and the tax credit was called “franked investment income” (“FII”): section 238 (1) ICTA. Where a UK company received FII, it was liable to pay ACT in relation to its own dividends only to the extent that those dividends and the ACT referable to them (i.e. its franked payments) exceeded the FII: section 241 ICTA.

Companies within a group could elect to pay dividends up the corporate chain without ACT, by making a “group income election”: section 247 ICTA. It was not possible, however, to make such an election on the payment of dividends to the ultimate individual shareholders. Thus, ACT would always be paid, at some stage, prior to the distribution of profits from a corporate group.

By contrast, a UK company receiving a distribution from a foreign company was not entitled to a tax credit, and the income did not qualify as FII. Such a company, therefore, was required to pay ACT on paying a dividend, unless it made a group income election, but in that case ACT would be paid by a company higher up the chain when it paid a dividend.

THE CASE V TAX ISSUES

Dividends paid by a foreign company to a UK company subject to corporation tax in the hands of the UK company under Case V.

Issue 1 is concerned with the extent to which a Case V charge is unlawful under EU law where a foreign company pays a dividend (or other “qualifying distribution”) to its UK holding company. The debate between the parties is whether, as the Claimants submit and as Henderson J held, domestic law must give credit for the higher of (i) the actual tax paid by the foreign subsidiary in its own country on the profits distributed by way of the dividend (the effective rate) and (ii) the nominal rate of tax applicable on those profits in its own country (the “foreign nominal rate” or “FNR”) or, as HMRC submitted below and in this court, only at the foreign nominal rate. It is common ground that in either case the credit cannot exceed the nominal rate of corporation tax in the UK.

It should be noted that the impact of this issue as regards Case V tax is limited, but it has a very substantial impact as regards ACT. This is because the UK system provided tax credits against Case V tax that were normally equal to the tax paid by the foreign company on the distributed profits in its own country. Tax credits were provided either in accordance with bilateral double tax treaties to which effect was given under section 788 ICTA or in accordance with section 790 ICTA. The benefit of such credits was taken at the time when the Case V tax otherwise became due for payment. By contrast, ACT was payable at an earlier time and might be irrecoverable because no Case V tax, or insufficient Case V, became payable.

The issue arises because dividends received by a UK company from a UK subsidiary were exempt from corporation tax in the hands of the receiving company, whereas dividends received from a foreign subsidiary attracted a credit, whereby the UK parent company would generally receive a credit for the corporation tax paid by the foreign subsidiary on the distributed profits in its own country. In a case where the tax paid by the foreign subsidiary was less than the nominal rate of UK corporation tax, this discriminated between dividends paid by a UK subsidiary and dividends paid by the foreign subsidiary. This was held to contravene articles 49 and 63 in FII CJEU1.

In FII CJEU1, the court explained why such discriminatory treatment contravened the freedom of establishment and the free movement of capital. The freedom of establishment entails for companies the right to carry on business in another member state through a subsidiary (see [39]) and thus prevents a member state “from treating foreign-sourced dividends less favourably than nationally-sourced dividends, unless such a difference in treatment concerns situations which are not objectively comparable or is justified by overriding reasons in the general interest” ([46]). As regards the free movement of capital, the court explained at [64]:

“Such a difference in treatment has the effect of discouraging United Kingdom-resident companies from investing their capital in companies established in another member state. In addition, it also has a restrictive effect as regards companies established in other member states in that it constitutes an obstacle to their raising of capital in the United Kingdom. In so far as income arising from foreign-sourced capital is treated less favourably from a tax point of view than dividends paid by companies established in the United Kingdom, shares in companies established in other member states are less attractive to United Kingdom-resident investors than those of companies having their seat in that member state.”

In FII CJEU1, the court held that a member state was entitled to deal with the problem of economic double taxation of profits through different systems applicable to domestic and foreign dividends, provided there was equivalent treatment of those dividends. It was therefore permissible for the UK to exempt domestic dividends from tax in the hands of the receiving company and to apply an imputation system to foreign dividends, provided that “[i] the tax rate applied to foreign-sourced dividends is not higher than the rate applied to nationally-sourced dividends and [ii] that the tax credit is at least equal to the amount paid in the member state of the company making the distribution”, subject to a limit of the nominal rate of tax charged in the UK (FII CJEU1 at [57] and repeated at [73]). The purpose of the limit is to prevent the UK being required to give credit for more tax or a higher rate of tax than would be chargeable on UK profits.

As a general proposition, therefore, articles 49 and 63 require the treatment of domestic and foreign dividends to be equivalent and preclude a more favourable treatment of domestic dividends. Member states are nonetheless free to apply different systems to the differently-sourced dividends provided they produce equivalent treatment.

There could be no serious room for argument that the effect of the judgment in FII CJEU1 was that the UK corporation tax regime contravened EU law to the extent that the tax credit provided against Case V tax on a dividend paid by a foreign company was less than the amount of tax paid on the distributed profits by the foreign company. The question is whether that conclusion was in effect modified by the court in FII CJEU2.

The genesis of the second reference decided in FII CJEU2 is found in FII CJEU1 at [54]-[56]:

“54.

The Claimants none the less point out that when, under the relevant United Kingdom legislation, a nationally-sourced dividend is paid, it is exempt from corporation tax in the hands of the company receiving it, irrespective of the tax paid by the company making the distribution, that is to say, it is also exempt when, by reason of the reliefs available to it, the latter has no liability to tax or pays corporation tax at a rate lower than that which normally applies in the United Kingdom.

55.

That point is not contested by the United Kingdom Government, which argues, however, that the application to the company making the distribution and to the company receiving it of different levels of taxation occurs only in highly exceptional circumstances, which do not arise in the main proceedings.

56.

In that respect, it is for the national court to determine whether the tax rates are indeed the same and whether different levels of taxation occur only in certain cases by reason of a change to the tax base as a result of certain exceptional reliefs.”

When the case returned to the High Court, there was agreement between the parties that in many cases, and indeed in the great majority of cases, the effective rate of tax of UK companies was significantly less than the nominal rate and in some cases the subsidiary would not be chargeable to tax at all. This results in particular from the different treatment of deductions from profits, for example capital expenditure, for the purposes of determining distributable profits and taxable profits; nothing turns on the detail of the different treatment.

The effect was that in a case where the effective rate of tax on the profits of a company was, say, 10% and the UK nominal rate of corporation tax was, say, 25%, there would be a marked difference of treatment in the hands of the holding company of a dividend paid by the company depending on whether it was resident in the UK or elsewhere. If it was a UK company, the dividend would be exempt from tax in the hands of the holding company, whereas if it was a foreign company the holding company would be entitled to credit for the tax actually paid by the foreign company on the distributed profits but would have to pay corporation tax at a rate equal to the difference (15%). The second reference to the CJEU in effect asked whether this difference constituted a breach of the rights under articles 49 and 63: see FII CJEU2 at [36].

In FII CJEU2, the court repeated at [38] that those articles required “a member state which has a system for preventing economic double taxation as regards dividends paid to residents by resident companies to accord equivalent treatment to dividends paid to residents by non-resident companies”. It repeated at [39] that the exemption and imputation systems were equivalent provided that foreign-sourced dividends were not taxed at a higher rate than nationally-sourced dividends and “the tax credit [in respect of the foreign-sourced dividend] is at least equal to the amount paid in the state of the company making the distribution, up to the limit of the tax charged in the member state of the company receiving the dividends”. The adoption of different systems must not “entail discrimination prohibited by the FEU treaty” ([40]).

At [46], the court held that the adoption of the exemption system for nationally-sourced dividends and an imputation system for foreign-sourced dividends which gives credit only for the tax actually paid “cease to be equivalent if the profits of the resident company which pays dividends are subject in the member state of residence to an effective level of taxation lower than the nominal rate of tax which is applicable there”. This leads to an additional tax liability on the receiving company: [47]. Another way of looking at it was set out at [48]:

“Unlike the exemption method, the imputation method therefore does not enable the benefit of the corporation tax reductions granted at an earlier stage to the company paying dividends to be passed on to the corporate shareholder.”

The fact that in the UK the effective level of tax on the profits of UK companies was lower than the nominal rate in the majority of cases had, the CJEU held, the consequence set out at [52]:

“It follows that application of the imputation method to foreign-sourced dividends as prescribed by the legislation at issue in the main proceedings does not ensure a tax treatment equivalent to that resulting from application of the exemption method to nationally-sourced dividends.”

This difference in treatment was not justified by any relevant difference in the situation of UK companies and foreign companies: [53]. It therefore constituted a breach of articles 49 and 63: [54].

The CJEU went on to consider whether the difference was justified by an overriding reason in the public interest and, if so, whether it was proportionate. The need for cohesion in the national tax system provided a sufficient justification for the difference ([56]-[59]), but it failed the test of proportionality:

“60.

As to the proportionality of the restriction, whilst application of the imputation method to foreign-sourced dividends and of the exemption method to nationally-sourced dividends may be justified in order to avoid economic double taxation of distributed profits, it is not, however, necessary, in order to maintain the cohesion of the tax system in question, that account be taken, on the one hand, of the effective level of taxation to which the distributed profits have been subject to calculate the tax advantage when applying the imputation method and, on the other, of only the nominal rate of tax chargeable on the distributed profits when applying the exemption method.”

This led the CJEU to the following analysis:

“61.

The tax exemption to which a resident company receiving nationally-sourced dividends is entitled is granted irrespective of the effective level of taxation to which the profits out of which the dividends have been paid were subject. That exemption, in so far as it is intended to avoid economic double taxation of distributed profits, is thus based on the assumption that those profits were taxed at the nominal rate of tax in the hands of the company paying dividends. It thus resembles grant of a tax credit calculated by reference to that nominal rate of tax.

62.

For the purpose of ensuring the cohesion of the tax system in question, national rules which took account in particular, also under the imputation method, of the nominal rate of tax to which the profits underlying the dividends paid have been subject would be appropriate for preventing the economic double taxation of the distributed profits and for ensuring the internal cohesion of the tax system while being less prejudicial to freedom of establishment and the free movement of capital.”

Having previously argued that EU law required only that credit be given for tax actually paid on the profits distributed by the foreign subsidiary, HMRC submitted to Henderson J on the basis of the decision in FII CJEU2 that EU law required only that credit be given for the foreign nominal rate of tax and that no credit need be given for tax actually paid, if that were higher.

Henderson J rejected HMRC’s case, for the reasons that he set out at [32]-[36]. As HMRC largely repeats its submissions on this appeal, it is worth citing those paragraphs in full:

“32.

I must now explain why I do not accept this submission. In the first place, I do not agree that a sharp distinction can, or should, be drawn between the initial stage of determining whether there is a restriction on freedom of establishment, and subsequent stages when issues of justification, cohesion and proportionality are considered. Although it is helpful for analytical purposes to sub-divide the question in this way, and the ECJ frequently does so, the single question which always has to be answered is whether there has been a breach of the relevant freedom. I would find it very surprising if a prerequisite for a compliant hybrid system which the Court has repeatedly identified at the first stage of the analysis were then to become completely irrelevant at the stage when justification is considered.

33.

Secondly, I agree with the Claimants that the reason why the Court's analysis in [FII CJEU1] stopped at the restriction stage is probably that the Court was satisfied, subject to confirmation by the national court of the question on tax rates remitted to it in paragraph 56 of the judgment, that the two conditions for a compliant hybrid system (as then stated by the Court) were indeed satisfied. There was no doubt that the UK provided a credit for underlying tax on dividends paid by foreign subsidiaries; and it is fairly clear that the Court expected its query on domestic tax rates to be answered in the affirmative (i.e. in the sense that the tax rate applied to foreign-sourced dividends was not higher than the rate applied to nationally-sourced dividends). It was therefore unnecessary for the Court to prolong its analysis, in what was anyway an exceptionally long and complicated judgment, to consider what the position would have been if a restriction were found to exist. That only became necessary in [FII CJEU2], in the light of the (probably) unexpected answer returned by the national court to the question which (as it then thought) had been remitted to it.

34.

Thirdly, I am satisfied that the reason for the almost exclusive focus on rates of tax in the Court's discussion of justification is not that the need to provide a credit for underlying tax had suddenly become irrelevant, but (again) that it was not in issue, because nobody disputed that the UK tax system did provide such a credit. It is clear, to my mind, that the Court had not lost sight of the additional requirement of a credit for underlying tax, not least because of the express reference to such tax in the answer to the first question in paragraph 65 of the judgment (and see too paragraph 71, in the context of the second question relating to ACT). If the Court had intended to hold that the only prerequisite for a compliant hybrid system was the grant of a credit at the FNR, it would surely have said so in terms, and also explained why its standard jurisprudence, to which reference was made in paragraph 39, was no longer applicable. I find it particularly implausible that such a radical restatement of principle was intended, in view of the fact that the judge rapporteur in [FII CJEU2] was Vice-President Lenaerts, who had also been the rapporteur in [FII CJEU1], and in the Portfolio Dividends case.

35.

Fourthly, Mr Ewart developed a submission that, in laying down a single requirement for the grant of a credit at the FNR, the ECJ was adopting a solution propounded by the European Commission in its written observations. In paragraph 31 of its observations, the Commission had said this:

"31.

In such circumstances there seem to the Commission to be two ways of ensuring equal treatment. One is to exempt both domestic and foreign dividends. That solution has the drawback, as outlined above, that it may permit excessively favourable treatment of foreign dividends where the tax rate in the source State is lower than in the United Kingdom. The other, which is wholly consistent with the Court's reasoning in [FII CJEU1], is to have regard solely to the nominal rate of tax in calculating the tax credit on foreign dividends."

Mr Ewart placed emphasis on the word "solely". The problem with that submission, however, is that no equivalent to "solely" can be found in the Court's discussion of nominal rates in paragraphs 60 to 65 of the judgment. On the contrary, the Court said in paragraph 62:

"For the purpose of ensuring the cohesion of the tax system in question, national rules which took account in particular, also under the imputation method, of the nominal rate of tax to which the profits underlying the dividends paid have been subject would be appropriate for preventing the economic double taxation of the distributed profits and for ensuring the internal cohesion of the tax system while being less prejudicial to freedom of establishment and the free movement of capital."

The words which I have italicised ("in particular") are a translation of "notamment" in the original French text. The meaning of "notamment" given in the European Communities Glossary (8^th edition, 1984), published by the Council of the EU, is "for example, in particular, including, inter alia". Thus the Court clearly did not intend to say "solely", even though it evidently derived considerable assistance from the Commission's submissions. The word which the Court chose ("notamment") was in my view apt to make the point that a credit at the FNR would normally suffice, but a credit for underlying tax also had to be given in case it was higher.”

At [36]-[39], Henderson J considered an example of what HMRC submitted was an anomalous outcome of accepting the Claimants’ case but rejected it as a ground for favouring HMRC’s submissions, because other anomalies could be instanced pointing in the opposite direction, and the issue could not be decided on the basis of competing anomalies.

In this court, the Claimants refine and repeat the submissions made below and submit that the appeal on this issue should be dismissed, first, because Henderson J was right for the reasons he gave and, secondly, because the same issue was determined against HMRC by the Court of Appeal in the recent decision in Portfolio Dividends CA.

HMRC submit that the effect of the decision in FII CJEU2 is that the charge to Case V tax on dividends paid by foreign subsidiaries would have complied with EU law if the UK had granted a credit for tax at the relevant foreign nominal rate on the gross amount of the dividend, subject to a limit of the UK nominal rate. This, they submit, is exactly what the CJEU decided in FII CJEU2.

HMRC’s analysis of the reasoning in FII CJEU1 and FII CJEU2 is as follows. First, the UK corporation tax system did not satisfy both the requirements set out in FII CJEU1 at [57]. It therefore prima facie contravened articles 49 and 63. Secondly, that conclusion was displaced because the need for cohesion in the tax system justified some restrictions on the freedom of establishment and the movement of capital. But, thirdly, the restrictions went further than was necessary to achieve that purpose and so were disproportionate. The UK provisions were therefore unlawful only to the extent that they went beyond what was proportionate. Fourthly, the CJEU (unusually) spelt out the modification required to make the system proportionate and therefore compliant with EU law in its judgment at [62], cited above.

HMRC submit that it follows that, in order to ensure compliance with EU law, credit did not need to be given for tax actually paid on the distributed profits if that was more than the amount produced by application of the foreign nominal rate to those profits.

We are unable to accept this submission. The CJEU correctly proceeded in both FII CJEU1 and FII CJEU2 on the basis that the UK corporation tax system complied, or very largely complied, with the second requirement set out in FII CJEU2 at [39], repeating its conclusion in FII CJEU1, that “the tax credit is at least equal to the amount of tax charged in the state of the company making the distribution”. This requirement was satisfied by the double taxation arrangements to which we have referred. In very large part, credit was given for the amount of tax paid on the distributed profits in the foreign company’s home state, whether that was less than, more than or the same as tax on those profits at the foreign nominal rate. In considering the problem posed by the difference between effective rates and nominal rates, the court in FII CJEU2 did not need to, and in our view did not, re-visit that requirement. It was concerned to examine only the need to comply with the other condition as a result of the finding by the English court that in a majority of cases in the UK the effective rate was less than the nominal rate. As Henderson J observed in the judgment under appeal at [33], this was probably not the finding that the CJEU had anticipated.

Like Henderson J, we consider it implausible that the CJEU intended to abandon one of the two key features of a compliant dual credit scheme identified by it in FII CJEU1 and expressly repeated in FII CJEU2, and to do so without any reference to it or explanation for it. The CJEU’s understanding of the task facing it in FII CJEU2 is made clear by it at [2]: the reference “is designed to obtain clarification regarding various paragraphs” of its judgment in FII CJEU1. See also the judgment at [21]-[35]. It was not to re-visit and radically re-formulate the fundamental features of its decision as set out in that judgment. This appears to be clear from the clarification provided by the court in FII CJEU2 at [65]:

“65. In light of the foregoing, the answer to the first question is that articles 49FEU and 63FEU must be interpreted as precluding legislation of a member state which applies the exemption method to nationally-sourced dividends and the imputation method to foreign-sourced dividends if it is established, first, that the tax credit to which the company receiving the dividends is entitled under the imputation method is equivalent to the amount of tax actually paid on the profits underlying the distributed dividends and, second, that the effective level of taxation of company profits in the member state concerned is generally lower than the prescribed nominal rate of tax.”

It is difficult to imagine a clearer statement that, in order to comply with EU law, the tax credit to which the receiving company is entitled under the imputation system must be “equivalent to the amount of tax actually paid on the profits underlying the distributed dividends”. Exactly the same point is made at [80], and again at [85]:

“It is apparent from the answer to the second question that national rules, such as those at issue in the main proceedings, which seek to prevent the economic double taxation of distributed profits are incompatible with European Union law in so far as those rules, in the context of a group taxation scheme, do not take account, as regards dividends from other states, of the corporation tax already paid on the profits out of which those dividends have been paid.”

HMRC place reliance on what the court said at [87] about the claim that a UK company compelled to pay ACT on profits from foreign-sourced dividends can bring:

“The answer to the third question therefore is that European Union law must be interpreted as meaning that a parent company resident in a member state, which in the context of a group taxation scheme, such as the group income election at issue in the main proceedings, has, in breach of the rules of European Union law, been compelled to pay ACT on the part of the profits from foreign-sourced dividends, may bring an action for repayment of that unduly levied tax in so far as it exceeds the additional corporation tax which the member state in question was entitled to levy in order to make up for the lower nominal rate of tax to which the profits underlying the foreign-sourced dividends were subject compared with the nominal rate of tax applicable to the profits of the resident parent company.”

The statement is the consequence of the conclusion reached by the court on the issues before it in FII CJEU2, but it is not addressed to a case where the company has been denied a credit for tax actually paid in excess of the foreign nominal rate.

We do not accept Mr Ewart’s submission that to compare the actual tax paid and the nominal rate of tax is to compare apples and pears. It is precisely what the CJEU did in FII CJEU1 and it is precisely what the UK submitted should be done in FII CJEU2 in preference to giving credit for the foreign nominal rate and moreover what HMRC submitted should be done in Portfolio Dividends HC and CA.

Mr Ewart relied before us, as he had before the judge, on the example set out in the judgment below at [36]. Mr Ewart is right that in such a case the UK company could in effect obtain a double advantage. However, Mr Ewart accepted that there were other realistic examples where the effective rate of tax would be greater than the foreign nominal rate without any scope for double counting in favour of the taxpayer. Whatever the answer to issue 1, it is capable of producing unfair results to one side or the other and we do not consider that it can be answered by weighing those various results.

In our view, the issue is to be answered by reference to the basic underlying principles. If a foreign company pays a dividend and it has paid tax on its profits in its home state at a higher rate than the nominal rate in that state, the principle of equivalence under EU law requires that credit must be given for that higher rate on receipt of the dividend by a UK company, in circumstances where the receipt of a dividend from a UK company would be exempt from tax. The fact that, in the case of non-portfolio dividends, the UK company was generally entitled to a credit against its charge to Case V tax on the dividend at a rate equal to the rate paid by the foreign company on its profits does not displace the requirement but shows how, in the case of Case V tax, the UK generally met it.

We accordingly agree with the judge’s decision on issue 1. We would reach this conclusion even without the decision of this court in Portfolio Dividends CA.

Issue 1 arose for decision in Portfolio Dividends CA in the context of what were called “portfolio dividends”. These are dividends or other qualifying distributions paid by a foreign company to a UK corporate shareholder which holds less than 10% of the voting power in the foreign company. In FII CJEU1, it was held that article 49 was contravened by the discriminatory treatment of portfolio dividends paid by a foreign company to a UK company as against those paid by a UK company, in exactly the same way as non-portfolio dividends: see [61] to [71].

Under the heading “Nominal rate or effective rate?” this court considered at [45] to [80] of the judgment in Portfolio Dividends CA the issue of the amount of the tax credit to which a UK company is entitled against the charge to corporation tax on foreign dividends under Case V. The court identified the issue in [47]:

“The dispute is whether EU law requires, as the judge held, a tax credit for the higher of tax actually paid and the foreign nominal rate of tax of the dividend paying company capped at the UK corporation tax rate, or, as HMRC claim, a credit for the actual tax rate paid capped at the UK corporation tax rate.”

It can therefore be seen that, in contrast to their position in the present case, HMRC were submitting that credit need not be given at the foreign nominal rate but only for the actual tax paid. In his judgment at first instance in that case, as in the present case, Henderson J concluded from his analysis of FII CJEU1 and FII CJEU2 that credit for tax actually paid and credit at the foreign nominal rate were to be treated as alternatives “with credit to be granted for whichever was the higher up to the limit of the Case V charge reduced by withholding tax”: [62] of this court’s judgment.

At [65] this court held Henderson J’s analysis of FII CJEU2 to be well-founded. It rejected HMRC’s contention that the credit was to be restricted to the tax actually paid as inconsistent with the decision and reasoning in FII CJEU2. At [76] it reiterated that whether the credit should be at the nominal rate or the actual rate does not depend on whether the dividends are portfolio or non-portfolio dividends, and that the reasoning of the CJEU applied equally to all types of dividend. It observed that “the critical question is whether there is equivalence between the treatment of foreign and UK-sourced dividends”.

Turning to the issue as stated at [47], the court said at [78]:

“There remains for consideration whether or not the judge was right to hold that the credit should be at the higher of the actual or nominal rate (up to the limit of the Case V tax charge after a credit for withholding tax: see Declaration 1 C). This issue is unlikely to arise in practice although there can be some rare cases in which the actual rate will exceed the nominal. In our view the judge was right to take the higher of the two rates since the foundation of the jurisprudence is that there should be a credit for at least the tax actually paid. The controversy over whether, if the nominal rate is higher than the actual, there should be credit at the higher rate does not, in our view, mean that less than the actual rate is ever appropriate.”

HMRC submitted that we should not consider ourselves bound by the decision of this court in Portfolio Dividends CA for two reasons.

First, it was submitted that the conclusion that the UK taxpayer was entitled to a credit of the higher of the tax actually paid and tax at the foreign nominal rate was obiter. In Portfolio Dividends CA, Prudential and other companies receiving portfolio dividends accepted that they could not prove the amount of tax actually paid on the distributed profits received by them, and therefore the real issue was whether the credit should be equal to the tax actually paid or the foreign nominal rate, not whether it should be the higher of them. This is not a submission open to HMRC, as Mr Ewart was ultimately inclined to accept. Not only did the court express the dispute in [47] as being a choice between (i) the higher of the tax actually paid and the foreign nominal rate of tax and (ii) the tax actually paid, but that was the express effect of the order made by Henderson J and it was the issue stated for this court to decide in the agreed list of issues. The relevant declaration made by this court in its order of 19 April 2016 is:

“The effect of the rulings of the CJEU is that the foreign dividend should be afforded equivalent treatment, taking the form of the imputation method according to which credit should be given for the relevant foreign tax at the effective rate (i.e. the actual tax paid) or the nominal rate (whichever is the higher), subject to a cap at the rate of the UK’s appropriate nominal rate (i.e. the corporation tax rate or the rate of ACT as applicable).”

The issue is stated in these terms in HMRC’s grounds of appeal to the Supreme Court and, for good measure, HMRC specifically invited this court to decide this issue in Portfolio Dividends CA, rather than leave it for decision on the present appeal.

