Thin Cap Group Litigation, Test Claimants In v Revenue and Customs

Introduction 32-34

Lankhorst-Hohorst 35-47

The decision of the ECJ in the present case 48-64

Discussion 65-75

Introduction 76-81

Is a conforming interpretation possible? 82-90

Disapplication and its effect 91-99

Introduction 100–103

Volvo 104–116

Lafarge 117–135

IBM 136–160

Siemens 161–172

Standard Bank 173–179

The evidence of Mr Brooks and Mr Black 180–187

Introduction 196–198

The categories of claim 199–208

The guidance given by the ECJ 209–212

Discussion 213–218

Does English law provide an adequate remedy? 219–226

Are any claims which are not San Giorgio claims

Change of position 236

Introduction 237–245

Disclosure of documents by the Revenue 246–254

Evidence 255–309

Discussion 310–338

Introduction 339–340

Section 32(1)(c) of the Limitation Act 1980 341

When does the limitation period start to run? 342

Section 320 FA 2004 and section 107 FA 2007 343

Section 33 TMA 1970 344

Introduction 345-348

The relevant provisions of the US Treaty 349-351

Discussion 352-362

What is thin capitalisation (or “thin cap” for short)? A helpful introduction to the subject may be found in the opinion of Advocate General Geelhoed in the reference for a preliminary ruling in the present case to the Court of Justice of the European Communities (“the ECJ”). Under the heading “Context and rationale of “thin cap” rules”, he said this:

“3.

There are two principal means of corporate finance: debt and equity finance. Many Member States draw a distinction in the direct tax treatment of these two forms of finance. In the case of debt finance, companies are generally permitted to deduct interest payments on loans for the purpose of calculating their taxable profits (i.e., pre-tax), on the basis that this constitutes current expenditure incurred for the pursuit of the business activities. In the case of equity finance, however, companies are not permitted to deduct distributions paid to shareholders from their pre-tax profits; rather, dividends are paid from taxed earnings.

4.

This difference in tax treatment means that, in the context of a corporate group, it may be advantageous for a parent company to finance one of the group members by means of loans rather than equity. The tax incentive to do so is particularly evident if the subsidiary is located in a relatively “high-tax” jurisdiction, while the parent company (or indeed an intermediate group company which provides the loan) is located in a lower-tax jurisdiction. In such circumstances, what is in substance equity investment may be presented in the form of debt in order to obtain a more favourable tax treatment. This phenomenon is termed “thin capitalisation”. By thus manipulating the manner in which capital is provided, a parent company can effectively choose where it wishes profits to be taxed.

5.

Many States, viewing thin capitalisation as abusive, have implemented measures aimed at countering this abuse. These measures typically provide for loans which fulfil certain criteria to be regarded for tax purposes as disguised equity capital. This means that interest payments are recharacterised as profit distributions, so the subsidiary cannot deduct all or part of the interest payment from its taxable income, and the payment is subject to any applicable rules on dividend taxation.”

The UK tax system draws the distinction noted by the Advocate General in paragraph 3, and although the UK’s national tax legislation does not, in terms, deal with “thin capitalisation” as a separate topic, it has nevertheless sought in various ways to counter the perceived opportunity for abuse to which the Advocate General refers. The questions that I have to consider in the present case arise out of challenges to the compatibility of the UK legislation and rules with Community law, made by multinational groups which have invested in UK-resident subsidiaries by means of loan finance.

The potential vulnerability of the UK legislation to such a challenge first became apparent to taxpayers and their advisers when the ECJ delivered judgment on 12 December 2002 in Case C-324/00 Lankhorst-Hohorst GmbH v Finanzamt Steinfurt, [2002] ECR I-11779, [2003] STC 606 (“Lankhorst-Hohorst”). In that case, which I will need to examine in more detail later in this judgment, the ECJ held that the German thin cap rules breached Article 43 EC (freedom of establishment). The German rules considered in that case were in many respects very different from the UK rules, but the judgment made it clear that Article 43 was likely to be engaged in cases where a member state’s thin cap rules did not apply to similar lending by a resident parent company, and that the grounds upon which a national measure which in principle breached Article 43 could be justified were likely to be fairly narrow. In particular, paragraph 37 of the judgment suggested that any justification based on the risk of tax avoidance would have no hope of success unless the relevant national rules were specifically targeted and went no further than was necessary to achieve that purpose:

“37.

As regards more specifically the justification based on the risk of tax evasion, it is important to note that the legislation at issue here does not have the specific purpose of preventing wholly artificial arrangements, designed to circumvent German tax legislation, from attracting a tax benefit, but applies generally to any situation in which the parent company has its seat, for whatever reason, outside the Federal Republic of Germany. Such a situation does not, of itself, entail a risk of tax evasion, since such a company will in any event be subject to the tax legislation of the state in which it is established … .”

Following Lankhorst-Hohorst, numerous claims were brought in the High Court by UK-resident subsidiaries of multinational groups, and on 30 July 2003 a Group Litigation Order (“GLO”) was made by Chief Master Winegarten for the orderly management and disposal of the claims. The proceedings are known as the Thin Cap Group Litigation, and the order of 30 July 2003 as the Thin Cap GLO. Within the group litigation appropriate test claimants have been identified. Two corporate groups, Lafarge and Volvo, with their headquarters in France and Sweden respectively, have been test claimants from an early stage, and agreed statements of facts relating to them were included in the order for reference to the ECJ which was made by Park J on 21 December 2004. The other three test groups of companies whose claims are now before me have been added since the date of the reference. They are IBM, Siemens and Standard Bank, which have their respective headquarters in the USA, Germany and the Republic of South Africa.

The questions posed in the order for reference were, in summary, as follows. Question 1 asked (in effect) whether the UK’s thin cap rules were contrary to Articles 43, 49 or 56 EC, in circumstances where the loan finance to the UK-resident borrowing company was granted by a parent company resident in another member state. Question 2 asked what difference (if any) it would make to the answer to Question 1 in various factual situations where the states of residence of the lending and/or the direct or indirect parent company were not member states (“third countries”). Question 3 asked whether it would make any difference to the answers to Questions 1 and 2 if it could be shown that the borrowing constituted an abuse of rights or was part of an artificial arrangement designed to circumvent the tax law of the member state of the borrowing company, and (if so) what guidance the ECJ thought it appropriate to provide as to what constituted such an abuse or artificial arrangement. Question 4 may for present purposes be ignored, because it proceeded on the footing that there was a restriction on the movement of capital between member states within Article 56 EC, and the ECJ has now held that the only Article engaged in the present context is Article 43. Questions 5 to 10 then asked a number of detailed questions relating to remedies, very similar to the questions asked in the FII group litigation with which I have already dealt at considerable length in my judgment in Test Claimants in the FII Group Litigation v Revenue and Customs Commissioners [2008] EWHC 2893 (Ch), [2009] STC 254 (“FII Chancery”).

The Advocate General delivered his opinion on 29 June 2006, and the ECJ gave judgment on 13 March 2007: Test Claimants in the Thin Cap Group Litigation v IRC, Case C-524/04, [2007] ECR I-2107, [2007] STC 906. The case was heard by the Grand Chamber of the Court, and both the Advocate General (Geelhoed) and the Juge Rapporteur (Lenaerts) were the same as in the FII case, in which judgment had been given four months earlier on 12 December 2006: Test Claimants in the FII Group Litigation v IRC, Case C-446/04, [2006] ECR I-11753, [2007] STC 326 (“FII”).

I will not at this stage analyse the judgment of the ECJ in any detail, but will confine myself to some fairly brief observations.

First, as I have already indicated, on the issues of liability (Questions 1 to 3 in the order for reference) the Court held that the only freedom of movement which applied was the freedom of establishment in Article 43 EC. The basic reason for this was that the relevant UK legislation was “targeted only at relations within a group of companies”: see paragraph 33 of the judgment and the cases there cited, Case C-196/04 Cadbury Schweppes and Cadbury Schweppes Overseas [2006] ECR I-7995, [2007] Ch 30 (“Cadbury Schweppes”) at paragraph 32 and FII at paragraph 118.

Secondly, the Court held that the UK’s thin cap rules at all material times involved a difference in treatment between resident and non-resident borrowing companies which constituted a restriction on freedom of establishment; that the restriction could not be justified by the need to ensure cohesion of the UK’s tax system; but that the restriction would be justified on the ground of prevention of abusive practices so long as it did not go beyond what was necessary to attain that objective: see paragraphs 36 to 78 of the judgment.

Thirdly, the Court held that the question whether the latter test was satisfied depended on the answers to two further questions which it was for the national court to determine (paragraphs 79 to 87 of the judgment).

Fourthly, on the assumption that liability was established in the paradigm situation envisaged in Question 1 in the order for reference (i.e. where the loan was granted by a parent company resident in another member state), the Court held that the same result would follow where the lending company and the parent company were resident in different member states, but that Article 43 was not engaged where the lending company (wherever it was resident) did not control the UK borrowing company, and where the ultimate parent company of both the lending and the borrowing companies was resident in a third country (paragraphs 93 to 102 of the judgment).

Fifthly, on the questions relating to remedies, the Court gave guidance very similar to that which it had given a few months earlier in FII. In relation to the issue of “sufficiently serious breach” (in the context of a Factortame damages claim), it expressly directed the national court to take into account the fact that the consequences in the field of direct taxation arising from the freedoms of movement guaranteed by the EC Treaty had only gradually been made clear, and said that until delivery of the judgment in Lankhorst-Hohorst “the problem raised by the current reference for a preliminary ruling had not, as such, been addressed in the Court’s case-law”: see paragraph 121 of the judgment (the section on remedies runs from paragraph 106 to 128).

In the light of this mixed success for both sides, further directions for the conduct of the test claims were given by Rimer J (as he then was) on 6 July 2007, and a case management conference was held before me on 4 February 2008. Directions were given for the trial of the following issues:

In general terms, the intention was that all issues of principle relating to liability and remedies should be determined, but that questions of causation and quantum should be deferred for decision (if necessary) at a later date. In the event, no jurisdictional issues of an Autologic nature have been raised by either side, so the focus of the trial was on issues (b) to (e) as amplified in a helpful summary of issues (“the Summary of Issues”) which was agreed between counsel and handed to me on the first day of the trial, 20 January 2009. The hearing extended over nine days to 30 January, and for reasons which I will explain later a further hearing on the issue of sufficiently serious breach took place (after a number of postponements) on 2 and 3 July 2009.

As in FII Chancery, I do not propose to refer to or analyse the parties’ statements of case in any detail. I will instead begin by describing the legislative background, and will then deal with the issues in the order set out in the Summary of Issues. The issues are for the most part questions of law, but there are two areas (the question whether any of the test claimants were in fact engaged in abusive tax avoidance, and the issue of sufficiently serious breach) where it will be necessary for me to make detailed findings of fact.

This section is largely based on the description of the legal background in the order for reference, and I have also appended to my judgment the same schedule containing the text of the main relevant provisions as was incorporated in the order for reference. I do this partly for convenience, but also because it may be relevant to the first and fundamental issue on liability to know in detail what information the ECJ was given by the parties about the legal background and the relevant legislative and DTC provisions.

The position needs to be considered at three separate times: first, the rules applicable until 1995; secondly, the amendments which were made in 1995 and which remained in force until 1998; and thirdly, the transfer pricing rules which were introduced in 1998 and which remained in force, in a potentially discriminatory form, until 2004.

The rules applicable until 1995

Prior to their amendment by the Finance Act 1995, the main relevant domestic provisions were contained in section 209(2) of the Income and Corporation Taxes Act 1988 (“ICTA”).

Section 209(2)(d) provided that any interest paid by a company on a loan which represented more than a reasonable commercial return on the loan was to be regarded as a distribution of profits to the extent that it exceeded such return. The provision was in the following terms:

“(2)

In the Corporation Tax Acts “distribution”, in relation to any company, means –

…

(d)

any interest or other distribution out of assets of the company in respect of securities of the company, where they are securities under which the consideration given by the company for the use of the principal thereby secured represents more than a reasonable commercial return for the use of that principal, except so much, if any, of any such distribution as represents that principal and so much as represents a reasonable commercial return for the use of that principal;”

“Security” was defined in section 254(1) as including securities not creating or evidencing a charge on assets, and it was also provided that interest paid by a company on money advanced without the issue of a security for the advance, or other consideration given by a company for the use of money so advanced, should be treated in the same way as if a security had been issued. So the scope of the paragraph extended to simple unsecured loans.

The effect of this provision was that the excess amount was not deductible as interest in computing the company’s taxable profits, but was treated as a distribution (or in other words a dividend) paid out of post-tax profits. The fact that the excess interest was treated as a distribution also meant that the company was liable to pay advance corporation tax (“ACT”) under section 14 of ICTA on making the payment.

Section 209(2)(d) applied without distinction to payments made to both resident and non-resident lenders. However, section 209(2)(e)(iv) and (v) laid down a further rule which in effect treated as a distribution any interest (other than interest already treated as a distribution under paragraph (d)) paid to any lender not resident in the UK which was a member of the same group of companies, as defined in the legislation. Accordingly, under the domestic provisions applicable until the legislation was amended in 1995, interest payments made by a UK-resident company to another group member (as defined) outside the UK were always treated as a distribution, even where the interest represented a reasonable commercial return on the loan.

However, the position under the domestic provisions had to be read subject to the arrangements in certain DTCs which prevented the application of these rules and thus ensured that the interest was allowed as a deduction from profits for tax purposes in certain circumstances. By virtue of section 788(3) ICTA such arrangements had effect, notwithstanding any provisions to the contrary in domestic UK legislation, in so far as they provided for relief from corporation tax in respect of income or chargeable gains, or for determining the income or chargeable gains to be attributed to persons not resident in the UK (and their agencies, branches or establishments in the UK), or to persons resident in the UK who had special relationships with persons not so resident. In other words, where and in so far as provisions contained in a DTC were given effect in UK domestic law by section 788(3), they prevailed over any provisions in domestic tax legislation which were inconsistent with them.

The wording of the relevant provisions in the UK’s DTCs varies, but broadly they fall into two categories.

The first category of provisions, which are based on the original draft of the OECD (the Organisation for Economic Co-operation and Development), focus on whether the interest rate is commercial having regard to the amount of the debt. They do not enquire whether the amount of the debt itself is commercial. Such provisions are to be found, for example, in the treaties concluded with Luxembourg, Japan, Germany, Spain and Austria. Article 11(7) of the treaty with Luxembourg may be taken as typical:

“Where, owing to a special relationship between the payer and the recipient or between both of them and some other person, the amount of the interest paid, having regard to the debt-claim for which it is paid, exceeds the amount which would have been agreed upon by the payer and the recipient in the absence of such relationship, the provisions of this article shall apply only to the last-mentioned amount. In that case, the excess part of the payments shall remain taxable according to the law of each Contracting State, due regard being had to the other provisions of this Convention.”

The second category of provisions involve a more general enquiry into whether the amount of the interest exceeds for any reason what would be paid on an arm’s length basis. This includes the question whether the amount of the loan itself exceeds the amount which would have been lent on an arm’s length basis. Such provisions are to be found, for example, in the treaties concluded with the USA, the Republic of Ireland, Switzerland, the Netherlands, France and Italy. Article 11(5) of the treaty with the USA may be taken as typical:

“Where, owing to a special relationship between the payer and the person deriving the interest or between both of them and some other person, the amount of the interest exceeds for whatever reason the amount which would have been paid in the absence of such relationship, the provisions of this article shall apply only to the last-mentioned amount. In that case, the excess part of the payments shall remain taxable according to the law of each contracting state, due regard being had to the other provisions of this convention.”

The critical difference lies in the contrast between the words “for whatever reason” in the second category of treaty provisions, and the words “having regard to the debt-claim for which it is paid” in the first category. The broader scope of the second category of treaty provisions was confirmed by section 808A of ICTA, which was inserted by section 52 of the Finance (No. 2) Act 1992. Section 808A(2) said that the special relationship provision, in treaties of this type, should be construed as requiring account to be taken of all factors, including:

“(a)

the question whether the loan would have been made at all in the absence of the relationship,

(b)

the amount which the loan would have been in the absence of the relationship, and

(c)

the rate of interest and other terms which would have been agreed in the absence of the relationship.”

Subsection (3) then provided that the special relationship provision should be construed as requiring the taxpayer to show that there was no special relationship, or (as the case might be) to show the amount of interest which would have been paid in the absence of the special relationship. These provisions apply to interest paid after 14 May 1992.

Section 808A was not itself overridden by section 788(3), because it is contained in the same Part of ICTA (Part XVIII) as section 788, and section 788(3) is itself made “[s]ubject to the provisions of this Part”. Accordingly section 808A had the effect of modifying the interpretation of the terms of DTCs in the manner which it provided.

The 1995 amendments

The domestic thin cap provisions were amended by the Finance Act 1995 with effect, generally, for interest paid after 28 November 1994. Section 209(2)(d) of ICTA remained unaltered. Section 209(2)(e)(iv) and (v) were, however, repealed and replaced by section 209(2)(da). Under this provision, interest paid between group members (as defined) which exceeded an amount that would have been paid on an arm’s length basis was to be treated as a distribution. However, under section 212(1) and (3), as amended, section 209(2)(da) did not apply if the payer and the recipient of the interest were both within the charge to UK corporation tax.

Section 209(2)(da) provided:

“(2)

In the Corporation Tax Acts “distribution”, in relation to any company, means –

…

(da) any interest or other distribution out of assets of the company (“the issuing company”) in respect of securities issued by that company which are held by another company where –

(i)

the issuing company is a 75 per cent subsidiary of the other company or both are 75 per cent subsidiaries of a third company, and

(ii)

the whole or any part of the distribution represents an amount which would not have fallen to be paid to the other company if the companies had been companies between whom there was (apart from in respect of the securities in question) no relationship, arrangements or other connection (whether formal or informal),

except so much, if any, of any such distribution as does not represent such an amount or as is a distribution by virtue of paragraph (d) above or an amount representing the principal secured by the securities;”

Section 209(2)(da) was amplified by subsections 209(8A) to (8F), which were enacted at the same time. Section 209(8B) specified the criteria to be used in determining whether interest payments were to be treated as distributions. Section 209(8A), in conjunction with subsections (8B) to (8F), determined how far companies could be grouped together for the purposes of assessing the levels of their borrowing on a consolidated basis. In essence, the rules did not allow the consolidation of separate UK sub-groups which were part of a wider foreign group. The borrowing capacity of each UK sub-group had to be considered independently. Thus, for example, by virtue of section 209(8D)(c), where the borrowing company (“the issuing company”) is an effective 51 per cent subsidiary of a UK-resident holding company, the issuing company is to be taken to be a member of a UK grouping of which the only members are the UK holding company and the effective 51 per cent subsidiaries of the UK holding company.

The transfer pricing rules introduced in 1998

Finally, schedule 28AA to ICTA, introduced by the Finance Act 1998, introduced a detailed code of rules on transfer pricing which also applied to interest payments. The transfer pricing rules applied where there was “provision by means of a transaction”, or a series of transactions, between two companies under common control and the terms of the provision were different from what they would have been if the companies had not been under common control. For this purpose, control includes direct or indirect participation in the management, control or capital of any company concerned. The rules applied where the provision gave one of the affected persons an advantage in relation to UK tax. However, until the rules were amended by the Finance Act 2004 this was deemed not to be the case where the other party to the transaction was within the charge to UK tax and certain other conditions were fulfilled. The rules were then amended by the Finance Act 2004 so that they applied where both parties to the transaction were within the charge to UK tax. It was the making of these amendments which finally eliminated the possibility of any legislative discrimination between borrowers depending on whether or not they were resident in the UK.

