The Test Claimants in the FII Group Litigation v HM Revenue and Customs

The trial of the test claims of UK-resident companies in the British American Tobacco (“BAT”) group, within the litigation proceeding under the Franked Investment Income (“FII”) group litigation order (“GLO”) originally made in the Chancery Division of the High Court on 8 October 2003, embarked on its Odyssey over ten years ago when, in June 2004, Park J directed that a reference be made to the European Court of Justice (“the ECJ”). The order for reference was dated 13 October 2004, and the Grand Chamber of the ECJ gave its judgment on 12 December 2006 (“FII (ECJ) I”, Case C-446/04, [2006] ECR I-11753, [2012] 2 AC 436 (Note), [2007] STC 326).

With the benefit of the guidance given by the ECJ, the resumed trial of all issues relating to liability (but not causation or quantification) in the BAT test claims took place before me in July 2008. I had meanwhile succeeded Rimer J, who had himself succeeded Park J, as the designated managing judge of the FII GLO. The trial lasted for 13 days, most of which were devoted to complex legal argument, although I also heard factual and/or expert evidence on some important issues, including the defence of change of position (to the claimants’ restitutionary claims) and the question whether a sufficiently serious breach of EU law by the UK had been established to ground liability for damages in accordance with the Factortame criteria. I handed down my judgment, to which I will refer as “FII (High Court) I”, on 27 November 2008: see Test Claimants in the FII Group Litigation v HMRC [2008] EWHC 2893 (Ch), [2009] STC 254. The judgment was unavoidably lengthy, running to 450 paragraphs. The order giving effect to it was dated 12 December 2008.

Since then, there have been the following main developments in the FII group litigation.

Against this background, I gave directions at a case management conference on 14 May 2013 for the trial of the BAT test case to be resumed to determine all remaining issues of liability and quantification, with the sole exception of certain issues upon which the Supreme Court had in 2010 deferred a decision on whether to grant permission to appeal from FII (CA). The directions included a timetable for further amendments to be made to the pleadings, for the exchange of a set of worked examples dealing with the factual situations upon which the court would be asked to rule, for service by the claimants of a schedule giving particulars of the quantification of their claim, for the exchange of witness statements and for expert evidence. The trial was subsequently fixed to begin on 29 April 2014, with a four week time estimate.

As one would expect with a case of this complexity, the working out of those directions gave rise to a number of disputes and applications to the court. I ruled on some of these questions in a judgment which I handed down on 29 November 2013: see [2013] EWHC 3757 (Ch), [2013] STC 826. One of the matters in issue was whether it was still open to the Revenue to argue that the foreign income dividend (“FID”) regime introduced in 1994 could benefit from the “standstill” protection then contained in Article 57(1) EC (now Article 64(1) TFEU). I held that it was not, on the basis that the point had already been conclusively determined in the claimants’ favour by the Court of Appeal: see my judgment at [31] to [39]. An appeal by the Revenue from my decision on this issue was heard and dismissed by the Court of Appeal on 27 March 2014, for reasons subsequently given in judgments handed down on 2 September 2014: see [2014] EWCA Civ 1214.

In the event, oral argument in the resumed trial before me began on 6 May 2014, and then continued over the course of 17 sitting days until 12 June. The parties had sensibly agreed a timetable for the trial which, while it may have looked rather leisurely on paper, in fact had built into it a number of gaps for reading and the preparation of written submissions, as well as the late Spring vacation. In this way, the evidence and argument on the multifarious issues which I had to consider were broken down into relatively digestible segments. I certainly found this beneficial, and it must also have lessened the burden on counsel of presenting an exceptionally difficult and complicated case. I emphasise, too, that the timetable was adhered to without difficulty, and the case was concluded within its revised five week time estimate.

By the start of the trial, the parties had all but agreed a statement of facts and disputed issues, together with 11 appendices. By the end of the trial, it was common ground that I could treat this as an agreed document. In the course of argument, however, a number of more detailed sub-issues had emerged, and each side’s position on various computational issues had also evolved or been modified. I therefore asked the parties to agree a fuller list of issues in tabular form, cross-referenced to the transcript and the relevant passages in the many written submissions which were provided to me both before and during the hearing. This exercise took rather longer to perform than I had expected, but the resulting document (which I will call “the List of Issues”) was of real value. It was finally submitted to me in mid-July, over a month after the end of the hearing. It lists no fewer than 29 issues for me to determine, some of which are sub-divided. I will in general use the List of Issues as the template for my judgment, and I will refer to the individual issues in the form “Issue 1(a)”, etc.

Before I move on, it is convenient to record that over the last 18 months I have heard two other very substantial cases which raised several issues which also arise in the present case. Those cases are:

Rather than attempt an explanation of the issues in those two cases at this early stage, I will refer to the judgments as and when it is relevant to do so. I will, however, say now that, in general, I accept the submission of the test claimants that, where I have already decided an issue after full argument in either or both of the cases, I should not revisit it in the present case unless persuaded that I was clearly wrong, or that I have something to add which may be of assistance to a higher court. The relevant issues are ones of importance and difficulty. They are also likely to be considered by the Court of Appeal in the relatively near future. Where I have already expressed my considered opinion on them, I think that I should exercise restraint before adding to the discussion of them at first instance. This is particularly so in relation to issues decided in Portfolio Dividends (No. 2), where I heard argument from substantially the same teams of counsel as in the present case, and where the issues arose in a closely comparable context (the main significant difference being that the FII group litigation is concerned with the payment of dividends by foreign subsidiaries to their UK-resident parents, whereas Portfolio Dividends (No. 2) was concerned with the payment of dividends on foreign portfolio shareholdings to UK-resident companies).

Because of the general importance of nearly all the issues, the status of the claims as test cases in group litigation, the practical certainty of appeals whatever I decide, and the huge amounts of money at stake, I have been asked by both sides to evaluate the evidence and state my conclusions on all of the issues raised by the parties, even if a decision on them is not strictly necessary to my ultimate conclusions. I agreed that I would where possible adopt this approach, as I did in Littlewoods (No. 2). Apart from the factors which I have mentioned, there are two other main reasons for doing so. First, I wish to minimise the need for further hearings if a higher court differs from me on questions of law. Secondly, I hope that some of my reasoning on questions of law, even if technically obiter, may be of assistance to a higher court when it comes to consider them.

The relevant factual background is mainly set out in FII (High Court) I at [29] to [38]. Paragraph [29] reproduces the statement of facts agreed between the parties in 2008 in relation to the BAT test case. The main test case has throughout been that of the BAT group. The claim of the Aegis group was previously joined as an additional test case in relation to certain limitation issues, but those issues have now been conclusively determined and the Aegis claim therefore played no part in the present trial.

On the other hand, however, two further test claims have been joined for the purposes of the present trial, because they raise certain factual situations relevant to quantification which do not arise in the BAT claim. They are the claims of the Ford group and the GKN group. The Ford claim extends to profits received from EU branches of the business carried on by the UK claimant, as well as dividends received from EU subsidiaries, and also raises some technical issues about credits received by ultimate shareholders. The GKN claim includes dividends received from EU-resident holding companies which incorporated the distribution of profits from a variety of jurisdictions (both EU and non-EU, or “third country”), and again raises a number of technical issues which are not covered (or are arguably not covered) by the facts of the BAT group claim.

Other findings of fact relevant to liability issues which have a bearing on the present quantification trial may be found in FII (High Court) I at [29] to [38], [64], [104] to [105], [107], [277] to [302] and [349] to [352]. My detailed findings of fact on the issue of “sufficiently serious breach” at [376] to [404] may, however, mostly be ignored at this stage, because that is one of the issues on which the Supreme Court has yet to decide whether to grant permission to appeal.

As to the underlying data relevant to the main quantification issues, the parties have been able to agree the details of the subject dividends, the tax payments and the applicable rates of UK tax. This information is contained in exhibits “AMHC 9”, “12, “13” and “14” to the sixth witness statement, dated 11 November 2013, of Anthony Cohn, who is now the regional tax manager for the BAT group with responsibility for Eastern Europe, the Middle East and Africa. The information contained in those exhibits was itself based on the schedules to the particulars of claim and the claimants’ particulars of quantification dated 23 September 2013.

In particular, the parties agree that the exhibits in general correctly record:

There are some minor respects in which it has become clear that the details as originally set out in the exhibits to Mr Cohn’s sixth statement need to be amended, but I do not understand any of these to be controversial, and I need not take up time examining them.

In relation to FIDs, details of the ACT paid in respect of them were included in the statement of agreed facts for the liability trial: see paragraphs 1.29 to 1.31, reproduced in FII (High Court) I at [29], and schedule 3 to the particulars of claim. Details of the distributable foreign profits against which the FIDs were matched under the FID regime may be found at pp 140 to 151 of exhibit “AMHC 5” to Mr Cohn’s fifth statement dated 8 May 2013. Those details, too, are agreed. With one exception, all of BAT’s FIDs were matched under the FID regime against distributable foreign profits which carried an actual rate of underlying tax paid in excess of the relevant UK corporation tax and ACT rates. The exception is the German distributable foreign profits matched with the 1994 FID referred to in paragraph 1.31 of schedule 3, where the actual rate of underlying tax and the foreign nominal rate of tax (on the claimants’ method of calculation) both exceeded the ACT rate but not the UK corporation tax rate.

There are a number of other specific areas in which the parties have been able to agree relevant facts, but it will be more convenient to refer to them as and when I come to the issues to which they relate.

The relevant legislative background is set out in FII (High Court) I at [12] to [28].

Issue 1: in what respects was the Case V charge unlawful under EU law?

The first group of issues in the List of Issues concerns the charge to corporation tax under Case V of Schedule D which the ECJ has held to be unlawful. It is now common ground that the effect of the ECJ’s decisions in FII (ECJ) I and FII (ECJ) II is that the Case V charge imposed on dividends which a UK parent company received from an EU-resident subsidiary was, at least to some extent, unlawful; but the parties remain divided on the precise extent of the illegality.

In short, the first question is whether:

Another way of expressing the same question is to ask whether the unlawfulness of the Case V charge lay in its failure to provide:

Since the UK tax system did in fact always provide a credit for the underlying foreign tax actually paid on EU dividends, by way of claims for double taxation relief (whether under double taxation agreements, or unilaterally under section 790 of the Income and Corporation Taxes Act 1988 (“ICTA 1988”)), and since in the majority of cases the alternative of a credit at the FNR would not have exceeded a credit for the underlying foreign tax actually paid, the practical significance of this dispute may at first sight appear to be limited. It is, indeed, true that the claims to recover unlawfully levied Case V tax form a relatively small part of the claimants’ overall claims. But the question is nevertheless an important one for a number of separate reasons. First, the extent to which ACT was lawfully charged, in respect of the onward distribution of the foreign dividends, can only be answered when the unlawfulness of the Case V charge on the dividends has been correctly identified. Secondly, and more generally, the rival approaches of each side to quantification take as their starting point the nature and extent of the illegality thus identified. It is only when the disease has been accurately diagnosed that an appropriate remedy can be fashioned. Thirdly, the two types of credit are conceptually quite distinct, and it is necessary to understand how they fit in with the reasoning of the ECJ as developed in its two decisions (the subsequent decision in FII (ECJ) III being for present purposes irrelevant).

Subject to one point, the parties’ agreed formulation of Issue 1 is reflected in the slightly fuller paraphrases which I have set out in paragraphs [20] and [21] above. The qualification is that the agreed formulation of the claimants’ contention includes the grant of a credit for foreign withholding tax (“WHT”) as a requirement of compatibility of the Case V charge with EU law. It is, however, common ground that EU law does not itself require a credit to be granted for WHT, the reason being that the relief of true, or “juridical”, double taxation of the same income in both the source State and the State of receipt is a matter for individual Member States to determine, in the absence of any general rules for the elimination of double taxation within the EU: see Portfolio Dividends (No. 2) at [54]. In these circumstances, I prefer to deal with the question of WHT in the context of Issue 4 below.

I should also note a reservation made by the claimants. They wish to reserve the right to argue in a higher court that the unlawful restriction of the tax credit in section 231(1) of ICTA 1988 to dividends paid by UK-resident companies should be disapplied, and cannot be remedied by a process of conforming construction. For the purposes of the present trial, however, they are content to accept, and indeed positively rely on, the guidance given by the Court of Appeal in FII (CA) at [105] and [107]:

“105.

… All the court is entitled and bound to do is to see whether s 231 can be read so that the right to a credit which is conferred extends not only to those expressly mentioned in s 231, namely resident companies, but also takes into account the rights of persons under Community law. Once that question is answered and the interpretation is given, the task of conforming interpretation is at an end. There is no further test to be applied about the ease of enforcing the rights thereby conferred or protected because persons entitled to tax credits as a matter of Community law are put onto the same footing under the section as other persons entitled to rely on the section as a matter of domestic law.

…

107.

It therefore falls to this court to determine the appropriate conforming interpretation. In our judgment, a conforming interpretation can be achieved simply by reading in words that make it clear that it is not just resident companies that can claim a credit under s 231 but also other persons entitled to do so by Community law to the extent that they are so entitled. The extent of that entitlement can then be investigated when the section falls to be applied, rather than the difficulties more properly arising at the point of application being erected as an objection to conforming interpretation. It will apply even if the extent of the entitlement is not fully ascertained until after the ECJ has answered any question put to it in a further reference.”

In essence, the question raised by Issue 1 is the same as the question which I have already considered and ruled on, in the context of portfolio dividends, in Portfolio Dividends (No. 2). The only factual distinction of any significance is that in the case of portfolio dividends, unlike dividends paid by subsidiaries, the UK legislation did not at any material time provide a credit for foreign underlying tax: see Portfolio Dividends (No. 2) at [83], where the relevant provisions of section 790 of ICTA 1988 are set out. This alone made it clear that the Case V charge on portfolio dividends was unlawful, given the ECJ’s repeated insistence in the case law that a credit for underlying tax was a prerequisite if a Member State chose to operate a “hybrid” system which granted exemption from corporation tax to domestic dividends, but operated an imputation system (granting credits for foreign tax) in the case of foreign dividends. Indeed, the ECJ had already held in its reasoned order in that case (in April 2008) that the Case V charge on portfolio dividends was, for that reason, unlawful, and that the defence of justification then advanced by the UK government had to be rejected: see Case C-201/05, Test Claimants in the CFC and Dividend Group Litigation v HMRC (Note), [2008] ECR I-2875, [2008] STC 1513, at paragraphs 35 to 43 and 64 to 69 of the reasoned order, and Portfolio Dividends (No.1) [2010] EWHC 2811 (Ch), [2011] STC 2014, at [8] to [10], [16] to [21] and [24].

Although the unlawfulness of the Case V charge on portfolio dividends had been thus established at an early stage, it was necessary in Portfolio Dividends (No. 2), as in the present case, to go on to examine precisely in what respects the charge was unlawful, because the adjourned trial before me in July 2013 was a full trial of the action, including issues of principle relating to quantification. After hearing full argument, from the same leading counsel as in the present case, and analysing the judgment in FII (ECJ) II, as well as earlier European case law, I concluded that the “dual credit” solution propounded by the test claimants was correct, and that the ECJ could not have intended in FII (ECJ) II to lay down the principle that the grant of a single credit at the FNR would have sufficed to ensure compliance with EU law: see Portfolio Dividends (No. 2) at [80] to [96].

The matter has now been argued before me a second time, with the Revenue in particular advancing several submissions which I do not recall having been made in Portfolio Dividends (No. 2). But Mr Ewart QC has not succeeded in persuading me that my previous conclusion was wrong. On the contrary, although I recognise that the question is by no means an easy one, I have upon reconsideration again come to the same conclusion. In those circumstances, I do not propose to treat the question at any great length (see my comments at [9] above), but will mainly confine myself to dealing with certain points which are not reflected in my earlier judgment.

Mr Ewart could hardly gainsay the settled line of European case law, including paragraph 39 of the judgment of the ECJ in FII (ECJ) II itself, which states that a Member State is in principle free to prevent (or mitigate) the economic double taxation of distributed profits by operating a hybrid system of exemption for domestic dividends and imputation for foreign dividends, provided that two conditions are satisfied. First, the “tax rate” applied to foreign dividends must be no higher than the rate applied to domestic dividends; and secondly, the tax credit must be at least equal to the underlying tax paid in the State of the company making the distribution, up to the limit of the tax charged in the recipient State. Mr Ewart’s primary submission was, instead, that this test was relevant only at the initial stage of determining whether the national legislation of the home State constituted a restriction on freedom of establishment and the free movement of capital which prima facie infringed Articles 49 and 63 TFEU.

Mr Ewart pointed out that in FII (ECJ) I the ECJ had not taken its analysis beyond this point, whereas in FII (ECJ) II the Court went on to consider the questions of justification and proportionality. After holding (at paragraphs 55 to 59) that the restriction in the UK legislation could be objectively justified by the need to ensure the cohesion of the national tax system, because the necessary direct link existed between the two limbs of the hybrid system, the Court then discussed the proportionality of the restriction. It was only at this stage (paragraphs 60 to 64) that the Court introduced its thesis that the exemption granted to domestic dividends had to be regarded as equivalent to the grant of a tax credit at the domestic nominal rate of tax, with its corollary that cohesion of the tax system could be maintained if a tax credit at the FNR were also granted for foreign dividends. Since the UK tax system failed to provide a credit for foreign dividends at the FNR, and since (as my findings in FII (High Court) I had established) the effective rate of corporation tax in the UK was generally lower than the nominal rate, it followed that the defence of justification failed (for a fuller exposition of my views on this difficult passage in the judgment, see Portfolio Dividends (No. 2) at [66] to [77]).

The Revenue’s case on this point is lucidly summarised in paragraph 37 of their skeleton argument, as follows:

“It can be seen from this analysis [of FII (ECJ) II] that the actual foreign tax paid on the underlying profits is only relevant at the stage of determining whether there is a restriction on freedom of establishment. Having concluded that there is such a restriction, the actual tax paid becomes irrelevant. At that stage, EU law requires the UK company to receive a credit computed by reference to the nominal rate of tax applied to the underlying profits in question in the relevant foreign jurisdiction. If the actual tax paid is less than tax at the nominal rate then the amount of the tax credit would have to be topped up to an amount calculated by reference to the nominal rate up to the level of the tax charged in the UK. Conversely, if the tax paid was greater than the tax computed at the nominal rate then the amount of the tax credit would be reduced to an amount calculated by reference to the appropriate nominal rate as EU law does not require credit to be given at any level beyond the foreign nominal rate. It is therefore pointless to refer to the actual tax paid in computing the tax that could lawfully have been charged.”

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.

Secondly, I agree with the claimants that the reason why the Court’s analysis in FII (ECJ) I 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 (ECJ) II, 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.

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 (ECJ) II was Vice-President Lenaerts, who had also been the rapporteur in FII (ECJ) I and in the Portfolio Dividends case.

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 Case C-446/04 [i.e. FII (ECJ) I], 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.

Mr Ewart sought further support for the Revenue’s argument in an example, which is set out as follows in the Revenue’s skeleton argument (immediately after the passage quoted in [31] above):

“38.

A simple example demonstrates that this must be correct. Suppose an EU company makes accounting profits of 1,000 in each of two consecutive years. The FNR is 20% and the rate of UK corporation tax is 30%. It makes a provision in its accounts for 200 in each year and so 800 is distributable in each year. However, due to timing differences in the tax computation, it pays 150 tax in the first year and 250 tax in the second year. It pays 800 of a dividend to its UK parent in each year.

39.

The [Revenue] would compute the lawful tax as 100 in each year. This is achieved by grossing up the dividend by the FNR to obtain 1,000 and applying the difference between the UK rate and the FNR (30-20=10%) to that grossed up amount. This totals 200 over the two years which is 10% of the total profits over those two years (2,000). This is exactly as expected.

40.

The Test Claimants would compute the lawful tax for the first year by taking a credit of 190 (20% of 950 which is the dividend of 800 grossed up by the tax paid of 150) instead of the actual tax credit of 150. This produces lawful tax of 95 (30% of 950, less 190). However, in the second year the Test Claimants would take a credit for the tax paid of 250 as it is higher than a credit at the FNR. This produces lawful tax of 65 (30% of 1050, less 250). This results in lawful tax over the two years of 160. This is 40 less than would be expected. It results from the Test Claimants’ method of taking the higher of the FNR and the foreign tax actually paid. This double counts part of the FNR for year one as it is reflected in the tax paid in year two.”

The response of Mr Aaronson QC to this example in his oral submissions was to confess and avoid. He accepted that the example accurately reflected the positions of the parties, and that the result might be regarded as anomalous. But he pointed out that it was an example of a fairly unusual nature, depending on timing differences in the UK company’s tax computation. He argued that any hybrid system of cross-border relief for economic double taxation was almost bound to give rise to anomalies at the margins, and stressed that this was the only anomaly that the Revenue had been able to find with all the resources at their disposal. He also countered the anomaly with one of his own, which (by contrast) he submitted was of a systemic nature. If the Revenue’s argument were correct, he said, it would follow that relief could never be granted for foreign underlying tax when it exceeded tax at the FNR on the relevant portion of the accounting profits of the company making the distribution. This could happen on a regular basis, particularly where the FNR was relatively low and the tax base was substantially greater than the company’s accounting profit (for example because certain types of deduction were disallowed for tax purposes).

Mr Ewart’s answer to this point was to say that in such cases the less favourable tax treatment was caused by the foreign tax system, and that the possibility of this happening was expressly recognised by the ECJ in paragraph 64 of its judgment. He also submitted that, in any event, the UK does not cause economic double taxation of the higher amount which is taxed in the source State. The UK only taxes the amount distributed which represents the accounting profits.

On balance, I do not consider that the competing arguments on anomalies take the matter much further. I agree with Mr Aaronson that some anomalies are only to be expected in a hybrid system, and that the Revenue’s example is one of a relatively unusual nature which does not show the dual credit solution to be fundamentally flawed. I am also inclined to agree with him that it would be surprising if the ECJ’s solution to the problem did not require credit to be given for underlying tax if it exceeded tax at the FNR. It is true that such cases are likely to arise only where the foreign tax base is larger than the distributable profits, and inequality of treatment of this nature was envisaged by the court in paragraph 64 of its judgment; but the dual credit solution is more likely to reduce such inequality than the single FNR credit, and that seems to me a point in its favour. It is important not to lose sight of the fact that the whole point of relieving economic double taxation is to ensure, as far as possible, that the same profits are in substance taxed only once in the hands of the company and its shareholders. To that end, relief is routinely given by the UK for underlying tax which is attributable to foreign dividends, and the ECJ will doubtless have had this in mind when repeatedly laying down the requirement that credit must be given for underlying tax if a hybrid system is to be compliant with EU law. At this point, however, the argument is in danger of becoming circular, which is why I do not think it ultimately takes matters any further.

To conclude, I am satisfied, for the reasons given in Portfolio Dividends (No. 2), as amplified in the present judgment, that EU law required the dual credit test to be satisfied if the Case V charge on the dividends in issue in the present case was to be valid, and Issue 1 should be answered accordingly.

Issue 2: what is the appropriate FNR?

I have so far discussed the “dual credit” required by EU law as if it were obvious which FNR should be adopted when a dividend was paid by a water’s edge EU subsidiary to its UK parent. There should indeed be no room for doubt in the relatively unusual case where the profits distributed were themselves generated in the EU water’s edge company. In such a case, the only candidate for the FNR for which a credit must be granted is the FNR applicable to the company which paid the dividend. But what of more complex cases, for example where the dividend was paid out of profits which had originally been subject to “underlying” tax in one or more countries further down the offshore corporate structure (or “corporate roots”), and which may not have been subject to tax at all in the water’s edge country?

A particular example of this kind of case which often arose in practice was the so-called “offshore pooling” in a “mixer” company (typically incorporated in the Netherlands) of dividends from countries with rates of corporation tax which were both higher and lower than the UK rate, so that blended dividends could be transmitted on to the UK carrying an average rate of creditable underlying tax which was as close as possible to the cap set at the UK nominal rate: for a brief explanation of the system, see FII (High Court) I at [102] to [103]. BAT Nederland BV was an example of a Netherlands holding company of this type. Furthermore, where the underlying dividends were paid by subsidiaries (or any other companies in which the “mixer” company had a holding of more than 10%), they would typically be exempt from corporation tax in the hands of the mixer company by virtue of a so-called “participation exemption”. Thus the FNR applicable to the blended profits from which the onward dividends were paid would be nil, at least if it is right to regard the exemption from tax as equivalent to taxation at a nominal rate of zero.

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:

In Portfolio Dividends (No. 2) I adopted the first of these solutions, for the reasons briefly given in [108]. I said that this seemed to me to be the comparison which the ECJ had in mind when in FII (ECJ) II it contrasted the absence of such a credit for foreign portfolio dividends with the (notional) credit at the full UK nominal rate which was implicit in the exemption for domestic dividends contained in section 208 of ICTA 1988. I then said:

“I cannot imagine that the ECJ envisaged the enquiry into nominal rates extending beyond the state of residence of the source company. The enquiry should in principle be a simple one, which can normally be answered by looking at the published tax legislation of the source state.”

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.

Nor, in my view, can any very clear guidance be obtained from the reasoning of the Court in FII (ECJ) II. Given the largely theoretical nature of the Court’s analogy between the exemption of a dividend from tax and the grant of a credit at the relevant FNR, it could well be that the ECJ was content to take the water’s edge FNR as the one for which credit should be given, regardless of whether the distributed profits were themselves subject to tax at that stage. This solution has the great merit of simplicity, and (as I said in Portfolio Dividends (No. 2)) seems to me the situation which the ECJ primarily had in mind. This may, however, merely reflect the fact that, when considering question 1, in the order for reference, the ECJ (as it had done in FII (ECJ) I) concentrated on the simple water’s edge scenario, before widening its focus to consider the “corporate tree” structures in its discussion of question 3.

There is also force in the point made by the Revenue that the reference in paragraph 62 of the judgment to “the nominal rate of tax to which the profits underlying the dividends paid have been subject” (my emphasis) suggests that the relevant FNR is one which applied when the profits were in fact subject to tax (the same language is then repeated in paragraphs 72 and 86). This in turn would fit in with the Court’s analysis of the vice which it perceived in the UK’s system of corporate taxation, namely the failure of the imputation system for foreign dividends to mirror the fact that the effective rate of UK corporation tax was normally significantly lower than the nominal rate. Since the ECJ considered that the appropriate solution to this problem was to top up the credit for actual underlying tax with a credit at the level of the FNR, it would make good sense to take as the FNR for that purpose the FNR which applied when the profits were in fact charged to tax.

On balance, I have come to the conclusion that in principle the appropriate FNR to apply is that of the jurisdiction in which the distributed profits were in fact subject to tax, in preference to a blanket application of the FNR applicable to the water’s edge company which made the distribution. I consider that this solution better accords both with the language of the ECJ (in paragraphs 62, 72 and 86) and with the rationale for requiring a credit at the FNR at all. In the case of portfolio dividends, the practical simplicity of solution (a) was conclusive; but in the very different context of dividends paid up through a single multi-national group, the appeal to simplicity should not in my judgment be decisive.

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.

In a little more detail, the claimants’ methodology builds on and adopts the existing machinery (contained in section 801 of ICTA 1988) which provided relief for actual underlying tax by (in effect) treating all tax paid on the underlying profits further down the corporate “roots” of the group structure as if it had been paid by the company making the distribution, and then applying a weighted average to ascertain the global rate of actual underlying tax attributable to the blended dividend. If the result of applying these rules was that no Case V tax was in fact charged, the claimants say that the need to grant a further credit at the FNR cannot be implemented in a way which would undermine the relief which had actually been granted. The Revenue, however, 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 difference between the two approaches is well brought out by a simplified example given by counsel for the Revenue in their written closing submissions. The example assumes an EU holding company (which I will call H) which has one subsidiary resident in Germany (where the FNR is 40%) and another subsidiary resident in Ireland (where the FNR is 10%). Each subsidiary has profits of 100, on which the German subsidiary bears tax of 40 and the Irish subsidiary bears tax of 10. H therefore receives dividends of 150 (60 from the German company and 90 from the Irish company). The dividends are exempt from tax in the hands of H, but the FNR applicable to H would be 30%. H then pays on the blended dividend to its UK parent, P, in whose hands the gross dividend of 200 is in principle subject to corporation tax at 30% (i.e. the UK nominal rate).

The result of the claimants’ approach is that the higher FNR on the German profits is combined with the lower Irish FNR, before comparison is made with the UK rate of corporation tax. The lawful Case V charge would therefore be 10 (i.e. 200 charged at the UK rate of 30%, less the combined tax credit of 50). The Revenue, by contrast, say that if the German and Irish subsidiaries had paid the dividends directly to the UK, then the lawful Case V tax would have been nil in respect of the German profits (because the FNR of 40% exceeds the UK corporation tax rate of 30%), but 20 in respect of the Irish profits (100 charged at 30%, less the Irish tax credit of 10). According to the Revenue, the same result should ensue when the dividends are paid indirectly to the UK through H.

Incidentally, it is also worth pointing out that, on the facts of this example, the claimants’ first approach (which I have rejected) would provide a more favourable outcome for them than their alternative case. If it were right to take the FNR applicable to the water’s edge company, H, which is assumed to be the same as the UK corporation tax rate of 30%, the credit would eliminate the whole of the UK charge. This may be thought a rather surprising result, given that the dividends were in fact exempt from tax in the hands of H.

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.

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 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 section 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.

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 Article 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 Article 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.

In appendix III to their skeleton argument, the claimants formulate two further principles which they submit should be followed on their alternative approach, although they do not arise in practice in the test cases because of the amounts of EU actual underlying tax. The first principle is that, if the profits are taxed more than once in the EU on their way to the UK, the appropriate nominal rate is that of the intermediate country in the event that it is higher than the rate in the country where the profits were first taxed. The second principle relates to third-country income, and says that where it is paid up through an EU company where it is taxed (as opposed to being exempt), that income should carry a credit at the FNR of the country where it is taxed. My provisional view is that both these principles are correct, and should be followed in any case where they arise in practice.

To conclude, for the reasons which I have given, and subject to the qualifications which I have mentioned, I consider that the appropriate FNR to adopt in calculating the credit required by EU law is the FNR ascertained in accordance with the claimants alternative case, i.e. solution (b) in paragraph [43] above.

The relevant FNRs: issues of fact

The claimants have asked me to make findings of fact on what the relevant, or potentially relevant, FNRs actually were. For that purpose, they have adduced evidence of the nominal rates of corporation tax which were in force at the material times in all the potentially relevant countries, including those from which the foreign profits were ultimately derived. The List of Issues does not include any specific factual questions asking me to determine what particular FNRs actually were, as opposed to questions about methodology or about the special cases which I will come on to consider in the next section of this judgment. I infer that there is no real dispute between the parties about the headline nominal rates of corporation tax which were actually in force in the countries concerned at the relevant times.

This impression is reinforced by paragraph 26 of the agreed statement of facts, which records the parties’ agreement that, apart from the exceptions mentioned in paragraph 28, the FNRs which appear in the tables at exhibit “AMHC 9” to Mr Cohn’s sixth statement dated 11 November 2013 “are a correct statement of a foreign nominal rate in the jurisdictions and for the accounting periods indicated”. The purpose of those tables was to quantify the claims of the BAT test claimants by amending and updating the spreadsheets attached to the claimants’ particulars of quantification dated 23 September 2013. Those particulars used the nominal rates applicable to companies lower down the corporate roots, rather than those applicable to the water’s edge companies. In practice, much of the necessary work had already been performed for the purposes of the claims in Portfolio Dividends (No. 2): see my judgment in that case at [109] to [111]. The further work carried out for the purposes of the present case is described in the evidence, and was further reviewed by a team at PricewaterhouseCoopers LLP (“PwC”) working under the direction of a senior corporate tax partner, Mr Nicholas Woodford. Mr Woodford is one of the claimants’ expert witnesses, and in Part 1 of his report dated 9 December 2013 he explains the approach which was taken to the assessment and verification of the nominal rates. The fruits of his team’s work are set out in appendix 2 to his report, supported by notes for each country.

The exceptions set out in paragraph 28 of the agreed statement of facts relate to the special cases which I have mentioned, and to the dividends received by foreign holding companies which were subject to a participation exemption. That apart, no issue is taken by the Revenue with the work done by the claimants in ascertaining the relevant FNRs, apart from a bald statement in paragraph 26 of the agreed statement that the Revenue “do not agree that these were the FNRs applicable to dividend (as opposed to trading) income”. This reservation was not developed or amplified during the hearing, save for a statement by Mr Baldry QC on day 7 (Transcript, pp 38-39) that it was intended to cover the Revenue’s position that the relevant rate to take is the one that was actually applied to a dividend. On this point, Mr Woodford confirmed in cross-examination that a schedule of “special nominal rates” appended to his report was intended to show whether there were special rates attributable to particular classes of income, but the information given was not exhaustive and might depend on how the question had been put to PwC’s local office in the country concerned. In general, the main focus of the exercise carried out by PwC had been on the headline nominal rates of corporation tax applicable to companies generally.

Subject to what I have just said about special nominal rates, and subject to the special cases which I will now consider, I see no reason to doubt that the work carried out by the claimants to ascertain the relevant nominal rates of tax has been as thorough and accurate as could reasonably be expected. In the absence of any specific challenges to it by the Revenue, I consider that it should be accepted, and that the relevant headline rates of nominal tax should (unless the parties agree otherwise in any particular case) be taken as those set out in appendix 2 to Mr Woodford’s report.

Issue 3: special cases

I must now deal with the specific situations, or special cases, which are identified in paragraph 28 of the agreed statement of facts. I will begin with the last of them, namely the treatment of dividends received by EU holding companies which were subject to a participation exemption, both because it is taken first in the List of Issues (Issue 3(a)) and because it forms a natural sequel to the treatment of EU mixer companies, which I have already discussed.

Participation exemptions, and the GKN test claim

As noted above, the GKN test claim has been added as a supplementary test case in order to deal with the position where dividends are paid to the UK from non-resident EU intermediate holding companies which benefit from either a full or a partial participation exemption, and where the dividends in question are attributable to underlying profits earned in various jurisdictions, both in the EU and third countries. (Here, as elsewhere, where I refer to the EU, the reference should be read as including countries in the EEA.) The facts relating to the GKN test claim are agreed, and are set out in paragraphs 55 to 58 and 60 to 67 of the agreed statement of facts. Those paragraphs are reproduced, without the accompanying schedules, in appendix 1 to this judgment. The summary which follows is mainly based on appendix V to the claimants’ skeleton argument.