Secondly, HMRC submitted that this court had misunderstood the effect of the decision and reasoning in FII CJEU2 and we were therefore not bound by its decision. We very much doubt that this is a course open to us and we are satisfied that in any event, in the light of so recent and careful a consideration of FII CJEU2, it would not be appropriate for us to do so. It is, however, unnecessary for us to consider those questions, because we are in entire agreement with the analysis of FII CJEU2 given by this court in Portfolio Dividends CA.

For these reasons, we consider that the Judge was correct in his conclusion on issue 1.

As Henderson J described in his judgment at [41], this issue arises because the number of cases in which profits distributed to a UK company by an EU subsidiary (the water’s edge company) were originally made and taxed in that subsidiary are comparatively small. In such a case, the FNR would clearly be that applicable to the water’s edge company.

In most cases, however, the water’s edge company will be a “mixer” company to which dividends are paid out of profits made and (in many cases) taxed lower down the corporate chain. Typically, multi-national groups with UK holding companies will establish mixer companies resident in countries, such as the Netherlands, where income derived from dividends is not subject to tax. So far as possible, their affairs will be arranged so that they receive dividends out of profits which have borne tax in different countries at rates which, when blended and appropriately weighted, equate to the tax payable in the UK on the dividends paid by them. In this way, maximum advantage can legitimately be taken of the credit for tax actually paid, calculated in accordance with section 801 ICTA which is itself applied to dividends paid by mixer companies on the basis of blending and weighting the tax that has been paid overseas on the underlying profits.

Three possibilities were canvassed before Henderson J as to the appropriate approach to determining the FNR in the case of mixer companies. He listed them in [43]:

“Against this background, Issue 2 asks how to determine which FNR is appropriate in respect of each foreign dividend received by a UK company. The three possibilities canvassed in argument were:

(a)

the FNR of the foreign water's edge company paying the dividend to the UK (this being the Claimants' primary case);

(b)

the FNR of the jurisdictions where the income had been subject to tax, applying where necessary a weighted average of those FNRs following the system of s 801 of ICTA 1988 (the Claimants' alternative case); and

(c)

the FNR of each jurisdiction where the underlying profits had been subject to tax, so that where a dividend was derived from a variety of sources of profit carrying differing FNRs the dividend must be disaggregated into the parts which carried different FNRs (the Revenue's case).”

Although in Portfolio Dividends HC1 Henderson J had adopted the first of these solutions, it was a choice heavily influenced by the practical impossibility in nearly all cases of portfolio dividends of ascertaining the tax rates applicable to the underlying profits. This was not the case with non-portfolio dividends such as those in the present case. Henderson J said at [45]:

“In reaching that conclusion, however, I was heavily influenced by the practical impossibility in nearly all cases of pursuing an enquiry into FNRs beyond the water's edge company paying the dividend where the holdings in question were all of less than 10%, and the UK recipient would normally be in no position to trace the course which the distributed profits had previously followed. Those practical considerations do not apply where the dividends have throughout remained in a single multi-national group of which the UK recipient is a member. It is therefore not suggested by the Claimants that either the second or the third solutions would in practice be unworkable. Indeed, it is common ground that the necessary information can be retrieved without undue difficulty, because it had to be collected and submitted to the Revenue for the purposes of claiming double taxation relief for the actual underlying tax on the dividends. I do not, therefore, start with any predisposition to hold that solution (a) is the correct one in cases of the present type.”

The judge was satisfied that no clear guidance could be obtained from the court’s reasoning in FII CJEU2, and it has not been suggested to us that he was wrong in that view.

The judge concluded that the first possible solution should be rejected and that “in principle the appropriate FNR to apply is that of the jurisdiction in which the distributed profits were in fact subject to tax”, which he considered better accorded both with FII CJEU2 at [62], [72] and [86] and with the rationale for requiring a credit at the FNR at all.

He therefore rejected the Claimants’ primary case. Although the Claimants cross-appealed against this part of the judge’s order, it was not pursued before us. Mr Aaronson told us that it was a solution that would be pursued only if it was, on the facts of any particular case, the only way of giving practical effect to the CJEU’s ruling that credit should be given at the FNR. In such a case, he submitted, the principle of effectiveness would require its adoption.

As between solutions (b) and (c), Henderson J preferred solution (b), which was the Claimants’ alternative case.

At [49], the judge considered that the choice between these solutions “reflects a fundamental difference of approach between the two sides”:

“The problem of how to treat mixer companies, and analogous arrangements, is to my mind more difficult, and the choice between solutions (b) and (c) reflects a fundamental difference of approach between the two sides. In essence, the Claimants' approach is to take the UK system as it actually was, and to make only such modifications to it as are appropriate to accommodate the ECJ's ruling. The Revenue's approach, by contrast, is to start from first principles and then try to work out how far the UK could lawfully have imposed a Case V charge on the various streams of income comprised in the blended dividends paid by the mixer company.”

He explained at [50] that the Claimants’ methodology built on and adopted the machinery under section 801 ICTA, to which we have referred above. Section 801 is the machinery, in respect of dividends received by a UK company from an overseas related company, for giving credit against UK corporation tax on the dividends for any tax payable on the distributed profits, wherever down the corporate chain they were earned. It is applicable to profits earned in non-EU countries as well as in EU countries. By contrast, HMRC

“…start by separating out, or disaggregating, the component parts of the blended dividend, and then ask in respect of each such part whether (and, if so, to what extent) the UK could lawfully have levied a Case V charge on it. Only to the extent that a Case V charge could not lawfully have been levied on the separate income streams, they submit, is the charge to tax on the blended dividend to be regarded as unlawful.”

The judge explained his reasons for preferring the Claimants’ approach:

“54.

As to the difference in approach which is reflected in the example which I have discussed, I consider that the claimants' approach as set out in their alternative case is to be preferred. The question of the unlawfulness of the Case V charge does not arise in a legislative vacuum. It has to be considered in the context of the actual tax system operated by the UK, which was binding as a matter of domestic law and has to be applied by the English court subject only to any disapplication or conforming construction which may be needed in order to make it compliant with EU law. The introduction of a credit for tax at the FNR should therefore be implemented in a way which, as far as reasonably possible, reflects and goes with the grain of the existing UK legislative scheme. It seems to me that the claimants' approach respects this principle more closely than the Revenue's, because it adapts and builds on the existing machinery for giving credit for underlying tax. It is not an objection to this approach, in my judgment, that the grant of relief from juridical double taxation of cross-border dividends, of which the s 801 machinery forms part, is not itself required by EU law. The point is, rather, that the machinery formed an integral part of the UK's existing system for taxation of cross-border dividends which has to be made compliant with EU law.

55.

Another aspect of the same point, in relation to mixer companies, is that the claimants' approach does less violence to the actual facts than the Revenue's approach. The main (if not the only) purpose of such companies, before the introduction of the eligible unrelieved foreign tax ('EUFT') rules in 2001, was to blend income streams which carried different rates of creditable underlying tax in such a way that the higher rates were offset by the lower, and the weighted average rate was no higher than the UK nominal rate. This is what actually happened, and the onward dividends received by the UK water's edge company were blended ones to which the s 801 machinery applied. By contrast, the notional disaggregation of the blended income streams in accordance with the Revenue's methodology contradicts the actual blending which the mixer company was designed to achieve, and seeks to undo (in relation to nominal rates of tax) the legitimate tax planning which led to the actual blending of the dividends in the first place. The significance of this becomes apparent when it is recalled that, if the Revenue's argument on Issue 1 were correct, the only credit required by EU law would be a single credit at the FNR. The Revenue say that this credit should be calculated on a basis which, in effect, posits a lawful Case V charge even where such a charge could not in fact have been imposed because of the availability of double taxation relief for underlying tax. A reconstruction of this nature seems to me to go well beyond the proper bounds of conforming interpretation.”

HMRC submit that the judge reached the wrong conclusion and should have adopted solution (c).

HMRC point out that if a UK company had two direct EU subsidiaries, the relevant FNR applicable to dividends paid by them respectively would be the FNR applicable to the distributed profits of each of them. There would be no question of a blended rate and HMRC submit that it should make no difference if the two subsidiaries are held as direct subsidiaries of a water’s edge mixer company.

HMRC gave a simple example to illustrate this point. Assume that the rate of UK corporation tax is 30%. If EU subsidiary 1 is subject to corporation tax at 40% in its home state, the UK could not under EU law charge corporation tax on a dividend paid by it, because its FNR is higher than the UK rate. If subsidiary 2 is subject to corporation tax at 10% in its home state, the UK would be entitled to charge corporation tax at 20% on a dividend paid by it. If the two dividends are mixed and a blended rate of FNR applied, the UK is prevented from charging on the profits derived from subsidiary 2’s dividend the corporation tax that, as a matter of EU law, it is entitled to charge.

HMRC further point out that if a water’s edge company had two subsidiaries, one resident in an EU country and the other resident in a non-EU country, the dividends received from each would be disaggregated for the purpose of determining the appropriate FNR in order to comply with EU law as held in FII CJEU2. The relevant FNR would be that applicable to the EU subsidiary’s profits and the tax, and rate of tax, paid on the non-EU subsidiary’s profits would be disregarded for these purposes. The UK would remain free to charge the dividends received from the non-EU subsidiary to UK corporation tax without any credit. This important point was made clear by Henderson J at [56]:

“There is, however, a further refinement which I need to mention. It is only in respect of distributions of profits originating in the EU that the operation of the Case V charge needs to be modified in order to make it compliant with EU law. This follows from the fact that, in relation to distributions by subsidiaries, in contrast to portfolio dividends, the only fundamental freedom under EU law which was infringed by the Case V charge was freedom of establishment. Although the Case V charge was in principle capable of engaging art 56 EC (free movement of capital) in relation to distributions by third country subsidiaries, any infringement of it, at least prior to the introduction of the EUFT rules, was admittedly protected by the standstill provisions of art 57(1) EC: see FII (CA) at [72]. Accordingly, I believe it to be common ground, and (if not) it appears to me correct in principle, that any third-country source income which was paid up through an EU mixer company in whose hands it was exempt from tax should be disregarded in calculating the FNR credit required by EU law. There can be no objection in principle to the imposition of the unmodified Case V charge on distributions of profits which originated in a third country, and which have not been subjected to tax in the EU, even if they formed part of a blended dividend paid up by an EU mixer company.”

The answer to issue 2 depends on the correct approach to arriving at a conforming interpretation of UK law to ensure compliance with EU law. This is essentially a matter of domestic law, not EU law. All that EU law requires is that the result complies with EU law. As Sir Andrew Morritt C observed in Vodafone 2 v Revenue and Customs Commissioners [2009] EWCA Civ 446; [2010] Ch 77, at [34]:

“The jurisdiction of the ECJ to give preliminary rulings relates to the interpretation of the EC Treaty and the other matters referred to in Article 234 EC. They do not include the interpretation of the legislation of a member state, see Pfeiffer v Deutsches Rotes Kreuz [2004] ECR I-8835 para 115 and Criminal ProceHoerchstedings against Pupino [2005] ECR I-5285 para 47. Further, as those citations show, the obligation of the national court is to examine the whole of the national law to consider how far it may be applied so as to conform to enforceable Community rights.”

The basic principle is that, so far as possible, domestic legislation will be interpreted so as to conform with EU law. But, as the authorities concerning EU law and the analogous area of human rights establish, the court must examine the whole of the relevant legislation and is not confined to construing the existing words of the legislation.

An agreed statement of applicable principles was adopted by the Court of Appeal in Vodafone at [37], which emphasises that the court is not constrained by the conventional rules of statutory construction but is permitted not only to depart from the strict application of the words of the statute but to imply words necessary to comply with EU law. The precise form of words to be implied does not matter, although, as Sir Andrew Morritt C observed at [39], “it undoubtedly assists in the consideration of whether or not it is a permissible interpretation to see on paper how it is suggested that it would be effected, whether by interpolation, deletion, rewording or otherwise”.

What is required is that the interpretation arrived at should “go with the grain of the legislation” and be “compatible with the underlying thrust of the legislation being construed”, as it was put by Lord Nicholls in Ghaidan v Godin-Mendoza [2004] UKHL 30, [2004] 2 AC 557, at [33] in the context of compatibility with the European Convention on Human Rights. A substantial departure from the language used need not involve a departure from the fundamental or cardinal features of the legislation: see Revenue and Customs Commissioners v IDT Card Services Ireland Ltd [2006] EWCA Civ 29, [2006] STC 1252 at [89] per Arden LJ who added:

“It is possible to read the legislation up (expansively) or down (restrictively) or to read words in to the legislation.”

Nonetheless, the process of arriving, if possible, at a conforming interpretation of domestic legislation is essentially a process of construction, albeit “a highly muscular” one (FII SC at [176] per Lord Sumption). It is not a process of legislation. Section 2 of the European Communities Act 1972 requires legislation to “be construed and have effect subject to” community rights. The process of interpretation cannot create a wholly different scheme from any scheme provided by the legislation: Vodafone at [70] per Longmore LJ, citing Lord Rodger of Earlsferry in Ghaidan v Godin-Mendoza at [110].

In Vodafone itself, an authority on which HMRC heavily relied before us, the Court of Appeal implied an additional exception, by way of an extra sub-paragraph, into the relevant charging section in ICTA in order to render it compliant with EU law.

We see force in HMRC’s submission that the right approach is to identify the FNR at the level at which the distributed profits have borne tax, as indeed did the judge (see the judgment below at [47]), and we accept that there is nothing in the CJEU’s judgments requiring a blended rate to be applied to dividends received by a mixer company before distribution to a UK company.

If the disaggregation of the dividends received by a mixer company were impractical, that would be a powerful reason for a different approach but, as the parties are agreed and as the judge recorded, no such difficulties exist in the present or similar cases.

However, the very broad approach taken in the submissions of HMRC goes far wider than anything previously accepted as a conforming interpretation and is contrary to authority. HMRC do not propose an interpretation, however broad, of any provisions in the ICTA. They seek to introduce a scheme that is simply not in the legislation, whereas the Claimants seek to adapt an existing provision, that regulates tax credits to avoid double taxation, to give effect to the CJEU’s decisions. As Mr Ewart put it in his oral submissions, HMRC’s approach “is not an attempt to go into the interstices of particular provisions and mould them but is simply to say that there is a provision to be read in, which gives effect to Community law rights”.

The same approach was adopted by HMRC in Portfolio Dividends CA in the context of ACT and was rejected by this court:

“110.

Mr Ewart sought to suggest that this court's ruling at [107] in FII (CA) was not tied to interpretation of section 231(1) but produced a sort of power of amendment which roamed at large across all the ACT provisions, leaving it open to HMRC to propose a different methodology within the interstices of those provisions taken as a whole for giving effect to the requirements of EU law. In substance, he wished to introduce a distinct crediting methodology into section 241.

111.

In our view this is an unsustainable submission. It rests on a misconception regarding the operation of the Marleasing interpretive principle and of the ruling given in FII (CA). Application of that principle does not make it irrelevant which particular statutory provision in a group of provisions is being interpreted. On the contrary, it is a principle of interpretation which is applied to give a specifiable and specific meaning to a particular provision (or series of provisions, taken one by one), even if it allows considerable latitude as to the wording which may be read into the provision (or provisions). In considering whether a particular conforming interpretation can be given to a particular provision, the court has to check to see that that proposed interpretation does not go against "the grain" of the legislation in question or conflict with its cardinal features. This was the exercise performed by this court in FII (CA) to arrive at the particular conforming interpretation it identified for a specific section, namely section 231(1) of ICTA.”

We endorse the approach of Henderson J in the judgment under appeal at:

“[54] As to the difference in approach which is reflected in the example which I have discussed, I consider that the Claimants' approach as set out in their alternative case is to be preferred. The question of the unlawfulness of the Case V charge does not arise in a legislative vacuum. It has to be considered in the context of the actual tax system operated by the UK, which was binding as a matter of domestic law and has to be applied by the English court subject only to any disapplication or conforming construction which may be needed in order to make it compliant with EU law. The introduction of a credit for tax at the FNR should therefore be implemented in a way which, as far as reasonably possible, reflects and goes with the grain of the existing UK legislative scheme. It seems to me that the Claimants' approach respects this principle more closely than the Revenue's, because it adapts and builds on the existing machinery for giving credit for underlying tax. It is not an objection to this approach, in my judgment, that the grant of relief from juridical double taxation of cross-border dividends, of which the section 801 machinery forms part, is not itself required by EU law. The point is, rather, that the machinery formed an integral part of the UK's existing system for taxation of cross-border dividends which has to be made compliant with EU law.”

For completeness, we add that, with respect to the judge, we do not accept the force of the point advanced by the Claimants and accepted by him at [55], that HMRC’s approach contradicts the actual blending achieved by the mixer companies and seeks to undo the legitimate tax planning that led to it. Section 801 applies to tax actually paid and on any footing would continue to do so. The Claimants’ tax planning would not have been undone if HMRC’s approach had been adopted.

For these reasons, we affirm the decision of the judge on issue 2.

Five special cases relating to the appropriate FNR, each turning on its own facts, were argued before the judge. He held in favour of HMRC on three of these cases and against HMRC on the other two. HMRC appeal against the judge’s decision on those two cases, relating to what are described as the Henri Wintermans sale and Arenson Group plc (“Arenson”).

The Henri Wintermans sale refers to dividends paid by BAT Nederland BV (BV) out of capital gains realised on the sale of two subsidiaries. The gains formed part of the distributable profits of BV but, as the judge explained at [75], “did not form part of the tax base for corporation tax in the Netherlands, apparently on the basis of some form of participation exemption”.

HMRC argued that, as the gains did not form part of the taxable profits of BV, they were not subject to economic double taxation and should therefore be subject to a full Case V charge. The claimant argued that the gains formed part of the distributable profits to which the Dutch FNR applied, although they were removed from the tax base.

Henderson J held in favour of the Claimants for the reason given at [77]:

“With some hesitation, I agree with the Claimants on this point. In the absence of any evidence about the precise nature of the relevant exemption, I think it preferable to regard the capital gain as a receipt which prima facie formed part of the taxable profits of BAT Nederland BV and was in principle subject to tax at the nominal rate, even though the effect of the exemption was to narrow the tax base by removing it from charge.”

HMRC submitted that the judge should have treated the gains as wholly outside the tax charge, so no FNR applied to it, unless the claimant could establish by evidence that it was within the tax charge but subject to a relief. We are not satisfied that this is the right approach and we consider that the judge was entitled to come to the conclusion stated by him in [77]. Moreover, we accept Mr Aaronson’s submission that the correct approach is to apply the FNR to BV’s distributable profits, of which the gains clearly formed part. This is a good example of a circumstance in which the effective rate of tax may be less than the FNR and this very example was provided by the Claimants in their submissions to the court in FII CJEU2. On 5 October 2016, Henderson J gave judgment in Six Continents Ltd v HMRC [2016] EWHC 2426 (Ch), in which he heard expert evidence on the Dutch tax law relevant to this issue and accepted the evidence given on behalf of Six Continents, and on that basis reached the same conclusion as in this case.

Arenson was a UK trading subsidiary of a Danish group in which BAT had a minority holding. Dividends paid by the Danish holding company included a relatively small amount of income received as a dividend from Arenson out of profits chargeable to UK corporation tax. HMRC argued that a claim could not be brought in respect of so much of the Danish dividend as was attributable to the dividends paid by Arenson on the grounds that it was a UK company and its dividends were therefore subject to UK tax, even when distributed as part of a dividend paid by its Danish holding company.

The judge rejected HMRC’s case. While Arenson was itself subject to UK corporation tax on its profits and to ACT on any dividends paid by it, that was irrelevant when it came to the credits to be allowed against a Case V charge on dividends paid by the Danish parent company (save to the extent that the amount of such tax could form part of the credit that BAT was entitled to claim on the dividends paid to it by the Danish company). Once received by the Danish company, the dividends paid by Arenson formed part of the Danish company’s profits chargeable to tax in Denmark in accordance with Danish law and dividends paid out of such profits were no different from dividends paid by any other EU company.

We consider that the judge was right for the reasons he gave.

The Claimants cross-appeal against the judge’s decision in relation to German “silent partnership” profits. A UK company (Overseas) in the BAT group held interests in two German partnerships from which it derived income that was chargeable to tax in its hands under Case V. Overseas was entitled to credit against its Case V tax for withholding tax and trade tax paid in Germany but not in respect of the ACT paid when Overseas paid dividends out of the income received from the partnerships. Henderson J rejected any claim in respect of the ACT. There was, he said at [97], no answer to the simple point that there is no objective comparability between investment in a German partnership and investment in a German subsidiary. Legally and economically they are different forms of investment. In the absence of objective comparability, no question of difference of treatment can arise.

The cross-appeal was supported by only very brief written and oral submissions. Mr Aaronson’s main point was that the CJEU’s decisions should be read broadly as applying to any income stream received from a foreign source that has been subject to tax abroad. We do not accept this and we think the judge rightly rejected BAT’s case.

The judge considered two issues under this heading. The first was whether, when calculating the FNR credit, the dividend should be grossed up not only for withholding tax and underlying tax paid in the state of residence of the foreign company, but also, as HMRC argued, at the FNR. The judge rejected HMRC’s case on this and there is no appeal. The second was whether, in calculating the amount of Case V tax that the UK could lawfully charge, credit must be given for withholding tax. The judge rejected HMRC’s case that credit need not be given for withholding tax, and HMRC appeal against that decision.

HMRC submit that credit need not be given for withholding tax because, as is common ground, there is no obligation as a matter of EU law to do so. So, in calculating the amount of tax that the UK could lawfully charge under EU law, the UK tax rate should be applied to the gross amount of the dividend received by the UK company (i.e. the dividend received is grossed up to include both withholding tax and the underlying foreign tax) and then compared with the FNR or foreign effective tax. The difference is the amount of tax that may be charged as a matter of EU law. If the difference is equal to or more than the amount of Case V actually charged, which amount would be calculated after giving credit for withholding tax, HMRC submit that no tax will have been unlawfully charged as a matter of EU law. In other words, in making this calculation, the UK company’s entitlement under domestic law to a credit for withholding tax is ignored but HMRC takes it into account in calculating whether it has charged too much tax.

The judge rejected HMRC’s case on two grounds, reflecting the Claimants’ submissions. First, there is an illogical symmetry in grossing up the dividend to include withholding tax on the one hand but not giving credit for it on the other. Secondly, the conforming interpretation of UK legislation required to give effect to the Claimants’ EU rights does not permit HMRC, or the court, to ignore their domestic rights to credits which are unrelated to and unaffected by their EU rights. As the judge put it at [112], the Claimants’ approach “better accords with the system the UK actually operated”.

The parties repeated their submissions before us. Mr Aaronson neatly summarised his argument by saying:

“You do not remove a tax credit given by Parliament which is the withholding tax credit, in order to take advantage of a tax credit by reference to underlying tax, which EU law demands.”

We consider the judge was right for the reasons he gave.

THE ADVANCE CORPORATION TAX ISSUES

Henderson J answered this question in the affirmative. The parties are agreed that the answer to this questions follows from the answer to issue 1. As we agree with the judge on issue 1, it follows that we agree with him on issue 5.

HMRC accept that this issue goes together with issue 4, and accordingly we dismiss the appeal on this issue.

The answer given by Henderson J was that lawful ACT was to be calculated by granting a notional tax credit in respect of an EU dividend, capped at the UK rate of ACT then in force, such credit being the higher of (a) the underlying tax and withholding tax actually paid and attributable to the dividend and (b) a nominal rate credit calculated by applying the relevant FNR to the gross foreign profits distributed by way of the dividend.

This answer follows from the answers given by the judge to issues 1 and 5. As we agree with those answers, it follows that we agree also with his answer to this issue.

As Henderson J remarked in his judgment at [141], this is an issue of central importance in the case, with a potentially large impact on the sums recoverable by the Claimants.

It is for the Claimants to establish the method by which the extent of the overpayments by them of ACT is to be calculated. HMRC have put forward a different approach to that espoused by the Claimants but it is not a question of choosing the better system. HMRC's approach may be rejected without it following that the Claimants' system is correct. Nonetheless, there must be linkage if effect is to be given to the Claimants' rights, and HMRC may be assumed to have put forward the most appropriate alternative to the Claimants' approach.

The Claimants put forward two possible approaches at trial, one of which (“the FID method”) was rejected by the judge and is not pursued on this appeal.

The essential features of the principal methods put forward by the parties were clearly explained by the judge at [118]-[120]:

“118.

In very broad terms, the rival methodologies for dealing with these problems reflect a fundamental difference of approach which I have already noted in relation to some aspects of the Case V charge. The claimants' approach, in outline, is to take the UK system as it actually operated, and make adjustments to it designed to accommodate the tax credit required by EU law at the point of entry of the EU dividend into the UK. The dividend is thus regarded as carrying a tax credit in the same way as a dividend received by the water's edge company from a UK subsidiary would have done, although the quantum of the credit is not necessarily the same because the credit required by EU law may in certain circumstances be less than a domestic credit at the ACT rate. The credit thus identified is then treated as available for use within the group in the same way as a credit attaching to domestic FII, and is set against the actual ACT payable on subsequent distributions (usually at the level of the top company in the group, because of group income elections).

119.

The Revenue's methodology, by contrast, starts from the proposition that it is only when ACT was actually paid that it is necessary to determine whether any of it was unlawful. It is therefore necessary to begin by identifying the EU source income comprised in the dividends which trigger the actual charge to ACT. Since the charge was usually imposed at the top of the group, after payment up of all or part of the EU income originally received by the water's edge company through one or more intermediate holding companies, and since all the UK companies involved usually had other sources of income, the Revenue have to devise a mechanism for tracing the original EU income as it passed up the group until the stage when ACT became payable. This is an exercise which has no analogue in the UK ACT system, and it involves the making of a number of sometimes arbitrary assumptions, as well as calculations of very considerable complexity.