Introduction

I begin with the critical question whether the UK thin cap rules during the three periods which I have identified were compatible with Article 43 EC. In the light of the judgment of the ECJ, the parties are agreed that the answer to this question turns on whether the UK rules were proportionate to achieve the purpose of preventing abusive tax avoidance (that being Issue 1 in the Summary of Issues). The parties disagree, however, on the proper interpretation of the crucial passage in the judgment of the Court which sets out the criteria by reference to which the issue of proportionality is to be judged. It is common ground that the arm’s length test applied by the UK (via DTCs prior to 1995, and directly by section 209(2)(da) of ICTA thereafter) is an appropriate, and Community law-compliant, tool to use for this purpose. The dispute, in a nutshell, is whether (as the Revenue submit) the arm’s length test is the only one that needs to be applied, or whether (as the claimants submit) the taxpayer must also have the opportunity to show that there was a commercial justification for the impugned loan arrangement, even if it fails the arm’s length test. The Revenue do not deny that considerations of commercial justification may be relevant, but submit that their relevance is confined to, and forms part of, the arm’s length test. According to the Revenue, the arm’s length test has been endorsed by the ECJ as an exhaustive test, or as it was sometimes put in argument as a complete proxy, for determining whether the impugned arrangement constituted abusive tax avoidance.

If the Revenue’s contention is correct, the claimants accept that the UK rules are compatible with Article 43 and the present claims cannot succeed. Conversely, if the claimants are right, it is clear that the UK rules have never provided for a separate test of commerciality, and the consequence must be that at all material times before 2004 they infringed Article 43.

Before I come on to the judgment of the ECJ in the present case, I will first examine the earlier decision of the Court in Lankhorst-Hohorst. It is helpful to do so for at least two reasons. First, it can now be seen as a precursor of the fuller analysis of the subject by the Court in the present case, and some understanding of its facts and of the decision is necessary by way of background. Secondly, and perhaps more importantly, its facts provide a striking illustration of a case where no arm’s length test could have been satisfied (or at any rate no arm’s length test of the type enshrined in the UK legislation after 1995), but where the loan by the parent to its subsidiary nevertheless had a convincing commercial justification, and no question of objectionable tax avoidance, in any ordinary sense of the term, arose. These features of the case give rise to what Mr Graham Aaronson QC in argument aptly termed “the Lankhorst-Hohorst question”: if a similar factual situation had occurred in the UK, would the taxpayer have had any defence to a thin cap ruling by the Revenue which sought to recharacterise the whole (or indeed part) of the interest paid as a distribution?

Lankhorst-Hohorst

The taxpayer company, Lankhorst-Hohorst GmbH, was a wholly-owned German subsidiary of a Dutch parent company, which was itself wholly owned by another Dutch company (“LT BV”). The German company had substantial bank borrowings in excess of DM 3.7 million, and was insolvent on a balance sheet basis. Paragraph 10 of the judgment of the ECJ records that for 1996, 1997 and 1998 the company’s balance sheet showed a deficit not covered by equity capital, and in 1998 the deficit amounted to more than DM 1.5 million. In December 1996 LT BV granted Lankhorst-Hohorst a loan of DM 3 million repayable over 10 years in equal instalments from 1 October 1998. Interest at a variable rate was payable at the end of each year. LT BV duly received interest payments of DM 135,000 in 1997 and DM 109,695 in 1998. The loan, which was intended as a substitute for capital, was accompanied by a letter of support (“Patronatserklärung”) under which LT BV waived repayment if third party creditors made claims against the company. The loan enabled the company to reduce its bank borrowing to less than DM 1 million, and also to reduce its interest payments by approximately 50% (see paragraph 15 of the judgment). It appears that the rate of interest payable by the company on the loan from its parent was about half the rate which it previously had to pay on its bank borrowing.

Paragraph 8a(1) of the German Law on Corporation Tax then in force provided that:

“Repayments in respect of loan capital which a company limited by shares subject to unlimited taxation has obtained from a shareholder not entitled to corporation tax credit which had a substantial holding in its share or nominal capital at any point in the financial year shall be regarded as a covert distribution of profits …

(2)

where repayment calculated as a fraction of the capital is agreed and the loan capital is more than three times the shareholder’s proportional equity capital at any point in the financial year, save where the company limited by shares could have obtained the loan capital from a third party under otherwise similar circumstances or the loan capital constitutes borrowing to finance normal banking transactions … ”

In other words, as I understand it and at the risk of some over-simplification, interest on a loan from a non-resident parent company (which had no entitlement to a corporation tax credit, in common with certain German corporations which were exempt from corporation tax: see paragraph 4 of the judgment) would be regarded as a distribution if the lender’s debt to equity ratio in the borrowing company exceeded 3 to 1 at any time in the financial year, unless one or other of two exceptions was satisfied. The relevant exception for present purposes is the first. In essence it imposes an arm’s length test, and asks whether an unconnected third party would have been prepared to make the loan in otherwise similar circumstances.

The national court found, unsurprisingly, that the exception could not apply. Having regard to its indebtedness and its inability to provide security, Lankhorst-Hohorst could not in fact have obtained a similar loan from a third party, granted without security and covered by a letter of support: see paragraph 12 of the judgment.

The German tax authorities assessed Lankhorst-Hohorst to corporation tax on the basis that paragraph 8a of the corporation tax law applied. The company appealed to the national court, contending that the loan by the parent company constituted a rescue attempt, and that regard should be had to the purpose of the statutory provision, which was to prevent tax evasion. The sole purpose of the loan had been to reduce the company’s expenditure on bank interest. The case was “accordingly not a case of a shareholder with no right to a tax credit seeking to avoid tax chargeable on true distributions of profits by arranging for the payment of interest to itself” (paragraph 15 of the judgment).

The German tax authority accepted that application of paragraph 8a might exacerbate the situation of companies in difficulty, but submitted that the legislature had taken this into account in providing for the relevant exemption, even though the exemption was not, on the facts, applicable. It also submitted that the wording of paragraph 8a did not suggest that the existence of tax avoidance was a condition for its application, a submission which was accepted by the national court (paragraph 16 of the judgment).

The national court asked the ECJ, on a reference for a preliminary ruling, whether Article 43 EC was to be interpreted as precluding the national rule contained in paragraph 8a of the German statute. The case was heard by the Fifth Chamber of the Court, presided over by Judge Wathelet. Three of the other four judges were subsequently to be members of the Court in the present case. Advocate General Mischo delivered his opinion on 26 September 2002, and the Court delivered its judgment on 12 December 2002.

The Court began its analysis by holding that there was, in principle, a breach of Article 43, because the legislation treated a German resident subsidiary of a foreign parent less favourably than if it had a German parent company entitled to receipt of a tax credit. The Court then went on to consider whether the restriction on freedom of establishment could be justified:

“33.

It must still be established whether a national measure such as that in Paragraph 8a(1) … pursues a legitimate aim which is compatible with the Treaty and is justified by pressing reasons of public interest. In that event, it must also be such as to ensure achievement of the aim in question and not go beyond what is necessary for that purpose …

34.

First, the German, Danish and United Kingdom Governments and the Commission submit that the national measure at issue in the main proceedings is intended to combat tax evasion in the form of the use of “thin capitalisation” or “hidden equity capitalisation”. All things being equal, it is more advantageous in terms of taxation to finance a subsidiary company through a loan than through capital contributions. In such a case, the profits of the subsidiary are transferred to the parent company in the form of interest, which is deductible in calculating the subsidiary’s taxable profits, and not in the form of a non-deductible dividend. Where the subsidiary and the parent company have their seats in different countries, the tax debt is therefore likely to be transferred from one country to the other.

35.

The Commission adds that Paragraph 8a(1) … does indeed provide for an exception in the case of a company which proves that it could have obtained the loan capital from a third party on the same conditions, and fixes the permissible amount of loan capital in comparison with equity capital. However, the Commission points to the existence, in the present case, of a risk of double taxation since the German subsidiary is subject to German taxation on interest paid, whereas the non-resident parent company must still declare the interest received as income in the Netherlands. The principle of proportionality requires that the two member states in question reach an agreement in order to avoid double taxation.

36.

It is settled law that reduction in tax revenue does not constitute an overriding reason in the public interest which may justify a measure which is in principle contrary to a fundamental freedom …

37.

As regards more specifically the justification based on the risk of tax evasion, it is important to note that the legislation at issue here does not have the specific purpose of preventing wholly artificial arrangements, designed to circumvent German tax legislation, from attracting a tax benefit, but applies generally to any situation in which the parent company has its seat, for whatever reason, outside the Federal Republic of Germany. Such a situation does not, of itself, entail a risk of tax evasion, since such a company will in any event be subject to the tax legislation of the state in which it is established …

38.

Moreover, according to the findings of the national court itself, no abuse has been proved in the present case, the loan having been made in order to assist Lankhorst-Hohorst by reducing the interest burden resulting from its bank loan. Furthermore it is clear from the case-file that Lankhorst-Hohorst made a loss in the 1996, 1997 and 1998 financial years and its loss largely exceeded the interest paid to LT BV.”

The Court then rejected a further submission, advanced by the German and UK governments, that the legislation was justified by the need to ensure the coherence of the applicable tax systems. It also rejected a submission by the UK government that the legislation could be justified “by the concern to ensure the effectiveness of fiscal supervision”. Accordingly, the answer to the question referred was that the national legislation was incompatible with Article 43.

The judgment of the ECJ in Lankhorst-Hohorst was clearly a significant one, because it established for the first time that national thin cap rules would in principle breach Article 43 if their effect was to treat resident subsidiaries of a non-resident parent in another member state less favourably than if the parent were resident in the same state as the subsidiary. The implications were no doubt particularly worrying for national tax authorities precisely because thin cap rules were normally only needed, and therefore only applied, in cross-border situations. If both the parent and the subsidiary were resident and taxable in the same state, it would not normally matter if an uncommercial loan were made, or if excessive interest were charged, because the interest would be taxable in the hands of the recipient at the usual national rate, and the taxable receipt would match the deduction allowed to the subsidiary. There would therefore be no room for the mischief referred to in paragraph 34 of the judgment to operate.

In these circumstances, the question of justification naturally loomed large; and here, it seems to me, the Court clearly indicated that a relatively strict approach should be adopted. This is hardly surprising, given that one of the fundamental freedoms was involved, and the question of justification arises only when a prima facie infringement of the freedom has been established. Thus, although the Court implicitly accepted that prevention of tax avoidance could, in certain circumstances, justify what would otherwise be an infringement, it pointed out that the relevant German legislation was general in its application, and did not specifically target “wholly artificial arrangements, designed to circumvent German tax legislation”. The Court also pointed out, obviously correctly, that the mere existence of a cross-border arrangement is not, in itself, objectionable, if only because the interest is prima facie taxable in the parent company’s state of residence. There will normally be no point in transferring profits from one state to another in the form of interest unless the rate of tax applicable to the interest is lower in the latter state, or there is some other specific tax advantage to be obtained. Furthermore, in the instant case not only was there no identified risk of tax avoidance, but the loan had in fact been made for good commercial reasons which were accepted by the national court.

What the judgment did not do, no doubt because it was unnecessary to the decision and did not arise on the facts, was to explore the relationship between the arm’s length test in the first exception to paragraph 8a(1) and the concept of tax avoidance. The Court did, however, draw attention to the need to avoid double taxation in cases where application of thin cap rules would lead (in effect) to the same interest being taxed in both member states. This is just one aspect of the principle of proportionality referred to in paragraph 33 of the judgment.

The Court did not state, in terms, that a defence of justification could not be relied upon by the national taxing authority in cases where the taxpayer was able to establish that the loan in question was commercially justified. However, that appears to me to be the clear implication of the judgment, and in particular of paragraph 38. Having held that the legislation itself did not specifically target tax avoidance in paragraph 37, the Court then turned to the particular facts of the case in paragraph 38. It would surely not have done so unless it considered that those facts provided a further reason why application of the legislation could not be proportionate.

The decision of the ECJ in the present case

The decision of the ECJ in the present case contains a much fuller discussion of the subject than Lankhorst-Hohorst. The Court considered Questions 1 and 3 in the order for reference together. Having held that the questions fell to be answered in the light of Article 43 alone, and that the difference in treatment between UK resident subsidiaries depending on the place where their parent company had its seat constituted a restriction on freedom of establishment, the Court then turned to the question of justification. It said by way of introduction, in paragraph 64, that such a restriction is permissible “only if it is justified by overriding reasons of public interest”, and that the application of the restriction must also “be appropriate to ensuring the attainment of the objective in question and not go beyond what is necessary to attain it” (or, in other words, it must be proportionate). The authorities cited for these basic propositions were the recent decisions of the Court in Marks & Spencer (Case C-446/03, [2005] ECR I-10837, [2006] Ch.184, paragraph 35) and Cadbury Schweppes (paragraph 47), both of which post-date Lankhorst-Hohorst. The Court then recorded (paragraph 65) that the UK government, supported by the German government, argued that the UK thin cap provisions were justified on two grounds, cohesion of the national tax system and the prevention of tax avoidance. According to the UK government, the true position was that they were both aspects of the same objective, namely to ensure fair and coherent tax treatment.

The argument based on cohesion of the national tax system was then considered by the Court and rejected: paragraphs 66 to 70. The Court repeated its by then familiar doctrine that for an argument based on cohesion to succeed “a direct link must be established between the tax advantage concerned and the offsetting of that advantage by a particular tax levy” (paragraph 68). No such direct link had been demonstrated by the UK. I note in passing that, although the possibility of a justification based on cohesion had been recognised as long ago as 1992 in the case of Bachmann (Case C-204/09, [1992] ECR I-249, [1994] STC 855), its scope had been so narrowly defined by the Court in subsequent case law that in practice it had never successfully been invoked since Bachmann itself. This was expressly recognised by the Advocate General in the present case, and in an interesting discussion he suggested that the defence, as developed, had become a formalistic one which in practice added nothing to the basic non-discrimination principles applied by the Court: see paragraphs 85 to 91 of his opinion. The Court, however, did not adopt or refer to this passage in its judgment, and it is only in more recent case law that the defence has again begun to develop.

Having rejected the cohesion defence, the Court then turned to the defence based on “the fight against abusive practices”. For present purposes this is the crucial passage in the judgment. It starts at paragraph 71 and runs through to the answer to Questions 1 and 3 given in paragraph 92. I will begin by citing the Court’s conclusion:

“92.

The answer to Questions 1 and 3 must therefore be that Article 43 EC precludes legislation of a Member State which restricts the ability of a resident company to deduct, for tax purposes, interest on loan finance granted by a direct or indirect parent company which is resident in another Member State or by a company which is resident in another Member State and is controlled by such a parent company, without imposing that restriction on a resident company which has been granted loan finance by a company which is also resident, unless, first, that legislation provides for a consideration of objective and verifiable elements which make it possible to identify the existence of a purely artificial arrangement, entered into for tax reasons alone, and allows taxpayers to produce, if appropriate and without being subject to undue administrative constraints, evidence as to the commercial justification for the transaction in question and, secondly, where it is established that such an arrangement exists, such legislation treats that interest as a distribution only in so far as it exceeds what would have been agreed upon at arm’s length.”

The reasoning which led the Court to reach this conclusion proceeded by the following stages.

First, the Court recorded the UK’s submissions that:

The Court then repeated two familiar principles established by earlier case law: first, a national measure restricting freedom of establishment may be justified where it specifically targets wholly artificial arrangements designed to circumvent the legislation of the member state concerned; but, secondly, the mere fact that a resident company is granted a loan by a related company established in another member state cannot be the basis of a general presumption of abusive practices and justify a measure which compromises the exercise of a fundamental freedom guaranteed by the Treaty (paragraphs 72 and 73).

The Court then continued:

“74.

In order for a restriction on the freedom of establishment to be justified on the ground of prevention of abusive practices, the specific objective of such a restriction must be to prevent conduct involving the creation of wholly artificial arrangements which do not reflect economic reality, with a view to escaping the tax normally due on the profits generated by activities carried out on national territory (Cadbury Schweppes and Cadbury Schweppes Overseas, paragraph 55).

75.

Like the practices referred to in paragraph 49 of the judgment in Marks & Spencer, which involved arranging transfers of losses incurred within a group of companies to companies established in the Member States which applied the highest rates of taxation and in which the tax value of those losses was therefore the greatest, the type of conduct described in the preceding paragraph is such as to undermine the right of the Member States to exercise their tax jurisdiction in relation to the activities carried out in their territory and thus to jeopardise a balanced allocation between Member States of the power to impose taxes (Cadbury Schweppes and Cadbury Schweppes Overseas, paragraph 56).

76.

As the United Kingdom Government observes, national legislation such as the legislation at issue in the main proceedings is targeted at the practice of thin capitalisation, under which a group of companies will seek to reduce the taxation of profits made by one of its subsidiaries by electing to fund that subsidiary by way of loan capital, rather than equity capital, thereby allowing that subsidiary to transfer profits to a parent company in the form of interest which is deductible in the calculation of its taxable profits, and not in the form of non-deductible dividends. Where the parent company is resident in a State in which the rate of tax is lower than that which applies in the State in which its subsidiary is resident, the tax liability may thus be transferred to a State which has a lower tax rate.

77.

By providing that that interest is to be treated as a distribution, such legislation is able to prevent practices the sole purpose of which is to avoid the tax that would normally be payable on profits generated by activities undertaken in the national territory. It follows that such legislation is an appropriate means of attaining the objective underlying its adoption.”

The Court was therefore satisfied, at a fairly high level of generality, that thin cap legislation of the type enacted in the UK was an appropriate means of combating the mischief identified in paragraph 76, that is to say the now familiar risk of profits being artificially transferred from a higher tax to a lower tax jurisdiction. The next step was to determine whether or not the UK legislation went beyond what was necessary to attain that objective: paragraph 78.

The judgment continues:

“79.

As the Court held in paragraph 37 of its judgment in Lankhorst-Hohorst, that requirement is not met by national legislation which does not have the specific purpose of preventing wholly artificial arrangements designed to circumvent that legislation, but applies generally to any situation in which the parent company has its seat, for whatever reason, in another Member State.

80.

By contrast, legislation of a Member State may be justified by the need to combat abusive practices where it provides that interest paid by a resident subsidiary to a non-resident parent company is to be treated as a distribution only if, and in so far as, it exceeds what those companies would have agreed upon on an arm’s length basis, that is to say, the commercial terms which those parties would have accepted if they had not formed part of the same group of companies.

81.

The fact that a resident company has been granted a loan by a non-resident company on terms which do not correspond to those which would have been agreed upon at arm’s length constitutes, for the Member State in which the borrowing company is resident, an objective element which can be independently verified in order to determine whether the transaction in question represents, in whole or in part, a purely artificial arrangement, the essential purpose of which is to circumvent the tax legislation of that Member State. In that regard, the question is whether, had there been an arm’s length relationship between the companies concerned, the loan would not have been granted or would have been granted for a different amount or at a different rate of interest.

82.

As the Advocate General stated at point 67 of his Opinion, national legislation which provides for a consideration of objective and verifiable elements in order to determine whether a transaction represents a purely artificial arrangement, entered into for tax reasons alone, is to be considered as not going beyond what is necessary to prevent abusive practices where, in the first place, on each occasion on which the existence of such an arrangement cannot be ruled out, the taxpayer is given an opportunity, without being subject to undue administrative constraints, to provide evidence of any commercial justification that there may have been for that arrangement.

83.

In order for such legislation to remain compatible with the principle of proportionality, it is necessary, in the second place, that, where the consideration of those elements leads to the conclusion that the transaction in question represents a purely artificial arrangement without any underlying commercial justification, the re-characterisation of interest paid as a distribution is limited to the proportion of that interest which exceeds what would have been agreed had the relationship between the parties or between those parties and a third party been one at arm’s length.”