The ultimate parent company of the GKN group was GKN Holdings Plc (“GKN Plc”), which was incorporated and resident in the UK. The group operates through subsidiaries established throughout the world, largely in the automotive, aerospace and defence industries. In the period relevant to this claim, the group operated through a large network of subsidiaries, both in the UK and abroad. The participation of GKN Plc in its foreign subsidiaries was either direct or indirect, through UK resident intermediate parents or offshore holding companies. Four sample dividends have been taken to illustrate the questions to which this structure gave rise.

Three of the sample dividends were paid to GKN Plc by GKN Netherlands BV, an intermediate holding company resident in the Netherlands, in the accounting periods ended 31 December 1980, 1981 and 1982. Those dividends were mainly sourced in the profits of subsidiaries earned in other jurisdictions, both within and outside the EU. GKN Netherlands BV was a mixer company which, as I understand it, performed a similar role to BAT Nederland BV. The underlying tax rate attributable to the dividends which it paid to GKN Plc was established by a weighted average pursuant to section 801, and was sufficient (in combination with other reliefs) to negate a Case V charge on the dividends in the hands of GKN Plc.

The fourth sample dividend was paid to A P Newall and Co Ltd, a UK resident company within the GKN group, by Uni-Cardan AG, an intermediate holding company resident in Germany, in the accounting period to 31 December 1986. The profits thus distributed were the profits of Uni-Cardan AG for the preceding accounting period ending 31 December 1985. Those profits incorporated dividends which had been received directly or indirectly from a variety of companies resident in Germany, other EU Member States and third countries. The profits underlying the dividends had been earned either by the distributing companies themselves, or by their subsidiaries resident in the EU or third countries.

Uni-Cardan AG was the top company in a German consolidated tax group known as an “organschaft”, and as such it benefited from a partial participation exemption. Some of the dividends which it received were taxable in Germany, but others were exempt under German tax treaties, while under the treaties with Italy and Austria the dividends remained taxable but with the grant of a tax credit. A further complication in the case of the Austrian and Italian subsidiaries was that they were only half-owned directly by Uni-Cardan AG, with the balance of the shares being owned by other German intermediate subsidiaries. In consequence, the portion of the dividends of the Austrian and Italian subsidiaries paid to those other German subsidiaries was also subject to tax in Germany. In addition, part of the 1985 profits of Uni-Cardan AG included its own earnings upon which it paid German tax. As the top company in the organschaft, it also paid German tax on the profits of the other companies in the organschaft.

I received evidence on the operation of the German tax system in relation to foreign-sourced dividend income, both before and after the introduction in Germany of a dividend exemption system in 2000, and about the way in which tax consolidation was effected through an organschaft, from Dr Axel Karl Bödefeld. Dr Bödefeld is a partner in a German law firm in Cologne, with a doctorate in tax law from the University of Bochum. Despite the expert nature of some of his evidence, Dr Bödefeld was called as a witness of fact. He gave his oral evidence by video-link, and was briefly cross-examined by Mr Baldry. He gave his evidence with great clarity, in excellent English, and I have no hesitation in accepting it. Among other matters, he explained that:

The issues raised by the first, second and third sample dividends are essentially the same as those which I have already considered in relation to the FNR, in cases where dividends were channelled through an EU mixer company which itself benefited from a full participation exemption from tax on dividends paid to it by its subsidiaries. As I have explained, I consider that the nominal rate credit for EU-source dividends should be given at the level where the relevant profits were in fact charged to tax, and not at the level of the mixer company (here GKN Netherlands BV) where they were exempt, but that the Revenue’s methodology should not be permitted to generate a (notional) liability to Case V tax when such a charge was in fact negated by the credit given for actual underlying tax pursuant to the machinery of section 801.

There is one further argument advanced by the claimants in favour of their primary case which I have not yet noted. The argument is that the FNR of the mixer company should be taken, despite the participation exemption, because the right way to regard the exemption is as a narrowing of the relevant part of the tax base to zero, but still subject to the headline nominal tax rate. Whether or not it is theoretically possible to characterise the effect of a participation exemption in these terms – a question which it seems to me could not be answered without expert evidence, and might yield different answers in different jurisdictions – I would reject the argument. I consider that the ECJ’s reasoning requires the nominal rate credit to be ascertained at a level where the EU profits were in fact charged to tax, and not at a level where any charge, assuming it to be otherwise applicable, was automatically negated by an exemption which reduced the tax base to nil.

There is a further point relating to evidence, which the claimants have raised in the context of the GKN test claim. On the footing that credit should in principle be given on the basis set out in appendix III to the Revenue’s skeleton argument (i.e. the Revenue’s alternative case, which I have accepted), it is possible that evidence would be required in relation to lower tier companies going beyond the evidence that was required (and agreed with the Revenue) for the grant of double tax relief. In such circumstances, the claimants submit that the nominal rate of the lowest tier company for which they do have evidence should be taken as a proxy. The submission is based on pragmatism rather than principle, but is perhaps none the worse for that. Bearing in mind the very lengthy period of the claims, the fact that the need to grant a credit at the FNR has only recently emerged in the ECJ’s case law, and the desirability of avoiding complex and expensive factual enquiries, I am prepared to accept the submission, subject to two qualifications. The first is that the nominal rate must be that of a company in whose hands the relevant profits were in fact charged to tax. The second is that if a compelling case can be made out for taking the enquiry any further in a particular case, and if this can be done at a proportionate cost, the parties may agree to do so (or, in default of agreement, apply for permission to do so).

As to the fourth sample dividend, it seems to me that calculation of the FNR credit will require separate treatment of (at least):

I am conscious that I have given this guidance at a fairly high level of generality, but I hope that in the light of it the parties will be able to agree how the exercise should be performed. Here, as elsewhere, there will of course be liberty to apply if agreement cannot be reached.

The Henri Wintermans sale

This issue concerns dividends paid by BAT Nederland BV which were associated with the sale of two subsidiaries outside the group, Henri Wintermans Sigarenfabrieken BV and Velasques Sigarenfabrieken NV. These subsidiaries manufactured cigars, and were sold in or around 1996 to the Danish Skandinavisk group. The 1996 interim and final dividends of BAT Nederland BV included the distribution of a capital gain arising on the sale. This capital gain did not form part of the tax base for corporation tax in the Netherlands, apparently on the basis of some form of participation exemption. In any event, the Revenue contend that the gain should be excluded from the present claim. Since it did not form part of the taxable profits of BAT Nederland BV, it was never subject to economic double taxation, and the Case V charge should therefore apply to it without modification.

The claimants dispute this analysis, on the basis that the nominal rate of corporation tax in the Netherlands applied to the gain, even if it was removed from the tax base. They point out that, in his expert evidence for the liability trial, Mr John Whiting of PwC identified untaxed capital profits as one of the causes which commonly enabled UK companies to pay corporation tax on their accounting profits at an effective rate below the nominal rate: see paragraph 3.7 of his report, where he found that this was a cause of a rate reduction in approximately 3% of the sample cases which he had examined. Since the purpose of granting a credit at the FNR is to reflect the difference between nominal and effective rates of tax, it would be paradoxical, say the claimants, to disallow their claim in a case which illustrates that distinction.

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.

The Revenue also challenge the inclusion in the claim of a further dividend in 1998, to the extent that it represented a release to BAT Nederland BV’s distributable reserves, following the sale of the two subsidiaries, of a statutory reserve which it had been obliged to set up in 1990. The creation of the reserve came about as a result of a change of accounting policy required by Netherlands law. Until the end of 1989, BAT Nederland BV accounted for its investments on an equity accounting basis, which meant that the balance sheet value of its investment in the two subsidiaries reflected its share of their net asset value, including retained profits. From 1 January 1990, BAT Nederland BV had to account for the investments in its subsidiaries on a historic cost basis, determined on the basis of their net asset value at 31 December 1989. The difference between that net asset value and the original historic cost was accounted for by creating a statutory reserve.

According to the claimants, the reserve represented BAT Nederland BV’s share of the accumulated undistributed profits of the subsidiaries at that date, and any dividends subsequently paid out of pre-1990 profits would reduce the reserve. The undistributed profits comprised in the reserve would originally have been subject to tax in the Netherlands, although in the hands of the subsidiaries rather than BAT Nederland BV. Those profits were represented in the 1998 dividend payment made by BAT Nederland BV, and should accordingly carry a rate of tax equivalent to the Netherlands nominal rate.

The Revenue disagree, arguing that the statutory reserve did not represent the retained profits of the subsidiaries, but rather represented the increase in value of the subsidiaries in the balance sheet of BAT Nederland BV from the date of their original acquisition until the end of 1989. The exempt gain on the subsequent sale of the subsidiaries was equal to the accounting profit made on the sale, together with the release of the statutory reserve. It follows, say the Revenue, that the statutory reserve did not represent the taxed profits of the subsidiaries, but was rather part of the gain on disposal which was exempted from corporation tax. The profits of the subsidiaries were indeed subject to tax in the Netherlands, but those profits were not distributed to BAT Nederland BV either before or as part of the sale. The 1998 dividend was therefore paid out of exempt profits, and does not qualify for any credit.

I am again handicapped in dealing with this point by the lack of any clear evidence about the precise nature of the exemption which applied on the sale of the subsidiaries, or the nature of the statutory reserve. In addition, neither side dealt with the question in oral argument. It seems reasonably clear, however, that the statutory reserve was included in the gain on disposal which was exempted from corporation tax. If that is right, I can see no good reason for treating the reserve differently from the rest of the gain, i.e. as part of the taxable profits but subject to a base adjustment in the form of the exemption. I agree with the Revenue that it is probably better to regard the reserve as a component in the computation of the gain on the disposal, but if I am right about how the 1996 interim and final dividends should be treated, the same conclusion should in my judgment follow in relation to the 1998 dividend.

Belgian co-ordination centres

Belgian co-ordination centres were companies established in Belgium which benefited from extremely favourable tax treatment, both for themselves and their shareholders. I was told that these advantages were later held to contravene EU law as a form of unlawful State aid, but before then such centres attracted much foreign investment. Two are relevant for the purposes of the BAT test claim: a company called Haseldonckx Logistics SA (“HXL”), which paid dividends to its minority shareholder BAT Benelux SA (“BAT Benelux”, also resident in Belgium) in 1988 and 1989; and BAT Co-ordination Centre NV (“BCC”), which paid dividends to its shareholders in subsequent accounting periods. At the material times, both HXL and BCC were wholly-owned subsidiaries of the BAT group. The majority of the shares in HXL were owned by the Wiggins Teape group, which left the BAT group in 1990. This explains why the dividends paid by HXL to Wiggins Teape do not feature in the present claim. It was the departure of Wiggins Teape from the BAT group which led to the formation of BCC as a successor co-ordination centre.

The immediate parent companies of BCC were at all material times BAT Benelux (which held between 100% and 79.7% of the shares at different times) and Velasques Sigarenfabrieken NV (“Velasques”), resident in the Netherlands, which held the balance of the shares in BCC until its sale outside the group in 1996 (see [75] above), when it was replaced as the minority shareholder by the UK-resident BAT Co Ltd. The majority shareholder of both BAT Benelux and Velasques was BAT Nederland BV, which was itself a wholly-owned subsidiary of Westminster Tobacco Ltd (UK resident); the minority shareholder in each case was BAT Co Ltd. In summary, therefore, the profits distributed by the two co-ordination centres (HXL and BCC) mostly reached the UK water’s edge companies (BAT Co Ltd or Westminster Tobacco Ltd) indirectly, via companies resident in Belgium (BAT Benelux) and/or the Netherlands (Velasques and BAT Nederland BV). After 1996, BCC also paid dividends directly to its minority shareholder, BAT Co Ltd.

The investment of the BAT group in HXL attracted the attention of the Revenue, in the context of an investigation under the controlled foreign companies legislation, and on 28 April 1992 Mr Hardman wrote to the inspector of taxes dealing with the matter, Mr M J Pigott, in order to explain the role of HXL in the group. Mr Hardman said that HXL was engaged in the provision of administration and support services, and the provision of financial credit, to group companies, mostly operating in Belgium, France and Italy. Most of HXL’s profit came from loans made to group companies, and nearly all of its interest income in 1989 came from credit extended to its Belgian shareholders, which were BAT Benelux and a company in the Wiggins Teape group. Those companies had invested in the share capital of HXL, and HXL had then lent the funds back to the shareholders. The letter continued:

“The Belgian companies obtained tax deductions for the interest on their external borrowings and for the interest on the borrowings from HXL. Under the Belgian tax rules applicable to Co-ordination Centres HXL did not pay tax by reference to the conventional base of net income. Instead it was taxed by reference to its expenses excluding salaries and wages and interest costs, i.e. a very small taxable base. The interest income of HXL was not taxed.

Dividends payable by HXL to its Belgian shareholders were largely exempt in the hands of the recipients. Only 10% of the dividend was taxable to the Belgian shareholder.

Furthermore the 1989 dividends of HXL qualified for a fictitious withholding tax credit (FWT). Although no withholding tax was payable by HXL on its dividend payments the dividends were treated in the hands of the recipients as if there had in fact been a withholding tax. The FWT was available to set against not only the tax payable on the 10% of the HXL dividend which was taxable but also the other taxable income of the Belgian shareholders.

The combination of these factors provided a very cheap and effective mechanism for financing Belgian trading companies and BAT Benelux became a shareholder in HXL to avail itself of this financing mechanism.”

This letter was put to Mr Hardman in cross-examination, and he confirmed that it accurately reflected his understanding of how the system operated. He also agreed that the system would have applied to BCC in the same way.

Further details of the way in which the co-ordination centres operated and were taxed may be found in a report of the Code of Conduct Group (Business Taxation) of the EU Council submitted to the ECOFIN Council on 29 November 1999. Annexe A001 to the report contains a description of the rules with which the centres had to comply, and the tax benefits which they enjoyed. The rules required a co-ordination centre to form part of a multi-national group of specified size, and it was only permitted to conduct certain authorised activities, all of which had to be within the group. The permitted activities were either of a preparatory and auxiliary nature, or consisted of the rendering of financial and other similar services for the sole benefit of other members of the group. The centre was not permitted to acquire shares in other companies, and it had to employ at least 10 full-time employees in Belgium. The description of the tax benefits included this passage:

“Belgian co-ordination centres are liable to corporate income tax at the normal (40.17%) rate, but (instead of the actual profits as shown in its financial statements) only on a notional tax base determined as a percentage of certain operating costs incurred by them. Certain items, such as personnel costs, financing costs and taxation are excluded from this base.

The percentage depends on the mark-up charged to the affiliated group companies and on the type and nature of the co-ordination centre’s activities. In the absence of objective criteria, the mark-up percentage will be fixed at 8 per cent.”

It can be seen from this description that the notional tax base, although subject to charge at the normal Belgian rate of corporation tax, was of a highly artificial and restricted nature. It was calculated as a fixed percentage of certain operating costs, and excluded from charge altogether the loan interest which provided HXL with most of its profits. Mr Hardman referred in his letter to “a very small taxable base”, and consistently with this the Revenue have calculated, in a paper which was produced during the hearing as part of their written submissions, that the underlying effective rate of tax for BCC for 1994 was only 2.12%.

Against this background, I now have to decide what FNR credit would have been required by EU law in respect of the relevant dividends paid by HXL and BCC. Two points can be disposed of at the outset. First, for the reasons I have already given, I would reject the claimants’ primary case that the relevant FNRs are those of the Belgian or Dutch water’s edge companies. The dividends were wholly exempt from tax in their hands, whether as a result of BAT Nederland BV’s participation exemption, or as a result of the available withholding tax credit in Belgium (which, as we have seen, was designed to match and extinguish the reduced Belgian tax charge on the dividends). Secondly, there can be no question of going any further down the corporate roots than the co-ordination centres, because they were themselves (as Mr Hardman put it) the bottom tier company. The profits which they distributed were their own, and they were not permitted to own shares in other companies, so they had no dividend income.

The claimants’ alternative submission is that the appropriate nominal rate is the Belgian nominal rate, on the ground that this was the rate applicable to the profits of the two centres in Belgium, albeit with a very restricted tax base. This submission gains some support from the description of the charge to tax, both in the EU Council report and in Mr Hardman’s letter of 28 April 1992. As a matter of analysis, the centres were charged to tax at the full nominal rate although on a hugely reduced tax base. On the other hand, it may be said that the tax treatment of the accounting profits under the co-ordination centre regime went far beyond any conventional restriction of a tax base on account of reliefs or other causes of the kind identified by Mr Whiting in his report in 2008. In truth, the centres were taxed on a wholly artificial tax base which bore no direct relation to the accounting profits, but was instead ascertained by applying a fixed percentage mark-up to certain limited categories of costs. Circumstances such as these seem to me very far removed from anything which the ECJ would reasonably have contemplated when laying down its requirement for a credit at the FNR in addition to a credit for underlying tax.

It would in my judgment be unacceptable if the BAT group, having enjoyed all the fiscal advantages of investing in the Belgian centres, were also to be entitled to a credit at the full Belgian nominal rate of 40.17% on the full amount of the distributed profits. The situation is on any view an anomalous one, and I think it calls for a bespoke solution. Two are advanced by the Revenue. The first is that the FNR should reflect the effective rates of tax applicable under the special regime. The second is that the full Belgian FNR should be applied, but only to the proportion of the profits which were subject to tax in Belgium under the regime. The remainder should be regarded as exempted profits which have not been charged to tax. I am not attracted by the former of these proposals, because the whole point of a nominal rate credit is that it should compensate for the fact that corporation tax is usually levied at an effective rate lower than the nominal rate. The suggestion that the FNR should reflect the effective rate is therefore conceptually incoherent. On the other hand, the second proposal seems to me to provide an appropriate solution to the problem, and I would adopt it. It preserves the grant of a credit at the full FNR, while recognising the economic reality that the only profits brought into charge to tax were those comprehended in the artificially restricted tax base.

German “silent partnership” profits

This issue concerns income received by a UK company in the BAT group, International Tobacco (Overseas) Ltd (“Overseas”), from its interest in two German partnerships, the first with BAT Cigaretten-Fabriken GmbH (“BAT CF”), the second (and subsequent) one with BATIG GmbH (“BATIG”). BATIG was the top company of BAT’s German sub-group and its tax-consolidation “Organschaft” (on which see [68] to [69] above). BAT CF was a wholly-owned subsidiary of BATIG, and a member of the Organschaft.

The partnerships were of the type known as a “stille gesellschaft”, or silent partnership. A description of the main features of such a partnership may be found in Memec Plc v IRC [1998] STC 754 (CA) at 759f-h per Peter Gibson LJ, quoting the description given by Robert Walker J (as he then was) at first instance, [1996] STC 1336 at 1345. The effect of a silent partnership was to relieve the German owner of the business from liability to federal German corporation tax on the silent partner’s share of the profits, although the owner remained subject both to German withholding tax and (if itself a trading company) to German trade tax (imposed at the Land, or municipal, level) on the profits.

By an agreement dated 7 January 1974, as amended on 22 April 1974, Overseas agreed with BAT CF to make a capital contribution to BAT CF of DM 80 million, in consideration of which Overseas would receive 20% of the after tax profits of BAT CF every year (but capped at a return of 6% per annum over bank base rate upon its contribution) and would assume 10% of the losses of BAT CF (up to the amount of the contribution) until the capital contribution was paid out. I assume that the agreement subsequently entered into between Overseas and BATIG was of a similar nature.

The income received by Overseas under the agreement was taxable in the UK under Schedule D Case V, as income arising from a foreign possession, with credit allowed for withholding tax and the German trade tax, but not for German corporation tax on the underlying profits. As the decision in Memec makes clear, the silent partnership was not “transparent” for UK tax purposes, and the agreement itself was the source of Overseas’ share of the profits: see [1998] STC 754 at 766a-e. The profits were, however, “dividends” within the wide definition of that term in Article VI of the double taxation agreement between the UK and Germany, and as such were subject to withholding tax as permitted by Article XVIII (1)(a): see Memec at 760g-761f and 768d-e. Relief for corporation tax on the underlying profits was excluded, because the profits were not “dividends” for the purposes of Article XVIII (1)(b), where the term has its ordinary meaning in UK law: see Memec at 768g-769b. The basis upon which the Revenue allowed credit for the German trade tax was not explained to me, but perhaps it was regarded as sufficiently analogous with federal withholding tax to justify the grant of relief in order to avoid what, before Memec, might have appeared to be juridical double taxation of the same trading profits.

In these circumstances, the Revenue argue that EU law did not require any credit to be given by the UK in respect of the shares of profit received by Overseas. The relief of juridical double taxation is a matter for national law to determine. Nor is there any objective comparability between a shareholder in a company and a silent partner in a German partnership, so there can be no objection in principle under EU law if the UK chose to treat the two situations differently. Accordingly, it is said, the claimants have no claim under EU law in respect of these receipts.

The claimants’ answer to this argument is to submit that EU law is relevantly engaged, because the UK charged ACT on the onward distribution in the UK of profits originating from the German partnerships. In the eyes of EU law, ACT is an advance payment of MCT payable by the company which makes the distribution. This charge therefore represented economic double taxation of the relevant profits, once in Germany and again in the UK, comparable with the economic double taxation of EU dividends. It is immaterial for this purpose that the foreign tax borne by the profits was the local trade tax rather than federal corporation tax. What matters is the imposition of two separate levels of taxation on profits which in a purely UK context would have been subject to only one charge to corporation tax within the group. Nor could it be right, say the claimants, to regard the ACT as somehow paid on account of the German trade tax, because under the UK/German double taxation agreement the right to tax the profits of the partnership was exclusively allocated to Germany.

In my judgment the Revenue’s submissions on this question are to be preferred. Even though the relevant profits were, from one point of view, subject to economic double taxation in Germany and the UK, I can see no answer to the simple point that there is no objective comparability between investment in a German subsidiary on the one hand, and investment in a German silent partnership on the other. Both legally and economically, they are different forms of investment. For example, the former engages the EU freedom of establishment, whereas the latter does not (in the absence of any shareholding by Overseas in either BAT CF or BATIG) and can engage only the free movement of capital. As a matter of substance, the distribution of profits by a subsidiary to its parent is a very different matter from the payment of a share of profits to a partner pursuant to the terms of a silent partnership agreement. It is for individual Member States to determine how those two different forms of economic activity should be taxed, and since they are not objectively comparable no question of difference of treatment can arise.

I therefore conclude that the claimants have no maintainable claim in respect of the receipts from the silent partnerships. I would only add that, if I am wrong in this conclusion, it is common ground that the appropriate nominal rate credit should be calculated by reference to the nominal rate of the trade tax imposed on the profits, rather than the nominal rate of federal corporation tax.

Arenson Group Plc

This issue lies within a very narrow compass. It concerns the profits of a UK company, Arenson Group Plc (“Arenson”), which ran a furniture wholesale business. Arenson was at the material time a wholly-owned subsidiary of the Danish Skandinavisk group in which BAT held a minority equity share of about 30%. The dividends paid by Skandinavisk to BAT in 1989 and 1990 incorporated a relatively small amount of income originating from Arenson, which Arenson had paid up as dividends to its Danish parent. The question is whether the claimants are entitled to bring a claim in respect of that portion of the 1989 and 1990 dividends.

The claimants say that there is no reason to distinguish between the portion of the dividends which originated from Arenson and the rest of the distributed income which was of Danish origin. In each case the income had been subject to underlying tax in the EU, and it was liable to economic double taxation by virtue of the subsequent imposition of the Case V charge in the UK on the BAT water’s edge recipient. The Revenue disagree, arguing that the UK was entitled to charge the full amount of corporation tax and ACT in respect of profits made in the UK by a UK subsidiary, even where those profits passed through an intermediate EU holding company.

I agree with the claimants. Nobody could question the right of the UK to tax the profits made by Arenson in its hands, or to levy ACT on Arenson when it distributed those profits by way of dividend to its shareholders. But that is not the point. The question in the present case is whether the UK was entitled to subject to the Case V charge the dividends paid by Skandinavisk to BAT, and to levy ACT on the onward distribution of those dividends in the UK, without granting (at least) a credit at the Danish nominal rate of corporation tax for EU source income comprised in the dividends. For those purposes, it seems to me entirely irrelevant whether the EU income comprised in the dividends originated in the UK or in any other Member State. The defect of the UK system, as diagnosed by the ECJ, lay in its failure to grant a nominal rate tax credit for BAT’s EU source dividends. The only relevance of the original charge to UK corporation tax on Arenson’s profits is that the amount of such tax could form part of the credit for underlying tax which BAT was entitled to claim on the dividends which it received from Skandinavisk.

For these short reasons, the claimants are in my judgment entitled to maintain their claim in respect of the profits which originated in Arenson.

Issue 4: how should the lawful Case V charge be computed?

Issue 4 asks how to compute the amount of Case V tax which could lawfully have been charged in respect of each foreign dividend. To a large extent, the answer to this question depends on the answers to questions which I have already considered. Thus I have already decided that EU law required a credit to be given for actual underlying tax as well as a credit at the FNR, and I have dealt with various issues about which FNR or FNRs should be adopted. But there remain two questions of principle which I have not yet addressed, and which are raised under Issue 4.

Grossing up

The first question relates to grossing up when calculating the FNR credit. The essential difference between the parties is whether (as the claimants say) the credit should be ascertained by applying the relevant FNR to the income as originally calculated for the purposes of the Case V charge, or whether (as the Revenue say) the income should be grossed up at the FNR before the credit is calculated by applying the FNR to the grossed up amount.

The claimants’ approach bases itself on the actual Case V computation, under which the net dividend received by the UK water’s edge company was grossed up by (i) any foreign WHT, and (ii) any foreign underlying tax, in order to ascertain the gross Case V income. The UK corporation tax rate was then applied to this amount, before giving credit for WHT and/or underlying tax, leaving a net liability to UK corporation tax. This procedure followed the rules governing the grant of relief by way of credit in Chapter II of Part XVIII of ICTA 1988: see in particular the grossing-up provisions in section 795(1), and section 793 which states the basic principle (reflected in sections 788 and 790) that double taxation relief is given in the UK by means of a reduction in the UK tax chargeable, setting the creditable foreign tax against the UK corporation tax.

All that is needed to make this system compliant with EU law, submit the claimants, is an increase in the creditable foreign tax of such amount (if any) as would reflect application of the relevant FNR to the relevant foreign income, that is to say the income as it was actually computed for Case V purposes. This approach, they say, builds on, and appropriately modifies, the actual UK system for taxing foreign dividends, and involves a conforming construction of the domestic UK legislation of the kind envisaged by the Court of Appeal in FII (CA) (see [25] above).

The Revenue disagree, arguing that since EU law requires the grant of a credit at the FNR, the relevant income must first be grossed up at the FNR. They identify the relevant income for this purpose as the amount of the net dividend received in the UK plus WHT. At first sight this approach may appear logical, but in my view it is mistaken. The grant of a credit at the FNR is not like the grant of a credit for WHT or underlying tax. Its purpose is to mirror the exemption granted by the UK to domestic dividends, and to recognise the fact that in the UK corporation tax was routinely paid at an effective rate lower than the nominal rate. The credit should therefore be applied to the gross amount of the dividend for UK tax purposes. The net dividend received in the UK had to be grossed up for WHT and underlying tax, because those taxes had actually been paid out of (or in respect of) the distributed profits. The credit for tax at the FNR served a different purpose, however, and did not reflect any further actual charge to tax on the distributed profits. I can therefore see no warrant for grossing up the Case V income at the FNR before applying the FNR credit.

I am reinforced in this conclusion by an example given by the claimants in their submissions in reply, pointing out that where the relevant FNR was lower than the UK corporation tax rate, the effect of grossing up the Case V income at the FNR would be to inflate the amount of the lawful Case V charge. The reason for this is that the lower FNR credit would not fully cancel out the notional Case V charge on the additional gross profit produced by the grossing up at the nominal rate. I should record that the Revenue did not have an opportunity to respond to this example, but the claimants’ point seems sound to me.

At what stage should credit be given for withholding tax?

The second question concerns the stage in the computation at which credit should be given for WHT, and specifically whether the lawful Case V charge should include, or exclude, any credit given for WHT. Two matters are not in dispute: first, credit must be given for WHT at some stage; and secondly, the giving of a credit for WHT (which relieves juridical double taxation) is not in itself a requirement of EU law. As I have explained, the way in which credit for WHT was in fact given was by setting it against the Case V charge: see [105] above.

The Revenue argue that the amount of Case V tax which the UK could lawfully have charged must be ascertained without giving any credit for WHT, precisely because the grant of such a credit was not a requirement of EU law. They then compare this amount with the amount actually charged, which of course included the grant of a credit for WHT at the final stage of the computation. Only if the latter amount exceeds the former, they say, can the claimants maintain a claim for unlawfully levied Case V tax.

The claimants submit that this is the wrong comparison. The true comparison, they argue, is between (i) the amount actually charged under the UK provisions, and (ii) the amount which should have been charged under the UK provisions adjusted so as to comply with EU law. On this basis, the amount of tax which should have been charged would have allowed a credit at the FNR, in order to secure compliance with EU law, but would still have granted a credit for WHT at the final stage. Accordingly, the claimants say they have a good claim in any case where the adjusted charge to tax, after the grant of a credit for WHT, would have been less than the actual charge.

In my judgment there is an illogical asymmetry in the Revenue’s methodology, because WHT is ignored in computing the lawful tax charge, but included in the actual tax charge with which the notional lawful charge is compared. In principle, the credit for WHT should either be included, or excluded, on both sides of the comparison. Otherwise the effect, as the claimants submit, is to net off the WHT credit against the nominal rate credit required under EU law. In my view the correct approach is to include the WHT credit on both sides of the comparison, because that better accords with the system which the UK actually operated.

I would add that the combined impact of the two questions which I have discussed under Issue 4 can make a very significant difference to the quantum of the claims. One of the examples helpfully appended to the Revenue’s written closing submissions is of a dividend in excess of £3 million (including WHT) which was paid in 1977 by BATIG to one of its five BAT parent companies, Weston Investment Co Ltd (“Weston”). At that date, corporation tax was charged in the UK at 52%. The relevant FNR was 35.42%, and the composite rate of underlying tax was 23.47%. The original Case V computation, on grossed up income of £4,167,225, yielded a net corporation tax liability (after allowing credit for both underlying tax and WHT) of £710,533.

The Revenue’s methodology ignores underlying tax entirely, but grosses up the dividend (inclusive of WHT) at the FNR of 35.42% to produce Case V income in excess of £4.9 million. This is then charged to tax at 52%, and a tax credit at the FNR on the gross income is then deducted, on the footing that this was the only credit required by EU law, leaving a maximum lawful Case V charge of £818,866. Since this was greater than the actual charge of £710,533, it follows that Weston would have no claim for unlawfully levied tax.

By contrast, the claimants’ computation starts from the original Case V income of £4,167,225; charges it to tax at 52%; then deducts underlying tax and WHT, as in the original computation; and then deducts a further credit of £497,882, representing a “top up” to a credit at the FNR on the Case V income. The result of this exercise is a lawful Case V liability of £212,651. Since the amount actually charged was £710,533, it follows that Weston has a claim to recover unlawfully levied tax of nearly £500,000, together with interest from 1977.

Introduction

The next group of issues concerns the methodology for calculating unlawful ACT, and for linking the ACT actually paid by UK companies in the BAT group with EU source income. The starting point for consideration of these issues must be the rulings of the ECJ in FII (ECJ) I and FII (ECJ) II. To a considerable extent, the answers to them follow on from the answers which I have already given to the questions concerning the unlawfulness of the Case V charge. That is only to be expected, since the ECJ has consistently regarded ACT as a payment in advance of MCT. Consideration of ACT does, however, involve further complexities, not least because ACT was only charged if and when a UK company made distributions outside a group income election. It was the normal practice within the BAT group for such elections to be in place, with the consequence that ACT only became payable when the top holding company in the group distributed dividends to its external shareholders.

This feature of the UK system as it operated in practice gives rise to obvious, and difficult, questions about how the actual charge to ACT, when it was made, should be related to EU source dividends originally received by UK water’s edge companies, often many years previously. As one would expect, the BAT group never had any system for ring-fencing its income from EU sources, and once dividends had been received in the UK they were generally mingled with other income. Equally, when a UK company paid a dividend to its parent, it would normally not specify the particular profits from which the dividend was paid, although it may be possible to deduce from the company’s accounts whether it was paid from the current year’s profits, or profits of an earlier year, or shareholder reserves, or a combination of those sources.

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).

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.

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.

With this brief introduction to the rival methodologies, I will now consider the issues which arise under this heading in turn.

Issue 5: does EU law require a credit to be given within the ACT computation for underlying tax as well as for tax at the FNR?

I will deal with this issue briefly, because I have already answered the corresponding question in relation to the Case V charge in favour of the claimants (see Issue 1 above), and I have also considered the same question in relation to ACT in Portfolio Dividends (No. 2) at [124] to [128]. I expressed my conclusion in Portfolio Dividends (No. 2) at [128] as follows:

“In the light of paragraph 72 of FII (ECJ) II, it appears to me that section 231(1) [of ICTA 1988] would have been compliant with EU law if it had also provided a credit in respect of dividends received from non-UK resident companies for (a) underlying tax to which the distributed profits had been subject, and (b) the nominal rate of corporation tax in the source State, but subject to an upper limit equal to the UK nominal rate of ACT.”

Subject to the qualification that, in the present context, it would be more accurate to refer to “EU resident” rather than “non-UK resident” companies, I adhere to that conclusion, which appears to me to be firmly based on the reasoning of the ECJ in FII (ECJ) I and FII (ECJ) II. I was taken again, by both sides, through the relevant passages in the judgments of the Court, and the opinions of the Advocates General in both cases. I remain unpersuaded, for the reasons already given in relation to Issue 1, that the Court in FII (ECJ) II intended to depart in any way from its oft-repeated insistence that a dual system of exemption for domestic dividends and imputation for foreign dividends, if it was to be compliant with EU law, required at least the grant of a credit for actual underlying tax. If that conclusion is correct in relation to the Case V charge, it must also be correct, in my judgment, in relation to the ACT charge, given the ECJ’s assimilation of the two taxes.

It is true that the language of paragraph 62 of FII (ECJ) II, read in isolation, refers only to the grant of a nominal rate credit; and see too the similar language in paragraphs 72 and 82, in the context of the Court’s discussion of the “corporate tree” issues raised by question 2 in the order for reference. The Revenue naturally make much of this point, but I do not find it convincing. First, it is clear from paragraph 72 that it was intended to reflect paragraph 62, where the language does not suggest that a credit at the FNR was seen by the Court as always sufficient in itself to ensure compliance with EU law (this is the point about “in particular” and “notamment” discussed in [35] above). Secondly, paragraphs 71 and 80, which both form part of the Court’s reasoning on question 2, refer expressly to the need for a Member State’s legislation to take account of “the corporation tax already paid on the distributed profits”, i.e. underlying tax. Thirdly, the concentration on nominal rates in FII (ECJ) II is in my view readily explained by the fact that this was a refinement of the basic conditions for compliance which the Court had already laid down in FII (ECJ) I and other authorities.