120.

Once the EU portion of the dividend has been identified, it next has to be decided how much of the ACT actually payable in respect of the dividend (at the rate of ACT then in force) was unlawful. For this purpose, the Revenue calculate the maximum rate of ACT which they say could lawfully have been charged in respect of each stream of EU income represented in the dividends, and express that rate as a percentage which is then compared to the actual ACT rate charged on the dividend. Only to the extent that the ACT actually charged on the EU income components of the dividend exceeds the maximum lawful percentages attributable to those components is the charge unlawful.”

Henderson J held that the method proposed by the Claimants was the right method to adopt. He identified at [143] three reasons to favour the Claimants' general approach of, in broad terms, feeding the credit required by EU law into the ACT system at the UK water’s edge and then treating the aggregate of the dividend and the credit as if it were domestic FII. First, this appeared to be the method mandated by the Court of Appeal in FII CA1. Secondly, it did the least violence to the domestic ACT system. Thirdly, it was readily justifiable as the product of a conforming construction of section 231 ICTA. We will return to the decision of this court in FII CA1 but, while acknowledging that it might well bind him, Henderson J considered the question afresh, particularly as there had been the decision in FII CJEU2 and other developments since then.

The judge accepted the fundamental submission of Mr Aaronson, which he summarised at [145]-[146]:

“145.

In arguing for the claimants' approach, Mr Aaronson rightly went back to first principles. That which needs to be remedied, he submits, is the failure of the ACT system to provide a credit for underlying tax and/or tax at the FNR in the case of EU dividends, because only in that way (according to the analysis of the ECJ) could the UK have justified its decision to counter economic double taxation of corporate profits by the adoption of a dual system of exemption for domestic dividends and imputation for EU dividends. The remedial tax credit needs to be given, and fed into the ACT system, at the UK water's edge, rather than at the level when ACT in fact became payable, because when the dividend reaches the UK it has already been subjected to foreign tax on the distributed profits. The essence of the domestic ACT system is that, when a dividend is paid out of profits which are subject to corporation tax, ACT is payable and the aggregate amount of the dividend and ACT then constitutes FII in the hands of the recipient company. Thus in order to achieve parity of treatment of EU dividends in the hands of the UK water's edge company they must be treated in the same way as dividends received from UK subsidiaries in respect of which ACT has already been paid. The critical point is that, at the stage of receipt in the UK, the EU dividends have already borne actual foreign tax. In order to remedy the unlawful discrimination identified by the ECJ, the foreign tax must be regarded as equivalent to domestic ACT, and the sum of the dividend and the foreign tax credit must be treated as equivalent to FII.

146.

If the matter is viewed in this way, submits Mr Aaronson, it can be seen to be irrelevant that EU dividends were in fact normally paid up the UK group under group income elections, with ACT only becoming chargeable at the top level. In order to remedy the unlawful discrimination under EU law, the dividends must be treated as if they had already borne a tax equivalent to ACT when they were received at the UK water's edge. In order to avoid a double charge to tax, the dividends and their associated tax credits must thenceforth be treated as equivalent to FII. For remedial purposes, it is too late for the dividends to be the subject of notional group income elections, because the payment of ACT which such an election would postpone has already taken place.”

The judge noted as important at [147] that in FII CJEU2 the court had rejected the UK's argument that where the UK water’s edge company paid dividends under a group income election there was no discrimination at that stage, because there was no actual charge to ACT. In holding, contrary to the UK's case, that a restitutionary claim could be made by the parent company which in fact paid the ACT, the court was focusing on the ACT system as it operated in practice, and in particular on the way in which tax credits generated at a lower level would be passed up so as to frank dividend payments at a higher level. Dividends at the higher level would be franked by those tax credits, regardless of whether a later dividend was paid out of the same profits as the original dividend and regardless of any other sources of income which may have fed into the parent company's profits and the dividend paid by it. Henderson J described this as a cardinal feature of the ACT system, which was respected by the Claimants' method but contradicted by HMRC's method.

The judge rejected HMRC's approach, which was very complex and involved making numerous (and often unverifiable) assumptions, precisely because it bore no relation to the actual system of ACT in force in the UK. He considered it be conceptually unsound, in particular because it ignored the operation of the FII system and lacked the essential element of assimilating EU income into the FII system from the water’s edge company. This fatally undermined HMRC's method.

The Claimants also submitted to Henderson J that HMRC's method involved practical difficulties that made it impossible to implement in a coherent and verifiable manner. He was taken through the details of these objections. He concluded at [171] that, in view of the conclusions that he had already reached, it was unnecessary for him to explore these submissions in any detail but he considered them to be, in general, well-founded.

The fundamental submission for HMRC before us, as before the judge, was that no issue of unlawful ACT arises until ACT is charged in respect of a dividend which comprises or includes profits derived from EU-sourced dividends. At that stage, the amount of unlawful ACT can be calculated. In the context of a group of companies where a group income election is in place, as was in fact the case with the Claimants, the unlawful ACT is not charged until a dividend is paid by the ultimate holding company to its shareholders. By reason of the group income election, no ACT would be payable on dividends paid by subsidiaries up the corporate chain. These submissions were summarised by the judge at [119]-[120], cited above.

HMRC submit that the amount of unlawfully charged ACT is determined by identifying the sources of the profits comprised within the dividend declared by the ultimate holding company. In the broadest of terms, this would involve analysing the accounts of each of the subsidiaries in the corporate chain to determine, in the case of each dividend paid by them, the profits out of which the particular dividend was paid. As Mr Ewart emphasised, this is not an exercise in tracing the cash out of which the dividend is paid but tracing the profits comprising the dividend.

To illustrate the point, Mr Ewart took the simplest of examples. A UK company has two subsidiaries, one incorporated in an EU member state and one incorporated in a country outside the EU. Each of those subsidiaries pays a dividend of 100 to the UK company. The UK company pays a dividend of 100 to its shareholders. Unless the UK company has for some reason designated the dividend paid by it as coming from one or other of the dividends received from its subsidiaries, HMRC submits that the dividend will be assumed to be derived equally from the two dividends paid to it.

In practice, of course, the exercise would be very much more complicated. As dividends pass up a corporate chain, income will be received from multiple sources. A company will receive dividends from one or more subsidiaries and income derived from its own trading or other business activities. In order to determine the distributable profits of any particular company, there must be deducted from that company’s various sources of income, its own costs, depreciation and so on. The result is a single figure of its profits, available for distribution in such amount as the company acting by its directors or shareholders decide. HMRC’s approach involves a notional apportionment of each company’s profits to its various sources of income on a pro rata basis, net of all appropriate deductions.

HMRC submit that, contrary to the judge’s view, this does not involve assumptions applied in the absence of evidence to the contrary. They submit that, unless a company has elected to pay a dividend out of profits derived from a particular source, the dividend must “as a matter of fact” be paid rateably out of all its sources of income contributing to its profits. This is not an assumption “but simply a reflection of the facts”.

For these reasons, HMRC submitted, the Claimants’ approach was wrong and should not be adopted.

On behalf of the Claimants, Mr Aaronson provided an example of how the approach of HMRC would operate. Assume a corporate chain of three companies, an ultimate holding company (A), a directly owned subsidiary (B) and a further subsidiary (C) owned by B. C pays a dividend to B of 100, on which it pays ACT of 25. In addition to receiving the dividend of 100 from C, B also receives other income of 100 from a different source. B pays a dividend of 100 to A. If HMRC’s apportionment method were to be adopted, it would follow that the dividend of 100 paid by B would comprise an equal amount drawn from the dividend it had received from C and its other income. On that basis, it would be entitled to a tax credit of 12.5 against its gross liability to ACT of 25. It would therefore be liable to pay ACT of 12.5. If A also had its other income of 100 and also paid a dividend of 100 to its shareholders, precisely the same result would follow.

As Mr Aaronson submitted, this is not how the ACT system operates in a purely domestic context. If C paid ACT of 25 in respect of its dividend of 100 to B, a tax credit of 25 would be available both to B and to A when they respectively paid dividends of 100. Neither A nor B would pay any ACT, notwithstanding that each of them had other income of 100. It was a fundamental feature of the UK system that, whenever a company paid a dividend of an amount which was equal to or less than the amount of dividends it had received on which ACT had been paid, the dividend carried a tax credit.

HMRC accept that in the domestic context the ACT system operated in the way described above. They submit that it is appropriate to adopt the different methodology proposed by them in the case of an EU dividend received by a UK company, precisely because it is not a dividend which was subject to the ACT system.

We accept, as did the judge, the submission for the Claimants that the approach of HMRC fails to provide the equivalent treatment required by EU law. A UK company receiving a dividend from an EU subsidiary paid out of profits chargeable to corporation tax in its home state will be in a worse position than a UK group receiving a dividend from a UK subsidiary. As Mr Aaronson submitted, and as the judge accepted, the critical point is that, at the stage of receipt in the UK, the EU dividends have already borne actual foreign tax. In order to remedy the unlawful discrimination identified by the CJEU, the foreign tax must be regarded as equivalent to domestic ACT, and the sum of the dividend and the foreign tax credit must be treated as equivalent to FII.

Like the judge, we consider that the Claimant’s approach is, as a matter of principle, the correct approach and equally that HMRC’s approach is the wrong approach.

This makes it unnecessary for us to consider in detail the other objections raised against HMRC’s method which, even if accepted, would not in any event mean that the Claimants’ approach was therefore the correct approach, the burden being on them, not HMRC, to establish the correct approach. Nonetheless, we agree with the judge that HMRC’s method does not, as they submit, rest on facts but on assumptions and that it bears no relation to the actual system of dividend taxation in force in the UK during the relevant period: see the judgment below at [166]-[167].

Since Henderson J gave judgment under appeal in the present case, this court has given judgment in Portfolio Dividends CA. The question of linkage was not argued before Henderson J at first instance in that case. The claimants had pleaded the case that is advanced by the Claimants in the present case, but HMRC had not responded to it and the judge ruled that it was too late for them to do so at trial. The judge made a declaration giving effect to the Claimants' case, that foreign-sourced dividends must be treated as having been fed into the domestic ACT system at the water’s edge carrying the tax credit to the extent required by EU law.

Notwithstanding the proper case management decisions taken by the judge, HMRC submitted to this court in Portfolio Dividends CA that they should be permitted to argue their case on the underlying point of principle. The court had of course read Henderson J's judgment in the present case and his detailed treatment of this issue and it knew that the present appeal would soon be heard. Nonetheless, it concluded that it was necessary and appropriate to address the merits of the point. It observed that it had heard full argument and was "as well-placed as this court will be in June to decide this question".

Having considered the submissions of the parties, the court said at [101]:

"In our judgment, the decision of this court in FII (CA) has given the answer to this question. The answer which has been given bears out Prudential’s submission on this point."

The court cited from paragraph [107] of the judgment in FII CA1:

“a conforming interpretation can be achieved simply by reading in words that make it clear that resident companies can claim a credit under section 231 in respect not only of qualifying distributions made by resident companies (domestic-source income) but also distributions made by other companies (foreign-source income) to the extent that Community law requires a tax credit to be given in respect of that income too. The extent of that entitlement can then be investigated when the section falls to be applied. …”

The court continued at [105]:

“This means that the relevant conforming interpretation of the ACT provisions is achieved by modifying the interpretation of section 231 by application of the Marleasing interpretive obligation so as to create a tax credit of the relevant amount in respect of foreign dividends assessed by reference to the relevant foreign nominal or effective rate of tax (whichever is the higher), capped at the UK nominal rate of tax. No other change to the interpretation of the ACT provisions in accordance with their ordinary meaning was suggested by the court and none is necessary to give effect to the requirements of EU law. It may be observed that although HMRC succeeded in overturning the more generous approach (from the taxpayer's point of view) favoured by the judge in FII (High Court), the solution arrived at leads to the conclusion that Prudential is correct in its submissions on this question in the present appeal.”

In its judgment the court considered in some detail the application of other provisions in the ACT legislation, particularly sections 238 and 241: see [92]-[112]. The court’s conclusion at [109] was:

“Once the conforming interpretation of section 231(1) given by this court in FII (CA) and by the judge’s declaration is applied, the other ACT provisions simply apply in accordance with their ordinary meaning, precisely as Prudential submits on the present appeal.”

This part of the court’s decision provides the answer to another of HMRC’s objections to the Claimants’ method, that it requires further conforming interpretations of provisions beyond just section 231. Even if it did, and Mr Ewart instanced section 247 in addition to those expressly considered in Portfolio Dividends CA, we see no difficulty in providing an interpretation that gives effect to the requirements of EU law and goes with the grain of the legislation.

Having invited this court to rule on this issue in Portfolio Dividends CA, rather than wait until the hearing of this present appeal, it seems surprising that HMRC should now seek to re-argue it. It is a course that is not only wasteful of court's resources, to say nothing of the parties’, but it runs counter to the doctrine of precedent on which our system depends.

HMRC seeks to meet the latter point by submitting that this court in Portfolio Dividends CA misunderstood the earlier decision in FII CA1 and that therefore we are free to reach our own view. Moreover, it was submitted that, properly understood, the earlier decision was to the opposite effect to the decision in Portfolio Dividends CA and therefore we are in the position of choosing between two conflicting decisions of this court.

We do not accept these submissions. First, the decision in FII CA1 clearly does not conflict with the decision in Portfolio Dividends CA. Secondly, we see no basis for suggesting that the earlier decision was misunderstood by this court in Portfolio Dividends CA. The court in FII CA1 expressed itself with complete clarity, leaving no room for misunderstanding. We need not, and do not, decide whether it would have been open to us to differ from the judgment in Portfolio Dividends CA, if we thought the decision of the CJEU had been misunderstood, only noting that Starmark Enterprises Ltd v CPL Distribution Ltd [2001] EWCA Civ 1252, [2002] Ch 306 provides arguable grounds for saying that it would have been open to us.

We conclude that the judge was right but we also consider ourselves bound by this court's decision in Portfolio Dividends CA.

Claims in respect of Foreign Income Dividends were both determined as an issue in the FII Test Claimants case and were the subject of the summary judgment applications by Evonik Degussa and others referred to at [23] above. In both cases Henderson J decided the issue in favour of the Claimants. In the former, permission to appeal was refused both by him and, on the papers, by Patten LJ. In the latter, permission to appeal was refused by Henderson J. HMRC’s renewed oral applications for permission to appeal were directed to be heard with the present appeal, with the appeal to follow if permission were given. In fact, we heard full argument and reserved the applications for permission to be decided as part of this judgment.

It is necessary first to give some description of the regime governing FIDs. It was introduced by the Finance Act 1994 as a response to the ever-increasing problem of surplus ACT caused by the large volume of foreign dividends received by UK-based groups and the lower volume of domestic profits giving rise to mainstream corporation tax against which the ACT could be set off. The relevant provisions were enacted as sections 246A to 246Y ICTA.

A description of the principal features of the regime was given in the order for reference to the CJEU in FII CJEU1 and summarised as follows in its judgment:

“23.

From 1 July 1994, a resident company receiving dividends from a non-resident company could elect that a dividend which it paid to its shareholders be treated as a “foreign income dividend” (“FID”). ACT was payable on the FID but, to the extent to which the FID matched the foreign dividends received, the resident company could claim repayment of the surplus ACT.

24.

While ACT was payable within 14 days of the end of the quarter in which the dividend was paid, surplus ACT was repayable when the resident company became liable for mainstream corporation tax, namely nine months after the end of the accounting period.

25.

When a FID was paid to an individual shareholder, the latter ceased to be entitled to a tax credit, but was treated for income tax purposes as having received income which had borne tax at the lower rate. Tax-exempt shareholders, such as United Kingdom pension funds which received a FID, were also not entitled to a tax credit.

26.

For dividends paid after 6 April 1999, the ACT system and the FID regime were abolished.”

For present purposes, the important feature is that ACT was payable on a dividend that a company elected to treat as a FID, but was repayable to the extent that it was matched to qualifying foreign dividends received by it or its subsidiaries. Repayment was generally made between approximately nine and seventeen months later.

The CJEU held that this represented a cash flow disadvantage which was discriminatory and contrary to articles 49 and 63, when compared with domestic dividends that were treated (with the associated tax credit) as franked investment income so that ACT was not paid by the recipient company when it paid a dividend. This was so held by the CJEU in FII CJEU1, subject to one point only – whether the “standstill” provision then in article 64 applied so that article 63 did not apply to the FID regime. This issue was remitted to the English courts for decision. The Court of Appeal, upholding Henderson J, held in FII CA2 that the standstill provision did not apply, so that the regime contravened both articles and it mattered not whether the company paying the foreign dividend was resident within or outside the EU. FIDs matched with dividends paid by non-EU companies are sometimes referred to as “third country FIDs”. There was no appeal from the Court of Appeal’s decision on this point.

Other important features in the present context should be noted. First, in order to qualify as a foreign dividend against which FIDs could be matched, the distributable foreign profits had to have borne a rate of underlying tax which exceeded the UK corporation tax rate. Once matched with a foreign dividend the full amount of ACT paid on the FID would be repayable. There would be no residual ACT not requiring to be repaid nor would there be any Case V charge on receipt of the foreign dividend.

Secondly, the matching rules enabled a FID to be matched with foreign dividends received by any subsidiary in the current or preceding accounting period, and an unmatched FID of the parent could also be matched with a foreign dividend received by a subsidiary in a subsequent period. It follows that it was not necessary to show that the profits distributed as a FID had their source in the foreign dividend with which it was matched.

Henderson J upheld the time value claims in respect of the entirety of the ACT paid (and later repaid) on FIDs. He dealt with the issue more fully in his judgment in Evonik, no doubt because, as he records in his judgment in FII HC2 at [182], HMRC’s submissions in the latter case were not developed orally.

The judge first emphasised, by reference in particular to Gloster LJ’s judgment in FII CA2, that in all the previous cases in which the FID regime had been considered, including in the CJEU and the Court of Appeal, it was the regime as a composite whole that had been considered and held to contravene EU law. He went on at [36] to emphasise the integrated nature of the FID regime:

“… the fact that ACT would only be repaid by HMRC to the company which had declared and paid a FID when it had been matched to HMRC's satisfaction with distributable foreign profits of the company which had already borne a rate of underlying tax in excess of the UK corporation tax rate. In the real world, no UK company would ever have declared a FID unless it was confident that it would be able to perform the necessary matching exercise and thereby obtain repayment of the relevant ACT. Furthermore, the effect of the matching requirements was that, once a repayment of ACT had been made on a final as opposed to a provisional basis, one could be certain that the ACT in question was directly attributable to foreign profits of the company which had already borne tax at a rate higher than the UK rate. The necessary linkage was provided by the matching exercise itself.”

He identified at [37] the central issue as to whether the (entire) ACT paid by the Claimants on their FIDs was unlawful under EU law, and whether this was so clearly established that HMRC had no real prospect of successfully defending the claims.

To this end, the judge traced the history of the preceding litigation on this issue, beginning with question 4 in the first reference to the CJEU, dealing with the FID regime and reformulated by the CJEU as follows:

“By question 4, the national court essentially asks whether Articles 43 EC and 56 EC … preclude national legislation, such as the legislation at issue in the main proceedings, which, while allowing resident companies receiving foreign-sourced dividends to elect to recover ACT accounted for on a subsequent distribution to their own shareholders, first, obliges those companies to pay the ACT and to reclaim it subsequently and, secondly, does not provide any tax credit to their shareholders, whereas those shareholders would have received such a tax credit if the resident companies had made a distribution based on nationally-sourced dividends.”

He then set out the paragraphs in FII CJEU1 in which the court reached its conclusion that the FID regime contravened articles 49 and (subject to the standstill point) article 63, in part because of the cash flow consequence of paying ACT and later reclaiming it. The court’s answer to question 4 was that articles 43 and 56 precluded legislation:

“which, while exempting from advance corporation tax resident companies paying dividends to their shareholders which have their origin in nationally-sourced dividends received by them, allows resident companies distributing dividends to their shareholders which have their origin in foreign-sourced dividends received by them to elect to be taxed under a regime which permits them to recover advance corporation tax paid but, first, obliges those companies to pay that advance corporation tax and subsequently to claim repayment …”

The judge stated at [44] that it was, in his judgment, clear that the CJEU “ruled on the lawfulness of the FID regime as a whole, and held that it infringed both Articles [49] and [63]”. Accordingly, the only live issues relating to FIDs in FII HC1 were the “corporate tree” points which were subsequently resolved definitively in favour of the Claimants in FII CJEU2.

On appeal from FII HC1, the Court of Appeal upheld Henderson J except that a conforming interpretation could be given to section 231 ICTA, rather than disapplying it. Paragraph 16 of the Court of Appeal’s order provided:

“The following claims are successful in relation to the GLO issues determined in the trial:

(a)

claims for the repayment of surplus ACT … or ACT refunded under the FID regime, paid on or after 1 January 1973, by claimants which received dividend income from subsidiaries established in other member states in so far as (i) the ACT paid was not due after taking into account the tax credit available under section 231 ICTA 1988 in respect of those dividends and (ii) the claims are made within the applicable limitation periods;

(b)

claims for the time value of ACT on third country FIDs paid on or after 1 July 1994 and refunded under the FID regime in so far as (i) the ACT paid was not due after taking into account the tax credit available under section 231 in respect of those dividends and (ii) the claims are made within the applicable limitation periods;

(c)

claims for interest based on claims under (a) and (b) above.”

The judge observed, correctly as we think, that the reference to “the tax credit available under section 231” must be a reference to the notional tax credit available under the conforming interpretation of section 231, rather than to any actual tax credit, because section 231 does not otherwise provide any tax credit for foreign dividends.

The judge continued at [53]:

“Once this point is understood, it seems clear to me that the restrictions contained in sub-paragraphs (a)(i) and (b)(i) of the varied Order 1 will have no impact on claims such as those brought by the present claimants. The unlawfulness of the FID regime as a whole under EU law has been definitively established by the ECJ. It follows that the ACT levied on the FIDs was unlawful, and the companies should have been placed in the same position as a UK company in receipt of FII from a UK subsidiary. The matching rules mean that all of the dividends matched with the FIDs will have been paid out of profits which bore an underlying rate of tax higher than the UK corporation tax rate. Accordingly, had the company which paid the FIDs instead received the same dividends in a purely domestic scenario, they would have been received as FII and would have franked the onward distribution in its entirety. Moreover, EU law would require credit to be given for the whole of the notional foreign tax, up to the UK corporation tax limit, because only thus could economic double taxation of the same underlying profits be avoided. In other words, in cases of the present type the notional tax credit required by EU law under the modified section 231 would in effect require the matched dividends to be treated as if they were domestic dividends, with the result that no ACT on the FIDs was payable.”

The judge proceeded to consider the submissions raised on behalf of HMRC in support of their position and concluded that none had any real prospect of success. He accordingly entered summary judgment and refused permission to appeal.

It is worth also citing the paragraph in his judgment in FII HC2 containing the crux of Henderson J’s reasoning:

“These claims are simpler than those relating to ordinary dividends because the FID regime was subject to a separate and self-contained legislative code which required the FIDs to be matched with identified foreign profits which had borne underlying tax in excess of the UK corporation tax rate. In those circumstances, there could be no question of EU law requiring any further tax credit to be granted for the FIDs, but equally there is no further linking exercise to perform before the extent of the illegality of the ACT charge can be ascertained. Since the whole of the matched foreign profits had already borne foreign tax at a rate above the corporation tax rate, it was clearly unlawful to subject the same profits to a second charge to corporation tax in the form of ACT.”

HMRC submit that the judge was wrong to decide that the unlawfulness of all of the ACT paid by the claimant on FIDs had been established and that the CJEU in FII CJEU1 had ruled on the lawfulness of the FID regime as a whole in such a way that any ACT paid on the FIDs was unlawful. Previous decisions of the CJEU and this court in the FII litigation had held the FID regime to be unlawful in principle but had not determined the extent of the unlawfulness or how any remedy should be quantified. On the contrary, those courts had treated ordinary dividends and FIDs in the same way, as can be seen from paragraph 16 of this court’s order set out above. They were all aspects of a single corporation tax system and the same principles and methods of quantification, when finally settled, should be applied to all claims, including those in respect of ACT paid on FIDs.

HMRC submit that claims in respect of FIDs should be quantified in accordance with its top-down “tracing” system which we have considered, and rejected, under issue 9. But even if that system is rejected, the method proposed by the Claimants and accepted by us, and by this court in Portfolio Dividends CA, should be applied, with the result that not all the ACT paid on FIDs will necessarily be unlawful. In this connection, HMRC emphasise the feature of the FID regime mentioned above, that the profits distributed by way of a FID need not have originated in any foreign dividend. The system simply involved matching foreign dividends with dividends that a company elected to treat as a FID for the purpose of reclaiming ACT paid on those foreign dividends.

HMRC submit that the fact that under domestic law a special regime applied to enable the Claimants to choose to match their FIDs with qualifying foreign dividends is “wholly irrelevant” in working out how much ACT the UK could charge on those FIDs. The purpose of matching was to enable ACT to be reclaimed, not to remedy the breach of EU law found by the CJEU to exist in the FID system. HMRC sought to support these submissions by reference to the reasons given by the judge at [160]-[162] for rejecting the Claimants’ attempt to introduce “the FID method” as an alternative in the very different context of non-FID ACT under issue 9.

