Test Claimants In the FII Group Litigation v HM Revenue & Customs

The Franked Investment Income (“FII”) Group Litigation with which this judgment is concerned was established by a Group Litigation Order (“GLO”) made by Chief Master Winegarten on 8 October 2003. The claimants are all companies which belong to groups which have UK-resident parents and also have foreign subsidiaries, both in the European Union (the “EU”) and elsewhere (“third countries”). In the broadest terms, the purpose of the FII GLO was to determine a number of common or related questions of law arising out of the taxation treatment of dividends received by UK-resident companies from non-resident subsidiaries, as compared with the treatment of dividends paid and received within wholly UK-resident groups of companies. The claimants argued that there were unjustified differences between the taxation treatment of UK companies with resident and non-resident subsidiaries respectively, which breached the provisions of Article 43 (freedom of establishment) and Article 56 (free movement of capital) of the EC Treaty and their predecessor articles, and which had caused them loss dating back (in at least some cases) to the accession of the UK to the EU in 1972 and the introduction of the advance corporation tax (“ACT”) regime in April 1973.

The system of UK corporate taxation relating to dividends which underlies the FII Group Litigation is the same as that which gave rise to the well-known decision of the Court of Justice of the European Communities (“the ECJ”) in the Hoechst case (Metallgesellschaft Ltd v IRC, Hoechst AG v IRC (joined cases C-397/98 and C-410/98), [2001] ECR I-1727, [2001] Ch 620) and the spate of litigation which has followed it, including the landmark decisions of the House of Lords in Deutsche Morgan Grenfell Group Plc v IRC [2006] UKHL 49, [2007] 1AC 558 (“DMG”) and Sempra Metals Ltd v IRC [2007] UKHL 34, [2007] 3WLR 354 (“Sempra Metals”). That litigation has been managed in England under a different GLO, namely the ACT Group Litigation Order. Whereas the focus of the ACT Group Litigation was on the UK domestic legislation which prevented UK-resident subsidiaries of foreign parents from making group income elections, thereby obliging them to pay ACT when paying dividends to their foreign parents, the focus in the FII Group Litigation has been on UK-parented groups with foreign subsidiaries, and on the tax treatment of dividends coming into the UK from abroad. At the simplest level, therefore, the present litigation is concerned with factual situations which are the opposite of those which gave rise to the questions considered in Hoechst and the ACT Group Litigation.

On 28 June 2004 the trial of selected test claims in the FII Group Litigation began before Park J, but it was immediately apparent that a preliminary reference to the ECJ would be needed on the numerous issues of Community law arising in the case. Without delivering a formal judgment, Park J directed that a reference be made to the ECJ and asked the parties to agree an order for reference setting out the facts and questions to be decided. The order was eventually made on 13 October 2004 (“the Order for Reference”), following detailed consideration of its terms by the parties and Park J. As well as setting out the questions on which the ECJ was asked to rule, the Order for Reference contained a useful statement of the legal and factual background to the case upon which I shall draw later in this judgment.

On 6 April 2006 the Advocate General (Geelhoed) delivered his opinion, and on 12 December 2006 the ECJ gave its judgment on the reference (Case C-446/04, Test Claimants in the FII Group Litigation v CIR, [2006] ECR I-11753, [2007] STC 326). Much of my judgment will be devoted to a detailed examination of the opinion of the Advocate General and the judgment of the ECJ. The case is notable, not only for its length and complexity, but also for the number of issues on which the ECJ differed, or arguably differed, from the opinion of the Advocate General. The decision does not lend itself to brief summary, but in general terms the claimants succeeded in many of their central complaints about the ACT regime, the Revenue was also successful on some important points, and a number of issues were left to be decided by the English court, including a crucial factual issue which will determine whether or not the UK corporation tax regime for taxation of foreign dividends under Case V of Schedule D is compatible with Community law, and some very important questions relating to remedies.

At a subsequent case management conference held on 5 July 2007 before Rimer J (as he then was), who had succeeded Park J as the judge designated to manage the FII Group Litigation, it was directed that the test claims should proceed to trial of all the issues in the FII GLO, including liability for restitution, “save in so far as those issues concern causation or quantification”. All issues relating to liability were therefore to be decided at the forthcoming trial, with questions of causation and quantum deferred for subsequent determination. Detailed directions were also given for service of amended pleadings and preparation for trial, including a direction for agreement of a list of questions to be decided by the court.

I do not think it would be useful to give a detailed synopsis of the pleadings and the issues at this stage, because what the parties mainly need is the decision of the court on questions of principle which can be stated in fairly abstract terms, without reference to the particular underlying facts. The pleadings play an essential role in defining the issues and laying the necessary factual foundations for the questions of law which have to be decided, but in group litigation of this nature the pleadings tend to recede into the background once the stage of trial has been reached. What I propose to do instead is to set out the basic legal background, and then to describe the factual background only to the limited extent necessary to explain how the main issues of law arise. I will then deal with the issues in the same general order as they were argued before me, and when questions of fact need to be determined I will explain what they are at the appropriate juncture and make my findings accordingly.

It may be helpful, however, if I say by way of introduction that the questions put to the ECJ, and the questions which I now have to decide, may be grouped under four broad headings:

Almost without exception, these issues are both difficult and important, and huge amounts of money are potentially at stake. I was told that the maximum amount of the claims advanced in the FII Group Litigation is of the order of £5 billion. I express my unfeigned gratitude to all counsel on both sides, and to their legal teams, for the high quality of the written and oral submissions which were addressed to me in the course of a hearing which lasted for 13 days in July 2008.

I will conclude this introductory section of my judgment with two minor points which it is convenient to clear out of the way.

First, I was informed that countries which are not members of the EU, but are within the European Economic Area (“the EEA”), are in essentially the same position as EU member states in relation to all the issues which I have to decide. In the interests of simplicity, I will not normally make separate references to the EEA in the remainder of this judgment, but my references to EU member states should, where appropriate, be read as including states which are within the EEA for periods after the EEA Agreement came into being on 1 January 1994.

The second point is that significant parts of the opinion of the Advocate General and the judgment of the ECJ in the present case are devoted to questions of compatibility of the relevant UK legislation with Articles 4 and 6 of the Parent-Subsidiary Directive, Council Directive 90/435/EEC of 23 July 1990. However, the ECJ held that the relevant provisions of the Directive were not engaged, with the result that for the purposes of this judgment it can be ignored.

This section of my judgment is largely based on the agreed material set out in the Order for Reference, and I shall begin by quoting the summary of the relevant legislation contained in paragraphs 4 to 9 thereof:

I will now describe the arrangements in more detail, again drawing to a large extent on the account given in the Order for Reference. Statutory references, unless otherwise stated, are to the Income and Corporation Taxes Act 1988 (“ICTA”), which consolidated equivalent provisions in the previous legislation.

Where a UK-resident company made a qualifying distribution it was liable to pay ACT on the distribution: section 14(1). The sum of the amount of the distribution and the ACT was called a franked payment: section 238(1). Before 6 April 1993, the rate of ACT was linked to the basic rate of income tax. For example, from 1988 to 5 April 1993, when the basic rate of income tax was 25%, the ACT rate was 25/75 (or 1/3) of the amount of the distribution. Between 6 April 1993 and 5 April 1994 the ACT rate was set at 22.5/77.5 (or 9/31). From 6 April 1994 until 5 April 1999 the ACT rate was linked to the lower rate of income tax: section 14(3). At that time the lower rate of income tax was 20%. The ACT rate was therefore 20/80 (or 1/4).

Income tax was charged under various “Schedules” for different types of income. This is a peculiar feature of the UK tax system. Under Schedule F (section 20) individual shareholders were liable to income tax on dividends and other distributions received. A UK-resident individual in receipt of a qualifying distribution from a UK-resident company was entitled to a tax credit equal to such proportion of the amount or value of the distribution as corresponded to the rate of ACT: section 231(1). Income tax was chargeable on the total of the distribution and the tax credit: section 20(1). The tax credit extinguished all or part of the taxpayer’s liability. Lower-rate taxpayers and non-taxpayers (e.g. taxpayers whose income did not exceed the personal allowances) could recover some or all of the tax credit in cash. Entities not subject to UK tax on investment income, e.g. pension funds, could before 2 July 1997 claim payment in full of the tax credit on dividends received.

A UK-resident company was subject not to income tax but to corporation tax: section 6(2). However, corporation tax was not chargeable on dividends and other distributions received from another UK-resident company, nor were such payments taken into account in computing the corporation tax liability of the company making the distributions. This follows from section 208, which provides as follows:

A UK-resident company was, by contrast, subject to corporation tax on dividends received from non-resident companies. Such tax was charged under Case V of Schedule D, set out in section 18, being

The standard clause in such treaties, reflecting section 790(4), is usually to be found in the “Elimination of Double Taxation” article. So, for example, Article 22(b) of the UK/Netherlands Double Taxation Treaty reads:

A UK-resident company receiving a qualifying distribution from another UK-resident company was entitled to a tax credit: section 231(1). The total of the distribution and the tax credit was called franked investment income (“FII”): section 238(1). A UK-resident company receiving a distribution from a non-resident company was not entitled to a tax credit, and the income did not qualify as FII. Where a UK-resident company received FII, it was liable to pay ACT in relation to its own dividends only to the extent that those dividends and the ACT referable to them (i.e. its franked payments) exceeded the FII: section 241. Special arrangements applied under section 247 to dividends paid between UK-resident members of groups of companies. Provided that they satisfied certain minimum holding requirements – broadly speaking, the requirement was that more than 50% of the shares of the company paying the dividend had to be held by the parent – the UK-resident subsidiary and its UK-resident parent could make an election (called a group income election) under which dividends could be paid to the parent by the subsidiary without its having to account for ACT. Where a group income election was in force, the payment of dividends under it did not entitle the parent company to a tax credit, and the dividends were not included within its FII. The effect of a group income election was to postpone the payment of ACT until a distribution was made by the parent company.

A company was entitled to set ACT paid in respect of a qualifying distribution during an accounting period against its mainstream corporation tax (“MCT”) liability for that and future periods. There was, however, a limit on the amount which could be set off based on the income tax rate (see paragraph 14 above). Since the UK operated a partial imputation system, so that the UK corporation tax rate exceeded the ACT set-off rate, the company always faced a marginal corporation tax liability on its profits. Moreover, where a company received credit for foreign tax, this reduced the amount of the corporation tax liability available for set-off of ACT: section 797(4). Unrelieved ACT, known as “surplus ACT”, could be carried back or forward for set-off against MCT of other periods: section 239.

A company was also permitted to surrender to its subsidiaries the benefit of ACT payments it had made: section 240. The subsidiaries to whom the surplus ACT could be surrendered were restricted to subsidiaries resident in the UK: section 240(10). The subsidiaries were then able to set the surrendered ACT against their own UK MCT liability.

A company with surplus FII (that is, FII which exceeded franked payments) could, if it had losses, set the amount of those losses against the surplus FII under section 242 and obtain a payment in cash of the tax credit comprised in that amount of surplus FII. This provision was abolished with effect from 2 July 1997.

Experience with the arrangements described above showed that companies receiving significant foreign dividend income generated surplus ACT. This was because:

The reclaimed ACT became repayable at the same time as the MCT became payable, i.e. nine months after the end of the accounting period, and was set first against any MCT liability for the period and any excess was then repaid. As ACT was paid 14 days after the end of the quarter in which the dividend was paid, and MCT was payable nine months after the end of the accounting period, this meant that ACT would remain outstanding under the FID system for between 8½ and 17½ months depending when the dividend was paid in that accounting period.

A FID did not constitute FII, although a corporate shareholder could use a FID received by it to frank a FID paid, so that ACT was payable only on the excess of FIDs paid over FIDs received. Because a FID did not constitute FII the shareholder receiving the FID was not entitled to a tax credit under section 231(1); but an individual receiving a FID was nevertheless treated as receiving income which had borne tax at the lower rate for the year of assessment. However, no repayment was made to individual shareholders of income tax treated as having been paid, nor could a tax exempt shareholder such as a UK pension fund reclaim a tax credit similar to that which would have been payable on a non-FID qualifying distribution.

For distributions made on or after 6 April 1999, the ACT system was abolished. Companies no longer had to pay or account for ACT on shareholder dividends and other qualifying distributions. The FID rules were also abolished.

For companies with brought-forward surplus ACT, a “shadow ACT” system was introduced. The shadow ACT regulations allowed companies access to their surplus ACT in an amount broadly similar to the relief allowable under the old rules. This meant that surplus ACT could only be utilised after the shadow ACT was notionally used and exhausted.

UK-resident individuals now receive dividends with a tax credit equal to one ninth of the dividend. The tax credit extinguishes lower and basic rate income tax liability on the dividend.

The test claimants for all of the FII GLO issues bar one are members of the British American Tobacco (“BAT”) multinational group of companies. The parties have agreed a statement of facts relating to the BAT test case, which I will now reproduce.

The payments of dividends by non-resident subsidiaries to the claimants were made in the period beginning in financial years ending in 1973 and up to the present.

The payments of dividends from the UK-resident parent to its public shareholders were made in the period beginning in the financial years ending in 1973 and ending in the quarter ending 31 March 1999. Dividends paid by it to its public shareholders after that time are not relevant.

The payments of ACT made by the Claimants were in the period commencing in the financial years ending in 1973 and ending on the 14 April 1999.

FID payments were made between the quarter ending 30 September 1994 and the quarter ending 31 March 1999.

The Claimants are all companies within a group of companies known as the BAT group. They comprise the publicly owned ultimate parent company of the group, the UK resident intermediate parent companies through which non-resident subsidiaries were held, other UK resident subsidiaries and one current and one former Dutch resident holding company. Throughout the Relevant Period the BAT group traded through subsidiaries in most countries of the world.

The essential features of the group structure of the Claimants relevant to their claim did not change throughout this time, in that:

the parent company (the ultimate parent) was UK resident, publicly owned and its shares were publicly traded;

each of the UK resident intermediate parents was the wholly owned subsidiary of the ultimate parent either held directly or indirectly through another of the UK resident claimant companies;

the ultimate parent and each of the UK resident intermediate parents wholly owned either directly or indirectly numerous subsidiaries resident in most, if not all, EU Member States and EEA countries and many third countries. The names and seats of those non-resident subsidiaries are set out in Schedule 1 to the Eighth Amended Particulars of Claim.

While those essential features remained constant throughout the Relevant Period the identity of the ultimate parent and group structure changed (although the ultimate parent was always one of the claimant companies). References to the ultimate parent identifies the relevant company at the time.

The BAT Group’s traditional interests were in tobacco and at the commencement of the Relevant Period (1973) the ultimate parent company was BAT Co Ltd, now British American Tobacco (Investments) Ltd (the 2^nd Claimant) (“BAT Co”).

The group had prior to the Relevant Period commenced diversifying into other businesses so that by 1978 it also owned interests internationally in retailing, cosmetics and paper and packaging. At the commencement of the Relevant Period BAT Industries plc (the 1^st Claimant) (“BAT Industries”) was a direct wholly owned subsidiary of BAT Co.

In 1976 the group restructured and BAT Industries became the ultimate parent and BAT Co its wholly owned subsidiary. BAT Co remained the UK intermediate holding company for much of the group’s international tobacco interests.

In the 1980s the BAT group expanded into financial services acquiring the Allied Dunbar and Eagle Star groups.

In 1989 BAT Industries was the subject of a hostile takeover bid by Hoylake Investments. In response the Board of BAT Industries announced an increased dividend, a share buy back scheme and BAT Industries divested its interests in its businesses other than tobacco and financial services.

In 1998 BAT Industries divested itself of its financial services businesses which companies were merged with Zurich Insurance AG leaving the group with interests exclusively in tobacco.

In 1998 British American Tobacco plc (the 5^th Claimant) (“BAT plc”) became the ultimate parent and BAT Industries its direct wholly owned subsidiary.

Beneath the ultimate parent company the interests in other companies within the BAT group resident both in the UK and elsewhere were owned through UK intermediate holding companies (the UK resident intermediate parents) in the manner described above. Relevant to this claim are:

BAT Co and BAT Industries for the periods they were not the ultimate parent company;

British American Tobacco (Holdings) Limited (the 3^rd Claimant), formerly Staines Investments Ltd, which was acquired by BAT Industries from the Tesco group on or about 14 February 1991 and became the intermediate parent company of BAT Co.

Alsterufer Investments Ltd whose assets and liabilities including its rights to bring claims such as this one were transferred to the 4^th Claimant before liquidation in 2001/02.

In 1999 BAT plc acquired the tobacco business of the Rothmans group. In consideration for shares issued in BAT plc to the holding company of the Rothmans group (R&R Holdings a company with its seat in Luxembourg), BAT plc acquired through a wholly owned UK resident intermediate holding company the shareholding of Rothmans International BV, the Dutch holding company through which the worldwide Rothmans tobacco interests were held. Rothmans International BV was re-named British American Tobacco International (Holdings) BV (the 7^th Claimant) (“BAT International BV”).

In the period 1999 to 2000 the shareholdings in most of BAT’s interests worldwide were transferred to BAT International BV.

BAT International BV transferred its residence for tax purposes to the UK with effect from 14 January 2002. Prior to that migration in the period from July 2001 to January 2002 the holdings of BAT International BV were restructured to ensure that the dividends from trading subsidiaries which routinely paid dividends exceeding £5 million did not pass through intervening holding companies which held interests in trading companies in other States before those dividends reached BAT International BV.

The circumstances of the restructuring referred to in paragraphs 1.17 and 1.18 are set out in the first witness statement of Robert Fergus Heaton (27 March 2008).

British-American Tobacco Company (Nederland) BV is and was a Dutch resident holding company which was formerly owned by UK resident intermediate parent companies. In 1999 it became a subsidiary of BAT International BV as part of the restructuring referred to in paragraph 1.17.

Schedule 5 of the Eighth Amended Particulars of Claim lists other UK resident subsidiaries of the ultimate parent indirectly owned through UK resident intermediate parent companies which are relevant to claims of the claimants relating to losses surrendered and expenses applied in circumstances where the underlying liability to tax is challenged as unlawful.

The BAT group was parented in the UK by the ultimate parent and, under it, the UK resident intermediate parents. Schedule 1 to the Eighth Amended Particulars of Claim shows the financial years in which dividends were paid by non-resident subsidiaries to the UK resident intermediate parents, the recipient and paying companies, the country of residence of the paying company, the year in which the profits were declared which that dividend distributed and the amount.

Those dividend receipts attracted a liability to corporation tax under Schedule D Case V which was met either by the payment of tax, the utilisation of ACT, the surrender of losses by way of group relief or the use of expenses to reduce accounting profits. Schedule 2B shows for dividends received from companies resident in the EU/EEA and Schedule 2C for dividends received from companies resident outside the EU/EEA:

the recipient and paying company and the latter’s state of residence;

the amount of the dividend;

the underlying rate of tax on the profits and any withholding tax;

how the liability to tax under Case DV was met (whether by payment, surrender of losses, utilisation of ACT or otherwise) and in what amount

Schedule 2 of the Eighth Amended Particulars of Claim shows the distributions made by the ultimate parent to its public shareholders and the ACT incurred upon those distributions. It identifies the claimant which paid the ACT.

Schedule 2 to the Eighth Amended Particulars of Claim also shows the amount of ACT set off by the claimants and the balance which became surplus.

Schedule 2A shows some of the profits of EU resident subsidiaries in the BAT group together with the underlying tax paid upon them which the claimants contend should have been taken into account to mitigate against surplus ACT (see paragraphs 10 and 11 of the Eighth Amended Particulars of Claim).

Details of the amounts of ACT surrendered to subsidiaries and the amount of that surrendered ACT which was carried forward appear at Schedule 2 of the Re-Amended Particulars of Claim.

The Claimants contend that the inability of their shareholders to receive tax credits for dividends paid under the FID system caused the Claimants to “enhance” such dividends for shareholders for whom the tax credit was valuable by an amount equivalent in cash to the tax credit in the period 1994 to 1997.

Details of the payments made under the FID regime, the ACT paid and the date of its repayment and the amounts by which the Claimants say they enhanced those payments to compensate shareholders for the non-receipt of a tax credit are shown in Schedules 3 and 4 of the Eighth Amended Particulars of Claim and are as follows:

FID payments were also made following the abolition on 2 July 1997 of the ability of tax exempt shareholders to receive repayable credits with dividends. Those payments were made without enhancements as follows:

The FID of July 1994 had been matched in part against German sourced profits which were the subject of a tax refund in 1997. The effect of that tax refund was to reduce the underlying rate of tax on those matched profits and which meant that there was insufficient dividend income attracting double tax relief to match against the entire 1994 FID. In consequence the ACT reclaimed on 1 October 1995 exceeded that reclaimable which was agreed to form part of the reclaim for 1995. This required the Claimants to pay interest for the early receipt of the reclaim. On or after 15 July 1997 the Claimants paid to the Defendants £357,621.02 to meet that liability.”

The basic facts set out above were fleshed out in a number of witness statements. Five of the BAT group’s witnesses also gave oral evidence and were cross-examined by counsel for the Revenue. Those witnesses were:

The impressive credentials of these gentlemen speak for themselves, and I need hardly say that I found them all to be helpful and reliable witnesses. I will refer to their evidence as and when necessary when I come on to the various issues. At this stage, I will confine myself to describing in a little more detail how dividends were usually paid within the group and how the problem of surplus ACT developed.

In his third witness statement Mr Hardman described the ACT problem, as he termed it, as follows:

Mr Hardman went on to explain that the group’s accounting treatment of ACT recognised that it was, in effect, a permanent tax, and unrelieved ACT was written off through the profit and loss account and not recognised in the balance sheet as an asset, or even as a deferred tax asset. The build up of surplus ACT was illustrated in a schedule annexed to Mr Hardman’s statement, and he described the problem of surplus ACT as being “an unavoidable cost of international expansion”. The reason for this was twofold. First, as profits increased, the size of the annual dividend in respect of which ACT was payable also increased. Secondly, proportionately less and less of the group’s profits which made up the annual dividend came from UK trading activities, so there was proportionately less and less MCT against which to set the ACT. The result was that ACT took longer and longer to utilise. In the early 1970s, most of the ACT was utilised in the same year and only a small proportion was carried forward. By the end of that decade, the position was reversed and proportionately less and less of the ACT was being utilised in the same year. By 1986 ACT was taking nine years to utilise, whereas ten years earlier it had taken four years. By 1987 it was becoming permanently surplus. The position was then exacerbated in 1989, when the group’s response to a hostile takeover bid by Hoylake Investments (the consortium created by Sir James Goldsmith, Jacob Rothschild and Kerry Packer) involved the announcement of an increased dividend to shareholders and a share buy back scheme, each of which generated ACT liabilities. The group’s ACT bill therefore jumped from £59 million in 1988 to £115 million in 1989 and £236 million in 1990, before levelling out at an average of £135 million in the following two years. It was in this way that the surplus ACT mountain reached its peak of about £416 million in 1992.

Mr Hardman then explained in more detail why it was that ACT was usually incurred at the level of the ultimate parent company resident in the UK. The availability of group income elections meant that a group of companies like BAT could choose at what level within the group to pay ACT, following receipt of foreign income by a water’s edge company or the generation of taxable profits by trading within the UK. ACT would have to be paid at some stage before profits were distributed to the public shareholders of the ultimate parent, but the group could choose which UK company paid the ACT. A critical consideration in making the choice was the fact that ACT could only be surrendered to a subsidiary. It could not be surrendered to a parent company, nor could it be passed to a sister company within the group. Another relevant consideration was that a water’s edge company in receipt of foreign dividend income would seldom, if ever, be able to utilise the ACT it would have to pay if it did not enter into a group income election when paying a dividend up to its parent. It would have no UK subsidiaries to which the ACT could be surrendered, and double tax relief would severely limit the UK corporation tax liability of the water’s edge company against which ACT could be set. Accordingly, as Mr Hardman explained in paragraph 36 of his statement:

Mr Hardman went on to say that the only downside to paying ACT at the top company level and surrendering it down was that the recipient of a surrender of ACT could not carry it back, whereas a company which itself paid ACT could do so. However, this was not generally seen as a major consideration because group companies did not normally have unrelieved MCT in earlier years which could be used to absorb ACT generated in later years. In short, paying ACT at the level of the ultimate UK parent company was, generally speaking, the solution which gave the group the maximum flexibility in deciding how to utilise the ACT.

In cross-examination, Mr Hardman agreed that from the Revenue’s perspective the purpose of ACT was to fund the tax credits that were given to the shareholders (both exempt and non-exempt) who received the dividends, and that because the group paid very little MCT within the UK the bulk of the tax credits would have been unfunded by the group (and their cost would therefore have been borne by the general body of taxpayers) had ACT not been payable. He also confirmed that the BAT group did not attempt to control the conduct of its subsidiaries from the centre, and respected the corporate governance of all its subsidiaries around the world. It would therefore be for the board of the companies concerned to determine the level of dividends. The centre could set a policy, but it would be up to the individual companies to decide how to apply that policy. Similarly, matters were not organised so that the course of a single dividend could readily be traced as it made its way up the group. The position was much more complex than that, and there was no sequence of automatic bank transfers “from end market all the way through to the Plc on one single day”. Each board of directors would make its own decision as to what it was going to do with the dividend it received, and a tracing exercise would in practice be very difficult to perform because there were many tiers of companies and, the further up the hierarchy one gets, the more dividends will be received and mingled from various sources before onward transmission.

Most of Mr Bilton’s evidence was devoted to the introduction of the FID system, but he also described how the Schedule D Case V corporation tax liability of the water’s edge companies in receipt of foreign dividends was computed between 1990 and 1997. During those years he had prepared some of the computations himself, and overseen the preparation of others. He confirmed that in order to compute the liability account was taken of:

In cross-examination, Mr Bilton agreed that the water’s edge companies would have paid their dividends to their immediate UK parents under group income elections, with the result that the dividends received by the ultimate UK parent company were not FII.

The first main question that I have to decide arises from the decision of the ECJ on the issue whether the charge to tax under Case V of Schedule D (“Case V”) is compatible with Community law. As will appear, the outcome of this issue turns on the answer to a question which the ECJ left it to the national court to determine. Unfortunately, however, there is a fundamental disagreement between the parties about what that question actually is. Depending on how the question is interpreted, the answer to it is not (in the light of the expert evidence) controversial. If the question is interpreted in the manner for which the claimants contend, the result is that the Case V charge infringes Community law, while if the question is interpreted in the manner for which the Revenue contends, the result is that the charge is compatible with Community law.

The relevant question which the ECJ had been asked to answer was Question 1 in the Order for Reference, which reads as follows:

Translated into specifically UK terms, the focus of the question was therefore on the distinction between, on the one hand, the exemption from corporation tax afforded by ICTA section 208 for dividends received by a UK resident parent from a UK resident subsidiary, and, on the other hand, the charge to corporation tax under Case V on dividends received by a UK resident parent from a subsidiary resident in another member state, but with the benefit of double taxation relief for any withholding tax on the dividends and (subject to certain conditions) for the underlying tax paid on the profits which were distributed.

At the heart of the matter lies an important question of principle, namely whether (and, if so, subject to what conditions) Community law permits a member state to apply an exemption system for domestic-source dividends and a credit system for foreign-source dividends.