Issue 6: does EU law require credit also to be given against ACT for withholding tax?

This is again very similar to the question which I have already considered in relation to the Case V charge under Issue 4 above. My answer to it is essentially the same. Although the grant of a credit for WHT is not in itself a requirement of EU law, I have explained why WHT must in my judgment be taken into account in the calculation of the lawful Case V charge, and in particular why it is not legitimate to take it into account when looking at the actual charge to tax which was imposed but then to ignore it when constructing a notionally lawful charge. This creates a mismatch, which has the effect of inflating the amount of the lawful charge.

Another way of making the same point is to say that a compliant charge to tax cannot be constructed in a legislative vacuum, but must fit in with the actual structure of the UK charging provisions, including the credit given for WHT (whether under double taxation agreements, or unilaterally under section 790 of ICTA 1988). By granting such credit, the UK has restricted its taxing rights in relation to the foreign income of a UK water’s edge company so as to recognise the WHT imposed by the source State. This restriction must therefore be taken into account when comparing the actual charge to tax imposed by the UK with a notional charge which would have complied with EU law. Since, in the eyes of EU law, ACT is merely a prepayment of MCT, the same principles must then apply when calculating the lawful ACT charge.

It is no answer to this point to say, as Mr Ewart did more than once in his oral submissions, that the WHT is never received as part of the dividend in the UK, so it can never form part of the income to which the ACT charge is eventually applied. That, of course, is true, but the submission in my judgment confuses the measure of the income to which the ACT charge is applied (which necessarily excludes the WHT) with the computation which determines the rate at which the tax may lawfully be levied (when the WHT must be taken into account, on both sides of the comparison which needs to be made).

In fairness to Mr Ewart, I should also record that his submissions did pinpoint a significant flaw in the claimants’ calculations which was accepted by Mr Aaronson when he came to make his submissions in reply. In their original calculations, the claimants had included WHT in the amount of the foreign-source dividends to which the net ACT rate was applied. This was clearly wrong, because the WHT was deducted at source. It was never received by the UK water’s edge company, and therefore could never have attracted a charge to ACT.

Issue 7: how is the lawful ACT to be calculated?

The claimants’ primary case on how the lawful ACT is to be calculated, as refined in the course of argument, may be summarised as follows. A notional tax credit should be granted in respect of EU dividends paid to a UK water’s edge company, capped at the UK rate of ACT then in force. The credit should be for the higher of:

The practical effect of this formula is that the section 231 credit attaching to the EU dividend will be the same as the credit at the ACT rate which would have attached to a UK source dividend of the same net amount, unless the ACT rate then in force exceeded the higher of the FNR and the rate of credit for WHT and foreign tax actually paid, in which case the UK would have been entitled to levy ACT, on the onward distribution of the net dividend by the UK recipient, at the difference between the two rates. In the majority of cases, these conditions will not be satisfied and the notional section 231 credit will be capped at the ACT rate, with the result that the net EU dividend plus the section 231 credit can be regarded as equivalent to domestic FII. It will be recalled that section 238(1) of ICTA 1988 defines FII as income of a company resident in the UK which consists of a distribution in respect of which the company is entitled to a tax credit under section 231 “and which accordingly represents income equal to the aggregate of the amount or value of the distribution and the amount of that credit”.

The Revenue’s approach to computation of the lawful ACT is very different. Their method does not involve the grant of a section 231 credit, nor does it give credit for foreign tax actually paid (whether WHT or underlying tax). It does, however, share two important features with the claimants’ method. First, it looks at the position when the EU dividend was first received in the UK; and secondly, it grosses up the net dividend for both WHT and underlying tax to produce an amount of gross profits. The Revenue then apply to that amount of gross profits a rate equal to the difference between the UK rate of MCT in force at the time the dividend was declared and the relevant FNR. The product of that calculation is then expressed as a percentage of the net dividend received in the UK, to give the lawful rate of ACT applicable to it if and when it comes to be distributed outside a group income election.

So, for example, if the UK rate of mainstream corporation tax was 50%, and the relevant FNR was 30%, the Revenue say that the UK was entitled under EU law to levy ACT on the gross profits represented by the dividend at the maximum rate of 20%. Assuming the net amount of the dividend to have been 100, representing gross profits of 150, the maximum amount of lawful ACT would then be 20% of 150, or 30. Expressed as a percentage of the net dividend of 100, the lawful rate of ACT applicable to the dividend would be 30%.

The Revenue support their methodology on the grounds that, as a matter of EU law, ACT is merely a prepayment of MCT, so there is no reason why the lawful rate should be capped at the actual ACT rate which happened to be in force in the UK at the relevant time. In principle, ACT could have been lawfully levied at any rate up to the full MCT rate. The Revenue also allow no credit for tax actually paid, even though they include it in their computation of the gross profits, because their case is that the only credit required by EU law is a credit at the FNR. As before, they rely in particular on FII (ECJ) II at paragraphs 62 and 64.

The Revenue’s methodology is in my judgment open to a number of serious objections. First, it fails to allow a credit for underlying tax, which I have held to be a requirement of EU law. Secondly, it posits a lawful charge to ACT at a rate in excess of the rate actually in force in the UK at the relevant time. I cannot see any sound basis for making such a counter-factual assumption, which departs from the system which was actually in operation in favour of a purely hypothetical reconstruction of a system which the UK could in theory have chosen to adopt. A speculative exercise of this nature seems to me to go far beyond a legitimate conforming construction of the legislation governing the UK’s imputation system. Thirdly, the inclusion of actual foreign tax in the computation of gross profits, combined with the failure to give any credit for such tax, produces a mismatch similar to the one I have already considered in the context of the Case V charge, and again has the effect of hugely inflating the amount of lawful ACT.

This last point is well brought out by the first of the examples in appendix 3 to the Revenue’s written closing submissions. It concerns a dividend paid in 1977 by BATIG (then known as Interversa GmbH) to Durham Investment Co Ltd (“Durham”). The dividend was paid out of 1975 profits made in Germany. The amount of the dividend was £746,769, from which WHT of £112,015 was deducted, so Durham received a net dividend of £634,754. The underlying gross profits were £1,002,913, on which underlying tax of £256,144 had been paid. The relevant tax rates were:

On these figures, the Revenue calculate the amount of lawful ACT as £166,307. They reach this figure by applying the difference between the UK corporation tax rate and the FNR (namely 16.58%) to the gross profit of £1,002,913. Since £166,307 is less than the amount of the maximum lawful Case V charge (as calculated by the Revenue), there is no need to restrict it. Expressed as a percentage of the net dividend of £634,754, the calculation therefore yields a lawful ACT rate of 26.2%.

The claimants submit, and I agree, that this would be a most surprising result, in circumstances where the actual foreign tax paid (WHT of £112,015 and underlying tax of £256,144, totalling £368,159) exceeded the maximum amount of ACT which the UK could in fact have charged on the distribution in the UK of a net dividend of £634,754 (£334,336, i.e. 34.5/65.5 x £634,754). Furthermore, the FNR of 35.42% was higher than the ACT rate of 34.5%, which would naturally lead one to suppose that, on even the narrowest interpretation of the ECJ’s ruling, any charge to ACT on the onward distribution of the dividend would have been cancelled out by the credit at the FNR required by EU law.

The Revenue’s failure to allow a credit for underlying tax, while taking it into account as part of the gross profits, leads to yet a further anomaly. If that were the correct methodology, it would mean that the more underlying tax was actually paid, the greater the amount of ACT which the UK could lawfully have charged on the dividend. The reason for this is that the higher amount of underlying tax would correspondingly increase the amount of gross profits to which the difference between the UK corporation tax rate and the FNR would be applied, leading in turn to a higher lawful ACT percentage. In other words, the effect of the mismatch in the Revenue’s calculations is to increase the amount of lawful ACT proportionately with any increase in the amount of the actual underlying tax.

For all these reasons, I consider that the claimants’ method of calculating lawful ACT is to be preferred to the Revenue’s. This accords with the conclusion which I reached in Portfolio Dividends (No. 2), but I emphasise that on this part of the case I have considered the matter afresh with the benefit of the much fuller submissions advanced to me by each side.

Issue 8: how is the FNR in respect of which credit must be given to be determined?

The parties are in agreement that this question should be answered in the same way for ACT as for the Case V charge. Reference should therefore be made to the answers which I have already given to the questions raised under Issues 2 and 3 above.

Issue 9: how should ACT paid by UK companies be linked with EU-source income to give effect to the judgments in FII (ECJ) I and FII (ECJ) II?

This issue asks what methodology (or methodologies) should be used for linking ACT paid by UK companies with foreign-source income to give effect to the judgments in FII (ECJ) I and FII (ECJ) II. It is an issue of central importance in the case. I have already given a brief introduction to the rival methodologies advanced by the parties at the beginning of this section of my judgment: see [116] to [120] above.

At the outset, I would make two preliminary comments. First, as Mr Ewart rightly reminded me, it would be wrong to approach the question as if I merely had to choose between the rival methods. I must not lose sight of the fact that the onus lies squarely on the claimants to satisfy the court that their method is the correct one, if they are to be entitled to recover restitutionary compensation for their losses on the strength of it. Thus rejection of the Revenue’s methodology would not, of itself, lead to the conclusion that the claimants’ methodology is correct. Secondly, Mr Ewart accepted that it does not matter for the purposes of the Revenue’s argument on this issue whether the claimants are entitled under EU law to a credit for underlying tax as well as a nominal rate credit at the FNR. That question affects the amount of the credit required by EU law, but not the methodology for linking the credit with the ACT system in force in the UK.

In the context of other Issues I have already given a number of reasons which may be thought to favour the claimants’ general approach, that is to say (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. In summary, those reasons (which overlap) are:

In my judgment the cumulative force of these points is powerful, and it is arguable that the question is anyway settled in principle at first instance, as well as in the Court of Appeal, by the guidance already given in FII (CA). I have, however, considered the question on its merits, rather than merely as one of precedent, because matters have come a long way since the Court of Appeal gave its judgment in February 2010, the question has been fully argued before me, and the Court of Appeal did not have the benefit of the judgment of the ECJ on the second reference.

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.

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.

It is important to note in this context that in FII (ECJ) II the ECJ decided the “corporate tree” questions in the claimants’ favour, and rejected the United Kingdom’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. Conversely, when ACT was paid higher up the corporate tree, it was argued that the payment resulted from the decision of the relevant companies to make an election. The ECJ was unimpressed by these arguments, holding that this structure did not nullify the discrimination identified in FII (ECJ) I and that a restitutionary claim could be maintained by the parent in respect of the ACT which it eventually paid. In this part of its judgment the ECJ was of necessity focusing on the ACT system as it operated in practice in the UK, and 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: see in particular paragraphs 75 to 80 of the judgment.

An important aspect of the FII system is that it applied only to dividends (or, more accurately, qualifying distributions), and was unaffected in its operation by other sources of income of the companies concerned. Thus the whole amount of FII generated by the payment of a UK dividend would frank the payment of a dividend of equivalent size by the recipient company, and so on up the group, regardless of whether the subsequent dividends were paid out of the same profits as the original one, and regardless of any other sources of income which the intermediate companies may have had. This cardinal feature of the system is respected by the claimants’ methodology, but contradicted by the Revenue’s, which uses various techniques to identify the original EU dividend in its passage up the group to the stage when ACT is paid. In general, where an intermediate company has sources of income other than the EU dividend, the Revenue treat any dividend of equivalent size paid by that company as drawn on a pro rata basis from all of its available sources of income, with the consequence that the EU income component of the dividend is diluted, and the balance of the EU income is treated as retained.

Because the Revenue’s approach only grants credit at the FNR when ACT is finally payable, the effect of this identification technique is greatly to reduce the amount of EU income for which credit has to be granted when compared with the amount of EU income which was originally received by the water’s edge company. It is a technique for which no warrant can be found in the FII system, or in any other UK legislation relating to dividends, and it could only begin to make sense, in my judgment, if the Revenue were right in their submission that EU law was not relevantly engaged until ACT was actually paid. For the reasons advanced by the claimants, however, I am satisfied that this approach looks at the question the wrong way round, beginning at the top rather than the bottom, and that what needs to be found is a way of feeding EU dividends into the FII system at the water’s edge.

For completeness, I should note that the claimants also relied, in support of their general approach, on the numerous references in the ECJ’s judgments to the UK “system” of dividend taxation, and on the decision of the ECJ in the case of Salinen, Case C-437/08. The references to the UK system in the judgments do not in my view take the matter much further forward, because the ECJ inevitably had to describe and analyse the UK system in some detail in order to diagnose what was wrong with it. It does not necessarily follow that the appropriate remedy is to be found by modification of the system, although I think it is natural to begin by asking how the system could have been modified in order to render it compliant, and if such a method can be found by a process of conforming interpretation, that is then the obvious solution to adopt. Salinen provides a helpful example of the ECJ itself applying this kind of remedial approach at a systemic level.

I now need to describe the claimants’ methodology in a little more detail. Although not the method they first propounded, by the date of the trial the methodology which the claimants placed first in their submissions was one based on the machinery of the ACT system. They call it the “CT61” method, that being the specified form on which companies made their quarterly ACT returns. The following description of the main features of the CT61 method is given in a written “walk through” of the method, prepared by the claimants’ lawyers and dated 23 April 2014:

“The CT 61 Method

13.1

The method employed is to adapt the machinery used by the ACT system to perform the same task for domestic sourced income.

13.2

Under that system a company in receipt of income which carried a s231 credit would record the receipt as FII (the dividend plus the s231 credit) on its quarterly ACT return, known as a CT61. Where the company paid a dividend outside a group income election, it would likewise record the franked payment (the dividend plus the ACT liability) and deduct from it any FII [it] had received. Any excess FII would be carried forward. The corporate recipient of that franked payment would in turn complete the same return. That process was simplified of course by the option of paying dividends through a corporate group under a group income election so that only the ultimate parent need, in that case, prepare the CT61 return.

13.3

Importantly, although CT61s were quarterly returns, they could be aggregated over the annual accounting period concerned (paragraph 4 Schedule 13 ICTA). Thus a receipt of FII in a subsequent quarter could be carried back and credited against a franked payment in an earlier quarter provided both were in the same corporation tax accounting period. Where this occurred the company’s CT61 for the quarter in which it received FII would record the FII deduction but no franked payment and it would receive a refund of the relevant amount of ACT paid in the earlier quarter.

13.4

The Claimants’ “CT61 method” envisages adapting that process to trace the passage of “EU s231 credits” through the group. The Claimants envisage a virtual CT61 with an entry for an EU s231 credit just like FII and which is passed on with each dividend payment at each corporate level until it is received in an accounting period (see paragraph 13.3) in which the recipient pays ACT. It is then set against that ACT payment in the same way as a domestic s231 credit. The excess ACT so relieved was therefore unlawfully levied.

13.5

Two adjustments must be made to the domestic framework to accommodate this method:

(a)

where the foreign tax rates are below the ACT rate, the EU income will not carry a full s231 credit. If it were treated as FII (namely the dividend plus the reduced s231 credit) excess credit would be received [the explanation for this is given in a footnote]. Three approaches could accommodate this complication:

(i)

regard the EU income as FII but reduce the value of the EU dividend for which credit is given in the same proportion as the EU s231 credit is to a full s231 credit; or

(ii)

to reconstruct the dividend flows that would have occurred upon a CT61 by attributing a particular source of EU dividend to each onward dividend payment so that only the level of s231 credit attributed to that source of income is brought into account as a credit against ACT paid higher up in the group; or

(iii)

allocate to each dividend so much of the pool of EU s231 credits held by the dividend paying company as it may hold capped at the ACT rate upon that dividend and trace through the corporate hierarchy by this route only the EU s231 credits (i.e. as opposed to the equivalent of FII – the dividend plus the s231 credit).

These approaches should produce the same outcome. For historic reasons only [again explained in a footnote] the Claimants’ methodology [in an earlier version of the walk through] adopted the more cumbersome second approach. For this revision of that methodology the simpler third method has been adopted.

(b)

To identify the accounting period in which EU dividend income was received is straightforward. It is recorded in the accounts of the relevant companies. To identify the precise quarter of receipt is much more cumbersome and would require investigation of bank records. This has not been done and is not necessary for the reason explained above: FII received later than franked payments in the same accounting period could be carried back to relieve the earlier payment. The date on which EU dividends were received is therefore irrelevant. The quarterly CT61s for an accounting period in effect function as part of a single composite return.

The Claimants’ method therefore considers only the CT61s in a composite annual form. The only complication in doing so arises where actual FII has been received in a subsequent quarter to a franked payment so as to produce a repayment (the few instances where this arises are shown on Table 2). In that circumstance credit for the actual FII is considered to have been taken in priority to the application of the EU s231 credits (i.e. the most conservative approach is applied).”

The “walk through” document went on to explain, by reference to a number of tables, how the EU section 231 credits were passed up the group, in parallel with qualifying distributions on which ACT was actually paid (as recorded in the relevant CT61s). The EU credits were then set against the ACT payments, and the element so relieved was identified as “unlawful ACT”.

In their written and oral submissions, the Revenue successfully challenged one significant aspect of this methodology. They pointed out that what it purported to do was to pass tax credits up the group, whereas it should have passed up a modified version of FII, franking subsequent distributions of equivalent amounts but not (in such a case) generating any further entitlement to a tax credit. In these respects, the claimants’ method failed to replicate the domestic FII system and generated too much credit. The claimants rightly accepted the force of these criticisms, and by the time of their submissions in reply they had recast their computations so as to convert the EU section 231 credits into equivalent amounts of FII. This adjustment resulted in a reduction of approximately 3% in the claim. On a further point of detail, the claimants then dealt with the problem identified in paragraph 13.5(a) of the walk through (where the foreign tax was less than the ACT rate) by reducing the amount of the EU dividend for which credit is given in the same proportion as the EU credit would bear to a full section 231 credit.

The claimants’ revised calculations also took into account another, less significant, flaw in their original methodology which the Revenue had identified. The purpose of the revision was to take proper account of any changes in the ACT rate in the relatively rare cases where a dividend was paid up the group in the form of franked payments instead of under a group income election.

Subject to:

The alternative methodology now relied on by the claimants, and the only one which they originally advanced, is based on the FID legislation. It is unnecessary at this stage to describe the legislation in any detail. An adequate introduction to it for present purposes may be found in FII (High Court) I at [23] to [25] and [165] to [170].

The FID method builds on the fact that, in the FID legislation which came into force on 1 July 1994, Parliament provided statutory rules which enabled dividends paid in cash by UK companies, on which ACT was payable, to be matched with foreign dividends received by the company or any of its subsidiaries in the same or the preceding accounting period. What the method does, in outline, is to match dividends paid by higher-tier companies in the BAT group with available EU-source income received by their UK water’s edge subsidiaries in the same accounting period, and then link the ACT payable on the dividends with the underlying tax credits on the matched income.

The method is described more fully in a “walk through” document similar to that provided by the claimants for the CT61 method. I will not delve into it more deeply, however, because I consider the method to suffer from several obvious defects in comparison with the CT61 method.

First, the FID regime was only in force from 1994 until the abolition of ACT in 1999. It had no precursor in the earlier years back to 1973 which account for most of the present claims, whereas the CT61 system was in force throughout the whole of the period covered by the claims.

Secondly, the FID regime was expressly designed to relieve the problem of unutilised ACT, in a way which was (wrongly) thought to be fully compliant with EU law. It was not a system designed to remedy the breach of EU law which the ECJ has held to exist in the present case.

Thirdly, the matching rules under the FID regime were entirely arbitrary, in the sense that there did not need to be any connection between the dividend on which ACT was paid and the FID or FIDs with which it was matched. Providing the necessary conditions were satisfied, the group could perform the matching in whatever way suited it best. By contrast, the CT61 method looks at each EU dividend at the point of receipt at the UK water’s edge and then treats it as modified FII, the use of which is then governed as far as possible by the existing FII regime.

Finally, it seems particularly arbitrary to borrow the FID matching rules in isolation, when under the FID regime itself the matching process did not give rise to any tax credit, and the company was only entitled to a refund of ACT to the extent that it was not set off against MCT for which the company was liable. The matching therefore gave rise to a limited form of statutory relief which bears no relation to the unlawful levy of ACT which now needs to be remedied.

It is true that the FID method has the merit and attraction of simplicity, but it seems to me conceptually flawed, and I have little hesitation in holding that the CT61 method should be preferred to it.

The Revenue’s method is described in paragraphs 80 to 101 of their skeleton argument, and in much greater detail in their own “walk through” of selected dividends dated 6 May 2014. They emphasise that the UK was entitled to charge ACT at whatever rate it chose on distributions made out of either EU or non-EU sourced profits; and, as I have already explained, they concentrate on the stage when ACT in fact became payable, usually at the top of the group. It is therefore necessary for the Revenue to identify the EU-source component of the distributions on which ACT was charged, and for that purpose they have devised a so-called “tracing” approach based on information in the UK companies’ accounts.

In the Revenue’s skeleton argument, the tracing approach is described as follows:

“95.

The [Revenue’s] basic approach is to look at the current year profits or reserves from which the distribution has been made and estimate:

(1)

EU income included within those profits (including the source country of the income and the year in which the income arose)

(2)

Non-EU income (including UK income) included within those profits.

96.

The [Revenue’s] methodology begins with the UK “water’s edge” company to determine how much of the distributions declared for the period is made up of the EU income it received.

97.

The methodology assumes that the company was capable of making distributions out of both non-EU and EU income during the relevant period. Therefore, in determining how much of the company’s distributions are made up of EU and non-EU income, in general no preference has been given to either source of income. Where, as was usually the case, the dividend is less than the distributable profits or reserves (rather than equal to the whole of the distributable profits or reserves) the dividend is treated as comprising the various sources of income within the distributable profit or reserves on a pro-rata basis.

98.

The methodology then looks at the shareholder of the UK “water’s edge” company to determine how much of the EU income forming part of the dividend declared is recognised as income in the respective shareholder’s accounts.

99.

The process is repeated at each company level to establish how much of the dividend the intermediary declares is made up of EU income received from its subsidiaries.

100.

The final part of the methodology looks at the ACT paying company (typically the group’s ultimate parent) to determine how much of the distribution, upon which ACT was paid, is made up of EU income that flowed up from the UK water’s edge company.

101.

To facilitate the tracing exercise based on the general rules outlined above, a spreadsheet was set up for each UK BAT Group company noted in the abridged group structure included with the [Revenue’s] revised computations dated 12 March 2014. These are the companies through which EU income flowed. The spreadsheets for each of the relevant companies are included with the [Revenue’s] revised computations dated 12 March 2014.”

Despite the apparent simplicity of this general description, the “tracing” approach is in fact of enormous complexity, and depends on the making of numerous (and often unverifiable) assumptions. The basic reason for this is that the approach bears no relation at all to the actual system of dividend taxation in force in the UK, which passed FII up the group and ensured that ACT was charged only once on UK-source dividends. The Revenue are therefore obliged to devise a methodology which has no precedent, in order to answer a question which was in nobody’s mind at the time. The main raw material available to them for this purpose is information contained in the relevant UK companies’ accounts. The purpose of the statutory accounts, however, was to provide a true and fair view of the activities and financial position of the company concerned; and although there were also tax computations, these were prepared in the context of the system as it was then understood to operate. It would not have crossed anybody’s mind that the accounts and computations might one day have to be used as a quarry to provide raw material for the exercise which the Revenue have now undertaken.

It also needs to be appreciated that the exercise does not involve tracing in any recognisable legal or equitable sense of that term. Rather, it involves a series of ad hoc assumptions which are applied, in the absence of evidence to the contrary, in order to identify an EU-source component of a dividend subject to ACT. One of the most important of these assumptions is the default assumption that dividends were paid on a pro rata basis out of different sources of income comprised in the distributing company’s profits. But there is, in truth, no empirical basis for making such an assumption. It is just a convenient rule of thumb. The difficulty with it, in the present context, is that it runs counter to the self-contained system of dividend taxation operated by the UK, and has the effect of greatly diluting the stream of EU income subject to potentially unlawful taxation.

I recognise that a huge amount of hard work and ingenuity has gone into the development of the Revenue’s methodology. In what I have said so far, I have barely scratched its surface. I do not, however, intend to examine it in any more detail. Whereas the claimants’ CT61 method seems to me in principle the appropriate way to identify unlawfully levied ACT in the light of the ECJ’s judgments, I consider the Revenue’s method to be conceptually unsound precisely because it fails to replicate the key features of the CT61 model. In particular, it ignores the operation of the domestic FII system, and it seeks to trace EU income backwards from the point of charge to ACT when what is required is an appropriate assimilation of EU income into the FII system from the water’s edge upwards. These points of principle are in my judgment so fundamental that the failure of the Revenue’s method to incorporate them is alone enough to discredit their whole system.

The claimants submit, for good measure, that not only is the Revenue’s method conceptually wrong, but it is also open to practical objections of a structural nature which are so serious as to make it impossible to implement in a coherent and verifiable manner. These objections were elaborated by the claimants in their written closing submissions, and I was helpfully taken through some of them by junior counsel, Mr Bremner. There are three main areas where, according to the claimants, the Revenue’s tracing methodology clearly breaks down:

The position is made still worse, submit the claimants, by three related features of the Revenue’s approach. First, the calculations and descriptions relied upon were unsupported, despite their length and complexity, by the evidence of any witnesses (whether expert or factual) who could explain the methodology employed and be cross-examined on it. Secondly, the approach was incapable of being checked for accuracy and robustness by software tools of the kind used by Mr Woodford in his audit of the CT61 method, or in any other way short of a complete reconstruction from the basic data. Thirdly, the detailed spreadsheets produced by the Revenue were found by the claimants to contain well over 20,000 “hard coded” entries, that is to say bespoke calculations which had been inputted by hand and the accuracy of which had to be taken on trust.

In view of the conclusions I have already reached, it is unnecessary for me to explore these objections in any detail. It is enough to say that, in general, I consider them to be well founded.

Issue 10: FIDs

Three questions are raised under this heading, each of them relating to FIDs. The questions are these:

This section of my judgment needs to be read in conjunction with:

In the light of all this material, it is in my judgment now clear beyond argument that the FID regime not only breached Articles 49 and 63 in relation to EU FIDs (i.e. FIDs matched with distributions made by companies resident in the EU), as the ECJ expressly held in FII (ECJ) I, but that the regime also breached Article 63 (and could not benefit from the standstill provision) in relation to third country FIDs. It follows that the claimants are in principle entitled to recover the time value of all of the ACT which they were obliged to pay under the FID regime, from the dates of payment until the dates when the ACT was repaid to them.

This entitlement was made explicit in paragraph 16(b) of the Court of Appeal’s order dated 19 March 2010, which so far as material provided that:

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

…

(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 …”

The factual background relevant to the claimants’ FID time value claims is set out in Mr Cohn’s fifth witness statement. I emphasise that this evidence relates to the actual FIDs which were declared and paid by the ultimate parent companies of the BAT group in the six accounting periods ending 31 December 1994 to 31 December 1999. These claims are quite separate from, and do not in any way depend on, the so-called FID method for linking ACT with EU-source income which I have considered and rejected in the context of Issue 9.

As Mr Cohn explains, the companies which paid the FIDs (BAT Industries Plc in the first five of the accounting periods, and British-American Tobacco Plc in the sixth) duly paid ACT upon them. The FIDs were matched against qualifying distributable foreign profits received by UK-resident subsidiaries within the group. In order to qualify, it was necessary for the distributable foreign profits to have borne a rate of underlying tax which exceeded the UK corporation tax rate (a requirement which can be extracted, with some effort, from section 246I of ICTA 1988: fortunately it is not in dispute). When the matching had been agreed with the Revenue, the ACT was repayable. In fact, provisional repayments of the ACT were always made nine months after the end of the accounting period in which the FID had been paid.

There was a practical problem associated with the operation of the FID regime, described by another of BAT’s witnesses, Mr Thomas Bilton, in his evidence for the first liability trial in 2008. At the time when the election to treat a distribution as a FID had to be made, the underlying tax liabilities on the matching distributable foreign profits were most unlikely to have been finally settled. Thus the group had to make an educated guess what the underlying rate of tax would turn out to be. The consequences of under-estimating the rate could be severe, because if the matching foreign profits failed to satisfy the tax rate criterion the dividend would cease to qualify as a FID, the ACT would no longer be repayable, and by declaring the FID the group would in fact have exacerbated the problem of surplus ACT which the regime was intended to relieve.

Two consequences flowed from this. First, the group would aim to err on the safe side by matching FIDs with a greater amount of distributable foreign profits than it thought it would probably need. Secondly, the repayments of ACT which the Revenue made nine months after the end of the accounting period were (as I have said) provisional.

With one exception (which is the subject of the third question under this Issue), the claimants’ FID time-value claims are based on matching with foreign profits which was eventually agreed with the Revenue. It follows that the underlying rate of foreign tax paid on those profits equalled or exceeded the UK corporation tax rate, and the associated repayments of ACT became final. Mr Cohn has also calculated that it would make no difference if the only credit required by EU law were one at the FNR. The reason for this is that the BAT group always had a large enough pool of surplus distributable foreign profits, pursuant to its “safety first” policy, to absorb the effects of any difference between the FNR and the UK corporation tax rate.

I now come to the first and second questions under this Issue. The Revenue submit in their skeleton argument that FIDs should be treated in the same way as ordinary dividends. They say that the unlawfulness of the ACT charge was in principle the same for FIDs and non-FIDs, so the remedy should likewise be the same. Accordingly, the matching of FIDs with particular distributable foreign profits (which was in fact agreed with the Revenue) must be ignored, and the EU income component of FIDs must be determined in accordance with the Revenue’s standard “tracing” methodology.

These submissions were not developed by the Revenue in oral argument. In their written closing submissions, they sought to justify their stance thus (paragraph 67A):

“There is no reason in principle to apply any different treatment. The task is to identify the amount of tax that the UK could lawfully charge on a FID. That amount must be determined on the same basis as for a non-FID. The fact that, under domestic law, a special regime applied to enable the Claimants to choose to match their FIDs with certain foreign profits is wholly irrelevant in working out how much tax the UK could lawfully charge on those FIDs.”

In my judgment, it is not open to the Revenue to advance these arguments. The unlawfulness under EU law of the FID regime has already been conclusively established, in relation to EU FIDs by the judgment in FII (ECJ) I, and in relation to third-country FIDs by the judgment in FII (CA). The UK was therefore not entitled to levy any ACT on FIDs, and the claimants are entitled to recover the time value of the ACT which they paid, from the dates of payment until the dates of repayment. This entitlement is reflected in paragraph 16(b) of the Court of Appeal’s order.

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.

For these reasons, I am satisfied that FIDs must be treated separately from ordinary dividends. It follows that:

As to third-country FIDs, they must in my judgment be treated in the same way as EU FIDs. The distributable foreign profits with which they were matched must be brought into account accordingly, and must be excluded from the foreign income available to be matched with ordinary dividends. Because of the self-contained nature of the FID regime, and the separate rulings of the ECJ in relation to it, I consider that these conclusions would hold good even if the Revenue’s arguments in relation to the treatment of ordinary dividends were to be accepted.

The third question which I have to consider concerns only the first of the FIDs, which was declared and paid in 1994. BAT Industries Plc made an initial FID election, on the strength of which the Revenue provisionally repaid all of the FID ACT on 2 October 1995. This repayment was made on the footing that the profits against which the FID was matched qualified as distributable foreign profits.

Part of the 1994 FID was matched against dividends received by UK-resident claimants in the BAT group from the 1993 profits of BATIG. In 1997 BATIG received a tax refund attributable to 1993, the effect of which was to reduce the underlying rate of tax on BATIG’s 1993 profits to a level which meant they could no longer qualify as distributable foreign profits. Since BAT had meanwhile distributed FIDs in 1995 and 1996, it was able to reach an agreement with the Revenue whereby the shortfall in distributable foreign profits for 1994 was covered by an allocation of profits received in 1995. An agreement to this effect was reached in correspondence in July 1997. As a result of the re-matching, however, BAT Industries Plc had received part of the repayment of ACT attributable to the 1994 FID one year earlier than it should have done. Accordingly, BAT Industries Plc was required to pay to the Revenue interest on the early repayment in the sum of £357,621.02. The payment was made on 18 July 1997. BAT Industries Plc now wishes to claim that amount, together with interest on it, as well as the time value of the ACT for the period from payment until the original repayment date.

I am unable to accept this argument. The problem for BAT is that the 1994 FID did not in fact comply with the relevant rules to the extent that it was matched with BATIG’s 1993 German profits, although this did not become apparent until 1997. It follows that the ACT levied on that part of the FID never became repayable at all under the FID rules. BAT was only able to treat it as repayable, albeit one year early, because of the extra-statutory agreement that it was able to reach with the Revenue in 1997. It seems to me that the payment of interest flowed from, and was caused by, that agreement. It was, in effect, the price which the group had to pay in order to secure a favourable outcome which permitted the whole of the 1994 FID to be treated as valid. The interest payment is therefore not infected with the invalidity of the ACT charge, and BAT Industries Plc should in my view count itself fortunate that the 1997 agreement with the Revenue enables it to claim restitution of the time value of the ACT attributable to the matched BATIG profits.

My answer to the third question is therefore that the claimants are unable to make a claim in respect of the interest paid to the Revenue in July 1997. Furthermore, even if I had considered such a claim to be sustainable in principle, I am very doubtful whether it could have grounded a San Giorgio restitutionary claim under EU law. At best, the circumstances might have given rise to a consequential claim for damages.

Issue 11: utilisation of unlawful ACT

Once the amount of unlawful ACT has been ascertained, it is next necessary to consider how and when it was utilised, or rather how and when it should be treated as having been utilised for the purposes of the present claims. There is no disagreement about the dates and manner of utilisation of the ACT which was in fact paid within the group, usually by the ultimate holding company. That is a matter of historical record, and can be reconstructed from the documents. But how is the portion of each ACT payment that is now identified as unlawful to be treated? This is the basic question which I now have to consider.

It is common ground that the question has to be answered by the making of appropriate assumptions. The legislation tells us what could be done with ACT once it had been paid. In outline, ACT (if not repaid, for example under the FID regime) had to be set against the paying company’s liability to MCT under section 239(1) of ICTA 1988, unless it was surrendered to a subsidiary in accordance with section 240, in which case it was treated as ACT paid by the subsidiary and set against the subsidiary’s liability to corporation tax. Surplus ACT could be carried back for up to six years under section 239(3), or carried forward to the next accounting period under section 239(4), in which case it was treated as if it were ACT paid in respect of distributions made by the company in that period (and so on in future accounting periods, if the ACT remained surplus). This legislative scheme naturally took it for granted that the ACT had been lawfully charged and paid in the first place. It therefore offers no guidance on how to treat unlawfully paid ACT. Hence the need to make appropriate assumptions.