I read this passage as saying, in effect, that an arm’s length test is an appropriate and proportionate method of testing whether the parties to an impugned transaction have entered into an arrangement which is artificial, either in its entirety (because on an arm’s length basis no loan at all would have been granted) or in part (for example because an uncommercially high rate of interest is charged, or because the loan is of a larger amount than the lender would have been prepared to advance on an arm’s length basis). The Court was clearly impressed by the fact that a test of this nature is objective and can be independently verified (paragraph 81). No doubt the Court also had in mind that the arm’s length principle is one which has wide international currency and is endorsed by the OECD, as the UK government had submitted and the Advocate General had pointed out in his opinion. However, the Court went on to state two important qualifications. First, as the Court said in paragraph 82, there must be an opportunity, on each occasion on which the existence of an abusive arrangement cannot be ruled out, for the taxpayer to provide evidence of any commercial justification that there may have been for the arrangement. Furthermore, the opportunity must be one that is afforded in good faith and of which the taxpayer can in practice take advantage: that seems to me the force of the words “without being subject to undue administrative constraints” in paragraph 82. Secondly, as the Court said in paragraph 83, where the conclusion has been reached that the transaction is a purely artificial one without any underlying commercial justification, the recharacterisation of interest paid as a distribution must be limited to the excess over what would have been agreed had the parties been at arm’s length.

It seems to me clear almost beyond argument, from this passage in the judgment alone, that the Court regarded these two qualifications as additional conditions that had to be satisfied if the test of proportionality was to be met. The arm’s length test provides an objective and verifiable filter, by the application of which potentially objectionable transactions may be identified, and the quantum of interest which may be liable to recharacterisation as a distribution may be ascertained. However, it is essential that there should also be an opportunity, both in theory and in practice, for the taxpayer to show that the arrangement had a commercial justification (as, for example, in Lankhorst-Hohorst). It is also essential that any recharacterisation should be confined to interest which exceeds the amount which would have been paid on an arm’s length basis.

Any possible doubt on the point is, in my judgment, dispelled by reference to paragraphs 66 and 67 of the opinion of the Advocate General, which the Court was in my view clearly following (in more compressed terms) in this section of its judgment. In view of the importance of the point, I will quote those paragraphs:

“66.

On this point [i.e. the question of proportionality], it is my view that, depending on its formulation and application, legislation aimed at avoiding thin capitalisation may in principle be a proportionate anti-abuse measure. It is true that the idea that companies have the right to structure their affairs as they wish means that, in principle, they should be allowed to finance their subsidiaries by equity or debt means. However, this possibility reaches its limit when the company’s choice amounts to abuse of law. It seems to me that the arm’s length principle, accepted by international tax law as the appropriate means of avoiding artificial manipulation of cross-border transactions, is in principle a valid starting point for assessing whether a transaction is abusive or not. To use the reasoning of the Court developed in the indirect tax sphere and other non-tax spheres, the arm’s length test represents in this context an objective factor by which it can be assessed whether the essential aim of the transaction concerned is to obtain a tax advantage. Moreover, it is in my view valid, and indeed to be encouraged, for Member States to set out certain reasonable criteria against which they will assess compliance of a transaction with the arm’s length principle, and in case of non-compliance with these criteria for them to presume that the transaction is abusive, subject to proof to the contrary. The setting out of such criteria is, to my eyes, in the interests of legal certainty for taxpayers, as well as workability for tax authorities. This approach is to be contrasted for example, with the use of a single fixed criterion to be applied in all cases – such as a fixed debt-equity ratio – which does not allow other circumstances to be taken into account.

67.

However, the formulation and application in practice of such a test must also satisfy the requirements of proportionality. This means in my view that:

(1)

It must be possible for a taxpayer to show that, although the terms of its transaction were not arm’s length, there were nonetheless genuine commercial reasons for the transaction other than obtaining a tax advantage. In other words, as the Court noted in its Halifax judgment, “the prohibition of abuse is not relevant where the economic activity carried out may have some explanation other than the mere attainment of tax advantages” [Halifax Plc v Customs and Excise Commissioners (Case C-255/02) [2006] Ch.387, paragraphs 74 and 75]. An example that comes to mind is the situation on the facts in Lankhorst-Hohorst, where the purpose of the loan, as accepted by the Court, was a rescue attempt of the subsidiary via minimising the subsidiary’s expenses and achieving savings on bank interest charges. One could imagine, however, that similar situations (i.e., where a transaction was not concluded on arm’s length terms, but was nonetheless made non-abusively and not purely to obtain a tax advantage) would be relatively exceptional;

(2)

If such commercial reasons are put forward by the taxpayer, their validity should be assessed on a case-by-case basis to see if the transactions should be seen as wholly artificial [and] designed purely to gain a tax advantage;

(3)

The information required to be provided by the taxpayer in order to rebut the presumption should not be disproportionate or mean that it is excessively difficult or impossible to do so;

(4)

In cases where the payments are found to be abusive (disguised distributions) in the above sense, only the excess part of the payments over what would have been agreed on arm’s length terms should be re-characterised as a distribution and taxed in the subsidiary’s state of residence accordingly; and

(5)

The result of such examination must be subject to judicial review.”

It is clear from this important passage that the Advocate General considered the underlying principle in this area to be the Community law principle of abuse of law, which the Court had recently discussed in the Halifax case. In assessing whether the way in which a parent company has chosen to finance its subsidiary amounts to abuse of law, application of the arm’s length principle is “a valid starting point”. It provides an objective test, by reference to which “it can be assessed whether the essential aim of the transaction concerned is to obtain a tax advantage” (in a footnote at this point, the Advocate General referred to paragraph 86 of the Halifax judgment). Furthermore, it is legitimate for a member state to have a set of reasonable criteria which it will apply in assessing whether or not a transaction complies with the arm’s length principle, and if those criteria are not met it may properly “presume that the transaction is abusive, subject to proof to the contrary”.

However, the Advocate General went on to say, in paragraph 67, that the practical application of an arm’s length test in this way must also (my emphasis) satisfy the requirements of proportionality. For this purpose, the very first principle formulated by the Advocate General is that the taxpayer must be able to show “that, although the terms of its transaction were not arm’s length, there were nonetheless genuine commercial reasons for the transaction other than obtaining a tax advantage”. In my judgment it is perfectly clear from this, and from the Advocate General’s quotation from paragraph 75 of the Halifax judgment, that he regarded the test of commerciality as a separate test that it must be open to the taxpayer to satisfy, even if the terms of the transaction are not arm’s length. The correctness of this interpretation is then confirmed by his reference to the facts of Lankhorst-Hohorst, because that was a case (as I have explained) where the arm’s length test was not satisfied. The Advocate General thought that such cases would be “relatively exceptional”, but he also said in terms that if such commercial reasons are advanced by the taxpayer, “their validity should be assessed on a case-by-case basis”.

Returning to the judgment of the ECJ, the Court then considered the relevant UK provisions as they stood both before and after the 1995 amendments:

“84.

In the present case, the documents before the Court show that, prior to the amendments made in 1995, the legislation in force in the United Kingdom provided that interest paid by a resident subsidiary in respect of a loan granted by a non-resident parent company was treated, in its entirety, as a distribution, with no assessment of whether the loan satisfied a relevant criterion, such as that of being granted at arm’s length, and without that subsidiary being given any opportunity to provide evidence as to any valid commercial justifications there may have been for the loan.

85.

However, those documents also show that that legislation did not apply in cases involving a DTC which prevented the application of those rules and thus ensured that the interest in question was allowed as a deduction for tax purposes, provided that the rate of interest did not exceed what would have been agreed upon on an arm’s length basis. Under such a DTC, only that proportion of the interest which exceeded what would have been paid on an arm’s length basis was treated as a distribution.

86.

Whilst a tax regime such as the regime which arises, in cases to which they apply, under the DTCs concluded by the United Kingdom appears initially to be based on a consideration of objective and verifiable elements which make it possible to determine whether a purely artificial arrangement, entered into for tax reasons alone, is involved, it is for the national court to determine, should it be established that the claimants in the main proceedings benefited from such a regime, whether that regime gave them an opportunity, if their transactions did not satisfy the conditions laid down under the DTC in order to assess their compatibility with the arm’s length criterion, to provide evidence as to any commercial justification there may have been for the transactions, without being subject to any undue administrative constraints.

87.

The same applies to the national provisions in force after the legislative amendments introduced in 1995 and 1998. It is a matter of agreement that, under those provisions, it is only interest which exceeds what would be paid on an arm’s length basis that falls to be re-characterised as a distribution. Whilst, at first sight, the criteria laid down by those provisions appear to require a consideration of objective and verifiable elements in order to determine whether a purely artificial arrangement, entered into for tax reasons alone, is involved, it is for the national court to determine whether those provisions allow taxpayers, where the transaction does not satisfy the arm’s length criterion, to produce evidence of the commercial justifications for that transaction, under the conditions referred to in the preceding paragraph.”

I comment at this stage that the questions remitted to the national court in paragraphs 86 and 87 again plainly envisage that the question of commercial justification is a separate question that has to be considered in cases where the relevant transaction does not satisfy the arm’s length test. The Court is saying that the UK rules, both before and after 1995, appear to provide for the application of an arm’s length test, but (apart from confirming that to be the case) the national court should also determine whether the rules allowed taxpayers to produce evidence of commercial justification in cases where the arm’s length test was not satisfied. The clear implication is that the rules were disproportionate unless, first, such an opportunity was in fact provided, and, secondly, where satisfactory evidence of commercial justification was adduced, the rules would not be applied.

In the remainder of this section of the judgment, paragraphs 88 to 91, the ECJ held (disagreeing on this point with the Advocate General) that, where thin cap rules are applied by a member state, it is not the responsibility of that member state to ensure that there will be no double taxation in the state of residence of the parent company. In so deciding, the Court followed the principles which it had recently enunciated in its decision in Case C-374/04 Test Claimants in Class IV of the ACT Group Litigation [2006] ECR I-11673, [2007] STC 404, at paragraphs 59, 60 and 70.

Discussion

As will already be apparent, I am unable to accept the Revenue’s submission that the ECJ regarded the question of commercial justification as no more than an aspect of the arm’s length test. In my judgment it is abundantly clear that the ECJ regarded them as separate tests, each of which had to be satisfied if the thin cap rules of the UK or any other member state were to meet the criterion of proportionality. The proposition that the ECJ regarded the arm’s length test alone as sufficient for this purpose, or as a complete “proxy” for determining whether there was abusive tax avoidance, is in my opinion impossible to reconcile with the clear terms of the Court’s judgment, in particular at paragraphs 82, 83, 86, 87 and 92.

Mr Ewart argued for the Revenue that this interpretation of the judgment could not be correct, because almost every loan granted by a parent to its subsidiary will have at least some commercial justification (for example the provision of finance for a specified purpose, such as the acquisition of an asset or the payment of debts). The mischief is likely to lie not in the making of the loan, or the absence of a commercial purpose of the loan, but rather in the terms on which it is granted. Accordingly, it is only some elements of the total package agreed between the parties which are likely to be objectionable. Moreover, the ECJ expressly recognised in paragraph 81 that a transaction may be abusive only in part, for example where an arm’s length loan would have been granted in a different amount or at a lower rate of interest. Therefore, in the context of loans between connected parties, the function of the arm’s length test is to determine whether there is, in whole or in part, a “purely artificial arrangement”. In cases where the focus is on part of the transaction, the taxpayer must be permitted to provide evidence of a commercial justification for the amount of the loan and/or the rate of interest which demonstrates that they would have existed in a commercial transaction between unconnected parties. This, submits Mr Ewart, is all that the ECJ intended when it referred to the provision of evidence of any commercial justification for the arrangement.

Mr Ewart also submitted that a single arm’s length test is perfectly satisfactory, so long as it is not operated mechanistically and is applied to the circumstances of each case. The ECJ must have been aware, from the summary of the UK legal background and the relevant statutory provisions supplied to them in the order for reference, that the only legislative test in the UK was the arm’s length test. Nobody had ever suggested that the UK thin cap rules contained any sort of motive test, and this was reflected in the summary of the UK’s case in paragraph 71 of the judgment.

With regard to Lankhorst-Hohorst, Mr Ewart submitted that the ECJ had dealt with the issues in that case in a very cursory manner, and that it had effectively been superseded by the much fuller treatment in the present case. At one stage in his oral submissions he went so far as to describe the decision in Lankhorst-Hohorst as containing “no reasoning whatsoever”. He suggested that the actual decision in the case might have been correct, although not for the reasons given by the Court. The correct reason, according to Mr Ewart, would have been that the German legislation failed the second limb of the proportionality test, because if a taxpayer failed to comply with the legislation the whole of the interest was always disallowed, not the excess over the arm’s length amount.

The fact that Mr Ewart had resort to such a crude caricature of the decision in Lankhorst-Hohorst is in my judgment symptomatic of the threadbare nature of the Revenue’s arguments on this part of the case. As I have already said, the guidance given by the ECJ in the present case is clear, and I cannot discern any ambiguity in the requirement for a separate test of commercial justification. The details of how that test is to be applied are, of course, a matter for the national court, subject to the requirement that there should be no “undue administrative constraints”. Similarly, it seems to me that the question of how to deal with transactions which have some underlying commercial justification, but which in one respect or another fail to satisfy the arm’s length test, must be left to be worked out by the national court. It is only in (probably rare) cases like Lankhorst-Hohorst itself, where the transaction as a whole was commercially motivated, and no parts of it were uncommercial, that the defence must be allowed to succeed. In such a case there is simply no element of objectionable tax avoidance present at all, even though the arm’s length test is not satisfied.

Nor am I impressed by the argument that the ECJ cannot have meant what it appeared to say, because it must be taken to have known that the UK thin cap rules contained no test apart from the arm’s length test. In the first place, the Court was concerned, in the usual way, to give general guidance to the national courts of all member states, so the scope of their observations was not in any way confined to the UK rules. Secondly, the Court is of course dependent for its knowledge of the national rules on the information supplied to it by the parties. The Court may well have been in some doubt whether the description of the UK rules in the written procedure was exhaustive, and it would not necessarily have known that there was in fact no way in which an independent defence of commercial justification could have been relied upon by a UK taxpayer. Thirdly, I have to say that the Revenue’s presentation of its case to the ECJ seems at times to have come perilously close to giving a misleading impression of what the UK rules actually provided. I will give two examples.

First, the contention that there was an “advance clearance procedure” (to be found on page 15 of the UK’s written observations, and reflected in paragraph 71 of the judgment) was not true in any formal sense. There was never any designated statutory procedure similar to that found, for example, in section 707 of ICTA in connection with the obtaining of tax advantages from transactions in securities. The reference was, instead, to the statutory machinery which enabled a UK subsidiary to apply for permission to pay the interest on a loan from its non-resident parent without deduction of tax. Such applications were in practice the way in which the Revenue usually found out about loans which it might wish to investigate, and they provided an opportunity for the issues to be debated. (For a fuller description of this process, see paragraphs 181 to 184 below). However, the ECJ might well have been under the impression that there was a formal procedure under which a general defence of commercial justification could always be advanced.

Secondly, according to a detailed note of the oral proceedings before the ECJ at the hearing on 31 January 2006 prepared by the claimants’ solicitors, leading counsel then appearing for the UK, Mr David Anderson QC, submitted that the interest paid on the loan in Lankhorst-Hohorst, if paid in the UK, might well have escaped the UK’s thin cap rules, because the application of those rules “depends on the context in each case”, and the question would have been considered “during the advance clearance procedure”. I note in passing that it is a curiosity of the ECJ’s procedure that there appears to be no way of obtaining a transcript of the oral proceedings before the Court. In the absence of a transcript, I do not intend to criticise Mr Anderson QC in any way, but merely to point out that, if his submissions were understood by the Court in the same way as they were understood by the claimants’ solicitors, the Court may have gained the impression that the UK thin cap rules were more flexible than in fact they were.

If I am right in my interpretation of the ECJ’s judgment, there cannot in my view be any doubt about the answers to the questions remitted to the national court in paragraphs 86 and 87. First, the relevant DTCs with other member states before 1995 all included an arm’s length test in one or other of the forms described in the order for reference (see paragraphs 23 to 26 above). There is no disagreement about this between the parties. Secondly, however, neither the DTC regime before 1995, nor the rules in force after 1995, gave any opportunity for the taxpayer to advance and rely upon any commercial justification that there may have been for the transaction, on the assumption that the arm’s length test was not satisfied. Again, there is no dispute about this and the Revenue has never sought to argue the contrary.

One aspect of this, to which Mr Aaronson attached considerable importance both in his oral submissions and in his cross-examination of the Revenue’s witnesses, is that the statutory arm’s length test in section 209(2)(da) of ICTA requires the borrowing capacity of each UK sub-group to be considered independently. It follows that the wider financial circumstances of the group, including for example the strength of a covenant provided by the ultimate holding company of a group such as IBM or Volvo, cannot be taken into account in applying the arm’s length test. In their written submissions, the claimants argued that this and certain other features of the UK legislation, as applied by the Revenue, meant that it did not provide an appropriate objective test based on the arm’s length principle. However, this argument was not pursued by Mr Aaronson in his oral submissions, and I do not think that it is open to the claimants in the light of the ECJ’s judgment. In my view the judgment must be taken to have endorsed the use of an arm’s length test for this purpose, and although the concept of an arm’s length transaction is not without its difficulties, it cannot be said that the points relied on by the claimants deprive the test in section 209(2)(da) of its character as an appropriate objective test. Indeed, Mr Ewart argued that the test arguably went further than was necessary, because the concept of an arm’s length test could be interpreted as requiring the borrower to be regarded as a stand alone company with only its own assets (including its subsidiaries) to be taken into account. However, Mr Aaronson’s point about the UK sub-groups does in my judgment have relevance in relation to the separate test of commercial justification, because I can see no grounds for saying that this question has to be considered from an artificially narrowed group perspective. On the contrary, it seems to me that the question of commerciality is one that has to be considered by reference to the full facts as they actually are in the real world.

In the result, for the reasons which I have given I conclude that the UK thin cap provisions in issue in the present proceedings were not proportionate to achieve the purpose of preventing abusive tax avoidance, because they did not allow for a separate defence of commercial justification. It follows that the provisions breached Article 43, and that the test claimants succeed on the fundamental issue of liability.

Introduction

The next question which I have to consider is the effect on the UK thin cap regime of the breach of the freedom of establishment in Article 43 which I have found to be established. In the Summary of Issues the alternatives are said to be (a) that “the regime is disapplied”, or (b) that “the application of the regime is restricted to cases of actual abusive tax avoidance”. The claimants submit that the first of these alternatives is correct, while the Revenue argue for the latter. As will appear, however, I consider that to reduce the question to these two alternatives is to over-simplify a complex question.