The distinction between such systems, and the existence of other possible solutions to the problem of economic double taxation which arises in the context of distribution of dividends, were lucidly explained by the Advocate General at the beginning of his opinion under the heading “Overview of context of dividend taxation”. Mr Aaronson QC for the claimants laid much emphasis on this background analysis as a key to the reasoning of the Advocate General and the ECJ on the first question, so I will start by citing it in full:

After setting out the relevant UK and Community legislation, the factual background and the questions referred, the Advocate General began his analysis by considering which of Articles 43 and 56 EC was applicable. He concluded (para 30) that the relevant UK legislation might in principle fall within the ambit of either article, depending on whether the parent company’s holding in the capital of the subsidiary established in another member state gives it “definite influence over the company’s decisions” and allows it to “determine its activities”. In such cases the parent company is exercising its right of establishment under Article 43, and it is therefore the compatibility of the UK legislation with Article 43 that needs to be assessed. In cases where Article 43 was not engaged, Article 56 (free movement of capital) would be (see para 32). The Advocate General proceeded on the footing (see para 31) that Article 43 clearly applied to all the BAT test claimants, because their non-UK resident subsidiaries were all wholly-owned, and although the exercise of their freedom of establishment in setting up the subsidiaries also inevitably involved the movement of capital out of the UK, that was a purely indirect consequence of the exercise of freedom of establishment. However, he also found it necessary to consider the question of compatibility with Article 56, because of the possibility that there might be other claimants within the FII GLO which had no more than an “investment” holding in a relevant subsidiary (para 33). I will need to return to this aspect of the matter later. It is enough to note at this stage that the Advocate General proceeded on the footing that both articles were potentially engaged.

The Advocate General’s discussion of Question 1 is to be found in paragraphs 36 to 56 of his opinion. His reasoning proceeds by the following main stages:

Having set the scene in this way, the Advocate General then described the competing arguments and stated his conclusions as follows:

The Advocate General then went on to consider foreign portfolio holdings (i.e. holdings of less than 10% of the voting power of the company paying the dividend), and concluded that the UK provisions relating to them were clearly discriminatory, because credit was given only for the foreign withholding tax and not for any underlying foreign corporation tax.

In order to understand the Advocate General’s conclusion, it is necessary to examine with some care the argument for the claimants which is recorded in paragraph 49 of his opinion, and to note the distinction between the nominal rate at which tax is imposed (e.g. 30% in the case of UK corporation tax for many of the years in question) and the effective rate at which tax is borne after available reliefs etc have been taken into account. A simple example should help to make the distinction clear. If a company has pre-tax accounting profits of £100, and the nominal rate of corporation tax is 30%, tax of £30 is prima facie payable; but the company may well have reliefs available to it, such as group relief or carried forward losses, which reduce the taxable amount (or “tax base”) from £100 to (say) £50. The tax charged will then be £15, and the effective rate of tax on the original profit of £100 will be 15%, not 30%.

One of the arguments advanced by the claimants to the ECJ contrasted the treatment, under the exemption and credit systems respectively, of dividends paid out of profits which had borne corporation tax (or its foreign equivalent) at an effective rate lower than the nominal rate. Suppose, for the sake of simplicity, that the relevant foreign tax is substantially similar to UK corporation tax, and levied at the same nominal rate of 30%. Suppose further that a UK-resident parent company has two subsidiaries, one resident in the UK and the other resident in a member state. Each subsidiary has a trading profit of £100, and bears corporation tax (UK or foreign, as the case may be) at an effective rate of only 15%. When the parent receives a dividend from its UK subsidiary, the dividend is exempt from UK corporation tax, so the benefit of the lower effective rate of (UK) corporation tax is preserved, and no further (UK) corporation tax is payable on the dividend. Accordingly, the overall corporation tax burden is only 15. By contrast, when the parent receives a dividend from the foreign subsidiary, the dividend is in principle subject to UK corporation tax at the full nominal rate of 30% with a credit for the foreign tax paid of 15. This will lead to payment of UK tax of 15, and an aggregate tax burden of 30, which is twice as much as the tax burden of 15 under the exemption system. The important point is that there is no way in which the benefit of the lower effective rate of foreign corporation tax can be preserved. Thus one way of looking at the overall result is to say that the effective rate of foreign corporation tax is “topped up” to the standard UK level of 30%.

It is clear, in my judgment, that the Advocate General accepted this argument, and agreed with the claimants that cases could arise where application of the credit system by the UK in relation to foreign-source dividends will produce a less favourable result than the pure exemption system applied to domestic dividends. As he said in paragraph 50, “the effect of this could be seen as the application by the UK of a different (lower) tax rate to domestic-source dividends than to foreign-source dividends”.

A point which the Advocate General did not mention, perhaps because it did not occur to him or more probably because he considered it irrelevant, is that the UK parent company will not necessarily pay corporation tax at an effective rate of 30% on its foreign-source dividends, because it may well have reliefs of its own available to it. It is therefore not strictly accurate to say that the aggregate tax burden will “always” be topped up to the standard UK rate. However, this possibility does not in my judgment detract from the force of the basic point, which is that under the credit system there is no way in which the benefit of the lower effective rate of foreign corporation tax can be passed up to the parent, whereas under the exemption system such benefit will automatically be passed up. In drawing this contrast, the fact that the parent company may pay corporation tax on its Case V dividend income at an effective rate of less than 30% is neither here nor there. The point may be tested by hypothesising a parent company which is a pure holding company with no sources of income apart from its foreign dividends, and no other reliefs available to it which would reduce its Case V tax base.

I now turn to the decision of the ECJ on the question, which is contained in paragraphs 33 to 74 of the judgment. For present purposes the critical part of the judgment is at paragraphs 43 to 57, which form part of the discussion of the position under Article 43. This passage needs to be read as a whole, and I will therefore set it out in full, omitting only those parts which refer to the Parent/Subsidiary Directive:

The wording of the conclusion in paragraph 57 of the judgment was then repeated in paragraph 73, and again in paragraph 1 of the formal ruling (or “dispositif”) at the end of the judgment.

The passage which I have quoted appears to me to present no particular difficulties down to the end of paragraph 53. In broad agreement with the Advocate General, the Court are saying that a member state may choose between a number of different ways of dealing with double taxation of dividends (para 43), but that whatever mechanism is adopted for this purpose the freedoms of movement guaranteed by the EC Treaty preclude a member state from treating foreign dividends less favourably than domestic dividends, unless the difference in treatment concerns situations which are not objectively comparable or is justified by overriding considerations of national interest (para 46). Each member state must organise its system for taxing distributed profits in compliance with Community law, and must in particular “define the tax base and the tax rate” which apply to the company making the distribution and the shareholder who receives it (para 47). The reference to the tax rate in this paragraph must be to the nominal rate of tax prescribed by the legislation, because that is the only rate which a member state can prescribe. There is no reason in principle why a member state should not apply an exemption system to domestic dividends and an imputation (or credit) system to foreign dividends (para 48). However, if an imputation system is to be compatible with Community law in such a situation, it must satisfy two conditions. First, the member state may not subject foreign dividends to “a higher rate of tax” than the rate which applies to domestic dividends (para 49). Secondly, credit must be given for the amount of tax paid by the distributing company against the amount of tax for which the recipient company is liable, up to the limit of the latter amount (para 50). Provided that those conditions are satisfied, it does not matter if tax is payable in the source country at a “higher level” than the tax levied in the home state, and there is no obligation to repay the difference (para 52). Similarly, the mere fact that an imputation system imposes administrative burdens on taxpayers is not an objection, because it is an intrinsic part of the operation of a tax credit system that evidence has to be provided of the amount of tax actually paid in the source state (para 53).

In my judgment it is natural to read the reference to “a higher rate of tax” in paragraph 49 as meaning a higher nominal rate of tax, both because that is the natural meaning of the term if it is used without qualification, and because it comes only a few lines after the reference to “the tax rate” in paragraph 47, where (as I have said) the reference must be to the nominal rate. I find some confirmation for this view in the fact that the original French text of the judgment uses the same term (“taux d’imposition”) in both paragraphs. Accordingly, if paragraph 49 is read in isolation, the only condition which it requires to be satisfied is that the member state should not subject foreign dividends to a higher nominal rate of tax than the rate which it applies to domestic dividends. Equally, if that is the relevant question, it is not disputed that the condition was satisfied in the UK throughout the period in issue. The nominal rate of corporation tax in the UK has always applied uniformly to a company’s taxable profits, and different rates of tax are not applied to income from different sources.

I should not disguise, however, that there is a minor difficulty in applying even the test laid down by paragraph 49, because domestic dividends are of course in general exempt from corporation tax by virtue of ICTA section 208, so they are not as such subjected to a rate of corporation tax at all. Fortunately, however, it is common ground that what the test is really directed at is the imposition of a nominal rate of tax on foreign dividends which is higher than the rate applied to taxable income from domestic sources, such as trading profits, out of which domestic dividends are paid.

It is at paragraph 54 of the judgment that the real difficulties begin. Paragraph 54 recites a further argument advanced by the claimants. It does so in very compressed form, but the argument evidently seeks to draw a distinction, in the domestic UK context, between the exemption accorded to dividends in the hands of the recipient and the fact that the company paying the dividend may have paid corporation tax at an effective rate which is lower than the nominal rate, or may even have paid no tax at all, because of the availability of various reliefs. The words “at a rate lower than that which normally applies in the United Kingdom” must in my judgment be meant to refer to an effective rate, and to contrast it with the nominal (or “normal”) rate. Again, I find some confirmation of this in the French text, which refers to situations where the paying company “n’est pas débitrice d’impôt ou paie un impôt sur les sociétés inferieur au taux nominal applicable au Royaume-Uni”. The contrast appears to be between the tax actually paid (“impôt”) on the one hand, and the nominal rate of tax (“taux nominal”) on the other hand.

So understood, this is in substance one half of the argument which was more fully recorded by the Advocate General in paragraph 49 of his opinion, the other half of it being that the UK credit system does not permit the benefit of a lower effective rate of foreign corporation tax to be passed up by a company resident in another member state which pays a dividend to its UK parent: see paragraphs 48 to 51 above.

Paragraph 55 of the judgment then says, and for convenience I will set it out again:

Small companies relief operates by imposing a lower nominal rate of corporation tax on profits below a certain level, with taper provisions which increase the rate towards the normal level where the profits fall within an intermediate band below the level at which the full rate begins to apply: see ICTA section 13.

In my view the argument attributed to the UK government in paragraph 55 of the judgment must have been meant to refer to the passage in the UK’s written observations which I have quoted above, even though those arguments were not directed to the contrast between nominal and effective rates of tax upon which the claimants relied, and even though it is only in the context of an argument about nominal rates of tax that one could sensibly describe the small companies rate as “highly exceptional”. There is obviously nothing exceptional about the proposition that companies often pay corporation tax at an effective rate lower than the nominal rate, because the existence and availability of reliefs which reduce the tax base (such as group relief, or the carry forward of trading losses) are commonplaces of UK corporate taxation. The UK government could not rationally have argued that such cases were exceptional. That, however, as I read paragraph 55, is the argument which the ECJ understood the UK to be advancing.

It also seems to me to be implicit in paragraph 55 that, had the ECJ been satisfied that differences between the nominal and effective rates of UK corporation tax were not highly exceptional, they would have agreed with the Advocate General that the exemption system for UK domestic dividends treated them more favourably than foreign dividends, and therefore breached Article 43. As I have already explained, the ECJ agreed with the Advocate General in their general approach to the question, and I cannot believe that they would then have dealt so briefly with the argument based on effective rates of tax, which he had accepted as determining the question in the claimants’ favour, if they thought it was either irrelevant or mistaken. If they thought either of those things, they would surely have explained why the argument was irrelevant or mistaken, or conceivably they would just have ignored it. Having introduced the argument into their discussion, however, without any adverse comments, I think the ECJ must be taken to have endorsed it, and the conclusion which flowed from it, subject only to the question of fact which they thought (wrongly, in my view) had been put in issue by the UK government.

If, as I think, that is the right way to interpret paragraphs 54 and 55, paragraph 56 then falls into place. Lacking the competence to decide questions of national law for themselves, and believing that there might be a dispute about the factual premises of the argument which the Advocate General had accepted, the ECJ left it to the national court to determine:

If these are the right questions to ask, there is no longer any dispute about how they should be answered. With regard to question (a), Mr Aaronson QC for the claimants accepts that the tax rates are the same, and does not argue that the existence of small companies relief makes any relevant difference. With regard to question (b), Mr Ewart QC for the Revenue accepts that UK companies frequently pay corporation tax at an effective rate which is lower than the nominal rate of corporation tax because of the availability of reliefs which reduce the tax base. If evidential support for the answer to question (b) is needed, it may be found in the expert report of Mr John Whiting OBE of PricewaterhouseCoopers LLP. After a detailed survey of the levels of corporation tax paid by UK companies as a percentage of their UK accounting profits over the years from 1973 to 2005, he concluded that:

Mr Ewart’s principal argument for the Revenue was that the ECJ did not agree with the Advocate General, and in the view of the Court the only conditions which needed to be satisfied by the UK in order to secure compliance with Article 43 were those stated in paragraphs 49 and 50 of the judgment. He submitted that the second question referred back to the national court in paragraph 56 was nothing to do with effective rates of tax as contrasted with nominal rates, but was merely asking whether the existence of small companies relief meant that different nominal levels of taxation occur only in highly exceptional circumstances. For the reasons which I have given, I am unable to accept these submissions. It follows that in my judgment the claimants succeed on this issue, and the UK system of taxation of dividends received from member states has at all material times infringed Article 43 EC.

I should say, finally, that I am not deterred from reaching this conclusion by the fact that the conclusions stated in materially identical terms in paragraphs 57 and 73 of the judgment, and in the dispositif, refer only to the “tax rate” (“taux d’imposition”) applied to foreign dividends, and do not reflect the second enquiry remitted to the national court in paragraph 56. If I have interpreted the judgment correctly, the conclusion is over-compressed; but since it purports to be no more than the conclusion which flows from what has gone before, I do not think that it can throw much light on the substance of the discussion which precedes it.

The next group of issues that I have to consider concerns dividends received by a UK parent company from a subsidiary which is resident in a third country, i.e. a country other than a member state of the EU. The basic question which arises is whether the subjection of such dividends to the Case V charge infringes Article 56 EC. There could be no question of infringement of Article 43, because that article protects freedom of establishment in member states alone and has no application to the establishment of subsidiaries in third countries. Article 56, however, prohibits restrictions on the free movement of capital not only between member states, but also between member states and third countries, although in its application to third countries it is subject to an important “standstill” proviso (now contained in Article 57 EC) which preserves the effect of restrictions under national or Community law which existed on 31 December 1993.

It will be convenient at this stage to set out the text of Articles 56 and 57:

Before I come on to the third country issues, I will first say a little about the overlap between Articles 43 and 56 in cases where the subsidiary which pays the dividend is resident in a member state. I have already held that in such cases the Case V charge, and the system of credit for foreign tax associated with it, infringes Article 43. I have also pointed out that the Advocate General proceeded on the footing that the first four questions referred to the ECJ could in principle fall within the ambit of either Article 43 or Article 56, depending on the size and nature of the parent company’s holding in the foreign subsidiary, but that Article 43 took priority over Article 56 in any case where Article 43 was engaged. In paragraph 34 of his opinion, the Advocate General said that “the substantive principles for analysis of whether a breach has occurred are the same for both articles”, although “the geographic and temporal scope of Article 56 EC differs from that of Article 43 EC”. He then explained that while Article 43 applied only to restrictions on the exercise of freedom of establishment between member states, and had entered into force as part of the Treaty of Rome, the restriction in Article 56 on the movement of capital between member states and third countries entered into force only on 1 January 1994, and was subject to the standstill provision in Article 57(1). The principle of free movement of capital within the Community had first been implemented by Council Directive 88/361/EEC of 24 June 1988.

In the context of the present case, it is common ground that an infringement of Article 43 would in principle also entail an infringement of Article 56 if the latter provision were applicable. As the Advocate General said, the substantive principles for determining whether a breach has occurred are the same under both articles. The matters relied upon by the claimants as constituting unlawful discrimination are in each case the same. The standstill provisions of Article 57(1) cannot apply, because no third country is involved. As the ECJ said in paragraph 184 of the judgment, the case-law of the Community establishes “that any less favourable treatment of foreign-sourced dividends in comparison with nationally-sourced dividends must be regarded as a restriction on the free movement of capital in so far as it is liable to make the acquisition of holdings in companies established in other Member States less attractive”. The cases cited as establishing that proposition include Verkooijen (Case 35/98, Staatssecretaris van Financien v B. G. M. Verkooijen [2000] ECR I-4071), paragraph 35, and Manninen (Case C-319/02, Proceedings brought by Manninen [2004] ECR I-7477), paragraph 23.

In cases where a UK parent company receives foreign dividends but Article 43 is not engaged, it is necessary to distinguish between two situations, namely:

The Advocate General and the ECJ discussed cases in category (a), because they were unsure on the facts whether all the relevant shareholdings of all the claimant companies within the FII GLO were of such a nature as to engage Article 43: see the opinion of the Advocate General at paragraph 33, and the reference to it in paragraph 38 of the judgment of the Court. My understanding is that all the relevant shareholdings within the GLO are in fact thought to be of such a nature that Article 43 applies to them. Certainly that is true of all the relevant companies within the BAT group. It may, therefore, have been unnecessary for the ECJ to deal with category (a) cases. On the other hand, the ECJ was not asked to deal with, or give guidance in relation to, category (b) cases, even though many of the claimant companies have subsidiaries in third countries from which they have received dividends. The reason for this omission is no doubt that the case-law of the ECJ relating to the circumstances in which a company with third country subsidiaries may rely on Article 56 has developed very significantly since the date of the reference in 2004. At that time, it was generally thought that reliance could not be placed on Article 56 in any third country situation where Article 43 would apply if the subsidiary in question were resident in a member state. According to the claimants, however, that is no longer the position.

Accordingly, the first question that I have to decide under this heading is whether it is in principle open to a UK company in receipt of dividends from a subsidiary resident in a third country to rely on Article 56. If it is, the next question is whether the protection of Article 56 is removed by Article 57, and (if so) over what periods. To the extent (if at all) that Article 56 does apply, and is not disapplied by Article 57, it is common ground that the answer to the question whether the relevant UK legislation infringes Article 56 will be the same as the answer to the question whether it infringes Article 43.

Mr Aaronson submitted that this question should be answered in the affirmative. He accepted that the position had been uncertain, in situations where the third country shareholding was of a nature which would engage Article 43 if the company paying the dividend were resident in a member state, i.e. (for short) where the shareholding was one which conferred control as opposed to an investment or portfolio shareholding. He submitted, however, that the recent decision of the ECJ in the case of Holboeck (Case C-157/05, Holboeck v Finanzamt Salzburg-Land, [2008] STC 92) had clarified the law, and established the proposition that Article 56 could be relied upon even where the companies concerned were in such a relationship, provided that the national legislation in question was not intended to apply exclusively to companies in such a relationship. An instance of national legislation of the latter type is the UK legislation relating to controlled foreign companies, recently considered by the ECJ in the Cadbury Schweppes case (Case C-196/04, Cadbury Schweppes and Cadbury Schweppes Overseas [2006] ECR I-7995). By contrast, said Mr Aaronson, the Case V charge applies to all foreign dividends regardless of the size and nature of the shareholding from which they are derived. Accordingly, Article 56 may be relied upon even where (as in the present case) the companies concerned are all in a group relationship, and are mostly (if not all) wholly-owned subsidiaries.

The relevant facts of Holboeck were simple. Mr Holboeck was resident in a member state, Austria, and he owned two thirds of the shares in a company resident in a third country, Switzerland. The dividends which he received from the Swiss company were taxed in Austria at the full income tax rate. He claimed that this tax charge infringed Article 56, because dividends distributed to natural persons by companies resident in Austria were subject to tax at only half the average rate. A reference to the ECJ was made by the Austrian court, asking whether Article 56 precluded national legislation to this effect.

The case was heard by the Fourth Chamber of the ECJ, presided over by Judge Lenaerts who was also the Juge Rapporteur. Judgment was delivered on 24 May 2007, the court having decided, after hearing the Advocate General (Y Bot), to proceed straight to judgment without an opinion of the Advocate General.

In paragraph 19 of the judgment, the ECJ recorded a submission by the French and Netherlands Governments “that the Austrian legislation can be considered only in terms of the rules regarding freedom of establishment, not of those regarding the free movement of capital”. If that submission were accepted, the result would of course be that Mr Holboeck would be unable to rely on Article 43 because of the territorial limitation on its scope. The ECJ then dealt with the point in paragraphs 22 to 24 of the judgment as follows:

In my judgment the passage from the judgment which I have quoted does indeed clarify the position in the way contended for by Mr Aaronson, and unless the judgment has itself been superseded or put in doubt by subsequent case-law of the ECJ it is now clear that Article 56 may apply in a case such as the present one, where the national legislation in question (here the Case V charge) applies “irrespective of the extent of the holding which the shareholder has in the company making the distribution” (paragraph 24 of the judgment), and where the exemption for domestic dividends likewise applies to shareholdings of all sizes.

It is interesting to note that the authority cited in paragraph 24 of the judgment includes three paragraphs in the judgment of the ECJ in the present case, which had been delivered a few months earlier. With the benefit of hindsight, those passages can indeed be read as providing strong support for the proposition enunciated in Holboeck; but read in the context of the FII judgment alone they cannot be regarded as determinative of the question, not least because the ECJ had not been asked to address the question of third country dividends. The importance of Holboeck is that the point was raised in a simple and stark form, and received a clear answer.

For the Revenue, Mr Ewart advanced two main arguments on this part of the case. First, he submitted that as matters stood after Holboeck the position was still unclear in cases where the national legislation in question is not targeted solely at group relationships, but where the facts of the case involve a group of companies. He said that if Holboeck were still the last word on the question, it would ultimately be necessary to make a further reference to the ECJ to clarify the position. Secondly, however, he submitted that the ECJ had returned to the question in a very recent judgment delivered on 26 June 2008, Finanzamt Hamburg-Am Tierpark v Burda GmbH (Case C-284/06, “Burda”). He argued that Burda confirmed for the first time that even where the national legislation was not targeted solely at group relationships, Article 56 was excluded where on the facts of the case the dispute related exclusively to the effect of the legislation on a group. He suggested that Holboeck was out of line with both earlier and later ECJ authority, and that it should now be treated as superseded by the fuller discussion of the question in Burda. He sought in various ways to marginalise Holboeck, pointing out that it was not referred to at all in Burda, and the ECJ had made its decision without the benefit of an oral hearing or an opinion of the Advocate General.

I am unable to accept these submissions. As to the first point, I do not agree that Holboeck left the position unclear. On the contrary, as I have tried to explain, I consider that it clarified the position and resolved earlier uncertainties. The fact that the ECJ reached its decision without an oral hearing and without an opinion of the Advocate General (although they heard him, as the preamble to the judgment records), does not in my view detract in any way from the authority of the case. If anything, it suggests that the Court regarded the case as a relatively straightforward one, covered by existing principles which needed no more than clarification and restatement.

As to the second point, Mr Ewart’s submission faces the obvious difficulty that Burda was not a case which involved a third country at all. It concerned a German company, which was owned in equal shares by a Dutch company and another German company: see paragraph 24 of the judgment. The factual background, in very simplified terms, was that the German company had made profit distributions to its two parents, and the question arose whether various set-off arrangements under German national law, which gave rise to a liability to German corporation tax, infringed Article 43 because their effect was to treat the Dutch parent company less favourably than the German parent. In particular, German law did not permit a non-resident shareholder to set-off the German tax liability against its own national tax liability, whereas a set-off of this nature was available to the German shareholder.

For present purposes, the relevant part of the judgment is contained in paragraphs 68 to 75, under the heading “The applicable freedom”. In view of the significance which Mr Ewart attached to this passage, I will quote it in full:

In my judgment it is clear that this passage is directed solely at intra-Community situations where Article 43 is capable of applying. The judgment reiterates that Article 43 takes priority, although there may be scope for Article 56 to apply in cases where Article 43 is not engaged (i.e. cases in my category (a) in paragraph 71 above). The facts of Burda did not involve any third country, and there was no reason why the ECJ should consider the applicability of Article 56 to shareholdings in third countries. Had they for any reason wished to do so, it is to my mind inconceivable that they would not have referred to Holboeck, which had been decided less than a year earlier by the same Fourth Chamber. Judge Lenaerts presided over both hearings, and four of the five judges in each case were the same. The omission of any reference to Holboeck in Burda shows, in my judgment, that the decision in Burda has nothing to do with the application of Article 56 to third country shareholdings, and neither expressly nor by implication does it cast any doubt on the principles stated in Holboeck.

For these reasons I accept Mr Aaronson’s submissions and hold that those claimant companies which have subsidiaries in third countries are entitled to rely on Article 56, but subject to the provisions of Article 57 to which I shall now turn.

The starting point here is that the Case V charge was already in existence on 31 December 1993. Indeed, in its basic form as a charge to income tax on income from foreign possessions it dates back to the nineteenth century. Furthermore, the claimants accept that the restriction on dividends received from third country subsidiaries imposed by the Case V rules constituted a restriction relating to direct investment for the purposes of Article 57(1), with the result that the application of Article 56 to such dividends from 1 January 1994 was disapplied by the standstill provision in Article 57(1). However, although the basic Case V charge remains the same today as it was in 1994 (albeit that the charge to tax on dividends from non-UK resident companies is now to be found in section 402 of the Income Tax (Trading and Other Income) Act 2005), there have been a number of changes to the rules associated with the charge since 31 December 1993. In their skeleton argument, counsel for the claimants submitted that two of those changes were so substantial as to constitute a new restriction not permitted by Article 57(1), namely:

It is fortunately unnecessary for me to deal with the introduction of the shadow ACT regime in 1999, because in the course of his reply Mr Aaronson abandoned this point. As he disarmingly said, it was very complicated, it was unlikely to find favour with the court, and if I were unwise enough to decide it in his favour the chances were that I would be reversed on appeal. However, Mr Aaronson maintained his submissions based on the introduction of the EUFT regime, and submitted that from 31 March 2001 onwards the Revenue could no longer rely on the Article 57(1) standstill.

Before I come on to the EUFT rules, it is first necessary to examine the jurisprudence of the ECJ on the meaning of the words “any restrictions which exist on 31 December 1993 under national or Community law” in Article 57(1). The ECJ has considered Article 57(1) in a number of judgments, including the present case, and in their skeleton argument counsel for the claimants submitted that two relevant principles have been established:

The question was considered by the ECJ in the present case, not in the context of the Case V charge, but in the context of their discussion of the FID regime (which was introduced with effect from 1 July 1994). The reason why the ECJ did not consider the question in relation to the Case V charge is that the Order for Reference had been framed on the assumption that Article 57(1) would apply, because the Case V charge had been in existence on 31 December 1993, and would thus exclude the application of Article 56: see paragraph 26 of the Order for Reference, which said that:

The relevant part of the judgment of the ECJ is contained in paragraphs 189 to 192, which read as follows:

Konle concerned the interpretation of Article 70 of the Act of Accession of Austria which allowed Austria to maintain its “existing legislation regarding secondary residences for five years from the date of accession”. The restrictive measure in question had been adopted after the accession of Austria, but replaced an earlier measure which had been in existence before Austria’s accession. The ECJ said that although the adoption of a measure after the date of accession was not in itself fatal to the application of Article 70, on the facts of the case the measure was sufficiently different from its predecessor to lead to the conclusion that Article 70 could not apply. The crucial paragraphs in the judgment are paragraphs 52 and 53:

It will be noted that both in Konle and in the present case the ECJ was considering new legislation which had been introduced after the relevant standstill provision came into force, and was addressing itself to the question whether the new legislation could properly be regarded as a continuation of a restriction which had been in existence on the earlier date. The Court was therefore not directly considering the question of whether, and if so in what circumstances, a restriction which existed at the standstill date might itself subsequently lose the protection of Article 57(1). Mr Ewart fastened on this distinction in order to submit that the principles stated by the ECJ in Konle and the present case are of no assistance in relation to the Case V charge, because that was a restriction which was already in existence on 31 December 1993. For his part, Mr Aaronson submitted that the same principles should apply in both situations, and that an existing restriction would lose the protection of Article 57(1) if it was subsequently amended in a way which was based on an approach which differed from that of the previous law and established new procedures.