Mr Ewart devoted a good deal of his submissions on behalf of HMRC to seeking to demonstrate that the CJEU and the CA had not decided the issue of how to calculate the unlawful amount of ACT paid on FIDs and that the judge was wrong to think that he was bound by authority on this issue. We consider that Mr Ewart misunderstood the judge’s reasoning in this respect. He was not, as we read his judgment, saying that the earlier decisions had expressly or necessarily considered and decided this issue, leaving him simply to follow those decisions.

The judge’s reasoning in summary was that all the ACT paid in FIDs was recoverable, because of a combination of (i) the CJEU’s decision, confirmed in FII CA2, that the FID regime was unlawful in the respects identified by the CJEU, (ii) the linkage provided by the FID system itself between the FID and the matched foreign dividend, and (iii) the requirement of the FID regime that the distributable profits had to bear a rate of underlying tax exceeding the UK corporation tax rate.

We are entirely satisfied that the judge adopted the right approach to this issue. As we have earlier said, it is the task of the domestic court, in accordance with domestic law, to identify and apply the means of giving effect to the Claimants’ EU rights. If possible, under English law, this is achieved by a conforming interpretation. As it seems to us, there is no difficulty in producing a conforming interpretation of the FID regime, as the judge has done. Far from ignoring the existence of the FID regime as HMRC’s approach does, it simply makes the necessary adjustment of requiring a credit for the tax paid on the foreign profits, which inevitably results in the entire ACT being unlawful. It is an approach that clearly goes with the grain of the legislation.

We are not impressed by Mr Ewart’s submission based on the possibility that in some cases it could subsequently transpire that the profits out of which the matching foreign dividend had been paid had not borne tax at the necessary level, for example because of a repayment of the foreign tax, so that all or part of the ACT on the FID had been lawful. Mr Ewart submitted that a system which permitted payment of ACT to be considered unlawful only for it subsequently to become lawful was not coherent. We see no force in this point. It is frequently the case that tax payments have subsequently to be repaid and tax reliefs subsequently to be disallowed.

We consider it right to give HMRC permission to appeal on issue 10 in the main case and against the summary judgments in Evonik, but we dismiss the appeals.

This part of the judgment addresses the remedies issues decided by the judge under issues 15 and 17-27. As we have already said, once it is established (as it has been) that the Claimants are entitled under EU law to recover unlawfully levied Case V tax, and to be compensated for the use value of prematurely paid ACT, questions arise as to how those San Giorgio rights are to be given effect under English law. The Claimants have advanced both Woolwich claims and mistake-based restitutionary claims for that purpose. In broad terms, the judge dealt with three arguments advanced by HMRC. First, that they were entitled under English law to set off certain credits against the overpaid tax; secondly, that the Claimants were only entitled to recover the “actual benefit” that HMRC had received from having obtained the overpaid or prematurely paid tax; thirdly, that HMRC were entitled to defend the mistake-based restitutionary claims under English law on the basis that they had a valid “change of position” defence. In relation to each of these arguments, however, the Claimants responded by contending that none of them was open to HMRC as a matter of EU law.

In considering remedies, it is important to understand when the discussion concerns English law and when it concerns EU law. The position is, however, complicated by the fact that the Claimants’ primary claims are their mistake-based restitutionary English law claims brought to enforce their EU law San Giorgio rights. The logical approach in this case requires an understanding of the EU law right before considering the English law remedies that are being deployed to give it effect. We acknowledge that Lord Walker took a different view in FII SC at [36], but that was because the issues were different on that appeal. Here, the two most important remedies issues concern the evaluation or quantification of HMRC’s enrichment (i.e. the set-off and actual benefit arguments) and the question of whether EU law prevents HMRC arguing that they should not provide restitution of the entirety of the overpaid tax, the objective use value of that tax, and the objective use value of the prematurely paid ACT. Whilst the enrichment issues (and the change of position issues) are primarily matters of English law, they have to be understood against the background of the San Giorgio right to which the English law remedies are giving effect.

Our treatment of remedies will, therefore, be taken in the following order:

An introduction to the way the debate about remedies has progressed in this and other related litigation. Many of the decisions relied upon by the parties can only properly be understood against the background of the legal position as at the date of the specific decision. Not only is the English law of restitution developing, but the CJEU has developed and explained the relevant EU law as the cases have progressed. We have tried to keep this section brief, but it will give the reader the opportunity to see the chronology of decisions at a glance.

The enrichment issues in English law. These issues concern the proper evaluation of the amount of restitution to which the Claimants are entitled. They fall into two categories:

The defences to mistake-based restitutionary claims in English law: change of position. This section comprises issues 19 and 20. Issue 19 asks whether the change of position defence is available to HMRC, and issue 20 asks whether it was made out on the facts.

Effect of EU law. This section considers whether all or any of the English law conclusions under (2) and (3) are precluded by EU law. These are issues 18, 24, and 21, asking respectively whether each of the set-off argument under issue 17, the actual benefit argument under issue 22, and the change of position defence under issue 19, is precluded by EU law.

INTRODUCTION TO REMEDIES

One of the main remedies questions is whether the judge was right to decide that EU law required the mistake-based claims made by the Claimants to be treated as San Giorgio claims, even when the Claimants were also theoretically able (subject to the applicable six-year limitation period) to make Woolwich claims to recover the same overpaid tax. HMRC’s central contention is that the availability of Woolwich claims is sufficient as a matter of EU law to vindicate the Claimants’ San Giorgio rights. HMRC accept, as the judge recorded at [249], that the Woolwich claims allow the Claimants to recover both the full principal amounts of overpaid Case V tax and unutilised ACT, in addition to the time value of prematurely levied ACT measured in accordance with Sempra Metals principles (see paragraph 213 below). The only defence available to such Woolwich claims would be one of “passing on”, or unjust enrichment of the claimant, which is not alleged in the present case. It is common ground that none of the arguments raised by HMRC to reduce the quantum of, or as defences to, the mistake-based claims (namely set-off (issues 15, 17 and 18), change of position (issues 19-21) and actual benefit (issues 22-24)) would be available to reduce or defeat the Woolwich claims.

The judge decided that HMRC’s argument that only the Woolwich claims were required to vindicate the claimants’ San Giorgio rights was plainly wrong. He dealt with the question directly or indirectly in a number of parts of his judgment (particularly [247-249], [253-263], [288-289], [293-296], [303-308], [400-407], and [425-430]), though HMRC would say that he never really addressed it head-on. In short, HMRC submitted to us, as they had to the judge, that none of the authorities he relied upon, and certainly none of the CJEU decisions, have concluded that, where there are two available domestic remedies to vindicate a San Giorgio right, both must be moulded equally to satisfy the requirements of EU law. HMRC accept that Littlewoods CJEU establishes that an available domestic remedy cannot be retrospectively made less favourable by curtailing a previously existing limitation period without proper transitional arrangements, but not that each of two available domestic remedies must be made equally effective to vindicate the San Giorgio right, if one of them already does so. The reason, of course, why the argument is so important is that the Claimants’ Woolwich claims can only go back six years, whilst the mistake-based claims can go back much further, and are therefore far more valuable.

The nature of the San Giorgio right available to these Claimants is also directly relevant to the set-off issues we have described.

It is useful, therefore, first to track through the relevant decisions chronologically. We will include both the important domestic decisions as well as those of the CJEU. This is because the development of the law both domestically and at EU level has been complex and intertwined, and we do not believe one can properly understand the present position without seeing what changes occurred in which order. We will severely limit our citations from these authorities, merely summarising their effect, so they can be seen in context. We hope, however, that the chronological summary will form a sound basis for our consideration of the issues that follows. Before embarking on this exercise, however, we would wish to emphasise the following underlying points that need to be borne in mind.

First, there is a fundamental difference between claims in restitution and claims for compensation or damages. In EU law, a damages claim is generally subject to the three conditions explained by the CJEU in Francovich v. Italy (case C-6/90) [1991] ECR I-5357, as formulated by the CJEU in Brasserie du Pêcheur S.A. v. Federal Republic of Germany and R v Secretary of State for Transport ex p. Factortame Ltd (joined cases C-46/93 and C-48/93) [1996] Q.B. 404 (the “Francovich conditions”), namely that (i) the rule of law infringed must be intended to confer rights on individuals, (ii) the breach must be sufficiently serious, and (iii) there must be a direct causal link between the breach of the obligation and the damage sustained by the injured party. The CJEU, however, does not always make a clear distinction between damages claims, or “claims for compensation for loss”, on the one hand, and restitution on the other. This has given rise to some confusion particularly in understanding FII CJEU1. Whatever the descriptions used by the CJEU, it is clear that the San Giorgio right in EU law to recover overpaid tax is not itself a damages claim to which the Francovich conditions apply.

The second point to bear in mind is that there is an inherent tension between the English law remedies in restitution, whether they are Woolwich claims or mistake-based claims, and the San Giorgio right to recover overpayments of tax. The San Giorgio right to recover overpaid tax is not a right that takes any account of the level of enrichment of the tax authority receiving that overpaid tax. The English law of restitution considers the level of actual enrichment of the recipient, which is irrelevant to both the existence and extent of the San Giorgio right.

The third point is related to the second, in that it is clear that if the domestic law remedy is not adequate to vindicate the EU law right, the remedy must be moulded or reformed in order to do so. Thus, the exercise that the domestic court is engaged upon is to “mould” a restitutionary remedy, which focuses on the enrichment of the defendant, in order to provide full compensation for overpaid tax. That might be regarded more accurately as a total transformation than any kind of moulding process.

We turn now to deal with the cases in chronological order. In each case, we provide the date of the actual decision in parentheses, so that the chronology can be better understood.

In San Giorgio itself (9 November 1983), the CJEU decided that a taxpayer’s entitlement to repayment of charges levied by a member state contrary to EU law was a consequence of the rights conferred on individuals by EU law prohibiting the discriminatory application of internal taxes [12]. The CJEU also held [13] that EU law did not prevent a domestic legal system from disallowing such repayment where the claimant would thereby be unjustly enriched, as where the unduly levied taxes had been “passed on” to other persons. No other possible defences to domestic claims in unjust enrichment were mentioned by the CJEU.

In Woolwich (20 July 1992), the House of Lords reformulated the English law of restitution so as to establish that “money paid by a citizen to a public authority in the form of taxes … pursuant to an ultra vires demand by the authority is prima facie recoverable by the citizen as of right” (per Lord Goff at page 177F). Claims in restitution of this kind have been known as “Woolwich claims” ever since.

In Kleinwort Benson v Lincoln City Council [1999] 2 A.C. 349 (“Kleinwort Benson”) (29 October 1998) the House of Lords held for the first time that money could be recovered in restitution on the ground that it had been paid under a mistake of law subject to the defences available to claims in restitution. Claims of this kind have been referred to in the jargon of this case as “mistake-based claims”.

In Metalgesellschaft Ltd v Inland Revenue Commissioners and Hoechst AG v Inland Revenue Commissioners (joined cases C-397/98 and C-410/98), [2001] Ch 620, (“Hoechst”) (8 March 2001), the CJEU was concerned with another aspect of the group litigation concerning ACT. The issues related to EU companies and their UK subsidiaries, which were denied the right to make group income elections that UK companies with UK subsidiaries could make. The EU companies with UK subsidiaries claimed, therefore, that the rules concerning group income elections discriminated against them, because UK subsidiaries were thereby excused from paying ACT on distributions to their UK parents. The CJEU was asked four questions. The first was answered by saying that it was contrary to article 49 on freedom of establishment for UK legislation to afford UK companies the possibility of making group income elections which allowed them to pay dividends to their UK parent company without paying ACT, but to deny that possibility to UK companies with EU parents [76].

The second question put to the CJEU in Hoechst was whether the EU treaties gave the EU parent and/or the UK subsidiary a restitutionary right to claim interest on the ACT paid too early, or whether they only had a claim in damages. That question was answered by saying that EU law required the claimants to have an “effective legal remedy in order to obtain reimbursement or reparation of the financial loss which they have sustained” [96], and that was not altered by the fact that the claim was only for interest on the ACT paid too early. It will be recalled, of course, that at the time that Hoechst was decided, English law prevented recovery of interest on a principal sum that was not due. The CJEU reframed the question, but stressed that it was not for the CJEU to assign a legal classification to domestic claims. The claimants had to classify their claims subject to the supervision of the national court [81]. The breach of EU law arose from levying ACT prematurely [87], and the award of interest represented the “reimbursement” of what was improperly paid [87]. Accordingly, the treaty entitled those claimants to such interest by way of restitution [89]. The same applied if the claim were brought as a claim for damages [92-95].

The third question asked whether the EU treaties allowed one member state to deny any tax credit to a company resident in another member state, when that credit was granted to resident companies and to companies resident in some other member states by double taxation conventions. If not, the fourth question asked whether the member state was obliged to accord tax credits to the EU companies on the same basis as they were accorded to resident companies. The CJEU said it was unnecessary to answer these questions in the light of the answer given to the first question. Mr Aaronson placed great emphasis on that aspect of the decision.

The next important test case was Pirelli Cable Holding NV v Inland Revenue Commissioners [2006] UKHL 4, [2006] 1 WLR 400, (“Pirelli”) (8 February 2006), where the House of Lords had to consider the practical effect of Hoechst on the assessment of the “compensation” due to group companies. Dividends had been paid by UK subsidiaries to Italian and Netherlands resident parents, but they had not been able to make a group income election. The UK subsidiaries paid ACT, but the EU parents did not receive tax credits under section 231, although they did receive convention double taxation tax credits that were immediately paid by HMRC to those parents. The basic issue was whether, in paying “compensation”, HMRC could claim credit for the convention tax credits it had had to pay the EU parents. The first question looked at the position that would have pertained had the group been entitled to make a group income election, and asked whether the EU parents would then have been entitled to the convention tax credits. The House of Lords put a purposive construction on the relevant legislation so as to conclude that the EU parents would not, in that situation, have been entitled to convention tax credits (Lord Nicholls at [12-15], Lord Hope at [35-39], and Lord Walker at [103-106]). The second question focused on the distinction between the legal personalities of the subsidiary and the parent. The question was whether HMRC could have credit against the “compensation” payable to the UK subsidiary for a payment they had made to the separate EU parent. The House of Lords held that HMRC were able to take that credit because, as a matter of commercial reality, the loss of the opportunity to make a group income election was the loss of a group fiscal package (see Lord Nicholls at [20-22]). It is important to note that the speeches in the House of Lords did not address the type of “compensation” with which they were dealing.

The decision of the House of Lords in DMG (25 October 2006) followed swiftly on the heels of Pirelli. Again, the House of Lords had to consider the consequences of the Hoechst decision. In DMG, however, the UK subsidiary of a German parent was claiming “compensation or restitution” for the premature payment of ACT, under a mistake of law, as a result of its inability to make a group income election. The House of Lords held that a taxpayer who had wrongly paid tax under a mistake of law was entitled, under English law, to a restitutionary remedy, notwithstanding that it might have another common law claim (a Woolwich claim) with a shorter limitation period. HMRC had unsuccessfully argued that Kleinwort Benson did not extend to the payment of tax under a mistake. The decision is of some importance to the limitation question under issue 28, to which we will come later in this judgment. For present purposes, we have mentioned it as an important step in the establishment of the common law cause of action to recover tax paid under a mistake of law.

The next significant development was the CJEU’s decision on the first reference in this case in FII CJEU1 (12 December 2006). The decision in that case has already been summarised. They concerned the lawfulness of the corporation tax and ACT payable in the circumstances of this case. Those circumstances are obviously different from the factual situation in the cases already mentioned which all concerned the consequences of a UK subsidiary paying ACT, because it was unable to make a group income election where it was paying dividends to an EU parent. The CJEU did, however, make a comparison with the Hoechst situation that is relevant to the remedies that may be available in this case. In answering the second and fourth questions about the lawfulness of ACT paid by UK companies with EU subsidiaries which had made distributions to it, the CJEU relied on Hoechst as demonstrating that the failure to give the UK company a tax credit when it received dividends from its EU subsidiary was not justified by the fact that the EU subsidiary was not required to pay ACT. In that situation, the EU subsidiary had paid foreign corporation tax instead ([84-93] and [152]). On the same day (12 December 2006) the CJEU delivered its decision in Test Claimants in Class IV of the ACT Group Litigation v Commissioners of Inland Revenue (case C-374/04) [2007] STC 404 including the cases brought by Pirelli and others. It held there that it was compatible with EU law for the member state only to grant a corporation tax credit to resident companies receiving a dividend, and not to non-resident companies receiving a dividend, because the non-resident companies are not liable to UK corporation tax.

After FII CJEU1, Sempra Metals Ltd v Inland Revenue Commissioners [2007] UKHL 34, [2008] 1 AC 561, (“Sempra Metals”) (18 July 2007) was decided. It was the third House of Lords decision directly consequent on Hoechst (Sempra Metals was formerly known as Metalgesellschaft and was a party to the Hoechst decision in the CJEU). The decision is most well-known for overruling President of India v La Pintada Cia Navigation SA [1985] AC 104, which had prevented the award of interest at common law for late payment of debts, but that is not the aspect that is relevant to this case. The House upheld the decision to award compound interest by way of restitution for the premature payment of ACT under a mistake of law in circumstances where the group had been deprived of the opportunity to make a group income election. It treated the matter as one of English law. Lords Hope, Nicholls and Walker all considered the approach in English law to a claim for restitution of the use value of money, holding that it was open to a recipient including HMRC to show that there was no actual enrichment once the money fell into their hands (Lord Hope at [48], Lord Nicholls at [117-119], with whom Lord Walker agreed).

It was then more than a year before Henderson J gave judgment in FII HC1 (27 November 2008). In terms of remedies, his decision can be summarised briefly. He held that the San Giorgio principle extended as a matter of EU law to all claims for repayment of unduly levied tax and associated interest and loss of use claims, and that the EU law principles of equivalence and effectiveness required English law to provide both Woolwich and mistake-based restitutionary claims ([240-1], [260], and [441-2]). Henderson J, however, considered (wrongly, as FII CA1 held at [157-174]) that a Woolwich claim still required a demand, which did not exist in claims for unlawfully paid ACT.

The Court of Appeal in FII CA1 (23 February 2010) held that the judge had been wrong to reject HMRC’s submissions that a Woolwich claim alone provided a sufficient domestic remedy for the Claimants’ San Giorgio claims [157].

In Littlewoods HC1 (19 May 2010) Vos J had to consider whether the Woolwich claim was sufficient to vindicate the claimants’ San Giorgio right to recover overpaid VAT and the use value of that VAT. It is clear from [82-92]] that, in referring the matter to the CJEU, Vos J felt himself bound by FII CA1 to hold that only the Woolwich claims were needed to vindicate the claimants’ San Giorgio rights.

On 20 December 2010, Henderson J referred five specific further questions to the CJEU seeking clarification of FII CJEU1. FII CJEU2 (13 November 2012) did not, however, deal in any significant way with remedies issues.

The decision of the CJEU in Lady & Kid A/S v Skatteministeriet (C-398/09), [2012] STC 854, (6 September 2011) did not concern corporation tax. Instead it concerned the unlawfulness under EU law of a Danish business tax (the “Ambi”). The Danish Ministry of Fiscal Affairs resisted repayment of unlawfully levied Ambi, where the claimants had in fact taken advantage of even greater benefits as a result of the abolition of employer social security contributions undertaken as the quid pro quo for the introduction of the Ambi. In [20] and [25] the CJEU made clear that the “direct passing on of the tax wrongly levied [to] the purchaser constitutes the sole exception to the right to reimbursement of tax levied in breach of [EU law]”.

The Supreme Court in FII SC (23 May 2012) held that the question of whether EU law required both Woolwich and mistake-based claims to be available was not acte claire (though the majority thought that the Court of Appeal in FII CA1 had been wrong on the point) and referred the matter to the CJEU alongside the questions about whether retrospective statutory limitation periods contravened EU law (paragraphs 14-16 and 23 of Lord Hope, paragraphs 102, 115 and 120 of Lord Walker, paragraph 191 of Lord Sumption, and paragraphs 212 and 222-227 of Lord Reed).

The judgment of the CJEU in Littlewoods CJEU (19 July 2012) did not deal expressly with detail of the remedies needed to vindicate a San Giorgio right. The CJEU repeated that it was for the internal legal order of each member state to lay down the conditions in which interest on overpaid tax levied in breach of EU law was to be paid. Those national rules should not lead to the taxpayer being deprived of an “adequate indemnity” for the loss occasioned through the undue payment of (in that case) VAT [29].

In Benedetti v Sawiris [2013] UKSC 50, [2014] AC 938 (“Benedetti”) (17 July 2013) the Supreme Court considered what approach should be taken to the valuation of services for which the provider was entitled to restitution.  Although the actual issue is not directly material, some of the discussion of principle is nevertheless relevant for our purposes.

In Portfolio Dividends HC2 (24 October 2013), Henderson J decided that the portfolio dividends claimants (owning less than 10% of the shares in various companies) were entitled to compound interest on their claims including the prematurely paid ACT, following Sempra Metals. He concluded at [167-189] that HMRC’s change of position defence to the mistake-based claims failed as a matter of EU law. At [179-184], Henderson J rejected HMRC’s argument that only the Woolwich claims were needed to satisfy the claimants’ San Giorgio rights. He relied on the fact that the majority in FII SC had provisionally held that both remedies were needed as follows:-

“[181] … The view of the majority is perhaps most clearly stated by Lord Reed, who said …:

[212] ... Where an action for the recovery of taxes under domestic law can be based either on the ground of mistake or on the ground of unlawful demand (or, as in the present case, on both grounds), it follows from the principle of equivalence that both grounds of action should also be available in similar circumstances to enforce an analogous right under EU law. So long as they must both be available, they must also both be effective. The principle of effectiveness therefore applies to both grounds of action.' (Emphasis added.)

[182] Mr Ewart criticised the final sentence which I have italicised as a non-sequitur, but he accepted that all of the majority agreed with it. Nor do I accept the implausible suggestion that Lord Reed was guilty of a logical fallacy. It seems to me he was making the simple point that, where domestic law provides two remedies for the recovery of unlawfully levied tax, the EU law principle of effectiveness must apply to both of them alike. Given the existence of the Woolwich remedy, effectiveness does not require the ingredient of a mistake to be removed from the other remedy; but it does require claimants to be allowed to choose between them while they both remain in force”.

At the time that Henderson J decided Portfolio Dividends HC2, only Advocate General Wathelet’s opinion was available in FII CJEU3, but Henderson J relied on that opinion at [47], [48] and [53] as support for his view.

The judgment in FII CJEU3 was delivered on 12 December 2013. The CJEU did not approach the questions referred in the same way as the Supreme Court had done. It did not, for example, think, that the primary question was whether, if two domestic remedies were available, both were to be treated as San Giorgio claims necessary to vindicate the claimants’ EU law right to recover unlawfully overpaid tax. Instead, the CJEU held simply that, in such a situation, the EU law principles of effectiveness, legal certainty and the protection of legitimate expectations precluded national legislation curtailing a limitation period otherwise available to one of the remedies retroactively and without notice [49]. According to HMRC’s submissions, there remained uncertainty about whether what was prohibited was the curtailment of the limitation period of any existing claim that could be used to vindicate an EU law right, or whether the ratio of the decision was really that both available and existing causes of action had to be regarded as San Giorgio claims.

In Littlewoods HC2 (28 March 2014), Henderson J revisited the questions that Vos J had considered in the light of FII SC and Littlewoods CJEU. He concluded at [328-341] that the claimants had to be allowed to pursue both their Woolwich and their mistake-based claims to recover the use value of the overpaid VAT. Henderson J relied also on his two recent judgments in Investment Trust Companies (in liquidation) v. Revenue and Customs Commissioners [2012] EWHC 458 (Ch), [2012] STC 1150, and Portfolio Dividends HC2 on the same point.

The appeals from both Littlewoods HC1 and Littlewoods HC2 were determined by Littlewoods CA (21 May 2015). The Court of Appeal said this at [130]: “[[37-39] of FII CJEU 3 seem] to us to be a clear rejection of the Court of Appeal’s view that the test Claimants in FII were not entitled to protection in respect of both existing domestic causes of action which were capable of giving effect to their EU San Giorgio rights. In the Supreme Court which made the reference views differed on this point but Lord Hope, Lord Walker and Lord Reed, all considered that the summary removal of the DMG claims by the disapplication of s 32(1)(c) was incompatible with the principles of equivalence and effectiveness”. HMRC, however, relied on [141-2] of the Court of Appeal’s judgment (dealing with Henderson J’s [328-341] in Littlewoods HC2) as demonstrating that, provided there was one remedy able fully to vindicate the San Giorgio right (in this case the Woolwich remedy), the other remedy (in this case the mistake-based claims) did not have to be moulded or reformed so as to do so, always so long as it was not amended retrospectively or without notice so as to make it less favourable. The passages particularly relied upon were as follows:-

“[141] The statement of principle in para 33 [of Littlewoods CJEU] confirms that the process of disapplying any domestic rule of law in favour of EU law leaves the national court with procedural autonomy in relation to available remedies. But consistently with that, it does not give to the national court any power of selection which it does not have under domestic law. The national court is left to apply its ordinary domestic rules in the form of the causes of action which are available to a taxpayer seeking to enforce its EU claims. The difficulty with HMRC's argument on this point is that it seeks to attribute to or invest the national court as a function of the principle of effectiveness with the power to select which remedies the Claimant should be permitted to pursue when the object of the principle of effectiveness as explained in [Littlewoods CJEU] is to ensure that the taxpayer’s San Giorgio rights are enforced. The ECJ in para 33 and the earlier cases there referred to has made it clear that the choice and availability of remedies is exclusively a matter of domestic law subject only to their being effective remedies for the purpose of enforcing the EU rights in question … Once it is clear that the domestic law rules for the recovery of overpaid tax are incapable of providing recovery in accordance with the San Giorgio principle, they fall to be disapplied in favour of the Claimant’s EU rights. The national court has no power in our view to disapply the domestic bar to the enforcement of those rights on a selective basis. The procedural autonomy it is granted under EU law simply requires it to make available to the Claimant the remedies which domestic law would give him had the claim for overpayment been a purely domestic one. …

[142] Henderson J was therefore right in our view to reject HMRC's argument that the court should determine which of the otherwise available domestic causes of action Littlewoods should be permitted to rely upon. That is a matter of domestic law and the decision of the House of Lords in DMG confirms that this is a matter of choice for the Claimant” (emphasis added).