The Revenue’s approach to this problem has the merit of simplicity. In the absence of any statutory guidance, they submit that, subject to limited exceptions, the only rational solution is to treat all the payments or applications of ACT which actually took place as having been comprised of lawful and unlawful ACT on a pro rata basis. To this general rule, they admit of only three exceptions:

The first of these exceptions is not controversial, and it is enough to say that I agree with it.

The second exception is explained as follows in paragraph 147 of the Revenue’s walk through:

“For this part of the computation the same assumption has been adopted by both the Claimants and the [Revenue] in that total ACT paid is treated as being utilised on a FIFO basis (that is if ACT was surrendered to a subsidiary company by the ultimate parent company in year 1 and further ACT surrendered to that same subsidiary company in year 2 and the subsidiary company then utilised some of the total ACT surrendered to it by the ultimate parent company in year 3, then the [Revenue] and the Claimants have both assumed that the ACT surrendered to the subsidiary company in year 1 was utilised before the ACT surrendered in year 2.”

In oral argument, Mr Ewart said that the Revenue’s reason for adopting this approach was that the claimants had done a lot of work on it, and “it seemed a reasonable method” (day 14, pp 31-32).

The third exception was presented by the Revenue, both in their walk through (at paragraph 138) and by Mr Ewart in his oral submissions (day 14, pp 30-31), as a pragmatic expedient adopted for ease of calculation only. Mr Ewart confidently assured me that, assuming consistent rates of compounding of tax, it produced the same end result as the pro rata approach.

The Revenue’s basic justification for their pro rata approach is given in paragraph 136 of their walk through. They say that the total ACT paid “forms part of a fungible pot of ACT in the hands of the company retaining the ACT or the company receiving the ACT by way of an ACT surrender”. In those circumstances, “although it is possible to say how much of a particular ACT payment is lawful and unlawful, it is not possible to state, without the use of hindsight, which part of the ACT payment is lawful and unlawful and how it was ultimately used”.

The claimants’ preferred approach, by contrast, has three main stages:

The above order of priority is reflected in the claimants’ two walk throughs, at paragraphs 14.1 to 14.7 in each case. It may aid comprehension to set it out in a table.

It will be appreciated that the effect of this order of priority is, in general, to treat lawful ACT as utilised before unlawful ACT, and in particular to postpone for as long as possible the utilisation of unlawful ACT against MCT. Since there was usually far more lawful than unlawful ACT in any given accounting period, the overall effect is to postpone utilisation of the unlawful ACT for considerably longer than the Revenue’s pro rata method. This has the result of considerably increasing the amount of the claimants’ time value claims for utilised unlawful ACT. Mr Ewart was able to demonstrate this by taking me through a number of worked examples. A vivid indication of the scale of the difference between the two approaches is provided by some comparative calculations performed by the Revenue, which are described in paragraphs 92 to 95 of their written closing submissions. On the basis of the claimants’ original CT61 methodology, but using the Revenue’s approach to utilisation, the amount of the claims in respect of unlawful ACT comes down from approximately £1.073 billion to approximately £818 million, a difference of the order of £255 million. These figures do not take account of subsequent changes to the claimants’ CT61 methodology, but they serve to show how much is at stake on what at first sight might appear to be an arid dispute about the assumed utilisation of unlawful ACT.

Mr Aaronson accepted in oral argument that a pro rata approach could be adopted, but submitted that the claimants’ approach was preferable because it was more principled. He submitted that lawful ACT should where possible be matched with lawful Case V tax, because the ECJ regards ACT as a prepayment of corporation tax. Conversely, unlawful ACT should where possible be matched with unlawful Case V tax. This would also have the advantage of isolating the illegality, and reducing the risk of double recovery for both unlawful ACT and unlawful Case V tax in respect of the same EU income (it needs to be remembered at this point that ACT and MCT were payable at different times, and the actual charge to ACT often came long after the charge to MCT on the relevant Case V income).

Moving on to the later stages in the order of priority, Mr Aaronson submitted that the general objective should be to restore as far as possible what would have happened in the absence of any unlawful charges to tax. On that hypothesis, he said, lawful ACT would have been matched with any remaining lawful Case V tax, and then with other MCT. Only when the lawful ACT had been exhausted could it be appropriate to match unlawful ACT with MCT.

Mr Aaronson also criticised the Revenue’s methodology because of the exceptions which it admitted to the general pro rata approach. He submitted that there was no logical basis for the exceptions, and accused the Revenue of cherry picking. The pro rata approach should either be applied across the board, or it should not be applied at all. Indeed, the claimants’ alternative submission, which only emerged in the course of the hearing, is that a pro rata approach should be applied throughout. Mr Aaronson said that if this were done, and if the FIFO approach to timing were abandoned, the end result was likely to be similar to that of the claimants’ preferred approach.

In considering the rival submissions, I begin with the justifications advanced by the claimants in support of their preferred approach. I have to say that I find them unconvincing. It seems to me a hopeless enterprise to attempt to reconstruct what would have happened in the absence of unlawful charges to tax which were unknown to everybody at the time. It would also be a highly artificial exercise, involving cash flows which could never have been replicated in the real world, since payments and surrenders of ACT (both lawful and unlawful) usually long post-dated the Case V charge to tax on the relevant EU profits. In truth, the methodology amounts to little more than a series of assertions to the effect that lawful ACT should where possible be matched with lawful Case V tax, that unlawful ACT should likewise be matched with unlawful Case V tax, and that lawful ACT should then be treated as utilised before unlawful ACT. The first two of these assumptions may seem superficially plausible, but they have no empirical foundation. The fact that the ECJ regards ACT as a pre-payment of MCT seems to me irrelevant in this context. As to the third assumption, the claimants advance no independent justification for it, although it is extremely favourable to them.

For these short reasons, I remain wholly unpersuaded that the claimants’ preferred approach has any advantages of principle over a pro rata approach. The latter approach, on the other hand, reflects the indisputable fact that lawful and unlawful ACT were both mingled in a single pot in every accounting period, with no way of distinguishing one component from the other. In those circumstances, I think the only rational solution to the problem is to regard all the payments, surrenders and applications of ACT which actually took place as having been composed of both lawful and unlawful ACT on a pro rata basis. I would further hold, in agreement with the claimants’ fallback submission, that the pro rata approach should be applied across the board, without any exceptions apart from the payments and repayments of ACT under the self-contained FID regime.

The end result, therefore, is that I accept the claimants’ alternative approach. The extent to which it actually produces a different result from their preferred approach remains to be seen when the calculations have been reworked. If the claimants’ preliminary view that there is no major difference turns out to be correct, this will presumably be because elimination of the FIFO approach to the timing of payments, which was included in the Revenue’s comparative calculations, more or less offsets the advantages to the claimants of postponing the utilisation of unlawful ACT.

Issue 12: carry back of excess FII in a single accounting period

This issue concerns a small point which was barely touched upon in the numerous written submissions presented to me, and (unless I am mistaken) was not mentioned at all in oral argument. I will therefore deal with it briefly.

The question arises where, within a single accounting period, actual FII has been carried back from a later quarterly ACT return period to offset ACT paid in an earlier quarter, with the result that ACT is repaid.

The relevant statutory provisions are contained in paragraph 4 of schedule 13 to ICTA 1988, which provides as follows:

“4.

(1) This paragraph shall have effect where –

(a)

a return has been made of franked payments made in any return period falling within an accounting period and advance corporation tax has been paid in respect of those payments; and

(b)

the company receives franked investment income after the end of the return period but before the end of the accounting period.

(2)

The company shall make a return under paragraph (1) above for the return period in which the franked investment income is received whether or not it has made any franked payments, or paid any foreign income dividends, in that period, and, subject to sub-paragraph (3) below, shall be entitled to repayment of any advance corporation tax paid (and not repaid) in respect of franked payments made in the accounting period in question.

(3)

If no franked payments were made by the company in the return period for which a return is made by virtue of sub-paragraph (2) above the amount of the repayment shall not exceed the amount of the tax credit comprised in the franked investment income received; and in any other case the repayment shall not exceed the amount of the tax credit comprised in so much of that franked investment income, if any, as exceeds the amount of the franked payments made in that return period.”

The effect of these rather densely worded provisions may be summarised by saying that FII received in a later quarterly return period must first be applied in franking any dividends paid by the company in that period, but that any surplus may then be carried back to frank unrelieved dividends paid in an earlier quarter, thus generating a repayment of ACT. If there has been a change of ACT rates in the meantime, the repayment is not to exceed the amount of the tax credit comprised in the FII which is carried back.

The question which arises relates to the resulting repayment of ACT. Is the repayment to be treated as:

In my judgment the Revenue are correct on this point. Although the repayment is generated in its entirety by the receipt of actual FII, I can see no good reason why that fact should alter the characterisation of the ACT which is repaid, or create an exception to the general pro-rata approach to utilisation which I have held to be appropriate.

Issue 13: under the CT61 method, how are EU section 231 credits received in an accounting period to be attributed to quarterly ACT payments made in that accounting period?

This is a timing issue which arises under the claimants’ CT61 method. The problem is caused by the lack of information about the precise dates of receipt of EU dividends by the UK water’s edge companies. There is no difficulty about identifying the accounting period in which the dividends were received, because it is recorded in the accounts of the relevant companies. To identify the precise quarter of receipt, however, would require detailed investigation of bank records. The claimants have not attempted to do this, and I do not know whether such an exercise would have been feasible. It would certainly have been very labour intensive, even assuming the necessary data still to exist.

The claimants’ original approach to this issue was to set the credits for EU income received in an accounting period against the earliest quarterly payments of ACT made in that period. The effect of this approach was, of course, to maximise the period for which the credits remained unutilised. The claimants’ justification for its adoption appears to have been that the date on which EU dividends were received is irrelevant, because of the rules which enabled FII received later than franked payments made in the same accounting period to be carried back to relieve the earlier payments. There is, however, an obvious fallacy in this explanation, as the Revenue pointed out. The rules did enable FII to be carried back, if certain conditions were satisfied, but they did not alter the actual date of receipt of the dividends and their associated tax credits.

It is unnecessary for me to pursue this issue any further, because the claimants accepted in the course of the hearing that their approach could not be justified. Accordingly, it is now common ground that credit for EU income received in an accounting period is to be set pro rata against ACT paid in each of its quarters.

Issue 14: must credit for foreign corporation tax incurred upon the profits of foreign branches be given against ACT?

This is one of two Issues raised by the supplementary claim of the Ford group. It concerns the situation where a UK company had a branch in an EU Member State, and asks whether credit must be given for foreign corporation tax on the profits of the branch against ACT for which the UK company became liable when it distributed those profits to its shareholders. The important difference from the BAT test claims is that the relevant foreign profits are not the profits of an EU-resident subsidiary which were paid up by way of dividend to a UK water’s edge parent company, but are instead the profits of a foreign branch of a UK company. As such, the branch profits were in principle within the charge to UK corporation tax on the profits of the UK company. Section 6(1) of ICTA 1988 charges corporation tax “on profits of companies”, and section 8(1) provides that, subject to any exceptions provided for by the Corporation Tax Acts, “a company shall be chargeable to corporation tax on all its profits wherever arising”. The territorial scope of the profits chargeable to tax was therefore in principle unlimited.

The facts relevant to this Issue are agreed, and for present purposes may be shortly stated. The Ford group is the well-known manufacturer of motor vehicles which has its headquarters in Dearborn, Michigan, USA. The claimant is FCE Bank Plc (“FCE”), a UK-registered bank regulated by the Prudential Regulatory Authority and the Financial Conduct Authority. FCE is resident in the UK. It offers a variety of retail, leasing and wholesale automotive financial products and services to Ford dealers and customers through the Ford Credit and Ford Bank brand names. It is authorised by the Financial Conduct Authority to carry out a range of regulated activities in the UK and other European countries. The test claim concerns ACT paid by FCE in its four accounting periods ending 31 December 1995 to 1998. During that period, about 78% of the ordinary shares of FCE were owned either by the ultimate US parent company of the group, Ford Motor Company, or by US-resident intermediate holding companies. The remainder of the shares were owned by Ford Werke AG, a German-resident company, which was in turn wholly owned by Ford Motor Company.

In the relevant accounting periods FCE received income from branches in some 13 EU Member States. That income arose from trades in each of the Member States concerned. Under the terms of the double taxation agreements between the UK and each of those Member States, the income in question was taxable in the UK with credit for the tax paid in the source State. Thus the effect of the double taxation agreements was to relieve juridical double taxation of the branch profits. The UK did not surrender its right to charge the profits to corporation tax, but instead agreed to grant a credit for the foreign corporation tax paid on those profits in the branch State.

In 1995, 1996 and 1997 FCE paid dividends on its ordinary shares, and in 1996 and 1997 it also paid dividends on its preference shares (all of which were owned by Ford Motor Company). The payment of these dividends gave rise to the payment of ACT in the usual way, none of which has subsequently been utilised.

FCE argues that the reasoning of the ECJ in relation to the ACT charge on the onward distribution of profits received from EU subsidiaries should be applied by analogy to EU branch profits. Reliance is placed on FII (ECJ) I at paragraphs 85 to 90 of the judgment of the Court, where the ECJ regarded the charge to ACT on such profits, which had already borne foreign corporation tax, as the imposition of a second layer of taxation on those profits; whereas in a purely UK context ACT was payable once only (under the FII system) and was merely an advance payment of tax which could be set against the corporation tax charge on the distributed profits. This disparity of treatment discriminated unjustifiably against EU-resident subsidiaries. By parity of reasoning, submits FCE, the ACT charge on the distribution of EU branch profits involves the imposition of an independent second layer of tax on those profits. The first layer of tax is the foreign corporation tax charged in the branch State, and no credit is allowed for that tax against the subsequent ACT charge. By contrast, the profits which a UK company derives from its UK branches are subject to only a single layer of taxation, namely the charge to UK corporation tax of which ACT is a prepayment. Thus the system discriminates against EU branch profits when compared with UK branch profits, and infringes the UK company’s freedom of establishment elsewhere in the EU.

There is also a second limb to the argument. Relying on the principle of EU law that a company “cannot be required to pay in advance a tax to which it will never be liable” (FII (ECJ) I at paragraph 90), FCE submits that the UK, by entering into the relevant double taxation agreements with the branch States, has relinquished the right to charge corporation tax on the branch profits in favour of the source State. Since the UK will never be able to charge corporation tax on the branch profits, it cannot charge ACT when they are distributed.

The Revenue’s answer to these submissions is straightforward. They submit that there is no relevant difference of treatment between a UK company with EU branches and one without such branches. In each case, the UK company is chargeable to corporation tax on its worldwide profits under sections 6 and 8 of ICTA 1988. When the profits are distributed as dividends, each company is required to account for ACT on the same basis, under section 14. It is immaterial to this analysis that the UK has chosen to give credit for local corporation tax paid by a foreign branch against the UK charge to corporation tax on the branch profits. In so doing, the UK was providing relief for juridical double taxation of the same profits, but there was no requirement under EU law to provide such a credit: see Portfolio Dividends (No. 2) at [54]. Accordingly, the UK was entitled as a matter of EU law to charge corporation tax on the branch profits, and it must follow that it was likewise entitled to charge ACT on them.

In my judgment the Revenue’s submissions are well founded. There was no relevant difference of treatment under the UK tax system between the UK branches and the EU branches of FCE, or any other UK company. In each case, the worldwide profits of the company were subject to corporation tax, and ACT was chargeable on their onward distribution. The double taxation agreements did not negate the charge to corporation tax on the branch profits, but merely provided a credit for foreign corporation tax to set against the UK tax charged on them. This was a matter for negotiation between the UK and the Member States concerned, often in return for corresponding reliefs given when a company resident in the other State had a UK branch. The position differs fundamentally from the cross-border taxation of dividends, where the UK applied different systems of rules to dividends from UK and foreign subsidiaries.

Mr Aaronson argued that, even though the relief of juridical double taxation was not required by EU law, it is still necessary to have regard to the impact of the relevant double taxation agreements when answering the question whether there was a difference of treatment between UK and EU branch profits. In support of this argument he referred me to the decision of the ECJ in Case C-80/94, G. H. E. J. Wielockx v Inspecteur der Directe Belastingen [1995] ECR I-2508, [1995] STC 876. But Wielockx appears to me to be authority for a very different proposition, namely that a Member State cannot rely on the defence of fiscal cohesion to justify discrimination where such cohesion is secured only by the conclusion of double taxation agreements: see the judgment of the Court at paragraphs 23 to 25. I cannot find here any encouragement for the notion that double taxation agreements may be relied upon in order to discover discrimination under EU law where none would otherwise exist, and least of all where the discrimination would flow from measures taken to relieve juridical double taxation which is not a concern of EU law.

As to the second limb of FCE’s argument (see [220] above), there is in my judgment an equally short answer to it. If EU law is not engaged by a difference in treatment which infringes the freedom of establishment of UK companies with EU branches, the principle of EU law that “a company cannot be required to pay in advance a tax to which it will never be liable” cannot come into play at all. In the present context, application of the principle would depend on the jurisprudence of the ECJ which regards ACT as a prepayment of MCT. Unlike EU law, however, UK law has never characterised ACT in that way. Despite its name, ACT was a separate tax charged when a UK company made qualifying distributions, and it remained chargeable even if the company had no, or insufficient, MCT against which to set it off.

For these reasons, I conclude that FCE’s claims to recover ACT charged on the distribution of profits earned by its EU branches must be dismissed.

Issue 15: does it make any difference that the UK group had a non-resident parent which received double taxation treaty credits?

This is the second question raised by the Ford test claim. The only additional facts necessary for its resolution are set out as follows in the agreed statement of facts:

“49.

Upon the payment of the dividends which attracted those ACT liabilities [i.e. the ACT paid by FCE in the accounting periods ending 31 December 1995 to 1998], the US parent companies of FCE were entitled to receive a tax credit pursuant to Article 10(2) of the UK-US double taxation convention of 31 December 1975 (SI 1980/568). The credit was calculated as one half of the ACT paid less UK tax computed at 5% of the dividend plus the half ACT credit. Ford Werke AG was not entitled to such a tax credit as the terms of the UK-German double taxation convention made no provision for any like credit.

…

51.

In the accounting periods ending 31 December 1995 to 1997 FCE also received dividends from its wholly owned subsidiaries in Germany, Denmark and Austria. Those subsidiaries held no interests in companies outside their territory and the dividends distributed the profits earned in those jurisdictions and upon which, in every case, tax was paid at a rate equal to or exceeding the UK corporation tax rate and full double tax relief was afforded.

52.

FCE incurred no withholding tax on these dividends.”

The “half tax credits” referred to in paragraph 49 were duly paid by the Revenue to the US parent companies of FCE. The question is whether, in computing the restitution which FCE is entitled to claim for so much of the ACT as was unlawfully levied, credit should be given for the half tax credits paid to the US parent companies.

In arguing for an affirmative answer to this question, the Revenue rely heavily on the decision of the House of Lords on the first main issue in Pirelli Cable Holding NV v IRC [2006] UKHL 4, [2006] 1 WLR 400 (“Pirelli”). That issue concerned the calculation of the compensation payable to companies in the Pirelli group for the failure of the UK to permit group income elections to be made (as established in the Hoechst case) when dividends were paid by a UK-resident subsidiary to its parent companies resident in Italy or the Netherlands. Because no elections could be made, ACT was paid by the subsidiary in respect of the dividends. The parent companies, being non-UK resident, were not entitled to tax credits under section 231 of ICTA 1988, but they were entitled under the relevant double taxation agreements with Italy and the Netherlands to half tax credits similar to those payable under the UK/US agreement with which I am now concerned.

In those circumstances, Lord Nicholls framed the first question in the following terms at [7]:

“In the present case the first question arising on the assessment of compensation is this. If the United Kingdom legislation had permitted parent companies resident in other member states of the European Community to make a group election, and if an election had been made in respect of the dividends in question, would the parent companies have been entitled to payment of the convention tax credits they in fact received under the double taxation conventions? If they would have been so entitled then no deduction should be made in respect of these tax credits when calculating the compensation. If, however, the parent companies would not have been so entitled, then in principle – and subject to the other issues raised on this appeal – due allowance should be made in respect of these tax credits when calculating the compensation. Due allowance should be made because, in this event, the convention tax credits received by the parent companies comprise financial benefits they would not have received had the group been able to make a group income election and had the group duly done so.”

Reversing the courts below, the House of Lords held unanimously that, on the true construction of the relevant double taxation agreements, the convention tax credits would not have been payable if group income elections had in fact been made. The underlying reason for this was the “unspoken linkage” (as Lord Nicholls termed it in [1]) between liability to pay ACT and the grant of a tax credit, both in the UK tax legislation and (by extension) in double taxation agreements which conferred a half tax credit by reference to section 231. Thus Pirelli is authority for the proposition that, under the UK imputation system, the payment of ACT was a prerequisite for the receipt of a tax credit, including tax credits granted under double taxation agreements. See generally the speeches of Lord Nicholls at [8] to [15] and [19] to [21], Lord Hope at [35] to [39], Lord Scott at [67] to [72] and Lord Walker, who found it “a short but very difficult point of statutory construction”, at [103] to [107].

The House of Lords also held, in answer to the second question in the case, that credit should be given for the tax credits received by the parent companies in assessing the compensation due to the subsidiary for the payment of the unlawfully (or prematurely) levied ACT, notwithstanding the separate legal personalities of the companies concerned: see per Lord Nicholls at [17] to [22], Lord Hope at [40] to [44] and Lord Scott at [73] to [82]. In very simplified terms, the central point was that a group income election could only be made by the parent and subsidiary jointly, so the group had to be treated as a single unit for the purpose of assessing the compensation. As Lord Hope put it at [41]:

“I do not think that the revenue’s approach falls into the trap of ignoring the companies’ separate identities. What it does is to look at them instead as members of the group. I agree with Lord Scott that the relevant harm was the harm that the group suffered by reason of the breach of the parent companies’ right to freedom of establishment under article 52 of the Treaty. The breach deprived the group of the benefit of a joint election which, if there had been no breach, would have been available under section 247(1) of the 1988 Act for the benefit of the group as a whole. It would have been exercisable by the paying and the recipient members of the group jointly. It is the joint nature of the exercise that makes it appropriate to look at the group as [a] whole when the compensation is being assessed rather than the effect of the breach on each company taken in isolation.”

In their written closing submissions, the Revenue seek to apply the reasoning of the House of Lords in Pirelli to the present case in the following way:

FCE argues, by contrast, that the payment of the tax credits to its US shareholders is irrelevant. FCE submits that the question has to be considered in the context of the particular breach of EU law identified by the ECJ, which is fundamentally different from the breach which underlay the claims in Pirelli. In the present case, the breach of EU law was the failure of the UK to grant a (modified) section 231 tax credit to FCE in respect of the dividends received from its EU subsidiaries. If the UK tax system had operated in conformity with EU law, that credit would have to have been passed up to FCE’s shareholders in order to relieve the economic double taxation of the underlying EU profits. This objective was partially achieved by the grant of the half tax credits to the US shareholders, so it would be wrong to take them into account as a deduction when computing the restitution due to FCE for the ACT which was unlawfully levied. On the contrary, the grant of the tax credits was an example of the system operating, at least in part, as it ought to have done.

In considering these submissions, it is first necessary to analyse the nature of FCE’s EU law claim. I believe it to be common ground that, under EU law, FCE should have been granted modified section 231 tax credits in respect of its EU dividends. Since the profits underlying those dividends all originated in the relevant Member States, and were subject to national corporation tax at rates equal to or greater than the UK rate, FCE was not liable to Case V tax on the dividends. The charge was extinguished by the double taxation relief allowed for the underlying foreign tax. Further, the modified section 231 credits required by EU law, even if confined (as the Revenue submit) to credits at the relevant FNRs, would have been more than sufficient to extinguish any liability to ACT on the onward distribution by FCE of the EU dividends. However, ACT was in fact charged when FCE distributed the relevant income to its shareholders. Thus FCE’s basic claim is for restitution of the unlawful ACT which it was obliged to pay in respect of the income from its EU subsidiaries comprised in the dividends which it paid to its shareholders in its 1995 to 1998 accounting periods. Since none of this ACT was utilised, the claim is for restitution of the full amount of the unlawful ACT together with interest running from the dates on which it was paid.

In the eyes of EU law, the grant of modified section 231 tax credits to FCE would have been enough to remedy the defects in the UK system of taxation of EU dividends as it applied to FCE. EU law was not directly concerned with the grant of tax credits to FCE’s non-EU shareholders, and FCE’s US-resident parent companies would have had no rights under EU law to claim tax credits from the Revenue when they received dividends from FCE, had no credits been provided under the UK/US double taxation agreement.

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.

At this stage in the analysis, the second issue in Pirelli becomes relevant. Again differing from the courts below, the House of Lords was prepared to treat the Pirelli group as a single unit for the purpose of assessing compensation, on the basis, broadly speaking, that the making of a group income election would have benefited the group as a whole. But it was critical to this approach, in my judgment, that a group income election could not be made by the subsidiary alone. An election had to be made by the parent and the subsidiary jointly, because (as Lord Nicholls put it in [19]) the advantage to the dividend paying subsidiary (non-payment of ACT) came at a price to the recipient parent (deprivation of a tax credit).

Lord Nicholls continued:

“20.

Accordingly the loss sustained by the subsidiary cannot fairly be assessed in isolation. The commercial reality is that by not having the opportunity to make a group income election the group lost the opportunity to take advantage of a fiscal package: a package which affected the parent in one way and its subsidiary in a different way. In calculating the loss suffered by the group, that is, the parent and the subsidiary, regard must be had to both elements in the package. The effect on the parent must be considered as well as the effect on the subsidiary. The subsidiary lost the use of the money paid as ACT. In some instances, where the ACT was not set off against its mainstream corporation tax, the subsidiary lost the money altogether. But this cannot be treated as the amount of compensation payable by the Inland Revenue Commissioners without also taking into account any adverse consequence a group income election would have had on the parent. By the same token, assessment of the compensation must take into account any countervailing fiscal benefit received by the parent which would not have been available had a group income election been made.

21.

Pirelli sought to side-step this difficulty by presenting their claim as a claim by the subsidiary alone. But this difference in presentation cannot make any material difference in the outcome. A group income election cannot be made, or continued in force, by a subsidiary acting alone. Pirelli cannot, by presenting their claim in this way, alter the fundamental nature of the wrong for which compensation has become payable, namely, the loss of an opportunity for the parent and the subsidiary jointly to take advantage of a single fiscal package having different effects on the parent and the subsidiary.

22.

It is for this reason that assessment of the group’s overall loss, rather than the loss of the subsidiary alone, does not involve departure from the basic principle of company law that a parent company and it subsidiary are separate legal entities. Assessment of the overall loss represents the only fair way to assess the amount of loss suffered where a subsidiary and its parent have been denied the opportunity jointly to obtain a single package of this nature.”

The other two members of the House who dealt with this issue, Lord Hope and Lord Scott of Foscote, also placed emphasis on the fact that the infringement of Article 52 EC which grounded the claims for compensation was breach of the parent companies’ right to freedom of establishment. Lord Scott summed up this aspect of the matter at [80], as follows:

“80.

The correct analysis, in my opinion, is that the article 52 infringements of which the Pirelli claimants complain were infringements of the Pirelli parent companies’ right to freedom of establishment in the United Kingdom, that the primary measure of the loss thus caused was the financial detriment to their UK subsidiary caused by its liability to pay ACT but its inability to make group income elections, but that the benefit of the tax credits that the Pirelli parent companies would not have received but for the article 52 infringements should be brought into account. Logically, as it seems to me, the compensation should be compensation to the group for the detriment suffered by the group. Compensation that concentrates on the financial detriment of the subsidiary and ignores the financial benefit of the parent seems to me to overlook the nature of the article 52 infringement. The identity of the recipient or recipients of the compensation payable for the loss suffered by the group can be left to be decided by the group, but cannot, in my opinion, affect the quantum of the compensation.”

Lord Scott’s identification of the harm suffered by the group was expressly accepted by Lord Hope in the passage from his speech, at [41], which I have already quoted.

In the present case, by contrast, neither of the two main reasons given by the House of Lords for assessing compensation on a group basis is applicable. First, there is no parallel to the group income election which could be made only by the parent and subsidiary acting jointly, for the benefit of the group as a whole. Secondly, the US parent companies, unlike the EU-resident parent companies in Pirelli, have no rights of establishment protected by EU law. The only persons who can have a right of establishment under the Treaty are nationals of a Member State, including companies or firms formed in accordance with the law of a Member State: see the judgment of the ECJ in Hoechst at paragraphs 41 and 42, cited by Lord Scott in Pirelli at [76]. (This explains, incidentally, why it was never open to the Ford group to seek to recover the ACT paid by FCE on the basis that group income elections could not be made).

Despite the absence of any parallel with these two cardinal features of the facts in Pirelli, it may still be said that there was an indissoluble link between the payment of ACT by FCE and the payment by the Revenue of the half tax credits to the US parent companies. That is true, but in my judgment the existence of such a link cannot in itself justify disregarding the separate corporate identities of the companies, so as to take account of the credits paid to the US parents in assessing the restitution due to FCE for the ACT which it unlawfully paid. Indeed, it seems to me implicit in the judgments on the second issue in Pirelli that, in the absence of the two special features which I have examined, there would have been no justification for assessing the compensation on a group basis.

I conclude, therefore, that FCE is entitled to restitution of the ACT which it paid on the onward distribution of its EU dividend income, without having to give credit for the half tax credits paid to its US parent companies.

Issue 16: do any further adjustments need to be made to the claimants’ calculations?

As framed by the parties, Issue 16 is a residual question which asks:

“Do the Claimants’ and/or the Defendants’ calculations give effect to the principles decided above? If not, in what respects do they not do so and what adjustments need to be made?”

In my answers to Issues 1 to 15, I have set out the principles which should in my judgment be followed in quantifying the claimants’ claims in respect of unlawful Case V tax and unlawful ACT. In very broad summary:

In addition, I have dealt with the treatment of actual FIDs (under Issue 10), and I have answered a number of specific questions, in particular under Issues 3(a) to (e), 14 and 15.

I do not think it would be a useful exercise for me at this stage to try to work out whether any further adjustments need to be made to the claimants’ calculations. There are none which immediately occur to me, and I do not wish to embark on a time-consuming and unfocused exercise which might yield no useful results. I have now done my best to answer all the questions on the calculation of the unlawful tax put to me by the parties. The next step should be for the parties themselves to apply my answers to those questions, and try to agree the basic quantum of the claims. If any further questions arise, on which the parties are unable to agree, they should be formulated as precisely as possible and I will rule on them in due course.

Introduction

The issues which I have so far examined relate to quantification of the test claims. The remaining issues in the case assume that the claims have been correctly quantified, and relate in one way or another to the remedies sought by the claimants for the breaches of EU law which have been established.

I am satisfied that (with one possible exception noted in [190] above) all of the claims in the present trial are comprehended, as a matter of EU law, within the San Giorgio principle: that is to say, they are either claims to recover tax levied in breach of EU law, or claims which relate directly to such tax (including the time-value claims for utilised ACT, where according to the ECJ the charge to tax was not in itself unlawful, but the tax was levied prematurely: see the Hoechst case at paragraphs 87 and 88, and FII (ECJ) I at paragraph 205). The claimants also have other claims of a less direct nature (based, for example, on the use or surrender of various tax reliefs) where the classification of the claims under EU law has not yet been finally resolved, because it is still subject to outstanding applications for permission to appeal to the Supreme Court. Such claims have been deliberately excluded from those brought forward for determination in the present trial, so they may yet have to be decided at some (probably distant) future date.

In relation to the San Giorgio claims, it is elementary that EU law requires the national court to provide the claimants with an effective restitutionary remedy to vindicate their rights under EU law: see, for example, FII (CA) at [132] and [133]. As a matter of English law, the claimants rely on two causes of action in the law of restitution: the Woolwich cause of action (as it has now been explained and restated by the Supreme Court in FII (SC)) and the cause of action to recover tax paid under a mistake of law which was definitively recognised by the House of Lords in the DMG case (Deutsche Morgan Grenfell Group Plc v IRC [2006] UKHL 49, [2007] 1 AC 558).

It is convenient to record at this point that the Revenue accept that the claimants have good claims under the Woolwich cause of action to recover:

It is, however, no longer open to the Revenue to submit that the claimants’ mistake-based DMG claims are also confined to a six year limitation period, by virtue of section 320 of the Finance Act 2004 and/or section 107 of the Finance Act 2007. As a result of the decisions of the Supreme Court in FII (SC) and of the ECJ on the ensuing third reference, it is now clear that those statutory provisions were contrary to EU law in so far as they purported to curtail the extended limitation period under section 32(1)(c) of the Limitation Act 1980 which otherwise applied to the mistake-based claims. The Revenue have therefore been obliged to rely on different arguments in their quest to minimise the potentially huge financial impact of those claims, dating back as many of them do to 1973.

One of the most significant of those arguments forms part of the next issue which I have to consider, namely the extent to which the Revenue were enriched as a matter of English law by the payments of unlawful (or prematurely levied) ACT. In short, the argument is that the Revenue were never enriched by the payments of ACT because of their statutory obligation to grant corresponding tax credits to UK-resident recipients of the relevant dividends. Before coming to that question, however, I will first say a little about the underlying principles of the English law of restitution, and the impact upon them of EU law when it comes to providing the claimants with an effective remedy for their San Giorgio claims.

The restitution required by EU law

The Revenue prefaced their submissions on this part of the case with a summary of some of the main principles of that part of the English law of restitution which provides a remedy for unjust enrichment. I was referred in this context to the change of title, and the exclusion of “restitution for wrongs”, in the latest (eighth) edition of Goff & Jones, The Law of Unjust Enrichment, (2011) edited by Professor Charles Mitchell and others: see in particular paragraphs 1-01 to 1-05. As the Revenue point out, it has now become standard practice to analyse a claim for unjust enrichment in English law by reference to the four elements identified by Lord Steyn in Banque Financière de la Cité v Parc (Battersea) Ltd [1999] 1 AC 221 at 227, viz:

From the perspective of English law, it is axiomatic that, where unjust enrichment is established, it must (subject to defences) be reversed by requiring the defendant to give up the enrichment which he has unjustly obtained. As the Revenue put it in their written submissions, restitution for unjust enrichment is about reversing the enrichment of the defendant (at the claimant’s expense); it is not about compensating the claimant for his loss. There is ample authority for this proposition at the highest level: see, for example, Sempra at [30] to [32] per Lord Hope. Speaking of the claims to recover the time value of prematurely levied ACT in that case, Lord Hope said at [31]:

“The essence of the claim is that the revenue was unjustly enriched because Sempra paid the tax when it did in the mistaken belief that it was obliged to do when in fact it was being levied prematurely. So the revenue must give back to Sempra the whole of the benefit of the enrichment which it obtained. The process is one of subtraction not compensation.”