Both sides referred me in this context to the decision of the ECJ in another recent case, C-101/05 Statteverket v A, delivered by the Grand Chamber of the Court on 18 December 2007. This was a reference for a preliminary ruling about the interpretation of Articles 56 to 58 EC, and it concerned the refusal of the Swedish tax authorities to grant to a Swedish-resident individual, A, an exemption from tax on dividends distributed in the form of shares in a subsidiary by a company established in a third country. In 1992 Sweden had introduced an exemption for such distributions made by Swedish companies, provided that certain conditions were satisfied. The exemption did not extend to distributions by non-Swedish companies, whether resident in the EU or in third countries. In 1994 the exemption was repealed, but in 1995 it was reintroduced in Swedish national law. There matters rested until 2001, when the exemption was extended to distributions by foreign countries resident in the EEA or in any state with which Sweden had concluded a double tax treaty containing a provision for exchange of information.

A owned shares in a Swiss company, and was contemplating procuring a distribution of the shares which it held in one of its subsidiaries. He applied for a preliminary decision on whether such a distribution was exempt from income tax. There was a dispute whether the relevant double taxation arrangements between Sweden and Switzerland fell within the scope of the 2001 legislation. A also contended that the limits on the exemption in any event amounted to a restriction on the free movement of capital which could not be justified. Without deciding the former question, the Swedish appeal court referred the matter to the ECJ for a preliminary ruling, asking whether it was contrary to Article 56

In his opinion, Advocate General Bot considered (among other matters) the question whether the restriction of the exemption to Swedish companies fell within Article 57(1) as a restriction which existed on 31 December 1993. He held that it could not be so regarded, because the restriction had been repealed in 1994 and then reintroduced in 1995. He said (paragraph 110) that Article 57(1) must be interpreted strictly, because it constitutes a derogation from the principle laid down in Article 56(1), and then continued in paragraph 111:

The ECJ agreed with the Advocate General that the wording of Article 57(1) presupposes that the restriction in question has formed part of the legal order of the member state concerned continuously since 31 December 1993: see paragraph 48 of the judgment, where the ECJ pointed out that, if this were not the case,

The ECJ disagreed, however, with the Advocate General on the conclusion which followed from application of this principle. The reasoning of the court is to be found in paragraphs 50 to 52, as follows:

This decision is in my judgment important for at least two reasons. First, the ECJ was examining a restriction which had been in place on 31 December 1993, and had to consider whether it still enjoyed the protection of Article 57(1). Secondly, the ECJ identified the relevant restriction, and answered the question whether it had remained continuously in force, by looking at the matter from the point of view of companies which were excluded from the benefit of the exemption even after the 2001 changes to the legislation. The inability of such companies to make tax-free distributions of shares in their subsidiaries had remained unchanged since 31 December 1993, so from their perspective the restriction had been continuous. The ECJ, differing from the Advocate General, clearly thought it irrelevant to this analysis that Swedish companies had themselves been unable to make such tax-free distributions in the interval between the repeal of the exemption in 1994 and its reintroduction in 1995. They must have thought it equally irrelevant that the scope of the exemption had subsequently been extended in 2001, thereby relaxing the original restriction to a corresponding extent. The important point, as it seems to me, is that they did not identify the relevant restriction and test its continuity by reference to the small print of the legislation as it stood on 31 December 1993, and as it was subsequently varied by repeal, reinstatement and amendment, but instead simply compared the position of a company in a third country which always fell outside the exemption with the position of a Swedish company to which the exemption (when it was in force) applied.

If that approach is applied to the facts of the present case, it seems to me that the relevant restriction must be identified as the exclusion of third country-source dividends from the exemption given to UK-source dividends by ICTA section 208. That exclusion was in place on 31 December 1993 and has remained in place ever since, so Article 57(1) applies and prevents a recipient of such dividends in a member state from complaining that the exclusion infringes Article 56. The fact that there may have been changes since 1993 in the Case V regime which subjects the dividends to tax in the UK is, on this analysis, irrelevant. Such changes might be relevant to the question whether Article 56 had been infringed, and (if so) to the quantification of the loss suffered by the claimant; but the changes are not relevant to the continued existence of the basic restriction itself, which is simply the inability to take advantage of the exemption in section 208.

Mr Aaronson submitted that the relevant restriction should be identified by reference to the Case V charge, and said that the Case V charge should not be viewed in isolation but in conjunction with the rules for relieving it by the grant of a credit for underlying tax. Accordingly, if there were a major change in the relief rules (such as, on his argument, the introduction of EUFT), the restriction would no longer be the same as it was on 31 December 1993 and the protection of Article 57(1) would be lost. I would see considerable force in this argument if it were not for the decision of the ECJ in Statteverket v A, and if the question which I am now considering had to be answered in the same sort of way as the question whether a new provision introduced after 1 January 1994 falls within the scope of an existing restriction. However, it appears to me that Statteverket v A requires a different approach to be adopted in identifying restrictions which existed on 31 December 1993 and deciding whether they have remained in force continuously since that date. If the approach for which Mr Aaronson contends were correct, I would expect the ECJ to have answered the question by reference to the legislative history, and to have reached the same conclusion as the Advocate General.

In the light of the conclusion which I have reached, the introduction of the EUFT regime for dividends paid from the end of March 2001 was irrelevant to the continuing application of Article 57(1). I will, however, proceed to consider it, as briefly as I can, in case I am wrong in my conclusion and the continuity of the Article 57(1) restriction instead falls to be tested by reference to the terms of the Case V regime and the criteria laid down by the ECJ in Konle and in paragraphs 189 to 192 of its judgment in the present case.

Before the introduction of EUFT, credit for foreign taxes was in principle given on a source-by-source basis, and was subject to an upper limit equal to the amount of UK corporation tax chargeable on the dividend. The relief extended both to any withholding tax charged when the dividend was paid by the foreign water’s edge company and (where, as is usually the case, the relevant double tax treaty so provided, as well as in all relevant cases of unilateral relief under ICTA section 790) to underlying tax on the profits out of which the dividend was paid, whether charged on the water’s edge company paying the dividend or on companies further down the chain. The detailed rules were contained in Chapter II of Part XVIII of ICTA, including in particular section 797 (the upper limit) and section 799 (underlying tax). It was a consequence of this system that, in principle, the underlying foreign tax on a dividend from a country with a higher rate of local corporation tax than the UK rate (“a high tax country”) would be capped at the UK rate, and any excess tax from one source could not be credited against the UK liability on a dividend from another source, such as a country with a local corporation tax rate lower than the UK rate (“a low tax country”).

However, the pre-EUFT rules permitted this disadvantage to be mitigated to a very considerable extent, because UK groups were free to arrange their affairs so as to pay dividends from low tax and high tax countries to the UK through a non-resident “mixer” company. A single dividend could then be paid to the UK from the mixer company with an averaged rate of tax, thereby minimising the loss of double tax credit. This process was commonly known as “offshore pooling”, and became standard procedure for most UK multi-national groups. Following the merger of the BAT group with Rothmans in the late 1990s, the claimants adopted this structure by arranging to pay dividends to the UK via a Netherlands holding company, namely the seventh claimant, BAT International BV.

The procedure which the BAT group adopted is described in more detail in the witness statement of Robert Fergus Heaton, who is a chartered accountant and was a senior tax manager reporting to Mr Hardman from 1997 onwards. He was not cross-examined on his written evidence, which I accordingly accept. He explains that prior to the abolition of ACT in 1999 the group held most of its interests in its end market trading subsidiaries directly from the UK, and maximising the use of excess tax credits from subsidiaries in high tax countries was not seen as a major concern, because the group’s liability to tax on Case V income was one of the few sources of corporation tax against which ACT could be utilised. However, the abolition of ACT for dividends paid after 4 April 1999, and the shadow ACT regime which replaced it, exposed the group for the first time to liabilities to tax on its Case V income which it would be unable to shelter with surplus ACT. Accordingly, the group was rearranged following the merger with Rothmans in the way which I have described. It was group policy that subsidiaries should, where possible, distribute 100% of their profits, and for the most part these would ultimately be received by BAT International BV. A calculation would then be made on a quarterly basis of the underlying rate of tax on the dividends received – which was by no means a straightforward exercise – and a decision would then be made on the amount of the dividends to be paid from BAT International BV into the UK. Mr Heaton would review the calculations, and also take into account any other reliefs that might be available in the UK (such as group relief) to shelter any liabilities to tax upon receipt of the Case V income which would not be covered by double tax relief. It was always group policy to pay dividends through the group and into the UK as quickly as possible, and Mr Heaton would only raise concerns about the proposals if it looked likely that there would be a significant adverse tax impact.

The EUFT system had a complex and prolonged gestation, which is described by Mr Heaton. On 21 March 2000 the Chancellor announced the introduction of legislation which would make it no longer possible for UK multi-nationals to mix high and low tax dividend income offshore in order to produce dividends carrying a rate of underlying tax similar to the UK corporation tax rate. A period of consultation ensued, in response to which the Revenue announced that amendments would be made to the draft rules and their introduction would be deferred until 31 March 2001. In June 2000, the Revenue then announced that, while the Government was committed to the abolition of offshore mixing, it would introduce a limited form of onshore pooling. Further lobbying continued, which led to further announcements of changes to the proposals in November 2000 and March 2001, and the promulgation of draft regulations concerning the surrender of EUFT at the end of March 2001. The new rules then took effect for dividends paid from the end of that month.

The EUFT rules are complex in detail, but the broad principles can be stated quite briefly. The advantages of offshore mixing were largely nullified by the introduction of a “mixer cap”, which operated to restrict the amount of underlying foreign tax which could be credited against the UK corporation tax liability on foreign dividends. The cap applied not only where a dividend was paid by a non-resident company direct to the UK, but also, and critically, where a cross-border dividend was paid at any earlier stage within the group by one non-resident company to another. The mixer cap limited the creditable underlying tax to the UK corporation tax rate. Any unrelieved foreign tax (EUFT) would then be eligible for onshore pooling, and could be offset against the UK corporation tax payable on certain dividends from low tax countries. However, the dividends against which EUFT could be offset (“qualifying foreign dividends”, or “QFDs”) excluded certain important categories of dividend, including in particular:

Within the BAT group, it was estimated that if nothing was done the effect of the EUFT rules would be to increase the group’s UK tax liability by at least £60 million per year. A complex restructuring operation was therefore undertaken, and given the name “Project Frankenstein”. In very broad terms, the purpose of the exercise was to undo the restructuring which had followed the merger with Rothmans, and to ensure that the separate streams of dividends from the end market subsidiaries did not mix before they reached the UK, and then to transfer the residence of BAT International BV from the Netherlands to the UK. Mr Heaton was the manager of the project, and it occupied him full time for six months. A team of external advisers was brought in to assist, and the overall cost of the exercise came to about £5 million.

Against this background, can it be said that the introduction of the EUFT regime changed the legislative approach upon which the Case V charge was based and established new procedures (compare Konle at paragraphs 52 and 53, and the judgment of the ECJ in the present case at paragraph 192)? The EUFT system certainly established new procedures for the relief of foreign tax, but I do not consider that it changed the general approach upon which the Case V charge was based. As before, the general philosophy of the system was to grant relief in the UK for foreign withholding and underlying tax attributable to the dividends received, subject to an upper limit fixed by reference to the UK corporation tax rate. All that changed were the detailed rules relating to the utilisation of foreign tax which would otherwise have been unrelieved, and the introduction of a new system of onshore pooling to replace the offshore pooling which no longer carried any fiscal advantages. In some respects the new system was less attractive to UK multi-national groups than the system which it replaced, but the new system also introduced important elements of flexibility which had not previously been available. If one stands back from the small print and the detail, the overall nature of the Case V charge was in my judgment still in substance the same after the introduction of EUFT as it had been before. It is true that the changes would have cost the BAT group a substantial amount in extra tax, if Project Frankenstein had not been undertaken to mitigate the position. But the reorganisation was no greater than that which followed the merger with Rothmans and the abolition of ACT, and the claimants no longer contend that the abolition of ACT itself forfeited the protection of Article 57(1). Moreover, nobody suggests that a mere increase in the nominal rate of UK corporation tax would have been fatal, so the increased tax cost of the changes to a typical UK multi-national could not of itself be determinative. The test is essentially a structural one, and in my view the focus should be on the main features of the structure rather than the fine points of detail. Accordingly, if it were necessary for me to decide the question, I would hold that the introduction of the EUFT system did not cause the protection of Article 57(1) to be lost.

The result of the foregoing discussion, in my judgment, is that the third country Case V claims fail in their entirety.

The next group of questions that I have to consider concerns the ACT regime, and the second and third questions in the Order for Reference.

In general terms, Question 2 in the Order for Reference was concerned with the contrast between the ACT regime applicable to domestic UK dividends and the absence of any comparable regime applicable to dividends received from non-resident companies. The question was posed in the following terms:

The contentions to which Question 2 relates are now pleaded in paragraphs 2 to 11D of the Ninth Amended Particulars of Claim. The claimants’ basic complaint focuses on the stage when dividends are finally paid outside the group by the ultimate UK parent company to its individual and institutional shareholders, both in the UK and abroad. At that stage, no group income election could be made (as a matter of UK domestic law) and ACT was in principle payable, from its inception in 1973 until its abolition in 1999. However, ACT was only payable to the extent (if at all) that the UK parent company did not have FII available to it to frank the dividends. Such FII would be available in respect of dividends which the UK parent company had itself received from its UK-resident subsidiaries, and in respect of which ACT had already been paid. By contrast, however, dividends which had been received from foreign subsidiaries could never generate any FII, because a tax credit under ICTA section 231, and hence FII, could arise only in respect of distributions made by companies resident in the UK. Furthermore, this would be the case even if foreign corporation tax had been paid by the foreign subsidiary which paid the dividend.

The key contention in paragraph 11A of the Ninth Amended Particulars of Claim is framed as follows:

Paragraph 11A(b) then went on to claim that, while the ACT regime provided a mechanism for groups which were largely UK resident to mitigate the effect of ACT, no mechanism for mitigation at all was made available to UK parented groups whose subsidiaries were not UK resident. Where the UK holding company and its subsidiaries had liabilities to UK corporation tax, then ACT would not normally represent an absolute cost to the group as the ACT would be treated as a prepayment of corporation tax. Where, however, the profits of the group were largely earned by subsidiaries resident outside the UK, and those profits were distributed to the UK holding company, the ACT could not in practice be offset and was not refundable, even where the profits in question arose from a source within an EU member state. In such circumstances, the UK holding company had to bear the ACT “as, effectively, an absolute and permanent cost”.

For these reasons, it was alleged that the ACT provisions provided a disincentive for UK companies to establish subsidiaries in other EU member states or to move capital or make payments between the UK and subsidiaries wherever situated.

The ECJ dealt with Question 2 in paragraphs 75 to 112 of the judgment, and answered the question as follows in paragraph 112:

In the light of this answer, the Revenue admit in their Amended Defence that there is discrimination within the meaning of Articles 43 and 56 to the extent that a resident company which has received dividends from another resident company may deduct the amount of ACT paid by the latter company from the amount of ACT for which the former is liable, whereas a resident company which has received dividends from a non-UK resident company within the EU is not entitled to make such a deduction in respect of the corporation tax which the latter company is obliged to pay in its state of residence: see paragraph 11(a). The Revenue also admit that ACT was unlawful and unduly levied in circumstances where a resident company received a dividend from an EU-resident subsidiary and was obliged to account for ACT when distributing that dividend to its own shareholders, to the extent that the ACT would not have been payable had a deduction been allowed in the amount of the corporation tax levied on the underlying profits distributed by the latter company plus the withholding tax levied on that distribution in the source state: see paragraph 13(a). The Revenue expressly deny, however, that there is any breach:

It is these contentions which give rise to the questions which I must now consider, and which the parties have, for convenience, compendiously dubbed the “corporate tree” points. In short, the issue is this. The ECJ has clearly held that ACT is unlawfully levied where the water’s edge foreign company which pays the dividend has itself paid foreign corporation tax on its distributed profits, and where the water’s edge UK company which receives the dividend has itself paid ACT when paying the dividend on to its immediate parent company. But does the same result follow:

The claimants contend that both these questions should be answered in the affirmative, with the result that it does not matter (to revert to the tree metaphor) how far down the foreign roots of the corporate tree the underlying profits have borne foreign tax, or how far up its UK trunk and branches ACT has been paid. In financial terms, the question is a very significant one. If the Revenue are right, the claimants’ victory on Question 2 in the ECJ will have been a hollow one, because I was told that it is very rare for the twin conditions of payment of foreign tax by the foreign water’s edge company and payment of ACT by the UK water’s edge company to be satisfied. Indeed, those conditions have never been satisfied at any time within the BAT group itself. On the other hand, if the claimants are correct the maximum claim under this head (including compound interest going back as far as 1973) would be of the order of £1.2 billion.

The core reasoning which led the ECJ to reach the conclusion which I have already quoted is to be found in paragraphs 82 to 93 of the judgment. It proceeds by the following stages:

Mr Aaronson submitted that the reasoning of the ECJ depended on three fundamental principles which it laid down in rejecting the arguments advanced by the UK government. The first principle is that ACT, from the perspective of Community law, is simply an advance payment of UK corporation tax. The second principle is that local corporation tax paid overseas is equivalent to corporation tax paid in the UK. The third principle is that once corporation tax is paid or payable in respect of certain profits, no further corporation tax is to be paid on those profits as they are distributed up the chain. It is this third principle which avoids a succession of charges and which, according to the ECJ, is reflected in the UK domestic system of corporate taxation.

Mr Ewart agreed that there could be no dispute about Mr Aaronson’s first principle. He also accepted that his second principle was correct in certain contexts, including the present one. Mr Ewart took issue, however, with Mr Aaronson’s third principle, and argued that it was not correct. He submitted that the true principle is set out in paragraph 55 of the judgment of the ECJ in another case which was delivered on the same day as the judgment in the present case, Test Claimants in Class IV of the ACT Group Litigation v IRC, Case C-374/04, [2007] STC 404 at 445:

Mr Ewart submitted that the unequal treatment in the present case only arose where the dividend paid by the UK subsidiary was FII. In such a case, the UK subsidiary would be liable for ACT because the payment would be a “qualifying distribution”: see ICTA sections 238(1), 231(1) and 14(1) and (2). That is why the ECJ regarded the UK system as providing for a deduction for ACT paid by the UK subsidiary. Where, however, the UK subsidiary was “exempt” from ACT on the dividend because it was paid under a group income election, the dividend was not FII: see ICTA section 247(2). In those circumstances, the UK parent did not receive any relief from ACT, and it was irrelevant if ACT had been paid on an earlier dividend in the chain. Only dividends actually received by the UK parent could count as FII in its hands. Furthermore, ACT could not be surrendered up a group, but only to a subsidiary.

Mr Ewart went on to submit that the ECJ assimilated the ACT paid by the UK subsidiary with foreign corporation tax paid by the foreign subsidiary. Where the UK subsidiary was exempt from ACT (because of the making of a group income election), the UK parent would not receive any credit against its ACT liability. Accordingly, where a foreign subsidiary was exempt from corporation tax on its profits, there was no difference in treatment if its UK parent did not obtain any credit against its liability for ACT when it received a dividend from that subsidiary. In the absence of any difference in treatment, there could be no breach of either Article 43 or Article 56.

Mr Ewart further submitted that there was no breach of Article 43 in not giving credit against ACT for foreign tax paid by any company other than the immediate foreign subsidiary, because there was equally no credit given domestically (through the ACT rules) in such a situation. The claimants’ argument amounted to saying that there ought to have been a system equivalent to the system for giving credit for foreign tax against UK corporation tax, rather than a system equivalent to the ACT regime. That argument was flawed, because the claimants are complaining about the incompatibility with Community law of the domestic system of giving credit against ACT. The only relevant comparison, says Mr Ewart, is with the domestic ACT system.

In evaluating these submissions, it is in my view necessary to keep firmly in mind the distinction between two different aspects of the UK ACT regime. The first aspect is the availability of FII to frank distributions made by a company. The second aspect is the ability within a group to make a group income election.

FII first arises within a group where a UK-resident company has made a distribution and actually paid ACT. According to the ECJ, the payment of ACT is a payment of corporation tax in advance, albeit at a rate lower than the mainstream rate and on an amount calculated by reference to the amount of the dividends. It is that payment of tax which confers, in the eyes of Community law, a corresponding exemption from ACT upon the recipient of the distribution, provided that the recipient is also another UK-resident company. The exemption is conferred by treating the amount of the distribution and its associated tax credit as FII, which is then set off against the franked payments made by the recipient company when it in turn makes onward distributions: see ICTA sections 238(1) and 242(1). If a company has surplus FII, it can carry it forward to subsequent accounting periods (see section 241(3)) or use it in various other specified ways. One thing that a company has never been able to do with surplus FII, however, is to surrender it to its parent. FII can only be generated in the hands of a company which has itself received a dividend which in principle triggered a liability to ACT in the hands of the paying company. However, if the same dividend is passed up a UK group, the ACT paid at stage one will in effect frank the ACT which becomes payable at each subsequent stage, with the end result that ACT is paid only once within the group.

By contrast, a group income election enables dividends to be passed up within a group without triggering the payment of ACT, and without generating any FII: see in particular ICTA section 247(2). A group income election is, as its name suggests, a matter of choice for the companies concerned. It enables, but does not compel, the companies within a group to choose the stage at which ACT will be payable. So, for example, as Mr Hardman explained in his evidence, the standard (but not invariable) practice within the BAT group has been to arrange matters so that dividends were passed up within the group under group income elections and ACT became payable only at the final stage when dividends were paid to external shareholders by the ultimate UK holding company. In such a situation, there has been no prior payment of ACT, and no FII is generated within the group at all.

Against this background, the critical question, it seems to me, is whether the reasoning of the ECJ should be read as confined to the avoidance of a series of charges to tax in situations where ACT has actually been paid by the UK company which receives the foreign dividend, or whether the reasoning should be read as applying more broadly to all cases where a combination of the rules relating to FII and the ability to make group income elections in practice enables the group to ensure that ACT is paid no more than once. It is only the former type of situation which involves an actual charge to ACT which is subsequently matched by a corresponding exemption with the group in the form of FII. On the other hand, both scenarios may reasonably be seen as reflecting a general policy of avoiding, or at least mitigating, the economic double taxation of distributed UK profits, and it may be thought artificial to look at the ACT regime without having regard to the ability to make group income elections which is, in practice, such an important part of it.

One difficulty in dealing with this question is the fact that, as I understand it, both the written and the oral arguments before the ECJ focused on the simple situation (probably hardly ever replicated in practice) where a foreign dividend is paid out of profits which have borne corporation tax in the paying company’s state of residence, and is received by a UK company which is itself liable to ACT when it distributes the dividend to its immediate UK parent. It is easy to understand why this course should have been adopted in the interests of simplicity and ease of exposition, but it did have the unfortunate effect, it seems to me, of obscuring the corporate tree points with which I now have to grapple. For example, the ECJ was never expressly asked to consider what difference (if any) it would make to its analysis if the UK company which received the dividend did not itself pay ACT, but instead made a group income election when it paid its own dividends, or if the foreign company which paid the dividend did not itself suffer corporation tax on the underlying profits out of which the dividend was paid, but corporation tax had been paid further down the corporate chain.

In these circumstances I think it would be unwise for me to speculate about how the ECJ would have answered hypothetical questions which were never put to it, and I should instead begin by looking at the precise terms of the question which they were asked and their answer to it.

Question 2 in the Order for Reference, which I have quoted in paragraph 111 above, is in my view concerned only with situations in which FII arises, and where that FII is used to frank dividends paid by the recipient company. That follows from the reference to a member state which keeps in force and applies “measures which provide for the resident company to pay dividends to its shareholders without being liable to pay ACT to the extent that it has received dividends from companies resident in that Member State”. That is a description, in suitably abstract terms, of the way in which the FII system operates. The words that then follow in brackets, “(either directly or indirectly through other companies resident in that Member State)”, show that the UK company concerned cannot be the bottom company in the UK hierarchy, because such a company cannot itself have UK subsidiaries; but subject to that, the UK company may be at any level within the group. The contrast that is then drawn is with the inability of the company which pays the dividends and incurs liability to ACT to obtain relief from ACT “to the extent that it has received dividends from non-resident companies”. The comparison therefore envisages that the company will have received the foreign dividend itself, and the implicit complaint is that no FII is available to the company in respect of the foreign dividend, although the company is entitled to FII in respect of dividends which it has received from UK subsidiaries. The question says nothing at all about group income elections, and the focus seems to me to be entirely on situations where liability to ACT is in principle triggered by the payment of dividends, either because no group income election is possible or because the parties have chosen not to make one.

Read in the context of the question which they were asked, and the submissions which were addressed to them, it seems equally clear to me that the answers given by the ECJ in paragraphs 94 and 112 of the judgment, the latter of which I have quoted in paragraph 116 above, are likewise confined to situations where the UK company is itself in receipt of FII. Furthermore, the comparable situation which involves a breach of Articles 43 and 56 requires two separate conditions to be satisfied. The first condition, foreshadowed in the language of Question 2, is that the UK company should itself have received a foreign dividend, but without the benefit of any corresponding FII. The second condition is that the company is not entitled to a deduction from its own ACT liability “in respect of the corporation tax which the last mentioned company is obliged to pay in the State in which it is resident”. These last words are in my judgment clear and unambiguous, and they focus attention solely on corporation tax which the foreign company paying the dividend has been obliged to pay in its own state of residence.

It must follow, in my judgment, that if no corporation tax has been paid in the state of residence of the company which pays the foreign dividend, the discriminatory treatment identified by the ECJ cannot have occurred. If local corporation tax has been paid further down the foreign corporate chain, that would undoubtedly be relevant in computing the credit for foreign tax which is to be allowed against the Case V charge in the hands of the recipient UK company; but it is far from obvious, to my mind, that the payment of such tax is also relevant to the entirely separate question whether the UK ACT regime contravenes Articles 43 and 56. It will be remembered that the ACT regime only permits FII generated by dividends received from immediate subsidiaries to be credited against ACT falling due when an onward dividend is paid. As I have already pointed out, the system does not permit FII generated any further down the line to be set against ACT. Accordingly, the ACT regime contains no obvious equivalent to the double taxation relief afforded for underlying tax on dividends paid by lower-tier companies.

In short, it seems to me well arguable that Mr Aaronson’s third principle elegantly begs the question in the claimants’ favour by assuming that one should assimilate the double taxation rules on credit for underlying tax with the much simpler ACT regime. Mr Aaronson submitted that this must be what the ECJ had in mind in paragraph 91 of the judgment; but read in context I think that the focus of that paragraph is much narrower, and is merely making a comparison between cases where both the company paying the dividend and the company receiving it are subject to corporation tax. This conclusion is reinforced by paragraphs 92 and 93, in each of which an argument advanced by the UK government is rejected because it disregards the fact that the foreign company which pays the dividend is subject to corporation tax in its state of residence (my emphasis). Nothing is anywhere said about corporation tax payable at any previous stage.