We will deal with the judge’s approach to these issues when we come to each of the EU law remedy issues in turn.

THE ENRICHMENT ISSUES IN ENGLISH LAW (ISSUES 17, 15, 22 & 23)

We are dealing with the enrichment issues first, because they affect the quantification of the mistake-based claims in restitution. HMRC’s claim to have changed their position is a true “defence” to the claim in restitution, so does not come under this heading.

Issues 17 and 15, however, relate as we have said to the question of whether or not HMRC can set off payments they claim to have made against the restitution they must make. We will deal with them in the order that the judge dealt with them, even though issue 17 is of far greater financial significance than issue 15, which only concerns convention half tax credits provided to FCE’s US parents.

Issues 22 and 23 concerning “actual benefit” also relate to the quantification of the restitutionary claim. The argument on “actual benefit” is that HMRC as the recipients of the use value of the prematurely paid tax were not actually enriched by it, because the premature payments were spent in the same way as all other receipts, most of which was in respect of income expenditure which is irretrievable and of no lasting benefit.

In broad terms, this issue asks whether HMRC should be able to deduct the double taxation treaty half credits that they have had to pay to the US parent companies (whom we shall call collectively “Ford US”) of FCE from the restitution that HMRC must make to FCE in respect of unlawfully levied ACT. The ACT in question is recoverable by FCE because FCE paid ACT on its distributions to Ford US, having itself received dividends from FCE’s subsidiaries in Germany, Denmark and Austria. Those subsidiaries distributed profits made in those jurisdictions, upon which the foreign corporation tax had been paid at a rate equal to or exceeding the UK rate of corporation tax, for which full double taxation relief was afforded.

The judge decided that HMRC had to make restitution for the ACT paid by FCE without deducting the double taxation treaty half credits allowed to Ford US. His reasons are lucidly explained at [226-241] which we will not repeat here. But in essence he applied the decision of the House of Lords on the first issue in Pirelli, but distinguished the House of Lords’ decision on the second issue in that case. In applying the House of Lords’ ruling on the first issue in Pirelli the judge said at [235] that:-

“Since it is the unlawful charge to ACT on the onward distribution of the EU dividends for which FCE has to be compensated under EU law, and since (applying the decision on the first issue in Pirelli) the US parent companies would have had no entitlement to the treaty half tax credits if FCE had not paid the relevant ACT, I think it would in principle be correct to regard the tax credits as countervailing financial benefits received by the US parents which would not have been available if the UK tax system had operated in accordance with EU law. So viewed, the credits should be taken into account in calculating the compensation due to FCE, unless the fact that they were paid to the parent companies, and not to FCE, prevents this.”

The judge then distinguished the decision on the second issue in Pirelli to the effect that the Pirelli group was to be treated as a single unit when considering the “amount of compensation” that should be payable by HMRC. In Pirelli, of course, a group income election was not available because the parent companies were resident in Italy or the Netherlands. The House of Lords held that the fact that the double tax credits were due to the EU parent companies rather than the English company paying the ACT did not prevent HMRC from claiming credit for them when providing restitution of the ACT for two main reasons: first because, according to the judge’s analysis, the group income election could not have been made by the subsidiary alone (see Lord Nicholls at [21] in Pirelli and the judge at [236]), and secondly because the Pirelli group was complaining about breaches of the freedom of establishment provisions in article 52 which had caused detriment to the group as a whole (see Lord Scott at [80] and Lord Hope at [42] in Pirelli, and the judge at [238]). The judge distinguished the House of Lords’ decision on the second issue in Pirelli by pointing out that neither of these two reasons applied to the position of FCE and Ford US. FCE and Ford US were never in a position to say they ought to have been allowed to exercise a group income election (as the Pirelli parent and subsidiary had been), and Ford US had no right of establishment under EU law (to which the Pirelli parents had been entitled). Moreover, even though there was an indissoluble link between FCE’s payment of ACT and the HMRC’s payment of the half tax credits to Ford US, such a link could not justify disregarding the separate corporate identities of FCE and Ford US.

Against that background, HMRC submit in support of their appeal on issue 15 that the judge’s analysis overlooked the fact that FCE was making claims in restitution, and that such claims look to the enrichment of the defendant rather than any loss to the claimant. It is, therefore, necessary, in evaluating an appropriate remedy, to consider not only payments made to the claimant, but also payments necessarily made to other associated parties. That was the position in Investment Trust Companies v. HMRC [2015] EWCA Civ 82, [2015] STC 1280 (“ITC CA”) where certain closed end investment trusts claimed that HMRC had been unjustly enriched by VAT mistakenly paid by the trusts on fees that their investment managers charged for their services. The CJEU held that the services were exempt from VAT under EU law. The Court of Appeal held (reversing Henderson J) that HMRC were not enriched to the extent that they had subsequently repaid input VAT to the investment managers. It said the following at [37-39]:-

“[37] This means in our view that the focus of a claim for restitution of overpaid tax based on the principle of unjust enrichment has necessarily to recognise and respect the methodology adopted for the recovery of the charge and to assess the claim that HMRC have been unjustly enriched in the light of that process. This requires the court to decide whether there has been unjust enrichment at the end of that process and not merely at the time of the original mistaken payment.

[38] Since the Managers had, by common consent, a directly effective right to treat the services supplied to the claimants as exempt and were entitled to enforce that right retrospectively through the medium of s 80, it must follow that the making of those claims operated as an election to treat those supplies as exempt and the waiver of the right to deduct input tax: see Becker v Finanzamt Münster-Innenstadt (Case 8/81) [1982] ECR 53.

[39] … The Managers are to be placed into the position which they would have been in had the domestic legislation properly implemented the provisions of the Sixth Directive. This would have been that no output tax was payable in any of the accounting periods in question but that no recovery could be made in respect of the £25 input tax. The reversal of the tax position created by VATA 1994 has to be carried out on a global basis … The [Managers] must, as Lewison LJ has said, take the rough with the smooth: see Birmingham Hippodrome Theatre Trust Ltd v Revenue and Customs Comrs [2014] EWCA Civ 684, [2014] STC 2222, [2014] 1 WLR 3867 (at [35]).”

HMRC accordingly argue in this case that the judge was right to recognise at [235] that Ford US would have had no half tax credits if FCE had enforced its EU law rights by not paying ACT. Accordingly, FCE cannot enforce its EU law rights now and claim restitution without giving credit for the treaty half tax credits that HMRC paid to Ford US.

In response, the Claimants rely on the judge’s approach. They submit that ITC CA can be distinguished on the basis that the input tax there was directly and immediately returned to the managers, which was not the case here, and argue that HMRC have not proved what credits they are entitled to claim. More importantly, however, Mr Aaronson relies on his respondents’ notice and makes three separate points of distinction with Pirelli. First, the tax credit in this case was not given, as it was in Pirelli, because of the payment of ACT, but to alleviate the economic double taxation caused by the payment of foreign corporation tax by the EU subsidiaries as explained at [85-89] and [159] of FII CJEU1. Secondly, Pirelli was one of the Hoechst line of cases concerning the unavailability of a group income election. In such a case, section 231 (1) says that a tax credit will not be obtained if a group income election is made. But in this case, the tax credit arises from the payment of foreign corporation tax by the EU subsidiaries and had to be paid regardless of whether ACT was paid on the dividend paid by FCE. It was said, therefore, that the link between the ACT payments and the tax credit in Pirelli was not present here, because the claim has nothing to do with the inability to make a group income election. The third point concerned the difference of legal personality between the shareholders to whom the tax credit was provided, and the UK subsidiary which had paid the ACT.

In reply, Mr Ewart submitted, in effect, that Mr Aaronson was making distinctions without a difference. In this case, the discriminatory treatment was as between foreign dividends and domestic dividends, but the result was the same as in Pirelli, namely that ACT was not chargeable contrary to the expectation of the domestic system. The problem in Pirelli was the same as in this case, namely how do you construe UK legislation when it has to apply to a situation that nobody would have assumed when it was enacted. Lord Nicholls held that there was an unspoken link between the ACT and the tax credit, applying a purposive construction, so that HMRC had to be able to set off the tax credit that would never have been paid in practice, had the unlawful ACT not been paid.

The first question, as it seems to us, is whether the judge was right in [235] to regard the treaty half tax credits as “countervailing financial benefits received by Ford US which would not have been available if the UK tax system had operated in accordance with EU law”. He thought that Ford US would have had no entitlement to the treaty half tax credits “if FCE had not paid the relevant ACT”. That much, as it seems to us, was clear from the agreed statement of facts: “[u]pon the payment of the dividends which attracted those ACT liabilities … [Ford US] were entitled to receive a tax credit pursuant to Article 10(2) of the UK-US double taxation convention … calculated as one half of the ACT paid less …”. The question, however, is whether the judge was right to say that the tax credits would not have been available if the tax system had operated in accordance with EU law. Mr Aaronson says they would have been available anyway, because economically they were the “reciprocal” of the foreign corporation tax paid by FCE’s EU subsidiaries, not of the unlawful ACT, as the reasoning in FII CJEU1 made clear.

In our view, the answer to this question follows from our conclusions on the nature of the unlawfulness determined under the tax issues dealt with above. The relevant unlawfulness is the failure to treat a UK company in receipt of EU dividends on which EU tax has been paid in the same way as a UK company in receipt of domestic dividends on which ACT has been paid. Thus FCE ought to have been in a position in which it could distribute its dividend income (even if derived from EU subsidiaries) to its US parents without paying ACT. Had it done so, however, the agreed facts make clear that the treaty half tax credit would not have been received by Ford US. So, it seems to us that the first question in this case must be answered in the same way as it was in Pirelli. The fact that the reason for the unlawfulness of the payment of ACT is different in the two cases seems to us to be, as HMRC submitted, nothing to the point in the particular circumstances being considered under this issue. We shall, however, return to consider the broader implications of Mr Aaronson’s three points under issue 17 below.

As regards the second aspect of this issue, the judge said at [235] that the credits should, therefore, be taken into account “in calculating the compensation due to FCE unless the fact that they were paid to the parent companies, and not to FCE, prevented this”. He held, as we have said, that this did prevent HMRC claiming credit for the tax credits, relying on both the corporate personality point and the fact that in two respects there was no proper analogy between Pirelli and this case (a group income election was not available to a group with a US parent, and Ford US was, unlike the Pirelli parent not able to complain about breaches of the freedom of establishment provisions in article 49).

It does not seem to us that the appropriate question is really whether the decision on the second issue in Pirelli can be distinguished. It is rather whether the proper restitutionary award as a matter of English law should or should not allow HMRC to take credit for something they would not have had to pay but for the unlawful payment of ACT. In that regard, the question of legal personalities should be considered first. Pirelli allowed HMRC the credit because, as a matter of commercial reality, the loss of the opportunity to make a group income election was the loss of a group fiscal package. That reasoning has no read-across to the facts of this case, where the question is simply one to be determined under the law of restitution. The Court of Appeal in ITC CA did, however, make it clear that the court had to decide “whether there has been unjust enrichment at the end of the process and not merely at the time of the original mistaken payment”. We think that the situation in this case is akin to the situation in ITC CA, because the credit paid to Ford US was paid (or at least credited) immediately the ACT was paid by FCE. ITC CA demonstrates that restitution looks to the enrichment of the recipient, not the loss sustained by the claimant. For that reason, the Claimants cannot, it seems to us, rely on the fact that the credits were given to legal persons distinct from FCE as a reason for denying HMRC the right to point to those credits as reducing the amount by which they were enriched. Just as “the amount of the overcharged tax levied on the managers in ICT [CA] was never more than £75”, the amount here was the ACT paid by FCE less the tax credit granted to Ford US. In these circumstances, whilst we agree with the judge that there are the two important points of distinction between the House of Lords’ reasoning on the second issue in Pirelli and this case, we do not agree that those distinctions are fatal to HMRC’s case.

In our judgment, therefore, HMRC’s appeal on this issue should be allowed and we should declare that HMRC can claim a credit against the repayment to FCE of overpaid ACT in respect of double taxation treaty credits granted to its parents, Ford US.

This issue asks more generally whether HMRC should be able to deduct the tax credit that they have had to allow the shareholders to whom relevant distributions were made from the restitution that HMRC must make to the Claimants in respect of unlawfully levied ACT. The question is similar to that posed under issue 15, but the judge answered it somewhat differently. He held that HMRC’s case that they were entitled to deduct the tax credit from the restitution due to the Claimants was “hopeless”, because HMRC were immediately enriched by the full amount of the relevant receipts of ACT, so that the link to the tax credits had no impact on HMRC’s initial enrichment. The right of the recipient of the dividend to a tax credit derived from section 231, whilst the liability to pay ACT on the UK claimant was to be found in section 14, and the one was in no way conditional on the other. The liability to pay ACT was triggered only by the making of a qualifying distribution. Moreover, HMRC’s simplified model made no allowance for three complexities: (a) the fact that ACT may be paid low down in a chain of companies and the tax credit may arise high up in that chain, (b) there will often be a considerable delay between the payment of the ACT and the allowance of the tax credit (whether by way of repayment or allowance against tax due), and (c) the recipient of the dividend may be outside the UK and therefore not entitled to a tax credit in any event. Finally, the argument was inconsistent with HMRC’s admission that the claimants had a valid Woolwich claim to recover the unreduced amount of the unlawfully levied ACT.

HMRC argue that the judge was wrong to look at the receipt of ACT in isolation, bearing in mind the decision in ITC CA and the House of Lords’ decision in Pirelli at [1], [36-37], and [103-5], and the clear link between the receipt of the ACT and the associated tax credits. Conversely, Mr Aaronson raises the same points in relation to this issue as he did under issue 15. In our judgment, however, as will now appear, the points have more force in relation to this issue.

It is as well to remind ourselves first about the nature of the question. The question, as under issue 15, concerns the proper evaluation, as a matter of English law, of HMRC’s enrichment as a result of the overpayment of ACT. In this case, however, the tax credits were granted to a multitude of shareholders in a large variety of differing positions. Some corporate shareholders will have utilised the dividend plus the tax credits against ACT otherwise due on their own distributions within months of the grant of the credits, while others may never have utilised them. Non-corporate shareholders may have claimed repayment almost immediately (for example, in the case of pension funds) or at a later date, while others will have used the tax credit against income tax otherwise due, normally over the following 9 to 21 months. There is, therefore, force in the judge’s central point that the HMRC was “immediately enriched by the full amount of the relevant receipts of ACT”.

HMRC’s response is to point to the dicta we have already mentioned in Pirelli that make it clear that the central principle of the ACT regime was that part of the corporation tax paid by a company was imputed to its shareholders by giving them an appropriate tax credit, even though the legislation did not itself make the direct connection (see Lord Nicholls in Pirelli at [1]). The link, submit HMRC, between the ACT and the credits is inextricable, as held in Pirelli, and ITC CA makes clear that it does not matter that the credit is given to a different legal person.

In our judgment, HMRC’s case is successfully met by Mr Aaronson’s principal submission on this issue. He submits that there is no proper analogy with Pirelli. In Pirelli, what was lost by the breach of EU law relied upon was the ability to make a group income election that meant that no ACT was payable; it followed inexorably from that that, in such a case, section 231 meant that no credit was obtained. But here the position is different. The vice that EU law seeks to cure is the failure to treat UK companies with EU subsidiaries in the same way as UK companies with UK subsidiaries. If they had been treated equally, the ACT on the distribution up to shareholders would not have been payable because the (EU) subsidiary had paid its (foreign) corporation tax. But if the foreign corporation tax is to be treated as UK ACT, then the UK company to whom the dividend was paid by the EU subsidiary would have obtained an ACT credit, and that credit would have passed on to the shareholders receiving the ultimate distribution. Thus, the ACT paid in this situation is not properly analogous to the ACT paid in Pirelli, where the double taxation tax credit was, as a matter of construction, inextricably linked with the payment of the ACT. Here the tax credit granted to the shareholders is properly to be regarded, as a matter of EU law, as the quid pro quo or what Mr Aaronson called the “reciprocal” for the payment of foreign corporation tax at the bottom of the chain by the EU subsidiaries making the original distribution (see FII CJEU1 at [85-89]).

It should, of course, be noted that this argument depends on EU law, when the question being determined is one of English law. But that is only because one is asking the English law question of whether HMRC were truly enriched at the expense of the claimants, and to answer that question, one has to consider the reason why the credit was given. If the credit would have arisen even without the payment of ACT in the reformulated world of compliance with EU law, then it follows that it cannot be set off against the proper value of the restitution HMRC must make.

This submission, which we hold to be well-founded, does not appear to be reflected in the judge’s reasons, and indeed Mr Ewart commented that it had not been advanced before the judge. It was, however, foreshadowed in Mr Aaronson’s skeleton argument for this appeal and no objection was taken to it being argued before us.

The principal reasons for the judge’s conclusion are set out in his judgment at [277]-[279]. They are addressed to the submission made by HMRC that they “were not enriched at all” by the receipts of ACT because of the obligation to grant corresponding tax credits. The judge was right to regard that as a hopeless submission, assuming that there was always a delay between the payment of the ACT and the use of the tax credit. At the very least, HMRC had the use of the money in the intervening period. In fact, it appears from what we were told that, at least in the case of pension funds (which had significant holdings of shares in those Claimants with publicly-traded shares), payment of the amount of the tax credits would be claimed and made simultaneously, or virtually simultaneously, with the distributions and the payment of ACT.

The judge, however, went further as we read his judgment and rejected HMRC’s case on two main grounds.

First, there was an insufficient connection between the ACT and the tax credits, because they “were the subject of independent statutory provisions, neither of which was made conditional upon the other” [278-279]. While literally true, we do not consider that this gives sufficient weight to the link between the payment of ACT and the grant of tax credits that lay at the heart of corporation tax regime in force between 1972 and 1999, as clearly recognised by the House of Lords in Pirelli. In our judgment, it was not necessary, as the judge thought, that HMRC should be “a mere channel for the transfer of money from companies to their shareholders” [279]. The link between ACT and tax credits was, in practice and in law, so strong that in our judgment it would be right to take account of tax credits in determining the benefit received by HMRC through the unlawful payment of ACT, were it not for the reason given above for rejecting HMRC’s case on this issue.

Secondly, the judge appears to have considered that because HMRC was immediately enriched by the full amount of ACT paid to them, no account is to be taken of any later credit or payment that it had to give in respect of the tax credits, even though the tax credits were granted immediately, and the obligation to give effect to them in due course arose immediately, under the relevant statutory provisions. We do not accept this. The link exists, even though HMRC will in most cases not have to give effect to the credit for some time after its receipt of the ACT. We readily accept that HMRC must make restitution of the time value of the money for the relevant period, but we do not see why HMRC’s enrichment should be calculated on a basis that wholly ignores their obligation to give effect to the credits.

As for the three points raised by the judge at [280], we do not regard them as, individually or collectively, providing grounds for rejecting HMRC’s case in principle. The first point was that the company paying the ACT may not be the company that pays the dividend attracting the tax credit, and the absence in some cases of “a close correspondence in time and amount between the payment of ACT and the grant of tax credits to individual shareholders”. Given the inextricable link between the payment of ACT and the grant of tax credits, we do not consider that this point can defeat HMRC’s case in principle. The second was that there will usually be a period of time between the payment of ACT and the tax credits taking effect. We have earlier addressed this point and while we accept that HMRC should make restitution in respect of the time value of the payment of ACT, it does not follow that HMRC are entitled to no credit at all in respect of the tax credits. The third point was that some shareholders will have been non-resident and therefore not entitled to a tax credit. The effect of this point is to reduce the credit to which HMRC might be entitled, rather than to eliminate it.

Accordingly, we would dismiss HMRC’s appeal on this issue, albeit for different reasons than those given by the judge, and hold that, as a matter of English law, HMRC were enriched by the full amount of the payments of ACT even when they were required to give corresponding tax credits to the recipients of the dividends.

This section deals with the English law “actual benefit” arguments under the “enrichment” heading for reasons already explained. The judge acknowledged at [413] that, though the “actual benefit” arguments bear a close similarity to the defence of change of position, they actually concern the quantification of the government’s enrichment. Vos J had previously observed in Littlewoods HC1 at [129] that there were two potentially significant differences between “actual benefit” and “change of position”. First, the legal burden of proof lies on the claimant to establish the quantum of HMRC’s enrichment, whereas the burden lies on HMRC to establish the defence of change of position. Secondly, the causal link required by the defence of change of position as between the initial enrichment and the subsequent disenrichment does not have to be demonstrated to reduce the restitution on the grounds of a reduced actual benefit. Therefore, even if the answers may often be the same, “actual benefit” arguments are legally distinct from the defence of “change of position”.

Each of the parties placed great reliance on Littlewoods CA, which was decided after the judge’s decision, in relation to both issues under this heading. It is, therefore, useful to consider what that decision actually decided before turning to the individual issues. It will be recalled, of course, that the actual benefit arguments were considered in both Littlewoods HC1 (before Littlewoods CJEU) and in Littlewoods HC2 by Henderson J. Vos J had decided in Littlewoods HC1 that an “exhaustion of benefits defence”, as it was then pleaded by HMRC, was available in principle to the claimants’ mistake-based claims, but not to the claimants’ Woolwich claims. Vos J also concluded at [131] that it was open to HMRC to argue on the assessment of quantum (and to call evidence to support the case) that the government had not benefited from the overpayments of VAT after the first fiscal year, so that no use value should thereafter be awarded. Henderson J decided in Littlewoods HC2 that as a matter of English law the correct approach to quantification was to ascertain the objective use value of the overpaid tax, which was properly reflected in an award of compound interest, and that the claimants were content to receive the use value of the overpaid tax to the government (which was less than the use value to the claimants), and that the actual benefit derived by the government from the overpayments was irrelevant to the objective use value.

Littlewoods CA dealt with the actual benefit issues under issue 6A which asked whether the government’s benefit from the overpayments of tax were correctly measured (a) by the “objective use value” measured by reference to the government’s cost of borrowing, or (b) by reference to the actual use made by the government of the overpayments and the “actual benefit” which government derived from them. The Court of Appeal dealt in detail with Sempra Metals and Benedetti before concluding at [187] that “since HMRC proved to the satisfaction of the judge that the actual benefit it obtained from Littlewoods’ overpayments was less than market value of the time value of that money, actual use value could be relevant to valuing the benefit which the government received as a result of the overpayments”. It, therefore, went on to consider whether the point was in fact relevant, by considering the judge’s findings about the use by the government of the overpayments.

In this context, the claimants in Littlewoods CA accepted that HMRC were innocent recipients, meaning that HMRC were not at fault in receiving the overpayments [192]. The Court of Appeal continued at [193] by saying that “the starting point in principle ought in our judgment to be that an innocent recipient of an overpayment should not have to make restitution of more than he actually received, not knowing it was an overpayment, unless he freely accepted the benefit of having an overpayment and the obligation to pay for it at market rates” (see Lord Clarke in Benedetti at [18]). Thus far, HMRC contend that the decision was correct.

Mr Burrows, who argued this part of the case for HMRC, did, however, take issue with the paragraphs immediately following [195-7], which held that those principles were not applicable to the circumstances of the present case. They read as follows:-

“[195] In the present case, there is no evidence that HMRC actually knew Littlewoods had made what in law were overpayments or could have rejected them. Nonetheless, we do not consider that HMRC should be treated as if it were an involuntary recipient of overpayments of tax. Taxpayers have to pay tax even though they may not have received any assessment from HMRC. In this case there were also assessments. Even if HMRC had no idea at the time that Littlewoods was making overpayments of tax, it still cannot in our judgment be said to be in the position of an involuntary recipient of a benefit. It is obvious that, under a system of self-assessment, tax will from time to time be paid in error and that that tax will have to be repaid. That is an inherent risk of a system of self-assessment. Statistically a certain percentage of tax receipts will have to be repaid, and we consider that government should not be able to discharge its obligations in restitution to the taxpayer by choosing to take a course which would dilute its repayment obligations.

[196] In any event, the government was not an involuntary recipient of the benefit in another sense. The government was well able to make a decision as to whether to use the money. It is not compelled to use the money for government spending or reducing taxation. It was free to use the money for greater financial benefit by repaying borrowings or (subject to any restrictions on government investment) by placing the money on deposit. No doubt it would have repaid government borrowings if government borrowing rates were particularly high. The fact that it did not do so suggests that the rates may have been particularly favourable to government at the time.