It is first necessary, in my judgment, to identify the precise nature of the breach of Article 43 which has been established. The ECJ has in my view endorsed the arm’s length test, both as embodied in DTCs before 1995 and in its statutory form after 1995, as an appropriate, but not in itself sufficient, means of identifying abusive transactions, the abuse in question being the adoption within the group of an arrangement which is, in whole or in part, purely artificial, and the essential purpose of which is to circumvent the tax legislation of the UK by allowing the subsidiary to transfer profits to its parent company in the form of tax-deductible profits instead of non-deductible dividends: see in particular paragraphs 76-7 and 81-2 of the judgment. The defect in the UK rules, both before and after 1995, was that they did not allow any opportunity for the taxpayer subsidiary to show that there was in fact an underlying commercial justification, whether for the arrangement as a whole or for the part of it which was prima facie invalidated by the arm’s length test. In simple terms, the arm’s length test, which the ECJ regarded as essentially objective and capable of independent verification, needed to be, but was not, supplemented by an essentially subjective motive test, in order to filter out and save from counteraction those transactions which, although they failed the arm’s length test, nevertheless had (either in whole or in the relevant part) a genuine commercial justification. Had the UK rules provided for such a motive test, and had the test been operated in practice without “undue administrative constraints”, the regime would have been compliant with Article 43. The second of the conditions laid down by the Court in its discussion of proportionality, namely that only the excess amount of interest over what would have been agreed at arm’s length should be liable to counteraction, would not have been a problem, because it is common ground that the arm’s length test, as interpreted and applied by the Revenue, has at all material times targeted and sought to re-characterise as a distribution only such excess amount.

Section 2(4) of the European Communities Act 1972 requires domestic legislation to “be construed and have effect subject to” Community rights. Those rights include the freedom of establishment under Article 43, which is of direct effect. Accordingly, the first question is whether it is possible to construe the UK thin cap regime, as it was in force at the material times, in such a way as not unlawfully to restrict the freedom of establishment of a parent company based in another member state which had power (directly or indirectly) to influence the manner in which the relevant UK subsidiary was funded. It is only if such a conforming construction is not possible that the offending legislation must be disapplied.

Both sides were content to adopt my summary in FII Chancery of the basic principles which should guide the national court in this area: see paragraphs 142 to 152 of my judgment in that case. However, the argument on this part of the present case concluded at the end of January 2009, and the Court of Appeal has subsequently dealt with the issue of conforming construction at much greater length, and after a full consideration of all the relevant authorities, in Vodafone 2 v Revenue and Customs Commissioners (No. 2) [2009] EWCA Civ 446, [2009] STC 1480 (“Vodafone 2”), judgment in which was handed down on 22 May 2009. The leading judgment in Vodafone 2 was delivered by Sir Andrew Morritt C, with whose reasoning and conclusions on this issue both Longmore LJ (in a short concurring judgment) and Goldring LJ agreed.

It appears to have been common ground in Vodafone 2 that the principles to be observed in looking for a conforming interpretation are the same under section 2(4) of the European Communities Act 1972 as they are under section 3 of the Human Rights Act 1998, which requires the court to read and give effect to legislation in a way which is compatible with Convention rights “so far as it is possible to do so”. The relevant principles were summarised by counsel appearing for the Revenue in that case (Mr David Anderson QC, Mr Ewart QC and Ms Sarah Ford) in terms from which counsel for the taxpayer (Mr Ian Glick QC and Ms Kelyn Bacon) “did not dissent”: see the judgment of the Chancellor at paragraphs 37 and 38.

Since the principles were not controversial, and since they were accepted and endorsed without qualification by the Court of Appeal (and are, therefore, binding on me), it may be helpful if I set out the Revenue’s summary as it appears in those two paragraphs, but omitting the references to authority:

“In summary, the obligation on the English courts to construe domestic legislation consistently with Community law obligations is both broad and far-reaching. In particular:

(a)

it is not constrained by conventional rules of construction …;

(b)

it does not require ambiguity in the legislative language …;

(c)

it is not an exercise in semantics or linguistics …;

(d)

it permits departure from the strict and literal application of the words which the legislature has elected to use …;

(e)

it permits the implication of words necessary to comply with Community law obligations …; and

(f)

the precise form of the words to be implied does not matter …

The only constraints on the broad and far-reaching nature of the interpretative obligation are that:

(a)

the meaning should “go with the grain of the legislation” and be “compatible with the underlying thrust of the legislation being construed” … An interpretation should not be adopted which is inconsistent with a fundamental or cardinal feature of the legislation since this would cross the boundary between interpretation and amendment … ; and

(b)

the exercise of the interpretative obligation cannot require the courts to make decisions for which they are not equipped or give rise to important practical repercussions which the court is not equipped to evaluate …”

Is a conforming interpretation possible?

The claimants submitted that it was not possible to construe the relevant DTCs and domestic legislation so as to comply with Article 43. Under the domestic provisions in force until 1995, interest payments made by a UK-resident company to another group member outside the UK were always treated as a distribution, even where the interest represented a reasonable return on the loan. However, the DTCs with member states mitigated this strict position by reference to the amount of interest which would have been paid in the absence of a special relationship between the payer and the recipient. After 1995, 209(2)(da) of ICTA provided a statutory arm’s length test. Accordingly, both before and after 1995 the provisions directed enquiry to the arrangements which would have been entered into between unconnected parties. The legislation left no room for an enquiry into what, according to the ECJ, is ultimately the decisive question, namely whether there was a commercial rationale for the arrangements and what that rationale was. In applying these provisions, the Revenue were concerned only with what a third party UK lender might have done if the UK sub-group had been a stand alone group, and the commercial rationale for the arrangements was regarded as relevant only to the extent that it would have been taken into account by such a lender. To construe the legislation as having allowed taxpayers to demonstrate the full commercial rationale for their arrangements, and not to apply the legislation where such a rationale existed, even if the arm’s length test was not satisfied, would be contrary to the wording of the legislation and would be tantamount to a wholesale re-writing of the provisions.

The Revenue did not advance any positive case in response to these arguments in either their written or their oral submissions, so I take it that they do not dissent from the proposition that a conforming construction is not possible. Moreover, the Revenue have made no application to re-open the question and make further submissions on it in the light of the decision of the Court of Appeal in Vodafone 2. In any event, whether the Revenue concede the point or not, I have considered it, and come to the clear conclusion that in the present case a conforming construction is indeed not possible. In my judgment there is no process of construction, even allowing for the width and potency of the principles identified by the Court of Appeal in Vodafone 2, which could treat the arm’s length test in the UK thin cap rules, either before or after 1995, as supplemented by a separate test of commercial motive or purpose. It is of the essence of the UK rules that the arm’s length test alone is decisive. The rules are not explicitly concerned with tax avoidance at all, and if the arm’s length test is not satisfied by the taxpayer that is the end of the enquiry, whatever the underlying rationale for the arrangements may have been. Conversely, the ECJ starts from the position that thin cap rules prima facie involve a restriction of freedom of establishment, and says that such rules can be justified only if (relevantly) they are a proportionate means of combating abusive tax avoidance. For that purpose, it is essential that the rules should provide for a subjective motive test as well as an objective arm’s length test.

Put simply, the UK arm’s length test cannot in my judgment be interpreted as importing, or as itself being subject to, a further subjective test of commercial justification. Such an interpretation neither “goes with the grain” nor is it “compatible with the underlying thrust” of the UK rules. To adopt it would be “inconsistent with a fundamental or cardinal feature of the legislation”, namely the use of the arm’s length test as the sole criterion, and would cross the boundary between interpretation and amendment.

An instructive contrast may be drawn with Vodafone 2 itself, where the question, following the judgment of the ECJ in Cadbury Schweppes, was whether the UK controlled foreign companies (“CFC”) legislation could be interpreted in a way that did not infringe the freedom of establishment of the UK-resident holding company of the Vodafone group, which had in March 2000 incorporated a Luxembourg subsidiary to act as the vehicle for the takeover of Mannesmann AG. The Revenue sought to assess the profits of the Luxembourg company to tax in the UK under the CFC legislation for the accounting period ended 31 March 2001, giving credit for the tax already paid on those profits in Luxembourg. The CFC legislation set out in ICTA section 748(1) a number of circumstances in which the Revenue’s power to apportion the profits of a CFC and charge the sum so apportioned to tax in the UK did not apply. The exceptions were, in summary, where the CFC (a) pursued an acceptable distribution policy, (b) engaged in certain exempt activities, (c) satisfied a public quotation condition, (d) made profits of less than £50,000 or (e) was resident in a territory specified in regulations to be made by HMRC subject to any conditions HMRC might specify.

In addition, section 748(3) contained a further exception which applied where (broadly speaking) the reduction in UK tax achieved by the relevant transactions was minimal, or was not the main purpose, or one of the main purposes, of the transactions. This exception depended in large part on the subjective intention behind the relevant transactions, whereas the exceptions in section 748(1) depended on the objective existence of specific circumstances. Moreover, the exceptions in sections 748(1) were elaborated in considerable detail, and were not mutually exclusive.

The facts of Cadbury Schweppes, and the guidance given by the ECJ in that case, are fully set out in Vodafone 2 at paragraphs 11 to 23. The ECJ held, in barest outline, that the CFC legislation prima facie infringed Article 43, and although capable of justification on the ground of prevention of abusive practices:

“… the specific objective of such a restriction must be to prevent conduct involving the creation of wholly artificial arrangements which do not reflect economic reality, with a view to escaping the tax normally due on the profits generated by activities carried out on national territory”

(paragraph 55 of the judgment of the ECJ). The ECJ considered that the exceptions in sections 748(1) and (3) were, as they stood, insufficient to justify the restriction on freedom of establishment, but left it to the national court to determine whether (as the UK government then submitted) the motive test could be interpreted in a way that would restrict the application of the CFC legislation to wholly artificial arrangements (paragraph 72 of the judgment).

The Special Commissioners were divided on the question. On appeal, Evans-Lombe J held at first instance that attention should be confined to section 748(3), and that it was impossible to construe section 748(3) conformably with Article 43. In the Court of Appeal the Revenue appear to have changed their tack, and argued (a) that the whole of the CFC legislation should be considered in ascertaining whether it was amenable to a conforming interpretation, and (b) that the appropriate way to achieve such an interpretation was to introduce an additional exception into section 748(1) in respect of a CFC:

“if it is, in that accounting period, actually established in another member state of the EEA and carries on genuine economic activities there”

(see paragraph 39).

These submissions were accepted by the Court of Appeal, the core of whose reasoning may be found in paragraph 44 where the Chancellor said:

“To my mind the extension of the exceptions to the CFC legislation for which counsel for HMRC contends is as permissible as either of those which found favour in [Ghaidan v Mendoza [2004] UKHL 30, [2004] 2 AC 557 and Revenue and Customs Commissioners v IDT Card Services Ireland Ltd [2006] EWCA Civ 29, [2006] STC 1252]. It does not alter the impact on other CFCs which are not excepted by any other exception. Certainly it provides an additional exception but, as counsel for HMRC submitted, the grain or thrust of the legislation recognises that the wide net cast by section 747(3) is intended to be narrowed by section 748. Further the terms of various exceptions were not intended to be either mutually exclusive or immutable as the ability to amend the conditions contained in various parts of Schedule 25 and the terms of para (e) of section 748(1) show.”

The Chancellor also rejected a number of other arguments advanced by the taxpayer in a passage of his judgment (paragraphs 40 to 60) which is too long to cite in full, but all of which repays careful study.

For present purposes, I think it is enough to say that there are some crucial differences between Vodafone 2 and the present case. In Vodafone 2 the scheme of the legislation was to cast the initial net widely, but then to narrow down its embrace by a number of overlapping and (in some instances) flexible exceptions. Importantly, in my view, the exceptions expressly included a subjective motive test as well as the objectively verifiable exceptions, and the basic purpose of the motive test was to exclude from the net transactions which did not have a tax avoidance purpose. Thus the grain or thrust of the legislation was, explicitly, to counter tax avoidance, and all that was needed to achieve conformity with Article 43 was the addition of one further objective exception to the list in section 748(1). In the present case, however, there has never been a subjective exception of any kind, and to introduce one would amount to a radical restructuring of the legislation.

Disapplication and its effect

Since a conforming construction is not possible, the legislation must be disapplied. But what precisely does that mean, and what does it entail? As I explained in FII Chancery at paragraphs 143 and 145-6, the concept of disapplication requires the relevant provisions of national law to be treated as being without prejudice to the directly effective Community rights of EU nationals, and as incorporating a notional proviso to that effect. 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.

The claimants’ primary submission is that the UK thin cap provisions should be disapplied in their entirety in the context of all of the test claims. So, for example, the Volvo test claimants seek a declaration that the UK provisions (as they applied to certain transactions from October 1999 onwards)

“… in so far as they restrict the ability of a company resident in the United Kingdom to deduct interest on loan finance granted by a company resident in another member state and which for the purposes of the Thin Cap Provisions is treated as controlling the UK resident company (a “Parent Company”), or by a company which for the purposes of the Thin Cap Provisions is treated as controlled by such a Parent Company, are contrary to Article 43 … of the EC Treaty and therefore unlawful.”

Similar declarations are sought by the other test claimants, including (in the case of Standard Bank) in relation to certain pre-1995 transactions.

The problem with this approach, in my judgment, is that it goes too far. The analysis of the relevant principles of Community law by the ECJ establishes that the only deficiency in the UK provisions, both before and after 1995, was their failure to provide for a defence of commercial justification, and thus to confine the scope of the provisions to cases of abusive tax avoidance. In those circumstances, it seems to me that a more focused approach is called for, and that the Community right of UK subsidiaries in the position of the test claimants which needs to be preserved, and to take priority over the UK provisions as they stood, is the right to have the provisions invoked against them only in cases where the relevant transactions constituted (either wholly or in part) abusive tax avoidance. To disapply the provisions in their entirety, even in cases where the transactions constituted the most blatant tax avoidance and had no commercial justification, would go far further than is necessary to protect the claimants’ Community rights, and could in theory at least give them a wholly undeserved windfall, if they had in fact indulged in activities of that nature.

The right solution, in my view, is to disapply the national rules only in relation to transactions which satisfy the test of commercial justification, either in whole or in any relevant part. Furthermore, since the rules should at all material times have afforded taxpayers in the position of the claimants an opportunity to demonstrate the existence of a commercial purpose, but did not do so, it seems to me that the onus should now be on the Revenue to show (by positive evidence, and not merely as an inference from the fact that the arm’s length test is not satisfied) that the transactions in question lacked any commercial purpose. To require the claimants to demonstrate that there was a commercial justification for the transactions, or the relevant part of them, possibly many years after the event, and when the issue did not arise under the rules as they were then understood and applied, would in my judgment infringe the Community principle of effectiveness.

I find some support for this reversal of the burden of proof in the observations of Advocate General Jacobs, admittedly in a rather different context (the applicability of a “passing on” defence to claims for recovery of wrongly levied local beverage taxes in Austria), in Case C-147/01 Weber’s Wine World v Abgabenberufungskommission Wien [2003] ECR I-11365 at paragraphs 70 and 71 of his opinion:

“70.

However, the question of retro-activity is relevant to the question of the burden of proof since, as Mr Schlosser points out, claimants who know that they will have to establish certain facts are more likely to ensure that they have specific evidence of those facts than are those who do not.

71.

To compile and retain such evidence may be a cumbersome task in situations such as the present, so that a trader might feel justified in not carrying it out if there were no current or foreseeable need to do so in order to be able to obtain reimbursement of any tax which he considered to be clearly incompatible with Community law and the imposition of which he knew caused him a loss. In particular, retail price calculations may not have taken the amount of tax specifically and separately into account if the trader did not expect to have to provide proof of his loss.”

See too the judgment of the ECJ at paragraphs 109 to 117.

Mr Aaronson QC relied on these observations as a reason for disapplying the UK provisions in full, but in my view practical considerations of this nature can be accommodated in a more proportionate way by reversing the burden of proof on the issue of commercial justification.

It may be objected that the solution which I propose looks more like the kind of result which, in a different case, might be achieved by a process of conforming construction, and is less radical than the concept of disapplication would naturally suggest. However, the two concepts (construction and disapplication) are conceptually different, and the concept of disapplication, properly understood, does not necessarily require an all or nothing approach. It is therefore not surprising if, in some cases, the practical difference between their effect is small, even though they start, so to speak, from separate directions. The crucial point, so far as disapplication is concerned, is to identify the nature of the infringement of Community law which the ECJ has found to be established, and then to perform the necessary surgery on the offending national legislation so as to give effect to the claimant’s Community rights. Such surgery, to pursue the metaphor, need not always consist of amputation of a limb or the removal of a diseased organ, but may in appropriate cases be of a reconstructive nature.

I find further encouragement for my approach in two matters. First, in Vodafone 2 Sir Andrew Morritt C briefly discussed the question of disapplication in paragraphs 61 to 66 of his judgment. The discussion was obiter, and he declined to reach any concluded view on the issue (see paragraph 67). Nevertheless, the discussion is of some value, because he records the alternative forms of disapplication suggested by the parties, and points out (in paragraph 65) that:

“the essential difference between the two versions put forward by each party is whether the limitation on the freedom of establishment constituted by the justification found by the ECJ to be permissible is to be read into it so as to restrict the class of person who, in the words of Lord Walker of Gestingthorpe in [Fleming (trading as Bodycraft) v Revenue and Customs Commissioners [2008] UKHL 2, [2008] 1 WLR 195] at [49], “are so circumstanced that the offending provisions must not be invoked against them, either in particular cases or at all.””

The Chancellor went on to say (in paragraph 66) that he “would need a good deal of persuading” that it was appropriate simply to disallow the CFC legislation in favour of Vodafone in relation to the Luxembourg company in the relevant accounting period,

“for that would be to ignore the justification for the CFC legislation the ECJ upheld in Cadbury Schweppes and to disallow the CFC legislation more extensively than the ECJ considered to be necessary to preserve the necessary freedom of establishment.”

Secondly, Mr Aaronson QC himself suggested, in the course of his oral submissions, that even if the UK provisions were disapplied, the Revenue might still be able to rely upon the principle of abuse of law in a case of wholly artificial tax avoidance. That may well be so, but I think the better way to accommodate the point is as part of the process of disapplication itself.

Introduction

In view of the conclusions which I have reached, the only issue of fact which I strictly need to resolve is whether any of the relevant transactions entered into by the test claimants were, either wholly or in part, purely artificial arrangements devoid of any commercial justification. If that is the right question to ask, there can be no doubt about the answer. I am satisfied that none of the relevant transactions was, even remotely, of such a character, and to be fair the Revenue have never sought to argue that they were.

Conversely, if the Revenue are right in their interpretation of the guidance given by the ECJ, and the only test that needs to be applied is the arm’s length test, the evidence clearly establishes that the claimants had every opportunity to adduce evidence of any commercial justification upon which they wished to rely in the context of the arm’s length test. Furthermore, if agreement had not been reached they would have been free to test the issue by bringing an appeal against an assessment or closure notice to the General or Special Commissioners in the usual way. There is no suggestion anywhere in the evidence that there was any systematic inhibition or constraint in the way in which the thin cap provisions were in practice operated by the Revenue, such as to amount to an “undue administrative constraint” of the type contemplated by the ECJ in paragraphs 82, 86 and 92 of the judgment. Nor, for their part, have the claimants ever contended that there was.

In these circumstances, the considerable volume of factual evidence adduced by the test claimants and the Revenue, and the further evidence adduced by the Revenue about the general background to the administration of the thin cap provisions, is of only limited relevance to the issues which I have to determine at this stage of the proceedings. No issue arises about the credibility of any of the witnesses who gave oral evidence – I am satisfied that they were all conscientiously doing their best to assist the court – and cross-examination was limited both in scope and duration.

I have considered what findings of fact I should make, bearing in mind that this case is almost bound to go to appeal, and a higher court may differ from me on the legal issues involved. I cannot readily envisage any circumstances in which it would be either necessary or helpful for me to make exhaustive findings, which would in any event do little more than repeat what is to be found in the witness statements. On the other hand, if I were to say no more about the evidence, I might run the risk of failing to perform the primary task of a trial judge, which is to find all the relevant facts. I therefore propose to adopt a middle course, giving a relatively brief description of the evidence in each of the five test claims and making specific findings where I think that it might be helpful to do so.