Mr Aaronson also suggested that support for his interpretation could be found in the opinion of the Advocate General, particularly at paragraphs 68 to 75. The discussion in those paragraphs is somewhat compressed, perhaps partly because the Advocate General (unlike the ECJ) decided to consider Questions 2 and 3 together. However, I cannot find any clear indication in them that he was considering anything more than the contrast between foreign distributing companies which were themselves liable to corporation tax in their state of residence on the one hand, and UK companies which were themselves liable to ACT on the other hand. He clearly had group income elections in mind (see paragraph 66, where he refers to them in the context of Hoechst), but the less favourable treatment which he identified in paragraph 72 appears to me to focus only on the contrast which I have mentioned. Similarly, when he said in paragraph 73 that “[d]ue to the grant of the corporate tax credit, ACT only had to be paid once in the “chain” of distribution”, he was not in my view thinking of group income elections, but rather referring to the situation where the same profits are paid up a UK chain of companies without any elections being made, so that at each stage after the first one a liability to ACT arises but it is covered by a corresponding amount of FII.

For these reasons, I conclude that Mr Ewart’s submissions are substantially correct in their analysis of what the ECJ has actually decided, and the judgment establishes only that there is a breach of Articles 43 and 56 in situations where:

In my view it cannot safely be deduced or inferred from the Court’s limited reasoning what its answer would be to the corporate tree points, and having myself heard detailed argument on the question I certainly do not regard it as so clear that I can confidently provide the answer myself. Accordingly it seems to me that a further reference to the ECJ on the corporate tree points will, unfortunately, be necessary. My regret at reaching this conclusion is, however, tempered by the fact that both sides agree these questions to be very important, and over £1 billion may turn on their resolution.

My own view, which I express briefly and with considerable hesitation, is that as a matter of Community law Mr Aaronson’s submissions are to be preferred. The governing principle of relieving or mitigating economic double taxation of dividends is a powerful one, whereas the corporate tree points are technical in nature and depend, it may be said, on a rather literal and blinkered approach to the ECJ’s judgment. In the light of the governing principle, the situation expressly considered by the ECJ may be seen as a paradigm one, and a similar line of reasoning should in my judgment lead to the conclusion that there is also a breach of Articles 43 and 56 in situations where:

As regards situations of type (a), the UK rules on double taxation relief provide a credit for such underlying tax, so if one accepts the ECJ’s premise that the ACT system has the object of relieving economic double taxation, it seems logical to compare it with all factual situations where double taxation relief would in principle be available to the UK company which receives the foreign dividend. It is true that the UK rules for relief of double taxation are more sophisticated than the ACT regime, but the benefit of a payment of ACT can in practice be passed up a group in such a way that ACT is paid once only in respect of the same underlying profits. That provides a close enough analogy, to my mind, to justify a comparison with situations where underlying tax is paid by a lower-tier company and a higher-level company subsequently pays a dividend with the benefit of a credit for that underlying tax, even though it does not pay corporation tax itself.

As regards situations of type (b), the ability to make group income elections only applies in situations where ACT would otherwise be payable, and its basic function is to enable the group to decide at what level in the corporate hierarchy ACT is to be paid. In my view the ability to make such elections cannot sensibly be segregated from the rest of the ACT regime, because it serves the same basic purpose of ensuring that ACT is paid once only within the group in respect of the same profits. It should not therefore matter, in the present context, whether the company which receives the foreign dividend itself pays ACT, or whether it pays the dividend to its parent under a group income election. Equally, it should not matter whether or not the company which receives the dividend happens to have FII available to it from other sources. The important point is that in all cases the foreign dividend is incapable of generating any relief from ACT in the hands of the recipient company, yet it will inevitably generate a liability to ACT at some point before it ultimately leaves the group. Viewed from a group perspective, therefore, it will give rise at some stage to an unrelieved charge to ACT within the group, in addition to the foreign tax which it has already borne at source. By contrast, UK profits which are paid up through the group are in practice liable to ACT once only, and they are then liable to a reduced charge to MCT because the ACT discharges an equivalent part of the MCT liability at the level at which it is paid. In other words, the former situation involves full economic double taxation, whereas the domestic UK regime does not.

Whether the ACT provisions of UK domestic law are incompatible with Article 43 only in the limited factual circumstances expressly identified by the ECJ, or in the much wider circumstances contended for by the claimants, the question then arises of how that incompatibility should be remedied under English law, so as to give full effect to the directly effective rights of the claimants which have been infringed.

The basic principles which should guide the national court are not in doubt. If at all possible, the offending national legislation should be construed or interpreted so as to make it “conform to the superior order of EU law”, to use the vivid phrase of Lord Walker of Gestingthorpe in Fleming v Revenue and Customs Commissioners [2008] UKHL 2, [2008] 1 WLR 195, at paragraph 25. If, however, a conforming construction is not possible, and if the national legislation infringes directly effective Community rights, “the national court is obliged to disapply the offending provision” (ibid, paragraph 24). As Lord Walker went on to explain, the provision of national law is not made void, but it has to be treated as being (in the words of Lord Bridge of Harwich in R v Secretary of State for Transport, ex parte Factortame Limited [1990] 2 AC 85 (“Factortame No. 1”), at 140) “without prejudice to the directly enforceable Community rights of nationals of any Member State of the EEC”.

I need not spend more time on the concept of conforming construction, and the difficulties to which it gives rise, because neither side suggested that the ACT provisions in UK domestic legislation could be construed in a way which achieved conformity with Article 43. It is enough to say that I agree with that view. I will, however, say a little more about the concept of disapplication, which has been developed in a number of important cases both in the ECJ and in the English courts.

The seminal decision of the ECJ was in the case of Simmenthal (Amministrazione delle Finanze dello Stato v Simmenthal S.p.A., Case C-106/77, [1978] ECR 629), where in paragraphs 20 to 24 the Court emphasised the duty of national courts to “set aside” any provision of national law which conflicted with directly effective Community rights, or which might “impair the effectiveness of Community law”.

In the domestic context, the speech of Lord Bridge in Factortame No. 1 (loc. cit), and the subsequent speech of Lord Nolan (with whom the rest of their Lordships agreed) in ICI v Colmer (Inspector of Taxes) (No.2) [1999] 1 WLR 2035 at 2041, indicate that the process of disapplication is not a mechanical, blue-pencil exercise, but rather requires the court to assume the incorporation of a proviso in the offending legislation stating that its provisions are to be without prejudice to the directly effective Community rights of the complainant. Depending on the circumstances, the effect of such a notional proviso may be achieved by striking out certain words, or by a process of remoulding or adapting the statutory language to the necessary extent (see the speech of Lord Nicholls of Birkenhead in Autologic Plc v IRC [2005] UKHL 54, [2006] 1 AC 118 at paragraph 17, with whom both Lord Steyn (paragraph 47) and Lord Millett (paragraph 61) concurred).

Mr Aaronson submitted that the simplest and most appropriate way to achieve compliance in the present case would be to remove the UK territorial limitation on FII, so that it included foreign dividends received by a UK company from a company resident in a member state. This could be done by striking out the words “resident in the United Kingdom” in ICTA section 231(1), which confers a tax credit on the recipient

I emphasise that these alterations do not, in any sense, involve an amendment of the UK statute, which is something only Parliament can do, but are simply a convenient way of expressing how the statutory language is overridden by the superior order of Community law in cases where the Article 43 right of establishment of a company resident in a member state is engaged. The statutory authority in English law for performing this exercise is, of course, to be found in section 2 of the European Communities Act 1972, as interpreted by the House of Lords in cases such as Factortame No. 1 and ICI v Colmer.

Alternatively, Mr Aaronson admitted that the same result could be achieved by reading in a proviso to the charge to ACT in ICTA section 14, so as to exclude from its ambit any qualifying distribution to the extent that it represented dividends received from a company resident in a member state.

In my judgment either of these methods would achieve the objective of ensuring that foreign dividends from member states are treated in the same way as domestic dividends for ACT purposes, and like Mr Aaronson I slightly prefer the first method. It seems to me more in accord with the structure and philosophy of the UK partial imputation system to confer a tax credit on the relevant foreign dividends and then treat them as FII in the hands of the recipient rather than to exclude them from the scope of the charge to ACT, although the practical result, so far as I can see, is in each case the same.

Mr Ewart argued, however, that each of Mr Aaronson’s solutions would run the risk of over-compensating the claimants, and that the only directly effective right of the UK parent company was to pay ACT after deducting the actual corporation tax paid by the foreign subsidiary. He submitted that the UK legislation was incompatible with Community law only in so far as it did not grant a credit for foreign corporation tax against ACT, and that the ACT charge should be disapplied only to that extent.

Mr Ewart’s submission gains some initial support from the terms in which the ECJ answered Question 2 in paragraphs 94 and 112, and also perhaps from paragraph 90 of the judgment where there is a reference to the UK company being deprived of “the opportunity to reduce the amount of ACT which it is obliged to pay on a subsequent distribution by way of dividend”. Nevertheless, I consider that Mr Ewart’s proposed solution is wrong in principle. Under the UK system ACT is charged, and a tax credit is conferred on the recipient of the distribution, at a fixed rate, regardless of the actual amount of corporation tax which is in fact paid on the distributed profits. This will be so even if, for example because of the availability of various reliefs, no corporation tax is paid at all. What matters is that the distributed profits are in principle liable to corporation tax in the hands of the company making the distribution, not the amount of tax which is actually paid. By parity of reasoning, therefore, the right way to bring foreign dividends into the ACT system is to confer a tax credit which will generate FII, at the same rate as that applicable to UK dividends, and not merely to allow a deduction from ACT for the foreign corporation tax which has been paid.

There are, in addition, great practical difficulties with Mr Ewart’s solution, at any rate if the claimants are correct on the corporate tree points. The dividends in issue go back, at least potentially, as far as 1973, and continued to be paid until 1999. The claimants never sought to match dividends paid by the UK ultimate parent to its shareholders with foreign income, nor did they trace the flow of dividend income within the group in such a way that its component parts could be separately identified, nor did they ever seek to calculate an ACT charge after deduction of foreign tax on the matched dividends. They had no reason to do so under the provisions applicable at the time, and they submit that it would be contrary to the Community principles of legal certainty, protection of legitimate expectations and effectiveness to impose such requirements upon them now, with retrospective effect, as a condition for succeeding with their claims. In my judgment there is much force in this submission, and I am satisfied that it would in practical terms be impossible for any kind of accurate reconstruction to be made of the course that foreign dividends followed through the group over a period of some 25 years. Accordingly, even if I thought the Revenue’s solution was the correct one in principle, I am very doubtful whether it would now be possible to implement it in a way which was fair to the claimants and satisfied the important principles of Community law which I have mentioned. However, I do not propose to spend longer on this subject because, as I have said, I consider that Mr Aaronson’s solution is in principle the correct one to adopt.

This is a curious question which arises from the answer given by the ECJ to Question 3 in the Order for Reference. Question 3 had two limbs. The second limb, which I need not set out, asked in effect whether it was contrary to Community law for the UK legislation to permit the (downwards only) surrender of surplus ACT to other UK members of a group, but also to provide double taxation relief for foreign corporation tax which would reduce the liability to MCT against which the ACT liability could be set. The ECJ discussed this question in paragraphs 118 to 128 of the judgment, and answered it in favour of the UK government. Their conclusion in paragraph 127, repeated in paragraph 138, was that Article 43

The first limb of Question 3 asked in effect, with reference to the same national provisions for the surrender of ACT, whether it was contrary to Community law for the UK not also to provide for the set off of surplus ACT (or some equivalent relief) in respect of profits earned, whether in the UK or in other member states, by non-UK resident group companies. The precise wording of the question, framed as usual in Community rather than UK-specific terms, was as follows:

The thrust of the question, therefore, was whether it was open to the UK to confine the surrender of ACT to UK-resident members of the group, or whether some equivalent relief should have been granted to group companies resident in other member states in respect of their profits. The question expressly envisaged a possible distinction between profits earned by such a foreign subsidiary (a) in the UK, and (b) in other member states, including most obviously its own state of residence. In practice, a foreign subsidiary would only earn profits in the UK if it carried on business in the UK through a branch or agency. Although perfectly possible in theory, such a situation is unlikely to arise in a UK-based multi-national group, because the group would in the normal way arrange for its UK trading activities to be carried on by its UK subsidiaries. I was told that there has never been any instance in the BAT group of a foreign subsidiary carrying on a branch business in the UK, but there may be cases in other groups within the FII GLO where this has occurred.

The ECJ discussed this question in paragraphs 129 to 134 of the judgment, and answered it as follows in paragraph 139:

It is in my judgment clear from that answer that the ECJ considered Article 43 to have been breached only where a non-resident subsidiary was taxable in the UK on profits which it had made in the UK. That is the force of the final words “where the latter are taxable in that Member State on the profits which they made there”. Accordingly, it would appear that the ECJ did not consider Article 43 to be breached in the much commoner situation where a UK company wished to surrender ACT to a foreign subsidiary so that it could be set off against the latter’s foreign profits.

That this limitation on the scope of the ECJ’s conclusion was deliberate is made abundantly clear by the discussion which preceded it in paragraphs 129 to 134. The whole of that discussion is clearly directed to the situation where the foreign company is itself liable to corporation tax in the UK: see for example the final words of paragraph 129 (referring to “the corporation tax for which they are liable in the United Kingdom”), and the references in paragraphs 132, 133 and 134 respectively to corporation tax “in that Member State”, “in the UK” and “in the Member State concerned”. In each context, it is clear that the reference is to the UK alone.

The ECJ nowhere deals with the question of set off of ACT against foreign profits, even though that question was clearly raised by Question 3, and even though that is the factual situation in which the question is in practice most likely to arise. The explanation for this otherwise surprising state of affairs is in my judgment to be found in paragraph 115, where at the start of their discussion of Question 3 the ECJ said this:

In the light of that statement, it seems clear to me that the ECJ proceeded on the footing that the question of set off of ACT against foreign profits was no longer in issue. I am satisfied, however, that this was not the case. The claimants, in their written observations submitted to the ECJ, did not limit the scope of Question 3 in this way, and it would be surprising if they had. There is no transcript of the oral hearing before the Court on 29 November 2005, but a detailed note prepared by the claimants’ solicitors shortly after the hearing (and based on written notes taken during the hearing) again gives no indication that any such concession was ever made. In his oral submissions to me, Mr Aaronson submitted that the ECJ must have misunderstood the claimants’ position, and he suggested that the mistake might have occurred when, for the sake of emphasis, he stressed to the ECJ during the oral hearing that no set off of ACT was permitted by the UK legislation even where the foreign subsidiary traded in the UK through a branch and actually had UK profits of its own. In other words, he was relying on that rare factual situation to bring into high relief the unfairness and illogicality (as the claimants would have it) of the UK system.

For his part, Mr Ewart did not dispute that Mr Aaronson kept open and advanced the argument that ACT should be available for set off against foreign profits, and he did not submit that the understanding of the ECJ as recorded in paragraph 115 of the judgment was correct.

In these circumstances, I cannot escape the conclusion that a misunderstanding must have occurred, and that it is for this reason alone that the ECJ did not consider the question of set off of ACT against foreign profits. What, then, is to be done? Unsurprisingly, each side sought to persuade me that the answer to the question was so clear that I should decide it here and now. However, as with the corporate tree points, I am satisfied that there is enough doubt about the answer to justify, and indeed require, a further reference to the ECJ. I can see that, if the claimants succeed on the corporate tree points, there would be strong grounds for saying that they should succeed on this question too. The relevant underlying considerations are very similar, and the UK system may again be seen as indefensibly insular from a Community point of view. On the other hand, if the claimants were to fail on the corporate tree points, I find it hard to envisage circumstances in which they would succeed on this question. It is also relevant to bear in mind that the written observations submitted by the Commission of the European Communities on Question 3 argued in favour of the conclusion which the ECJ actually reached, while the questions put to counsel at the end of the oral hearing by Judge Lenaerts appear to have been specifically directed to the question whether ACT could be set against the UK taxable profits of foreign subsidiaries. I therefore regard the question as an open one, which only the ECJ can resolve.

I have already given a brief description of the FID regime, which came into force on 1 July 1994, in paragraphs 12 and 23 to 25 above. The information contained in those paragraphs is, in substance, what the ECJ was told about the system in the Order for Reference.

The governing legislation in ICTA section 246A to 246Y was long and complex, but it is unnecessary for me to refer to it in any great detail. A helpful thumbnail sketch may be found in Moores Rowland’s Yellow Tax Guide for 1995/96, (Butterworths), Part I, Commentary on Statutes, pp120-122, to which I was referred by Mr Aaronson. I will, however, mention some aspects of the legislation upon which Mr Aaronson placed particular emphasis.

The starting point is that where a company resident in the UK paid a dividend in cash, it could elect that the dividend should be treated as a FID: section 246A(1) and (2). An election could not be made unless it was made in respect of all the dividends on the same class of share; and where a company had more than one class of share capital, ignoring any fixed-rate preference shares, it could only elect for a dividend to be a FID if it paid a dividend at the same time, and on the same terms, in respect of each share of each class, and opted for each of those dividends to be a FID: section 246A(4) to (9). It follows from these provisions that a company could not elect, for example, that a dividend payable on a particular class of shares should be a FID for recipients who were liable to UK income tax, but not a FID for exempt shareholders. An election had to be made on an all or nothing basis. This point is of central importance to the issue of enhancement of FIDs, to which I will come later in this judgment in my discussion of remedies.

The next important point is that an election had to be made to an inspector in an approved form no later than the time when the dividend was paid, and the election was then irrevocable after the dividend was paid: section 246B(1). The company therefore had to commit itself to the FID regime before the dividend was paid, even though in many cases the company would not yet know for sure whether it had sufficient distributable foreign profit to match with the dividend.

A FID did not carry a tax credit, nor was it treated as a franked payment made by the company: sections 246C and 246E. A FID received by an individual shareholder, personal representatives of a deceased individual or trustees of a discretionary trust was, however, treated as income which had borne income tax at the lower rate, although no claim could be made for repayment of any such tax: section 246D.

Because a FID did not carry a tax credit, it could not generate FII in the hands of a corporate recipient. However, a company had to pay ACT only on the excess of the FIDs which it paid over the FIDs which it received in an accounting period: section 246F(1) and (2). Furthermore, any excess of FIDs received over FIDs paid could be carried forward to the next accounting period and was then treated as a FID received in that period: section 246F(3). These provisions therefore produced a similar effect, within the FID regime, to the franking of franked payments by FII, and ensured that ACT had to be paid once only on a FID within the group.

The rules for the matching of FIDs with distributable foreign profit were contained in sections 246J to 246M. “Distributable foreign profit” was defined in section 246I as, in effect, the amount of the company’s foreign source profit after deducting the amount of foreign tax payable in respect of that profit or (if less) the amount of corporation tax payable in respect of the foreign profit before double taxation relief. The matching rules were relatively generous, and enabled a FID to be matched with distributable foreign profit of any subsidiaries in the current or preceding accounting period. An unmatched FID of the parent could also be matched with distributable foreign profit of the subsidiary in a subsequent period.

Sections 246N, 246P and 246Q contained the provisions for repayment or set off of ACT which had been paid on FIDs during an accounting period. The drafting of these sections was exceedingly dense, but in very broad terms they provided for the repayment to the company of whichever was the lesser of:

The basic structure of the provisions can be seen from the first four subsections of section 246N:

Question 4 in the Order for Reference asked, in effect, whether the FID regime was compatible with Community law. The question was framed as follows:

The ECJ dealt first with Article 43, and concluded in paragraph 164 that “Article 43 EC precludes a regime having the characteristics of the FID regime described by the national court in Question 4”. Their discussion of the question in paragraphs 144 to 163 began as follows:

The ECJ then went on to deal with various arguments advanced by the UK government. The first argument was that the differences identified by the Court did not involve any restriction on freedom of establishment, because the ACT regime applicable to domestic dividends was not comparable with the receipt of foreign dividends, because in respect of the latter no ACT had been accounted for. This argument was essentially the same as the one which the UK had unsuccessfully relied upon in relation to Question 2, and it received the same short shrift in paragraphs 152 to 154:

Another argument advanced by the UK government was that the absence of a tax credit under the FID regime could be justified by the fact that no ACT had been accounted for on foreign-source dividends, and the fact that the ACT which was paid on making a FID was subsequently repaid. Again, the response of the ECJ was firm:

The ECJ then proceeded to consider the compatibility of the FID regime with Article 56, in the context of foreign dividends received from companies resident in third countries. They held, in short, that the position was essentially the same as under Article 43, and that a point relied on by the UK government about the additional difficulty of obtaining information from companies resident in third countries as compared with companies resident in a member state was insufficient to justify the difference in treatment. The FID regime therefore breached Article 56 as well as Article 43.

The answer given by the ECJ to Question 4 raises the same corporate tree points, mutatis mutandis, as their answer to Question 2 in relation to ACT. Subject to one point, neither side suggested that the corporate tree questions involved any different considerations, or should receive different answers, in one context rather than the other. The only difference is that Mr Aaronson did briefly suggest, in his opening submissions, that the claimants were in an even stronger position in relation to FIDs than they were in relation to ACT, because the FID regime itself provided for matching of FIDs with distributable foreign profit arising in lower-tier companies. However, Mr Ewart pointed out, correctly in my view, that this distinction did not advance the claimants’ case, because the FID regime, like the treatment of foreign dividends outside or before the FID regime, still had to be compared with the basic UK regime for domestic dividends. In other words, the domestic comparator was in each case the same.

In the circumstances, I would give the same answers to the corporate tree questions in the context of the FID regime as I have already given to the corresponding questions in the context of the ACT regime.

I now move on to a more substantial question, which is whether, on the assumption that Article 56 applied and was breached in relation to third country FIDs, the introduction of the FID regime was a new restriction such that the UK could no longer rely upon the standstill provision in Article 57(1). I use the expression “third country FIDs” as a convenient, although slightly inaccurate, shorthand for FIDs which are matched with distributions made by companies resident in third countries.

This part of my discussion assumes that the ECJ’s conclusion that the FID regime breached Article 56 in relation to third country FIDs is correct, and has not been overtaken by its subsequent decision in Burda. On that aspect of the matter, I have nothing to add to what I have already said in paragraphs 80 to 95 above.

I will also not repeat what I have already said in paragraphs 88 to 98 about the jurisprudence of the ECJ in relation to Article 57(1). It will be recollected that where the question is whether new legislation introduced after 31 December 1993 can properly be regarded as a continuation of a restriction which was in existence on the latter date, the relevant test, as laid down in Konle and the judgment of the ECJ in the present case, is whether the new legislation is based on an approach which is different from that of the previous law and which establishes new procedures. If there is such a difference of approach, coupled with the establishment of new procedures, the new legislation cannot be regarded as “in substance identical” to the previous restriction. However, a member state will be able to continue to rely on an existing restriction where the new legislation “is limited to reducing or eliminating an obstacle to the exercise of Community rights and freedoms in the earlier legislation”. In other words, a relaxation of an existing restriction will not remove protection from what remains of the restriction. The principles which I have just identified were stated by the ECJ in paragraph 192 of the judgment, in the context of their discussion of Question 5 in the Order for Reference. I have already quoted part of Question 5 in paragraph 89, but for completeness I will now quote it in full:

The ECJ did not answer this question themselves, but remitted it to the national court for determination in the light of the guidance given in the judgment. The only remaining part of the judgment which I need to refer to in this context comes immediately after paragraph 192:

The important points which emerge from paragraphs 193 and 194 are in my judgment as follows. First, the object of the FID regime was to limit the restrictive consequences of the former legislation for UK resident companies in receipt of foreign dividends. If that were all that the FID regime did, there would be no reason why the relaxed restriction should not continue to enjoy the protection of Article 57(1). Secondly, however, shareholders who receive a FID are not entitled to a tax credit, even though the company pays ACT in the usual way after the dividend has been paid. This is prima facie a new restriction, based on an approach which is different from that of the previous law. However, it may be capable of justification on the basis that ACT is reimbursed to the company. Whether it can be so justified will depend on a detailed examination of the relevant legislation, which only the national court can perform. A relevant consideration, highlighted by the ECJ in the final sentence of paragraph 194, is that the company is only entitled to be reimbursed surplus ACT, whereas the shareholders in receipt of FIDs are not entitled to any tax credit.

Drawing on this guidance, Mr Aaronson submitted that the FID legislation was clearly based on a new approach, and the denial of any tax credit to recipients of FIDs could not be justified because the company was only entitled to reimbursement of surplus ACT after any remaining liability to MCT had been discharged by set off. He referred in this context to the provisions for repayment and set off of ACT which I have described in paragraphs 171 to 172 above, and submitted that the FID regime could only have continued to enjoy the protection of Article 57(1) if, as a minimum, it provided for the grant of tax credits in respect of any part of the ACT which was not reimbursed.

Mr Ewart’s riposte to these submissions was to argue that the question remitted to the national court by the ECJ in paragraph 194 of the judgment had been conclusively answered in the Revenue’s favour by the decision of the House of Lords in Pirelli Cable Holding NV v IRC [2006] UKHL 4, [2006] 1 WLR 400 (“Pirelli 1”). That case was concerned with the quantification of the compensation payable to certain companies under the ACT GLO, and one of the issues which arose was whether, on the hypothesis that a UK company and its foreign parent were able to make a group income election, and thus to avoid payment of ACT on dividends paid by the former to the latter, the foreign parent would nevertheless still be entitled to receive tax credits under the double taxation agreement entered into between its state of residence and the UK. The House of Lords, differing from Park J and the Court of Appeal, answered this question in the negative. They held, in essence, that there was an inextricable link between payment of ACT and the receipt of a tax credit, and that this principle should inform the construction of the relevant provisions in the double taxation agreement in such a way as to preclude payment of the tax credit in circumstances where ACT was not payable.

The link between payment of ACT and the receipt of a tax credit was expressed in varying ways, but to similar effect, by their Lordships. For example, Lord Nicholls said in paragraph 1 at 402G:

Accordingly, submitted Mr Ewart, if the original liability to pay ACT under the FID regime was subsequently extinguished by repayment of the ACT, no tax credit could have been payable, and in denying payment of a tax credit to the recipient of a FID the legislation was merely reflecting that basic principle.

Mr Ewart sought to derive further support for his submission from the decision of the Court of Appeal in the next instalment of the Pirelli litigation, Pirelli Cable Holding NV v Revenue & Customs Commissioners (No. 2) [2008] EWCA Civ 70, [2008] STC 508 (“Pirelli 2”). For present purposes, however, I do not think that the judgments of the Court of Appeal in Pirelli 2 add anything to the basic principle of linkage between the payment of ACT and the grant of a tax credit established by Pirelli 1. The focus of Pirelli 2 was on a different point, upon which the claimants had fastened following their defeat in Pirelli 1, namely whether entitlement to a tax credit was also triggered by the payment of MCT by the dividend-paying subsidiaries. That contention was rejected by Rimer J at first instance, and then by the Court of Appeal. I was told that the House of Lords subsequently refused permission to appeal, so that chapter in the litigation is now closed.