[197] In any event, if the money was spent on public projects, the court is entitled to take the view that it was well spent. It is difficult to value the benefit which the government obtains by spending the money for others’ benefit. In those circumstances the court should err in favour of applying the objective use value to the benefits obtained by government. HMRC has not sought to prove that the money was not well spent or that there was some particular reason on the facts why it would be unfair to impose on government the burden of repayment with interest reflecting the savings in government borrowings.”

Mr Burrows described those paragraphs as “opaque”. He contended that the only role of the distinction between voluntary and involuntary receipt was to distinguish between cases where a benefit had been freely accepted and where it had not; and he submitted that it necessarily followed from the statement at the start of [195] – that is, that HMRC were unaware that the payments in question were overpayments – that they could not have been freely accepted. He submitted that none of the factors identified in the rest of the passage justified the contrary conclusion. It was not enough that the Government knew that there were bound to be some overpayments from time to time, which is the point made in the rest of [195]. Nor was it relevant that the Government had a choice how to use the money [196] or that it could be assumed to have been well spent [197]. That did not get over the basic problem that the benefit of the overpayments had not been freely accepted.

The short answer to those submissions is that the paragraphs in question are part of the ratio in Littlewoods CA and we are accordingly bound by them. But in any event we do not think that Mr Burrows’ submissions fairly engage with the Court’s reasoning. As is shown by the preceding paragraphs on which he relies, the Court fully acknowledged, at [193], that the starting point should be that an innocent recipient of an overpayment should not have to make restitution of more than he actually received, not knowing it was an overpayment, unless he freely accepted the benefit of having an overpayment and the obligation to pay for it at market rates; and, further, [194] that “generally” where a person is the involuntary recipient of a benefit – as in Baron Pollock’s example of someone cleaning his shoes without being asked or the modern equivalent of the motorist whose car windows are cleaned while he is waiting at the lights – he cannot be regarded as having freely accepted it. But in the paragraphs with which we are concerned the court was explaining why the case of overpaid taxes justified a rather particular approach to “voluntariness”. That was as a result of a combination of two features. First, while the Government may not know that a particular payment was an overpayment, it is inherent in the system of self-assessment, which is a mechanism for discharging obligations imposed by the state, that overpayments will occur: that is the point made at [195]. Secondly, as a matter essentially of policy, it must be taken that the Government will have applied the money for its own – that is, for the public’s – benefit: that is the point made at [196-7]. That combination of features is peculiar to the case of overpaid taxes, and means that the simple proposition that free acceptance requires knowledge needs some modification in this context: the position of the Government cannot be equated with that of the motorist who has his windows cleaned at the lights. The Court was careful to acknowledge that it was not saying that the Government’s acceptance of these particular overpayments was in the most obvious sense voluntary. Rather, at [195] it says that it should not be “treated as if it were” an involuntary recipient; and at [196] it refers to the receipt not being involuntary “in another sense”.

We respectfully agree with that reasoning. In our view Littlewoods CA was right to conclude that HMRC were in that case to be treated as if they had freely accepted the benefit of the overpaid VAT for the reasons it gave. The court should not, we think, be over-ready to accept that government has not made good use of the tax it collects or that it can escape paying the objective use value of that tax, when it is shown not to have been due. We do not rule out that there may be cases where that would be the result, but they will not be frequent. With that introduction, we turn to deal with the specific actual benefit issues.

This issue relates only to the quantification of the Claimants’ claims for the use value of the utilised ACT. It does not relate to HMRC’s receipt of overpaid corporation tax, because “the availability of money to use is not unequivocally enriching in the same degree as the receipt of money”, as Lord Hope recognised in Sempra Metals at [33].

The judge answered this issue question negatively for three reasons:-

In our view, the judge’s approach to the availability of an actual benefit argument has already been considered by the Court of Appeal in Littlewoods CA. As we have already said, the Court of Appeal concluded in [187] that actual use value could be relevant to valuing the benefit which the government received as a result of the overpayment of tax. Again, as we have already said, we agree with that conclusion. It seems to us that it is a conclusion that is directly applicable in this case, and that we should, therefore, answer this issue by saying that as a matter of law it was open to HMRC to raise an actual benefit argument in respect of the Claimants’ mistake-based claims. Although the Claimants’ skeleton argument formally resists this proposition, we did not understand the Claimants to be arguing strenuously otherwise. They did, however, strenuously resist the appeal on the next issue as to whether the argument was made out on the facts.

The judge concluded at [421-424] that, even if the actual benefit arguments were available to HMRC, they failed on the facts. His reasons were as follows:-

HMRC’s argument relies here on the same arguments that they advance in relation to change of position. We will deal with those arguments under issue 20. HMRC point specifically in this connection to the presumed distribution of government spending between capital, income expenses, and repayment of debt. HMRC rely on the fact that Henderson J put that reasoning specifically to the Claimants’ expert, Dr Andrew Sentance. In essence, HMRC argue that they did show that the government did not benefit in the longer term from the time value of the money because the receipts were spent in precisely the same proportions as all other monies. Thus, the government only benefited from the capital expenses and debt reduction payments. The income expenses were dissipated and that proportion of the time value should not be repayable.

Mr Beazley, who argued this part of the case for the Claimants, relied on Littlewoods CA for the proposition that the government could not contend that it had wasted the money, or not had the benefit of it. The government was to be treated as a voluntary recipient, and cannot discharge its obligations in restitution to the taxpayer by choosing to take a course that dilutes its repayment obligations. It was not, therefore, open to HMRC to say that the tax overpayments were “wasted” on income expenses. HMRC must make restitution to the claimants at no less than the government borrowing rate as a voluntary recipient which was able to choose what to do with the overpayments.

As we have already said, we take the view that we are bound by the approach adumbrated in Littlewoods CA. In that case, the Court of Appeal held that, even though HMRC had not actually known of the overpayments, the government was to be regarded as a voluntary recipient, and, therefore, to have freely accepted the benefit. We can see no basis to distinguish that decision. The circumstances here are the same, even if, as the judge pointed out, the evidence was different. That evidential difference did not affect the fact that, in this case as in Littlewoods CA, the government was aware that ACT would be overpaid from time to time, and that the government could choose how to spend those overpayments and whether or not to repay borrowing. Most importantly, however, here as in Littlewoods CA, we agree that the court should err in favour of applying the objective use value to the benefits obtained by government, whose spending is exclusively for the public benefit.

The judge’s factual findings are not appealed on this or any point. We have seen nothing in the evidence that the judge accepted, which takes this case outside the parameters explained in Littlewoods CA. The fact that the government may have shown that, over the long term, there were fixed proportions spent on income, capital and debt repayment does not mean that the government did not choose how to spend its tax receipts, including those that were overpaid. We do not accept that the income expenses, which make up the vast majority of government spending, were “dissipated” in the way that HMRC submitted. Once it is established that HMRC have freely accepted overpayments of tax, the mistake-based restitution claim must be valued on the basis of the objective use value of the money. Here, that objective use value is accepted by the Claimants to be limited to compound interest at the government borrowing rate.

HMRC made trenchant criticisms of the judge’s detailed reasoning on this issue, but we do not think it necessary to address those specific arguments, having felt bound to follow the approach of a previous Court of Appeal.

We would dismiss the appeal on this issue. The “actual benefit” argument was not, in our judgment, made out on the facts of this case.

THE DEFENCE OF CHANGE OF POSITION IN ENGLISH LAW

This issue is raised by the Claimants’ cross-appeal against the judge’s decision in their respondents’ notice. It is necessary, first, however, to consider whether the argument is open to the Claimants, since HMRC contended below and the judge held that the issue was res judicata having been decided by him in FII HC1 and upheld in FII CA1. For that purpose, a description of the previous decision-making in this case is required.

The judge had originally dealt with the question of the availability of a defence of change of position at [335-342] of FII HC1 which led to declaration 14 in his order dated 12 December 2008 that “[t]he Defendants are entitled to maintain a change of position defence in respect of any mistake-based restitutionary claims which go beyond San Giorgio claims”. The judge’s reasoning can be summarised as follows:-

When FII HC1 reached the Court of Appeal in FII CA1, it addressed the matter under the heading of “[i]ssue 15: [i]s any change of position defence precluded by the wrongdoer principle?”. Its judgment recorded that it had been “common ground that, notwithstanding the apparently absolute terms of [declaration 14], the Judge was not making a final determination in respect of the change of position defence, but rather was acknowledging that, on the facts available at this stage, there is nothing to preclude [HMRC] from raising the defence at trial”. The Court of Appeal continued by recording that the Claimants had nonetheless appealed on the issue, and that the only proper basis for an appeal would have been that the judge should have ordered that “there would be no proper basis for the successful invocation of the defence at trial”. In fact, however, the Claimants were simply arguing that the judge had misstated the applicable law and should not have formed a preliminary view that the defence was likely to succeed. The Court of Appeal held that that was not a sound basis for an appeal. It then held that the defence of change of position did not anyway call for consideration, since it was common ground that the EU law principle of effectiveness precluded the application of any change of position defence to San Giorgio claims, though all the San Giorgio claims in the case fell within the Woolwich cause of action. The Court of Appeal said that the mistake-based claims were, as they had held, limited by section 33 of the Taxes Management Act 1970, section 320 of the Finance Act 2004, and section 107 of the Finance Act 2007 (a decision later held to be contrary to EU law in FII CJEU3). The court then declined to deal with the parties’ substantive submissions on change of position, as the defence was “highly fact-sensitive”, and anything said would be obiter, and raise important and difficult questions of law and policy.

Importantly, however, the Court of Appeal in FII CA1 actually amended the judge’s declaration 14 to read: “[t]he Defendants are in principle entitled to maintain a change of position defence in respect of any mistake-based restitutionary claims which go beyond San Giorgio claims” (emphasis added), by adding the words in bold.

The judge then considered the question again in FII HC2 at [316-341]. He recorded in [317] that he had held in FII HC1 at [339] that “the defence should in principle be available to the Revenue when restitution is sought of overpaid tax on the basis of mistake”, and that he had distinguished mistake-based claims from Woolwich claims because “the unlawfulness of the tax [formed] no part of the cause of action in mistake”. Though his discussion had been confined to restitution claims which went beyond San Giorgio claims, the position under English law could not, he held, depend on whether the defence was precluded by EU law. Thus the judge accepted that his reasoning in relation to mistake-based claims which were not San Giorgio claims must extend to the mistake-based claims which were San Giorgio claims. By implication, the Court of Appeal’s amended declaration 14 had to be taken to apply to all mistake-based claims. Moreover, the Claimants had not resisted these propositions in FII HC1, and there was no suggestion that this part of his decision in FII HC1 was of a preliminary nature. The question was one of law, and the judge thought his conclusion was unqualified. He therefore concluded in [318] and [339] that he had already decided the question in HMRC’s favour in FII HC1 and that he was “functus officio” and lacked jurisdiction “to revisit the question in the same proceedings”. It was anyway inappropriate to do so. The judge nonetheless thought that the point was in urgent need of an appellate decision.

Against this background, the first question for us is whether we are bound by FII CA1 (at [189-193], upholding Henderson J’s decision at [335-342] in FII HC1) to decide this issue 19 in HMRC’s favour. HMRC submitted that we were broadly for the reasons that the judge gave.

Mr Beazley submitted that this court was not so bound because:-

In our view, the judge clearly decided, as a matter of principle, in FII HC1 that a defence of change of position was available to HMRC in this case (see [336] of FII HC1). He gave the reasons that we have already set out. That decision was appealed by the Claimants. It is not now open to them to say that it was only appealed on a limited point. It is quite easy to see why the Court of Appeal headed the issue in FII CA1 by asking whether any change of position defence was precluded by the wrongdoer principle. That had been the thrust of the judge’s reasoning on the point in his judgment [337-339].

When it came to FII CA1, the Court of Appeal recorded that the claimants had appealed declaration 14 on the basis that “the Judge misstated the applicable law and should not have formed a preliminary view that the defence was likely to succeed”. The appeal was rejected because the Court of Appeal did not regard that as a sound basis for an appeal in the light of the judge’s order (which it held had simply been determining the matter “in principle” not finally).

It can be seen, therefore, that it would have been open to the Claimants in FII CA1 to have argued that the judge was wrong in law to have decided that the defence of change of position was available even “in principle”. But that was not the appeal that the Claimants mounted. In our judgment, they cannot have two bites at the cherry. If the Claimants wished to re-open this question in this case, they should have appealed the amended declaration made in FII CA1 to the Supreme Court.

In these circumstances, the judge was right to think that the question of whether the “change of position” defence to the Claimants’ mistake-based claims was available in principle to HMRC in this case was res judicata.

Although the substance of this issue was argued orally and in writing, we are unwilling to express our views on it. We acknowledge that the judge has said that the point is in need of an appellate decision, but we do not think that decision can be in this case unless it were still open to the Claimants to appeal FII CA1 to the Supreme Court. We prefer to say nothing about that, because we are aware that some issues have been “parked” when the matter last went to the Supreme Court. But either way, in our judgment, any appeal on the substance of this issue would be an appeal from FII CA1 and not from us. In any event, the arguments on either side are summarised succinctly by the judge in [319-338] of his judgment.

On this issue, HMRC attack the conclusion (at [399]) that a change of position defence was not available to HMRC on the facts that the judge found. There is no challenge to any of those factual findings. HMRC say, in essence, that the judge misapplied the law on the change of position defence to a restitutionary mistake-based claim to the facts that he found. HMRC submit that the evidence did show a long-term link between tax receipts and government expenditure and that HMRC should, therefore, have succeeded in their defence insofar as the proportion of that expenditure was in the nature of income, rather than capital or debt repayment expenditure. The defence was available generically for the government’s change of position resulting from its enrichment by all ACT paid by taxpayers, whether Claimants or not, as a result of the particular mistake relied upon. The defence is available in relation to mistakenly paid corporation tax, and also mistakenly paid ACT, which was later utilised (by far the largest amount as a matter of quantum).

It is necessary to summarise the judge’s decision in a little detail, even though HMRC’s argument is relatively straightforward as we have already explained.

Before the judge, it was common ground that the burden of proving the defence of change of position was on HMRC, that the defence could operate proportionately (i.e. in part or to the extent that it is made good), and that the law is correctly stated in section 23.1 of Mr Burrows’ Restatement of the Law of Restitution to the effect that: “[t]he defendant has a defence to the extent that:- (a) the defendant’s position has changed as a consequence of, or in anticipatory reliance on, obtaining the benefit, and (b) the change is such that the defendant would be worse off by making restitution than if the defendant had not obtained, or relied in anticipation on obtaining, the benefit”. The recipient’s loss is a consequence of obtaining the benefit if it is causally relevant, and though there is no relevant test for “causal relevance” in the cases, the relevant starting point is a “but for” test (see Scottish Equitable plc v. Derby [2001] EWCA Civ 369 per Robert Walker LJ at [30-31], where he had also endorsed the “wide view” of the defence). The judge rejected the Claimants’ submission that the government had a heavier burden to establish the defence, once it was held that it was available.

The judge clarified a number of preliminary points, all of which HMRC accepted before us. It is sufficient for us to mention only his third and fourth such points. The third was to the effect that it was not enough for a defendant to show merely that money had been spent. The expenditure had to be “extraordinary” in the sense only that it would not have been incurred but for the overpayment. The fourth point was to accept that the defence is essentially concerned with disenrichment, and that it may in that context be relevant to consider whether the relevant expenditure was reversible. The judge thought, however, that it would be wrong to elevate that consideration into a general test of irretrievability. The judge approached the matter on the basis that HMRC acted throughout in good faith.

The judge set out the scale of the relevant overpayments at paragraph 358, commenting that it was obvious that overpayments of that order of magnitude were “only a miniscule proportion of total government receipts”, representing the equivalent of only between two and five pence per week out of the average household budget.

The judge recorded at [365] the agreement between the experts as follows:-

“(a)

It is not possible to know with certainty either how the Government deployed the overpayments in question, or what it would have done had the overpayments not occurred.

(b)

The UK Government does not normally hypothecate revenue to particular uses and did not hypothecate the overpayments to a particular use.

The judge then considered the evidence of the three experts in detail at [366-392]. In short, he accepted the evidence of Dr Andrew Sentance [378-9] for the Claimants and thought that the evidence of Sir Jonathan Stephens for HMRC was undermined by some important answers that he gave in cross-examination.

The judge explained HMRC’s approach which was to say that all ACT was used to fund spending or reduce taxes, and was irretrievable unless spent on capital assets or to reduce government debt. Since the payments could not be traced, one could ascertain in any year the proportion that was irretrievable by looking at the breakdown of government expenditure as a whole as between capital spending, current spending, and reduction of debt (which Professor Myles undertook).

The point that undermined Sir Jonathan Stephens’ evidence was his express acceptance that the government might allow borrowing to change in the light of developments in receipts or spending, and that tax receipts and spending therefore did not need to move together (which was reflected in paragraphs 1(d) to (f) of the experts’ joint statement). In cross-examination, Sir Jonathan accepted that borrowing would normally take the strain in the short-term as a response to an unexpected reduction in taxation revenue. The default position was that borrowing would always take at least some of the strain. Having dealt in detail with Sir Jonathan’s evidence on the ACT system, the judge concluded at [391] that he was “left with no credible evidence about how the cash flow aspect of ACT was taken into account in government forecasts of receipts and expenditure”.

The judge’s conclusions are at [393-9] as follows:-

“… The methodology propounded by [HMRC’s experts] depends on the supposed existence of a fixed link between taxation revenue and government expenditure, such that any increase or diminution in forecast receipts is automatically reflected in a corresponding increase or diminution in government expenditure within the same fiscal year.

… The fundamental problem with this analysis, however, is that it depends on assumptions which all the experts now agree to be unsustainable. The only firm correlation between taxation receipts and government expenditure is a long-term one, and even then it has not been demonstrated by [HMRC] that there is a direct causal relationship between receipts and expenditure (as opposed to a link of a multi-factorial nature). In the short-term, the experts agree that “tax receipts, spending and borrowing can all fluctuate significantly and there are often variations from forecasts and plans” (paragraph 1(f) of the joint report). The correctness of this view was abundantly confirmed by the evidence which I heard from both Sir Jonathan and Dr Sentance. In particular, Sir Jonathan expressly accepted that, in the short-term, borrowing often takes the strain when tax receipts are lower than forecast.

… The precise extent of the “short-term” may be open to debate, but on any view it includes at least a period of two to three years.

… The model espoused by [HMRC] in their expert evidence also strikes me as implausibly rigid and unrealistic when regard is had to the detailed evidence about matters such as: (a) the timetable within which government spending decisions are taken; (b) the existence of political or economic pressures which may cause even firm spending targets to be departed from; (c) the inherent imprecision of forecasts, and the extent to which they were regularly not met in practice (forecasting errors are agreed to be of the order of 1% of GDP); and (d) the additional element of flexibility provided by the contingency reserve.

… Furthermore, even if [HMRC’s] methodology were otherwise reliable, I would not be satisfied on the evidence before me that it could be applied to the payments of utilised ACT. … The benefit to the government in such cases is of a cash flow nature, but I have no reliable evidence of the way in which this was reflected in government forecasts. Indeed, for all I know it may be the case that all receipts of ACT were routinely left out of account when planning future government expenditure … I consider, therefore, that a full explanation of the way in which payments of ACT were in fact taken into account in government forecasts would have been a prerequisite for any successful establishment of the defence on the facts in relation to the ACT claims, and that the absence of such evidence is in itself a fatal flaw in the Revenue’s case.

… I consider that the quantification model propounded by the Revenue is clearly unsound, both conceptually and on the facts”.

HMRC’s basic position on this issue is that the mistaken payments resulted in equivalent extra expenditure, so that the Government would be worse off by having to make restitution. They support this by relying on Sir Jonathan’s evidence that the planned overall level of borrowing is set first independently of expected tax revenues of this magnitude. Thus, when deciding on tax and spending, lower planned receipts are reflected in lower planned expenditure or higher taxes. Consequently, the mistaken payments would have been factored in to the Government’s plans for tax and spending. They would not have been “lost in the rounding”. Any excess tax payments were also used to increase spending rather than reduce borrowing. The judge was entitled to hold as a matter of fact that there was no link between taxation revenue and government expenditure in the short term, but was wrong to misunderstand HMRC’s case that the desired fiscal position is fixed before spending and tax plans are drawn up, so lower receipts result in lower spending or higher taxes, and the lack of a short-term relationship between tax and spending is not contrary to HMRC’s case. In any event, there was a long-term relationship between tax receipts and spending, which was enough for HMRC’s case. HMRC should not have to make restitution of the proportion of the vast majority of the overpayments that went towards increasing spending or reducing taxes, as opposed to capital investment or repayment of borrowing.

In oral argument, Mr Baldry, who argued this part of the case for HMRC, took us to three cases. First, he referred to Lord Goff’s famous passage in Lipkin Gorman v Karpnale Ltd [1991] 2 AC 548, at page 580F-G. Lord Goff said that he did not wish to state the principle any less broadly than “that the defence is available to a person whose position has so changed that it would be inequitable in all the circumstances to require him to make restitution, or alternatively to make restitution in full”, though the mere fact that the defendant has spent the money is not enough to make it inequitable that he should repay it, as the expenditure might in any event have been incurred in the ordinary course of things.

Mr Baldry then referred to Philip Collins Ltd. v. Davis [2000] 3 All ER 808 at pages 826-830, and Scottish Equitable plc v. Derby per Robert Walker LJ at paragraphs 30-31 and 33-35. The latter was referred to by the judge, as we have mentioned, but it is worth noting that in that case, Robert Walker LJ expressly approved Jonathan Parker J’s statement in Philip Collins Ltd (which concerned over-payments of royalties to members of a rock band) to the effect that it may be right for the court not to apply too demanding a standard of proof when an honest defendant says that he has overspent by improving his lifestyle, but cannot produce any detailed accounting. The Claimants, however, suggested that cases about personal overspending were hardly pertinent to issues of government spending and budgeting. Subject to that caveat, Philip Collins Ltd showed that a change of position defence could in some circumstances succeed, even where the musicians in question could not show that any of their specific additional expenditure was directly referable to overpayment of royalties. HMRC submitted that there was generally no need for such a specific connection and that the defence was available wherever the defendant had geared its expenditure to its income.

HMRC relied on two specific pieces of evidence: the judge’s finding at [365] that the overpayments were fully factored into the government’s spending plans, and the agreed expert evidence that “in the long-run, tax receipts and spending are related to each other” (see paragraph 291 above). HMRC pointed to three errors in the judge’s reasoning. They submitted that the judge had actually found the necessary long-term relationship between receipts and spending in [394] where he said that “[t]he only firm correlation between taxation receipts and government expenditure is a long-term one, and even then it has not been demonstrated by [HMRC] that there is a direct causal relationship between receipts and expenditure (as opposed to a link of a multi-factorial nature)”, but had not recognised it. The judge then, HMRC submitted, wrongly introduced the need for a direct causal relationship between receipts and expenditure. His treatment at [395] of the number of years it might take to establish a “statistically significant return towards equilibrium” misunderstood that it did not matter whether the relationship was short term or long term provided it was shown to exist. Professor Myles had shown that there was a long term 1:1 connection, and Dr Sentance had accepted that those calculations were valid. In the meantime, all that had happened was that extra borrowing had taken the strain.

Mr Beazley submitted that the key to a change of position defence was that it had to be inequitable or unjust to grant restitution. There was no such injustice here. Indeed, for the bulk of the claim, there was no basis for a change of position defence since there was no evidence as to how the government had used the short term payments of ACT that were “utilised” often only months later. The benefit here was freely accepted by the government. The burden was on HMRC to establish their defence pro tanto, and they failed to show what the quantum of their claim actually was.

The Claimants also sought to argue that the long term linkage between receipts and expenditure was neither pleaded nor seriously advanced at trial, but this argument seems to us to be rather flimsy. Whilst HMRC’s experts concentrated on establishing the short term linkage, they did not deny the possibility of a longer term linkage.

Mr Beazley took issue with the causation requirement for the defence of change of position that Vos J had explained in Littlewoods HC1 at [129] as one of the distinctions between a change of position defence and an actual benefit argument. He submitted that that decision was wrong, and that the judge had been wrong to accept it at [413-414] in this case. The government could not be compared to the rock band in the Philip Collins Ltd case, and a direct connection had to be shown between the enrichment and the expenditure to make good the defence. He submitted that [195-197] of Littlewoods CA were crucial to change of position as well as actual benefit.

On the facts, Mr Beazley submitted that Sir Jonathan’s simplistic picture that ACT was taken into account in deciding spending was rejected lock stock and barrel by the judge in [387], [391], [393], and [394]: no credible evidence of linkage was found. It was “just nonsense” that small payments of the kind in issue here affected the government’s short or medium term spending. But even if that were the case, it did not apply to utilised ACT which is the lion’s share of the claim (see [397]). The government had simply failed to adduce anything by way of credible evidence.