Volvo

The first and fourth Volvo test claimants are companies registered in the UK, Volvo Group UK Ltd (formerly known as Volvo Truck and Bus Ltd) (“VTB”) and VFS Financial Services (UK) Ltd (“VFS”). The second test claimant, Volvo Treasury AB, is registered in Sweden and, as its name suggests, is the treasury company of the group. The third test claimant, AB Volvo, also registered in Sweden, is the ultimate holding company. VTB and VFS are indirect wholly owned subsidiaries of AB Volvo, their immediate parent companies being resident in the Netherlands.

The main evidence on behalf of the group was given by Mr Malcolm Martin, who since April 2001 has been the UK director of tax and corporate finance of VTB. Between 1 January 2000 and 31 March 2001 he held the post of business control director at VTB, and before that he was financial controller and company secretary of VFS. VTB deals with the sale and marketing of Volvo truck and bus products, services and ancillary equipment in the UK and Ireland. It is separate from the Volvo cars division, which was acquired by the Ford group in 1999.

Mr Martin’s two witness statements are dated 26 October 2004 and 25 November 2008. His second statement sets out the commercial background to the relevant transactions. There is also a short statement from Mr Lars-Eric Ericson, who is the vice president of Volvo and has been responsible for group international tax issues since he joined the group as an employee in 1996. He endorses Mr Martin’s second statement, and confirms that he was directly involved in the main matters referred to in it.

The principal official who dealt with the case for the Revenue was Mr Anthony Crewe. At the relevant times he was an inspector in Revenue Policy International, previously known as International Division.

Mr Martin and Mr Crewe were cross-examined. The evidence of Mr Ericson was taken as read.

The loan in issue was a loan of $215 million by Volvo Treasury AB to VTB, initially granted on 18 October 1999 for a term of two years at a rate of interest of 6.705%. The commercial background to the loan, as explained by Mr Martin, was briefly as follows. Volvo wished to make an investment in Henlys Plc, a UK listed multinational company with extensive interests in the bus industry in the UK and North America. Henlys’ expansion plans involved the acquisition of a company known as Bluebird which was the manufacturer of the well-known yellow American school buses. Henlys approached Volvo for funding, and an agreement was negotiated at Volvo’s head office in Gothenburg.

The essential features of the deal were that Volvo would lend Henlys $240 million to facilitate acquisition of Bluebird, and Henlys would issue to Volvo unsecured loan notes which were convertible to equity in the form of ordinary shares. The conversion rights, if exercised, would increase the Volvo group’s shareholding in Henlys from 10% to 30% and would place it in a strong position to launch a takeover bid. These were important consideration on AB Volvo’s part in entering the deal.

One of the terms of the agreement reached at head office level was that the loan would be advanced to Henlys through a UK resident company in the Volvo group. This was a requirement of Henlys’, designed to ensure that the interest on the loan would be deductible by Henlys. It was established policy in the Volvo group that finance was raised by Volvo Treasury AB in Sweden, which would then lend the funds on to other companies within the group. By raising the funding at the level of the Swedish group, it was able to obtain the best credit terms and maintain the highest credit rating. Volvo Treasury AB already had lines of credit arranged with external banks, and in group terms $240 million would not be considered a large transaction. VTB was not itself in a position to provide the loan from its own liquid resources. It was also group policy at the relevant time that the terms of the intra-group loan, including the interest rate, should be effectively the same as those obtained by Volvo Treasury AB from its external lenders, with a small uplift to cover the costs and management expenses of Volvo Treasury AB in raising the funds.

The convertible loan notes issued by Henlys for $240 million were for a ten year term at a fixed interest rate of 5.5% per annum. From Volvo’s point of view, the relatively low interest rate was compensated for by a relatively low conversion premium for the shares. According to a letter dated 15 October 1999 from the group’s UK tax advisers, PricewaterhouseCoopers (“PwC”) to Mr Ericson, the conversion terms were considered by AB Volvo to be “very favourable”. In accordance with group policy, Volvo Treasury AB raised the necessary funds by borrowing from its external bankers, over a two year term and at an interest rate of 6.816% per annum. The reason for the two year term was that it was thought that Volvo would wish to consider exercise of its conversion rights before two years had elapsed. On the strength of tax advice from PwC, the decision was also taken that £15 million of the intra-group loan should take the form of an equity subscription in order to prevent the UK group from being thinly capitalised under the relevant UK rules. Accordingly it was agreed that the Swedish group would subscribe for additional share capital in VTB in return for £15 million, and that the amount of the loan would be reduced from $240 to $215 million. The term of the loan was to be two years, at a fixed interest rate of 6.705% per annum. The interest rate on the intra-group loan was therefore slightly lower than the rate which Volvo was paying to its bank lenders. Mr Martin was unable to say why this occurred, but pointed out that small differences of this nature are not unusual, for example where there is a change in interest rates between the date when the intra-group loan is set up and when the external borrowing is actually drawn down. He said that for internal purposes the two rates would be regarded as effectively the same. In any event, whatever the precise explanation may have been, the fact is that VTB paid interest to Volvo Treasury AB on the $215 million at a slightly lower rate than Volvo Treasury AB was paying interest on the same amount to its external lenders.

In December 1999 VTB applied for clearance under the DTC with Sweden to pay interest on the loan to Volvo Treasury AB without deduction of withholding tax. The Revenue then began to investigate VTB’s thin capitalisation position, and a process of discussion and negotiation ensued which culminated in an agreement reached in November 2001 between PwC and Mr Crewe. Mr Crewe was initially sceptical about the commerciality of the loan, but in the course of the discussions he was satisfied about the overall commerciality of the arrangement. In his note of a meeting with PwC on 7 November 2000, Mr Crewe recorded his acceptance that the interest rate on the loan “prima facie … seemed to be a reasonable deal”, and “indicated that, in his opinion, Volvo were not seeking to obtain an unfair tax advantage by means of this loan arrangement and he could see the commercial rationale”. However, it was his duty to apply the UK thin cap rules, and he considered that in certain respects the arm’s length test (which confined attention to the UK sub-group) was not satisfied. In particular he was concerned about VTB’s debt to equity ratio, and the commerciality from VTB’s perspective of the difference between the rate of interest which it received on the loan notes and the rate of interest which it had to pay to Volvo Treasury AB. As he says in paragraph 20 of his witness statement:

“I appreciated the overall commercial background to the Henlys transaction. However, the cost was clearly in the UK whereas there was no guarantee that the benefits and the profits therefrom would be received or taxable in the UK.”

The essential terms of the agreement reached in November 2001 were as follows:

Volvo’s investment in Henlys was not a success, and according to Mr Martin “to have converted the loan notes to equity would have been a disaster”. Accordingly the loan notes remained unconverted, and the loan from Volvo Treasury AB to VTB was extended for a number of short periods on terms mirroring those that Volvo Treasury AB could obtain from its external bankers. Henlys eventually went into liquidation, and VTB was able to use its position as a major creditor to salvage something from the wreckage by acquiring certain assets in specie, including part ownership of Bluebird. Mr Martin accepted in cross-examination that the deal had been a particularly bad one from VTB’s point of view, and that they were left “sadder and wiser after the event”.

It is now necessary to mention another transaction which had an impact on VTB’s thin capitalisation position, namely the transfer of the controlling interest in a company called BRS Ltd from VFS to VTB in November 2000. The British Road Services (“BRS”) business had originally been acquired by VFS in 1998, as part of the development and expansion of Volvo’s truck hire business in the UK. Initially 75% of the shares in BRS Ltd were owned by VFS and 25% by VTB. However, in 2000 there was a change in strategy at head office level in Sweden, as a result of which the business of BRS Ltd was transferred from majority ownership by VFS to 100% ownership by VTB. The transfer was effected on 28 November 2000 at par, for the book value of the investment in VFS. VTB was not required to produce any significant funding for the acquisition. Mr Martin described the transfer as “an organisational strategic decision by top management”, and any tax consequences arising were to be separately addressed. There was no prior consultation with VTB, which was effectively presented with the decision as a fait accompli. VTB took advice on the taxation consequences from PwC, who advised that the transaction might have a “significant and detrimental impact” on the thin cap position, although they thought Mr Crewe had indicated that he would be amenable to a proposal to ring-fence the Henlys loan and the VTB loan so that the acquisition of BRS could effectively be disregarded. As it turned out, however, Mr Crewe did not accept the proposal, and as Mr Martin explains in paragraph 41 of his second statement:

“The acquisition of BRS Ltd with its accumulated losses had the effect of increasing the gearing of [VTB] on a consolidated basis and depressing its consolidated earnings relative to its interest payments. As a result, [VTB] on a consolidated basis fell outside the ratios which HMRC had permitted in the agreements reached in December 2000 and interest disallowances resulted.”

Various proposals to deal with the problem were discussed, and eventually VTB transferred all its interest in BRS Ltd to Volvo Holdings Sweden AB on 6 February 2003 for a nominal consideration of £1. According to Mr Martin, the sole reason for this was to bring VTB within the permitted bands for its debt to equity ratio and interest cover ratio as agreed with the Revenue.

Lafarge

The Lafarge group is a major producer of building materials, based in France. The ultimate group holding company is a French resident company, Lafarge SA, as is the group finance company, Financière Lafarge SA, which is an indirect wholly owned subsidiary of Lafarge SA. The holding company of the UK sub-group is Lafarge Building Materials Ltd (“LBM”). Another UK resident company, Lafarge Minerals Ltd (“LM”), is a wholly owned subsidiary of LBM. These four companies, together with a number of other group companies, are all claimants.

Before 1997 Lafarge had only a relatively small presence in the UK. It had a small gypsum plant in Bristol, and LM owned a small aggregates operation. Both of these businesses traded at a loss. The group’s main operations were in the USA, Canada, Spain, Brazil, France and Germany.

In October 1997 Lafarge launched a hostile takeover bid for the Redland group, a UK parented group in the building materials industry with a major European roofing business and a worldwide network of subsidiaries. The offer price was increased on 28 November 1997, and on 5 December the bid became unconditional. By the end of December Lafarge had received acceptances for over 90% of Redland’s shares, and on 13 February 1998 it compulsorily acquired the outstanding issued capital. The eventual price paid was a little over £1.8 billion. The effect of this takeover was approximately to double the size of the Lafarge group.

In July 2001 the group nearly doubled in size again with a further major UK corporate acquisition, this time of the Blue Circle Industries group. Lafarge had made an earlier, unsuccessful, takeover bid for Blue Circle Industries in about February 2000, as a result of which it acquired just under 20% of the shares in Blue Circle Industries Plc. After the failure of the bid the UK takeover regulations required a year to elapse before a further bid could be made. During this period negotiations continued with the board of Blue Circle Industries Plc, and when a new bid was made in February 2001 it had the support of the board. The acquisition was completed in July 2001 for a price of £2.5 billion.

The funding for both of these major acquisitions included a large amount of intra-group lending, and it is in relation to this lending that the thin cap issues in the present case arise.

The Lafarge group’s main witness was Dorothy Ruth Taylor, who is the UK head of tax for the group. Her two witness statements are dated 26 October 2004 and 25 November 2008. There are also two further short confirmatory statements from Mme. Anne-Marie Girault, the head of tax for the Middle East and Africa of Lafarge SA, and M. Olivier Luneau, the group treasurer from 1991 to 1998 and now a senior vice president of the group.

The inspector in International Division with whom Lafarge conducted its thin cap negotiations was Mr Richard Gallacher, and he too has made two witness statements in 2004 and 2008 (although, as in the case of Mr Crewe, the second statement is almost entirely a verbatim repetition of the first and merely updates some minor details).

Both Ms Taylor and Mr Gallacher were cross-examined, although Mr Aaronson’s cross-examination of Mr Gallacher was very brief. Neither Mme. Girault nor M. Luneau were required to attend for cross-examination and their evidence was taken as read.

The underlying commerciality of the two acquisitions has never been in doubt, and was accepted without hesitation by Mr Gallacher in cross-examination. His only concern was with the application of the UK thin cap rules, and the extent to which the acquisitions were financed by debt rather than equity.

The company in the Lafarge group which actually made the bid for and acquired Redland Plc was LM. Shortly before the bid was made, Lafarge had introduced LBM as the UK resident holding company immediately above LM. Ms Taylor was unable to explain why this had been done, but for thin cap purposes it was irrelevant because the UK sub-group was always regarded from a consolidated perspective. By the time of the takeover bid, Lafarge SA already had in place a multi-currency credit facility with third party banks, which provided short term loans and operated in practice like an overdraft facility. In addition, Lafarge put in place a further FF 10 billion (approximately £1 billion) syndicated loan facility with external banks, guaranteed by Lafarge SA, against which authorised subsidiaries, including LM, could draw. This would enable LM to draw funds to purchase the Redland Plc shares as and when shareholders accepted Lafarge’s offer. In the event, LM drew £964.5 million from this facility for the purchase of Redland Plc shares. The balance of the acquisition cost was largely funded by a loan of £661 million from Financière Lafarge SA to LBM, who in turn lent the money to LM. Financière Lafarge SA drew the £661 million from short term loan facilities it had with other third party banks. The intra-group loans were intended to match the terms and produce a small margin over the interest rates paid on the loans between Financière Lafarge SA and its third party lenders. The remainder of the purchase price, amounting to about £199 million, was met during 1998 in the same way, namely by a combination of drawings on the syndicated loan facility by LM and an increase in the existing loan from LBM to LM, funded by additional amounts raised by Financière Lafarge SA from its external banking facilities.

Despite the massive size of the Redland acquisition, the share capital of LBM was only £9.5 million. On 20 January 1997 LM issued £9 million redeemable preference shares, but this was nothing to do with the Redland acquisition and was intended to repay the intra-group debt for another, much smaller, acquisition which was made at the same time, of further shares in Ennemix Plc. In March 1998 £500 million of LBM’s debt to Financière Lafarge SA was replaced with equity. LBM issued 500 million £1 ordinary shares, 300 million of which were acquired by Financière Lafarge SA and the other 200 million by Spanish group subsidiaries. This recapitalisation left the UK sub-group with borrowings from third party banks under the short term syndicated loan facility of just under £1 billion and internal debt to Financière Lafarge SA of about £333 million.

In her first witness statement Ms Taylor said that the reason for this recapitalisation was concern about the application of the UK thin cap provisions, pursuant to advice given to the group by PwC. In her second statement, however, she said that she had subsequently found a memorandum which suggested that, while thin capitalisation was in mind, it was not the only concern, and Lafarge had always intended to capitalise LBM in an amount of between £600 and £800 million. She said that thin cap was therefore “one of the considerations” for the £500 million injection in March 1998. In cross-examination it was put to Ms Taylor that there was no evidence that PwC, or anybody else, had advised Lafarge about thin capitalisation before March 1998 when the recapitalisation took place, and in subsequent correspondence with the Revenue PwC had described the purpose of the exercise as being to establish what in Lafarge’s view was “an adequate capital base of the UK group”. Ms Taylor said that the financing structure of the group was “very much planned and dictated by the Treasury department”, and she agreed with counsel’s suggestion that the capitalisation took place “because of Lafarge’s commercial reasons to give their subsidiary an adequate capital base”. If it matters, I find that this was indeed the true reason for the £500 million injection, and it was not prompted by concern about the UK thin capitalisation rules.

A further development in the course of 1998 was the replacement of the external syndicated loan facility with longer term loans within the group. The purpose of this change was a purely commercial one, namely to reduce the overall level of group expenditure on interest in a way that would best maintain the credit rating of the group. At the end of this restructuring exercise, LM’s external bank debt had been reduced from almost £1 billion to £187 million, while it owed intra-group debts of £832 million to LBM, of £478 million to Redland Plc and of £431 million to Lafarge SA. In turn LBM had equity of £500 million and debts of £333 million to Financière Lafarge SA.

In mid 1998 Lafarge reorganised its head office functions and transferred its UK tax compliance and tax advice work to PwC. In September 1998, if not earlier, PwC alerted Lafarge to the risk that the interest paid by the UK companies on the intra-group borrowing might not satisfy the arm’s length test under the UK thin cap rules. In December 1998, as permission had not yet been sought to pay the interest gross, withholding tax of £865,186 was deducted from an interest payment by the UK companies to the French lenders and duly accounted for to the Revenue. A request for clearance then followed, which in the usual way prompted the Revenue to make a number of enquiries. Correspondence, meetings and negotiations then ensued, the details of which may be found in Ms Taylor’s second statement, while Mr Gallacher also gives the history of events from his perspective. The negotiations eventually bore fruit in an agreement reached in November 1999, the main terms of which were as follows:

The application of the agreed ratios resulted in the disallowance of interest for the year ending 31 December 1999 of £5,481,600, and for the year ending 31 December 2000 of about £13 million of interest owed by LBM and a further £1.3 million of interest owed by LM. Where the interest was disallowed as a deduction it was not paid, but capitalised and added to the debt owed.

I now turn to the Blue Circle acquisition. As I have already said, the result of the unsuccessful first bid in February 2000 was the acquisition of 19.99% of the shares of Blue Circle Industries Plc. The company which made the bid was again LM, and the acquisition of those shares was funded by intra-group debt. PwC advised Lafarge that since the UK group was already outside the ratios set in the first thin cap agreement, it would not be possible to increase the level of debt of the UK group without the interest being disallowed. Accordingly the shares were sold by LM to a Lafarge group subsidiary in Spain, and the proceeds were used to repay the loan for their acquisition.

In October 2000 LBM’s capital was increased by £275 million subscribed by Financière Lafarge SA for ordinary shares. This sum was used to reduce LBM’s indebtedness to Financière Lafarge SA. These transactions were undertaken on the advice of PwC, in an attempt to bring the UK sub-group within the ratios for debt to equity and interest cover prescribed in the first thin cap agreement.

In view of the interest disallowances which had already been suffered following the first thin cap agreement, Lafarge SA decided to fund the Blue Circle acquisition with a combination of debt and equity. In February 2001, shortly before the second and successful takeover bid was made, Lafarge SA made a further rights issue which raised €1.1 billion (£673 million) of equity for this purpose. This fund was then passed to Financière Lafarge SA, who in turn advanced it as a capital contribution of approximately £700 million in July 2001 to LBM. A loan in the same amount was then provided by LBM to LM. The remainder of the purchase price was raised through external short term borrowings by Lafarge SA itself, and then lent by Lafarge SA to LM. Once the takeover had been successfully completed, Lafarge SA refinanced its debt by issuing to the market medium term loan notes in euros, dollars and sterling, the respective amounts in each currency being intended to match the underlying investments in Blue Circle Industries Plc. The money raised from these loan notes was then used to repay the external short term borrowings, and Lafarge SA put in place with LM loan agreements on terms which matched those of the external loan notes. This ensured that the intra-group borrowing matched the external borrowing both in terms of rate and amount.

The Blue Circle acquisition clearly required the first thin cap agreement to be reconsidered, and a process of negotiation again ensued which led to a second agreement which was concluded in April 2002. The agreement applied for the accounting periods ending 31 December 2001 to 31 December 2005. It provided a detailed methodology for calculating the debt to equity and interest cover ratios, and an agreement that interest on intra-group debt could be paid free of withholding tax. The application of the second agreement produced interest disallowances for 2001, 2002 and 2003 in the following amounts:

IBM

The IBM group needs no introduction. As one of the world’s largest multinational corporations, it operates in all areas of computer technology and consulting in some 170 countries. It has operated in the United Kingdom since 1951, and currently employs about 20,000 people in this country.

The ultimate parent of the group was at all material times, and still is, International Business Machines Corporation (“IBM Corp”), a company resident in the USA. IBM Corp has a wholly owned subsidiary, IBM World Trade Corporation (“WTC”), also resident in the USA, which was the immediate parent of the UK group from 27 October 1995 until the restructuring in 2001 mentioned below. Before 27 October 1995, the UK group’s immediate parent was another US resident corporation, IBM World Trade EMEA Corporation (“EMEA”).