In my judgment Mr Ewart’s argument does not meet Mr Aaronson’s contention, nor does it answer the question remitted to the national court by the ECJ. A central point of that question, although it is easy to miss it on a superficial reading, is that the FID regime denies any tax credit, even though it does not provide for automatic reimbursement of the ACT which is paid, and instead gives priority to setting off ACT which would otherwise be repayable against any outstanding liability to MCT. In an extreme case, the result could be that no ACT at all was repayable, because the company concerned had a sufficiently large outstanding MCT liability to absorb all of the ACT paid in respect of the FIDs. Mr Ewart had no real answer to this point, and in my view this feature of the FID regime has to be regarded as the introduction of a new restriction which had no precursor in the previous legislation. By introducing this new restriction, I consider that the UK forfeited the protection of Article 57(1) for the FID regime, even though the general purpose of the FID regime was to mitigate the adverse ACT consequences of the regime otherwise applicable to foreign dividends.

For these reasons, I conclude that the FID regime breached Article 56 in relation to third country FIDs, and that the breach was not negated by Article 57(1). It follows that liability in respect of third country FIDs is in principle established.

I hope that I have now dealt with all the outstanding issues relating to liability which were argued before me, and it is time to move on to the question of remedies. Before doing so, however, I will first summarise the position on liability as it has now been established by the judgment of the ECJ and (subject to any appeal) by the preceding sections of my judgment.

I have held that the Case V charge has at all material times infringed Article 43 in relation to dividends received from member states (paragraph 66 above). However, I have held that there is no infringement of Article 56 in relation to dividends received from third countries, because although Article 56 is in principle engaged it has at all material times been disapplied by Article 57(1) (paragraphs 85 and 109 above).

The ECJ has already held that the ACT regime infringes Articles 43 and 56, at least in situations where a UK company receives a dividend from a subsidiary resident in a member state and the circumstances are such that:

On the corporate tree points, I have held that there needs to be a further reference to the ECJ (paragraph 138 above), but expressed my own tentative view that as a matter of Community law the claimants’ arguments ought to succeed (paragraphs 139 to 141 above).

I have also held that, in circumstances where Article 43 is infringed, the appropriate way in which to secure compliance with Community law is to treat the foreign dividend as though it carried an entitlement to a tax credit and thus generated FII (paragraphs 150 to 153 above).

On the ACT surrender point, the ECJ has held that the ACT regime infringes Article 43 in that it does not permit the surrender of surplus ACT to non-UK resident subsidiaries which are themselves liable to corporation tax in the UK because they have UK trading profits (paragraph 158 above). I have held that the question whether the inability to surrender surplus ACT for set off (or equivalent relief) against the foreign profits of a foreign subsidiary also infringes Article 43 is an open one, which only the ECJ can resolve, with the consequence that on this point too a further reference will be necessary (paragraph 163).

The ECJ has held that the FID regime breached Articles 43 and 56 in relation to dividends received from other member states.

The same corporate tree points arise in relation to FIDs as in relation to the ACT claims, and my conclusions on them are therefore the same (paragraph 179 above).

In relation to FIDs received from third countries, I have held that Article 56 was breached and the UK is not entitled to rely on the protection of Article 57(1), with the result that liability under this head is in principle established (paragraph 191 above).

Against this background, the relief claimed by the claimants in the Ninth Amended Particulars of Claim falls under a number of headings which are compendiously summarised as follows in the prayer for relief:

The detailed contentions which lead to the above claims are set out over some 10 pages of the Ninth Amended Particulars of Claim, with supporting schedules. What follows is intended as no more than a brief summary, sufficient to explain the general nature of the issues of principle on liability which I have been asked to decide at this stage of the litigation.

Restitution is claimed in respect of ACT payments made from 1973 until 1999, and also in respect of ACT payments made and repaid under the FID regime from 1994 until 1999. It is alleged in paragraphs 15A and 15B that the claimants made these payments under a mistake, and that the Revenue were unjustly enriched at their expense:

Restitution is then claimed in paragraph 15C, principally of:

Paragraph 15D alleges that the purported use of surplus ACT against unlawful Case V corporation tax liabilities was of no effect, and that the claimants are now entitled to restitution in respect of such surplus ACT.

There is then a further or alternative claim (in paragraph 15F) for compensation for losses alleged to have been suffered “as a direct result” of the breaches of the ACT provisions. The first category of losses so claimed covers substantially the same ground as the restitution claim, but the second category is new and is said to arise (paragraph 15G):

By a yet further amendment made during the course of the hearing on 7 July 2008, a mistake-based restitutionary claim was added in respect of the disclaimed capital allowances.

This part of the claim relates to enhancements which were made to the FIDs paid to shareholders in 1994, 1995 and 1996, with the alleged object of compensating exempt shareholders for the lack of the repayable tax credit which they would then have received in respect of an ordinary dividend. The enhancements, set out in schedule 3, totalled £389 million. The claim is pleaded on a restitutionary basis, or alternatively as a claim for compensation. The restitutionary claim alleges that the enhancements were paid by the ultimate parent company

The first claim under this head is a restitutionary claim for payments of Case V corporation tax which were actually made. The alleged mistake, and the alleged unjust enrichment, are the same, mutatis mutandis, as in relation to the ACT claims.

A further restitutionary claim is then made in respect of surplus ACT, loss relief and/or other expenses, losses or reliefs which were set off against the unlawful Case V tax liability. The mistake which is relied upon for the purposes of this claim is pleaded in alternative ways. First, the tax itself was unlawful, so the purported use of the ACT etc to offset it was ineffective, and the surplus ACT, losses, reliefs and expenses remained unaffected. Secondly, if the surplus ACT etc was in fact validly used, and became unavailable for alternative use, then the relevant mistake was that the claimants were lawfully obliged to satisfy the Case V liability. Either way, restitution is claimed of the amounts of tax thereby satisfied by mistake, and also of the time value of those amounts from the dates of payment until the date of judgment.

The alternative compensation claim relies upon the same use of the surplus ACT etc as having occasioned loss to the claimants. There is also an alternative claim, in respect of the losses and reliefs so used, that they continue to be available to the claimants, or (if they have been extinguished in satisfaction of an unduly levied tax) that Community law requires their reinstatement.

The Revenue’s response to these claims is contained in paragraphs 28 to 65 of the Amended Defence, and again I will provide a brief summary in order to show what are the main points in issue.

It is not admitted that the claimants made any mistakes of law, or that they acted upon any mistakes in making the ACT payments. It is denied that the Revenue were unjustly enriched, save to the very limited extent that liability is admitted earlier in the pleading.

If and so far as the Revenue were initially unjustly enriched, a defence of change of position is then advanced. This is an important topic, to which I will return in a later section of this judgment. I will not at this stage set out the way in which it is pleaded, or the particulars relied upon to support it.

It is also denied that the claimants are entitled to compensation, whether as pleaded or at all, although the point is taken in relation to one of the restitutionary claims (for surplus ACT set against unlawful Case V corporation tax liability) that the claim “is properly characterised as a claim for compensation rather than restitution” (paragraph 38).

On the facts, it is denied that the ultimate parent company was in any way “required” to increase the value of the distribution payments made to its shareholders as FIDs. It is not admitted that the FIDs paid were in fact enhanced, or that the claimants would not otherwise have distributed such sums to shareholders. The restitutionary claim is denied, on the footing that there was no unjust enrichment of the Revenue, and the alleged mistakes of law are again not admitted. Change of position is again pleaded as a defence, and the particulars given in relation to the ACT claims are repeated. The alternative claim for compensation is denied.

Two preliminary points are taken in relation to this part of the claim. The first point is that the claim should be stayed as an abuse of process on the procedural ground that it is within the exclusive jurisdiction of the Special Commissioners: see the decision of the House of Lords in Autologic Holdings Plc v HMRC [2005] UKHL 54, [2006] 1 AC 118. The second point is that the claimants’ common law claims for restitution and/or compensation in respect of monies paid pursuant to a mistake of law are excluded by the statutory provisions for recovery of tax overpaid in section 33 of TMA 1970. The first of these points was not pursued before me at the hearing. The second point is one that I will need to deal with in due course: see paragraphs 432 to 439 below.

The restitutionary claim is again denied, and the alleged mistakes of law are not admitted. Change of position is again pleaded, and the same particulars are repeated. With regard to the claims in respect of utilisation of surplus ACT, losses, expenses and other reliefs, it is pleaded that their use was in all instances valid and effective. Further, no claim lies in restitution, because any tax paid by the claimants against which they would otherwise have used these reliefs (i.e. if they had not been used by way of set off against the Case V liability) was lawfully levied and not itself paid under any mistake of law. The claim under this head is, if anything, a claim for compensation, in respect of which the claimants must satisfy the requirement of sufficiently serious breach.

The alternative claim for compensation under this head is also denied.

Questions 6 to 9 in the Order for Reference dealt with remedies. At this stage, the most important questions are Questions 6 and 8. Question 9 dealt with the issue of sufficiently serious breach, in the context of the compensation claims, to which I will come in a later section of this judgment. Question 6 asked whether, in the event of any breach of Community law being established under Questions 1 to 5, “in circumstances where the resident company or other companies in the same group of companies” made certain specified types of claim, which were then listed,

The types of claim specified in Question 6 were listed in nine numbered sub-paragraphs. I will not set them out in full, partly to save space and partly because they have to some extent been overtaken by subsequent amendments to the pleadings. In summary, however, they are as follows:

Question 7 asked some further questions in the event that the answer to any part of Question 6 was that the claim was a claim for payment of an amount representing a benefit unduly denied. Question 8 then asked:

The Advocate General dealt with Questions 6 to 9 together. He began by setting out some important principles of Community law, which are well-established and not in dispute:

The authorities cited by the Advocate General in paragraph 126 go back to the landmark decision of the ECJ in the San Giorgio case (Case 199/82, Amministrazione delle Finanze dello Stato v SpA San Giorgio, [1983] ECR 3595). It will be convenient to refer to this principle as the San Giorgio principle, and to claims within the scope of the principle as San Giorgio claims.

The Advocate General went on to discuss the decision of the ECJ in Hoechst, where the Court had emphasised that it was not its task to assign a legal classification under English law to the actions brought by the claimants before the national court, but was for the affected companies to specify the nature and basis of their actions – whether for restitution or for compensation in damages – subject to the supervision of the national court (Hoechst at paragraph 81). He noted that no questions had been referred by the national court in Hoechst in respect of the interpretation of the Factortame general conditions for state liability for damages, and the Court had not considered whether those conditions were made out. Advocate General Fennelly had considered the issue briefly, but only in the alternative because his view was that it was “more correct and more logical to treat the plaintiffs’ claim as restitutionary rather than as a compensatory claim for damages”.

The Advocate General then stated his conclusions in relation to the present case, in a passage which I should cite in full:

The most significant point which emerges from these paragraphs, in my judgment, is the potential width which the Advocate General was prepared to attribute to the San Giorgio principle. He clearly regarded it as capable of extending not only to the recovery of sums of tax unlawfully paid, but also to “equivalent” claims for reparation of financial loss. For this purpose, a claim will be equivalent if the relief claimed is a “direct consequence” of the unlawful levying of tax. The Advocate General evidently considered that all the types of claim listed in Question 6 might satisfy this test, with the exception of the claim for the FID enhancements. However, he expressly left it to the national court to determine in each case whether the relief claimed was in fact a direct consequence of the unlawful tax charge. Finally, in relation to claims which do come within the San Giorgio principle, he stressed that it is the duty of the national court to provide an effective remedy, even in a situation where a Factortame claim for damages might not succeed.

I now turn to the judgment of the ECJ, the relevant part of which begins at paragraph 197. Like the Advocate General, they dealt with Questions 6 to 9 together. The Court began by recording the claimants’ argument that each of the claims referred to in Question 6 should be categorised as a claim for repayment, whereas the UK Government contended that each of them constituted a claim for damages which was subject to the Factortame conditions. The Court then restated the familiar general principles that it is for the national court to assign a legal classification to the actions brought, but they must give effect to San Giorgio claims in accordance with the principles of equivalence and effectiveness.

The ECJ then continued at paragraph 204:

The ECJ then went on to discuss the conditions for State liability in damages, and said no more about the restitutionary aspects of the claims. The relevant part of their answer to Questions 6 to 9, in paragraph 220 of the judgment, reads as follows:

Mr Ewart submitted that the ECJ adopted a very different approach from the Advocate General in this part of the case, but I do not agree. It seems to me that if the Court had disagreed in principle with the clearly stated views of the Advocate General about the potential width of the San Giorgio principle, they would have said so; and they would certainly not have confined their comments to the relatively narrow point whether the claims based on waiver of reliefs, as well as those based on the FID enhancements, fell outside the principle. The reasoning of the Court in paragraphs 204 and 205 makes the point that the San Giorgio principle, as applied in Hoechst, extends not only to the repayment of tax unduly levied, but also to amounts paid to the State, or retained by it, “which relate directly to that tax”. In Hoechst, the claim in question was for the time value of the ACT which had been levied prematurely. In my judgment, this passage reflects essentially the same approach as the Advocate General’s view that the San Giorgio principle should “extend to all direct consequences of the unlawful levying of tax” (paragraph 132 of his opinion), and so understood it leads on naturally to the discussion which follows.

Paragraph 206 of the judgment is very compressed, but I read it as directed principally to the recovery of surplus ACT which has been unlawfully levied and which the taxpayer company has been unable to utilise. The difference from Hoechst is that in these cases the claim is for recovery of the unlawfully levied tax itself, and is not a direct consequential claim for the time value of ACT which was levied prematurely (because UK domestic law did not allow the making of a group income election) but which the company concerned was subsequently able to set off against its MCT or utilise in some other way.

Paragraph 207 then sets out the one point of substance on which the Court differed from the Advocate General, namely how to characterise the waiver of relief claims. It will be recalled that these claims, as summarised in Question 6(v), relate to situations where a company disclaimed reliefs to which it would otherwise have been entitled (such as capital allowances) in order to set off surplus ACT against its MCT liability, and as a result of which it had to pay MCT on the relevant slice of its taxable profits at the difference between the MCT and ACT rates. The claims were therefore for recovery of part of the MCT which the company had paid, and which was itself (it must be assumed) lawfully levied.

It is interesting to note that the ECJ did not deal with this point by saying that a claim in respect of lawfully levied tax would automatically fall outside the scope of the San Giorgio principle. Instead, they founded their decision on the point that, as with the enhancement of the FIDS, the waivers were “the result of decisions taken by those companies”, and were not “an inevitable consequence” of the unlawful ACT regime. In other words, they saw the chain of causation as having been broken at the stage when these separate decisions were made.

On the other hand, it is also clear, to my mind, that the test of causation for the recovery of consequential loss under the San Giorgio principle must be a strict one. The Advocate General used the phrase “direct consequences”, and the ECJ used the phrase “an inevitable consequence” (in French, “une conséquence inévitable”). I do not regard these expressions as necessarily conflicting with each other, or as intended to provide an exhaustive definition. They reflect rather the need for a direct and unbroken causal link, and an approach to causation which is considerably more stringent than that applicable to a Factortame damages claim (where “it is for the national court to assess whether the loss and damage claimed flows sufficiently directly from the breach of Community law to render the State liable to make it good”: see paragraph 218 of the judgment).

It is common ground that, in so far as the claimants’ claims fall under the San Giorgio principle as a matter of Community law, they should be classified in English law as claims for restitution. It is also common ground that the English law of restitution, as interpreted and clarified by the House of Lords in DMG and Sempra Metals, in general satisfies the requirements of equivalence and effectiveness, and does not need to be further refashioned in order to give full effect to the directly effective Community law rights of the claimants.

The first question, therefore, is to determine which of the claims in the Eighth Re-amended Particulars of Claim fall within the San Giorgio principle.

The Revenue argue for a narrow answer to this question. They concede no more than that the claimants are entitled to restitution of unlawfully levied Case V corporation tax and unlawfully levied ACT, together with associated interest and loss of use claims, in cases where the claimant has itself paid the unlawful tax and none of the corporate tree points arise.

The claimants, by contrast, say that all of their claims fall under the San Giorgio principle, with the exception of those specifically excluded by the ECJ (that is to say, the FID enhancement claims and the waiver of relief claims).

Leaving aside the corporate tree points, the answer to this question is in my judgment fairly straightforward if my analysis of the guidance given by the ECJ is correct. The San Giorgio principle extends to all the claims for repayment of unlawfully levied tax by the companies which actually paid the tax, and to associated interest and loss of use claims. It also extends to all claims for losses which are a direct or inevitable consequence of the unlawful levying of tax. It is for the national court to determine whether this test is satisfied in relation to each of the types of claim described in Question 6 of the Order for Reference, with the exception of the FID enhancement and waiver of relief claims, which the Court has definitively ruled cannot be San Giorgio claims.

The classification of the corporate tree claims, as a matter of Community law, is in my view a question which should be referred to the ECJ, because it needs to be considered in conjunction with the question whether such claims involve a breach of Community law, which I have already held should be the subject of a further reference: see paragraph 138 above. My own view, however, is that the corporate tree claims should be classified as San Giorgio claims if liability under Community law is once established. I do not think that the making of a group income election can sensibly be regarded as a decision which breaks the chain of causation, where its effect is merely to defer the payment of ACT which is (on this hypothesis) unlawful at whichever level within the group it is paid. Similarly, I do not see how the level at which foreign corporation tax is paid can affect the question, if it is held that payment of tax by a lower-tier company, coupled with payment of a dividend to the UK by the water’s edge company, engages Articles 43 and 56. Once the dividend has reached the UK, its treatment in the hands of the recipient is essentially the same, whatever the level at which the foreign corporation tax may have been borne.

Before I give my answer to the question of causation remitted to me by the ECJ, I will first look more closely at the types of restitutionary claim which are available under English law. I will also need to consider whether any of the claims which are not San Giorgio claims may nevertheless come within the scope of the English law of restitution, or whether they sound only in damages.

As a preliminary point, I would emphasise that the distinction between claims which sound only in damages, and those which sound in restitution (whether or not they sound also in damages), is in practical terms a very significant one. There are two main reasons for this, neither of which has anything to do with the questions of legal classification and taxonomy which arise, interesting and important though those questions are. The first reason is that a restitution-based claim does not require the claimants to overcome the additional hurdle of establishing that there has been a sufficiently serious breach of Community law. The second reason is that almost the only way in which the claimants can hope to escape the usual limitation period of six years for their claims is by pleading a cause of action in mistake-based restitution and then relying on section 32(1)(c) of the Limitation Act 1980. As is well known, section 32(1)(c) provides that where the action is for “relief from the consequences of a mistake”, the period of limitation does not begin to run until the claimant has discovered the mistake “or could with reasonable diligence have discovered it”.

I say “almost” the only way, because the claimants have an alternative argument which seeks to attack the orthodox view that the running of time for limitation purposes under section 32(1)(c) can only be postponed where mistake is an essential ingredient of the cause of action. I commented briefly on this argument in my recent decision in a case under the ACT GLO, Europcar UK Limited v Revenue and Customs Commissioners [2008] EWHC 1363 (Ch), [2008] STC …. (“Europcar”) at paragraphs 59 to 64, and expressed the view that it was not suitable for determination at first instance. In the circumstances, the claimants have not sought to argue the point again before me, but reserve it, if necessary, for a higher court.

It is common ground that two types of restitutionary claim in English law are relevant in the present case. The first is a claim for restitution of tax unlawfully demanded under the principle established by the House of Lords in Woolwich Equitable Building Society v IRC [1993] AC 70 (“Woolwich”). The second, established by the decision of the House of Lords in DMG, is a restitutionary claim for tax wrongly paid under a mistake of law. The two types of claim are, at any rate as the law now stands, conceptually distinct and subject to differing requirements. As Lord Hoffmann pointed out in DMG at paragraph 21, English law as yet “has no general principle that to retain money paid without any legal basis (such as debt, gift, compromise, etc.) is unjust enrichment”. Accordingly, a claimant in England “has to prove that the circumstances in which the payment was made come within one of the categories which the law recognises as sufficient to make retention by the recipient unjust” (ibid). So, for example, mistake is not a necessary ingredient of the Woolwich cause of action, whereas it is obviously a crucial ingredient of the second type of claim. Conversely, a Woolwich claim must involve, at least in some sense, the making of a demand by the Revenue, whereas there is no need for a demand in cases of the second type.

The claimants’ arguments on this part of the case were presented to me by Mr Cavender. In opening, he virtually ignored the Woolwich cause of action (although he later dealt with it in his reply to the Revenue’s arguments), and put at the forefront of his submissions the cause of action in mistake-based restitution, which he said was wide enough to cover all of the claims in the present case. He submitted that a mistake had clearly been made about the lawfulness of the offending tax provisions, and that it would not be a profitable exercise to try to identify the precise nature of the mistake. As in DMG, it was obvious that the claimants paid the unlawful tax under a mistake, even though there was a difference of opinion between their Lordships about what the mistake had been. Mr Cavender submitted that the mistake had been caused by the Revenue adopting an unlawful taxation system, and implementing it over a long period of time.

On the issue of causation, he submitted that a simple “but for” test was appropriate, which he said found some support in the existing authorities, although he acknowledged that he was to some extent asking me to break new ground. He stressed that the underlying concept which should guide and inform the development of the law in this area was unjust enrichment, and argued that the claimants should be able to obtain restitution in any circumstances where the Revenue had been enriched and where the mistake remained an operative cause of the enrichment.

The existing authority deployed by Mr Cavender in support of a “but for” test of causation was as follows.

First, he relied on the “two step” analysis of the relevant mistake by Park J in DMG, which was then adopted by Lord Hope and Lord Scott in the House of Lords. Park J had held that the mistake of law made by DMG was not that the ACT in question was payable – in Park J’s view it was payable, and nothing said by the ECJ in Hoechst had altered that fact – but rather that DMG did not realise it could have made group income elections which would have prevented the ACT from being payable: see Deutsche Morgan Grenfell Group Plc v IRC [2003] EWHC 1866 (Ch), [2003] 4 All ER 645, at paragraph 22, and paragraph 25 where Park J said that “[t]he liabilities to pay ACT arose as secondary consequences of that primary mistake”.

That analysis was accepted by Lord Hope, who said in paragraph 62 of his speech in DMG:

Secondly, Mr Cavender referred me to the decision of the Privy Council in Dextra Bank & Trust Co Ltd v Bank of Jamaica [2002] 1 All ER (Comm) 193 (“Dextra Bank”), where the opinion of the Board was delivered jointly by Lord Bingham of Cornhill and Lord Goff of Chieveley. In paragraph 28 they dealt with Dextra’s claim to recover its money as having been paid under a mistake of fact, and said this:

Barclays Bank v W J Simms, loc. cit., was a decision of Robert Goff J, as he then was. The passage cited by the Board in Dextra Bank included this statement at [1980] QB 694E:

I now turn to the submissions for the Revenue. Mr Ewart submitted that the Woolwich cause of action is all that is needed to provide an effective domestic remedy for San Giorgio claims, and that the only other domestic remedy needed to provide compensation for breaches of Community law is the Factortame cause of action for damages for breach of statutory duty. He submitted that the domestic UK remedy of mistake-based restitution is not required by Community law, with the consequence that the claimants cannot rely on it for limitation purposes.

In the context of the Woolwich cause of action, Mr Ewart referred me to the discussion by the Court of Appeal in NEC Semi-Conductors Ltd v IRC [2006] EWCA Civ 25, [2006] STC 606 (“NEC Semi-Conductors”), where the Court rejected the broad approach urged on them by Mr Cavender (appearing there, as in the present case, for the taxpayers) and held that, in the absence of a group income election, ACT was lawfully payable, and could not be said to have been unlawfully demanded by the Revenue: see the judgment of Mummery LJ at paras 140 to 147 and 162. The other two members of the Court, Sedley and Lloyd LJJ, agreed with Mummery LJ’s analysis: see paragraphs 84 and 89.

On the strength of NEC Semi-Conductors, Mr Ewart argued that where Community law does not require a repayment remedy, a Woolwich claim will not arise merely because tax was paid in consequence of a different unlawful statutory provision. The effect of the judgment of the Court of Appeal is that the charge to tax in question must itself be unlawful. Mr Ewart relied on similar reasoning to submit that the corporate tree ACT claims, and the claims for restitution of MCT where losses or reliefs had been surrendered or used to offset unlawful Case V tax when they could otherwise have been set against lawful MCT, could not be brought under the umbrella of a Woolwich claim. In each case, the tax that it was sought to recover was lawfully due, and the fact that it might have been paid as an indirect consequence of a different unlawful statutory provision was irrelevant.

With regard to the mistake-based restitution claims, Mr Ewart stressed that the underlying principle of unjust enrichment is entirely independent of any wrong on the part of the defendant. He relied on the lucid exposition of the concept of unjust enrichment by the late Professor Birks in Chapter 1 of the second (2005) edition of his book, Unjust Enrichment, and the endorsement of that approach (and of similar views expressed by Professor Burrows) by Lord Mance in Sempra Metals in paragraph 231, ending with the statement that “restitution on the basis of unjust enrichment looks, carefully and advisedly, at the recipient’s actual benefit”. On the issue of causation, Mr Ewart relied on the analysis in the dissenting speech of Lord Scott in DMG, particularly at paragraphs 84 and 85, leading to the conclusion at the end of paragraph 85:

When I asked Mr Ewart to formulate the test that the Revenue would propound in relation to mistake, he said that there had to be direct causation, and the mistake had to operate on the payment itself, if restitution was to be available. If the money in question was due when it was paid, it could not be unjust for the recipient to retain it. He gave the example of inheritance tax received by the Revenue due to a failure to make an election in a deed of variation pursuant to section 142 of the Inheritance Tax Act 1984, and submitted that even if the failure to make the election was due to a mistake the taxpayer would be unable to recover the tax from the Revenue, although he might have a claim for damages against a negligent adviser. The reason why the claims in DMG succeeded was that the relevant mistake was made at the time of payment of the ACT, and the claimants paid it when they need not have done so. By contrast, once some independent decision is introduced, or if another party is involved, it is no longer possible to say that the payment was made under a mistake.

In his submissions in reply, Mr Cavender dealt with the decision of the Court of Appeal in NEC Semi-Conductors. He accepted that the Court had adopted a narrow approach to the scope of the Woolwich principle, and pointed out that the claimants were unable at that stage to rely on mistake of law, because the hearing took place between the decisions of the Court of Appeal and the House of Lords in DMG. He also reminded me that NEC Semi-Conductors was a case which only concerned third countries, and suggested that the claimants were in a more difficult position than in the present case because they had to rely on provisions in double tax treaties as incorporated into UK domestic law by TA section 788. I am not sure that this last point is correct, however, because the claimants also had a direct claim for infringement of their Community law rights under Article 56: see my summary of the issues, and how they were decided at each stage of the litigation, in Europcar at paragraphs 36 to 40. Be that as it may, Mr Cavender argued that the Woolwich principle was, on any view, much narrower than the San Giorgio principle, and reaffirmed that what the claimants are really asking for in the present case is “an application of DMG”. He submitted that the causation principles endorsed by a majority of their Lordships in DMG meant that the claimants’ consequential claims should succeed, and invited me to adopt “an incremental approach to the overarching principle of unjust enrichment”.

With the benefit of these submissions, and others which I have not recorded but have done my best to take into account, I will now state my conclusions on this part of the case.