Finally, Mr Beazley relied on the passage in Goff & Jones on the Law of Unjust Enrichment, 8^th edition, 2011, edited by Mr Charles Mitchell, expressing the strong view at paragraph 27-51 that change of position should not be allowed as a defence to mistake-based claims for the recovery of tax in English law:-

“… it is submitted that in principle the better view is that public bodies should never be allowed to plead change of position in response to a claim to recover money paid as tax, whatever the ground on which the claim rests. The reason is that allowing them the defence would seriously undermine the constitutional principle that taxation must not be levied without Parliamentary authority, and the wider principle that public bodies are constrained by the rule of law. It would also unfairly cast the burden of paying for the government's unlawful act onto an innocent taxpayer when this should properly be borne by the public at large. These are the reasons given in Canada for rejecting a defence of fiscal chaos, and in Germany for denying public bodies the change of position defence. The English courts would do well to follow suit …”

We should say at once that, it having been decided that the defence of change of position is available to HMRC in this case, we cannot support Mr Mitchell’s contrary views in Goff & Jones. We do, however, accept Mr Beazley’s submission that [195-197] of Littlewoods CA (which we have already set out at paragraph 260 above) are relevant to the defence of change of position in this case. Those paragraphs draw attention to the injustice of denying the taxpayer restitution of the objective use value of the overpaid tax, when government has freely accepted the benefit and made use of the money for the public benefit. They are relevant because injustice or inequity is at the heart of the defence of change of position as Lord Goff’s famous dictum in Lipkin Gorman explains (page 580). To make good the defence on the facts, HMRC have to persuade the court that it would be inequitable that they should be called upon to repay the overpaid corporation tax and to make good the objective use value of the prematurely paid ACT. It may be noted at the outset that, although the judge does not appear to have made a direct decision on whether it would be inequitable to require HMRC to make restitution, it may be clearly inferred from what he has said at [342-399] that his answer would have been that it would not.

The second point we would make is that, whatever may be the position in relation to overpaid corporation tax, the position on utilised ACT is in our judgment clear. The judge makes clear factual findings at [397] of his judgment which HMRC simply cannot undermine, having not appealed the facts he found. Those findings were that he was not “satisfied on the evidence before me that [HMRC’s methodology] could be applied to the payments of utilised ACT” because the benefit to the government was of a cash flow nature, and the judge had “no reliable evidence of the way in which this was reflected in government forecasts”. The judge continued by saying that “for all I know it may be … that all receipts of ACT were routinely left out of account when planning future government expenditure”. The judge’s view was that “a full explanation of the way in which payments of ACT were in fact taken into account in government forecasts would have been a prerequisite for any successful establishment of the defence on the facts in relation to the ACT claims, and that the absence of such evidence is in itself a fatal flaw in [HMRC’s] case”.

We turn then to deal with the question of whether HMRC were right to say that, on the judge’s findings, they had established a long term causative connection between overpaid corporation tax and government expenditure and that they should at least be relieved of the requirement to make restitution in respect of that proportion (the vast majority in fact) of the expenditure that was represented by increased spending or reduced taxes (as opposed to capital expenditure or borrowing repayment).

As we have already said, we think these arguments were open to HMRC. We do not, however, accept that the judge’s factual findings were sufficient to support the change of position defence even in relation to overpaid corporation tax, as opposed to prematurely paid ACT. There are three reasons.

First, HMRC places great reliance on two lines in [394] of the judge’s judgment. They say that he decided that “[t]he only firm correlation between taxation receipts and government expenditure is a long-term one”, and that there was a causative link of a “multi-factorial nature” between receipts and expenditure. These findings, if that is what they were, are a very tenuous basis for the defence that HMRC seek to run. The judge plainly thought that many other factors came into the mix, and one cannot help thinking that he would have found considerable force in Mr Beazley’s submission that it was “just nonsense” to think that HMRC had proved that these small overpayments of corporation tax had affected spending or tax rates even in the longer term. Vos J reached a similar conclusion on different evidence in Littlewoods HC1 at [112-125], where he said at [124] that “the change of position defence fails because [HMRC] have failed to prove, the burden being on them, that they increased their spending in any way on the basis of the expectation or the happening of the overpayment”, accepting Professor John Kay’s evidence that “had the Government not expected to receive these sums by way of VAT, it would not have changed its budgetary or spending decisions”.

As regards causation, we do not accept that government can be regarded in the same way as the members of the rock band in Philip Collins Ltd. The government needs to show, as a starting point, that “but for” these particular receipts, expenditure would not have been made or taxes would not have been reduced. HMRC were miles away from demonstrating that to be the case here. The reason why a proper link must be shown is because, otherwise, as it seems to us, it is very unlikely to be possible to say that it is inequitable for the government to make restitution. It may be difficult to think of plausible examples, but if, say, an unlawful windfall tax were imposed on the oil industry, and the government had announced that the monies from that tax were to be used to cut stamp duty on house purchases by 1%, and the figures showed that to have been done, one might imagine that the defence could achieve some airspeed. The facts found by the judge did not come close, in our judgment, to demonstrating a real link between these modest overpayments of ACT and government expenditure, even in the medium to long term. The link was “multi-factorial” which is simply not sufficient to make it unjust to require restitution.

Our third reason comes back to Littlewoods CA. There, the Court of Appeal made clear that the court should err in favour of applying objective use value to benefits obtained by government. We think that the court should normally err also in favour of awarding restitution to taxpayers who have overpaid tax as a result of a mistaken view of the facts or the law. That is not depreciating the existence of the defence or the fact that change of position plays a fundamental part in the law of unjust enrichment, enabling it to develop broadly. But if the government is to succeed in mounting a change of position defence, it must establish that it is truly inequitable for it to have to make restitution. In our judgment, that was not established here, and could not have been established on the basis of the kind and quality of the evidence that HMRC adduced. At the risk of repetition, there was simply no satisfactory evidence that, if the government had not had the modest levels of overpaid corporation tax and prematurely paid ACT, it would have cut back on expenditure rather than, for example, repaying borrowing.

In our judgment, HMRC’s appeal on this issue should be dismissed. HMRC failed to make out the defence of change of position on the facts, and failed to show that the judge had misapplied the law to the facts that he found (which were not themselves appealed). Apart from our three generally applicable reasons in [309-311] above, the position may be summarised as regards each of the elements of BAT’s claim that the judge described in [358].  In respect of the claim for utilised ACT (totalling about £228.2 million), the change of position defence failed because there was no short term correlation established between tax receipts and government expenditure, and there was no evidence of the way that any short-term cash flow advantage was reflected in government forecasts (see [397] of the judge’s judgment). In respect of the claim for unutilised ACT (totalling about £95.3 million), the change of position defence failed because no direct link was established between tax receipts and government expenditure even over the long term.  The general relationship of a multi-factorial nature was not sufficient to establish the defence.  Moreover, there was a fatal absence of evidence of how any ACT (utilised or not) was taken into account in government forecasts.  In respect of the claim for overpaid Case V tax (totalling about £700,000), the change of position defence failed for the same reasons as the claim for unutilised ACT, and also because of the relatively modest amounts involved.

DOES EU LAW PRECLUDE RELIANCE ON ANY OR ALL OF THE ANSWERS CONSIDERED UNDER (B) AND (C) (ISSUES 19, 21 AND 24)?

Despite the fact that we have found that the enrichment arguments and the defences of change of position fail as a matter of English law, we should deal with the important question of whether they are anyway precluded by EU law (issues 18, 21 and 24). It is to be noted that it was not suggested that EU law was an answer to the set-off that we have held should have been allowed to HMRC under issue 15, because, of course, the tax credits in question were paid to non-EU parents. We can, therefore, ignore issue 15 for the purposes of this discussion.

The judge’s answer was that the enrichment arguments and the defence of change of position were anyway precluded by EU law (issues 18, 21 and 24). This was in each case because, as we have already intimated, he thought that the EU law principles of equivalence and effectiveness required that the mistake-based claims as well as the Woolwich claims should be available to give full effect to the claimants’ EU law San Giorgio rights to be compensated for having overpaid tax. We have already dealt with the main authorities that provide a backdrop to these arguments under head A above.

Before dealing with the parties’ arguments on this point, we should first explain the judge’s reasoning. He dealt with it in various places in his judgment, predominantly (as we have said) [247-249], [253-263], [288-289], [293-296], [303-308], [400-407], and [425-430]. It is useful to look at these paragraphs together despite the fact that they are notionally addressing different specific issues.

In his introduction to the remedies section of his judgment, Henderson J stated first at [247] that he was satisfied that substantially all the claims that he was dealing with were comprehended, as a matter of EU law, within the San Giorgio principle. He said that it was elementary that EU law requires the national court to provide the Claimants with an effective restitutionary remedy to vindicate their rights under EU law. He recorded that HMRC accepted that the claimants had good Woolwich claims to recover unlawful Case V tax, unutilised ACT, and the time value of prematurely levied ACT, measured in accordance with Sempra Metals principles, but that all these claims were subject to the 6-year limitation period.

Later in the introduction, the judge recited what he regarded as the consistent jurisprudence of the CJEU to the effect that a claimant with a San Giorgio claim to recover unlawfully levied tax is entitled to recover “reimbursement not only of the tax unduly levied but also of the amounts paid to that State or retained by it which relate directly to that tax … and losses constituted by the unavailability of sums of money as a result of a tax being levied prematurely” ([87] and [88] of Hoechst and [205] of FII CJEU1). He referred to Littlewoods CJEU at paragraph 29 saying that the EU principle of effectiveness required, in relation to interest, that the national rules “should not lead to depriving the taxpayer of an adequate indemnity for the loss occasioned through the undue payment of VAT”.

The judge then analysed the character of the San Giorgio right in some detail at [259-263]. None of this analysis is substantively challenged, pointing as it does to the need for repayment regardless of whether the state has been correspondingly enriched. HMRC’s point is simply that the San Giorgio right in this case is adequately vindicated by the availability of the Woolwich claims.

When he came to deal with issue 18 at [288-289], the judge proceeded on an unspoken assumption that the mistake-based claims were necessary to vindicate the Claimants’ San Giorgio rights. He said that he had already explained in Portfolio Dividends HC2 at [171-178] and Littlewoods HC2 at [418-425] why he regarded it as “settled EU law that the only substantive defence which can be recognised to a San Giorgio claim is that of unjust enrichment of the taxpayer”. It can be commented that those particular passages make the powerful point that the CJEU’s decision in Lady & Kid emphasises that the only defence to a San Giorgio claim is that of “passing on”, but do not explain why a mistake-based claim is needed in addition to a Woolwich claim to vindicate the right in the first place.

The judge then returned to the point in his introduction to the change of position defence at [293-296] and [303-308]. He alluded first to FII HC1, in which he had held that “the defence was precluded in relation to the claimants’ San Giorgio claims, and that this was so whichever domestic restitutionary remedy the claimants relied upon to vindicate those claims”, explaining that that was “in particular because (as then generally understood) it did not extend to taxes (like ACT) which were not the subject of a demand”.

The judge then explained how FII CA1 had held at [152-174] that the Woolwich cause of action alone provided the Claimants with a sufficient UK remedy for their San Giorgio claims, since it did not depend on the making of a demand. The absence of a need for a demand had been upheld by FII SC at [79], and [171-174]. Having summarised the effect of FII SC and FII CJEU3, the judge summarised the position at the start of the trial as follows:-

“In relation to mistake-based claims which are brought to enforce San Giorgio rights under EU law, two important questions arise:

a)

Is the defence precluded as a matter of English law, for the same or similar reasons as it is admittedly precluded in relation to Woolwich claims? and

b)

If the defence is not precluded under English law, is reliance on it nevertheless precluded by EU law?”

The judge then said at [308] that he had heard argument on those questions in Portfolio Dividends HC2, but only decided the EU law point, holding at [171-189] that “reliance on the defence was precluded by EU law in relation to all claims which are brought in a member state to vindicate San Giorgio rights”. As mentioned at paragraph 222 above, Henderson J had indeed addressed the point in Portfolio Dividends HC2 at [180-184], but he had done so at a time before the CJEU in FII CJEU3 had delivered its judgment.

In dealing with issue 21 at [400-407], Henderson J referred back again to his conclusion in Portfolio Dividends HC2 at [178-189]. He referred again to the judgment in Lady & Kid which he held to be “of particular strength”, and to other CJEU decisions to similar effect. He dealt also with FII CJEU3 saying in [403] that it had made “no reference to the principle [enunciated in Lady & Kid], perhaps unsurprisingly because it was not directly engaged by the questions referred, but I can detect nothing in the judgment which would cast any doubt on the principle as restated by the Advocate General, or on the slightly different language which he used to justify it in paragraph 74 of his opinion: ‘the right to a refund of taxes levied in a member state in breach of EU law is the consequence and complement of the rights conferred on individuals by provisions of EU law prohibiting such taxes’”. The judge concluded at [405-406] that it was “abundantly clear from the whole of the judgment [in FII CJEU3] that it regarded the two remedies available under English common law as alternative remedies for the recovery of taxes levied in breach of EU law” (see [2-6], [24], [25], [39], [41], [43] and [46-49]). The judge then addressed HMRC’s argument and rejected it as “plainly wrong” for the reasons previously given: “for as long as English law permits claimants to choose between the two causes of action in order to recover tax unlawfully levied under EU law, both remedies must be taken to rank equally in the eyes of EU law, and claims of each type must be regarded as San Giorgio claims”.

In dealing with issue 24, the judge assumed at [426] that the mistake-based claims were properly characterised as San Giorgio claims. He then held that he should follow Littlewoods CJEU in holding that compound interest was required under EU law to remedy the claims relating to prematurely paid ACT.

HMRC have addressed to us an argument similar to the one addressed to Henderson J in Portfolio Dividends HC2. It is an argument that deals with all three of the issues 18, 21 and 24. HMRC acknowledge that there is a right in English law to repayment of unlawfully charged tax under the Woolwich principle, and that there can be no exception to that remedy (applying Lady & Kid). They submit that it is for the domestic legal system to lay down detailed procedural rules and forms of action. The EU treaties were never intended, they say, to create new remedies, unless no legal remedies existed at all. If there is such a remedy, EU law has nothing to say about other domestic remedies. EU law does not require the alteration of domestic remedies. Since the UK legal system provides the Woolwich remedy, which makes it possible for the Claimants’ individual rights to be protected, there is no need to reform or alter the mistake-based remedy.

In relation to Lord Reed’s reasoning at [212] in FII SC, Mr Ewart argued that Lord Reed had started from the principle of equivalence saying that, where domestic law offered two remedies, that principle required that both “should also be available in similar circumstances to enforce an analogous right under EU law”. Lord Reed had concluded that “so long as they must both be available, they must both be effective”. Mr Ewart submitted that Lord Reed had said nothing about the nature of the two remedies. The mistake-based claims are available to give effect to the San Giorgio rights, but the domestic defences apply and those defences do not require moulding or reformation. Mr Ewart also revisited these submissions in a series of written propositions, which we do not think we need to record expressly in this judgment.

In response, Mr Aaronson supported the reasoning of the judge, and relied particularly on [92-103] of Littlewoods CA. The Court of Appeal said in [93] that “it was now tolerably clear that EU law requires national law to reimburse the losses occasioned by the unavailability of money as a result of tax being levied unlawfully”. It said at [103] that “the EU law right is, in the terms in which it is expressed in the case law, a private or personal right of the taxpayer. National law must give effect to that right, and it is no answer to the individual taxpayer’s claim that national law has done so for other taxpayers, or even for the vast majority of them”. He relied also on [130], which we have already referred to at paragraph 226 above. Even if Littlewoods CA were not conclusive, the whole thrust was to follow FII CJEU1 and Hoecsht. The San Giorgio right is simply not vindicated by a Woolwich claim. Both claims have to be moulded to vindicate the San Giorgio rights. Since HMRC accept that none of the three arguments can succeed against a Woolwich claim, they cannot succeed against mistake-based claims either. That is clear from Lord Reed’s judgment in FII SC. Lady & Kid at [20] and [25] made it unequivocal that the only defence to claim for reimbursement of tax unlawfully levied is passing on. In Zuckerfabrik Julich AG v British Sugar (cases C113-10, C147-10 and C234-10) the question was whether EU law precluded the claimant from recovering interest on overpayments to the Rural Payments Agency where the Agency had recovered no interest itself when it was repaid by the European Commission. The CJEU made it clear at [64-67] that EU law required the RPA to repay interest, even though it had not itself been repaid with interest. The claimant’s San Giorgio right entitled the claimant to interest. The decision is informative as to the nature of the San Giorgio right.

In the course of argument, we raised the question of whether it would be appropriate to make a fourth reference to the CJEU on this point. Neither side was enthusiastic about the prospect. It was suggested that we should leave it to the Supreme Court (if the case goes that far) to make any such reference. For the reasons we now give, we do not think a further reference is necessary.

It should first be noted that, in the light of the decisions we have already reached on the enrichment arguments and on change of position, our decision on these issues 18, 21 and 24 does not affect the outcome of this appeal. The arguments and defences have failed in any event.

Nonetheless, since we have reached a clear conclusion, we think we should state it. We take the view that the Supreme Court in FII SC has decided the point in such a way that makes it inappropriate for us to depart from what the majority there concluded. We will return to this point in a moment. We also think that Littlewoods CA has decided the point.

We have considered the judge’s reasoning in detail because we do not think that he dealt with HMRC’s point very directly. The closest he got was in [405-406] where he said that the argument was plainly wrong. We agree with HMRC that Lady & Kid is not conclusive on this point, since all that case makes clear (as a line of other CJEU authorities do also) is that the only defence to a domestic claim to vindicate a San Giorgio right is a defence of “passing on”. It says nothing about whether each of two domestic remedies that could vindicate a San Giorgio right must be moulded so as to enable them to do so. Mr Ewart pointed out that all FII CJEU3 decided was that if such a remedy exists it cannot be amended retrospectively or without proper notice so as to curtail a limitation period. That case also says nothing about whether an existing right must be moulded to vindicate the San Giorgio right when another domestic remedy is already available to do so.

We agree with Henderson J in Portfolio Dividends HC2 that Advocate General Wathelet in FII CJEU3 clearly did support the proposition that both remedies had to be available to vindicate the San Giorgio claim. At [47], [48] and [53], he said:-

“47.

If, however, in application of the principle of procedural autonomy, a member state makes a number of legal remedies available to individuals, the second subparagraph of Article 19(1) TEU requires that each of those remedies ensure effective legal protection, and a legal remedy cannot offer "effective" protection unless the conditions in accordance with which it may be used and achieve a positive outcome are known in advance.

48.

Accordingly, as soon as taxpayers choose one of the national legal remedies available under national law (in the present case, the Kleinwort Benson remedy) or have recourse to the only national legal remedy available, they must come under the protection offered by the general principles of EU law.

…

53.

... the guarantees attaching to the principle of effectiveness apply to every legal remedy which national law makes available to claimants for the reimbursement of taxes levied in breach of EU law.”

This is not, however, something that has ever been said so clearly by the CJEU. We do not think that FII CJEU3 is entirely conclusive on the point, but we do accept that the Court of Appeal in Littlewoods CA thought that it was. We refer, in particular to [130], following the citation of [37-39] of FII CJEU3, where they said (as recorded already above) that: “[t]his seems to us to be a clear rejection of the Court of Appeal’s view that the test claimants in FII were not entitled to protection in respect of both existing domestic causes of action which were capable of giving effect to their San Giorgio rights” (and see also [136] of Littlewoods CA). We do not think that [141-142] of Littlewoods CA detract from what the Court of Appeal had said earlier, as Mr Ewart suggested it did. The national court’s “procedural autonomy” does not mean that a claimant seeking to enforce a San Giorgio right can be deprived of one of two available domestic remedies because of the application of defences or arguments that make it ineffective to give effect to that right.

We also think that the decision of the Supreme Court in this very case is of the greatest importance. The present situation is, we think, another example of HMRC running an argument that has already been run in the same case at another level. We find it very surprising that HMRC should take up so much time with arguing this point again at Court of Appeal level, when the Supreme Court has already disagreed with FII CA1, even if, at the same time, it decided to make a reference to the CJEU. We understand Mr Ewart’s argument that [212] of Lord Reed’s judgment does not expressly say that the mistake-based remedy has to be moulded to vindicate the claimants’ San Giorgio rights, even when the Woolwich claims are available to do so, but we do not agree with it for two simple reasons. First, Lord Reed said in terms that he disagreed with the argument that “the additional ground of action which English law provides [the mistake-based claim] falls outside the scope of the [principle of effectiveness]”. Secondly, Mr Ewart’s construction of Lord Reed’s [212] ignores the words “in similar circumstances” in his pre-penultimate sentence. There he said, again clearly we think, that “[w]here an action for recovery of taxes under domestic law can be based either on the ground of mistake or on [a Woolwich claim] (or, as in the present case, on both grounds) it follows from the principle of equivalence that both grounds of action should also be available in similar circumstances to enforce an analogous right under EU law. The principle of effectiveness therefore applies to both grounds of action” (emphasis added). Lords Hope at [20-23], Lord Clarke at [136], Lord Walker at [91], [102] and [115], and Lord Dyson at [140] either agreed with Lord Reed or were to similar effect.

In these circumstances, we do not think it is open to us to depart from the views of the Supreme Court. Moreover, even if it were open to us to do so, we would not have done so. The litigation concerning overpayments of various taxes has taken numerous twists and turns over the last ten years. But we think it is now reasonably clear, as the judge said, that where domestic law allows two possible remedies to vindicate a San Giorgio right the EU law principles of equivalence and effectiveness require that both those remedies are moulded so as to vindicate the Claimants’ San Giorgio rights to recover the overpaid tax and an adequate indemnity for the losses occasioned (see [29] in Littlewoods CJEU). As we have pointed out already, it is hard to “mould” a domestic restitutionary claim which looks only to the enrichment of the defendant to satisfy such an unequivocally compensatory right. But that is what must be done.

In our view, clear instances of this process mean that EU law precludes each of the arguments that HMRC has raised to reduce the restitution they must make. HMRC cannot argue for the set off of tax credits under issue 17; HMRC cannot argue to reduce the restitution by reason of an argument their actual benefit was less than the objective use value of the prematurely paid ACT; and HMRC cannot raise a “change of position” defence to the mistake-based claims.

Our conclusions on the remedies issues are, therefore, as follows.

HMRC’s appeal on issue 15 should be allowed and we should declare that HMRC are entitled to a credit against the repayment to FCE of overpaid ACT in respect of double taxation treaty half tax credits granted to its parents, Ford US.

HMRC’s appeal on issue 17 should be dismissed. As a matter of English law, HMRC were enriched by the full amount of the payments of ACT even when they were required to give corresponding tax credits to the recipients of the dividends.

HMRC’s appeal on issue 22 should be allowed. Littlewoods CA has already concluded in paragraph 187 that actual use value could be relevant to valuing the benefit which the government received as a result of the overpayment of tax. We agree with that conclusion. It is a conclusion that is directly applicable in this case. The answer to this issue is, therefore, that, as a matter of English law, it was open to HMRC to raise an actual benefit argument in respect of the claimants’ mistake-based claims.

HMRC’s appeal on issue 23 should be dismissed. The “actual benefit” argument was not, in our judgment, made out on the facts of this case.

The Claimants’ cross-appeal on issue 19 should be dismissed. The judge was right to think that the question of whether the “change of position” defence to the claimants’ mistake-based claims was available to HMRC in this case, in principle and as a matter of English law, was res judicata as a result of FII HC1 and FII CA1.

HMRC’s appeal on issue 20 should be dismissed. HMRC failed to make out the defence of change of position on the facts, and failed to show that the judge had misapplied the law to the facts that he found (which were not themselves appealed).

HMRC’s appeal on issues 18, 21 and 24 should be dismissed. We do not think it is open to us to depart from the views of the Supreme Court in FII SC. We would not anyway have done so. Where domestic law allows two possible remedies to vindicate a San Giorgio right, the EU law principles of equivalence and effectiveness require that both those remedies are moulded so as to vindicate the claimants’ San Giorgio rights to recover the overpaid tax and an adequate indemnity for the losses occasioned. HMRC are precluded by EU law from arguing under issue 17 for the set off of tax credits. HMRC cannot argue to reduce the restitution required by EU law on the basis that their actual benefit was less than the objective use value of the prematurely paid ACT. HMRC is precluded by EU law from raising a “change of position” defence to the claimants’ mistake-based claims.

None of the other remedies issues dealt with by the judge has been the subject of substantive argument, but we understand that the parties wish to reserve their position on one or more of them should the matter go, once again, to the Supreme Court.

It is now settled that the Claimants are entitled, in so far as their claim is formulated as a mistake-based restitutionary claim, to take advantage of the extended limitation period under section 32 (1) of the Limitation Act 1980. This reads (so far as material) as follows:

“… [W]here in the case of any action for which a period of limitation is prescribed by this Act, either—

(a)

…;

(b)

…; or

(c)

the action is for relief from the consequences of a mistake;

the period of limitation shall not begin to run until the plaintiff has discovered the … mistake or could with reasonable diligence have discovered it. …”

Time for bringing these mistake-based claims accordingly only began to run when the Claimants in question discovered the relevant mistake or could with reasonable diligence have discovered it.

The claims that the relevant payments of ACT and Case V tax by the BAT Claimants were made under a mistake were formulated in the Amended Particulars of Claim as follows. Paragraph 15A pleads:

“The Claimants made the ACT Payments by reason of their mistaken beliefs (i) that the ACT Provisions were lawful and enforceable, and/or (ii) that the Claimants were lawfully obliged to make those payments”.

Paragraph 17B pleads:

“The Claimants made the DV Corporation Tax Payments, to the extent of their unlawfulness, by reason of their mistaken beliefs (i) that the Dividend Provisions were lawful and enforceable, and/or (ii) that the Claimants were lawfully obliged to make those payments.”

(The various capitalised phrases in those paragraphs are defined terms but it is unnecessary to set out the definitions for present purposes.)