The principal companies in the UK group, before the 2001 restructuring, were IBM UK Holdings Ltd (“IBM Holdings”), the UK group parent, and its wholly owned UK subsidiary, IBM United Kingdom Ltd (“IBM UK”).

In 2001 a decision was taken at US board level to rationalise and restructure the IBM group. As part of this exercise, a new holding company was created to hold IBM’s UK and Netherlands operations. The new holding company was a UK-resident company called IBM North Region Holdings (“NRH”), and it was wholly owned by WTC. It purchased from WTC all the shares in IBM Holdings and the holding company of the Dutch group, IBM Nederland NV, so as to bring into its ownership the UK and Netherlands groups along with their subsidiaries. The purchase of these shares took place on 16 July 2001.

In a further reorganisation, in May 2002 a Dutch company, IBM Global Holdings BV (“IGH”) acquired the shares in NRH from WTC.

The claims made by the IBM group in the present case date back to the early 1990s, and fall broadly into two categories. The first category concerns thin cap negotiations with the Revenue in and around the period 1992 to 1995, and injections of equity into the UK group at that time. The second category of claims concerns the 2001 reorganisation, which gave rise to difficult and prolonged negotiations which eventually culminated in a further thin cap agreement.

The principal witnesses for IBM were intended to be Mr James Lamb, the former chief financial officer of the UK group, who made a witness statement dated 20 November 2008, and Ms Kathleen Bishop, the tax director (Europe, Middle East and Africa) for the IBM group, who made a statement dated 17 December 2008. In the event, for reasons which were not explained to me, Mr Lamb was not called as a witness, but Ms Bishop made a short further statement dated 21 January 2009 in which she effectively adopted Mr Lamb’s evidence as being true and accurate to the best of her knowledge, although she made it clear that she could not speak personally to events which occurred before she joined the group in 1999. The written evidence for IBM was completed by a short statement from Mr Andrew Ross, the then EMEA tax director at IBM UK, dated 27 October 2004, which was largely confined to the question when he first became aware of the possibility that the UK thin cap provisions might be open to challenge.

Evidence on behalf of the Revenue was given by Mr Arthur John Patterson, who is a Fellow of the Institute of Chartered Accountants in England and Wales and is currently employed by the Revenue as a transfer pricing specialist. He had conduct of the negotiations with IBM which followed the 2001 reconstruction, and he sets out their history, and his perception of the changing position, in his statement dated 24 November 2008.

Both Ms Bishop and Mr Patterson were cross-examined. Mr Ross was not required to attend, and his evidence was taken as read.

The background to the thin cap negotiations in the early 1990s is described by Mr Lamb. In the mid to late 1980s, the business of IBM in the UK grew rapidly, and by 1990 turnover had grown to £4.3 billion and over 20,000 people were in permanent employment. Large dividends were paid up the group to IBM Corp, and in 1990 the dividend paid actually exceeded after tax income for the year by some £30 million. Mr Lamb comments that “this distribution policy was to leave the UK group short of working capital when this profitable era came to an abrupt end”. The early 1990s was a period of global recession, and the position was made worse for the UK group by the fact that the UK IT market had become significantly more competitive. In 1991, the UK group made a loss before tax of £124.2 million and in 1992 this rose to a loss before tax of £766.8 million. According to Mr Lamb, by the end of 1991 IBM’s UK business was in crisis. Operating costs had to be significantly reduced, and the number of employees in the UK group fell to under 13,000 by 1995. There was also an alarming outflow of cash from the business. During the three years 1991 to 1993, the net outflow of cash before financing amounted to just over £1 billion.

In order to keep the IBM UK group afloat, a significant amount of additional financing was required. By September 1992, the UK group had total borrowings of £1.55 billion, made up of borrowing of £960 million from WTC and IBM International Finance BV (“IIF”), £57 million from other internal company sources and £138 million of external borrowings. According to Mr Lamb, these loans were essential to provide working capital for the business and to provide financing for the UK group’s customer financing business. The loans were made for purely commercial reasons to sustain a business in crisis. The UK group was further capitalised by three issues of preference shares by UK Holdings to EMEA on 22 December 1992, 31 July 1993 and 2 December 1993, in the aggregate amount of £550 million.

In August or September 1992 the UK group’s then tax advisers, Coopers & Lybrand Deloitte, wrote to the Revenue seeking clearance to pay interest on certain loans from IIF gross of tax. A thin cap enquiry ensued, which resulted in an agreement in November 1992 which set limits on the intra-group and overall borrowing limits of the UK group. If those limits were adhered to, the Revenue would grant clearance to all the loans previously made, and to all further loans taken out until 30 September 1993. It was in the light of this agreement, as I understand it, that IBM increased the capital of the UK group by £550 million between December 1992 and December 1993.

In 1993, the UK group made another loss (of £140.9 million), but the business had started to recover, and by the end of 1994 there had been a significant improvement in the group’s financial position and a pre-tax profit of £96.3 million was made. Growth continued in 1995 and 1996, and in 1995 IBM started to redeem the preference shares which had been issued in 1992 and 1993. They were fully redeemed by the end of the 1998 accounting period.

Further thin cap agreements were made between September 1993 and 1997, as detailed in Mr Lamb’s statement, but for present purposes the details do not matter.

On the evidence before me, I have no doubt that the borrowing by the UK group in the early 1990s was purely for commercial purposes. As Mr Lamb says in paragraph 40 of his statement, “[t]he reality of the situation was that the IBM UK group needed to raise finance to fund its solvency and continue to trade in an extremely difficult period in the early 1990s”.

I now turn to the 2001 reconstruction. According to Ms Bishop, the acquisition by NRH of UK Holdings and IBM Nederlands was carried out at arm’s length. The total consideration for the shares was £3.67 billion. They were transferred by WTC to NRH in exchange for a £2.25 billion loan note and shares in NRH issued at a paid up value of £1.4 billion. Accordingly, no actual cash changed hands. It was Ms Bishop’s job to ensure that the transaction was carried out in a way that was acceptable from a UK tax perspective, and she received advice throughout the relevant period from Mr Derek Jenkins, a tax partner at PwC. The Revenue were approached in advance of the proposed restructuring, in or about May 2001, but it did not prove possible to agree a debt to equity ratio for the funding before the transaction was implemented.

By November 2001 Mr Patterson was apparently satisfied about the underlying commerciality of the reconstruction, but he was concerned about the debt to equity ratio of the financing. One of the difficulties in the negotiations was that Mr Patterson was focusing, as the UK legislation required him to, on the UK group as a separate entity, whereas IBM argued that this did not accord with the reality of commercial borrowing. As Ms Bishop says in paragraph 17 of her first statement:

“If NRH had been seeking to borrow from an unconnected party, it is inconceivable that such a third party lender would not have taken into account the fact that it was the UK holding company for one of the world’s largest multinational corporations.”

I will not set out the history of the negotiations in detail. It can be traced, if necessary, in the statements of Ms Bishop and Mr Patterson together with their exhibits. In June 2002 an agreement in principle was reached, and by early December the Revenue had agreed the level of debt in NRH at £2 billion. This formed the basis of the concluded agreement, together with two further conditions upon which the Revenue had insisted:

When it became clear that the thin cap agreement was going to limit the debt in NRH to £2 billion, steps were taken to reduce the debt in NRH to that level. In the year ended 31 December 2002 NRH issued ordinary shares for approximately £317.8 million, of which £271.5 million was used to reduce the value of loans from group companies.

The thin cap agreement had made it clear that accrued interest of £15.6 million would be treated as a distribution. In fact the amount finally disallowed for 2002 was substantially larger, because NRH failed to comply with interest cover ratio requirements imposed by the agreement. The amount of the interest disallowance was finally agreed some years later at £71.6 million.

In 2003 further equity injections were made into NRH when it became apparent that NRH would not otherwise meet the £1.6 billion net worth requirement. A total of 105 million shares were issued by IBM UK for £382.1 million, and in addition NRH issued shares worth about £2.4 million to IBM Global Holdings BV to fund an acquisition. These shares issues brought the UK group above the £1.6 billion threshold. However, NRH again fell short of the interest cover requirements, and in the event all of the interest paid in 2003, amounting to £81.2 million, was disallowed.

In 2004, UK Holdings issued 250 million £1 preference shares to NRH, and also issued one share to IBM Global Holdings BV at a premium of £331 million. As in 2003, all of the interest payable to WTC on the £2 billion loan for the year was disallowed.

In June 2005, NRH entered into a new thin cap agreement with the Revenue covering the period from 1 January 2005 to 31 December 2009.

In cross-examination Ms Bishop explained that the price to be paid by NRH under the 2001 restructuring had deliberately been fixed as an arm’s length price, on the basis of calculations made by IBM’s corporate development team and advice taken from PwC. The intention was to have “the correct amount of debt and equity” that would be acceptable to the Revenue. The underlying commercial purpose of the reorganisation was to create a regional holding company, and it formed part of a similar series of transactions that were being set up at the same time world wide. The idea was to provide a degree of regional flexibility, and at the same time to incorporate a degree of devolution so that financial transactions would no longer have to be referred back to US board level. Although counsel for the Revenue suggested to Ms Bishop that the establishment of NRH was essentially a tax-driven exercise, intended to generate tax-deductible interest in the UK which would not be matched by taxable receipts in the USA, he made no headway with this line of questioning, and I see no reason to doubt Ms Bishop’s evidence that the reorganisation was essentially motivated by commercial considerations, and formed part of a world wide reorganisation.

I should also mention that Mr Patterson was asked by Mr Aaronson QC, in cross-examination, whether the 2001 reorganisation was a “wholly artificial scheme” of the type with which inspectors in International Division are all too familiar. This was a question confidently expecting the answer No, but instead Mr Patterson replied:

“I think it was. It was a typical scheme where there was a legitimate commercial purpose and legitimate commercial transaction. A fully artificial scheme was grafted on to it. I didn’t realise it at first when I was dealing with it, but by the end of the case I thought that was what happened. You may not welcome that as a view, but I am afraid that’s what I thought.”

Whilst I do not doubt that this is how Mr Patterson eventually came to view the transaction, and I can understand some of the features which may have led him to form that view, I am unable from the evidence before me to endorse it. If the Revenue had wished to advance a positive case of this nature, it would have been necessary to plead it with particularity, and much fuller evidence would have been needed on both sides. To conclude, therefore, I am satisfied that the 2001 reorganisation was commercially justified, and I do not consider that any part of it can be described as constituting abusive tax avoidance.

Siemens

The claims of the Siemens group concern interest deductions disallowed in Siemens Holdings Plc, its UK holding company, for the financial years ending 30 September 2001 to 2003, and an injection of capital in September 2003. The group’s main UK operating company was Siemens Plc, which is a wholly owned subsidiary of Siemens Holdings Plc. The ultimate parent company of the group is Siemens AG, a public listed company resident for tax purposes in Germany. Between 2000 and 2003 there was another German company interposed between Siemens AG and Siemens Holdings Ltd, but for present purposes nothing turns on this.

Most, but not all, of the UK businesses of the Siemens group were operated at the relevant time by subsidiaries of Siemens Plc. By the late 1990s the group had a very diverse portfolio of businesses in the UK in fields such as telecommunications, power generation, the supply of medical equipment, industrial automation, and electronics. One such subsidiary, and by turnover one of the largest single group industries in the UK, was Siemens Business Services Ltd (“SBS”). The business of SBS involved the undertaking of very large scale public and private sector outsourcing contracts for clients. The group also had some finance leasing interests, which were significantly expanded with the acquisition of Schroders Leasing Ltd in 2000.

Although the majority of the group’s business in the UK was conducted through Siemens Plc or its subsidiaries, there were a number of other UK subsidiaries owned directly from Germany. One example was Osram Ltd, the UK subsidiary of Osram GmbH, a German company which had previously been owned as a joint venture by Siemens AG and GEC. Siemens AG bought out GEC’s share some time before the mid 1990s.

The thin cap issues which arise in the present case all stemmed from major losses incurred by SBS, which by 2001 had entered into two public sector contracts, with the Immigration and Naturalisation Directorate (“IND”) and the National Savings Bank respectively, which it appeared SBS would be unable to fulfil at a profit.

Three witnesses gave evidence on behalf of the group: Mr Stephen John Beckett, who is the group head of tax for North West Europe, Russia and Central Asia, and who joined the UK tax department of the group in Spring 2000; Mr Gerard Thomas Gent, who is the general counsel and company secretary for the UK group; and Mr Roland Hummel, who joined Siemens AG in January 2001 as country tax manager for the group’s UK operations, and was based in Munich until January 2004 when he moved to the UK head office. In July 2006 he returned to Munich, and is now the director of taxes, mergers and acquisitions of Siemens AG. All of these witnesses were cross-examined, although none of them at any length. Evidence for the Revenue was given by Ms Christine Roots, who is now employed as a transfer pricing specialist and from 2000 to 2005 was an international specialist working in International Division. Her advice was sought in October 2001 by the Manchester Large Business office, which dealt with the corporation tax affairs of Siemens in the UK. She then had conduct of the negotiations on behalf of the Revenue, her counterpart at Siemens being Sharon Bowen, the then head of tax. Mr Beckett was her deputy, and he succeeded her as head of tax when she left the group in May 2004. Mr Beckett was aware of the discussions which had taken place before that date from regular briefings by Sharon Bowen and also from regular quarterly or half yearly meetings with the Revenue at which he too was present.

Before 2004 the policy of the group was to give local management of the regional business units a large measure of autonomy, but they did not have authority to raise funding from sources external to the group and had to make any funding requests to the finance department (Siemens Financial Services) of Siemens AG in Munich. It was also group policy that where a business unit wished to retain cash on deposit, it would deposit the money with Siemens AG itself. Accordingly, the UK region had no central treasury function of its own, and the business units negotiated and arranged their own financing directly with head office in Munich. This policy was changed in 2004, in response to the proposed introduction of new UK transfer pricing rules with effect from April 2004 which it was thought would impose transfer pricing restrictions on intra-group transactions between UK companies. In fact the consequences of the new legislation turned out to be less onerous in this respect than the group had initially feared, but the decision was taken to consolidate all cross-border debt in Siemens Holdings Plc and then use that company to lend to other UK resident Siemens companies. As a further consequence, a regional treasury department for the UK was established in Spring 2004.

At all material times Siemens AG sought to apply a global policy of financing its regional businesses on a debt to equity ratio of 2 to 1. There were clearly good commercial reasons for this general policy, which was operated and developed independently of the UK thin cap rules. However, it was not an immutable policy, as was shown by the acquisition of Schroders Leasing Ltd in 2000. That business (which now trades under the name Siemens Financial Services Ltd) was a licensed UK bank, and as such operated a debt to equity ratio well in excess of the group norm. Siemens wished to preserve this ratio, and the Revenue were asked to agree that the banking business should be treated separately from the rest of the UK group for thin cap purposes. Following a presentation given in February 2002, the Revenue agreed that much higher levels of gearing were appropriate for a business of this nature. They sought no thin cap adjustment, and further agreed to treat it separately from the rest of the UK group.

By mid 2001 the prospective loss over the duration of the two public sector contracts of SBS had been estimated to be of the order of £115 million. Under the accruals concept of accounting, full provision for that loss had to be made in the year when it became clear that the contract would be loss-making. The provision could then be adjusted as appropriate in future years. In fact the IND contract did eventually make a profit, and the provision relating to it was gradually released. However, provision was made for losses of £115 million in the accounts of SBS for the year ending 30 September 2001, and for a further loss of £73 million for the year ending 30 September 2003. A similar provision was made in the 2001 accounts of Siemens Holdings Plc, on the footing that the directors had agreed to provide financial support to SBS as required.

It was clear that these developments were likely to give rise to a thin cap problem which might lead to the disallowance of interest deductions on loans from Siemens AG. The UK region usually made an annual profit of about £100 million, whereas a loss of around £140 million was now expected. This would have an adverse impact on income cover ratios, and the loss would also reduce the equity value in the UK group. The matter was made worse by the fact that large amounts of cash had been spent on the contracts, which led to increased debt. As a result, said Mr Beckett, “the debt to equity ratio effectively took a double hit”.

Although it was clear that there was likely to be a thin cap problem, it was not immediately apparent how big it was. The first step was to construct a consolidated balance sheet for Siemens Holdings Plc. For various reasons this was not a particularly easy exercise. There were also a number of points which had not been agreed with the Revenue, including:

Once the matter had been referred to International Division, serious negotiations with the Revenue began. Ms Roots made various requests for information, and the first in a series of meetings took place in September 2002. The Revenue made it clear at an early stage that for thin cap purposes they were required to “replicate the obligations of a UK stand alone group”, and therefore had no discretion to include companies which were controlled directly by an overseas parent. On the question of cash deposits, the Revenue initially took the view that an independent lender would not necessarily agree to their being netted off, but by September 2003 they had given way on this point. However, the Revenue remained adamant that no allowance should be made for goodwill in preparing a consolidated account, and they also required the substantial pension deficit in Siemens Plc to be taken into account for 2002 onwards pursuant to the introduction of FRS 17. The reason for this was that consolidated accounts drawn up for the UK sub-group ought to comply with UK accounting standards, although Ms Roots said the Revenue would not look for “over-compliance” if any accounting adjustment were minor and could never have any effect on the final outcome of the thin cap position.

As before, I do not propose to trace the history of the negotiations in detail. They culminated in a meeting on 17 November 2003, when agreement was finally reached (subject to the eventual outcome of the present claim) on the basis that a debt to equity ratio of 1.2 to 1 would be applied, producing interest disallowances of £12.42 million in the three years to 30 September 2003. Meanwhile, an equity injection of £90 million had finally been made into the UK group in September 2003, with the sole purpose of ensuring that the Revenue’s required debt to equity ratio would be met. Siemens Holdings Ltd then used this £90 million, together with a further £40 million of its own available funds, to compensate SBS for a surrender to it of its losses of £140 million for 2001. Mr Beckett confirmed in cross-examination that the capitalisation had been specifically requested for thin cap purposes, and the use to which the money was put by Siemens Holdings Plc was a side effect rather than the motive for the exercise. He said that in his experience he had never known Siemens AG voluntarily to inject capital into a UK company unless it had been asked for.

Standard Bank

The claims of the Standard Bank group differ from those of the other test claimants in that no thin cap enquiries were ever raised by the Revenue and their applications for clearance to pay interest gross were allowed without further investigation. The reason for this was that at all material times the debt to equity ratio of the UK sub-group fell well within Revenue guidelines. Accordingly, the group suffered no disallowances of interest, and in order to succeed in their claims they will need to establish that the equity funding which was provided to the UK group during the relevant years would not have been provided at all, or not to the same extent, had they known that the UK thin cap rules were unlawful. I emphasise that I am not concerned with questions of causation of this nature at the present stage.

The South African parent company of the group is Standard Bank Group Ltd, previously Standard Bank Investment Corporation Ltd. At the relevant times the UK arm of the group consisted of a holding company, Standard Bank London Holdings Plc (“Holdings”) and an operating company which carried on a banking business, Standard Bank Plc (“Standard Bank”). Holdings was an indirect wholly owned subsidiary of the South African parent company, with two Luxembourg companies interposed between them, SBIC Investments SA (Luxembourg) (“SBIC”) and Standard International Holdings (Luxembourg) (“SIH”).