Lord Hope was there speaking of the correct characterisation of the claims under English law; but it is also necessary to consider whether a purely subtractive remedy of this nature will always be enough to satisfy a San Giorgio claim under EU law. Here it is important to note that, in accordance with the consistent jurisprudence of the ECJ, a claimant with a San Giorgio claim to recover unlawfully levied tax is entitled to receive compensation for the loss which it has sustained, in the form of (at least) reimbursement of the unlawful tax together with any amounts paid to the State or retained by it which relate directly to that tax.

Thus in FII (ECJ) I, in its discussion of remedies, the court at paragraph 203 repeated its standard learning that, in the absence of EU rules on the refund of unlawfully levied tax, it is for the domestic legal system of each Member State to lay down the detailed procedural rules governing actions for safeguarding rights derived from EU law, provided that those rules comply with the EU principles of equivalence and effectiveness. The court then continued:

“204.

In addition, the Court held in paragraph 96 of its judgment in [Hoechst] that, where a resident company or its parent have suffered a financial loss from which the authorities of a Member State have benefited as the result of a payment of [ACT], levied on the resident company in respect of dividends paid to its non-resident parent but which would not have been levied on a resident company which had paid dividends to a parent company which was also resident in that Member State, the Treaty provisions on freedom of movement require that resident subsidiaries and their non-resident parent companies should have an effective legal remedy in order to obtain reimbursement or reparation of the loss which they have sustained.

205.

It follows from that case law that, where a Member State has levied charges in breach of the rules of Community law, individuals are entitled to 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. As the Court held in paragraphs 87 and 88 of [Hoechst], that also includes losses constituted by the unavailability of sums of money as a result of a tax being levied prematurely.”

When stating its answer to the questions on remedies in paragraph 220, the Court said that national courts and tribunals are:

“… obliged to ensure that individuals should have an effective legal remedy enabling them to obtain reimbursement of the tax unlawfully levied on them and the amounts paid to that Member State or withheld by it directly against that tax.”

More recently, in the Littlewoods case, the ECJ has applied these principles in holding that the payment of interest on unlawfully levied tax is also a requirement of EU law, and not merely a procedural matter for the national court to determine: see Case C-591/10, Littlewoods Retail Ltd v HMRC, [2012] STC 1714 (“Littlewoods (ECJ)”) at paragraphs 23 to 26. The Court went on to say at paragraph 29 that the EU principle of effectiveness

“… requires that the national rules referring in particular to the calculation of interest which may be due should not lead to depriving the taxpayer of an adequate indemnity for the loss occasioned through the undue payment of VAT.”

I have considered the vexed question of what the ECJ meant by an “adequate indemnity” in Littlewoods (No. 2) at [253] to [302]. For present purposes, however, the important point is that the ECJ clearly considered that an adequate indemnity had to be provided under national law “for the loss occasioned through the undue payment of VAT”.

In the still more recent case of Irimie (Case C-565/11, Irimie v Administratia Finantelor Publice Sibiu and another, [2013] STC 1321), the Court has repeated and applied the same principles in the context of a claim to recover interest on the repayment of an unlawful Romanian pollution tax: see paragraph 26 of the judgment of the Court, and the discussion of Irimie in Littlewoods (No. 2) at [264] to [270].

In the light of this jurisprudence, I think it is now clear, as a matter of EU law, that the restitution required to give full effect to a San Giorgio claim may go significantly beyond the purely subtractive recovery which would be provided by the English law of unjust enrichment. There are at least three respects in which a San Giorgio claim may exceed a purely domestic claim to recover unlawfully levied tax:

To an English lawyer, these features of a San Giorgio claim may appear to sit uneasily in a no man’s land situated somewhere between the principles of unjust enrichment on the one hand, and compensation for loss caused by a legal wrong on the other hand. The remedy remains essentially restitutionary, in the sense that its core component is reimbursement of the unlawful tax, and repayment of amounts directly linked with it. Any wider loss caused to the claimant is recoverable, if at all, in a claim for damages on Factortame principles. But the measure of the restitution is the loss occasioned to the claimant by the payment of the tax, not the enrichment of the State by its receipt.

In my judgment the time has now come to recognise explicitly that a San Giorgio claim has a hybrid character in terms of classification under English law. Although such claims are most closely akin to claims in restitution for unjust enrichment, they also have important (but limited) compensatory elements, including (at least) the three which I have identified above. The phrase “restitutionary compensation” is not one which comes naturally to an English lawyer, but in my view it provides a convenient label, and I have already used it (or similar expressions) more than once in this judgment as a shorthand description of the nature of the remedy which the San Giorgio principle requires English law to provide. To the extent that the English law of unjust enrichment fails to provide such a remedy, the EU principle of effectiveness requires it to be moulded or adapted so that the remedy required by EU law is provided in full.

In his written and oral submissions to me, Professor Burrows stigmatised the approach which I have outlined above as one which involved the creation of a “hybrid monster”. With the greatest respect, I disagree. Recognition of the hybrid nature (in the eyes of English law) of San Giorgio claims seems to me no more, and no less, than the overriding effect of EU law requires. Indeed, in his Restatement Professor Burrows himself acknowledges on p 29 that “the law as set out in the Restatement is subject to any statute, or provision of EU law, to the contrary”. A taxonomy of the law of unjust enrichment which failed to recognise the possibility of English law being overridden in this way would, if I may be permitted the analogy, be as incomplete as a classification of equine quadrupeds which failed to recognise the existence of the mule, or a taxonomy of the genus Platanus which omitted hybrid species such as the London plane tree.

In their reply, the claimants submitted that the ECJ’s jurisprudence on the restitution of unlawfully charged taxes is an instance of the broader EU and civil law principle of the répétition de l’indu, i.e. the principle requiring the repayment of amounts paid or retained without any valid legal basis. This may well be correct, but the point was not further developed in argument, so I prefer to say no more about it.

Issue 17: 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?

This question focuses exclusively on the position under domestic English law, and raises the argument by the Revenue which I have already mentioned in [251] above, the gist of which is that the Revenue were not enriched by the receipts of unlawful ACT because of their obligation to grant corresponding tax credits to the recipients of the relevant dividends. The question also asks, more generally, whether the Revenue were enriched by the amounts of unlawful or prematurely levied tax received from the claimants, and (if so) to what extent.

Leaving aside the tax credit argument, the position is in my judgment as follows.

First, the Revenue were clearly enriched by the monetary value at the date of receipt of the sums of unlawful Case V tax and unutilised unlawful ACT which were paid to them. Those sums were all paid in money. As the current editors of Goff & Jones, op. cit., state at paragraph 4-12:

“In some circumstances, there is no danger that a defendant’s freedom to make his own spending choices will be compromised by ordering him to repay the market value of a benefit. The most obvious example is where the defendant receives money. Money is a universal means of exchange and defendants invariably desire things that money can buy. Hence they are invariably benefited by the receipt of money, and its face value is a reliable measure of their enrichment at the time when they received it.”

See too paragraph 4-37, where the authors explain why the defence of change of position does not detract from the rule that enrichment is tested at the date of receipt:

“It simply means that although the defendant was enriched by the value received, he need not repay the whole of this value because he has a partial defence to the claim – just as he need not repay it, for example, where the claim is subject to a time bar. To put this in another way, the claimant does not have to show both that the defendant received the value and that he still retains it: he only has to show that the value was received, and subsequent disenrichments must then be proved by the defendant seeking to rely on a change of position defence.”

The authors then quote the words of Millett LJ in Portman Building Society v Hamlyn Taylor Neck (a firm) [1998] 4 All ER 202 at 207, where he said that “the cause of action for money had and received is complete when the plaintiff’s money is received by the defendant”, and it “does not depend on the continued retention of the money by the defendant”.

Thus the claimants are prima facie entitled to recover the face value of the principal amounts of unlawful Case V tax and unutilised unlawful ACT which they paid to the Revenue, unless the Revenue are able to establish, the burden lying on them to do so, a valid defence of change of position to all or some part of the claim.

Secondly, the Revenue were also enriched by the time value of the use of the money paid to them, in the case of the unlawful Case V tax and unutilised ACT from the dates of payment until the eventual repayment of those sums to the claimants, and in the case of utilised ACT during the periods between the dates of payment of the ACT and the dates of its utilisation. See in particular Sempra at [33] per Lord Hope, [102] and [103] per Lord Nicholls, and [178] per Lord Walker, discussed by me more fully in Portfolio Dividends (No. 2) at [212] to [247], in the context of claims very similar to those in the present case, and in Littlewoods (No. 2) at [350] to [361].

Thirdly, I also consider that the Revenue were enriched by the time value of the continuing benefit arising from utilised ACT, from the dates of utilisation until the dates of eventual repayment to the claimants: see Portfolio Dividends (No. 2) at [245] to [246] and Littlewoods (No. 2) at [417].

As Lord Hope noted in Sempra at [33], referring to the views of the late Professor Birks, “the availability of money to use is not unequivocally enriching in the same degree as the receipt of money”. The same point was made by Lord Nicholls at [119], and by Lord Mance at [232]. In considering the extent of the Revenue’s enrichment in respect of the time-value claims, important and difficult questions therefore arise about the basis on which the enrichment should be ascertained, and the availability (both in law and on the facts) of the principle of so-called subjective devaluation. I will deal with these questions in detail when I come on to the Issues relating to change of position, the Revenue’s argument that English law only obliges them to make restitution of the actual benefit they obtained from the use of the money, and interest. I also need to remind myself that under English law it is for the claimants to establish the extent of the Revenue’s enrichment, at the first of the four stages of the standard analysis, although the distinction between the first stage and the defence of change of position can easily become blurred.

So far as the law is concerned, I have already considered many of the questions at some length in Portfolio Dividends (No. 2) and Littlewoods (No. 2). In particular, I have in both cases done my best to analyse the speeches in Sempra, and in Littlewoods (No. 2) I have discussed the question of subjective devaluation in the light of the judgments of the Supreme Court in Benedetti v Sawiris [2013] UKSC 50, [2014] AC 938 (“Benedetti”). My ultimate conclusion, in each case, was that the correct common law measure of the Revenue’s enrichment, under claims to recover the time value of unlawfully levied tax, is the objective market value of the use of the money in the hands of the Revenue, represented by an award of compound interest on the conventional basis adopted by the majority of the House in Sempra. To anticipate, my ultimate conclusion in the present case will be of a similar nature.

That being the general position, how (if at all) is it affected by the Revenue’s argument about tax credits? The argument mainly depends on the decision of the House of Lords on the first issue in Pirelli. It is a more ambitious version of the argument which I have already considered in the context of Issue 15. There, the Revenue sought to rely on the link between ACT and the grant of tax credits as a reason for reducing the restitution due to FCE. Now it is said that the Revenue were not enriched at all by the receipts of unlawful or prematurely paid ACT, because of their obligation to grant corresponding tax credits to shareholders. If the argument is sound, it would seem to follow that the claimants are left without any restitutionary remedy under English law in respect of their mistake-based claims to recover unlawful ACT or its time value. The claims would be defeated at the first hurdle.

In support of the argument, the Revenue submit that ACT and the shareholder’s tax credit formed the core of the partial imputation system introduced in the UK in 1973. These were the essential links between the company’s liability to corporation tax and the shareholder’s own tax liability: see the 1972 White Paper, Reform of Corporation Tax, (Cmnd 4955), at paragraph 9. This linkage is reflected in the reasoning of the House of Lords in Pirelli: see in particular the speeches of Lord Nicholls at [1], Lord Hope at [36] to [37] and Lord Walker at [103] to [105].

As Lord Hope said at [36]:

“The purpose of section 231 was to provide the receiving company with a tax credit equivalent in amount to the ACT payable on the dividend. The tax credit was designed to avoid tax having to be paid twice on the same dividend when tax was paid on its profits by the parent company.”

Lord Hope acknowledged in [37] that ACT would often remain surplus, because the paying company did not have sufficient profits at the end of its relevant accounting period to give rise to a liability to mainstream corporation tax. But, said Lord Hope:

“… the tax credit was always given, and given only, where the company making the qualifying distribution in respect of which it was given was liable under section 14(1) of the 1988 Act to payment of ACT on an amount equal to the amount or value of the distribution. As Mr Glick for the revenue put it, it was the payment of the ACT that made the giving of the tax credit to the recipient necessary.”

Similarly, Lord Walker said at [103]:

“… the clear scheme of the 1988 Act is that the payment of a dividend should be accompanied by a payment of ACT if a tax credit is to come into existence, and if exceptionally (because of a [group income election]) the payment of a dividend is not accompanied by a payment of ACT, the dividend would not give rise to a tax credit …”

Accordingly, say the Revenue, they received ACT with one hand but were obliged to pay it away again in the form of credits with the other, with the result that they were not enriched.

In my respectful opinion the argument now advanced by the Revenue is a hopeless one. I can see no answer to the simple point that the Revenue were immediately enriched by the full amount of the relevant receipts of ACT. Those receipts, like other tax receipts, were paid into the Consolidated Fund. There was no question of their being segregated, or held subject to an obligation to use them for any particular purpose. The link with the grant of tax credits to shareholders is of a different nature. It explains, at a conceptual level, how the partial imputation system was designed to operate; and in broad economic terms it explains how the tax credits were funded. But none of that has any impact on the Revenue’s initial enrichment. At best, it merely explains the part that the receipts were intended to play in the scheme of corporate taxation then in force.

The entitlement of UK-resident shareholders in receipt of qualifying distributions to tax credits under section 231 of the 1988 Act was not made conditional upon the levy of ACT in corresponding amounts, although the quantum of the credits corresponded to the rate of ACT in force for the financial year in which the distribution was made. Thus the right to tax credits derived from section 231, not from the payment of ACT. Similarly, the obligation of UK-resident companies under section 14 of the 1988 Act to pay ACT when they made qualifying distributions was not made conditional upon the grant of tax credits to shareholders, although the rate of ACT was fixed by reference to the basic or lower rates of income tax. The liability to pay ACT was triggered by the making of a qualifying distribution, and nothing else.

In the light of these statutory provisions, it is in my view impossible to regard the Revenue as a mere channel for the transfer of money from companies to their shareholders. The obligation to pay ACT on the one hand, and the entitlement to tax credits on the other, were the subject of independent statutory provisions, neither of which was made conditional upon the other.

Furthermore, the Revenue’s highly simplified model seems to me to break down in any event because of complexities for which it fails to make any allowance. I will mention three:

Finally, there appears to me to be an inconsistency between the Revenue’s argument on this point and their acceptance that the claimants have a valid Woolwich claim to recover unlawfully levied ACT. A Woolwich claim is part of the English law of unjust enrichment. The difference between Woolwich claims and those based on mistake lies in the nature of the “unjust factor” at the third stage of the traditional analysis. If such a claim is to succeed, it still needs to satisfy the first two stages. In other words, the claimant must be able to show that the defendant has been enriched at his expense. Thus enrichment of the defendant is a prerequisite of a Woolwich claim just as much as it is of a mistake-based claim. The Revenue did not offer any explanation, in either their written or oral submissions, of why they apparently accept that they were enriched by the unlawful ACT included in the Woolwich claims. In my view they were indeed enriched, but the same must also hold good for the mistake-based claims.

It is unnecessary for me to pursue these matters any further. As I have said, I am satisfied that, under English law, the Revenue were indeed enriched by the full amount of the unlawfully levied ACT.

There is, however, one aspect of the evidence on this part of the case which calls for comment. One of the Revenue’s expert witnesses was Sir Jonathan Stephens KCB. Sir Jonathan is a very distinguished public servant. He is of permanent secretary rank in the Civil Service, and currently works in HM Treasury reviewing arrangements for public sector pay. From 2006 to 2013, he was permanent secretary of the Department for Culture, Media and Sport. Before then, he worked on a variety of fiscal and spending matters in the Treasury from 1989 onwards, holding a number of senior posts from 2001 to 2006. As Director, Public Services, from 2001 to 2003, he oversaw the spending teams responsible for spending on health, transport, defence and foreign affairs, and police and criminal justice. As Director, Public Spending, from 2003 to 2004, he led the 2004 Spending Review. As Managing Director, Public Services, from 2004 to 2006, he was responsible for all public spending and a member of the Treasury’s departmental board.

With this experience, Sir Jonathan was unquestionably well qualified to give expert evidence on matters relating to public finance and spending, and much of his report dated 9 December 2013 was directed to such matters.

In the first section of his report, however, he discussed the operation of the ACT system and addressed the question whether the Government was enriched by receipts of ACT. He outlined the way in which he understood the system to operate, and concluded that the Government was not enriched. Under some justifiably severe cross-examination from Mr Aaronson, it soon emerged that this part of Sir Jonathan’s report was not based on his own knowledge or expertise, and amounted to little more than a presentation of arguments supplied to him by the Revenue. It was not a subject on which he was personally equipped to offer an expert opinion. As evidence, therefore, this part of his report was valueless, and in his closing submissions Mr Ewart rightly made it clear that he placed no reliance on it.

CPR rule 35.3 provides as follows:

“(1)

It is the duty of experts to help the court on matters within their expertise.

(2)

This duty overrides any obligation to the person from whom experts have received instructions or by whom they are paid.”

The general requirements of expert evidence are further explained in paragraphs 2.1 to 2.5 of the Practice Direction which supplements CPR Part 35. Reflecting earlier case law, this guidance includes the following:

“2.1

Expert evidence should be the independent product of the expert uninfluenced by the pressures of litigation.

2.2

Experts should assist the court by providing objective, unbiased opinions on matters within their expertise, and should not assume the role of an advocate.

…

2.4

Experts should make it clear –

(a)

when a question or issue falls outside their expertise;

…”

In the light of these well-known requirements, I find it hard to understand how Sir Jonathan and those advising him can have thought it appropriate to include the first section of his report, or how he could properly have subscribed his name to the expert witness declaration in standard form at the end of his report (where he said that he understood his duty to the court and had complied with it, and that he was aware of the requirements of CPR Part 35 and its associated Practice Direction). It ought to go without saying, but unfortunately needs to be repeated, that an expert witness should only give evidence on matters within his expertise; that he should not allow himself to be used as a spokesman for the party by whom he is engaged; and that eminence in one field, however great, does not qualify a person to proffer expert evidence in another field which lies outside his competence.

Issue 18: 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?

I can answer this question very briefly. If I have correctly analysed the nature of the restitution which is required under EU law in section (2) of this part of my judgment (see [252] to [263] above), it is clear that the EU principle of effectiveness requires the claimants to be given a remedy under national law which reimburses the whole of the unlawfully levied ACT. They have to be compensated, in the sense of receiving restitution, for the loss which they sustained by the payment of that tax. Viewed from the claimants’ perspective, their loss includes, at a minimum, the amounts of unlawful tax which they paid. Whether the Revenue were correspondingly enriched by the payments is irrelevant, because the remedy required by EU law is not a purely subtractive one. Accordingly, it must be contrary to EU law for the Revenue to contend that they are excused from making restitution of the unlawful tax on the basis that they were not enriched by its receipt.

In addition, I have already explained in Portfolio Dividends (No. 2) and Littlewoods (No.2) why I consider it to be 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: see Portfolio Dividends (No. 2) at [171] to [178] and Littlewoods (No. 2) at [418] to [425]. For that reason too, I consider the Revenue’s argument to be untenable as a matter of EU law, even assuming it to be correct as a matter of English law.

Change of position: introduction

I now come to the defence of change of position. I have already considered certain aspects of the defence in my judgment after the first liability trial in the present case: see FII (High Court) I at [303] to [352]. The Court of Appeal dealt briefly with an appeal from the relevant part of my decision in FII (CA) at [189] to [193]. More recently, I have returned to the topic, in a similar context, in Portfolio Dividends (No. 2) at [167] to [193]. Apart from the guidance given in FII (CA), the difficult issues to which the defence gives rise in cases of the present type have yet to be considered at an appellate level.

I will begin with an analysis of how matters stood in relation to the defence prior to the resumption of the present trial. For the full background, reference should be made to the passages cited above in FII (High Court) I, FII (CA) and Portfolio Dividends (No. 2).

The Revenue’s pleaded defence of change of position, as it stood in 2008, is set out in FII (High Court) I at [310]. Although renumbered, the relevant paragraphs remain substantially unaltered in the Revenue’s re-amended defence dated 11 February 2014. The core of the defence is now contained in paragraphs 35 and 36, which for convenience I will repeat:

“35.

Further and/or alternatively, if and to the extent that the [Revenue] were initially unjustly enriched, the [Revenue] have in good faith changed their position as a consequence of the payment by the Claimants and/or the Claimants in the FII GLO of the ACT Payments and/or the equivalent payments made by the other Claimants in the FII GLO such that it would now be inequitable and/or unconscionable to require the [Revenue] to make restitution of those sums.

PARTICULARS OF CHANGE OF POSITION AND OF INEQUITABILITY AND/OR UNCONSCIONABILITY

36.

The sums in question formed part of the United Kingdom’s tax revenue for the relevant year in which they were paid. Those sums have been irretrievably spent, in some cases decades ago.

…”

The above paragraphs relate to the ACT claims. The defence is pleaded in similar terms in relation to the other restitutionary claims.

In FII (High Court) I I held that, as a matter of EU law, the defence was precluded in relation to the test claimants’ San Giorgio claims, and that this was so whichever domestic restitutionary remedy the claimants relied upon to vindicate those claims. The Revenue had conceded at trial (although not in their pleaded defence) that they could not rely on change of position as a defence to claims based on the Woolwich cause of action, because this would breach the EU law principle of effectiveness. In other words, if the defence were otherwise made out on the facts, to allow it would deprive the claimants of the effective remedy which EU law requires a Member State to provide for the recovery of unlawfully levied tax. Since I took the view that the Woolwich remedy alone was not sufficient to enable the test claimants to obtain an effective domestic remedy for their San Giorgio claims, in particular because (as then generally understood) it did not extend to taxes (like ACT) which were not the subject of a demand, I concluded that the mistake-based DMG remedy was also needed to satisfy the claimants’ San Giorgio claims. Accordingly, change of position could not be permitted to defeat the claimants’ mistake-based claims any more than it could their Woolwich claims.

It followed from this reasoning that the potential scope which I perceived in 2008 for the defence to operate in the present case was very limited. It would only be open to the Revenue to rely upon it as a defence to any claims which (a) were properly classified as restitutionary claims in English law, but (b) were not, as a matter of EU law, San Giorgio claims: see FII (High Court) I at [304] and Portfolio Dividends (No. 2) at [170]. The remainder of my discussion of the subject in FII (High Court) I was therefore primarily directed to this narrow and as yet ill-defined category of claims, although my reasoning would have been equally applicable to all of the mistake-based claims had I been wrong in holding that they were protected by EU law in the same way as the Woolwich claims.

Furthermore, since the claims which have now been brought on for trial are (with the one possible exception already noted) all San Giorgio claims (see [247] above), it would seem to follow that, at any rate in the context of the present trial, the Revenue would be wholly precluded from relying on the defence unless either:

As to the first of those possibilities, the Court of Appeal has indeed held that the Woolwich cause of action alone provides the test claimants with a sufficient UK remedy for their San Giorgio claims, and that I was wrong to hold that a remedy in mistake is also required for that purpose: see FII (CA) at [152] to [174]. In reaching this conclusion, the Court of Appeal held that the Woolwich remedy does not depend on the making of a demand for the unlawful tax which it is sought to recover. The conclusion of the Court of Appeal on this point has since been definitively upheld by the Supreme Court in FII (SC), where Lord Walker reviewed the case law and academic writings on the subject before restating the principle as follows, at [79]:

“We should restate the Woolwich principle so as to cover all sums paid to a public authority in response to (and sufficiently causally connected with) an apparent statutory requirement to pay tax which (in fact and in law) is not lawfully due.”

See too the judgment of Lord Sumption at [171] to [174].

I will return to the second possibility noted above in the context of Issue 21 below. Meanwhile, it is enough to say I am satisfied that I ought to deal as fully as I properly can with all of the agreed issues in relation to the defence of change of position, and not decline to do so merely because I decided in 2008 that the defence is excluded in relation to San Giorgio claims.

A further reason for taking this course is that it is in principle desirable to consider the position under English law before looking at the impact on the defence of EU law. This point was made by Lord Walker in FII (SC) at [36], where he observed that the court had heard argument for the test claimants on issues of EU law, before hearing their submissions on issues of English law. Lord Walker then said:

“This sequence of argument may have been unavoidable, but it produced the result that the court heard submissions about the attitude of EU law towards national procedures and remedies – which is an important part of this appeal – before hearing submissions about the English remedies themselves. It is more helpful to start with the issues of English law, and then assess the impact that EU law has on them.”

Reverting to my judgment in FII (High Court) I, I observed at [303] that the question of the availability of the defence to the Revenue in respect of claims for the recovery of unlawfully levied tax was a novel one, on which there was no reported authority. I then reviewed the main English cases on change of position, beginning with the seminal decision of the House of Lords in the Lipkin Gorman case (Lipkin Gorman (a firm) v Karpnale Ltd [1991] 2 AC 548). I then stated my conclusions on the availability of the defence in the present case at [336] to [342]. For convenience, I will set out most of that passage:

“336.

In the first place, I can see no reason in principle why the defence of change of position should not be available to the Revenue, or indeed to any other category of defendant. No hint of any such limitation is to be found in any of the cases which were drawn to my attention, while there is, on the contrary, widespread recognition that a broadly based defence is needed in order to prevent injustice precisely because of the width and simplicity of the basic principle of unjust enrichment itself. As Lord Goff said at the end of his discussion of the subject in Lipkin Gorman, the availability of the defence will enable a more generous approach to be taken to the recognition of the right to restitution.

337.

On the question of what is meant by denial of the defence “to a wrongdoer” I have already indicated my view that what Lord Goff had principally in mind are cases where the cause of action relied upon by the claimant involves wrongdoing (in the sense of some recognised legal wrong) on the part of the defendant. On this approach, a defendant to a claim for restitution on the ground of mistake, whether of fact or law, is not a wrongdoer, because the cause of action does not depend on establishing any wrongdoing by him. It is, quite simply, not a fault-based claim at all.

…

339.

If that approach is applied to the present case, it suggests to me that the Revenue should not be entitled to rely on change of position as a defence to a Woolwich claim, because such a claim is founded on the unlawful levying of tax and therefore on the commission of a legal wrong, but that the defence should in principle be available to the Revenue when restitution is sought of overpaid tax on the basis of mistake. The critical distinction is that the unlawfulness of the tax forms no part of the cause of action in mistake. It is therefore irrelevant, just as the club’s conversion of the banker’s draft was irrelevant in Lipkin Gorman.

340.

The proposition that the Revenue should not be able to rely on change of position as a defence to a Woolwich claim gains further support, in my judgment, from the fact that no suggestion to this effect was made by Lord Goff in the Woolwich case itself, even though it was decided not long after Lipkin Gorman. Nor do I find this at all surprising. Subject to the need to bring his claim within a reasonable limitation period, the right of a private citizen to recover tax which has been unlawfully levied by the executive arm of government should, as a matter of principle, be unfettered. I can see no reason why corporate taxpayers, even if they are large multi-nationals, should be in any worse a position.

341.

The position is very different, however, where a taxpayer pleads his claim in mistake, and particularly where his motive in doing so is to take advantage of a more generous limitation period. As a matter of principle, it seems to me entirely appropriate that the defence of change of position should then be available, and that the Revenue should not be obliged to make restitution of mistakenly paid tax if it can establish the necessary factual ingredients of the defence.

342.

I appreciate that the question of who is a wrongdoer, in common with other aspects of the defence, needs to be worked out on a case by case basis, and Lord Goff intended his discussion in Lipkin Gorman to be no more than a starting point. However, I am not persuaded that the present case requires any broadening or redefinition of the concept beyond the simple test which I have set out. In particular, I do not consider that the Revenue should be regarded as wrongdoers, in the context of the mistake claims, merely because the mistake related to the lawfulness of the tax which was paid, and because it was the Revenue who had been responsible for the unlawful exaction of the tax. That is merely another way of saying that the unlawfulness which lies at the heart of a Woolwich claim should, as it were, infect the mistake claims too. I can only say that I see neither logic nor justice in such an extension.”

I then stated my preliminary conclusions on the factual elements of the defence, in the light of the evidence given for the Revenue by Mr David Ramsden. Mr Ramsden had been chief economic adviser to the Treasury since March 2008, and joint head of the government economic service since June 2007. I took the view that most of his evidence was directed to what I considered an irrelevant question, namely how a judgment against the Revenue for the full amount of the claims would be met, and the extent to which payment of such an amount would cause disruption to public finances. There is a summary of what I conceived to be the relevant parts of his evidence in FII (High Court) I at [349] to [352]. As to my conclusions on the factual aspects of the defence, it will again be convenient to cite most of what I said:

“343.

I now turn to the factual elements of the defence. I remind myself that the mere fact that the recipient has spent the money is not enough, and that a causal connection must also be shown, on at least a “but for” basis, between the receipt and any expenditure or other change of position upon which the defendant wishes to rely. Scottish Equitable v Derby suggests, albeit in a very different factual context, that the defence is not limited to specific identifiable items of expenditure, and that it will often be inappropriate to apply too demanding a standard of proof where an honest defendant has spent the money but cannot account for it in detail. How, then, should the court approach the question in a case where the Revenue has received many millions of pounds of mistakenly paid tax over a period stretching back for more than 30 years?

344.

To state the obvious, taxation is not imposed for its own sake, but in order to fund government expenditure. It is one of the two main ways in which public expenditure is funded, the other being public sector borrowing. One would expect government spending decisions, at a policy level, to be reached at least in part on the basis of the tax revenues which it has received in the past, and which it expects to receive in the future. Even if tax revenues are not spent immediately, common sense suggests that they will be used up over a fairly short period, and that it is probably safe to assume that tax receipts which predated the claims in the present case by more than six years, and therefore fell outside the scope of a Woolwich claim with its six-year limitation period, will have been exhausted well before the commencement of the action. As a matter of causation, no precise link can be demonstrated between particular receipts and particular items of government expenditure, but common sense again suggests that planned government expenditure would not have taken place at the level which it did but for the availability of the tax receipts which were taken into account in fixing departmental budgets. If all concerned, both the government and the taxpayers, proceeded on the footing that the tax was validly levied, I ask myself what is wrong with the argument that it would now be inequitable to require the Revenue to make restitution for the tax which was paid by mistake, because the money in question has long ago been spent in the public interest, and everybody assumed in good faith that it had been validly levied? I confess that, once the question is stated in these terms, the answer to it seems to me to be obvious. It would in my judgment be inequitable to require repayment in such circumstances, always bearing in mind that the claimants have a perfectly good separate San Giorgio claim for repayment of the unlawfully levied tax itself, free from any change of position defence.

345.

It may be objected to my analysis that the necessary causal link is not made out, because it would always have been open to the government to raise the necessary money in a different way (for example by borrowing, or by an increase in tax rates, or by a corresponding cut in expenditure elsewhere, or by a combination of those methods) if the mistakenly paid tax had not been available to it. However, although this argument has superficial attractions, it seems to me to miss the point. What matters is the existence of a causal link between the tax revenues which were in fact received and the expenditure which was actually made. The fact that the money could have been raised in a different way is, in my judgment, neither here nor there.

346.

A related objection is that, once departmental spending policy has been fixed and budgeted for, the various items or heads of expenditure should be regarded as commitments which have to be honoured, and are thus analogous with debts which have to be paid off in one way or another, rather like the mortgage in Scottish Equitable v Derby. In my view the answer to this point is again the same. Departmental spending plans are themselves likely to be predicated in part on the receipt of identifiable tax revenues, and (if so) they cannot be treated as purely extraneous obligations which the government would anywhere be under an obligation to fund.”

In the order which was drawn up on 12 December 2008 to give effect to my judgment in FII (High Court) I, it was declared in paragraph 14 that:

“The Defendants are entitled to maintain a change of position defence in respect of any mistake-based restitutionary claims which go beyond San Giorgio claims.”

In the Court of Appeal, the issues considered by the court (as set out in the schedule to the Court of Appeal’s order dated 19 March 2010, as amended) included the following:

“Issue 12 Is a remedy under Woolwich a sufficient remedy for EU law claims under San Giorgio, or is a remedy in mistake also required?

…

Issue 15 Is any change of position defence precluded by the wrongdoer principle?

Issue 16 Has the Revenue demonstrated a change of position as a matter of fact?

Issue 17 Is the defence of change of position available if a mistake remedy is (contrary to the Judge’s findings) not required by EU law for a San Giorgio claim? ”

In relation to issue 12, the Court of Appeal held, as I have already explained, that a remedy in mistake was not required under EU law to satisfy the claimants’ San Giorgio claims: see [296] above.

The Court of Appeal found it unnecessary to deal substantively with issues 15 to 17, because (put shortly) it was common ground that the conclusions which I had stated in the paragraphs quoted above were of an interim nature, despite the apparently unqualified terms of the declaration contained in paragraph 14 of the order under appeal: see the judgment of the court at [189] to [190]. The court then added these observations:

“191.

In any event, on our analysis and decisions, the defence of change of position does not call for consideration. It is common ground that the Community law principle of effectiveness precludes the application of any change of position defence to San Giorgio claims. For the reasons we have given, all the San Giorgio claims fall within the Woolwich cause of action. Claims for restitutionary relief based on mistake are, for the reasons we give, subject to the limitations imposed by s 33 of the Taxes Management Act 1970, the Finance Act 2004, s 320 and the Finance Act 2007, s 107 (see below). We did not understand the Revenue to be arguing for a change of position defence in relation to claims within those statutory provisions since, constrained by those limitations, they would mirror the San Giorgio claims enforceable under Woolwich.

192.

In those circumstances it is not necessary to consider the submissions made by each side as to the proper legal tests and approach on change of position and as to the judge’s analysis of the defence. The defence is highly fact-sensitive, and anything we were to say about it would be obiter. The submissions and the judge’s analysis raise important and difficult questions of law and policy: see generally the discussion by Professor Elise Bant in Restitution from the Revenue and Change of Position [2009] LMCLQ 166. In all those circumstances we do not consider it appropriate to comment further on the defence.

193.

Accordingly, on this Issue, we conclude for the reasons given above that the defence of change of position does not call for consideration.”

In the light of this reasoning, paragraph 11 of the Court of Appeal’s order stated that the appeals on issues 15, 16 and 17 did not arise for decision, and declaration 14 was affirmed subject to the insertion of the words “in principle” so that it now reads (with my emphasis):

“The 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.”