At paragraphs 18-18B of the Amended Particulars of Claim the BAT Claimants went on to plead their case as to the extended limitation period (prematurely, as a purist might think, since no limitation defence had yet been raised). In respect of the ACT claim, it was averred that:

“The Claimants discovered their mistakes relating to the ACT Payments when the ECJ gave its judgment [that is, in FII CJEU1] on 12 December 2006. The Claimants could not with reasonable diligence have discovered their mistakes any earlier than they did, alternatively, any earlier than when the ECJ gave its decision in Joined Cases C-397/97 Metalgesellschaft Ltd v IRC and C-410/98 Hoechst AG v IRC on 8 March 2001.”

In respect of the Case V claim it was averred that:

“The fact that the DV Corporation Tax Payments, to the extent of their unlawfulness, [and certain associated payments] were made by mistake depends upon the final determination of the issues in these proceedings. In the premises, the Claimants could not with reasonable diligence have discovered these mistakes at any other time or in any other way.”

In their Re-Amended Defence HMRC admitted the relevant paragraphs in full. That no doubt reflected the fact that even on the basis of the fall-back “reasonable discoverability” date of 8 March 2001 the BAT Claimants’ claims were comfortably in time since proceedings had been commenced in 2003. However, some of the other claimants in the FII group did not initiate proceedings until after the 8 March 2007, and HMRC in due course made it plain that, although pleadings had not been required in the cases of those claimants, it would be raising a limitation defence in them. Although we were not given details, it appears that there are about half a dozen claimants (or groups of claimants) in this class. In the “Statement of Agreed Facts and Disputed Issues” prepared in the run-up to FII HC2 HMRC contended that “the Claimants could reasonably have discovered their mistake on 8 March 2001”. It was agreed that that was an issue which Henderson J should determine. It was eventually designated as issue 28 and formulated as “when did the Claimants discover (or when could they with reasonable diligence have discovered) their mistake?”.

We have to say that the Judge did not get the help with issue 28 that he was entitled to expect, since it was not addressed in the written or oral closing submissions on either side. Even before us HMRC said nothing about the issue in their skeleton argument – a surprising omission in view of the sums at stake, but one for which Mr Ewart offered no explanation or apology.

We shall start our treatment of this issue by summarising the relevant parts of the authorities that the parties have relied upon.

It was not until the decision of the House of Lords in Kleinwort Benson that the question of what it meant to “discover” a mistake of law became live, but there is no substantial discussion of the question in that case. However the question did arise in DMG. It will be recalled that in Hoechst the CJEU had held that the non-availability of group income election to UK-resident companies with a parent company resident in another member state, and their consequent liability to ACT which would not have been payable if the parent had been UK-resident, infringed the right of establishment under the EC Treaty. That decision essentially endorsed the opinion of the Advocate General, which was promulgated on 12 September 2000. In consequence of Hoechst’s challenge, a large number of UK-resident subsidiaries of non-resident parents brought proceedings for restitution of ACT which they had paid when it was not in fact lawfully due. Some did not do so until the decision of the CJEU was known; but others anticipated it, including DMG which issued proceedings on 18 October 2000. Some of the payments in respect of which it claimed were made more than six years before that date.

DMG in due course became the lead claimant. Its claim was heard before Park J in May 2003, and his judgment was handed down on 18 July 2003 ([2003] EWHC 1779 (Ch), [2003] 4 All ER 645). He held that DMG was entitled to bring a mistake-based claim and that it could not reasonably have discovered its mistake until the decision of the CJEU in Hoechst. In the Court of Appeal ([2005] EWCA Civ 78, [2006] Ch 243) the majority (Buxton and Rix LJJ) held that DMG was not entitled to bring a mistake-based claim, and they accordingly did not have to consider the discoverability issue. Jonathan Parker LJ, however, agreed with Park J on both points.

The House of Lords restored the decision of Park J. The issue of when time started to run was considered in the speeches of Lord Hoffmann, Lord Hope and Lord Walker. We take them in turn.

At [23-30] Lord Hoffmann addressed the question of whether DMG had paid the ACT in question under a mistake, and in particular whether it was material that Hoechst’s challenge, of which DMG was aware from a comparatively early stage, might be regarded as having raised a doubt over its obligation to pay. He held that DMG had indeed made a mistake, albeit one of a special character because it arose from a change in the law. He continued at [32]:

“If DMG made a mistake about the law, when could they ‘with reasonable diligence’ have discovered it? On this question it is important to bear in mind the special nature of the mistake, namely that it was deemed to have been made because of the retrospective operation of a later decision of the Court of Justice. The ‘reasonable diligence’ proviso depends upon the true state of affairs being there to be discovered. In this case, however, the true state of affairs was not discoverable until the Court of Justice pronounced its judgment. One might make guesses or predictions, especially after the opinion of the Advocate General. This gave DMG sufficient confidence to issue proceedings. But they could not have discovered the truth because the truth did not yet exist. In my opinion, therefore, the mistake was not reasonably discoverable until after the judgment had been delivered.”

Lord Hope dealt with what he described as “the discovery issue” at [63-71]. In this passage he also addressed, like Lord Hoffmann, the closely related question of whether DMG could be said to have been mistaken in circumstances where it must or should have appeared, as a result of Hoechst’s challenge, that the obligation to pay ACT was in doubt. His conclusion and reasoning are sufficiently clear from [71], where he said:

“DMG's mistaken belief that a group income election was not available was not shown to be wrong until the issue which Hoechst had raised was determined by the European Court on 8 March 2001. The issue, which was one of law, was not capable of being resolved except by litigation. Until the determination was made the mistake could not have been "discovered" in the sense referred to in section 32(1) of the 1980 Act. In this situation I would hold, in agreement with Park J and Jonathan Parker LJ, that DMG was not in a state of doubt, when it paid the ACT. It was not then obvious that the payments might not be due. They were made in accordance with the law as it was then understood to be. There was nothing yet to be discovered to the contrary which could have been revealed by the exercise of reasonable diligence.”

Lord Walker addressed this issue at [144], where he said:

“I think the judge and Jonathan Parker LJ were correct in their views that the mistake was not discovered until the European Court of Justice gave judgment in the Metallgesellschaft/Hoechst case [2001] Ch 620. Perusal of the report in that case suggests that the United Kingdom Government tenaciously defended the ACT regime on every available ground. At no time before the judgment did the Government concede that the ACT regime was (in discriminating between national and multi-national groups) contrary to EU law and unlawful. It was the judgment that first turned recognition of the possibility of a mistake into knowledge that there had indeed been a mistake. I agree with the view of Lightman J in First Roodhill Leasing Ltd v Gillingham Operating Co Ltd (unreported) 5 July 2001, para 22 that there may be cases ‘where a party may be held to have discovered a mistake without there being an authoritative pronouncement directly on point on the facts of that case by a court, let alone an appellate court’. It all depends on the facts. But in this case it is, in my opinion, clear that the judgment of the European Court of Justice on 8 March 2001 was the decisive moment.”

Lord Scott dissented on the primary issue of whether DMG was entitled to base its claim on mistake and he accordingly said nothing about the question of when any mistake was reasonably discoverable.

Lord Brown agreed with the majority on the primary issue, but he differed as to the date when time started to run. He said at [165]:

“… I would hold that as soon as a paying party recognises that a worthwhile claim arises that he should not after all have made the payment and accordingly is entitled to recover it (or, as here, to compensation for the loss of its use), he has "discovered" the mistake within the meaning of section 32; and, by the same token, I would hold that if he makes any further payments thereafter, they are not to be regarded as payments made under a mistake of law.”

He added, at [169]:

“… I fail to see why the question whether moneys are paid under a mistake of law should turn on the degree of conviction or optimism which the parties hold upon the legal issue dividing them. Were the claimants in the Hoechst case (who issued their proceedings against the revenue in 1995) none the less to be regarded as having made all subsequent payments under a mistake of law? Surely not. Even DMG itself, it will have been noted, brought its claim in October 2000. Is it nevertheless to be said that their original mistake remained undiscovered until the European Court of Justice's actual decision in Hoechst some five months later?”

And at [172-3]:

“172.

On the present appeal, however, Lord Hope concludes his judgment on ‘the discovery issue’, at paras 63-71 of his speech, with the view that, when DMG paid the ACT, ‘it was not then obvious that the payments might not be due.’ I confess to some difficulty with that conclusion. Surely, when DMG learned in July 1995 that there was a serious legal challenge to the legality of the ACT regime, it must then have been obvious to them that these payments might not after all be due. Of course they could not be sure and of course nothing short of a final judgment from the European Court of Justice would have persuaded the revenue to accept any claim by DMG here for group income relief. But it does not seem to me to follow that DMG paid under a mistake of law-any more than Woolwich would be regarded as having paid under such a mistake simply because the revenue in that case were insisting on the validity of the contested regulations.

173.

I have the same difficulty with para 144 of Lord Walker's opinion. Again, I see no good reason why the revenue's tenacious defence of their position and their refusal to concede its unlawfulness means that DMG's mistake must be treated as undiscovered prior to the Metallgesellschaft/Hoechst judgment. The passage quoted by Lord Walker from Lightman J's judgment in First Roodhill Leasing Ltd v Gillingham Operating Co Ltd 5 July 2001, para 22, continues:

‘For this purpose it cannot be necessary that the party knows of the mistake as a certainty. There are gradations of knowledge. It may well be sufficient to constitute the necessary discovery when the claimant has good reason to believe that a mistake has been made (consider Earl Beatty v Inland Revenue Comrs [1953] 1 WLR 1090) or has been given “a line” on this question: see G L Baker v Medway [1958] 1 WLR 1216, 1224.’”

(We were shown the decision of Lightman J in the First Roodhill case to which Lord Walker and Lord Brown refer: its citation is [2001] EWHC Ch 397. It concerned an application for permission to amend and the discussion of the law about discoverability goes no further than the passages which they quote.)

The primary issue in FII SC concerned, as we have already said, the compatibility with EU law of section 320 of the Finance Act 2004 and section 107 of the Finance Act 2007 (“the statutory cut-off provisions”). The former purported to disapply the extended limitation period under section 32 (1) (c) of the 1980 Act in tax cases brought on or after 8 September 2003 (being the date when the intention to enact such legislation was announced), and the latter purported to extend that disapplication to cases brought prior to that date. Those provisions were said to infringe a number of different principles of EU law, but for present purposes we need be concerned only with the principle of protection of legitimate expectations: the Claimants’ case was that as at the cut-off date they had a legitimate expectation of being able to bring a mistake-based restitution claim in relation to payments of ACT going back more than six years. We are not in this case concerned as such with the principle of protection of legitimate expectations, but the consideration of that issue involved the Court in considering what the FII claimants could reasonably have understood their rights to be in 2003 and in 2007, and there is an obvious overlap between that question and the issue of when the Claimants could reasonably have discovered their mistake.

The legitimate expectation issue is most fully discussed in the judgments of Lord Walker and Lord Sumption. We take Lord Sumption’s judgment first. At [162-169] he reviewed the history of the relevant rights of recovery. At [167] he observed about the effect of DMG:

“The combined effect of the decisions on [the issues in DMG] was in one respect extremely remarkable. If tax was overpaid under a mistake of law, then provided that a claim to recover it was brought before six years had elapsed from the judgment establishing the correct legal position, there was no limit upon how far back the claim could go.”

He considered the protection of legitimate expectations at [197-204]. At [199] he said:

“Before Park J gave judgment in [DMG] on 18 July 2003, no one could reasonably have counted on being able to recover tax on the ground of mistake of law. They might have thought that there were strong arguments to that effect, but I do not believe that they could reasonably have assumed when deciding how long they had in which to bring their claims that those arguments would prevail. Even after Park J's judgment, the right to recover tax on the ground of mistake of law was being challenged on appeal on serious grounds. The existence of such a right was rejected by the Court of Appeal and was not definitively established until the judgment of the House of Lords on 25 October 2006.”

At [201] Lord Sumption emphasised that the Claimants’ legitimate expectations must be judged by reference to what the law was actually understood to be at the relevant time; and he went on to point out that this involved the same approach as the test of reasonable discoverability. He said, at [202]:

“It is right to point out that this is substantially the same principle as that on which the test claimants themselves rely when they say (with the support of the House of Lords in [DMG]) that they cannot be taken to have discovered their mistake about the lawfulness of the United Kingdom's corporation tax regime until the Court of Justice definitively decided the point. By the same token, the test claimants cannot be taken to have assumed that they had a right to recover the tax on the ground of mistake at a stage when they had arguments and hopes but no definitive decision.”

His decision was that the enactment of section 320 did not infringe the principle of the protection of legitimate expectations. (He reached a different decision about section 107, but we need not be concerned with that.)

Lord Walker addressed the compatibility of the statutory cut-off provisions with EU law at [100-115]. At [103-105] he said this:

“103.

… [U]ntil the decision of the Court of Justice in [Hoechst] there was no general appreciation that the UK corporation tax regime was seriously open to challenge as infringing the Treaty. Henderson J [that is, in FII HC1] did not make any detailed findings about this, since the principle of legitimate expectations does not seem to have been argued as a separate issue before him. But he did [2009] STC 254, para 267, make a general finding of fact about mistake:

‘The unlawful payments of ACT made from 1973 to 1999, and the unlawful payments of ACT made under the FID regime from 1994 to 1999, were in my view plainly made under a mistake about the lawfulness of the tax regimes under which they were paid. I am satisfied from the evidence, both written and oral, that this was not obvious to anybody within the BAT group at the time, since everybody proceeded on the footing that the tax in question was lawfully due and payable.’

104.

After 8 March 2001 a well advised multinational group based in the UK would have had good grounds for supposing that it had a valid claim to recover ACT levied contrary to EU law, with at least a reasonable prospect that the running of time could be postponed until then (but not subsequently) by the operation of section 32(1)(c) of the Limitation Act 1980.

105.

The next important date was 18 July 2003, when Park J gave his first instance judgment in DMG ... . This was the first judicial decision which positively upheld a claim for repayment of unduly levied tax with an extended limitation period under section 32(1)(c). But appeals to the Court of Appeal and the House of Lords were to follow and (as Henderson J observed, at para 406), ‘the outcome of those appeals was, at the time, impossible to predict with any confidence.’”

He went on, at [111-112], to agree with Lord Sumption’s conclusion in [199], which he characterises as being that as at 8 September 2003 “no one in the position of the test claimants could have had a reasonable and realistic expectation of recovering tax on the ground of mistake”. (He held, however, that the enactment of section 320 infringed the principle of effectiveness. The majority agreed with his conclusion on that point, but Lord Sumption and Lord Brown did not, which is why the further reference that led to FII CJEU3 was necessary.)

The other members of the Court dealt with the legitimate expectation issue less fully. We should, however, set out two passages because, as will appear, Henderson J attached some significance to them. At [131] Lord Clarke said:

“As Lord Walker explains at para 104, after 8 March 2001, when the Court of Justice decided [Hoechst], the Test Claimants would have had good grounds for supposing that they had a good claim to recover ACT levied contrary to EU law, with at least a reasonable prospect that the running of time could be postponed until then by section 32(1)(c) of the Limitation Act 1980.”

At [236] Lord Reed said:

“As Lord Walker has explained at paras 103-104, it has been established in this case that the payments were made under a mistake about the lawfulness of the tax regimes under which they were paid; and it was only after the Court of Justice issued its judgment in [Hoechst] that it was generally appreciated that the UK corporation tax regime was open to challenge as infringing Community law. A well-advised company in the position of the claimants would then have had grounds for considering that it was entitled to the repayment of tax which had been levied contrary to Community law, and that there was at least a reasonable prospect that it could rely upon the extended limitation period provided by section 32(1)(c) of the 1980 Act in order to recover any taxes paid more than six years before the proceedings were begun.”

The question referred to the CJEU in FII CJEU3 was, in summary, whether the enactment of s. 320 of the Finance Act 2004 infringed the EU principle of effectiveness. Resolving that question did not require the Court to consider the application of the reasonable discoverability test in the circumstances of the present claims. However, at [19], the CJEU said this:

“Under section 32(1)(c) of the 1980 Act, the limitation period applicable to that action began to run from discovery of the mistake of law giving rise to the payment of the tax, in the present case, the date of delivery of the judgment in [Hoechst], namely 8 March 2001.”

Henderson J dealt with issue 28 at [452-470] of his judgment. His reasoning can be summarised as follows.

The issue is introduced at [452-456]. We need not refer to anything save Henderson J’s reference at [454] to what he described as “at first sight … an insuperable logical difficulty” in the Claimants’ case, namely that it is hard to see how they can claim that the mistake on which they rely was not reasonably discoverable until 12 December 2006 when they had begun proceedings based on that very mistake three years ago.

The judge then summarised at [457-458] the evidence given by the Head of Tax at BAT, Mr Hardman, about when he knew what about the possibility that past payments of ACT had not been legally due. In short, his evidence was that he first became aware of such a possibility, and of the arguments being advanced by DMG, in early 2003. The judge found, at [458]:

“Mr Hardman therefore believed that, if DMG were successful, BAT could likewise bring a claim for recovery of ACT paid under a mistake, and that such a claim could extend back to 1973.”

At [459-460] the judge described the history of the DMG proceedings, summarising the view of the majority in the House of Lords as being that:

“… [I]t was only the definitive statement of the law by the ECJ in Hoechst which made the mistake reasonably discoverable within the meaning of section 32(1)(c). It was not enough that DMG had known before then that the Hoechst claim was being brought, or that it had reasonable prospects of success.”

He continued at [461]:

“By parity of reasoning, it seems to me strongly arguable that it was only when the House of Lords delivered its judgment in DMG that, in the present case, time began to run against the BAT test claimants. The existence of a mistake-based remedy to recover unlawfully levied tax had been established at first instance in 2003, but the issues were far from straightforward, the case went the other way in the Court of Appeal, and finality was only achieved when the House of Lords pronounced on the subject.”

The judge said that that view was supported by what Lord Sumption said in the passages from his judgment in FII (SC) which we have set out above at paragraph 361, with which Lord Walker had agreed. He concluded at [465]:

“In the light of the approach to section 32(1)(c) adopted by the House of Lords in DMG, and the history of the DMG litigation itself as described by Lord Sumption in FII (SC), I consider that the date when a mistake is discovered, or becomes reasonably discoverable, for the purposes of section 32 may, in principle, be a date later than that on which an action based on the mistake was started. A well-advised claimant may begin proceedings at a time when the law is still, objectively, unclear. That was the position, in my judgment, in relation to the claimants' mistake-based claims on 18 June 2003, and the uncertainty was not finally resolved until the House of Lords delivered its judgment in DMG on 25 October 2006.”

The judge started [466] by saying that the foregoing analysis had initially inclined him to hold, “by parity of reasoning”, that the Claimants did not discover, and could not reasonably have discovered, their mistake until 25 October 2006. But he then proceeded to hold that that reasoning is inconsistent with [104] of the judgment of Lord Walker in FII SC, which we have set out at paragraph 362 above. He emphasised the parenthesised phrase “but not subsequently”: he said that it showed that Lord Walker regarded the date of the decision in Hoechst as the latest date to which the running of time could arguably have been postponed. He observed that Lord Clarke and Lord Reed had referred approvingly to that passage, at least in general terms, in the passages which we have set out at paragraph 363 above. He concluded:

“468.

It seems to me that on a question of this nature, which is exclusively a question of English law, and mainly turns on an objective appraisal of a complex and evolving legal landscape, I should follow the guidance given by the majority in the Supreme Court, even though it was not an issue which they had to decide. I am further encouraged to do so by the fact that the same view is reflected in paragraph 19 of the judgment of the ECJ in FII (ECJ) III, where the Court said:

‘Under section 32(1)(c) of the 1980 Act, the limitation period applicable to that action began to run from discovery of the mistake of law giving rise to the payment of the tax, in the present case, the date of delivery of the judgment in [Hoechst], namely 8 March 2001.’

It was not, of course, within the competence of the ECJ to decide a question of English procedural law, but this wording presumably reflected the order for reference and the written observations of the parties in a way which was not perceived to be controversial.

469.

My answer to this Issue is accordingly that the date when the claimants discovered (or could with reasonable diligence have discovered) their mistake is 8 March 2001 when the ECJ delivered its judgment in [Hoechst]. If that is wrong, I would hold that the relevant date was 25 October 2006 when the House of Lords delivered its judgment in DMG.”

Although we have thought it right to set out the relevant authorities, and the Judge’s reasoning, quite fully, we can deal with the issue shortly.

The correct starting-point must be the reasoning of the majority in DMG, which was concerned directly with the issue of when a mistake of law can be regarded as “discoverable”. It is in our view quite clear that the majority held that in the case of a point of law which is being actively disputed in current litigation the true position is only discoverable, for the purpose of section 32 (1) (c) of the 1980 Act, when the point has been authoritatively resolved by a final court. That is wholly explicit in the speeches of Lord Hoffmann and Lord Hope quoted at paragraphs 355 and 356 above. Mr Ewart argued that Lord Walker’s reasoning – see paragraph 357 – was different because of his endorsement of Lightman J’s observations in the First Roodhill case and his statement that “it all depends on the facts”; and that accordingly there was no majority ratio. We do not agree. It is in our view clear that Lord Walker was acknowledging that in other circumstances the law may be regarded as established, and thus discoverable, even without a relevant judicial decision. But that was evidently not so where the correct legal position was being actively argued through the courts. We see no difference between Lord Walker’s reasoning and that of Lord Hoffmann and Lord Hope. That reasoning confronts but neutralises the apparent logical difficulty referred to by Henderson J that a mistake of law may not be discoverable in the relevant sense until the conclusion of proceedings even though the claimant has based his whole claim on that very mistake. Henderson J’s analysis at [465] of his judgment, quoted at paragraph 369 above, is in our view correct.

In our view it follows by parity of reasoning – as indeed Henderson J said – that the mistakes relied on by the Claimants were not discoverable until the decision of the CJEU in FII CJEU1. It was only at that point that it was authoritatively established that, to quote from the pleading, the ACT provisions were not lawful or enforceable, and that they had not been lawfully obliged to make the ACT payments. (The same goes for the Case V tax provisions, though those were not the focus of the argument.) The situation is substantially identical to that considered in DMG and we are bound by the reasoning of the majority.

The only basis on which the discoverability date could be 8 March 2001 would be if the mistake of law that was authoritatively exposed by the CJEU in Hoechst was in substance the same mistake as had led the Claimants to make the ACT payments. But on their pleaded case that was plainly not so. The provisions which they had mistakenly believed obliged them to make the payments are not the same as the provisions pursuant to which Hoechst had made the payments in that case; nor was it contended that the fact that the latter provisions infringed EU law necessarily meant that the former did also. There is no doubt that the decision in Hoechst prompted the question whether the ACT system infringed EU law in other respects, and if the opinion of Lord Brown in DMG had prevailed 8 March 2001 might indeed have been the crucial date; but it did not.

Henderson J was, as we have seen, deflected from the same conclusion by what Lord Walker said at [104] of his judgment in FII SC: see paragraph 362 above. We do not think that he should have been. The issue under discussion in that paragraph, and the longer passage of which it forms part, was whether the state of the law as it stood at the cut-off date of 8 September 2003 was such that the Claimants had a legitimate expectation of being able to rely on the extended limitation period in claiming restitution in relation to their payments of ACT (his conclusion in fact being, despite what he said in [104], that it was not). The importance of the reference to the date of the Hoechst decision was that it was arguable that time did not start to run until that date: the question of whether it might be postponed any later was of no significance. The phrase “(but not subsequently)” was thus unnecessary – which is no doubt why it was it was in brackets (and also perhaps why it was not picked up in the passages from the judgments of Lord Clarke and Lord Reed to which Henderson J referred). It is not possible to know why Lord Walker included it, but on any view it was obiter and it is highly unlikely that the question was the subject of any submissions.

The point is a fortiori as regards [19] of the judgment of the CJEU in FII CJEU3. The researches of the parties have been unable to identify from what source the Court derived the statement there made. But the issue of when time started to run was irrelevant to the issue before the Court and was the subject of no submissions to it.

We would accordingly answer issue 28 by holding that the Claimants’ mistake was discoverable on 12 December 2006.

We accordingly dismiss the appeals and cross-appeals, save that we allow HMRC’s appeal on issues 15 and 22 and the Claimants’ cross-appeal on issue 28. In the Evonik Degussa applications, we grant permission but dismiss the appeals.

In what respects was the Case V charge unlawful under EU law?

What is the appropriate Foreign Nominal Rate?

Special cases

How should the lawful Case V charge be computed?

Does EU law require a credit to be given within the ACT computation for underlying tax as well as for tax at the FNR?

Does EU law require credit also to be given against ACT for withholding tax?

How is the lawful ACT to be calculated?

How should ACT be paid by companies be linked with EU-source income to give effect to the judgment in FII CJEU1 and FII CJEU2?

FIDs

Does it make any difference that the UK group had a non-resident parent which received double taxation treaty credits?

Taking into account the interaction of ACT with shareholder tax credits, were the Revenue enriched as a matter of English law and, if so, to what extent?

Is the Revenue's argument that they were not enriched by reason of the interaction between ACT and shareholder tax credits precluded by EU law?

Is a change of position defence available to the Revenue as a matter of principle under English law in respect of the claimants' mistake claims?

Have the Revenue made out a defence of change of position on the facts?

Are the Revenue precluded from relying on a change of position defence by EU law?

Is the Revenue's "actual benefit" argument available to them in respect of the claimants' mistake claims under the English law of unjust enrichment?

If it is available, is the "actual benefit" argument made out on the facts?

Is the "actual benefit" argument permitted by EU law?

When did the claimants discover (or when could they with reasonable diligence have discovered) their mistake?