In the first place, I am unable to accept Mr Ewart’s submission that the Woolwich principle alone provides a sufficient UK remedy for claims which as a matter of Community law fall under the San Giorgio principle. As I have already explained, it seems to me that the ECJ has adopted a relatively wide view of the San Giorgio principle in the present case, whereas the Court of Appeal in NEC Semi-Conductors has strictly confined the ambit of the Woolwich cause of action to cases where the tax in question was itself unlawfully demanded. It has given no encouragement to attempts to broaden the concept of a “demand”, or to permit recovery of losses which go beyond the unlawfully demanded tax and interest. I therefore agree with Mr Cavender that the UK cause of action in mistake-based restitution is also needed in order to provide an effective UK remedy for many San Giorgio claims. I also agree with him that, since the decision of the House of Lords in DMG, the Woolwich cause of action is likely to play a subsidiary role in cases such as the present one. Woolwich was decided in 1992, several years before the House of Lords overturned the “mistake of law” rule in Kleinwort Benson Ltd v Lincoln City Council [1999] 2 AC 249 (“Kleinwort Benson”), and before the House of Lords held in DMG that the Kleinwort Benson principle extended to wrongly paid tax. Mistake is not a necessary ingredient of the Woolwich cause of action. As Lord Hoffmann said in DMG at paragraph 13, it is indifferent as to whether the taxpayer paid the tax because he was mistaken, or for some other reason. Where mistake is present, however, a restitutionary claim based on the mistake is likely to cover all the ground that a Woolwich claim could cover, but also has the potential to extend considerably further.

In theory at least, much may turn on how the relevant mistake is identified, and the differences of judicial opinion in DMG show how difficult this question can be. My own preference is to identify the mistake at a fairly general level, at least in cases where the ECJ has held that the domestic tax system breaches fundamental freedoms under Community law. I prefer to see the mistake as a mistake about the lawfulness of the charge to tax, in circumstances where there is an infringement of Community law, and not to analyse the position too closely in terms of the domestic machinery of the UK system, which is naturally predicated on the lawfulness of the system of which it forms part. To a considerable extent, the mistake in cases of this nature is a legal construct, and it can usually only be recognised after the event and with the benefit of hindsight. I find encouragement for this relatively broad approach to the question of mistake in the speeches of Lord Hoffmann and Lord Walker in DMG. I also respectfully agree with Lord Walker’s endorsement (in paragraph 143) of the observation of Neuberger J (as he then was) in Nurdin & Peacock Plc v D B Ramsden & Co Ltd [1999] 1 WLR 1249 at 1272:

For these reasons, I think that the mistake of law in the present case should be identified as a mistake as to the lawfulness of the ACT regime or the Case V charge, and it is probably not helpful to try to express it in more precise terms.

It by no means follows, however, that an equally broad approach should be adopted to the question of what can be recovered in a restitutionary claim based upon mistake. As Mummery LJ said in NEC Semi-Conductors at paragraph 175, the underlying principle is

These general reflections about the essentially subtractive nature of unjust enrichment claims lead me to be very wary of Mr Cavender’s siren call to extend the frontiers of the law of unjust enrichment by allowing recovery, on a so-called incremental basis, for claims which go beyond the repayment of mistakenly paid tax and the reversal of directly associated benefits to the Revenue which can reasonably be seen as arising from the mistaken payment. In my judgment the “but for” test of causation has no role to play in this context, except as a minimum requirement for linking the mistake made by the claimant with the types of recovery that, as a matter of general principle, a restitutionary claim can properly encompass. In other words, it cannot itself be used as a test to define the types of loss which are recoverable, and thus to bring within the scope of the law of restitution the same types of consequential loss as a claimant suing in contract or tort (or, I would add, for breach of statutory duty) might hope to recover.

In my judgment none of the existing authorities relied on by Mr Cavender require, or even encourage, me to take the step which he urges upon me. The test of causation in Barclays Bank v W J Simms, as I understand it, is not directed at all to the recovery of consequential losses, but rather to establishing the necessary link between the mistake and the payment which the claimant seeks to recover. I would make the same comment about the payment which was in issue in Dextra Bank, and also about the valuable discussion of mistake by Lord Hope in DMG. Lord Hope was not concerned with anything other than the actual payments of ACT which had been made in the absence of group income elections. His view was that it was only Park J’s analysis of the mistake which established “an unbroken causative link between the mistake and the payment” (paragraph 62 at 582F).

I will now attempt to apply these principles to the main heads of claim.

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.

It is equally clear, in my judgment, that these actual payments fall within the proper scope of a claim in restitution, together with associated claims for interest (and, in the case of the FIDs, for the time value of the ACT between the dates of payment and repayment).

I would adopt the same analysis in relation to any other payments of ACT, whether by water’s edge companies or by other UK group companies, which the ECJ may hold to have been unlawful when it decides the corporate tree points. The same result would also follow if a higher UK court were to hold, on appeal from my decision, that the corporate tree points should not be referred to the ECJ and should be decided in the claimants’ favour. In all these cases, there will have been an actual payment of unlawfully levied ACT, and the mistake (at the general level at which I have identified it) will have continued to be operative.

Further than this, however, as a matter of English domestic law, I am not at present prepared to go. Consequential steps that were taken within the group to utilise surplus ACT, for example by setting it off against unlawful Case V corporation tax, or by setting it off against lawful corporation tax increased in amount by the disclaimer of capital allowances, may in principle give rise to a claim for compensation, subject to proof of loss and causation, and subject to a sufficiently serious breach being established; but they do not in my judgment give rise to a restitutionary claim, because the whole point of these steps was that tax was not paid to the Revenue, and any consequential enrichment of the Revenue cannot in my view be seen as a direct result, or concomitant, of any unlawful tax charge. So, for example, if lawfully paid ACT was used to offset an unlawful liability to Case V corporation tax, the result was that the company concerned did not pay corporation tax which, if it had done so, it would admittedly now be able to recover. If the Revenue was enriched at all, it was certainly not by the receipt of unlawfully levied corporation tax, but rather as a result of the group then having less ACT available to use for other purposes. An enquiry into whether (and if so how) such ACT would otherwise have been used within the group can in my judgment only form part of a claim for consequential loss. I do not see how it could be said that the Revenue was immediately and directly enriched at the group’s expense by an amount equal to the ACT so used.

This conclusion is, however, subject to the important rider that, if a claim has been classified by Community law as a San Giorgio claim, the principle of effectiveness requires English law to provide an equivalent domestic remedy. In the present context, the ECJ has held that the waiver of reliefs in order to be able to offset a tax levied unlawfully, such as ACT, against an amount due in respect of another tax cannot form the basis of an action under Community law for reimbursement (paragraph 207 of the judgment). I do not find the intended scope of this exception altogether clear, and in particular I am puzzled by the reference to “unlawfully” levied ACT. If the ACT was levied unlawfully, it can be recovered in any event. It seems to me that the further claims under this head are concerned with the use of lawful ACT rather than unlawfully levied ACT. That point aside, however, it seems reasonably clear to me that this exception covers all of the claims in the present case which depend on a waiver of reliefs, or some other similar act, on the part of a claimant company. It therefore clearly covers the ACT claims which are based on disclaimed capital allowances; and I note that in the Eighth Amended Particulars of Claim these are in any event pleaded only as claims for compensation. The position with regard to the other ACT claims is less clear, but for the reasons which I have given I do not regard any of them as a direct, and still less as an inevitable, consequence of the unlawful charge to ACT. I therefore do not consider that they fall within the San Giorgio principle. It is important to note in this connection that, although the Advocate General said that all of the types of claim set out in Question 6 (apart from the FID enhancements) were potentially San Giorgio claims, he also said (at the end of paragraph 132 of his opinion) that in each case it would be for the national court to satisfy itself that the relief claimed “was a direct consequence of the unlawful levy charged”. I do not understand the ECJ to have dissented in any way from that proposition, and their use of the word “inevitable” in paragraph 207 suggests, if anything, an even more stringent test.

Accordingly, having considered the requirements of Community law, I see no reason to modify the conclusions which I have reached as a matter of English law.

It is clear from the decision of the ECJ that these claims are not San Giorgio claims, and Mr Cavender did not seek to argue the contrary. He did, however, submit that a restitutionary claim would lie under English law in respect of the enhancements of the FIDs paid to exempt shareholders, although he accepted that the enhancements paid to non-exempt shareholders could only give rise to a damages claim. I am unable to accept this submission. The parent company which paid the FIDs was under no legal obligation to enhance them, nor can it be said that the payment of the enhancement directly relieved the Revenue of any obligation which it might have been under to pay a tax credit to the recipients. Even if it be assumed that the FIDs should have carried tax credits, either on the whole or on part of their value, the Revenue was not itself obliged to pay the credits together with the dividends, but only to allow the recipient shareholders to make use of the credits in due course, and to repay them, upon a claim being made for that purpose, if the recipient was exempt or a non-taxpayer. Furthermore, any loss suffered as a result of the absence of tax credits was, on the face of it, suffered by the shareholders, not by the company which paid the FIDs. Unsurprisingly, such claims have been made by many shareholders, and they are being managed under a separate FID GLO which is proceeding before the Special Commissioners. Mr Cavender sought to persuade me that the shareholders could not have a valid claim in cases where their FIDs had been enhanced, because the enhancement served the same purpose as a tax credit and they have therefore suffered no loss. I will assume, without deciding, that that contention is correct. It does not, however, lead to the conclusion that the FID enhancements can somehow be treated as having been paid on behalf of the Revenue and as discharging a liability to which the Revenue would otherwise have been subject.

In truth, it seems to me that the enhancement of the FIDs was a form of self-help remedy, which the BAT group decided to adopt for good commercial reasons of its own. I will make my findings of fact about that aspect of the FID claims in the next section of this judgment: see paragraphs 277 to 302 below. In my judgment, however, the enhancement of the FIDs cannot possibly give rise to a restitutionary claim in English law, for essentially the same reasons that the Advocate General and the ECJ held it did not give rise to a San Giorgio claim.

If I am right in my approach to the ACT claims, the position with regard to the Case V claims is similar. The claims for unlawfully paid tax, by the water’s edge UK companies which received the foreign dividends, are in my view plainly good restitutionary claims. The same applies if, as a result of success on the relevant corporate tree points, the unlawful charge is held to extend to corporation tax on dividends received from foreign subsidiaries which had not themselves paid foreign corporation tax, although foreign tax was paid further back down the chain. In each case, it will still be the recipient UK company which is liable to the unlawful charge.

By contrast, I do not consider that any of the other Case V claims sound in restitution under English law, nor do I consider them to be either a direct or an inevitable consequence of the unlawful tax charge, so they are not San Giorgio claims under Community law. In short, they sound, if at all, in damages.

The claimants say that the only reason why the FIDs were enhanced was to compensate exempt shareholders for the absence of a tax credit. This proposition is disputed by the Revenue, who submit that the predominant reason for the BAT group’s decision to pay enhanced FIDs to its shareholders was the reduction in accumulated surplus ACT (i.e. the ACT mountain) which could thus be achieved.

The introduction of the FID regime, and the BAT group’s reaction to it, are dealt with by four of the claimants’ witnesses of fact in their written evidence: Mr Broughton, Mr Allvey, Mr Bilton and Mr Hardman. The first three of these witnesses were cross-examined by Mr Baldry on this part of their evidence. Mr Ewart’s cross-examination of Mr Hardman dealt with other topics, no doubt because Mr Hardman’s evidence in relation to FIDs added little of substance to that of his colleagues.

The two witnesses who were most closely involved were Mr Bilton and Mr Allvey. The proposals for what became the FID system were first announced by the Chancellor of the Exchequer in the March 1993 budget. From the outset, Mr Bilton (who was then a senior tax manager with the ultimate parent company of the group, BAT Industries, and reported to the Head of Tax for the whole group, Mr Ken Etherington) was involved with Mr Etherington in reviewing the proposed legislation and considering whether it would benefit BAT to adopt the system. Mr Etherington attended various consultation meetings with the Revenue, and Mr Bilton also attended a number of seminars given by professional firms explaining how the legislation was likely to work.

Mr Allvey was at this time the finance director of BAT Industries. He had been involved, together with the finance directors of the hundred largest UK companies (“the 100 Group”), in lobbying the Government for the repeal of the ACT system, and he was also a member of the tax committee of the CBI. It was Mr Allvey’s responsibility, as finance director, to recommend to the board of BAT Industries what dividends should be paid. The question of how to make use of the group’s ACT mountain was never far from his mind, and one of the prime objectives he set for Mr Etherington was to find any lawful ways in which surplus ACT could be utilised. One innovative idea, which had been successfully introduced after a good deal of hard work and research in the early 1990s, was the payment of enhanced scrip dividends. The advantage of such dividends from BAT’s point of view was that they did not attract an ACT liability. By offering shareholders the option of taking a dividend in this way, enhanced so that the net value of the dividend received in shares exceeded the value of a cash dividend carrying a tax credit, the ACT liability on distributions could be drastically reduced. Mr Allvey persuaded the main board that the idea was a good one, and secured approval for it. The enhanced scrip dividend was well received by investors, and produced good publicity for the group. Mr Allvey says that BAT Industries was the first company to pay such dividends, and when it did so the market saw the company as proactive and willing to investigate new ways of delivering shareholder returns.

BAT Industries had a policy of progressively increasing its dividends year by year, and Mr Allvey and his team worked to a five year plan. The procedure was that BAT Industries would usually pay one interim dividend and then a final dividend every year, with a bias in amount towards the final dividend. In the period before FIDs, the amount of the final dividend recommended by the board would be announced with the company’s final results in early March. The interim dividend was paid on a board resolution, but the final dividend required the approval of shareholders at the annual general meeting.

The stated purpose of the proposed FID regime was to mitigate the problem of surplus ACT for companies with substantial amounts of foreign income. It was therefore of obvious interest to the BAT group, and on 6 April 1993 Mr Etherington prepared a note for the Chairman which summarised the proposals contained in a consultative document issued by the Revenue. In this note Mr Etherington identified certain problem areas. He said the biggest problem was not a tax one, but rather

The technical problems to which Mr Etherington had drawn attention turned out not to be very serious, as Mr Bilton explains in his statement. In particular, the legislation when it was introduced did permit companies to pay both FIDs and ordinary dividends in the same year, although streaming of dividends paid at the same time was not permitted; and the legislation also allowed FIDs to be carried back, save that they could not be matched against foreign income received before 1 January 1994.

As the proposals for the introduction of FIDs began to develop, Mr Bilton and Mr Etherington did a lot of work on the effect that payment of FIDs would have on different shareholders, and how they might be increased to compensate exempt shareholders for the lack of a credit. One of the first things they did was to establish the proportion of shareholders who would be adversely affected, and Mr Bilton attended meetings with members of the group’s investor relations department on this issue. They were advised by the group’s banking advisers, Cazenove and Co, that 37.3% of the shareholder base, including the 15 largest shareholders, were tax exempt institutions who would suffer from the inability to receive a tax credit.

The attitude of Mr Bilton and Mr Etherington was that the FID system was “better than nothing”, because it would enable some surplus ACT to be used up, but the commercial problem remained that the tax exempt shareholders would not obtain any repayable tax credit if the dividend was paid as a FID. In paragraph 32 of his witness statement, Mr Bilton says this:

In August 1993 Mr Etherington produced a note for the Economic Affairs Committee of the CBI, which recommended that the FID proposals should be welcomed by the CBI even though the solution to the problem of surplus ACT was not perfect. He pointed out that payment of FIDs would be optional, but on the government’s reckoning the proposals would cost the exchequer £200 million. He then commented:

Proposals were worked up within the group to cover various possibilities, of which the most promising (and the one finally adopted) was to pay the 1993 final dividend in FID form. On 27 January 1994, Mr R H Lomax, the financial assistant to the Chairman of BAT Industries, sent a note to Mr Allvey and Mr Etherington in which he said:

On 15 March 1994 Mr Etherington prepared a note to the board of BAT Industries, in which he set out three proposals for using FIDs in 1994. The first proposal was that the final dividend for 1993 be paid as a FID, with the payment date being advanced from June to 1 July 1994, that being the date when the legislation would come into force. The date for payment of the next interim dividend would also be advanced from October 1994 to 1 January 1995, in order to avoid the payment of any ACT in 1994. Mr Etherington commented:

The benefit which Mr Etherington believed to be “unintended” was, as both Mr Bilton and Mr Allvey accepted in cross-examination, in fact part of the way that the system was meant to work, not least because ACT could always be carried forward. However, this was not how Mr Etherington and Mr Allvey perceived the point at the time, and their advice to the board was framed accordingly.

Meanwhile, further advice was sought from external advisers about the reaction of investors to the FID system, and on 2 March 1994 a memorandum was received from those advisers giving the results of a survey which they had conducted among institutional investors on the subject of enhanced scrip dividends and FIDs. Concerning FIDs, the memorandum said that the level of knowledge of them was generally low, but that when the system was explained:

The board meeting at which the final dividend for 1993 would be fixed, subject to shareholder approval, was due to take place on 24 March 1994. Mr Allvey prepared a paper for the meeting, which covered much the same ground as Mr Etherington’s earlier note of 15 March, but in slightly more detail. He referred to the same “unintended benefit” of the proposed legislation, and the fear that it might be removed by amending legislation once it became apparent to the Revenue. Of the proposal to pay a FID, he said:

This recommendation was duly accepted at the board meeting of BAT Industries on 24 March, and it was resolved to recommend payment of a final dividend of that amount at the forthcoming AGM on 6 May 1994. On the same day, a letter was sent to shareholders headed “25% Increase in final dividend”. The letter explained that BAT Industries intended to pay the 1993 final dividend of 12.2p, which it had previously announced on 9 March, as a FID, and to increase it to 15.25p per share. The increase was said to represent the savings that the company would make in ACT by paying the dividend as a FID. The advantages to shareholders were then explained, and they were asked to vote in favour of the proposal at the AGM on the basis that it was “clearly in the interests of BAT Industries and its shareholders”.

Mr Bilton agreed in cross-examination that this letter would have been the first that the BAT shareholders would have known of the proposal to pay the dividend by way of a FID, although he also agreed in re-examination that some of the major shareholders would have known about the proposals as a result of the review conducted by the group’s external advisers.

The AGM gave the necessary approval, and BAT Industries paid its first FIDs accordingly on 1 July 1994. Payments of enhanced FIDs then continued until 1997, when they were discontinued as a result of the legislative change which removed the entitlement to repayable tax credits from exempt taxpayers altogether.

Both Mr Broughton and Mr Allvey said in their oral evidence that, if the FIDs paid in July 1994 had not been enhanced, there was a real fear that exempt shareholders would have voted against it at the AGM. Mr Broughton also said that the “centre of focus” for BAT in relation to the enhancement of the FIDs was the lack of tax credit for the exempt shareholders, rather than the availability of surplus ACT. This answer was, however, given in re-examination, to a slightly leading question which invited him to choose between those two alternatives. For his part, Mr Allvey said in answer to another rather leading question in re-examination that, if the FID regime had allowed the tax-exempt shareholders a credit in the same way as it did other shareholders, there would have been no need to enhance the FIDs.

Having carefully considered all this evidence, I conclude that the main reason why BAT Industries decided to take advantage of the FID regime in the first place was the opportunity which it presented to make use of surplus ACT and thus reduce the group’s ACT mountain. This motivation was made even stronger by the fear, unfounded though it was, that the government might step in to prevent the use of surplus ACT carried forward from earlier years, once it realised that the effect of the FID regime would be to make surplus ACT available for use against MCT. Because there was virtually no other way in which surplus ACT could be used within the group, and because the ACT paid on FIDs was in principle repayable, the payment of dividends in the form of FIDs would substantially reduce the group’s overall tax burden. In these circumstances, the otherwise irrecoverable ACT that was saved by paying a FID rather than a conventional dividend could be seen as a bonus to the group at the Revenue’s expense, and the bigger the FID, the bigger that bonus would be. There was therefore everything to be said for making the maximum use of the FID regime, and doing so at the earliest opportunity.

It is against this background that the enhancement of the FIDs has to be evaluated.

The evidence establishes, in my judgment, that it would have been commercially unacceptable to pay unenhanced FIDs (by which I mean FIDs at the same level as the ordinary dividend that would otherwise have been payable) to the very large and influential body of exempt shareholders, which comprised nearly 40% in value of the total shareholder population. The team which was working on FIDs within the group, including Mr Etherington, Mr Bilton and Mr Allvey, were aware of this problem from a very early stage, and their assessment of it was confirmed by the advice which was received in March 1994 from their external advisers. The only way of solving this problem was to ensure that any FID paid to exempt shareholders would in practice be at least equal to the gross amount that they would have received upon payment of a conventional dividend plus tax credit. A further consequence was that the non-exempt shareholders also had to receive a similar enhancement, even though the FID regime gave them a non-repayable basic rate tax credit, because streaming of dividends was not permitted: they had to be enhanced for everybody, or not at all. This was something of an uncovenanted bonus for the non-exempt shareholders, although it was hoped that the increased earnings per share, together with the improved tax position of the group, would be reflected in an increase in share value, to the advantage of all shareholders. The package could therefore be presented as one that was of benefit to everybody, and Mr Broughton agreed in cross-examination that this was indeed the case.

It follows, in my view, that once the decision had been taken to go down the FID route, because of the ACT advantages which it offered, there was then no practical alternative to enhancement of any FIDs that were actually paid. In that sense, the enhancements would not have been made but for the unavailability of tax credits for exempt shareholders; and it is not surprising that, when tax credits became generally unavailable for exempt shareholders in 1997, the enhancements ceased. There was no longer any need for them in order to give the exempt shareholders the same cash benefit as they would have received from an ordinary dividend. This is in my view a telling point, and it reinforces my conclusion that the main cause of the enhancement of the FIDs, as opposed to the main cause of the decision to pay them at all, was indeed the unavailability of a tax credit for exempt recipients.

Mr Ewart submitted that the picture presented by the contemporary documents in 1993 and 1994 was rather different, and that the driving consideration in fixing the level of the first FIDs paid on 1 July 1994 was not the need to provide an enhancement for exempt shareholders, but rather a commercial decision to pass on to the shareholders the whole of the ACT saving which was expect to flow from payment of the 1993 final dividend in FID form. On this view of the matter, the dominant consideration was the size of the ACT saving, and the question was how much of it should be passed on to the shareholders. The solution adopted had the convenient result that the exempt shareholders would be no worse off, but that was an incidental advantage rather than a governing factor of the decision.

I accept that there are some passages in the documents which can be read as lending some colour to Mr Ewart’s analysis, and it is certainly true that the question of how to share the benefit of the projected ACT saving was one of the matters which were considered. However, Mr Allvey’s paper for the board on 22 March 1994 expressly dealt with the “shareholder perspective” as well as the “company perspective”, and the weight of the evidence, both oral and written, seems to me to provide decisive support for the conclusion that the question of enhancement was always a primary consideration, and not just a convenient side effect of a decision taken for other reasons.

I therefore find in favour of the claimants that the primary reason why the FIDs were enhanced in the BAT group was indeed to compensate exempt shareholders for the absence of a tax credit for them under the FID regime.

The basic issue for determination under this heading is whether a defence of change of position is available to the Revenue in respect of the claims in restitution, and (if so) by reference to what principles it may be relied upon by the Revenue. It is agreed on both sides that this is a novel question, in the sense that there is no reported case in which the Revenue, or indeed any other government department, has sought to rely on such a defence, although the existence of a change of position defence in the English law of restitution generally has of course been recognised since the decision of the House of Lords in Lipkin Gorman v Karpnale Ltd [1991] 2 AC 548 (“Lipkin Gorman”).

The claimants submit, and the Revenue accept at least for the purposes of the present case, that the defence is in any event not open to the Revenue in respect of San Giorgio claims, because to allow such a defence would breach the Community law principle of effectiveness. The Revenue’s express acceptance of this proposition is confined to Woolwich claims, since (as I have explained) they submit that the Woolwich cause of action is the only restitutionary remedy needed in English law to give effect to San Giorgio claims. I have rejected that submission (see paragraph 260 above), and held that the domestic remedy of mistake-based restitution is also needed for that purpose. In those circumstances, the logic of the Revenue’s concession in relation to Woolwich claims must in my view also extend to mistake-based restitution claims in so far as they are needed to give full effect to San Giorgio claims, but of course no further. In other words, the issue is a live one in relation to all of the restitutionary claims in the present case apart from those which I have held should be classified as San Giorgio claims (that is to say, broadly speaking, the claims for repayment of unlawfully levied tax and associated interest and loss of use claims, together with the corporate tree claims if the ECJ agrees with my view that they should be decided in the claimants’ favour and be treated as repayment claims within the San Giorgio principle).

In the light of the Revenue’s concession, I need not spend long on the question of what defences Community law recognises to San Giorgio claims.

The case law of the ECJ accepts that reasonable limitation periods may be fixed for repayment claims, in the interests of legal certainty: see for example Case C-62/00, Marks and Spencer Plc v Commissioners of Customs and Excise, [2000] ECR I-6325, [2003] QB 866, (“Marks and Spencer”) at paragraph 35. A Member State may also be relieved from the obligation to repay VAT or customs levies, subject to strict conditions, where the charge in question has been passed on to the customer, and reimbursement of the taxable person would unjustly enrich the latter: see for example Case C-147/01, Weber’s Wine World v Abgabenberufungskommission Wien [2003] ECR I-11365 at paragraph 94.

There is, however, no precedent for a defence of this nature in a case involving direct taxation and breach of one of the fundamental freedoms, and the Court has adopted a robust approach in holding that conflicting rules of national law must be disregarded if they would prevent full recovery of an unlawful tax. So, for example, in Hoechst the ECJ expressly ruled that the so-called Pintada principle of English domestic law, as it was then understood, could not bar recovery on the footing that at the date of issue of the writ no principal sum was outstanding. (For the coup de grâce now delivered by the House of Lords to the Pintada principle, see Sempra Metals passim).

Similarly, in a Danish case, Case C-188/95, Fantask A/S and others v Industriministeriet (Erhvervsministeriet) [1997] ECR I-6783, (“Fantask”) the ECJ held that Denmark could not refuse to make a repayment on the basis of a settled law defence available under Danish law. As the Court said in paragraph 40:

As I have already mentioned, the Revenue have pleaded change of position as a defence to the ACT claims, the FID claims and the Case V claims: see paragraphs 31, 48, 56 and 62 of the Amended Defence.

In each context the matters relied upon in support of the plea are the same, so it will be sufficient to quote paragraph 31 and the particulars that are then set out in paragraphs 32 to 35:

The facts of Lipkin Gorman are too familiar to need detailed recitation. A solicitor (Mr Cass) had stolen money from his firm’s client account, and used it to gamble at the defendant’s casino. Cass was convicted of theft, and the solicitors brought an action against the club seeking to recover the monies which he had stolen. They framed their main claim as a personal claim for money had and received. As Lord Goff emphasised at 574B – C, in claiming the money in this way the solicitors had not sought to make the club liable on the basis of any wrong; nor was there any suggestion that the club had acted in bad faith. The House of Lords held that this claim succeeded in principle, but one of the defences relied upon by the club was, in effect, a defence of change of position. It was argued that it would be unfair to hold the club liable, because it had acted throughout in good faith, it had honoured all the gaming contracts which it had entered into with Cass, those contracts were merely void and not illegal, and they were accordingly in the position of an innocent third party to whom money was given but who no longer retained it: see 578A – C.