It is also important to note that the reasons given by the Court of Appeal, at [191] of the judgment of the court, quoted above, for saying that the defence of change of position does not in any event call for consideration, are no longer sustainable. The Supreme Court has now held, unanimously, that the Revenue are unable to rely on either section 33 of TMA 1970 or section 107 of the Finance Act 2007 to exclude or curtail the mistake-based restitutionary claims. The Supreme Court was divided on the question whether section 320 of the 2004 Act also infringed EU law. On the ensuing third reference, the ECJ held that it did.

Against this background, the position at the start of the present trial may I think be summarised in the following propositions:

In Portfolio Dividends (No. 2) I heard argument on both of the questions identified in sub-paragraph (3) above, but I only found it necessary to rule on the second of them. I held that reliance on the defence is precluded by EU law in relation to all claims which are brought in a Member State to vindicate San Giorgio rights: see [171] to [189]. Since I also held that it was in any event not open to the Revenue to rely on the defence, because they had adduced no evidence, anything that I said about the first question would have been obiter and would also have lacked any factual foundation: see [193]. In the present case, by contrast, the Revenue have adduced some evidence relevant to establishment of the defence, and the first question is one of the matters I have been asked to decide. Before coming to it, however, I must first briefly explain why it is agreed that the defence is not available in relation to Woolwich claims.

Why is change of position not available in English law as a defence to Woolwich claims?

In FII (High Court) I the reason which I gave for holding that change of position is not available to the Revenue as a defence to Woolwich claims was that such claims are “founded on the unlawful levying of tax, and therefore on the commission of a legal wrong”: see [339], quoted above. I therefore regarded Woolwich claims as examples of the “wrongdoer” exception which Lord Goff had recognised, but not elaborated upon, in Lipkin Gorman.

This part of my reasoning has attracted a good deal of academic criticism. See, for example, Goff & Jones, The Law of Unjust Enrichment, at paragraphs 27-43 and 27-44; Shiers and Williams, [2009] BTR 365 at 375; and Elise Bant, Restitution from the Revenue and Change of Position, [2009] LMCLQ 166 at 172 (the article referred to by the Court of Appeal in FII (CA) at [192]).

Professor Bant has argued that a better explanation is to be found in the principle which she calls “stultification”. She explains this as follows, in the article cited above at 172:

“However, Henderson J’s conclusion on the inapplicability of the change of position defence to Woolwich claims can be justified by another route. This can be termed the “stultification” bar to the application of the defence. This arises whenever the application of a particular rule would undermine an overriding policy of the law. In the context of change of position, the bar entails that the defence should be denied where its recognition would stultify the policy reason for ordering restitution.

Stultification ought to be the most significant hurdle to public authorities seeking to rely on the defence in circumstances where they have received a benefit pursuant to an unlawful demand. The Woolwich principle rests on judicial recognition of a longstanding and vitally important public interest in prohibiting unlawful demands by public authorities. The policy of prohibiting such demands will be undermined if a public authority is effectively permitted to keep the benefit of the unlawful demand by relying on the defence of change of position.”

This approach is also adopted by Professor Burrows in his Restatement, section 23(2) of which states that the defendant does not have the defence if:

“(b)

the weight to be attached to the unjust factor is greater than that to be attached to the change of position (as, for example, where the unjust factor is the unlawful obtaining of a benefit by a public authority).”

In the commentary on p 122, Professor Burrows refers to my holding in FII (High Court) I and comments:

“The best explanation for this is that the justification for the Woolwich principle outweighs concern for the position of the defendant.”

See too Burrows, The Law of Restitution (3^rd edition, 2011) at pp 543 and 698.

Wearing his hat as an advocate in the present case, Professor Burrows argued that in FII (High Court) I I had reached the right conclusion on this point, but for the wrong reason. He submitted that the charging of unlawful tax, or ultra vires conduct by a public authority, is not a tort or other civil wrong as such, and that restitution on the Woolwich principle is not restitution for a wrong: it is restitution for unjust enrichment. The true reason for the bar on the defence therefore lies in the policy which underlies the Woolwich “unjust factor”, namely the constitutional importance of a citizen having a right to restitution of unlawfully levied tax. The potency of that policy outweighs any concern for the defendant’s change of position.

I consider that I was correct to regard the Revenue as a wrongdoer in the context of the present case. The Revenue committed a clear breach of EU law in levying the unlawful tax, and if sufficiently serious the breach could have founded a claim for damages against the State as well as claims for repayment of the tax. As Goff and Jones point out, however, at paragraph 27-44, “the receipt of money paid as tax that is not due is not always wrongful, in the sense that the defendant must always commit a breach of legal duty when receiving such payments”. In terms of the reformulation by the Supreme Court, it is enough to engage the Woolwich principle that sums should have been paid to a public authority in response to “an apparent statutory requirement to pay tax which (in fact and in law) is not lawfully due”. It is the unlawfulness of the apparent charge, rather than any breach of legal duty in its imposition, which lies at the heart of the principle.

On balance, therefore, I acknowledge the force of the criticisms made of my reasoning in FII (High Court) I on this point, and I now think that a better explanation for the bar on the defence of change of position to Woolwich claims is to be found in the stultification principle advanced by Professor Bant and other scholars. In essence, to allow scope for the defence would unacceptably subvert, and be inconsistent with, the high principles of public policy which led to recognition of the Woolwich cause of action as a separate one in the English law of unjust enrichment, with its own specific “unjust factor”.

Issue 19: 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?

This Issue squarely raises the question which, as I have explained, I found it unnecessary to answer in Portfolio Dividends (No. 2): see [308] above. Before coming to the substance of the debate, however, I must first deal with a preliminary point taken by the Revenue. They argue that I have already decided the question in their favour in FII (High Court) I, with the result that it is res judicata at first instance (and, arguably, in the Court of Appeal, given the terms of paragraph 11 of that court’s order: see [305] above).

The argument runs briefly as follows. In FII (High Court) I, I held at [339] (quoted above) that “the defence should in principle be available to the Revenue when restitution is sought of overpaid tax on the basis of mistake”. The reason which I gave for thus distinguishing mistake-based claims to recover unlawful tax from Woolwich claims was that “the unlawfulness of the tax forms no part of the cause of action in mistake” (ibid). I added in [341] that the position where the claim was pleaded in mistake seemed to me “very different” from Woolwich claims, particularly when the taxpayer’s motive in relying on his mistake in paying the tax was to take advantage of a more generous limitation period. It is true that my brief discussion of the question was technically confined to restitution claims which went beyond San Giorgio claims, but (so it is argued) the position under English law cannot depend on whether the defence is precluded by EU law. Thus my reasoning in relation to mistake-based claims which are not San Giorgio claims must extend to the much larger category of mistake-based claims which are San Giorgio claims. By implication, the declaration contained in paragraph 14 of my order after the first trial, as upheld and varied by the Court of Appeal, must be taken to apply to all of the mistake-based restitutionary claims.

Unless I am mistaken, the claimants did not advance any answer to this argument in their otherwise very comprehensive written and oral submissions. In any event, it seems to me that the argument is well-founded. I cannot see any sensible reason for distinguishing in this context between the mistake-based claims which, as a matter of EU law, either do or do not fall within the San Giorgio principle. The availability of the defence as a matter of English law cannot depend on the classification of the claims under EU law. Moreover, there is no suggestion that this part of my decision at the first trial was of a preliminary nature. The question is one of law, and my conclusion was unqualified. The question has therefore been finally determined by me at first instance in the BAT test claims (subject to the modification of my declaration by the Court of Appeal).

I will nevertheless go on at least to rehearse some of the substantive arguments which were addressed to me on this Issue, for the reasons given in [10] above. I am further encouraged to take this course by the fact that six years have elapsed since the first trial, and the question has been more fully argued (with the benefit of some significant new material, both in the case law and in academic writings) than it was in 2008.

The argument for the claimants, skilfully advanced by Mr Beazley, starts from the combination of constitutional and policy considerations which underpins the Woolwich principle and rules out change of position as a defence to Woolwich claims. In his written closing submissions, Mr Beazley grouped these considerations under four headings:

Widespread support for these principles can be found in the authorities, and their force in the context of the Woolwich remedy is not disputed by the Revenue, although they would not necessarily formulate them in quite the same way. It is therefore unnecessary for me to examine them in any detail. The main authorities relied upon by the claimants include:

The next step in the argument is that the development of the English law of unjust enrichment, as it applies to recovery of unlawful tax paid by mistake, should be moulded by the same considerations as apply to the Woolwich remedy itself. Considerations of policy and justice lie at the heart of the law of unjust enrichment, as Lord Goff emphasised in Lipkin Gorman at 579-580 and repeated in Woolwich at 171-172. The law should develop in the light of these guiding principles, said Mr Beazley, and dogmatic classification should be avoided. Mr Beazley commended to me the principled but cautious approach to the development of change of position adopted by the Court of Appeal in the Niru Battery case (Niru Battery Manufacturing Co v Milestone Trading Ltd [2003] EWCA Civ 1446, [2004] QB 985) and in Commerzbank AG v Gareth Price-Jones [2003] EWCA Civ 1663, [2003] 147 SJLB 1397). Thus, for example, Clarke LJ (as he then was) said in Niru Battery at [149]:

“In short, as I read the speeches in the Lipkin Gorman case, the essential question is whether it would be inequitable or unconscionable, and thus unjust, to allow the recipient of money paid under a mistake of fact to deny restitution to the payer.”

Mr Beazley went on to submit that the considerations which typically ground a defence of change of position in an ordinary private law claim to recover payments made by mistake are either irrelevant to tax recovery claims, or apply with very much less force. For example, a perceived need to protect settled transactions, or to protect a private defendant who has incurred extraordinary expenditure in reliance on a mistaken payment, is of no or little relevance in terms of policy or justice when a claim is brought against the State to recover mistakenly paid tax. In such cases, says Mr Beazley, the balance between the rights of the claimant and legal certainty in relation to past receipts by the Revenue lies in the balance struck by Parliament through the imposition of time limits for the bringing of claims.

If the matter is viewed in this way, the conclusion naturally follows that change of position should not be permitted as a defence to mistake-based claims for the recovery of unlawful tax, any more than it should for Woolwich claims. Strong academic support for this conclusion is provided by the current editors of Goff and Jones, who say at paragraph 27-51 (in a chapter for which Professor Charles Mitchell assumed responsibility):

“… 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 …”

See too paragraphs 27-43 and 27-44, criticising my reasoning in FII (High Court) I as “suspect for several reasons”.

The argument for the Revenue, deployed with equal skill by Mr Ewart and Professor Burrows, starts (in contrast) by emphasising the general nature of the defence. This was expressly recognised by the Court of Appeal in Haugesund Kommune v Depfa ACS Bank [2010] EWCA Civ 579, [2012] Bus LR 1. The case concerned two Norwegian local authorities which had entered into swaps contracts with the defendant bank which were held to be void under the relevant Norwegian legislation. One of the issues was whether the local authorities could rely on change of position as a defence to the restitutionary claims brought by the local bank. The leading judgment was delivered by Aikens LJ. After reviewing the development of the defence in the case law in [109] to [121], he said at [122]:

“I think that the following propositions are obvious from the foregoing analysis of the cases. First, the defence of change of position is a general defence to all restitution claims (for money or other property) based on unjust enrichment. Secondly, the question is always whether it is unjust to allow the claimant to have restitution in whole or in part. Thirdly, that itself depends on whether the defendant has so changed his position that the injustice of requiring him to repay outweighs the injustice of denying the claimant restitution (in whole or part). That can only be decided on the facts of each individual case. Fourthly, a defendant cannot rely on this defence if he has acted in bad faith or has failed to show good faith, which is not the same thing as having acted dishonestly.”

Pill LJ agreed with the conclusions and reasoning of Aikens LJ: see [155]. Although dissenting on certain issues, Etherton LJ also agreed that the change of position defence should be dismissed: see in particular his judgment at [152].

Next, the Revenue point to the existence of what is now a considerable body of authority in which it has been held, or apparently taken for granted, that it is in principle open to a public authority to rely on the defence. FII (High Court) I may have been the pioneer, but the trail has subsequently been followed by Hamblen J in Bloomsbury International Ltd v Sea Fish Industry Authority [2009] EWHC 1721 (QB), [2010] 1 CMLR 12, at [133] to [144] and by Vos J in Littlewoods (No. 1) (Littlewoods Retail Ltd v HMRC [2010] EWHC 1071 (Ch), [2010] STC 2072), at [112] to [131].

In a little more detail, the Bloomsbury International case concerned the lawfulness, under both UK delegated legislation and EU law, of the levy imposed by the Sea Fish Industry Authority (“the SFIA”) to finance its activities. Hamblen J held that the levy was valid under both domestic and EU law, and although reversed by the Court of Appeal, his conclusions were affirmed by the Supreme Court: see [2011] UKSC 25, [2011] 1 WLR 1546. Accordingly, the part of Hamblen J’s judgment which deals with remedies was obiter, and neither the Court of Appeal nor the Supreme Court dealt with change of position. Nevertheless, the judge accepted that the defence of change of position was open to the SFIA in relation to mistake-based claims which preceded the issue of proceedings by more than six years, and found that the defence had been made out on the facts. He followed the approach which I had adopted in FII (High Court) I, while recording at [140] that the claimants reserved the right to challenge that approach in a higher court.

In Littlewoods (No. 1), one of the issues which Vos J had to consider was whether change of position is an available defence in English law to Woolwich claims and/or mistake-based restitutionary claims. In dealing with this issue, Vos J rejected an argument advanced by Mr Jonathan Swift QC for the Revenue that the defence should be available to Woolwich claims, but accepted that the defence was available for mistake-claims: see [101] to [109]. It is fair to note, however, that the contrary had not been argued by Mr Laurence Rabinowitz QC for the claimants, who was content to accept for the purposes of the hearing that my approach in FII (High Court) I had been correct, while reserving his position in any higher court on the question whether change of position should ever be available in relation to claims concerning repayment of tax: see [103].

Having thus held that the defence was in principle available in relation to the claimants’ mistake-based claims, Vos J then considered whether it was made out on the facts. For the reasons given at [112] to [131], he held that the defence failed.

Mr Ewart also referred me to certain passages in DMG and Sempra where it appears to have been assumed that the defence of change of position would be available to the Revenue. Although there was no issue about the availability of the defence in either case, the relevant observations in DMG, in particular, are worthy of note, because that was the case which established the existence of a mistake-based remedy to recover unlawful tax from the Revenue. Lord Hope referred at [38] to the speech of Lord Goff in Kleinwort Benson Ltd v Lincoln City Council [1999] 2 AC 349 at 371-373, and summarised the grounds upon which Lord Goff had held that English law should recognise the right to recover money paid under a mistake of law. Lord Hope concluded his summary with these words:

“As [Lord Goff] put it, at p 373C, a blanket rule of non-recovery, irrespective of the justice of the case, could not survive in a rubric of the law based on the principle of unjust enrichment. Instead it was for the law to evolve appropriate defences which could, together with the acknowledged defence of change of position, provide protection where appropriate for recipients of money paid under a mistake of law in those cases in which justice or policy does not require them to refund the money.”

In my view the Revenue’s argument gains no real support from this passage, where Lord Hope was doing no more than setting the background by reference to what Lord Goff had decided in Kleinwort Benson. Furthermore, Lord Hope’s reference to “the acknowledged defence of change of position” is balanced by his reference to cases in which justice or policy requires the money to be refunded.

Still less, in my view, can the Revenue gain assistance from the speech of Lord Walker in DMG. In rejecting the argument that a “settled law” defence should be recognised by the courts, Lord Walker said at [145]:

“As Lord Goff recognised in the Woolwich and Kleinwort Benson cases, there are strong policy arguments in favour of a general common law right of recovery of wrongly exacted tax, and also strong countervailing arguments in favour of some restrictions on that right of recovery, especially as the defence of change of position could seldom, if ever, apply to wrongfully exacted tax. But the balancing of these high policy arguments is essentially a matter for the legislature.”

Lord Walker therefore clearly took the view that the defence could “seldom, if ever” apply to wrongly exacted tax, in a context where he must have had mistake-based claims firmly in mind. The most that can be said is that, in a case where the point was not in issue, Lord Walker did not entirely rule out the possibility of the defence applying.

The passages to which I was referred in Sempra do not in my opinion take the matter any further, because they do no more than reflect the general availability of change of position as a defence in cases of restitution for unjust enrichment. For the record, the passages in question are contained in the speeches of Lord Nicholls at [119] and Lord Scott at [132] and [151].

The Revenue argue that the difference between the Woolwich and mistake-based remedies is of a fundamental nature, and goes far beyond a technical distinction between the unjust factors relied on. The Woolwich remedy is founded on a policy-based unjust factor, whereas mistake belongs with other unjust factors concerned with impaired consent. The Woolwich remedy was developed to reflect high constitutional principles, and for that reason is not dependent on the making of a mistake. By contrast, when a mistake-based claim is made to recover unlawfully levied tax, the claim is not based on the policy considerations which underpin the Woolwich remedy, but merely on the impaired consent of the claimant. The claim would in principle be just as strong if there were no unlawful tax involved, for example where there has been a double payment of lawful tax by mistake.

Given the fundamentally different nature of the two remedies, it is not surprising, say the Revenue, if they have different limitation periods and admit of different defences. In particular, recognition of mistake of law as an unjust factor evolved hand in hand with recognition of change of position as a defence: see the speech of Lord Goff in Kleinwort Benson, and Portfolio Dividends (No. 2) at [185] to [187]. If, therefore, the claimants wish to take advantage of the mistake-based remedy, with its extended limitation period, they must take the remedy with its attendant disadvantages, which include the potential defence of change of position. There would be no injustice to the claimants in such a result, because they already have their Woolwich claims, free from any defence of change of position, within the limitation period laid down by Parliament for such claims.

The Revenue disclaim any suggestion that they are adopting a rigid, or formalistic, adherence to categories. They accept that particular heads of policy may apply to more than one cause of action, but submit that a careful assessment is always needed to see whether the relevant policy has a legitimate role to play in relation to the cause of action in question. Thus, for example, the policy which prevents exclusion of liability for fraudulent misrepresentation does not mean that liability for negligent misrepresentation may not legitimately be excluded. In the present context, English law has specifically devised the Woolwich cause of action to meet the relevant policy concerns, and it would be unfair to the Revenue (representing the public as a whole) if those same policy concerns were brought back to eliminate change of position as a defence to the mistake claims. To allow this would be to permit “policy double counting”.

In their written closing submissions, counsel for the Revenue go on to say:

“Moreover, it is imperative not to lose sight of the consequence of acceding to the Test Claimants’ argument. It would produce what reasonable people looking at this case would surely regard as the grotesque result in policy and principle that claimants can go back and back, right to the start of the ACT system, and reclaim tax and interest even though the Government was acting in good faith and has irretrievably spent the money years ago in the public interest. The Test Claimants present the policies as if the only legitimate way of protecting defendants is for Parliament to reform limitation periods (which of course it has tried to do albeit without success for the purposes of this case). But that ignores the established common law method of protecting defendants against restitutionary claims which is the change of position defence.”

With the benefit of these rival submissions, I am afraid I must begin by saying that it would in my judgment be wrong for me to express a concluded view on them. The question is one of difficulty and importance, on which there is much to be said on each side. If, however, I am correct in my belief that I have already finally answered the question in the Revenue’s favour at the first trial in 2008, it follows that I am now functus officio in relation to it in the context of the present action. It cannot make any difference to this analysis that the question arises in the context of group litigation managed under the FII GLO. I therefore consider that I lack jurisdiction to revisit the question in the same proceedings. And even if that is too stark a conclusion, I think that it would be inappropriate for me to do so. There would be a real risk that I might end up deciding the same question in different ways in the same case.

What is badly needed, in my judgment, is consideration of the question by an appellate court, at the earliest convenient opportunity. In so far as I have any continuing role to play at first instance, I think it should be confined to adding to my discussion of the subject six years ago any further material which I consider may help to inform the debate in a higher court.

To that end, I have thought it appropriate to record some of the arguments ably presented to me on each side, and to refer to and comment on some of the relevant case law and academic writings which have appeared in the meantime. Further than that, however, I am satisfied that it would be inappropriate for me to go.

Issue 20: have the Revenue made out a defence of change of position on the facts?

On the assumption that it is in principle open to the Revenue under English law to rely on change of position as a defence to the mistake-based claims, and subject to the question (considered under Issue 21 below) whether such reliance is precluded by EU law, the present Issue asks whether the defence has been made out on the facts.

Certain matters are common ground. It is agreed that the burden of proof lies on the Revenue to establish the defence. It is agreed that the defence is not of an “all or nothing” nature, and can operate to the extent that it is made good, or “proportionately” as it is sometimes said. The parties are also content (as am I) to take as correct the formulation of the defence in section 23(1) of Professor Burrows’ Restatement:

“The 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.”

At p 119, Professor Burrows explains that the loss is “a consequence of … obtaining the benefit” if it is causally relevant, and that it is normally unnecessary to go further and show detrimental reliance by the defendant. (The position is different where the loss is incurred in anticipation of an expected enrichment, because it is then “hard to see how there can be relevant loss unless there has been detrimental reliance on obtaining the benefit”, p 120.) As to the relevant test for “causally relevant loss”, it has yet to be laid down in the cases, but the appropriate starting point is a “but for” test (p 120, referring to Scottish Equitable Plc v Derby [2001] EWCA Civ 369, [2001] 3 All ER 818, at [31] per Robert Walker LJ).

It is worth noting that in Scottish Equitable the Court of Appeal clearly endorsed the so-called “wide” version of the defence. Robert Walker LJ (with whom Keene and Simon Brown LJJ agreed) said this:

“30.

The judge noted the view, put forward by Andrew Burrows (The Law of Restitution (1993) pp 425-428) that there is a narrow and a wide version of the defence of change of position, and that the wide view is to be preferred. The narrow view treats the defence as “the same as estoppel minus the representation” (so that detrimental reliance is still a necessary ingredient). The wide view looks to a change of position, causally linked to the mistaken receipt, which makes it inequitable for the recipient to be required to make restitution. In many cases either test produces the same result, but the wide view extends protection to (for example) an innocent recipient of a payment which is later stolen from him …

31.

In this court [Counsel for Scottish Equitable] did not argue against the correctness of the wide view, provided that the need for a sufficient causal link is clearly recognised. The fact that the recipient may have suffered some misfortune (such as a breakdown in his health, or the loss of his job) is not a defence unless the misfortune is causally linked (at least on a “but for” test) with the mistaken receipt. In my view [Counsel] was right to make that concession …”

On behalf of the claimants, Mr Beazley argued that where it is the government which seeks to rely on the defence, the court should adopt a stricter approach to the evidence than it would in an ordinary private law claim against an individual or corporate defendant. In support of this argument, Mr Beazley relied on the policy considerations which underlie the Woolwich remedy, and on the alleged difficulty of applying the usual tests of causation to a subject as opaque and impenetrable as the operation of government finances. The implication of my preliminary views on the facts in FII (High Court) I, he submitted, was that the defence would be virtually bound to succeed whenever the Revenue chose to invoke it in response to a claim to recover unlawfully levied tax. That could not be right, both as a matter of principle (there should normally be no obstacle to the recovery of mistakenly paid tax) and in the light of judicial observations such as that of Lord Walker in DMG at [145] that the defence “could seldom, if ever, apply to wrongfully exacted tax”: see [333] above.

As a specific example of the stricter approach for which he contended, Mr Beazley submitted that in assessing whether the government had changed its position as a result of receiving the overpaid tax, the court should adopt a test of detrimental reliance rather than the usual “but for” test. He also briefly touched on, but did not press, a more radical submission that, by analogy with the rules of tracing property in equity, the government should be treated as spending first tax and other money to which it was entitled, and as retaining payments of tax to which it was not entitled for as long as possible: compare In Re Hallett’s Estate (1880) 13 Ch D 696 (CA) at 709 to 711 and Foskett v McKeown [2001] 1 AC 102 (HL) at 129 to 133 per Lord Millett.

I am unable to accept these submissions. Once it has been held that the defence is in principle available to the Revenue, I do not think it would help the principled development of the law to hold that the government has a heavier burden than an ordinary defendant in seeking to establish it. The court will of course scrutinise the evidence with care, and will not make automatic assumptions that the question whether the government is left “worse off” can be answered in the same way, or by reference to the same criteria, as would be appropriate in a suit between private litigants. But the basic nature of the test should in my judgment be the same. Thus I reject the argument that detrimental reliance should be substituted for a “but for” test, or that the test needs to be satisfied to a heightened standard of proof.

Still less can I accept Mr Beazley’s tracing analogy, which seems to me to confuse two different legal categories, and to apply proprietary principles to a non-proprietary claim. As Lord Millett was at pains to point out in Foskett v McKeown at 129D-G, a claim to vindicate property rights is entirely different from a claim to reverse unjust enrichment. The two causes of action “have different requirements and may attract different defences”. In particular, a claimant who brings an action in unjust enrichment must show that the defendant has been enriched at his expense; but he is not concerned to show that the defendant is in receipt of property belonging beneficially to him, or its traceable proceeds. Conversely, a claimant who brings an action to recover misappropriated trust property “must show that the defendant is in receipt of property which belongs beneficially to him or its traceable proceeds, but he need not show that the defendant has been enriched by its receipt”. As to defences, a claim in unjust enrichment is normally subject to a change of position defence, whereas a claim to trace is subject to the defence of bona fide purchase for value.

I do, however, agree with Mr Beazley to this extent, namely that my preliminary views on the factual elements of the defence at the first trial were probably premature, and should not be read as implying that the defence is likely to succeed whenever the Revenue choose to invoke it. I therefore approach the question with an open mind, and with the benefit of the much fuller evidence and submissions now before me.

The basic question which I have to consider is whether I am satisfied, on the balance of probabilities, that the Revenue have changed their position as a consequence of the payments of unlawful tax by the BAT test claimants in such a way that they would be worse off by making restitution than if the overpayments had never been received. If that question is answered in the affirmative, the defence will succeed to the extent that the Revenue would be worse off, but no further.

It is convenient at this stage to clarify some further preliminary points. The first is at least in part one of nomenclature. In considering the defence, it is very easy to refer interchangeably to the Revenue (or HMRC), the Treasury, the government, or even the state. This looseness of language is natural enough, because the Commissioners for Her Majesty’s Revenue and Customs (to give them their full title) are a non-ministerial department closely linked to HM Treasury, which is itself an executive ministry of the government, which under our unwritten constitution is a component part, or arm, of the state. The use of these differing terms also reflects the fact that tax revenues are not hypothecated for particular purposes, or paid into a separate account. Like most other public revenues, they are paid into the Consolidated Fund. It is important, however, not to lose sight of the fact that the defendants to the present claims, who have to make good the defence, are the Revenue, which is part of the executive, not part of the legislature. Government has to be financed, and it raises the money which it needs through a combination of taxation and public borrowing. The main focus of the change of position defence therefore has to be on the part played by the overpayments of tax in the government’s finances, and in particular on the nature of the relationship between the overpayments and government expenditure.

The second point leads on from the first. At various times in their submissions, counsel for the claimants suggested that the concept of being “worse off” cannot sensibly be applied to the Revenue, because all government expenditure is intended to promote the public interest and is thus incurred for the public benefit. How then, it is asked, can the government be regarded as impoverished by the expenditure which it undertakes? The answer to this submission, in my judgment, is that it confuses the object or purpose of government expenditure with its cost. All government expenditure has to be funded, and if the government incurs expenditure which it would not have incurred but for the receipts of overpaid tax, and which does not confer a direct financial benefit on the government, I consider that the government is at least prima facie relevantly worse off. Mr Ewart drew the apt analogy of a charity, which would not be prevented from relying on the defence merely because the extraordinary expenditure in question was undertaken in furtherance of its charitable objects.

The third point is that it is not enough for the defendant merely to show that the money in question has been spent. The expenditure must be “extraordinary”, in the sense that it would not have been incurred but for the overpayment. The expenditure need not, however, be extraordinary in the sense of being of an unusual nature, either intrinsically or for the particular defendant. Thus increased expenditure of a routine nature can qualify, provided that the causative test is satisfied.

Fourth, I agree with Professor Burrows (Restatement, p 118, quoted above) that the defence is essentially concerned with “disenrichment”. In answering the question whether the defendant has been disenriched, it may be relevant to consider whether the expenditure or loss relied upon is reversible, and (if so) how easily the defendant could take steps to reverse it (ibid, p 119). But it would be wrong to elevate this consideration into a general test of irretrievability. Expenditure may well be irretrievable, for example because it is immediately consumed, or for some other reason cannot be recouped from the payee, but that fact alone does not stamp the expenditure as a relevant disenrichment. Among other things, it also has to satisfy the causal “but for” test if the defence is to be made out.

I emphasise this point because the Revenue’s expert evidence on government finances, and much of their submissions, focuses on irretrievability as a separate question. One of the specific questions which Sir Jonathan Stephens was asked to consider in his instructions was whether the sums spent by government between 1973 and 1997 were “irretrievable”, and section C of his first report (paragraphs 75 to 106) was devoted to this question. In the summary of his conclusions in paragraph 106, he described the question of how much of the enrichment was irretrievable as being “key to the question of change of position”. When cross-examined, Sir Jonathan confirmed that the concept of irretrievability was not one which he had ever encountered or discussed in the Treasury. He explained his understanding of the concept as being whether there was an ongoing benefit to the government which it could realise, measured by conventional accounting and economic definitions. Sir Jonathan was, of course, well aware that much government expenditure (for example on health or education) is intended to produce future social benefits, including benefits of an economic nature to both government and citizens alike. His point was that no reliable measures had yet been developed to capture and quantify what he called the “second and third round effects” of government expenditure.

Fifth, given the very long period covered by the claims, and the absence of hypothecation, the question inevitably has to be considered at a fairly high level of generality, and by reference to aggregated cash flows rather than the tracing of individual receipts. I do not consider that this necessity invalidates the defence, and I find some encouragement for adopting a global approach (although on a far smaller scale) in Lipkin Gorman itself, where the House of Lords looked at the aggregate result of all the payments passing between the fraudulent solicitor Cass and the casino, rather than the individual payments, in considering whether the defence was made out.

Finally, despite some occasional suggestions to the contrary by Mr Beazley and Mr Aaronson, I am satisfied that I should approach the question on the basis that the Revenue acted throughout in good faith. If the test claimants wished to rely on any allegation of bad faith as an answer to the defence of change of position, it would in my judgment have been necessary for them to plead the point distinctly and establish it by evidence. Furthermore, in the comparable context of the claimants’ Factortame claim for damages, it was not alleged that either the Revenue or the government had acted in bad faith: see FII (High Court) I at [401].

The scale of the relevant overpayments was approximately as follows. The total amount of the claims for overpaid Case V corporation tax, from 1973 to 1999, is just under £700,000, or an average of about £30,000 per annum. The claims for surplus (i.e. unutilised) ACT over the same period total about £95.3 million on the CT61 method which I have held to be appropriate, or an annual average of about £3.7 million. The claims for utilised ACT, again on the CT61 basis, total about £228.2 million, or an annual average of about £8.8 million. It follows that the amount of overpaid tax in any one year, averaged over the whole period to 1999, was of the order of £12.5 million.

It is obvious that overpayments of this order of magnitude represented only a minuscule proportion of total government receipts. According to the claimants’ expert witness on government finance, Dr Andrew Sentance, the average overpayment represented approximately 0.005% (or a two-hundredth of 1%) of government receipts, equivalent to approximately 2 pence per week when compared with the average household income of £23,200 in 2011/12. Sir Jonathan Stephens took issue with some of the assumptions made by Dr Sentance in performing this calculation, but even on his preferred basis (taking absolute average rather than arithmetical average figures, to reflect the fact that in some years there were net repayments by the Revenue to the claimants) the payments amount to only 0.011% of total receipts, equivalent to 5 pence per week of a household income of £430 a week.

The expert witnesses who gave evidence on government finances at the resumed trial were (for the Revenue) Sir Jonathan Stephens and Professor Gareth Myles, and (for the claimants) Dr Sentance. No separate evidence of fact was adduced on either side.

I have already referred to Sir Jonathan’s relevant experience and expertise in the context of what he had to say about the linkage between ACT and tax credits: see [283] to [287] above. The criticisms which I there made do not extend to the rest of his evidence, which fell within his area of real expertise and his extensive experience, at the highest levels, of the practical and policy aspects of government finance, particularly between 2001 and 2006. To the extent that much of his evidence is evidence of fact rather than expert evidence in the strict sense, the Revenue also seek to rely on it. I can see no reasonable objection to their doing so, even though the date for exchange of expert evidence followed the deadline set for exchange of witness statements of fact. An expert witness must be able to draw on his relevant factual experience, and it would serve no useful purpose to insist on such factual evidence being taken out from the body of his report and served separately in advance.

Professor Myles is an economist who has specialised in public economics and taxation. Since 1996 he has been professor of economics at Exeter University and since 1997 he has been a research fellow at the Institute of Fiscal Studies (“the IFS”). He has published widely on taxation and public expenditure. He is currently director of the Tax Administration Research Centre, an independent body which obtains grant finance from the government and is run, as I understand it, under the joint auspices of the University of Exeter and the IFS. For present purposes, the most relevant parts of Professor Myles’ evidence are those in which he endorses the approach of Sir Jonathan Stephens to change of position and calculates the quantum of the enrichment on that basis. It is worth noting, in this context, that the concept of irretrievable expenditure was as unfamiliar to Professor Myles as it was to Sir Jonathan: he said that he had never heard of the expression outside the present case.

Dr Sentance is currently the senior economic adviser at PwC, and a part-time professor at Warwick Business School. He is a business economist, with wide experience of analysing and advising on issues relating to economic policy and public finances. From 2006 to 2011, he was an external member of the Bank of England Monetary Policy Committee, and before then he held senior positions at the Confederation of British Industry, the London Business School and British Airways. During his time at the CBI, he was a founder member of HM Treasury’s Panel of Independent Forecasters, sometimes known as the “Seven Wise Men”. He too has published widely, mainly on macroeconomic policy issues and aspects of public finance.

Dr Sentance confirmed in cross-examination that he had never been involved internally in the workings of government, and in that sense his expertise was that of an outsider without direct personal experience of the way ministers take decisions. He said, however, and I accept, that his career had involved “an awful lot of interaction … with civil servants” who had been in such a position. He added (day 9, page 89):

“I have been observing and commenting on it and reading material about that process.”

It seems to me that, with his wealth of relevant experience, Dr Sentance is well qualified to comment from an expert and independent viewpoint on the issues which I have to consider.

In their joint statement dated 11 March 2014, the three experts record that they are agreed on the following issues:

“(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.

(c)

In general terms, the overpayments could have led to:

(i)

reduced borrowing;

(ii)

increased spending;

(iii)

discretionary tax reductions; or

(iv)

some combination of (i), (ii) and (iii).

(d)

There are a wide range of factors which the Government takes into account in setting its borrowing, tax and spending plans.

(e)

In the long-run, tax receipts and spending are related to each other.