The only witness who gave evidence on behalf of the group was Mr Ian Gordon Gibson, a South African chartered accountant who is currently chief executive of Standard Bank Offshore Group Ltd. He started work with the Standard Bank group in London in August 1992 as financial controller, becoming finance director in August 1996 and then working as chief operating officer of the international group until 2003. The inspector who dealt with the group’s treaty clearance applications was Mr Richard James Clayton, and he gave evidence for the Revenue. Both Mr Gibson and Mr Clayton were briefly cross-examined.

Standard Bank began its operations in London in the early 1990s with the acquisition of an existing financial services company, Ludgate Holdings UK Ltd. The London business was initially small, but it has grown substantially over the years and most of the group’s international operations started from the London base. Offices were established in Hong Kong and New York in or around 1995. All of the subsidiaries in other jurisdictions came under the umbrella of the Luxembourg companies, because Luxembourg was considered to have a favourable business regime. The number of people employed by the group in London grew from around 12 at the beginning to about 600 today. The London business remains the largest part of the group’s operations outside South Africa.

There was naturally a need for capital as Standard Bank expanded its business, and throughout the 1990s SBIC provided capital injections of equity and debt for Holdings and Standard Bank on an almost annual basis. As finance director, Mr Gibson was involved in the decisions to introduce new capital. He would make a pitch to the shareholders in South Africa for this purpose, and prepared three year plans. It was necessary to juggle a number of factors, and to ensure that capital was put into the business in the most efficient manner and in compliance with local tax requirements. Once a decision had been taken at head board level, SBIC was used as the vehicle for putting capital into the different jurisdictions as and when it was needed.

As a bank, Standard Bank was of course subject to banking and financial regulation in the UK and also had to meet various capital requirements. The rules allowed banks to be relatively highly geared. The other major constraint in the UK was the thin capitalisation rules, on which Standard Bank took advice on a continuing basis from KPMG. As a relative newcomer to business in the UK, the group relied heavily on the advice of KPMG, and the debt to equity ratios that the group applied to the UK companies were all established on the basis of that advice. As I have already said, these ratios fell well within established Revenue guidelines, and were not challenged or investigated by the Revenue. Indeed, Mr Clayton indicated in his evidence that he thought at least some of the advice was unnecessarily cautious for a banking group.

There is no suggestion anywhere in the evidence that any of the capital contributions made to the London companies were in any way uncommercial or motivated by considerations of tax avoidance. On the contrary, the group’s aim throughout was to comply with the UK thin cap rules, and it succeeded in doing so, quite possibly by a more comfortable margin than was in fact necessary.

The evidence of Mr Brooks and Mr Black

Two senior officials, Mr Martin Brooks and Mr Graham Black, gave evidence for the Revenue about the general background to the UK’s thin cap legislation and how it is administered. Mr Brooks was employed in International Division between June 1986 and July 1997, and was responsible for transfer pricing issues, including thin capitalisation, in policy, advisory, training and operational capacities. He is now a senior international tax specialist. Mr Black was employed as an inspector in International Division between 1990 and 1995 and again between 1999 and 2001. He is now employed by HMRC as a national business director.

One of the matters helpfully explained by Mr Brooks is the relationship between DTCs and the UK thin cap legislation, and the process for seeking permission to pay interest gross to a foreign lender. In relation to interest generally, many countries, including the UK, assert taxing rights both over payments received by a person resident in the country and over payments made from a source in the country. However, such rights are subject to the provisions of any relevant bilateral DTC. The interest article in many of the UK’s DTCs, including all those with EU member states, provides that interest paid from the UK to a resident of the other country is, to a greater or lesser extent, exempt from income tax in the UK, but also contains an anti-avoidance “special relationship” clause aimed at identifying interest that exceeds the amount that would be expected to have been paid if the parties to the loan had no such special relationship, or in other words if they were at arm’s length. Such “excessive” interest is denied the benefit of the interest article.

The general rule is that yearly interest (other than bank interest) paid from a UK source to a non-resident lender must be paid subject to deduction of tax at source: see section 349 of ICTA, and now section 874 of the Income Tax Act 2007. Section 791 of ICTA gives the Revenue power to make regulations for giving relief due under bilateral DTCs, and such regulations have been made in the form of the Double Taxation Relief (Taxes on Income) (General) Regulations 1970, SI 1970 No.488. However, double tax relief has to be claimed, and is not given automatically: see ICTA section 788(6). Accordingly if an overseas lender advancing a loan to a UK company wished to benefit (as in practice it undoubtedly would) from the provisions of the relevant DTC, it had to apply by a certified treaty claim to the Revenue for confirmation that the interest was taxed either only in the state where it was received, or at an agreed rate in both states. This was the commonest way in which thin cap issues were brought to the attention of the Revenue, and inspectors would then routinely request further information. Before 2000, if the loan was of less than £1 million, the local inspector would decide whether to recommend clearance and would consider whether the borrower was, or would become, thinly capitalised. If the loan was of more than £1 million, the claim and the additional information were forwarded to International Division for them to consider. Because of the relatively low financial limit, the vast majority of treaty claims would be reviewed by an inspector in International Division. After 2000, however, the limit was increased to £50 million, with the result that inspectors working in local and large business offices became responsible for reviewing the capitalisation of a much greater number of UK borrowers.

Apart from formal treaty claims, it was also common for groups and their advisers to contact International Division before making a formal claim and request that the process of obtaining clearance be started immediately. This practice was endorsed in the Revenue’s Tax Bulletin 17, and was also referred to by the Financial Secretary to the Treasury during the Committee Stage debate on the relevant provisions in the Finance Bill 1995.

Tax Bulletin 17 emphasised that there were no blanket exceptions for intra-group loans that satisfied a formulaic test, but in the majority of cases, if the interest rate and other terms of a loan did not give cause for concern, a company or UK sub-group of companies with a debt to equity ratio of less than 1 to 1 and with income cover of more than three times interest payable would not merit further attention. Mr Brooks explained this general practice as follows:

“40.

… This was because these ratios were considered an approximation of the arm’s length borrowing capacity of most UK companies. They were a rule of thumb that had been developed over the years by Inspectors experienced in dealing with thin capitalisation cases and represented an average between the high and low ratios seen in independent companies. It was also the view of HMRC, again developed through experience and reviewing numbers of loan agreements between independent parties, that UK lenders would seek a debt to equity ratio of 1:1 or less from independent borrowers. In the event of a borrower getting into difficulties, a receiver might be expected to realise only 50% or less of the value of assets on a [company’s] balance sheet. In these circumstances, if a company was financed equally by borrowed funds and share capital, then the bank would be able to recover its money.”

However, as Mr Brooks went on to explain, a number of factors other than financial ratios were taken into account in the risk assessment process, and there have never been any hard and fast rules.

Mr Brooks readily accepted in cross-examination that the UK grouping tests in the legislation after 1995 excluded reference to any higher parts of the group, although account could be taken of foreign subsidiaries held directly from the UK on the footing that they were assets of the UK company concerned. He also accepted that a hypothetical lender in the real world would normally look to the whole group, with the consequence that what was in fact a perfectly straightforward commercial arrangement could exceed an arm’s length debt to equity ratio as ascertained by application of the UK rules. Mr Aaronson QC explored this point with Mr Brooks by reference to a detailed example in the advanced thin capitalisation course which was prepared for inspectors working in local and large business offices following the 2000 increase in the threshold of loans that they would have to consider. According to Mr Black, this course represented the best practice that had been built up within International Division over a number of years. The course material explicitly recognised that “we must disregard completely any influence from a larger world-wide group”, and included a conclusion in the following terms:

“Cross-border finance is a wide subject, and we challenge not only abusive schemes, but also “genuine” group transactions which result in the UK interest draw being higher than it would otherwise be in a wholly arm’s length situation.”

Both Mr Brooks and Mr Black also confirmed that it had been the Revenue’s practice to disregard any non-UK grouping before 1995, when interpreting and applying the special relationship provision in the interest articles of DTCs. In this respect, therefore, the 1995 thin cap legislation did little more than codify existing practice.

Apart from putting the UK grouping test on a statutory footing, and introducing the arm’s length test as the basic criterion for the UK thin cap rules, two other changes were made by the 1995 Finance Act. The first change was to state explicitly that, in all cases where there was thin capitalisation, the effect of the rules was confined to the excess interest that was paid, and they did not apply to the whole amount. In other words, it was only the non-arm’s length amount of interest which was disallowed and recharacterised as a distribution. The second change was the introduction of a provision that the new rules would not apply where the lender was itself chargeable to UK corporation tax on the interest which it received, whereas previously the rules simply did not apply in cases where the lender was resident in the UK (because in that situation there was no obligation to pay the interest under deduction of tax, and no DTCs were engaged). Mr Aaronson was concerned to establish that neither of these changes was of any real practical significance, and to a large extent Mr Brooks and Mr Black agreed with him. The effect of the first change was to formalise the position which had previously obtained by virtue of the interest article in the relevant DTCs. As to the second change, both witnesses agreed that the only circumstance in which a non-resident lender could be subject to corporation tax on the interest would be where it made the loan through a UK branch or agency: see ICTA section 11(1) and (2). Neither Mr Brooks nor Mr Black, with their extensive combined experience, had ever come across a case where loans were made by a UK branch or agency of a non-resident lender, and while they were understandably unwilling to say that it never happened in practice, they both agreed that the practical significance of this change was minimal. As Mr Brooks said, multinational groups just did not operate in that way. The sort of loans that the Revenue might wish to challenge would probably be made by companies based in countries like the Netherlands or Luxembourg, and there was no practical reason why any group should wish to channel the loans through a UK branch or agency.

This is a self-contained issue that arises in relation to certain claims made by IBM. It was not considered by the ECJ, because it was not included in the permutations in Question 2 in the order for reference. Put shortly, the question is whether Article 43 is engaged in a situation where the UK subsidiary to which the loan is made has an EU resident parent, but the lending company is neither itself EU resident nor the subsidiary of an EU resident parent.

It will be remembered that, following a second reorganisation of the IBM international group in May 2002, the UK company NRH became the direct wholly owned subsidiary of the Dutch company IBM Global Holdings BV. The principal UK companies below NRH in the UK sub-group were IBM Holdings and IBM UK. In the years ending 31 December 2002, 2003 and 2004 NRH repeatedly fell short of the interest cover requirements set by the Revenue, and substantial disallowances of interest were made. The relevant loans were made by NRH’s indirect US parent company, EMEA, which was itself an indirect wholly owned subsidiary of IBM Corp. Thus the UK company which suffered the deduction, NRH, had an EU-resident Dutch parent, but the lending company was a US company under ultimate US control.

IBM argues that Article 43 applies in these circumstances, because the freedom of establishment of the Dutch parent company of NRH is adversely affected. NRH and its UK subsidiaries must be regarded as established and maintained by the Dutch parent, and the unjustified restriction inherent in the operation of the UK thin cap rules is suffered by the Dutch parent whether or not the lending in question happens to be made by an EU-resident company. The only relevant relationship is the relationship between the parent and the subsidiary under Article 43, and the critical question is whether the UK legislation restricts the ability of the borrowing subsidiary to borrow funds and thus places a restriction on the EU parent’s freedom to establish a subsidiary in the UK.

Mr Aaronson supported this argument by reference to paragraphs 94 and 95 of the judgment of the ECJ, where it began its consideration of Question 2 in the order for reference as follows:

“94.

In that regard, it must be noted, first of all, that, as was stated in paragraph 61 of this judgment, national legislation such as the legislation at issue in the main proceedings which, in treating interest paid by a resident subsidiary to a parent company as a distribution, applies a difference in treatment between resident subsidiaries which is based on the place where their parent company has its seat, constitutes a restriction on freedom of establishment, since it makes it less attractive for companies established in other Member States to exercise freedom of establishment and they may, in consequence, refrain from acquiring, creating or maintaining a subsidiary in the Member State which adopts such a measure.

95.

It follows that legislation of this kind constitutes a restriction on freedom of establishment which is prohibited, in principle, by Article 43 EC, both where a resident borrowing company is granted a loan by a company which is established in another Member State and has a direct or indirect holding in the capital of the borrowing company, conferring on it definite influence on the decisions of that company and allowing it to determine its activities, and where a borrowing company is granted a loan by another non-resident company which, irrespective of where it is resident, is itself controlled by a company which is resident in another Member State and which has, directly or indirectly, such a holding in the capital of the borrowing company.”

Mr Aaronson placed particular emphasis on the words “irrespective of where it is resident” in paragraph 95. He acknowledged, of course, that the ECJ was envisaging a situation where the lending company, wherever it happened to be resident, was itself controlled by an EU parent, but he submitted that this merely reflected the terms in which the question had been asked.

The Revenue submit that Article 43 is not engaged because there is no relevant difference of treatment in such a case. The restriction depends upon the residence of the lender, not the residence of the parent company which is exercising its freedom of establishment. The restriction identified by the ECJ was that the UK thin cap rules subjected subsidiaries of non-resident parent companies to different treatment in comparison to subsidiaries of resident parent companies: see paragraphs 59, 61 and 94 of the judgment. This was a restriction on the non-resident parent’s freedom to establish and maintain its subsidiary in the UK, because it could not lend to that subsidiary without the thin cap rules applying. The ECJ said in answer to Question 2 that the result should be the same where there is a loan from a non-resident company which is controlled by the non-resident EU parent. The reason for the extension is that it is equivalent to the EU parent making the loan, since it controls the non-EU resident lender. It is different, however, when the loan is made by a non-EU resident lender which is not itself controlled by the EU parent. In such circumstances there is no restriction on the EU parent’s freedom of establishment, and Article 43 is not engaged.

In response to the Revenue’s argument, Mr Aaronson made two main points. First, he submitted that the ECJ had decided, essentially as a matter of policy, to concentrate on freedom of establishment to the exclusion of other freedoms which might potentially have been engaged, notably the freedom of movement of capital under Article 56. Since the focus of the investigation has been narrowed in this way, the relevant enquiry is whether the UK subsidiary is unable to borrow funds from a lender outside the UK without suffering the burden of compliance with the UK thin cap rules. Secondly, that question should be answered in favour of the EU parent company, because in the normal way a UK subsidiary of an EU parent is more likely to borrow from a non-UK lender than would be the case if it had a UK parent. Mr Aaronson also submitted that if I felt any doubt on this question a further reference to the ECJ would be appropriate.

In my judgment the submissions of the Revenue are correct, and it is clear from the way in which the ECJ has dealt with Questions 1 and 2 that Article 43 is not engaged where the lender is neither itself EU resident nor controlled by an EU resident. The freedom of establishment which is infringed is not that of any EU parent which happens to have a UK subsidiary, but only that of an EU parent which itself lends to the UK subsidiary (Question 1) or which controls the lending company, wherever the lending company is resident (Question 2, and paragraphs 94 and 95 of the judgment). Conversely, there is no breach of Article 43 where the lending company is under third country control, even if the lending company is itself EU resident (paragraphs 98, 99 and 102 of the judgment). In my view these holdings are enough to show that the ECJ was not concerned with the freedom of establishment of an EU parent company in any wider or more general sense, or with potential discrimination based on nothing more than the alleged propensity of a UK subsidiary of an EU parent to borrow from an overseas lender (wherever resident) rather than a UK lender.

The point is a short one, and does not admit of much elaboration. For the reasons which I have given, I would answer it in favour of the Revenue.

Introduction

In this part of my judgment I deal with the questions of principle that arise in relation to the remedies of a restitutionary nature that are, or may be, available to the claimants if (as I have held) they succeed on liability and the UK thin cap rules have to be disapplied because they breached Article 43. I will deal with the alternative remedy in damages, and the hotly contested issue of “sufficiently serious breach”, in the next main part of this judgment.

Both the Advocate General and the ECJ dealt with the questions relating to remedies in the order for reference (Questions 5 to 10) fairly briefly. The main reason for this, I think, is that the same Advocate General and a similarly constituted Grand Chamber of the Court had already considered and answered very similar questions a few months earlier in FII. I have already expressed the view, in another case, that the Advocate General and the ECJ followed the same general approach in the present case as they had in FII, and said that I could find no indication that they sought either to modify or to develop that approach: see FJ Chalke and another v Revenue and Customs Commissioners [2009] EWHC 952 (Ch), [2009] STC 2027 (“Chalke”), at paragraph 118. I concluded:

“The [present] case therefore confirms, if confirmation were needed, that FII represents the current, and carefully considered, jurisprudence of the ECJ on these topics.”

That is a view to which I still adhere, and (despite the obvious risk of circularity) I find it of some assistance in interpreting the guidance given by the Court in the present case, and in considering how far the judgment of the Court on these questions departs from the views of the Advocate General.

I have already covered much of the ground that is relevant to this part of my judgment in FII Chancery at paragraphs 201 to 276, and I will not repeat it. In addition, I have explained in Chalke why I consider that the judgment of the ECJ in FII represented a significant advance in the jurisprudence of the Court on the scope of the San Giorgio principle: see paragraphs 94 to 107. (As in FII Chancery and Chalke, I refer to the right to a refund of taxes and duties levied in breach of rules of Community law as the San Giorgio principle, and to claims within the scope of the principle as San Giorgio claims: see Amministrazione delle Finanze dello Stato v SpA San Giorgio (Case 199/82) [1983] ECR 3595, [1985] 2 CMLR 658). At the end of my discussion in Chalke, I said that in my view:

“the San Giorgio principle must now be regarded as entitling a claimant who has paid tax levied in breach of Community law not only to repayment of the tax itself, but also to reimbursement of all directly related benefits retained by the member state as a consequence of the unlawful charge. It is only in this way that the claimant can obtain an effective remedy for its loss, and effect can be given to the underlying principle that the member state should not profit from the imposition of the unlawful charge.”

The categories of claim made by the claimants

Question 5 in the order for reference identified ten categories of claims advanced by the claimants, and asked how each category was to be regarded for the purposes of Community law, namely as a San Giorgio claim, as a claim for compensation or damages, or as a claim for payment of an amount representing a benefit unduly denied. I will set out the ten categories in the paraphrased form helpfully adopted by the Revenue in their submissions:

A preliminary point to note is that the claims in categories (a) to (d) and (f) to (g) are all claims in respect of corporation tax or ACT which was actually paid by the borrowing company. There was some debate about the intended scope of category (b), and whether it added anything of substance to category (a). It seems likely that category (b) was meant to cover situations where interest on the loan was paid in full, but the borrowing company decided not to claim a deduction for the whole of that interest because it took the view (whether as a result of negotiations with the Revenue, or on the basis of its own advice) that the amount of interest not claimed would fall to be disallowed under the thin cap provisions. So understood, category (b) adds nothing to category (a), save that the disallowance is voluntary rather than the result of a formal ruling by the Revenue. Mr Ewart suggested that category (b) might also cover cases where the voluntary reduction was based on a misunderstanding of the thin cap provisions and went further than was necessary. It seems unlikely that there should be any claims of such a nature, bearing in mind the imprecision of the arm’s length test and the scope for reasonable disagreement about precisely how much interest should be disallowed in any given case; but if it emerges that this is a real, as opposed to a purely theoretical, possibility in relation to a claimant, the full facts will have to be investigated and it may be that only a portion of the unclaimed interest can be the subject of a restitutionary claim.

Subject to the point which I have just made about category (b), there is no dispute about the classification of seven of the types of claim. It is agreed that categories (a), (b), (f) and (g) are properly characterised as San Giorgio claims for restitution. Equally, it is common ground that the last three categories, (h), (i) and (j), are properly characterised as claims for compensation. The categories of claim in dispute are therefore (c), (d) and (e), which the claimants contend are San Giorgio claims but which the Revenue say sound only in damages.

Before I come on to the guidance given by the ECJ, it is necessary to say a little more about the three categories of claim in dispute, and in particular category (e).