The only one of their Lordships to deal with this question at any length was Lord Goff. After reciting the club’s submission, he continued at 578C:

It is important to note from this preliminary passage that the defence of change of position has to be firmly based on legal principle, and is not a matter for the general discretion of the court. A claimant who has made good his claim to recover stolen money (or, I would add, money paid under a mistake) is prima facie entitled to recover it at common law, and if that right of recovery is to be denied, it can only be on the basis of a properly articulated legal principle.

After reviewing some earlier authority, and pointing out that estoppel was not an appropriate concept to deal with the problem, Lord Goff continued at 579E:

I would make two further comments at this point. First, the defence is not an all or nothing one. Depending on the facts, it may extend to the whole or any part of the claim. Secondly, Lord Goff’s articulation of why it would be unjust to allow restitution in cases of this nature does not depend on a balancing of the prejudice to the defendant if he has to repay with the unfairness to the claimant if restitution is denied. The focus is solely on the position of the defendant. If his position has so changed that he will suffer an injustice if called upon to repay or to repay in full, that injustice will then automatically outweigh the injustice of denying the claimant restitution. This may at first blush seem a little surprising, but is in my judgment fully justified, at least as a general proposition, by the innocence of the defendant who has received the money, and the absence of any requirement of fault in the cause of action for its recovery.

After referring briefly to some of the relevant authority in the common law world, Lord Goff summed up his conclusions at 580C in a paragraph which I must cite in full:

Lord Goff then added what he termed “two further footnotes”. The first distinguished the defence of change of position from the defence of bona fide purchase (the latter providing a complete defence, without in most cases any enquiry into the adequacy of the consideration, whereas the former “will only avail a defendant to the extent that his position has been changed”). The second was to welcome recognition of the defence, because its availability in appropriate cases “will enable a more generous approach to be taken to the recognition of the right to restitution”.

At the risk of redundancy, I will again add a few comments of my own arising from Lord Goff’s magisterial summary.

First, Lord Goff was himself the first to recognise that the new defence should be left to develop on a case by case basis, in the usual way. In saying this he had the express agreement of Lord Bridge of Harwich: see 558H.

Secondly, when Lord Goff said it is commonly accepted that the defence “should not be open to a wrongdoer”, it seems to me fairly clear from the context that what he had in mind was cases where the cause of action relied upon by the claimant was itself founded upon wrongdoing by the defendant. That is what he said a few sentences later in relation to the facts of Lipkin Gorman itself (“the action is not founded upon any wrongdoing of the respondents”), and it is a point which he had already emphasised at 578C – D. The focus upon the cause of action suggests that other wrongdoing by the defendant is unlikely to be relevant, and again some support for this inference can be found in Lipkin Gorman itself. Part of the money received by the club from Cass was in the form of a banker’s draft for £3,735 drawn in favour of the solicitors, which Cass then endorsed and the club accepted in exchange for chips. Lord Goff held, on this part of the case, that the solicitors had sufficient title to bring an action for damages for conversion of the draft, and that the club did not become holder in due course of the draft because it never gave value for it. Accordingly, that part of the solicitors’ claim succeeded, and in respect of the draft the club was, at least technically, a wrongdoer. Nevertheless, that wrong did not impinge in any way on the cause of action for money had and received, and Lord Goff clearly considered it irrelevant to the defence of change of position.

Thirdly, Lord Goff’s broad statement of the nature of the defence again focuses solely on the position of the defendant. The relevant question is whether his position has so changed “that it would be inequitable in all the circumstances to require him to make restitution, or alternatively to make restitution in full”. The relevant equities are those which arise from the defendant’s change of position. They do not have to be weighed up against any competing equities of the claimant’s, because the claimant is prima facie entitled to recovery as of right.

Finally, the mere fact that the defendant has spent some or all of the money does not make it inequitable that he should be called upon to repay, because the expenditure might have been incurred by him in any event. This is a point that is all too often overlooked, and Lord Goff reinforced it by saying that the defence “is likely to be available only on comparatively rare occasions”. Situations where the defendant has spent some or all of the money before the claimant brings his claim for repayment are likely to be anything but rare; but cases where the expenditure has not been in the ordinary course, or on the payment of debts which the defendant would anyway have to discharge, are likely to be much more infrequent. Lord Goff referred in this connection to the speech of Lord Simonds in Ministry of Health v Simpson. That case formed part of the explosion of litigation generated by the residuary gift to charity in the will of Caleb Diplock, which the House of Lords held to be invalid in Chichester Diocesan Fund and Board of Finance Incorporated v Simpson [1944] AC 341, and in respect of which he therefore died intestate. The decision in Ministry of Health v Simpson established that the next of kin had a direct claim in equity against the charities to whom the residuary estate had been wrongly distributed. In the passage referred to by Lord Goff at 276, Lord Simonds gave short shrift to the argument that a claim should not lie against a person who had received a legacy in good faith and then spent it, without knowledge of any flaw in his title. He said:

Mr Cavender referred me to a number of more recent authorities in which the defence of change of position has been considered. The most important of these are the decisions of the Court of Appeal in Scottish Equitable Plc v Derby [2001] EWCA Civ 369, [2001] 3 All ER 818 and Niru Battery Manufacturing Co v Milestone Trading Ltd [2003] EWCA Civ 1446, [2004] 1 All ER (Comm) 194 (“Niru Battery”), and the decision of the Privy Council in 2002 in Dextra Bank (to which I have already referred on the question of causation in mistake claims: see paragraph 251 above).

Scottish Equitable v Derby was a case in which a life assurance company had mistakenly overpaid retirement benefits on a pension policy. The amount of the overpayment was £172,451, of which £51,333 had been received by the defendant, Mr Derby, himself. Of this sum, he used £41,671 to reduce (by about two thirds) the mortgage on his matrimonial home. He used the balance to live a little better by improving the lifestyle of himself and his family in very modest ways. When Scottish Equitable discovered the mistake, it asked Mr Derby to repay the overpayment, but he declined to do so. The judge at first instance (Harrison J) decided the case in Scottish Equitable’s favour, and his judgment was upheld by the Court of Appeal. The leading judgment was given by Robert Walker LJ, the other members of the court being Simon Brown and Keene LJJ.

In the course of his judgment, Robert Walker LJ endorsed “a wide version” of the defence of change of position, which 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. The causal link must satisfy at least a “but for” test, as he said in paragraph 31:

Robert Walker LJ also said he would “readily accept” that the defence was not limited to specific identifiable items of expenditure: see paragraph 33, where he went on to say:

The other two members of the court agreed with Robert Walker LJ. Simon Brown LJ expressed sympathy with Mr Derby’s predicament (which included matrimonial and health problems, as well as his financial position), but held that these additional factors had to be ignored, and that to give effect to them would simply involve what Scrutton LJ once called “well-meaning sloppiness of thought”: see paragraph 53.

In Dextra Bank, one of the issues was the relevance of fault to the defence of change of position. The plaintiff bank submitted that, where the defence was invoked, it was necessary to balance the respective faults of the two parties, because the object of the defence was to balance the equity of the party deprived with that of the party enriched: see paragraph 40. This submission was decisively rejected by the Privy Council. They pointed out that nothing had been said by Lord Goff in Lipkin Gorman about the relevance of fault, and said that, if fault was to be taken into account at all, it would be unjust to take into account the fault of one party (the defendant) but to ignore fault on the part of the other (the plaintiff). The question therefore was whether it should be relevant to take into account the relative fault of the two parties: paragraph 42. Having considered the position in the USA and New Zealand, Lord Bingham and Lord Goff, delivering the opinion of the Board, concluded in paragraph 45:

I would only add that this rejection of the relevance of fault to the defence of change of position is of particular value, because it has the authority of Lord Goff himself.

Niru Battery was a case with complex facts and involved a number of different parties with interlocking interests. It arose out of a contract for the purchase of lead, and the defendants included two banks. One of those banks had paid a sum of money under a letter of credit under a mistake of fact, and the other bank, CAI, was held by the trial judge to be liable to repay that amount unless it had a good defence of change of position. The judge held that the defence was not available to CAI on the facts, and CAI appealed.

The leading judgment in the Court of Appeal was given by Clarke LJ, and he dealt with the change of position defence in paragraphs 143 to 170 of his judgment. Sedley LJ agreed substantially with Clarke LJ’s reasoning (see paragraph 172), but also gave his own reasons for dismissing CAI’s appeal on the change of position issue: see paragraphs 176 to 192. Dame Elizabeth Butler-Sloss P agreed with both judgments.

The discussions of the defence by Clarke and Sedley LJJ are long and interesting, but in the context of the present case they seem to me to be of only marginal importance. I say that because the main argument advanced by counsel for CAI in the Court of Appeal was that a defendant should only forfeit the right to rely on the defence if he has acted dishonestly, which in this context should be equated with bad faith. The Court of Appeal considered that this submission took too narrow a view of what Lord Goff had meant by “bad faith” in Lipkin Gorman, and they restated the test in terms of whether it would be “inequitable or unconscionable” to allow the recipient of money paid under a mistake of fact to deny restitution to the payer: see in particular paragraphs 149, 161 and 180-185. In the present case, however, bad faith, however interpreted, is not an issue, and Mr Cavender founded his submission that the Revenue were not entitled to rely on the defence, not on the proposition that the Revenue had acted inequitably or unconscionably, but rather on the simple proposition that they were wrongdoers, a category which Lord Goff had been at pains to distinguish from cases of bad faith.

I would only add that the Court of Appeal’s emphasis on equitable considerations, and the analogy which they drew with recent case law on liability for knowing receipt, may in my respectful opinion be difficult to reconcile with the underlying logic of Lord Goff’s approach in Lipkin Gorman, and with the rejection by the Privy Council in Dextra Bank of a balance of fault test. The decision of the Court of Appeal in Niru Battery has indeed attracted criticism from some distinguished commentators: see for example Professor Birks, Unjust Enrichment, second edition, pp.214-5.

One further recent case which I should mention is the decision of Laddie J in Barros Mattos jnr v MacDaniels [2004] EWHC 1188 (Ch), [2004] 3 All ER 299, where he held that the otherwise innocent transferees of the proceeds of a large bank fraud could not rely on the defence of change of position where they converted the money into local currency in breach of Nigerian law. The change of position relied upon was itself illegal, and the judge applied the principle that a person cannot base his defence upon his own wrongdoing.

In the light of the principles to be derived from the existing case law, I will now state my conclusions on the main questions that were debated before me.

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.

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.

I find some encouragement for thinking that this was the focus of Lord Goff’s attention in the fact that he must have regarded the club’s conversion of the banker’s draft, which was clearly a tort and therefore a legal wrong, as irrelevant to the club’s change of position defence to the restitutionary claim against it in mistake. The club was, to that limited extent, a wrongdoer, but not in a way that impinged on the separate mistake-based claim against it. The defence was therefore available.

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.

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.

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.

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 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?

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.

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.

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 anyway be under an obligation to fund.

I must also clear away one major misconception which, as it seems to me, has bedevilled much of the pleading and evidence on this issue. Most of the Revenue’s pleaded case in support of its change of position defence is directed to establishing the proposition that it would now cost an enormous amount, and would severely disrupt public finances, if the Revenue had to pay the claimants’ claims in full: see the particulars which I have quoted in paragraph 310 above. Similarly, much of the evidence of Mr Ramsden (to which I will come in the next section of this judgment), and most of his cross-examination by Mr Aaronson, was devoted to that aspect of the matter. But the relevant question is not what it would now cost the Revenue to meet a judgment, or how it could reasonably expect to do so. The relevant question is whether the Revenue has in the past changed its position, on the strength of the receipts paid under a mistake, in such a way that it would now be inequitable to require the Revenue to make restitution. That question is correctly pleaded in paragraph 31 of the Amended Defence, and is reflected in the simple and (to my mind) compelling point made in paragraph 32, namely that the sums in question formed part of the UK’s tax revenue for the years in which they were paid, and have since been irretrievably spent, in some cases decades ago.

The remaining particulars given in paragraphs 33 and 34 are in my judgment largely, if not wholly, irrelevant to the defence of change of position, as are the points made in paragraph 35 relating to the position of the claimants. The relevance, if any, of all these considerations is to an entirely different question, which is not a question of domestic English law at all, but rather one of Community law, viz. whether there should be a temporal limitation on the judgment of the ECJ. The UK government twice tried to persuade the ECJ that such a limitation should be imposed, but without success: see the judgment of the Court at paragraphs 221 to 225, and the separate order of the Court made on 6 December 2006 refusing the UK’s request to reopen the oral and written procedure.

Evidence was given for HMRC by Mr David Ramsden, who has been the Treasury’s chief economic adviser since March 2008, and managing director of the macro economic and fiscal policy directorate and joint head of the government economic service since June 2007. He was previously director of macro economic and fiscal policy, and director of budget and tax policy with responsibility for delivering budgets and pre-budget reports, from 2004 until 2006. Between 1999 and 2003 he led the Treasury’s work on the assessment of the five economic tests which formed the basis of the Government’s decision not to participate in the single European currency.

In his witness statement dated 9 April 2008 Mr Ramsden said that he had been asked to explain in general terms how the annual budget is drawn up, and what consequences might arise for public finances in the event that the Revenue had to satisfy a judgment in the amount of £2 billion to £5 billion. I will not set out Mr Ramsden’s description of the budget process, but reference may be made to his statement for a description, at a fairly high level of generality, of how government fiscal policy is announced and implemented, of the fiscal policy framework followed by the Labour government since May 1997, and of the spending review process whereby, once the annual Budget has determined the amounts available to be spent, those sums are then allocated to spending departments.

For present purposes, the most significant part of Mr Ramsden’s written evidence is what he says in paragraph 15, under the heading “Spending of tax revenues”:

I do not propose to make detailed findings of fact at this stage about the remainder of Mr Ramsden’s evidence, which was directed to the question of how a judgment against the Revenue for between £2 billion and £5 billion would be met, because (as I have explained) it seems to me irrelevant to the question of change of position. Unsurprisingly, Mr Ramsden said that a judgment of that order of magnitude would be a considerable shock to the system, and the Government would have to make policy decisions in order to deal with it. Although Treasury forecasts of public sector borrowing allow for an annual margin of error of roughly one percent of gross domestic product, which on current figures would be around £14 billion, a judgment of this magnitude would be what Mr Ramsden called “an exogenous shock”, in the sense that it would fall completely outside the normal system upon which the forecasts are based, and it would therefore require a specific policy response to deal with it.

This issue arises in the context of the claimants’ non-restitutionary claims for compensation under Community law. It is common ground that the principle of effectiveness requires such a remedy to be provided in English law, in addition to a restitutionary remedy for San Giorgio claims. It is also common ground that three conditions need to be satisfied before a Member State will be held liable to pay compensation for loss and damage caused to a person as a result of breaches of Community law. Those conditions were authoritatively laid down in Factortame, and were repeated by the ECJ in the present case in paragraph 209 of the judgment:

There is no doubt in the present case that the first of the three Factortame conditions is satisfied. As the ECJ says in paragraph 211, Articles 43 and 56 EC are directly effective and confer rights on individuals.

Equally, the third Factortame condition presents no difficulties of principle. The ECJ says in paragraph 218 that it is for the national court to assess whether the loss and damage claimed flows sufficiently directly from the breach of Community law to render the State liable to make it good. Mr Cavender submitted that a “but for” test of causation is appropriate for this purpose, and I did not understand Mr Ewart to dissent from that proposition. In any event, I agree with Mr Cavender that a “but for” test is appropriate.

The dispute between the parties is about the second Factortame condition, and whether on the facts of the present case a sufficiently serious breach of Community law by the UK has been established. The question is again one for the national court to determine, but Question 9 in the Order for Reference asked the ECJ to give any guidance it considered appropriate about the circumstances which the national court ought to take into account in determining whether there is a sufficiently serious breach, and

The guidance which the ECJ saw fit to give on this question is contained in paragraphs 213 to 217 of the judgment, as follows:

Mr Aaronson realistically accepted that this guidance gives what he termed a “strong steer” to the national court to conclude that the breaches of Community law in the present case could not be regarded as sufficiently serious until, at the earliest, the decision of the Court in Verkooijen, which was delivered on 6 June 2000. Mr Aaronson also accepted:

The Advocate General also dealt with the question of sufficiently serious breach at some length, in paragraphs 136 to 138 of his opinion. He began by saying that he had considerable doubts whether the condition of sufficiently serious breach was fulfilled for all aspects of the system which, in his view, breached Community law, and cited from the guidance given by the ECJ in Factortame at paragraphs 55 to 58, including the observation of the Court that:

The Advocate General then continued as follows:

It is clear from this passage, in my judgment, that the Advocate General took essentially the same view as the ECJ, and accepted that the developing jurisprudence of the ECJ on the impact of Articles 43 and 56 in the field of direct taxation had not become “wholly clear” until the decisions in Verkooijen and Manninen. He too gave a “strong steer” that breaches occurring at what were, at the relevant time, the boundaries of the development of the Court’s case-law should not be regarded as sufficiently serious.

The other point which I would emphasise is that both the Advocate General in the present case, and Advocate General Fennelly in Hoechst, considered that the question whether the breaches were excusable or inexcusable had to be viewed objectively. This is reflected in the speeches of at least two members of the House of Lords, Lord Hoffmann and Lord Clyde, in the leading domestic authority on sufficiently serious breach, R v Secretary of State for Transport, ex parte Factortame Ltd (No. 5) [2000] 1 AC 524 (“Factortame No. 5”), to which I will now turn.

The issue in Factortame No. 5 was whether the breaches of Community law resulting from the enactment of the Merchant Shipping Act 1988 were sufficiently serious to give rise to liability on the part of the government for any damage that might subsequently be shown to have been caused to the applicants. The trial of the issue took place after two references had been made to the ECJ, the first of which established that the registration conditions for fishing vessels introduced by the 1988 Act were incompatible with the EEC Treaty, and the second of which laid down the conditions under which a member state could be liable for damages. The 1988 Act, which came into force on 1 December 1988 with a short transitional period expiring on 31 March 1989, was intended to protect British fishing communities by preventing foreign nationals from fishing against the UK’s quota under the common fisheries policy adopted by the Community. The factual background is helpfully summarised as follows in the headnote at [2000] 1 AC 524:

It was unanimously held by the Divisional Court, the Court of Appeal and the House of Lords that the test of sufficiently serious breach was satisfied. Four members of the House of Lords (Lord Slynn, Lord Hoffmann, Lord Hope and Lord Clyde) gave reasoned speeches, while Lord Nicholls agreed with the reasons for dismissing the appeal given by Lord Slynn and Lord Hope.

Although their Lordships expressed their reasons in slightly different ways, the general thrust of the speeches was, to quote again from the headnote (at 525E):

The theme of deliberate risk-taking by the government was emphasised by both Lord Slynn and Lord Hoffmann. As Lord Slynn said at 542C:

Lord Slynn went on to observe that the nationality condition was obviously discriminatory and in breach of Article 52, and that the ECJ had stated bluntly in the second reference that “the nationality condition constituted direct discrimination which was manifestly contrary to Community law” (542E). He pointed out that officials and ministers were clearly aware that there was a risk that if the legislation was adopted it would be held to be contrary to Community law, and quoted from a letter written by the Secretary of State for Transport to the Attorney General on 22 October 1987 in the following terms (543A):

Lord Hoffmann dealt with the point trenchantly at 547G:

Lord Hoffmann went on to give important guidance about the relevance (or rather, as he saw it, the irrelevance) of legal advice, and the need for an objective approach to the question whether the error of law was excusable:

Lord Hope expressed his agreement with the speech of Lord Slynn, but also added some valuable observations of his own. He said of the guidance given by the ECJ that it provided a “helpful but not exhaustive” list of relevant factors to be considered by the national court (550A). He was also inclined not to attach much importance to expressions of opinion by the ECJ about the nature of the breach, because the Court made it clear in its judgment that it was for the national court to assess the seriousness of the breach. As he said at 550B:

Lord Hope then continued at 550C:

Having reviewed the facts, Lord Hope expressed the view at 551F that:

He went on to say that he was unimpressed by the argument that the government had taken and acted upon legal advice:

Lord Clyde observed, like Lord Hope, that the ECJ’s list of factors to be taken into consideration “does not pretend to be complete or exhaustive” (554C). He continued:

Lord Clyde then identified some of the potentially relevant factors, of which I will quote the first four because they seem to me to be of particular relevance in the present case:

I begin with an important negative point. The Revenue have adduced no positive evidence at all directed to the issue of sufficiently serious breach. Furthermore, in response to an application for specific disclosure of documents relating to the issue, the Revenue made it clear that no relevant documents of any description had been identified. Details of the searches conducted by the Revenue are given in a witness statement dated 16 June 2008 of Mr M J A Twyman, a senior lawyer in HMRC Solicitor’s Office who has had conduct of the FII GLO litigation on the Revenue’s behalf since October 2003. This evidence has not been challenged by the claimants. It is therefore clear that, if any consideration at all was given within the Revenue to the question of the compatibility with Community law of the ACT regime, the Case V charge and the FID regime, it was done in an informal manner which has generated no documentary record. In particular, there is no suggestion that formal legal advice on the question was ever sought, either internally or externally, before the commencement of the present proceedings which began with the issue of the BAT claim form on 18 June 2003. This is not a case where legal advice was obtained and privilege has been claimed, or (as must have been the case in Factortame No. 5) waived. The government appears to have been content to proceed on the footing that challenges to the lawfulness of the legislation could safely be left to be dealt with if and when they emerged. Nor can it sensibly be suggested that the matter might have been investigated elsewhere in the government machine, because given the subject matter it is virtually inconceivable that any relevant paperwork would not have been copied to the Revenue, even if it did not originate there.

Mr Aaronson touched briefly on the question of legal advice in his cross-examination of Mr Ramsden, but it soon became clear that Mr Ramsden was not able to speak to the position before the present proceedings were begun.

Against this background, Mr Aaronson sought to develop the thesis that the Revenue had been guilty of culpable negligence in turning a blind eye to the problem and failing to address the questions of compatibility with Community law which arose in connection with the receipt of foreign dividends, notwithstanding a series of events and warnings which (he submitted) should on any reasonable view have alerted them to the existence of the problem and its impact on groups like BAT.

Taking first the adverse impact of the ACT and Case V regimes, and later the FID regime, on UK-based groups with substantial foreign income, Mr Aaronson relied on the copious and unchallenged evidence of lobbying and complaints to be found in the witness statements of (in particular) Mr Hardman, Mr Allvey and Mr Bilton, together with the voluminous exhibits thereto. I have already referred to some of this evidence in the introductory section of this judgment, and I need not set it out in detail because it is in my judgment abundantly clear that, by the late 1980s if not before, the Revenue were well aware of the existence of the surplus ACT problem, and of the vociferous lobbying on the subject by major UK companies, including the 100 Group (composed of the finance directors of the 100 largest UK companies) and another informal group known as the “Surplus ACT Sufferers Club”, the members of which included BAT, ICI and BP.

To take just one example from those singled out by Mr Aaronson in his oral submissions, on 19 December 1989 a meeting took place between representatives of four multi-nationals, including BAT, and senior Revenue officials, including the then Deputy Chairman, Mr J G Isaac. The companies put their case that UK groups which invested heavily abroad were exposed to surplus ACT which became an additional tax suffered by shareholders, and argued that the problem could largely be eliminated within five years if the limit on the ACT set off against MCT were raised from the then proportion of 25/35 to 100%. In response, the Revenue said that the problem of surplus ACT had been foreseen when the Select Committee on corporation tax took evidence in 1971/72, and the imputation system introduced in 1973 had always been intended to be a semi-imputation system. To allow a 100% credit would therefore breach the integrity of the system, and induce foreign parents to declare higher dividends from their UK subsidiaries. Such a solution would “enhance the discrimination of the tax system in favour of UK investment”, and would reduce public revenue by an estimated £200 million a year.

It is interesting to note from the minute of this meeting, as Mr Aaronson observed, that the Revenue appear to have accepted that the system already discriminated in favour of UK investment. Correspondence then ensued with the Financial Secretary to the Treasury, Mr Peter Lilley, and in a letter dated 3 April 1990 Mr Lilley confirmed that the limit on the amount of ACT which could be set off would not be increased. It is perhaps also worth quoting what he said at the end of the letter:

I now turn to the other matters specifically relied upon by Mr Aaronson, and begin with the report of the Ruding Committee in 1992. This committee was established in October 1990 by the European Commission, with the mandate of evaluating the need for greater harmonisation of business taxation within the Community. Among the questions considered by the committee was whether differences in taxation among member states caused major distortions in the internal market, particularly with respect to investment decisions and competition, with a particular focus on those distortions which were considered to be discriminatory. The committee was also asked to consider what specific measures were required at Community level to remove or mitigate such distortions.

In the “executive summary” of their main conclusions, the committee said that one significant source of bias against inward and outward direct investment was “the discriminatory effect of unrelieved imputation taxes … related to distributions by parent companies from profits earned abroad”. The summary of their conclusions and recommendations in Chapter 10 of the report included the following:

The committee’s detailed recommendations for the reform of corporation taxes included measures to remove existing discrimination in the taxation of dividends from profits earned in another member state. To that end, it was recommended:

Mr Aaronson pointed out that these recommendations foreshadowed the approach taken by the ECJ in the present case to the lawfulness of the Case V charge and the denial of a tax credit to foreign-source dividends. However, the more important point, in my judgment, is that there is no suggestion anywhere in the Ruding Report that the discrimination which they identified was in any way contrary to Community law.

I was not referred to any domestic research in the UK which equalled the breadth and scope of the Ruding Report, but the bundles do include a first draft of a paper by the Adam Smith Institute, with the rather tendentious title “UK Tax Prejudice Against Trading Abroad”, which was sent to Mr Etherington on 26 July 1989 by the director of the Institute, Dr Eamonn Butler. The Adam Smith Institute describes itself as a pro-market think tank, which is independent, non-profit making and not affiliated to any political party. The draft paper discussed the problem of surplus ACT, and in Part VII, headed “The European Dimension”, argued that “[t]ax prejudice against dividends from overseas income is prejudice in favour of activities in the domestic market to the disadvantage of activities abroad”, and that the logic of the single European market was that the domestic market should now be treated as the market of the Community member states. The author then commented in paragraph 48:

In his budget speech in March 1993, the then Chancellor, Mr Norman Lamont, announced a series of changes to dividend taxation, accompanied by the publication of a Consultative Document on possible changes designed to alleviate the problem of surplus ACT. On 24 June 1993 the Institute for Fiscal Studies hosted a conference on the subject of surplus ACT, at which papers were delivered by a number of speakers, including the then Director of the Company Tax Division at the Inland Revenue, Mr Peter Lewis. He had been asked to speak on the subject from the government’s point of view. In his paper he referred to three proposals which the government had put forward to help with surplus ACT, including the introduction of a FID regime. He expressly said that although shareholders would be treated as having borne tax at the lower rate of income tax when it came to calculating their tax liability, FIDs would not carry a tax credit which could be paid to exempt, non-liable or overseas shareholders. That was the proposal which the government had put out for consultation, although without any commitment, at that stage, to legislate on it. He continued:

Dealing with the detail of the FID proposals, Mr Lewis acknowledged that dividend “streaming” was the major issue, and referred to a widespread view that the FID scheme “would be much more useful and attractive if streaming arrangements could be adopted”. The other side of the coin, however, was exchequer cost:

Finally, Mr Aaronson referred me to a number of articles in learned journals, including:

The last of these articles referred to a case in 1995 where the European Commission had addressed a note to the German Government claiming that the German system of taxation of foreign dividends, which gave credit for tax paid on German dividends but not for tax paid on foreign dividends, infringed Articles 43 and 56. The situation was subsequently remedied by the German legislature, which dealt with the problem by applying the same “half income” taxation system to both domestic and cross-border dividends. In his article, Mr Lodin pointed out that the German imputation system, like the Finnish, French and British systems, did not give dividends on foreign shares the same preferential treatment as it accorded to domestic dividends. He then addressed the question whether the difference in treatment constituted a violation of the EU Treaty, concluding that the relevant rules were not clear enough to make the outcome obvious, there was very little guidance to be found in earlier decisions of the ECJ, and it was difficult to predict what the outcome would be.