(f)

In the short-run, tax receipts, spending and borrowing can all fluctuate significantly and there are often variations from forecasts and plans.

(g)

In response to large external shocks, the Government relies on borrowing to take some of the strain in the short-term, with spending and/or taxes adjusting over time.

(h)

Borrowing and spending can each be adjusted by smaller amounts in the short-term.

(i)

The overpayments by BAT were, in any one year, very small in relation to total Government revenues or spending, but were of a size that is equivalent to or larger than the bulk of payments that the Government receives from companies and individuals.

(j)

The interest rate proposed by Professor Myles, based on 10-year bonds, is a reasonable measure of the cost of UK Government borrowing over the period since 1973/74.”

Against this background, the thesis developed by Sir Jonathan Stephens, and endorsed by Professor Myles, was in essence as follows. The analysis begins by considering the materiality of the ACT overpayments. Although minute as a proportion of total government receipts or expenditure, the overpayments were individually of a similar size to the majority of government taxation receipts, and as such were fully factored into the government’s spending plans, which are prepared and submitted to Parliament on a “very granular level”, with estimates rounded to the nearest £1,000 through most of the period under review (and to the nearest £1 in the earliest years). For the purposes of submissions on budgeting to ministers, for review and the taking of policy decisions, amounts were rounded to the nearest £5 million, and would only be recorded as negligible where the cost or yield in the relevant year was below £3 million. Thus relatively small amounts were accounted for and taken into consideration by ministers when making budget decisions.

The next step in the analysis asserts that throughout the relevant period there was, in general, a “strong relationship” between government spending and receipts, although there was also a tendency for spending to exceed tax revenues. On only two occasions was there a notable divergence between spending and receipts: first, in the mid-1970s, when the UK sought a loan from the International Monetary Fund and had to make significant reductions in public spending; secondly, in the early 1990s, when the government again “had to take measures to put the public finances on a sound footing”.

Sir Jonathan then considers the extent to which the government was enriched by the overpayments of ACT, and how far such enrichment was “irretrievable”, that being in his opinion the key to the question of change of position. He starts from the proposition that, over the relevant period, fiscal policy (i.e. the question whether the government should run a surplus or a deficit, known as the “Budget judgment”) had primacy over tax and spending policy. The Budget judgment would be based on a number of macro-economic considerations and forecasts, into which the overpayments of ACT would have been factored. The primacy accorded to the Budget judgment meant that “the relative levels of taxation and spending simply follow”. Accordingly, the overpayments of ACT “would have been included in plans for either higher public expenditure or lower taxation than otherwise would have been the case” (paragraph 84 of Sir Jonathan’s first report).

Sir Jonathan then considers, and rejects, a possible objection to this conclusion. Since the actual levels of government borrowing usually differed in the event from the planned levels, might this have led in practice to the overpayments of tax being used to reduce borrowing rather than to fund additional spending or reductions in taxation? To show that this was the case, he says, the deficits actually incurred would need to have been substantially lower than those which were forecast. That is not so, however, and over the relevant period the forecasts showed a distinct “deficit bias”, i.e. the actual outcomes were normally worse than forecast. Over the 26 years from 1973 to 1999, borrowing in the year following the forecast was higher than forecast in seventeen years, and below forecast in only nine years, with an average error of £2.7 billion.

Sir Jonathan summarises his conclusions at this point as follows:

“88.

… The answer therefore on what happened to the payments of ACT could be characterised as that it was all used elsewhere – to fund spending or reduce other taxes – to meet the Government’s objectives.

89.

My experience in Government over a number of years is that this is a realistic description of the situation. Ministers look to use the resources available to them to the maximum degree possible to meet their objectives. Although tax changes are generally (not always) only made at the time of the Budget, spending decisions and allocations are regularly changed several times within the year, in response to changing forecasts and events. Having approved the initial estimates, Parliament’s approval is regularly sought for Supplementary Estimates later in the year. The last such Supplementary is submitted in March, just one month before the Government’s financial year ends, allowing adjustments to be made to spending decisions in the final month.”

Next, Sir Jonathan considers how much of this expenditure should be regarded as irretrievable. In principle, he says, an addition to government resources could be used to increase “current” public expenditure, increase expenditure on capital items, reduce taxation, or lower borrowing. He defines “current” expenditure, by reference to a recent (2013) Treasury document, as “spending on items that are consumed in the process of providing public services”, including for example public sector wages and salaries, and the purchase of goods or services. He regards such expenditure as wholly irretrievable. By contrast, expenditure on capital items is arguably recoverable, in the sense that the asset in question will often have a realisable value, subject to depreciation. Sir Jonathan is prepared to treat all government capital expenditure in this way, even though he recognises that many assets in which government invests, such as transport infrastructure, may not have a realisable market value. He also accepts that the repayment of government debt will lead to a saving of future interest. He regards tax reductions as equivalent to increases in current expenditure, and therefore disregards them in the interests of simplicity. He therefore considers that there are two ways in which the enrichment derived from the overpayments of ACT would be recoverable: first, if the money was used to acquire capital assets, and secondly, if it was used to reduce government debt.

The next question is how much of the overpayments should be attributed to those two purposes, it being impossible to conduct any kind of tracing exercise. At this point, Sir Jonathan makes two key assumptions. First, he assumes that the overpayments should be split between capital and current spending in the same proportions as overall government expenditure. He describes this approach as “using the average of how receipts were put to use to estimate the marginal use of extra receipts”. Secondly, he assumes that a proportion of resources was also used to reduce government borrowing, even though in most years the government was a net borrower and “there was little net lending by government to observe”. He estimates this proportion as equal to the proportion of total borrowing to the total resources at the government’s disposal each year.

Sir Jonathan then concludes (paragraph 105 of his first report):

“This approach is arguably generous, especially given my experience that Ministers in practice use the resources available to them to the fullest extent possible. But this approach, in which the initial payments of ACT are allocated between current spending, capital expenditure and reducing debt in proportion to the overall use of resources, is relatively simple. In the absence of any means of tracing whether the specific payments in question were actually put to extra spending, reducing taxes or reducing borrowing, it is, in my opinion, a fair approximation. I consider it is more likely to be closer to reflecting the actual use than either of the more extreme assumptions that all the payments were put exclusively to extra spending or exclusively to reducing borrowing. It also allows us to address the issue of change of position. Therefore it is the approach that I am proposing.”

Professor Myles, in his first report, adopts Sir Jonathan’s methodology and uses it to construct a model for quantification of the claim. In particular, he explains that:

On this last point, Professor Myles says (paragraph 81 of his first report):

“This measure indicates the relative weight that the Government placed on the competing pressures to spend, [reduce] tax revenue and reduce borrowing. The relative amount of borrowing indicates the way in which the Government has chosen to resolve this tension at an aggregate level. And the larger the deficit, the more priority the Government would likely have given to avoiding still higher borrowing.”

The application of Professor Myles’ model to the amounts of tax allegedly overpaid by the test claimants, as then computed, yielded a total quantum for the claims of £34,190,073. I emphasise that this figure assumed in the Revenue’s favour that Sir Jonathan was also right on the question, which I have yet to consider, whether the time value of the claims should be computed by reference to his assessment of the actual benefit derived by the government from the overpayments rather than by an award of compound interest.

Fuller details of Professor Myles’ model are contained in annex 4 to his report. The most important point to note is that he allocates what he terms the net payments of ACT (i.e. the overpayments actually made, minus repayments of ACT and utilised ACT) to the three purposes of current spending, capital spending and deficit reduction in the same financial year as that in which the overpayments were made. Thus in the financial year 1993/94, which he takes as a worked example, he proceeds as follows:

In his single expert report dated 17 January 2014, Dr Sentance discussed and took issue with many of the key assumptions which underlie the approach of Sir Jonathan Stephens and Professor Myles. The second section of his report, headed “Summary and conclusions”, reads as follows:

“14.

In my evidence set out below, I have considered the approach to public finances in the UK over the period since 1973/74 and the statistical relationship between various measures of public spending and taxation. My main conclusions are as follows.

(1)

The level of public spending in the UK is influenced by a wide range of factors, including the health of the economy, future economic growth prospects, the political priorities of the government, and social and demographic trends, as well as the state of public finances.

(2)

Since 1973/74, the UK has been able to borrow frequently on the bond markets – maintaining its AAA sovereign borrowing rating until very recently …

(3)

As a result of its strong credit rating, UK borrowing and public debt levels have been able to fluctuate considerably over the period since 1973/74 without seriously impacting its borrowing status. The key issue influencing the response of financial markets to changes in UK public finances over this period has been broader confidence in economic policy and the medium term financial framework. While this broader confidence has been maintained, significant short-term movements in levels of public sector deficits and debt have been accommodated.

(4)

While there is clearly a long-term relationship between government revenues and spending, this operates in both directions – with the level of spending affecting the need to raise revenue as well as vice versa. When we have seen big disturbances in tax revenues due to changes in economic conditions – as in the mid-70s, early-90s and during the recent financial crisis – on average it has taken around 10 years for the full adjustment to take place. In the meantime, it is public borrowing which has taken the strain.

(5)

Statistical analysis shows that, over much shorter time-horizons, changes in tax revenues have a very limited impact on actual levels of government expenditure or spending plans. In periods from one to five years, the statistical relationship between changes in public spending and taxation is very weak. This view is also backed up by an analysis of how spending plans adjusted to changing revenue forecasts in the 1980s and 1990s.

(6)

The amounts of tax overpayments in this case do not appear to be material in terms of the planning of public finances in total. The overpayments were not sufficiently large to change the reported profile of tax receipts, and - to the government – represented on average a net receipt of 0.005% of tax receipts between 1973/74 and 2000/01 …

15.

This analysis informs my assessment of the evidence provided by Sir Jonathan Stephens and Professor Gareth Myles. An important element of their evidence is that the UK government spent the bulk of the tax revenues in dispute – within one or two years of the revenue arising. As a result, they argue that there is little benefit to the government remaining from the tax overpayments, supporting this with a detailed model of public finances. However, their approach and analysis rests on many arbitrary and questionable assumptions and does not represent a robust methodology for evaluating this claim.

(1)

Their methodology implies that changes in government revenue would be reflected in immediate variations in government spending, which is not consistent with the way in which public finances were managed in practice over the relevant period.

(2)

Their analysis does not recognise the significant flexibility that the UK government has exercised over the relevant period in allowing public borrowing and debt levels to “take the strain” in managing public finances.

(3)

They do not recognise any continuing benefit to the economy and to the government from current public spending. Money allocated to current spending is deemed to be “irretrievably spent” and only capital spending creates a lasting benefit for government. Yet there are significant benefits to the economy and society from current public spending which includes any elements which add to the productive capacity of the economy in the same way as capital spending – for example through its effect on the health, wellbeing and skill levels of the workforce.

(4)

Their treatment of debt interest is abnormal and counter-intuitive in terms of conventional economic thinking. Instead of the normal convention of compound interest, where the value of debt interest saved would accumulate over time, debt interest saved is assumed to boost current public spending – which is then assumed to be “irretrievably spent” by the government – and hence disappears from their modelled calculations. In my opinion, this is not a sound basis for evaluating the impact on government finances of past overpayments of tax.”

Dr Sentance’s report is too long to quote from extensively, but in general I found it thoughtful and well-argued, providing cogent support for the summary and conclusions set out above. Mr Ewart cross-examined Dr Sentance for little over an hour, without undermining his evidence to any significant extent. I will therefore content myself with a few points culled from Dr Sentance’s evidence, while stressing that it needs to be read and evaluated as a whole.

In the section of his report devoted to the management of UK public finances, Dr Sentance convincingly establishes that “public expenditure is planned on the basis of medium-term and long-term views about the state of the economy and the need to provide public services and cash transfers” (paragraph 30). He points out that the timetable for setting public expenditure is considerably ahead of the key decisions on taxation, which are normally set out in the March Budget, a few weeks before the start of the new fiscal year. Since the early 1980s, public expenditure plans for the year ahead have normally been announced in outline in the Autumn Statement, and have been the subject of a public expenditure White Paper published in December. This timetable suggests that it is public spending plans which drive the need to raise revenue through taxation, rather than the other way round, at least in the short term. Public spending plans are of course drawn up with reference to the resources available to finance them, but “the link between revenue raised and the spending that it supports is flexible and indirect” (paragraph 32). The only variation in this timetable for tax and spending decisions occurred in the mid-1990s, when the then Chancellor of the Exchequer, Kenneth Clarke, set out a unified budget in November which covered both spending and tax measures. This approach continued until 1996, but after the election of a Labour government in May 1997 the budget returned to its normal timing in the Spring.

More generally, Dr Sentance considers that the impact of economic factors on public spending plans is far more complex and nuanced than the Revenue’s model would allow. In his view the government can ride out short-term fluctuations in revenue by varying borrowing, as long as it maintains a credible economic policy and a sound medium-term financial framework. The management of public finances in the UK is a highly complex issue, in which political and economic factors combine and the government needs to focus on both short-term management and maintaining a sound medium-term position.

In the next section of his report, dealing with the relationship between revenue and expenditure, Dr Sentance tests the thesis that government spending would have been lower if the overpayments of tax by the claimants had not been received. He does this in two different ways. First, he examines the government’s own estimates of revenue and spending plans as presented in the annual Budget “Red Books”. These contain figures for at least three years ahead, so it is possible to assess how changes in estimates of revenue feed through into higher or lower expenditure. The result of this examination is that “very little relationship” can be found between the revenue discrepancies and changes in spending totals, particularly in the short term. Secondly, Dr Sentance performs a statistical analysis of the correlation between changes in real government revenues and real spending, using data compiled by the Office of Budget Responsibility going back to the early 1970s. The result of this analysis is that the statistical association is “very weak”, with nothing to suggest that changes in government spending in real terms are related to the variation in government receipts over a period of one, three or five years.

On the question of the materiality of the overpayments, Dr Sentance comments that there is “no unique metric that can be used to assess materiality”, but in the context of the present case he would view the tax overpayments as material if it was clear that “they would have been significant enough to change the government’s overall spending plans through the sustained impact that they had on overall public finances”. He then gives a number of reasons for concluding that the payments were not material in that sense. In short, the reasons are: the extreme volatility of ACT and MCT receipts; the very small size of the overpayments in comparison with total government revenues; and the fact that none of the annual overpayments was large enough to change the public presentation of government finances. In only one year (1994/95) did the overpayment of surplus (unutilised) ACT by the test claimants exceed £100 million, which is the most detailed level of reporting of projections of public spending and revenue in the Red Books for the relevant period.

In his discussion of the setting of UK fiscal policy, Dr Sentance disagrees with Sir Jonathan Stephens that it was primarily driven by the need to achieve deficit reduction. While the need to contain public borrowing was an important objective of policy from the late 1970s until the 1990s, Dr Sentance considers that it had a “variable significance” over this period, and was certainly not the sole driving force as Sir Jonathan contends. In this context, Dr Sentance emphasises the distinction between the need to maintain sustainable borrowing levels over the medium to long term, and how policy is set in the short term. He also points out that, if deficit reduction was indeed the sole objective in both the short and the long term, “it is hard to explain why government spending has risen consistently in cash terms in every year since the early 1970s and rose in real terms in most years”. He adds that it would also be hard to account for “the considerable fluctuations in borrowing through this period”.

According to Dr Sentance, Sir Jonathan’s evidence is consistent with a “Treasury view” of the management of public finances which was promoted by Treasury ministers and officials, particularly in the 1980s, but even then was questioned as a realistic description of how policy operated in practice. Dr Sentance cites a study of the operation of the Treasury and its interaction with other Whitehall departments (Thain and Wright, The Treasury and Whitehall: The Planning and Control of Public Expenditure, 1976-1993, OUP 1995) which describes how in the 1980s the Treasury Select Committee of the House of Commons were sceptical of the “Treasury view” that revenue determined expenditure.

On this last point, I think that Dr Sentance may have under-estimated the potency of the “Treasury view” in the formation of government fiscal policy. As Sir Jonathan said in his report in reply to Dr Sentance’s report, drawing on his own practical experience in Whitehall, Treasury ministers and officials are best placed to describe what actually affects the decision making process, and it is hardly surprising that there is a “Treasury view” informed by the practical experience of actually managing the public finances. Nevertheless, Sir Jonathan significantly modified his initial position, as it seems to me, by expressly accepting 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 (paragraphs 35 to 38 of his reply report). This change of position is reflected in paragraphs 1(d) to (f) of the experts’ joint statement, quoted in paragraph [365] above.

Mr Aaronson probed this aspect of Sir Jonathan’s evidence in cross-examination. As a result, it soon became clear that Sir Jonathan accepted that borrowing would normally take the strain in the short-term as a response to an unexpected reduction in taxation revenue. Sir Jonathan said that, when there were shortfalls of revenue, ministers were not always happy to see it absorbed by increased borrowing, and sometimes gave instructions for spending to be adjusted accordingly. However, the attitude of ministers would depend on all the circumstances, and the default position was that borrowing would always take at least some of the strain.

This can be seen most clearly from two passages in Sir Jonathan’s cross-examination. In the first (day 8, pp 69-72) he accepted that “borrowing does take the strain” in response to unexpected changes in taxation receipts or government expenditure, and clarified that by “unexpected changes” he meant no more than changes from the forecast levels. In the second passage (day 8, pp 77-80) he accepted that the “control totals” for government expenditure were in practice by no means immutable, and when ministers took actual decisions during the year in question their attitude to the relationship between available revenue and expenditure would depend on the circumstances. The questioning continued:

“Q. And when [ministers] were neither relaxed nor unrelaxed, borrowing would be the shock absorber for the time being?

A. In the short term, borrowing always – as I think I recognise in my statement – takes some of the strain.

Q. And how much of the strain, it all depends, does it? Is that what you are really saying?

A. Yes; yes.

Q. And just to make sure I understand the correlation between tax receipts and expenditure, you are talking about the correlation between tax receipts and total managed expenditure, rather than current government expenditure. Is that correct?

A.

Yes.”

Sir Jonathan then confirmed that “total managed expenditure” includes both capital expenditure and debt interest, as well as current expenditure.

A little earlier in his cross-examination, Mr Aaronson explored with Sir Jonathan the timetable and criteria by reference to which government fiscal policy was typically determined in the period under review. I think it is fair to say that Sir Jonathan agreed in general with the description of the timetable given by Dr Sentance, although the precise machinery varied from time to time. In very general terms, there were typically three stages. First, a planning or control total for expenditure in the ensuing fiscal year would be set by the Chief Secretary of the Treasury, after discussions with the Chancellor, and would be communicated to departmental ministers at a cabinet meeting, usually in July. This control total set the upper limit for total public spending, out of which allocations had to be made to the spending departments. The second stage, therefore, was the process of allocation to the various spending ministries, which would follow a period of often intensive negotiations, typically between July and November. The results of this procedure would be incorporated in the Chancellor’s Autumn Statement. This Statement (or its equivalent in years when the timetable was different) would set out the spending plans for the next fiscal year, and sometimes for ensuing years as well. The third stage in the process was the Chancellor’s annual budget speech, usually in March, and the associated fiscal and other measures announced at the same time.

Although the control total was, in Sir Jonathan’s words, “deliberately intended to set a rigid envelope within which ministers had to allocate their spending” (day 8, p 63), he accepted that in practice there were variations from it. He also agreed that there was a separate contingency reserve, built into the control total, which was not initially allocated to ministries, and access to which was strictly controlled. The amount of the contingency reserve was substantial, and was typically calculated on a rolling three year basis, with larger amounts reserved for the more distant years. Thus, to quote some figures mentioned by way of example in the evidence, the contingency reserve for 1993 was in excess of £4.5 billion, increasing to £7 billion for 1994 and £10 billion for 1995.

A further point to emerge from cross-examination concerned Sir Jonathan’s rather imprecise references in his reports to “net” amounts of ACT. Since he readily accepted that he was not an expert in the detailed workings of the ACT system, I doubt whether he fully appreciated that the bulk of the overpayments of ACT were payments of ACT which was subsequently utilised. In such cases, the overpayments conferred a cash flow advantage on the government, but the receipts in question were subsequently matched with MCT, sometimes in the same fiscal year but often considerably later. When asked how receipts of this nature would enter into budgetary planning, it was plain that Sir Jonathan had no first hand knowledge to draw upon, but he said he thought it unlikely that they would be taken into account (day 8, pp 48-49). In answer to a long and rather leading question in re-examination, however, Sir Jonathan said he thought that the receipts would in fact have been taken into account, if (as was usually the case) the set off against MCT did not occur in the same fiscal year, because government accounts and budgeting are done on an annualised basis. He did not explain, however, how utilised ACT was in fact dealt with in the context of government forecasts, and I infer that his answer was more of an educated guess about what he thought ought to have happened than reliable evidence about what actually did happen. I conclude that I am left with no credible evidence about how the cash flow aspect of ACT was taken into account in government forecasts of receipts and expenditure.

Finally, I should comment on some evidence of statistical calculations performed by Professor Myles given in his report in reply to the evidence of Dr Sentance. The calculations were designed to show whether the long-term relationship of equilibrium between government receipts and expenditure was statistically significant, and to apply a standard error-correction mechanism in order to capture the rate of re-adjustment towards the equilibrium after any deviation from it. In broad summary, and in layman’s language, my understanding of Professor Myles’ calculations is that government receipts and expenditure are indeed linked in a statistically significant long-term relationship of equilibrium, and short-term deviations from it will tend to be corrected in a manner that is again statistically significant. The statistical link between the two variables is stronger when total managed expenditure is compared with government receipts than when the comparison is made with current expenditure. It is fortunately unnecessary for me to describe the calculations in any greater detail, because Dr Sentance accepted in cross-examination that they were statistically valid. On the other hand, he pointed out, and I would accept, that the relationship thus demonstrated is a comparatively weak one, particularly where the comparison is with current expenditure. In other words, the process of adjustment back towards equilibrium is quite a slow one, and even in the case of total managed expenditure one would not expect more than about two thirds of a deviation to be eliminated over five years. Dr Sentance also stressed that equations of the type employed by Professor Myles in themselves tell one nothing about causation, although it may be possible to infer some degree of causation, depending on the circumstances.

In the light of all this evidence, I can now state my conclusions quite shortly. The methodology propounded by Sir Jonathan in his first report, and quantified by Professor Myles, 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 existence of this fixed link is in turn said to be the automatic consequence of the primacy accorded to the “Budget decision” on the appropriate level of deficit. As Sir Jonathan put it in paragraph 88 of his first report:

“In a world where decisions on the appropriate level of deficit take primacy, the relative levels of taxation and spending simply follow.”

Consistently with this approach, as we have seen, Professor Myles’ model assumes that the net overpayments of ACT would in fact have been reflected in actual government expenditure within the same 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 the Revenue 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. It is only after about five years that a statistically significant return towards equilibrium after short-term deviations can begin to be detected. Even Professor Myles asserted no more than that “there is a general trend towards a long-term relationship” (day 9, p 48).

The model espoused by the Revenue in their expert evidence also strikes me as implausibly rigid and unrealistic when regard is had to the detailed evidence about matters such as:

Furthermore, even if the Revenue’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, on the footing that it could not be known in advance when, or to what extent, they would be utilised by being set off against MCT. Had that been the case, it would follow that all the overpayments of ACT should be left out of account, whether or not they were subsequently utilised, leaving only the relatively trivial overpayments of corporation tax to which the defence of change of position could in principle apply. 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.

It is unnecessary for me to go any further. For the reasons which I have already given, I consider that the quantification model propounded by the Revenue is clearly unsound, both conceptually and on the facts. No alternative model has been put forward, so there is no alternative basis on which the defence could be made out. In these circumstances, I prefer to say no more about other objections raised by the claimants to the Revenue’s methodology, such as the nature of the distinction drawn between current and capital expenditure, the question whether account should be taken of indirect future benefits to the government from certain types of current expenditure, and the treatment of saved interest on government borrowing. These are all questions of some importance and difficulty, and their resolution should in my judgment be left to a case where it is clearly necessary to decide them. I will, however, need to return to the treatment of interest when I come on to some of the later Issues.

For now, it is enough to say that my answer to Issue 20 is that the Revenue have failed on the facts to make out a defence of change of position to the mistake-based claims, assuming such a defence to be open to them in principle under English law. The question whether the defence is in any event precluded as a matter of EU law is the subject of Issue 21, to which I now turn.

Issue 21: if the Revenue are entitled under English law to a change of position defence on the facts to the claimants’ mistake claims, are they precluded from relying upon it by EU law?

I will deal with this issue briefly, because I have already considered it, with the benefit of full argument from the same leading counsel as in the present case, in Portfolio Dividends (No. 2): see [171] to [189]. My conclusion was that the defence of change of position is precluded by EU law in relation to all San Giorgio claims, because the only defence recognised by EU law to such claims is the defence of unjust enrichment of the taxpayer. This principle had been consistently stated by the ECJ in its case law dating back (at least) to Weber’s Wine World in 2003 (Case C-147/01, [2003] ECR I-11365), and had recently been reaffirmed in uncompromising terms by the Grand Chamber in the Lady & Kid case (Case C-398/09, [2012] STC 854), judgment in which was delivered on 6 September 2011.

As I explained in Portfolio Dividends (No. 2) at [172] to [175], the judgment of the Grand Chamber in Lady & Kid is in my view of particular strength for a number of separate reasons. First, the relevant principle was stated more than once in terms which admit of no ambiguity: “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 European Union law” (paragraphs 20 and 25 of the judgment of the Court). Secondly, the decision was one of a Grand Chamber of 13 judges. Thirdly, the Grand Chamber was departing from the advice of its Advocate General (Cruz Villalón), who argued that “the fact that the tax has been passed on does not constitute the only possible means of refusing repayment” (paragraph 67 of his opinion). Fourth, the detailed factual background to the four claims in the national proceedings in Denmark would have provided an ideal opportunity for the enunciation of a more qualified principle, had the Court wished to do so.

Furthermore, the ECJ has also articulated the reason for its uncompromising stance on this issue. To admit of any further defences would place a limitation on “a subjective right derived from the EU legal order”, and the restriction must therefore be narrowly construed: see Lady & Kid at paragraph 20, and the decision of the First Chamber of the ECJ in Accor (Case C-310/09, [2012] STC 438), delivered a few days later on 15 September 2011, at paragraph 73. The relevant passages are quoted in Portfolio Dividends (No. 2) at [174] and [176] respectively. The meaning of “a subjective right” in this context must, I think, be a personal or private right, representing the French “droit subjectif”: see Littlewoods (No. 2) at [270].

In Portfolio Dividends (No. 2) I drew attention at [177] to the then very recent opinion of Advocate General Wathelet in FII (ECJ) III, where he too had repeated the relevant principle, describing it as “settled case-law”. At that stage the decision of the ECJ on the third FII reference was still awaited. The judgment of the Court was delivered on 12 December 2013. It makes no reference to the principle, 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”.

In the light of the decision of the ECJ on the third reference, Mr Ewart was constrained to accept that the test claimants’ mistake-based cause of action was protected by EU law at least to the extent that EU law prevented the UK from withdrawing it without notice and retroactively. In the circumstances, the Court held that the enactment of section 320 of the Finance Act 2004 had infringed the EU law principles of effectiveness, legal certainty and the protection of legitimate expectations: see paragraphs 38 to 43, and 44 to 49. Moreover, the Court expressly held that it made no difference that the taxpayer had a choice between two national causes of action to recover the tax levied in breach of EU law: see paragraph 38. The formal ruling of the Court at the end of the judgment is unambiguous on this point:

“1.

In a situation in which, under national law, taxpayers have a choice between two possible causes of action as regards the recovery of tax levied in breach of European Union law, one of which benefits from a longer limitation period, the principles of effectiveness, legal certainty and the protection of legitimate expectations preclude national legislation curtailing that limitation period without notice and retroactively.”

It is in my judgment abundantly clear from the whole of the judgment of the ECJ, as well as from the terms of its ruling, that it regarded the two remedies available under English common law (namely the Woolwich cause of action, and the remedy based on mistake of law which the Court called “the Kleinwort Benson cause of action”) as alternative remedies for the recovery of taxes levied in breach of EU law: see, in particular, paragraphs 2 to 6, 24 (setting out the questions referred by the Supreme Court), 25 (where the first question was reformulated), 39, 41, 43 and 46 to 49. The Court nowhere says, or implies, that one remedy takes precedence over the other, or that the protection afforded by EU law to the mistake-based remedy differs in any material respect from that given to the Woolwich remedy.

Nevertheless, Mr Ewart sought to maintain the submission that only a claim based on the Woolwich cause of action can properly be classified as a San Giorgio claim under EU law. The reasons for the supposed distinction, it was said, lie in the principle of national procedural autonomy, in the admitted fact that the Woolwich remedy by itself would provide an effective remedy for the recovery of unlawfully levied tax, and in the nature of the unjust factor (mistake) which underpins the mistake-based remedy. So put, the argument is a further variant of one which I considered and rejected in Portfolio Dividends (No. 2) at [180] to [184]. I rejected it then, without the benefit of the decision of the ECJ on the third reference. In the light of that decision, I can only say that I consider the argument to be plainly wrong. I feel no doubt that, 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.

My answer to Issue 21 is accordingly that, even if the Revenue were in principle entitled and able under English law to make good a defence of change of position to the mistake-based claims, EU law would preclude them from relying on the defence.

Actual benefit: introduction

The next group of Issues (22 to 24) arises from the Revenue’s argument that the restitution to which the test claimants are entitled in respect of their mistake-based claims is confined to the actual benefit derived by the government from the use of the overpayments of tax. The Issues are, in short: (a) whether the argument is available to the Revenue under the English law of unjust enrichment (Issue 22); (b) if so, whether it is made out on the facts (Issue 23); and (c) if it is made out under English law, whether it is nevertheless precluded by EU law (Issue 24).

In the present context, “actual benefit” is to be contrasted with a measure of the government’s unjust enrichment which, at least in relation to the claims for utilised ACT, is based on the cost to the government of borrowing equivalent amounts of money, and does not involve an investigation of what the government actually did with the overpaid ACT in the period between its receipt and its later repayment or set off against MCT. A measure of the latter type was adopted by the majority of the House of Lords in Sempra, where it led to an award of compound interest on the amounts of prematurely paid ACT at conventional rates reflecting the favourable terms on which the government can borrow money.

The potential extent to which the Revenue seek to rely on “actual benefit” was initially far from clear. It was not pleaded in the Revenue’s re-amended defence in February 2014, perhaps for the technical reason that the burden lies on the claimant to establish the extent of a defendant’s enrichment (although there will normally be an evidential burden on a defendant who wishes to displace the normal market-value measure of his enrichment). Be that as it may, the Revenue accepted in their skeleton argument for the resumed trial that the claimants’ Woolwich claims had to succeed in full, given the policy considerations which underpin that cause of action. They also accepted (ibid, paragraph 172(i)) that the argument could not be run in relation to the principal amounts of unlawfully levied tax (both corporation tax and ACT), presumably because the payment of money normally represents an incontrovertible benefit which has to be measured by reference to its face value: see [266] above. Thus the argument is reduced to one that can apply only to the claimants’ time-value claims, that is to say the claims in respect of utilised ACT (as in Sempra) and to the claims for “interest” on the principal sums of overpaid tax (or more accurately, the time value representing the use of those sums).

The Revenue do not dispute that the market value of money over a period of time is normally measured in the modern world by compound interest. I think they would also agree that market value is both the starting point, and the default measure, in quantifying the claimants’ time-value claims. But, they submit, it is always open to a defendant to show that the actual benefit to him from the use of the money was less than its market value, and in such a case the restitution which the defendant has to make to the claimant is reduced accordingly. As Professor Burrows put it in both his written and his oral submissions, the law of unjust enrichment is the law of unjust actual enrichment and not the law of unjust notional or hypothetical enrichment. Under English law, restitution for unjust enrichment is about reversing the enrichment of the defendant (at the claimant’s expense), and not about compensating the claimant for his loss. This must be correct as a matter of principle, submits Professor Burrows, because the law of unjust enrichment imposes strict liability subject to recognised defences such as change of position.

In the present case, the Revenue rely on the expert evidence of Sir Jonathan Stephens and Professor Myles, and the model of government expenditure prepared by Professor Myles, to argue that the only actual benefit derived by the government from the premature payments of utilised ACT, and from the continuing use of the other overpayments of tax, lies in the extent to which the payments were used either to fund capital expenditure or to reduce government debt. To the extent that the payments were used to fund current expenditure, or to reduce taxation, the money in question was irretrievably spent and the government derived no continuing benefit from it.

The argument is advanced as one about the quantification of the government’s enrichment, but it obviously bears a close similarity to the defence of change of position. Indeed, in cases of the present type it is really the defence of change of position by another name, as Vos J recognised in Littlewoods (No. 1) at [129], although subject to two potentially significant differences. First, the legal burden of proof lies on the claimants to establish the quantum of the Revenue’s enrichment under English law, whereas the burden lies on the Revenue to establish the defence of change of position. Secondly, the causal link required by the defence between the initial enrichment and the subsequent disenrichment does not have to be demonstrated. As Vos J put it, “the special requirement of showing that a particular expenditure would not have occurred but for the overpayment is abrogated”.

In practice, however, these distinctions are in most cases likely to be more theoretical then real. As to the burden of proof, an evidential burden will lie on the defendant to displace the normal market value measure of his enrichment. In relation to causation, the existence of a “but for” link may not be a formal requirement, but in order to show that he has derived no (or a reduced) actual benefit from his use of the money, the defendant will need to prove what in fact he did with the money, and how he benefited from the use to which he put it. The Revenue were therefore right, in my judgment, to submit that the answer to the question of quantification of the enrichment will normally be the same as, or co-extensive with, the answer to the question whether the defendant has changed his position. But if that is right, it suggests to me (in common with some academic commentators) that the preferable approach is to recognise that the argument finds its true home as part of the defence of change of position, and that the proper measure of the defendant’s enrichment in cases of the present type is the objective value of the use of the money: see Professor Elise Bant, The Change of Position Defence (Hart Publishing, 2009), pp 129-130, and Man Yip, “Use Value of Money” – the Defence of “Exhaustion of Benefits”, [2012] 20 Restitution Law Review, pp 99-111.

Since I have already rejected the Revenue’s defence of change of position on the facts, it will come as no surprise that I reach the same conclusion on the facts, for substantially the same reasons, in relation to their argument based on actual benefit. Furthermore, I have already examined the main legal aspects of the argument in considerable detail in Littlewoods (No. 2) at [348] to [375] and [413], which in turn need to be read with my analysis of the speeches in Sempra in Portfolio Dividends (No. 2) at [212] to [247]. I do not intend to traverse the same ground again in this judgment, so I will deal comparatively briefly with Issues 22 and 23.

Issue 22: is the Revenue’s “actual benefit” argument available to them in respect of the claimants’ mistake claims under the English law of unjust enrichment?

As I have explained, this is in fact a question about quantification of the claimants’ time-value mistake claims: see [410] above. The Revenue now accept that the claimants are in principle entitled to recover in full:

For each of three main reasons, I consider that this question must be answered in the negative.

First, the normal objective measure of the time value of money is compound interest. Accordingly, that is (unless displaced) the appropriate measure of the Revenue’s enrichment in respect of the “massive interest free loan” (to quote Lord Nicholls in Sempra at [102]) constituted by the overpayments of tax, regardless of the use to which the Revenue actually put the money. Unless earmarked for a particular purpose, or subject to a trust in the hands of the borrower, a loan of money may normally be spent or applied in any way the borrower chooses. The actual use to which he puts the money is irrelevant to his obligation to repay the loan with interest. Put another way, what the borrower has obtained is the opportunity to use the money, or otherwise turn it to account, throughout the term of the loan. The borrower has been enriched by that opportunity. In the commercial world, such an opportunity may only be obtained by borrowing money at compound interest. This point was made with pellucid clarity in Sempra by both Lord Nicholls and Lord Hope: see [102] and [33] respectively.

Secondly, it will normally accord with the principles of fairness and justice which lie at the heart of the English law of unjust enrichment to treat the recipient of money paid under a mistake (whether of fact or law) in the same way as a borrower of the money on commercial terms, particularly if the recipient has in fact put the money to use in the same way as he would use borrowed money. The fairness and justice of proceeding in this way seem to me irresistible if the recipient has (albeit innocently) demanded or requested the mistaken payment, and if he then resists the making of restitution when the mistake is drawn to his attention. That, in essence, is in my judgment the position of the Revenue in the present case. Why, then, should the government be in a better position when ordered to make restitution for the time value of the mistaken overpayments of tax, than it would be for the cost of borrowing equivalent sums of money on the market in the usual way? In either case, the use which the government actually made of the money between the dates of receipt and repayment is to my mind completely irrelevant.

Third, in a case of the present type, there can in my opinion be no room for application of the principle of subjective devaluation which the Supreme Court (by a majority) recognised in Benedetti. It could not seriously be contended that, in the hands of the government, the mistaken overpayments of tax were any less valuable to the government than the rest of the resources which it raised through taxation or public sector borrowing. The proposition that money somehow has less than its normal face value when it is spent by the government needs only to be stated for its absurdity to be apparent. On this part of the case, I have nothing to add to my discussion of Benedetti in Littlewoods (No. 2) at [362] to [374].

Issue 23: if it is available, is the “actual benefit” argument made out on the facts?

Even if it were open to the Revenue to run the “actual benefit” argument, I would reject it on the facts. In the first place, I can find no firm basis upon which to conclude that the actual benefit derived by the Revenue from the overpayments of tax was anything other than the opportunity to use and spend the money in the public interest. That opportunity has an objective time value, represented by compound interest from the date of receipt of the money until the date of its repayment.

The Revenue’s attempt to demonstrate how the money was actually spent is, on the evidence before me, doomed to failure. For the reasons which I have already given in relation to the defence of change of position, I am satisfied that no short term link can be established between the relevant receipts and government expenditure. The highly schematic model presented by Professor Myles is also fatally flawed in a number of respects. The most that can be said is that there was a long term correlation between government receipts and expenditure in the period under review, but even then the nature of any causal connection remains obscure.

I would add this. The practical difficulty of establishing a plausible model, and the arguable artificiality of applying concepts like financial benefit to government expenditure, as though the position of the government were comparable with that of an individual or company in receipt of a mistaken payment, all tend to suggest that a wrong turn has been taken if this is the road to a solution. In other words, the factual and conceptual difficulties associated with a search for the actual benefit obtained by the government from the use of the mistaken overpayments of tax provide in themselves a strong indication that, in cases of the present type, the normal objective measure of the use value of money should be adhered to.

In theory, I would accept that there might arguably be grounds for departing from the normal objective measure if the Revenue could prove that, in each relevant year, the use of the claimants’ money conferred no actual time-value benefit on the government because, in the absence of those receipts, the government would not have needed to borrow equivalent additional amounts. But a case of this nature could only get off the ground if government borrowing were rigidly fixed in advance, with the result that any unforeseen reduction in receipts would automatically have led to a reduction in expenditure rather than an increase in borrowing. In Littlewoods (No. 2), where I heard evidence from different expert witnesses which differed in a number of material respects from the evidence in the present case, I accepted that a rigid relationship of this sort had been made out: see in particular [395] to [396]. Whether or not I was right in so concluding, I emphasise that in the present case I have had regard only to the evidence now before me. On the basis of that evidence, I am satisfied that no such rigid relationship between borrowing and expenditure has been shown to exist, and (on the contrary) that in the short term borrowing was likely to take the strain if there was an unexpected shortfall in receipts.

Issue 24: is the “actual benefit” argument permitted by EU law?

On the twin assumptions that (contrary to what I have now held on Issues 22 and 23) the Revenue’s “actual benefit” argument were available to them under English law, and were also made out on the facts, it remains to consider whether the argument would nevertheless be precluded by EU law. For the reasons which follow, I consider that EU law would indeed preclude it.

Since the actual benefit argument is confined to the time-value mistake claims, another way of posing the question is to ask whether the claimants have a right under EU law to restitution of the full time value of the mistakenly overpaid tax. In considering this question, I proceed on the footing that these claims are properly characterised as San Giorgio claims under EU law: see [247] and [406] above.

Thus reformulated, the question turns on the correct interpretation of the guidance given by the Grand Chamber of the ECJ in Littlewoods (ECJ), and in particular on what the Court meant by the requirement (in paragraph 29 of the judgment) that the taxpayer should not be deprived of “an adequate indemnity” for the loss occasioned through the undue payment of tax. The tax in issue in the Littlewoods case was VAT, but nobody suggests that the principle (however it is to be interpreted) does not apply with equal force to the overpayments of directly levied tax contrary to EU law in the present case, always assuming that the mistake claims are indeed San Giorgio claims protected by EU law.

I have already considered this question at length in Littlewoods (No. 2). I discussed the adequate indemnity issue at [253] to [302], concluding as follows:

“302.

In sum, my overall conclusion on the difficult question of the meaning of the “adequate indemnity” test in paragraph 29 of the ECJ’s judgment is that it requires payment of an amount of interest which is broadly commensurate with the loss suffered by the taxpayer of the use value of the tax which he has overpaid, running from the date of payment until the date of repayment.”

I then held that the payment of simple interest (on claims which, as in the present case, dated back to 1973) would not provide the claimants with an adequate indemnity for their loss: see [303] to [310]. In broad terms, the quantum of the time value claims, even on the most favourable assumptions from the Revenue’s point of view, amounted to some £430 million more than the simple interest which had already been repaid to the claimants under section 78 of the Value Added Tax Act 1994. Against that background, I then considered at [418] to [420] whether the Revenue’s approach to quantification of the restitution due to the claimants, which relied on an “actual benefit” argument similar to that in the present case, was inconsistent with EU law. I held that it was.

For convenience, I will repeat what I said at [419] to [420]:

“419.

If I have correctly understood the judgment of the ECJ in the present case, it seems clear to me that the claimants will only receive an “adequate indemnity” for their loss occasioned by the overpayments of VAT if they are paid, at least, a sum which represents the use value of the overpayments in the hands of the Government. Strictly speaking, the use value which EU law entitles them to receive is the lost use value to themselves of the sums overpaid. It is common ground that such value would be greater than the use value to the Government, given the Government’s ability to borrow at below ordinary market rates. The claimants are, of course, not compelled to seek to recover the maximum remedy afforded to them by EU law, and nobody suggests that they have forfeited their right to claim compound interest merely because they are content to limit their claim to the use value of the overpaid tax in the Government’s hands rather than their own.

420.

It was necessary to award compound interest in Sempra so as to satisfy that company’s right under EU law to be paid the use value of the prematurely levied ACT. Now the ECJ has confirmed that there is a right to interest under EU law in respect of all repayments of taxes levied contrary to EU law, the same consequences must in my judgment follow. Only compound interest will suffice to satisfy the claimants’ EU law right to interest. I would therefore hold that the Revenue’s attempt to limit the restitution which it has to make to the benefit which it actually received from the overpayments is contrary to EU law, even it be assumed that, in a purely domestic context, this is all the English law of restitution would require.”

I remain of the same opinion, and although the Revenue disagree with various aspects of my reasoning and conclusions on the “adequate indemnity” issue in Littlewoods (No. 2), I think that I should simply follow and apply my approach in that case to the facts in the present case. On any view, the total amount of the test claimants’ time-value mistake claims, with compound interest dating back to 1973, is massively greater than the quantum of those claims which would result from application of the Revenue’s actual benefit methodology. I am satisfied that the latter amount would not begin to provide the claimants with an adequate indemnity for their loss, and that the EU law principle of effectiveness therefore prevents the Revenue from relying on such an approach.

Issue 25: what is the measure of restitution due to the claimants?

In the light of my answers to the Issues which I have already discussed, I consider that the restitution to which the claimants are entitled has three main elements.

First, the claimants are entitled to restitution of the full amounts of the principal sums of unlawfully paid tax. These sums comprise the relatively small amounts of overpaid Case V corporation tax, and the much larger amounts of overpaid and unutilised ACT.

Secondly, they are entitled to restitution of the time value of the prematurely paid (or utilised) ACT, from the dates of payment until the dates of utilisation. These claims are equivalent to those considered by the ECJ in Hoechst and by the House of Lords in Sempra. It is common ground that the time value of these claims has to be measured by reference to compound interest.

Thirdly, they are entitled to restitution of the time value of the amounts recoverable under each of the above headings, from the dates of payment (or the dates of utilisation of ACT payments in the second category) until the date when restitution is made.

In accordance with my understanding of the approach of the majority in Sempra, and my reasoning in both Portfolio Dividends (No. 2) and Littlewoods (No. 2), I consider that the measure of the time-value claims in the third category should likewise be measured by reference to compound interest. Only in this way can the claimants be provided with a fully effective restitutionary remedy under EU law for their loss caused by the payments of unlawful tax. There is no room, under either EU or English law, for the time value of these claims to be measured by reference to the actual benefit supposedly derived by the Revenue or the government from the overpayments. Nor is any distinction to be drawn, in terms of quantum, between the Woolwich claims and the mistake claims. Both categories of claim enjoyed equal protection under EU law throughout the period with which I am concerned. As a matter of English law, the measure of the Revenue’s enrichment is in my view the same whether the unjust factor relied upon to ground the claim is the payment of tax which is not lawfully due or mistake.

Issue 26(a): simple or compound interest?

Issue 26, as formulated by the parties, reads as follows:

“If it is correct to compute any component of the restitution due to the Claimants by reference to interest, what interest rates and rests are to be applied?”

In fact, however, the first question raised by the parties under this heading is whether interest should be simple or compound. I will therefore refer to this question as Issue 26(a). As will be apparent, I have already answered it in favour of the claimants; but this heading provides a convenient place in which to consider certain arguments advanced by the Revenue which I have not yet dealt with.

One argument, developed by Professor Myles in his first report, is that there is no accepted principle for determining the present value of a past monetary payment. He argues that, although the value of future payments can be discounted to the present by application of recognised techniques, there is no equivalent methodology which can be applied retrospectively. Professor Myles explains his central point as follows:

“46.

The key point is that at the present time there are financial transactions available that permit money to be carried forward into the future (saving) or to be brought from the future to the present (borrowing). In a developed financial system such as that in the UK these transactions will be almost risk-free. These financial transactions ensure that the rate of discount used in the present value calculation is equal to the rate of interest. The rate of discount is only affected by personal attitudes when there is risk about the future receipt of payments.

47.

Now consider the very different question of placing a present value on the receipt of £100 a year ago. Following the approach set out above to valuing a future receipt the principle is to find the amount today that is equivalent to the past receipt.

48.

The fundamental difference between the two situations is that there are no financial transactions that can transfer money backward in time. Nor are there any contracts one can enter into today to bring money forward from the past. This should be distinguished from the act of saving in the past to carry money into the present. When viewed from the present this is an action that cannot be undertaken.

49.

Therefore, it has to be concluded that the present value of a sum of money in the past cannot be determined by an arbitrage argument using financial transactions. For this reason there is no established principle to which appeal can be made to determine the present value of a past amount. The “time value of money” is not simply obtained by compounding.”

Professor Myles goes on to argue that the present value of a past sum depends on what the recipient did with the money. The relevant question is how much current benefit has been obtained from the receipt of the funds in the past. It is only if the recipient actually placed the money on deposit at a commercial rate of interest that compounding would produce an appropriate measure of the current benefit. Professor Myles concludes, in paragraph 65:

“The correct position is that the present benefit of past revenues received by the government can only be determined by analysing how the government used those revenues. Since there is no hypothecation it is not possible to identify the use of any particular funds. The question therefore requires an empirical answer based on the average use of resources by the government.”

Dr Sentance disagrees. A key passage in his report reads as follows:

“85.

The third substantive problem I see with the valuation model set out by the HMRC experts concerns their treatment of debt interest. The modelling carried out by Professor Myles is abnormal and counter-intuitive as it substantially reduces the size of the claim relative to more conventional economic methodologies. In normal circumstances, the value of money that has been received by an organisation or entity over a prolonged period would be expected to attract interest – and this interest would compound over time. The UK government is a large borrower, and so to the extent that it receives money – due to overpaid taxes – that it would otherwise have borrowed on the capital markets, it receives a benefit. In normal circumstances, that benefit will accumulate over time in money terms as the interest saved reduces the amount of debt that has to be rolled over. Effectively, compound interest reflects this process.

86.

Professor Myles adopts a very different approach. He assumes – like additional revenue – that the bulk of the interest saved allows a boost to public spending, with a small proportion allocated to deficit reduction. As the vast majority of this increase in public spending flows through to current expenditure, it is then assumed to be “irretrievably spent”, like the substantial part of the initial tax overpayment. As a result, the compounding effect is put into reverse. Professor Myles’ formula degrades, rather than compounds, debt interest. As a result of this approach, and the other assumptions Professor Myles uses in his model based on Sir Jonathan Stephens’ report, the valuation he arrives at is just over £34m. To put this into context, it is less than 10% the value of the overpaid tax uprated to 2012/13 prices …, even before any interest is added.”

In his oral evidence in chief, Dr Sentance was asked to comment on Professor Myles’ view that the principle of compounding cannot be applied retrospectively. Dr Sentance said that he would describe compounding as the conventional approach, and the natural and logical way of valuing something in the present case. He based this view on other cases that he had seen, and on the approach adopted by his economist colleagues at PwC. When asked what kind of valuation work his colleagues were undertaking, he replied:

“These are valuations where there has been a mispayment in the past of some form or some sum of money in the past that then needs to be compensated for in the present, which seems similar to what we have here.”

Professor Myles had pointed out that no explicit textbook authority could be found for retrospective compounding of interest as a means of ascertaining the present value of a past receipt, but equally he had been unable to find any textbook or academic support for his view that the process was in principle illegitimate. When this was put to Dr Sentance, he replied that he had not come across any specific statement that says “you can’t apply the time value of money looking back and you should only apply it for the future”.

I prefer the evidence of Dr Sentance on this issue. I can see no logical reason why the process of compounding should not be applied retrospectively when placing a present value on a past receipt of money, and I note that Dr Sentance and his colleagues regard this as the natural approach to adopt. It is, of course, also the approach which was adopted by the House of Lords in Sempra when ascertaining the time value of prematurely paid ACT. In principle, and subject to what I say below about section 35A of the Senior Courts Act 1981, it seems to me that the position must also be the same in relation to the overpayments of corporation tax and unutilised ACT. In each case, the Revenue has had the opportunity to use the money, and the way in which an opportunity to use money is purchased in the marketplace is through borrowing it at compound interest.

I should add that an argument similar to that advanced by Professor Myles was considered by me, and rejected, in Littlewoods (No. 2) at [381] to [384]. I rely on my reasoning in that passage as further support for the conclusion which I have reached in the present case.

I now turn to an argument based on statute and case law. In Portfolio Dividends (No. 2) I held at [241] to [247] that the approach of the House of Lords in Sempra should be applied to measure the time value of all the other claims for overpaid tax, including the post-utilisation period in respect of utilised ACT. The Revenue now submit that I was wrong so to conclude, because under English law the position is governed by section 35A of the Senior Courts Act 1981. A claim for the restitution of overpaid tax, it is argued, is a claim for the recovery of a debt. In such cases, in the absence of a trust relationship, it is the settled practice of the court to award simple interest under section 35A: see, for example, Guardian Ocean Cargoes Ltd v Banco do Brasil S.A. [1994] 2 Lloyd’s Rep 152 at 159-160 per Saville LJ. This was also the course followed by the House of Lords in the Westdeutsche case (Westdeutsche Landesbank Girozentrale v Islington London Borough Council [1996] AC 669) which was distinguished, but not overruled, in Sempra.

The Revenue accept that the decision in Sempra made an important step forward, by awarding compound interest at common law. They submit, however, that this step forward did not contradict either section 35A or the earlier line of authority to which I have referred. The critical distinction is said to be that Sempra was not concerned with a standard restitutionary claim for money paid, but was instead a restitutionary claim for the early payment of money. That was the principal sum claimed, and it was as the measure of that sum that compound interest was awarded.

In my judgment there are a number of answers to this submission. First, as I have already explained, I consider that an award of compound interest is required by EU law on all of the claims in the present case. Only thus can the claimants obtain an “adequate indemnity” for their loss. Secondly, as a matter of English law, I consider that section 35A does not apply, precisely because the time value of the overpayments of tax forms part of the principal sum for which restitution has to be given. This is a theme which runs through the speeches of Lord Hope and Lord Nicholls in Sempra: see, for example, [31] to [34], [46] to [48], [114] and [120] to [122]. The significance of this theme was recognised, and endorsed, by Lord Walker at [178] when he said:

“The crucial insight in the speeches of Lord Nicholls and Lord Hope is, if I may respectfully say so, the recognition that what Lord Nicholls calls income benefits are more accurately characterised as an integral part of the overall benefit obtained by a defendant who is unjustly enriched. Full restitution requires the whole benefit to be recouped by the enriched party: otherwise “the unravelling would be partial only” …”

The extent to which Westdeutsche is still good law, in the light of Sempra, is a difficult question on which I have not heard full argument in this, or any other, case. I will therefore merely say that, in my judgment, the approach of the majority in Sempra entails a departure from Westdeutsche at least to the extent of recognising the existence of jurisdiction at common law to award compound interest as part of the restitution due to a claimant who seeks to recover overpaid, or prematurely paid, tax. The key passages in the speeches are to be found at [35] to [36] (Lord Hope), [110] to [112] (Lord Nicholls), [150] to [153] (Lord Scott), [183] to [188] (Lord Walker) and [218] to [224] and [234] to [240] (Lord Mance).

It is true that the claim before the House in Sempra was confined to prematurely paid ACT, but the logic of the majority’s approach must in my judgment extend to claims for restitution of overpaid tax. Indeed, such claims are, in my view, if anything stronger, because the time-value part of the claim is not freestanding and is anchored to the claim for restitution of the overpaid tax itself. As Lord Hope explained at [36], commenting on the fact that the House had not been invited to overrule Westdeutsche:

“36.

Furthermore the interest in question in the present case is, as the Court of Justice stressed … the principal sum itself. In my opinion the decision in the Westdeutsche case does not address this point. We were not asked to overrule that decision, because it is distinguishable on this ground. Furthermore, the basis of Sempra’s claim, as the common law has now recognised, is unjust enrichment. I do not think that it is open to the common law, when it is providing a remedy in unjust enrichment, to decline to apply the principle on which that remedy is founded when the principal sum to be awarded is being calculated. As Lord Nichols points out (see para 99), there is now ample authority to the effect that interest losses which are recoverable as damages should be calculated on a compound basis where the evidence shows that this is appropriate. The same rule should be applied to the restitutionary remedy at common law.”

In his oral submissions, Professor Burrows courteously complained that in Portfolio Dividends (No. 2) I had unjustifiably driven a coach and horses through section 35A, and if my approach were right, little if any scope would be left for section 35A to apply in restitution claims. For the time being, I would not wish to extend my approach beyond the context of claims for the recovery of overpaid tax. I also readily acknowledge that Sempra is a very difficult case to analyse and apply. It seems to me, however, that Sempra was on any view a case of groundbreaking significance in the development of the common law, and I think the courts are still only beginning to work out its full implications.

Issue 26(b): what interest rates and rests are to be applied?

The parties are in agreement that a reasonable measure of the cost of UK government borrowing over the period since 1973/74 is provided by a ten year moving average of the yield rate on ten year gilts. Since such gilts pay a coupon twice yearly, the claimants contend that the interest rate so ascertained should be applied with six monthly rests. For their part, the Revenue rely on a table of annualised rates, apparently based on information provided to the Revenue by the Bank of England for the purposes of the Littlewoods case. Since the end of the present hearing, the parties have fortunately been able to agree that both approaches yield the same result, i.e. that the moving average rate compounded six monthly produces the same result as the annualised rates shown on the Revenue’s table.

As in the Littlewoods case, the claimants are content to take the cost of government borrowing as a reasonable proxy for the time value of their own loss. These are therefore the rates which in my judgment should be applied to all of the claimants’ time value claims. In the interests of clarity, I think it is preferable to express the rate as one which compounds six monthly, rather than by reference to the annualised rates, although in practice nothing appears to turn on the distinction.

Issue 27: in respect of which periods do the claimants have valid Woolwich claims?

The answer to this question is agreed. The limitation period in relation to the claimants’ Woolwich claims is six years from the date of each tax payment, with the result that for the BAT test claimants Woolwich claims are available for tax payments made on or after 18 June 1997. The claim form was issued six years later, on 18 June 2003.

Issue 28: when did the claimants discover (or when could they with reasonable diligence have discovered) their mistake?

This is an isolated question, which is raised because it was apparently left undecided in FII (High Court) I. The answer to it is of no practical significance for the BAT test claimants, but in so far as the answer turns on an objective assessment of when the claimants could with reasonable diligence have discovered their mistake, it may be of relevance to other claims in the FII GLO. The question is concerned with the limitation period applicable to the mistake-based claims, and the extension of the normal six year period effected by section 32(1)(c) of the Limitation Act 1980.

The claimants contend that they did not discover, and could not with reasonable diligence have discovered, their mistake about the unlawfulness of the Case V charge and the relevant parts of the ACT regime until the ECJ delivered its judgment in FII (ECJ) I on 12 December 2006. They further argue that I have in fact already made a finding to this effect in FII (High Court) I at [267]. By contrast, the Revenue argue that the claimants could reasonably have discovered their mistake on 8 March 2001, the date when the ECJ delivered its judgment in the Hoechst case.

Since the claim form was issued on 18 June 2003, the proceedings were clearly begun well within the relevant six year period, even if it had started to run on 8 March 2001. This explains why the question is of no practical significance in the present case. But it also brings out what may at first sight appear to be an insuperable logical difficulty in the claimants’ case on this issue. How can it be said that they neither had discovered, nor with reasonable diligence could have discovered, their mistake until 12 December 2006, when they had already started the present action (including the mistake claims) three and a half years earlier?

Before attempting to answer that question, I will first record what I said in FII (High Court) I at [267]:

“267.

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. There was no question of paying the tax under protest, as in Woolwich. It is only now, in the light of the decision of the ECJ, that a mistake can be seen to have been made.”

I would make the following comments about that paragraph. First, it related only to the ACT claims, and came in the part of my judgment where I was seeking to apply the general principles on remedies, which I had discussed at [201] to [265], to the main heads of claim. Secondly, it would in my view be wrong to read the paragraph as a finding about when time began to run under section 32(1)(c). As I have explained, that was not identified as an issue which I had to decide, because it made no difference to the outcome. The finding was, rather, directed to what I had said in [261] and [262] about the nature of the relevant mistake, and the fact that it could usually only be recognised after the event and with the benefit of hindsight. Thirdly, the decision of the ECJ to which I referred must, on any natural reading, have been the decision in FII (ECJ) I, not the decision in Hoechst.

I should also refer to some of the evidence about the state of mind of the claimants when the action was started in June 2003. In his third statement dated 10 April 2008, Mr Hardman says that it was only in early 2003 that he and his then deputy head of tax at BAT, Chris Powell, became aware of the possibility of recovering surplus ACT through litigation. They obtained copies of the materials used for a seminar which had been given by PwC in September 2002. Mr Hardman then says (paragraph 107):

“It is from these documents and the conversations with our tax advisers which followed that we became aware of the argument that ACT paid in the circumstances of BAT might be a breach of Community law and might be recoverable.”

It was also around this time that Mr Hardman became aware of the case brought by DMG against the Revenue in relation to ACT paid on distributions by UK resident companies to a German parent. Mr Hardman’s understanding was that DMG was claiming that it had paid ACT under a mistaken belief that it was lawfully due, and that the six year limitation period for such claims did not start to run until it discovered its mistake, which was when the ECJ had given its ruling in the Hoechst case. 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.

The hearing at first instance in DMG took place before Park J on 12 to 15 May 2003, and he handed down his judgment on 18 July 2003. The present claim was therefore begun about a month after the conclusion of the argument, and a month before judgment was delivered. Not long afterwards, on 8 September 2003, the government announced its intention to introduce legislation curtailing the extended limitation period available under section 32(1)(c). This announcement did not affect BAT, however, since the claim form had already been issued. Mr Hardman says he continued to believe that “the question whether BAT’s claim could go back to 1973, when ACT was introduced, would be decided by the DMG case”.

The decision of Park J in DMG’s favour was reversed by the Court of Appeal, but restored by the House of Lords which delivered its judgment on 25 October 2006. One of the matters decided by the majority (Lord Hoffmann, Lord Hope and Lord Walker) was that DMG’s mistake about the law was not reasonably discoverable until after the ECJ had delivered its judgment in the Hoechst case: see [31], [71] and [144]. In the view of the majority, it 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.

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.

Some support for this view of the matter may be found in FII (SC), in the context of the consideration by the Supreme Court of the EU law principle of protection of legitimate expectations. The court was unanimous in deciding that the enactment of section 107 of the Finance Act 2007 infringed the legitimate expectations of the claimants. The court was divided, however, on the question whether the enactment of section 320 of the Finance Act 2004 also infringed the principle. Five of the Justices held that it did, but Lord Sumption and Lord Brown of Eaton under Heywood dissented on this issue (which was accordingly referred to the ECJ for determination).

In his discussion of this subject at [197] to [202], Lord Sumption examined the position as the DMG litigation made its way through the courts. In paragraph [199] he said this:

“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 [2006] Ch 243 and was not definitively established until the judgment of the House of Lords [2007] 1 AC 558 on 25 October 2006.”

See too [166] to [168] and [201] to [202], where Lord Sumption ended his discussion by pointing out that the relevant principle was substantially the same as that on which the test claimants themselves rely:

“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 Deutsche Morgan Grenfell) 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.”

Lord Walker disagreed with Lord Sumption on the validity of section 320 of the 2004 Act under EU law, but he agreed with much of what Lord Sumption said in paragraph [199] of his judgment. Thus Lord Walker said at [111]:

“111.

Lord Sumption holds, at para 199, that reasonable persons in the position of the test claimants would not, until Park J’s judgment in DMG on 18 July 2003, have counted on being able to recover tax on the ground of mistake of law; and that even after that decision the existence of such a claim was being challenged on serious grounds. He concludes from that proposition that 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.

112.

I cannot disagree with that conclusion …”

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.

On the strength of my analysis so far, I would be inclined to hold that the claimants did not discover (and could not with reasonable diligence have discovered) their mistake within the meaning of section 32(1)(c) until 25 October 2006. But I still need to consider a further passage from Lord Walker’s judgment in FII (SC). In the course of his discussion of legitimate expectations, Lord Walker reviewed the sequence of events leading up to the enactment of section 320 of the 2004 Act and section 107 of the 2007 Act (which he called “the statutory cut-off provisions”). After stating that the first two key dates were the decisions of the House of Lords in Woolwich and Kleinwort Benson, Lord Walker continued as follows in paragraph [103]:

“After 1998 English lawyers knew that the recovery of money paid under a mistake of law (perhaps including a mistake of tax law, subject to arguments on exclusive remedies) had become a real possibility, although it was by no means a firmly established cause of action. But until the decision of the Court of Justice in [Hoechst] on 8 March 2001 there was no general appreciation that the UK corporation tax regime was seriously open to challenge as infringing the Treaty. Henderson J did not make any detailed findings about this, since the principle of legitimate expectations does not seems to have been argued as a separate issue before him. But he did … make a general finding of fact about mistake:

[Lord Walker then quoted paragraph 267 of FII (High Court) I down to the words “was lawfully due and payable”]

104.

After 8 March 2001 a well advised multi-national 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. During 2002 the opinion of the Advocate General and the judgment of the Court of Justice in M & S, while possibly not adding much to the earlier jurisprudence, spelled out very clearly, for UK companies and lawyers, both the capacity and the limits of national legislation in curtailing limitation periods in proceedings for recovery of tax levied in breach of EU law.”

It is, I think, clear from this passage that Lord Walker regarded the date of the ECJ’s judgment in Hoechst (8 March 2001) as the latest date at which time could have begun to run against the test claimants under section 32(1)(c), notwithstanding the uncertainty which still surrounded the cause of action based on mistake. Lord Walker cannot have been referring only to the Woolwich cause of action, because of his express reference to section 32(1)(c); and the words “until then (but not subsequently)” (my emphasis) show that in his view this was the latest date to which the running of time could arguably have been postponed. Furthermore, Lord Walker was not alone in taking this view. Both Lord Clarke (at [131]) and Lord Reed (at [236]) quoted or paraphrased what Lord Walker had said at [104] with evident approval, although it is fair to note that neither of them expressly reproduced the force of the words “but not subsequently”. Lord Walker’s reasoning in this part of the case also had the general agreement of Lord Hope (at [16] to [22]) and Lord Dyson (at [140]). Only Lord Sumption and Lord Brown disagreed.

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 the Metallgesellschaft case, 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.

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 the Hoechst/Metallgesellschaft case. 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.

There is a second limb to Issue 28, which asks in respect of which payments and periods the claimants have valid mistake claims. It is unnecessary for me to discuss this question, because it is common ground that on any view the claimants started their mistake claims within the extended limitation period. As a result, all of the mistake claims dating back to 1973 are in time.

Issue 29: what is the quantum of restitution (which includes restitution for continuing enrichment) to which the BAT claimants are entitled?

It is agreed that I should not attempt to quantify the claims at this stage, and that in the absence of agreement further submissions should be made in the light of this judgment.

The ultimate parent company of the GKN group was GKN Holdings Plc (“GKN Plc”), a publicly listed company incorporated and resident in the UK. The group operates through subsidiaries established throughout the world largely in the automotive, aerospace and defence industries.

In the accounting periods relevant to this claim (i.e. since 1 January 1973) GKN Plc held a large number of subsidiaries both in the UK and abroad in countries situated within and outside the EU/EEA. Its participation in its foreign subsidiaries was either directly or indirectly through UK resident intermediate parents or offshore holding companies. The claim of the GKN group has been appointed as a supplementary test case to address the circumstances of dividends paid to the UK from non resident EU intermediate holding companies benefiting wholly or partly from a participation exemption where those dividends are attributed to profits earned in various jurisdictions, both in the EU/EEA and beyond it. To address these circumstances 4 sample dividends have been taken. This statement of facts concerns only those sample dividends.

Three of the sample dividends were paid to GKN Plc by GKN Netherlands BV, an intermediate holding company resident in the Netherlands, and who benefited from a participation exemption for dividends, in the accounting periods ending 31 December 1980, 1981 and 1982. There is no indication that this Dutch company used equity accounting. The dividends for 1980, 1981 and 1982 were sourced mainly in the profits of subsidiaries earned in other jurisdictions. The sample dividends paid in the years ending 31 December 1980 and 1981 were paid out of specified profits. Schedules 7, 8 and 9 are diagrams describing the corporate structure of the companies in the group relevant to each of those dividends.

Tables 1-3 at Schedule 11 show for each of those dividends the amount, the underlying tax paid plus any withholding tax, the source of those profits and (where the source state was a member state of the EU/EEA) the foreign nominal rate.

Under s801 of ICTA the underlying tax rate was established by a weighted average of the tax paid on each component of the dividend income. The underlying rate established in accordance with those provisions as agreed between the parties is also stated.

The fourth sample dividend was paid to A P Newall and Co Limited, a UK resident company within the GKN Group, by Uni-Cardan AG, an intermediate holding company resident in Germany in the accounting period ending 31 December 1986. It was the distribution of the profits of Uni-Cardan AG for the accounting period ending 31 December 1985. Those profits incorporated dividends received directly and indirectly from a variety of companies resident in Germany, other EU Member States (Italy, Denmark, Belgium, France and the Netherlands) and third countries (Austria, Spain and Switzerland). In all cases the profits underlying the dividends were earned by the distributing company or its subsidiaries in the same jurisdiction and tax was paid on those underlying profits in that jurisdiction. At Schedule 10 is a diagram of the group structure.

Uni-Cardan AG was the top company in a German consolidated tax group known as an Organschaft and benefited from a partial participation exemption. In the case of the selected dividend, where a German corporation received a dividend from a foreign subsidiary this dividend was ordinarily taxable income. However, under the German tax treaties with Belgium, Denmark, France, Switzerland and Spain (subject to conditions) an exemption privilege applied. Under the treaties with Italy and Austria, instead of exemption, the dividends remained taxable with the granting of a tax credit.

The practical effect of an Organschaft is that:

The Austrian and Italian subsidiaries were only half owned directly by Uni-Cardan AG, the balance being owned by other German subsidiaries (see Schedule 10). Consequently the portion of the dividends of the Austrian and Italian subsidiaries paid to those other German subsidiaries of Uni-Cardan AG were also subject to tax in Germany.

Part of those 1985 profits of Uni-Cardan AG included its own earnings upon which it paid German tax. As the top company in the Organschaft, Uni-Cardan AG also paid German tax on the profits of the other companies in the Organschaft.

The amount of German tax paid on the 1985 profits used for the dividend paid in 1986 by Uni-Cardan AG exceeded the UK CT payable on the resultant DV income. Consequently, there was no need to agree the amount of superfluous tax on the underlying dividends paid to Uni-Cardan AG by companies resident outside Germany. On the aggregate basis determined in accordance with s801 ICTA, the DV income qualified for full double tax relief.

Table 4 of Schedule 11 sets out the amount of the 1986 dividend paid by Uni-Cardan AG, the amount of German tax paid on the 1985 profits underlying that dividend regarded as sufficient to qualify the dividend for full double tax relief, the sources of profit which comprised Uni-Cardan AG’s 1985 profits and the foreign nominal rate applicable to Germany.