The fact that Mr Lodin found the outcome so difficult to predict in 1998, a year after the last of the enhanced FIDs was paid, reflects the uncertainty acknowledged by the ECJ itself in the present case, and which continued at least until the decision in Verkooijen in June 2000. Moreover, it is the evidence of the claimants’ own witnesses that they paid all of the tax in dispute on the footing that they believed it to be lawfully due, and had no reason to suspect the contrary before June 2000 at the earliest. So, for example, Mr Anthony Cohn, who was a Tax Manager with BAT Holdings, said in his first witness statement dated 13 May 2004:

To similar effect, Mr Hardman repeatedly emphasises that he never considered the relevant provisions of UK law to be unlawful. In his first witness statement, dated 18 May 2004, he confirmed the accuracy of what Mr Cohn had said in the passage which I have quoted, and continued:

I see no reason to doubt this evidence, and find that nobody within the BAT group questioned the lawfulness of the relevant UK legislation at any time before June 2000. All the disputed tax which was paid up to that date was paid in the firm belief that it was lawfully due.

The most significant factor in the present case, in my judgment, is the very real uncertainty about the impact of Community law on the relevant UK legislation. That uncertainty has been recognised by the ECJ itself as having persisted at least until the decision in Verkooijen. I bear in mind Lord Hope’s warning that the national court should not be over-influenced by views expressed by the ECJ on a question which it is for the national court to determine; but where the view relates to the interpretation and development of Community law, and not to the evaluation of a domestic factor such as the conduct of the Member State, it seems to me obvious that it must be accorded the greatest respect. It would be a rash national judge who disagreed with the ECJ about the lack of clarity of its own case-law in an area as difficult and complex as the cross-border taxation of dividends.

It is also relevant to bear in mind certain features of the relevant legislation, both in the Community and in the domestic context. I take first the relevant rules of Community law. This is not a case where the claimants can point to specific infringements of clearly-defined provisions, for example in the Parent-Subsidiary Directive. Nor is it a case of blatant infringement of a basic principle of Community law, such as the UK government’s discrimination on grounds of nationality in Factortame No. 5. The unlawfulness of discrimination on grounds of nationality in the domestic tax law of Member States was firmly established by the landmark decision of the ECJ in Commission v France (Case 270/83, [1986] ECR-273), although even there the Court left open the possibility that a residence-based distinction might in certain circumstances be justified. However, the present case does not involve any form of direct discrimination on grounds of nationality, or on grounds of residence. The relevant claimant companies are all registered and resident in the UK, and what is in issue is the treatment in the UK of dividends received from foreign subsidiaries. There was admittedly discrimination between the way in which UK tax law treated domestic dividends and foreign dividends, with domestic dividends receiving the more favourable treatment, but whether this form of discrimination involved a breach of Articles 43 and 56 remained unclear until the decision in Verkooijen.

This was indeed recognised by Mr Aaronson, who accepted more than once in the course of his oral submissions that, if the Revenue had sought advice on the question, they could reasonably have been advised that the position was unclear, and that there was no need to amend the relevant UK legislation unless and until the position had been clarified by the ECJ. Mr Aaronson was in my view right to make this concession, and since the question whether the Revenue’s failure to amend the legislation to make it compatible with Community law has to be judged objectively, this seems to me a very significant point.

Turning now to the relevant provisions of UK law, it is in my view material to bear in mind that the legislation in issue (apart from that relating to FIDs) was of long standing, and none of it (including the legislation relating to FIDs) was introduced with questionable motives, or involved the taking of a deliberate risk that it might be held to infringe Community law. The Case V charge dates back to the nineteenth century, and the partial imputation system of corporate taxation introduced in 1973 goes back almost as far as the UK’s accession to the EEC (as it then was) in 1972. I am confident that it would never have occurred to anybody in the early 1970s, or for some considerable time thereafter, that the national law of direct taxation in a member state might in some way infringe the Community principles of freedom of establishment and free movement of capital. In contra-distinction to VAT, which was a creation of Community law, direct taxation was generally seen as quintessentially a matter for national sovereignty to determine, and an area where residence would almost inevitably be used as a discriminating factor.

The fiscal and legal landscape had of course changed dramatically by the time when FIDs were introduced in 1994, but the critical point here, in my judgment, is that the whole purpose of the FID regime was to alleviate the problem of surplus ACT for UK companies with substantial foreign income. The only thing wrong with the system was that, as the ECJ has now held, it did not go quite far enough to secure compliance with Community law. It is at this point that the allure of hindsight must be firmly resisted. The relevant principles of EU law were still, objectively, uncertain, and shrouded in a mist which only began to lift when Verkooijen was decided some six years later. Of course, the multi-national groups which had been lobbying the government would have liked the FID regime to go further, and to provide for streaming of dividends; but that is a quite separate point, and nothing to do with Community law as it was perceived at the time. Nobody could reasonably criticise the government for taking the cost of the proposals into account, and for adopting a solution which was judged to be affordable at a time when the country was still emerging from the last recession.

Where the Revenue can justly be criticised, in my view, is for their apparently total failure to give any serious consideration to the potential Community law issues which arose from the Case V charge, the FII regime and the introduction of FIDs. I am surprised, to put it mildly, by the complete absence of any documentary evidence to suggest that any advice was ever sought on these questions, either internally or externally. I find it even more surprising that nothing was apparently done even after the commencement of the Hoechst litigation in 1995, when it should surely have been obvious to the Revenue, even if it had not been before, that the compatibility with Community law of all aspects of the ACT and FII systems, and of the cross-border taxation of dividends, was likely to come under very close scrutiny. It must also have been obvious that huge amounts of public money were potentially at stake, and that the long-standing grievance of surplus ACT would encourage affected taxpayers to leave no stone unturned in their quest for compensation. Yet absolutely nothing appears to have been done by the government to prepare for this onslaught.

Perhaps naively, I would expect the Revenue to have initiated a detailed review and “health-check” of all the UK’s tax legislation with a significant cross-border element soon after the decision in Commission v France, and I would also expect the matter to have been kept under regular review thereafter. In the circumstances, I find it hard to escape the impression that the Revenue’s general approach to the whole question was one of insular insouciance.

These are matters which I think it is right for me to take into account, but ultimately they do not count for very much on an objective appraisal of the situation. If, with the benefit of competent legal advice, the government could reasonably have decided to leave the existing legislation unchanged, and to introduce the FID regime in exactly the same form as they actually did, I think it would be unrealistic to regard their failure to take legal advice as a decisive factor in the balancing exercise which I have to perform. It is not alleged that the Revenue or the government acted in bad faith, and although their conduct is open to criticism, it cannot in my view be said that they “manifestly and gravely disregarded the limits on [their] discretion”. It would also seem paradoxical to find that a sufficiently serious breach was established, in circumstances where the claimants themselves, who had the resources and the motivation to investigate the question in depth, took the view until at least June 2000 that the legislation in question was not open to challenge.

The point which I am making is neatly illustrated by the contrast between what the Advocate General said in the present case on sufficiently serious breach, and his comments about the conduct of the UK in the context of the government’s application for a temporal limitation on the judgment. He dealt with this aspect of the case in paragraphs 140 to 147 of his opinion. Having held that the plea of temporal limitation should be rejected, without more, on the basis of insufficient substantiation (paragraph 145), he then went on to make some observations on the merits in case the Court should disagree with his view on the first ground. His observations included the following, in paragraph 146:

It is only fair to add that the Advocate General said he was reluctant to make any observations on the merits, without having heard substantial argument. However, the important point for present purposes is that although the Advocate General regarded this observation as material to the issue of temporal limitation, he did not make any similar point in his discussion of sufficiently serious breach, and on the contrary gave the strong steer to the national court which I have already described. In a word, failure to appreciate that there is a risk of non-compliance with Community law is one thing; meeting the objective test of sufficiently serious breach, which is set at a fairly high level, is another thing.

For all these reasons, I conclude that the claimants have failed to establish any sufficiently serious breach on the part of the Revenue or any other organ of the UK government. It follows from this that the claims for Factortame damages must in my judgment fail.

The effect of section 32(1)(c) of the Limitation Act 1980, as I have already pointed out, is that a mistake-based restitution claim may be brought up to six years after the date on which the mistake either was, or could with reasonable diligence have been, discovered by the claimant. On 18 July 2003, Park J held in DMG that this cause of action was in principle available to a person who wished to recover tax paid by mistake, and dismissed the argument (apparently supported by a passage in the speech of Lord Goff in Kleinwort Benson) that overpaid tax could be recovered only by a Woolwich claim or under the relevant statutory regimes, notably that contained in TMA 1970 section 33. Woolwich claims are subject to the usual limitation period of six years, subject to the possible argument noted in paragraph 244 above that section 32(1)(c) can apply in any case where the claimant is in fact seeking relief from the consequences of a mistake, even if mistake is not an essential ingredient of his cause of action. The statutory provisions for repayment of tax are likewise subject to a time limit of approximately six years.

Against this background, the potential exposure of the public purse to mistake-based restitution claims for wrongly paid tax was obviously huge, once the cause of action had been recognised by the court. The potential exposure was particularly great in cases where the claimant had a San Giorgio claim to repayment of unlawfully levied tax under Community law, since in such cases the claim could (and often did) go back as far as 1973, with compound interest (the right to recover which was confirmed by the House of Lords in Sempra Metals) on top. Moreover, the Community law principle of effectiveness would apply in its full rigour to such claims. The Revenue appealed against the judgment of Park J, and were in fact successful in the Court of Appeal in February 2005 before ultimately losing in the House of Lords in October 2006. However, the outcome of those appeals was, at the time, impossible to predict with any confidence.

It is therefore hardly surprising that on 8 September 2003 the Paymaster General (Ms Dawn Primarolo) announced that legislation would be included in the Finance Bill 2004 with the object of limiting the period for claiming repayments of overpaid tax to six years from the date of the original payment. In a written ministerial statement released on the following day, she said:

The draft legislation was enacted as section 320 of the Finance Act 2004, which received Royal Assent on 22 July 2004. The most important parts of the section read as follows:

It will be seen that the effect of subsection (1) was to disapply section 32(1)(c) in relation to any mistake of law relating to a taxation matter under the care and management of the Commissioners of Inland Revenue, that is to say all types of direct taxation such as income tax, capital gains tax, inheritance tax and corporation tax. Subsection (6) made it clear that the section would apply to all actions or claims for relief from the consequences of a mistake of law, however formulated. There can be no doubt that the section covers the cause of action recognised in DMG.

As I have already mentioned, on 25 October 2006 the House of Lords gave judgment in DMG, reversing the decision of the Court of Appeal and restoring the decision of Park J. Meanwhile, on 6 April 2006, the Advocate General had delivered his opinion in the present case. On 2 October, the parties were notified that the ECJ would deliver its judgment on 12 December 2006. On 27 November, the UK government requested the Court to reopen the procedure. On the morning of 6 December 2006, this application was rejected.

Later on the same day, the UK government announced that further legislation would be introduced to extend the disapplication of section 32(1)(c) to actions brought before, as well as after, 8 September 2003 for relief from the consequences of a mistake of law in the field of direct taxation. Legislation to this effect was in due course enacted as section 107 of the Finance Act 2007, which received Royal Assent on 19 July 2007.

Section 107 provides as follows:

The basic issue which I have to determine in this part of the case is whether it is open to the Revenue to rely on section 320 of the Finance Act 2004 and/or section 107 of the Finance Act 2007 as defences to mistake-based claims for restitution which pre-date the commencement of the claim by more than six years. For technical reasons, the test claimant in relation to section 320 is not any member of the BAT group, but three members of the Aegis group. It is unnecessary for me to refer to the Aegis pleadings in any detail. It is enough to say that the claimants contend that section 320 is invalid and of no effect, in that it breaches Community law. No separate point arises in relation to section 107 on the Aegis pleadings, but section 107 is in issue in the BAT claims and raises essentially the same questions.

There was no discernible difference between the parties about the principles of Community law which I should apply in this context. For that reason, it is unnecessary for me to review the case-law of the ECJ in any detail, and although I was taken through a number of cases by Mr Aaronson I think it will be enough if I cite the statement of the relevant principles by the ECJ in Marks & Spencer (Case C-62/00, [2002] ECR I-6325, [2003] QB 866).

The issue in Marks & Spencer was whether the Commissioners of Customs and Excise could rely on a reduction in the limitation period for claims to refunds of VAT from six to three years, which had been announced in July 1996 and was enacted, without a transitional period, by section 47(1) of the Finance Act 1997. That section amended section 80(4) of the Value Added Tax Act 1994, enacted the new limitation period, and provided that it was to be deemed to have come into force on the date of the announcement in July 1996. The claimant, Marks & Spencer, made a claim on 31 October 1996 for a refund of VAT overpaid as from May 1991 in connection with certain voucher schemes. The Commissioners’ stance was that they were only prepared to pay such part of the claim as fell within the new three year limitation period.

On a reference for a preliminary ruling made by the Court of Appeal, the ECJ gave guidance which is helpfully summarised as follows in the headnote at [2003] QB 867, holding:

For present purposes, the most important part of the judgment is contained in paragraphs 36 and following, where after referring to the principle of effectiveness, and earlier case-law which had held that, in the interests of legal certainty, it is compatible with Community law to lay down reasonable time limits for bringing proceedings, the Court continued as follows:

Mr Ewart submitted, and I would accept, that the relevant principles can be summarised in the following propositions:

The claimants argue that the Revenue are precluded from relying on sections 320 and 107 for two separate reasons: first, the restrictions were introduced without any transitional arrangements; and secondly, they were introduced with the specific intention of limiting the consequences of judgments of the ECJ.

I will consider first the question of transitional arrangements, because it is a short and relatively simple point, and if it is decided in the claimants’ favour there is then no need to consider the much more difficult second argument.

It soon became apparent from the oral submissions of Mr Aaronson and Mr Ewart that there is a lot of common ground in relation to the first argument. The Revenue accept that the Community principles of effectiveness and protection of legitimate expectations require legislation which cuts down an existing limitation period not only to ensure that the new limitation period is reasonable, but also to provide adequate transitional arrangements: see Marks & Spencer, paragraphs 38 and 43 to 47. The Revenue also accept that both section 320 and section 107 had the effect of cutting down the limitation period which would otherwise have applied to restitutionary claims to recover tax paid by mistake, i.e. the cause of action recognised and upheld by Park J, and subsequently by the House of Lords, in DMG. The retrospective effect of judicial decision-making means that this remedy must now be treated as having always been available to the claimants, even though they were doubtless unaware of it until quite recently. As a matter of English legal theory, at all times until 8 September 2003 the claimants were in a position to claim restitution of tax paid by mistake, and to take advantage of section 32(1)(c) in relation to that cause of action. Once Park J had delivered his judgment, this right emerged from the shadows of legal theory and became a matter of hard practical reality. In (say) August 2003, the claimants had the right to start such proceedings in any case where the mistake upon which they relied either had been, or should have been, discovered within the last six years. Even more importantly, they also had the right to start such proceedings at any time in the future in respect of mistakes which had not yet come to light, subject only to the requirements of section 32(1)(c). Hence the government’s urgent response, the announcement of 8 September 2003, and the enactment of section 320.

The effect of section 320 was to reduce the limitation period applicable to such claims to a fixed period of six years from the date when the tax was paid, regardless of when (or even whether) the mistake of law either was, or could with reasonable diligence have been, discovered. Furthermore, no transitional arrangements of any kind were provided for the commencement of claims to which section 320 would apply. Thus a claimant who on 7 September 2003 had the benefit of section 32(1)(c) in respect of mistakes discovered within the last six years, and unknown mistakes which might emerge at any time in the future, and who could in principle claim restitution of tax levied in breach of Community law for periods going back to 1972, woke up on 8 September to find that all claims brought on or after that date would be limited to recovery of tax paid within six years before the commencement of the claim. Existing claims, however, were not affected; although from 20 November 2003 onwards, the scope of amendments to existing claims was severely restricted (on that aspect of the matter, see the discussion in my judgment in Europcar at paragraphs 65 to 78).

Section 107 then introduced a further, entirely retrospective, restriction by, in effect, extending the operation of section 320 so that it also covered all existing claims brought before 8 September 2003, subject only to the saving in subsection (3) which preserved the effect of the judgment of the House of Lords in DMG for the parties to that case and the relevant members of the ACT GLO. There was again no transitional provision of any description, even though the effect of the section was to deprive claimants who had started their proceedings before 8 September 2003 of the benefit of a statutory limitation rule which had been in force at the time, and in many cases to effect a drastic reduction in the value of their claims. For example, a claim begun in early 2003 seeking repayment of tax levied in breach of Community law since 1973, together with compound interest, would now be limited to the recovery of tax paid from early 1997 onwards, together with correspondingly reduced amounts of interest.

Finally, Mr Ewart realistically accepted that, if the two sections were held to breach Community law, they would simply have to be disapplied in cases where the claimant was entitled to rely on that breach. In other words, he accepted that there was no process either of construction, or of remoulding, of the legislation by which reasonable transitional arrangements could be read in, where Parliament had specifically and deliberately enacted the legislation without any such provisions, and where it would clearly go well beyond any proper exercise of the Revenue’s care and management functions to devise appropriate provisions on an extra-statutory basis. Nor is it a case where the court could properly make its own assessment of the form that reasonable transitional provisions should take. In making this concession, Mr Ewart had well in mind the views expressed by the majority of the House of Lords in Fleming v Revenue and Customs Commissioners [2008] UKHL 2, [2008] 1 WLR 195, to which I was referred by both sides.

Against this background, the point upon which Mr Ewart founded the Revenue’s argument was (of necessity) a very narrow one. He argued that sections 320 and 107 were untouched by the principles of Community law which I have mentioned, at any rate in the context of the present case, because the only domestic causes of action which are needed to satisfy the Community principle of effectiveness in relation to San Giorgio claims are the Woolwich cause of action in restitution and the right to claim damages for breach of statutory duty. However, I have already rejected that argument in the section of this judgment dealing with remedies, and held that the DMG cause of action for mistake-based restitution is also needed in order to satisfy the Community principle of effectiveness in relation to San Giorgio claims: see paragraph 260 above.

If that is right, it must then follow in my judgment that sections 320 and 107 breach Community law in that they purport to curtail the limitation period applicable to such claims without providing any, let alone adequate, transitional arrangements. Accordingly, the two sections must be disapplied in relation to such claims, because in the present context that is the only way in which effect can be given to the superior legal order of Community law.

For these reasons, I conclude that it is not open to the Revenue to rely on sections 320 and 107 as defences to the test claims, where such claims are properly classified in English law as claims in restitution based on a mistake of law, and where the claims are necessary in order to provide the claimants with an effective remedy under the San Giorgio principle.

Having reached this conclusion, I prefer to express no view on the claimants’ alternative argument that the two sections, or at any rate section 107, were specifically intended to limit the consequences of a judgment of the ECJ. I say that for two reasons in particular.

My first reason is that, if the Revenue were to succeed in persuading a higher court that the Woolwich cause of action alone is sufficient to satisfy the claimants’ rights to recover overpaid tax under Community law, I am unable to see any answer to Mr Ewart’s submission that the sections would be unaffected by Community law in the context of the present case. The claimants would therefore gain nothing by establishing their alternative argument which they could not gain far more simply by success on their first argument.

My second reason is that the alternative argument raises potentially very difficult questions, both of law and of fact, which Mr Aaronson frankly admitted would probably necessitate a further reference to the ECJ. To give a flavour of the issues which might arise, Mr Ewart submitted that the only material admissible in answering the question is material which would be admissible in a domestic context for ascertaining the legislative aim of the two provisions. He relied in support of this submission on the decisions of the House of Lords in Wilson v First County Trust [2003] UKHL 40, [2004] 1 AC 816, and R (Countryside Alliance) v Attorney General [2007] UKHL 52, [2008] 1 AC 719. Mr Aaronson, however, argued that the timing of the announcement of the legislation which became section 107 on 6 December 2006 speaks for itself, and places an evidential burden on the Revenue to displace the inference that section 107 was specifically (although not necessarily solely) intended to limit the consequences of the judgment of the ECJ in the present case. He went on to submit that a proper investigation of the question, as a matter of Community law, would involve an examination of the subjective intentions of all the branches of government which were involved in the decision to legislate, that correspondingly wide disclosure should therefore be ordered, and that adverse inferences could be drawn if disclosure of relevant documents were to be refused on grounds of privilege or public interest immunity, even though no such adverse inferences could be drawn in a purely domestic context.

It is obvious, to my mind, that these contentions raise issues of far-reaching legal and constitutional significance, and for that reason alone I think it is undesirable for the court to pronounce upon them unless it has to.

I come finally, almost by way of footnote, to an argument which was briefly touched upon by Mr Ewart at various stages in his submissions, but which I am satisfied does not really advance the Revenue’s case. It is not specifically an argument about limitation, and indeed Mr Ewart mainly relied on it as a possible defence to the Case V claims for restitution and/or damages; but this section of my judgment seems to me as good a place as any to deal with it.

The general point which Mr Ewart seeks to establish is that there can be no recovery at common law, whether in damages or in any kind of restitution, where the claim falls within the ambit of the statutory regime for repayment of tax paid by error or mistake in TMA 1970 section 33, or equivalent provisions.

Section 33 has been through a number of changes of form in the period covered by the present claims, but for present purposes nothing turns on the differences between the various versions. As the section stood in 1994, it provided as follows:

Subsequent amendments were necessary to deal with the introduction of self-assessment, first for individuals and then for companies, and the corresponding provisions relating to corporation tax are now to be found in paragraph 51 of schedule 18 to the Finance Act 1998.

It will be noted that the statutory regime under section 33 is subject to a number of limitations. For example, it applies only to tax paid under an assessment. It was therefore accepted by the Revenue in DMG that it could not apply to the payment of ACT, which is not in the normal way subject to assessment. Again, although the section provided a remedy for tax overpaid under a mistake of law for decades before the decisions of the House of Lords in Kleinwort Benson and DMG, there has always been a “settled practice” exception where the return “was in fact made on the basis or in accordance with the practice generally prevailing at the time when the return was made”.

The Court of Appeal has recently held in Monro v Revenue and Customs Commissioners [2008] EWCA Civ 306, [2008] 3 WLR 734, (“Monro”) that, where section 33 applies, it was intended by Parliament to provide an exclusive regime, and a taxpayer cannot circumvent it by bringing a common law claim in restitution. That conclusion drew on dicta by Lord Hoffmann and Lord Walker in DMG (which are conveniently quoted by Longmore LJ in paragraph 38 of the judgment) and is, of course, binding on me. However, it does not seem to me to assist the Revenue in the present case, for two main reasons.

First, the House of Lords in DMG rejected a more far-reaching argument advanced by the Revenue that the enactment of section 33 implicitly excluded any common law remedies for the recovery of overpaid tax even in circumstances which did not fall within the ambit of section 33. As Lord Hoffmann said in paragraph 19, that submission went “much too far”. It is no doubt for this reason that Mr Ewart did not seek to rely on this argument in relation to the claims for recovery of unlawfully levied ACT, because (as I have pointed out) ACT was payable without any assessment, and therefore fell outside the scope of section 33.

The second reason is that, even where the tax in question is payable under an assessment and section 33 applies, it must yield to the principle of effectiveness under Community law where the claimant seeks domestic redress for a San Giorgio claim. This is the point which the ECJ established in Fantask, and it is recognised, although slightly obliquely, by the Court of Appeal in Monro: see the judgment of Arden LJ at paragraph 34.

I have already summarised my conclusions on the questions relating to liability discussed in sections II to IV of this judgment: see paragraphs 193 to 200 above. I will now summarise my conclusions on the questions relating to remedies and defences which I have discussed in sections V to VIII.

I have held that the San Giorgio principle extends as a matter of Community law to all the claims for repayment of unduly levied tax and to associated interest and loss of use claims (paragraph 240), and that the classification of the corporate tree claims should be referred to the ECJ, although my own view is that if such claims are held to involve a breach of Community law they too should be classified as San Giorgio claims (paragraph 241).

I have held that the Community principles of equivalence and effectiveness require English law to provide two restitutionary remedies for San Giorgio claims, namely the Woolwich cause of action and the cause of action for recovery of tax paid by mistake which was recognised by the House of Lords in DMG (paragraph 260). However, as a matter of English law the DMG cause of action should in my judgment be limited to claims for the repayment of tax paid by mistake and the reversal of any directly associated benefits retained by the Revenue as a result of the mistaken payment (paragraph 264). I do not rule out the possibility that this principle could also extend to the discharge of any directly related obligations on the Revenue’s behalf, because in such a situation the Revenue could again be said to have been unjustly enriched at the claimants’ expense. However, the “but for” test of causation propounded by the claimants as the touchstone for recovery in the English law of restitution is in my judgment wrong in principle, and should be rejected (paragraphs 264 to 5).

I have attempted to apply the above principles to:

I have found, on the facts, that the primary reason why the FIDs were enhanced in the BAT group was to compensate exempt shareholders for the absence of a tax credit for them under the FID regime (paragraphs 277 to 302).

I have held that a defence of change of position is in principle available to the Revenue in respect of any mistake-based restitutionary claims which go beyond San Giorgio claims, although not in respect of Woolwich claims, and that where the defence is available it is likely to succeed on the facts (paragraphs 335 to 346). I have discussed the parts of the evidence of Mr Ramsden which relate to this issue in paragraphs 349 to 351. I have also expressed the view that the question of what it would now cost the government to meet an adverse judgment, and the evidence directed to that question, are irrelevant to the defence of change of position (paragraphs 347 to 8, and 352).

I have concluded that the claimants have failed to establish any sufficiently serious breach on the part of the Revenue or the UK government, and that the claims for Factortame damages must accordingly fail (paragraph 404).

I have held that the Revenue are precluded from relying upon the curtailed limitation period for mistake-based claims introduced by section 320 of the Finance Act 2004, and retrospectively extended by section 107 of the Finance Act 2007, because in breach of Community law the two sections purported to curtail the limitation period applicable to San Giorgio mistake claims without providing any transitional arrangements (paragraph 427).

I have declined to rule on the claimants’ alternative submission that the two sections were introduced with the specific intention of limiting the consequences of judgments of the ECJ (paragraphs 428 to 431).

I have held that, although section 33 (and other equivalent provisions in domestic UK tax legislation) by implication exclude any other domestic remedy in cases where the facts fall within the scope of the section, they do not exclude other domestic remedies in cases where for any reason the facts fall outside the scope of the section, and they are in any event overridden by Community law where it is necessary to provide an effective domestic remedy for San Giorgio claims (paragraphs 432 to 439).

Finally, I should make it clear that, where I have held that further questions should be referred to the ECJ, I do not mean to pre-empt the question of whether it is appropriate to make a further reference at this stage of the litigation. As I indicated more than once during the hearing, my provisional inclination would be not to order a reference at this stage, if the relevant part of my judgment is to be appealed. However, this is a question to which further consideration can, if necessary, be given after this judgment has been handed down.