Category (c) covers cases where equity capital has been injected into the borrowing company, or where existing debt has been replaced with equity, in response to the thin cap provisions, and where the level of loan funding would have been higher but for the actual or perceived application of the thin cap provisions. In these circumstances the borrowing company pays less interest than would otherwise have been the case, it therefore has less interest to deduct in its corporation tax computation, its profits are correspondingly larger, and it pays additional tax as a result. It is this additional tax which forms the subject matter of the claim. Questions of causation may well arise in relation to the increased equity funding – for example, was it really a direct consequence of the actual or feared impact of the UK thin cap provisions, or was it part of a wider funding strategy imposed for different commercial reasons by a distant head office? I again emphasise that I am not concerned with such questions at this stage, and the question of characterisation of the claim has to be answered on the assumption that the necessary direct causal link has been established.

Category (d) is conceptually similar to category (c), but concerns cases where the actual or perceived application of the thin cap provisions relates not to the ratio of debt to equity, but rather to the rate of interest charged on the loan. If, as a result, the rate of interest is reduced, or even eliminated, there will again be less interest to deduct than would otherwise be the case and additional tax will be paid. What I have said about causation in relation to category (c) applies, mutatis mutandis, in just the same way to these claims.

Category (e), on the other hand, is not concerned with cases where tax, or additional tax, has actually been paid, but with cases where the tax which would otherwise have been paid has been offset by some form of relief. Mr Aaronson submitted (and the Revenue did not dispute) that three main types of relief are relevant in this context: loss relief, capital allowances and group relief. The basic way in which these three reliefs operate needs to be understood, and is briefly as follows.

Loss relief for trading losses was at all material times provided for by ICTA sections 393 and 393A. The basic rule, found in section 393(1), is that where a company carrying on a trade incurs a loss in the trade, the loss is automatically carried forward and set off against any trading income from the trade in succeeding accounting periods until it has been fully absorbed. It is important to note that this treatment is both mandatory and automatic: it does not depend on the making of a claim. However, the company also has the right to make a claim for the loss to be set off (a) “sideways”, against profits of any description of the accounting period in which the loss is made, or (b) retrospectively, against profits of the same trade in the preceding 12 months (or, before 1997, the preceding three years): see section 393A(1) and (2). Where such a claim is made, the losses automatically carried forward are correspondingly reduced.

For capital allowances, the general rule in section 73(1) of the Capital Allowances Act 1990 was that any allowance made to any person was to be made “in taxing his trade”. By virtue of section 144(2), any such allowance was to be given effect by treating it “as a trading expense of the trade in that period”. Different rules applied to allowances made under section 61 of the 1990 Act (machinery or plant on lease). By virtue of section 73(2), any such allowance had to be made “by way of discharge or repayment of tax”, and had to be available “primarily against income from the letting of machinery or plant”. However, this general rule was subject to detailed provisions in section 145 analogous to the loss relief provisions in ICTA sections 393(1) and 393A. Thus, if there was a deficiency of letting income against which to set the allowance, it would automatically be carried forward, subject to a right to claim that it be set off sideways against other profits, or carried back within specified time limits.

Group relief is different from loss relief and capital allowances in that it always involves two companies (the surrendering company and the claimant company), and it requires both a claim by the claimant company and the consent of the surrendering company: see ICTA section 402(1) and schedule 17A paragraphs 6 to 10. The types of relief which may be surrendered between members of a group of companies include loss relief under section 393A and surplus capital allowances given by way of discharge or repayment of tax. The effect of the claim is that the benefit of the surrendered relief is taken in the claimant company instead of the surrendering company. If a payment is made by the claimant company in return for the surrender, the payment is not taken into account in computing the profits or losses of either company: see section 402(6).

The guidance given by the ECJ

The Advocate General dealt with the classification of the claimants’ claims in paragraphs 101 to 107 of his opinion. After explaining why in his view the issue of remedies should arise only in relatively limited circumstances, because the UK thin cap rules “by and large … comply with Article 43”, and saying that he had dealt with “very similar questions” in his opinion in FII, he continued as follows:

“104.

As I observed in my Opinion in the FII case, the Court has consistently held that the right to a refund of charges levied in a Member State in breach of rules of Community law is the consequence and complement of the rights conferred on individuals by Community provisions as interpreted by the Court. The Member State is therefore required in principle to repay charges levied in breach of Community law. In the absence of Community rules on the recovery of sums unduly paid, it is for the domestic legal system of each Member State to designate the courts and tribunals having jurisdiction and to lay down the detailed procedural rules governing actions for safeguarding rights which individuals derive from Community law, provided, first, that such rules are not less favourable than those governing similar domestic actions (principle of equivalence) and, second, that they do not render practically impossible or excessively difficult the exercise of rights conferred by Community law (principle of effectiveness).

105.

The question raised in the present case is precisely the same as that raised in the FII case; namely whether the plaintiffs’ claims should be characterised as claims in restitution, damages or for an amount representing a benefit unduly denied.

106.

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

107.

Applying this to the present case, it seems to me that the heads of relief put forward by the Test Claimants should be assessed under the principles set out in the Court’s case-law on recovery of charges unlawfully levied; that is, the UK should not profit and companies (or groups of companies) which have been required to pay the unlawful charge must not suffer loss as a result of the imposition of the charge. As such, in order that the remedy provided to the Test Claimants should be effective in obtaining reimbursement or reparation of the financial loss which they had sustained and from which the authorities of the Member State concerned had benefited, this relief should in my view extend to all direct consequences of the unlawful levying of tax. Prima facie, this would to my mind include: (1) repayment of unlawfully levied corporation tax (Question 5(a), (b), (c), (d)); (2) restoration of relief used to offset unlawfully levied corporation tax (Question 5(e)); and (3) repayment of un-utilised advance corporation tax paid on wrongly re-characterised distributions (Question 5(f)). I would emphasise, however, that it is for the national court to satisfy itself that the relief claimed was a direct consequence of the unlawful levy charged.”

It is clear from this passage that the Advocate General took a wide view of the scope of the San Giorgio principle and considered that it should “extend to all direct consequences of the unlawful levying of tax”. He explicitly stated that claims in the three disputed categories would prima facie qualify as San Giorgio claims, while emphasising that it is ultimately for the national court to decide whether the relief claimed was indeed a direct consequence of the unlawful tax.

In paragraph 106 the Advocate General also expressed the view that the principle of effectiveness requires the national court, in characterising claims under national law, to take account of the fact that the conditions for a damages claim (including, in particular, the need for a sufficiently serious breach) may not be satisfied in a given case. The implication appears to be that, where a remedy in damages is for any reason not available, this should encourage the national court to classify the claim as a San Giorgio claim, even if it would not otherwise have done so. I confess that I find this passage in the Advocate General’s opinion puzzling, and with the greatest respect I am unable to agree with it. The principle of effectiveness requires both San Giorgio claims and damages claims to be available, and the basic requirements of each type of claim have been laid down by the ECJ in a long series of cases. If, in a given case, the conditions laid down by the ECJ for a damages claim are not satisfied, I do not see how that can have any impact on the separate question whether the conditions for a San Giorgio claim have been satisfied. This passage in the Advocate General’s opinion was not referred to by the ECJ, and so far as I am aware it has no parallel in other cases. Unless and until the principle is endorsed and explained by the ECJ, I prefer to regard it as no more than a general exhortation to ensure that the principle of effectiveness is respected.

The ECJ considered the question of classification of the claims in paragraphs 109 to 114 of the judgment, which I should cite in full:

“109.

It must be stated in that regard that it is not for the Court to assign a legal classification to the actions brought before the national court by the claimants in the main proceedings. In the circumstances, it is for the latter to specify the nature and basis of their actions (whether they are actions for restitution or actions for compensation for damage), subject to the supervision of the national court (see Metallgesellschaft and Others, paragraph 81, and Test Claimants in the FII Group Litigation, paragraph 201).

110.

However, the fact remains that, according to well-established case-law, the right to a refund of charges levied in a Member State in breach of the rules of Community law is the consequence and complement of the rights conferred on individuals by Community provisions as interpreted by the Court. The Member State is therefore required in principle to repay charges levied in breach of Community law (Test Claimants in the FII Group Litigation, paragraph 202 and the case-law cited there).

111.

In the absence of Community rules on the refund of national charges levied though not due, it is for the domestic legal system of each Member State to designate the courts and tribunals having jurisdiction and to lay down the detailed procedural rules governing actions for safeguarding rights which individuals derive from Community law, provided, first, that such rules are not less favourable than those governing similar domestic actions (principle of equivalence) and, secondly, that they do not render virtually impossible or excessively difficult the exercise of rights conferred by Community law (principle of effectiveness) (Test Claimants in the FII Group Litigation, paragraph 203 and the case-law cited there).

112.

In addition, where a Member State has levied charges in breach of the rules of Community law, individuals are entitled to reimbursement not only of the tax unduly levied but also of the amounts paid to that state or retained by it which relate directly to that tax. As the Court held in paragraph 87 and 88 of the judgment in Metallgesellschaft and others, that also includes losses constituted by the unavailability of sums of money as a result of a tax being levied prematurely (Test Claimants in the FII Group Litigation, paragraph 205 and the case-law cited there).

113.

However, contrary to what the claimants in the main proceedings contend, neither the reliefs or other tax advantages waived by a resident company in order to be able to offset in full a tax levied unlawfully against an amount due in respect of another tax, nor the loss and damage suffered by such a company because the group to which it belongs saw itself as having to substitute financing by way of equity capital for loan capital in order to reduce its overall charge to tax, nor the expenses incurred by the companies in that group in order to comply with the national legislation at issue, can form the basis of an action under Community law for the reimbursement of the tax unlawfully levied or of sums paid to the Member State concerned or withheld by it directly against that tax. Such expenditure is the result of decisions taken by those companies and does not constitute, on their part, an inevitable consequence of the decision by the United Kingdom to treat certain interest paid to non-resident companies as a distribution.

114.

That being the case, it is for the national court to determine whether the expenditure referred to in the preceding paragraph represents, in the case of the companies concerned, financial losses suffered by reason of a breach of Community law for which the Member State in question is responsible.”

Discussion

The critical paragraphs in the judgment of the Court are paragraphs 112 and 113. In paragraph 112 the Court clearly regarded the San Giorgio principle as extending to “reimbursement not only of the tax unduly levied but also of the amounts paid to that state or retained by it which relate directly to that tax”. The emphasis in this passage appears to be exclusively on restitution (of sums paid to the state or retained by it), and unlike the Advocate General the Court does not couple with it the essentially compensatory principle that claimants which have been required to pay the unlawful charge “must not suffer loss as a result of the imposition of the charge” (paragraph 107 of his opinion). This is an important distinction, because it draws a line (at a high level of generality) between remedies of a restitutionary nature and remedies of a compensatory nature, and by implication leaves the latter to be pursued (if at all) in a claim for damages. To an English lawyer the distinction is a welcome one, because it avoids the conceptual blurring of restitution and compensation which seems at times to be present in the Advocate General’s opinion. I think, therefore, on reflection, that there is a significant difference in approach between the ECJ and the Advocate General, and I was wrong in paragraph 231 of my judgment in FII Chancery to say (in relation to their similarly worded views in that case) that the ECJ must have endorsed the Advocate General’s views about the potential width of the San Giorgio principle. Where the ECJ and the Advocate General do agree is in expressing the need for a direct relationship between the amount claimed and the unlawful levying of tax. The difference is that the Advocate General regarded compensatory claims which satisfied that test as falling within the San Giorgio principle, whereas the ECJ (if I am right) did not.

The other important point made by the ECJ is that expenditure cannot be recovered in a San Giorgio claim where it is the result of independent decisions made by the claimant company or another company in the same group, and is not “an inevitable consequence” of the recharacterisation of interest as a distribution (paragraph 113 of the judgment). The focus here seems to be on causation, rather than the characterisation of the claim, and a strict test is to be applied. On this aspect of the matter I do not resile from what I said in FII Chancery at paragraph 235, in relation to the similarly expressed views of the ECJ in that case:

“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 …”

Although the general principle is reasonably clear, there is unfortunately some difficulty in understanding precisely which of the claimants’ contentions the ECJ intended to reject in paragraph 113. The problem is that none of the claimants have “waived” any reliefs or tax advantages in order to “offset in full a tax levied unlawfully against an amount due in respect of another tax”. This language appears to have been rather uncritically transposed from paragraph 207 of the Court’s judgment in FII, where it made perfect sense in relation to one of the categories of claim in that case (broadly speaking, where a company waived entitlement to capital allowances in order to have a greater liability to corporation tax against which to offset unlawfully levied ACT). However, there can in my view be no doubt that the ECJ meant at least to exclude category (c) claims, because the second exclusion in paragraph 113 has no precursor in FII and can only have been intended to refer to this category (“nor the loss and damage suffered by such a company because the group to which it belongs saw itself as having to substitute financing by way of equity capital for loan capital in order to reduce its overall charge to tax”). I agree with the Revenue’s submission that in cases of this type the corporation tax in question is not unlawfully levied, and the decision to finance by way of equity rather than debt is self-evidently an independent decision taken by the companies concerned rather than an inevitable consequence of the thin cap provisions. There is nothing in the thin cap regime which positively requires the injection of equity, although that is one of the ways in which a perceived thin cap problem may be addressed (another way would be to reduce the amount of debt). Indeed, the thin cap rules do not compel companies to do anything: they merely prescribe the consequences, in terms of recharacterisation of interest, if the arm’s length test is not met.

The ECJ did not in terms deal with claims in category (d), where the group chose to reduce the amount of interest charged, but again I agree with the Revenue that the result must be the same. Here too there is nothing unlawful about the additional corporation tax paid by the borrowing company, and the loss is a result of decisions taken by the two companies concerned rather than an inevitable consequence of the unlawful regime.

That leaves the claims in category (e), which following the exclusion of categories (c) and (d) must now be confined to the use of losses or other tax reliefs to offset the additional liability to corporation tax arising from a recharacterisation of interest (whether enforced or voluntary). It may be that the first exclusion in paragraph 113 of the judgment was intended to apply to this category, but it is not worded in terms apt to do so. The answer therefore has to be found by application of the general test set out at the end of paragraph 113. On that basis, it seems to me that where the additional liability to tax has been offset by group relief, a San Giorgio claim will not lie. Group relief requires positive decisions to be taken by the two companies concerned, and its application cannot be regarded as an inevitable consequence of the recharacterisation. On the other hand, where the additional liability is offset by losses carried forward, or by capital allowances, I think that a San Giorgio claim can in principle lie for restitution of the losses or capital allowances so used. They will have been set off against corporation tax which, if it had been paid, the company would have been entitled as of right to recover; and the set off will either have been automatic, in the case of losses carried forward, or will at most have involved the making of routine decisions by the taxpayer company itself, without the intervention of any third party.

For these reasons I conclude that claims in categories (c) and (d), and category (e) claims based on group relief, cannot qualify as San Giorgio claims, but that category (e) claims based on the use of trading losses or capital allowances should be so classified.

Does English law provide an adequate remedy?

I can deal with this question relatively briefly. There is no dispute that the English law of restitution in its present form provides an adequate domestic remedy for claims which are properly classified as San Giorgio claims, and therefore satisfies the principle of effectiveness. The relevant, or potentially relevant, types of English restitutionary claim are the action for restitution of unlawfully levied tax based upon a mistake of law, as recognised by the House of Lords in Kleinwort Benson (Kleinwort Benson Ltd v Lincoln City Council [1999] 2 AC 349), and the action for restitution of tax unlawfully demanded under the Woolwich principle (Woolwich Equitable Building Society v IRC [1993] AC 70): see FII Chancery at paragraph 245.

The parties are divided, as they were in FII Chancery, on the question whether the Woolwich cause of action alone is enough to satisfy the principle of effectiveness. The question was of critical importance in FII Chancery, because the great majority of the claims were brought outside the normal limitation period of six years and reliance had to be placed on a mistake-based cause of action (at any rate as the law now stands) if advantage was to be taken of section 32(1)(c) of the Limitation Act 1980. The Kleinwort Benson cause of action, unlike a Woolwich claim, is based on mistake. However, if it was not needed to satisfy the claimants’ directly effective Community law rights the Revenue could, and did, argue that the Woolwich cause of action alone, with its six year limitation period, was all that the principle of effectiveness required English law to make available to the claimants in respect of their San Giorgio claims.

This argument also carried through into other issues: the availability of a change of position defence, the effect on the claims of the curtailed limitation periods introduced by section 320 of the Finance Act 2004 and section 107 of the Finance Act 2007, and the scope of the ouster of non-statutory mistake claims in section 33 of the Taxes Management Act 1970. The Revenue’s general point, deployed in all these contexts, was that once it is recognised that mistake claims are not required in order to give effective redress for a San Giorgio claim, there is no reason to allow claims to be made outside normal time limits, a defence of change of position should be available, sections 320 and 107 can take effect in accordance with their terms, and the section 33 ouster can operate.

I rejected this argument in paragraph 260 of my judgment in FII Chancery, relying in particular on the relatively strict view that the Court of Appeal had taken in NEC Semi-Conductors Ltd v Revenue and Customs Commissioners [2006] EWCA Civ 25, [2006] STC 606 (“NEC Semi-Conductors”) about the ambit of the Woolwich cause of action and the need to identify a demand for the unlawful tax. In the present case the Revenue returned to the attack. Mr Ewart argued:

These submissions were attractively advanced, but I do not find them convincing. In the first place, it is in my view wrong in principle to treat the Community law principle of effectiveness as though it were a limiting one, and to look for the minimum remedies in UK law that are needed to satisfy it. The true principle is, rather, that it stipulates a minimum content that UK law must provide if San Giorgio claims are to be satisfied. If, however, English law provides two alternative causes of action to which a claimant may have recourse, it is no part of the principle of effectiveness to say that only the more restrictive of those causes of action is needed. Thus I do not see how it would avail the Revenue to establish that the Woolwich cause of action would, by itself, or with appropriate extensions, be sufficient to satisfy the claimants’ San Giorgio claims, when they also have the alternative open to them of making a mistake-based restitution claim. On the contrary, my understanding is that where domestic law goes further than is strictly necessary to give effect to Community law rights, Community law requires the full range of domestic remedies to be made available, and not to be curtailed without due notice.

Secondly, the principle that a claimant is normally entitled to choose between alternative remedies is one that is well recognised in English law, and most notably in Deutsche Morgan Grenfell Group Plc v IRC [2006] UKHL 49, [2007] 1 AC 558 (“DMG”): see in particular paragraphs 50 to 51 (Lord Hope) and 136-137 (Lord Walker). It is worth noting at this point what Lord Goff actually said in Woolwich in relation to the San Giorgio principle at 177E:

“I only comment that, at a time when Community law is becoming increasingly important, it would be strange if the right of the citizen to recover overpaid charges were to be more restricted under domestic law than it is under European law.”

This comment was made some years before Kleinwort Benson was decided, and thus at a time when English law would provide no remedy for the recovery of unlawfully exacted tax unless an appropriate remedy was recognised. The comment was simply not directed to the issue whether the Woolwich cause of action alone would suffice to satisfy San Giorgio claims. If he had been faced with that question, it is not difficult to guess what his answer would have been, given his approving reference in Kleinwort Benson to “the usual preference of English law to allow either of two alternative remedies” ([1999] 2 AC 349 at 387B.)

Thirdly, and in any event, I adhere to the view that in NEC Semi-Conductors the Court of Appeal clearly held that an unlawful demand is a necessary ingredient of a Woolwich claim, and it was not satisfied because ACT was not collected through the machinery of demand: see the judgment of Mummery LJ at paragraphs 140 to 147 and 162, Lloyd LJ at paragraph 84 and Sedley LJ at paragraph 89. The Court of Appeal declined to spell a demand for ACT out of the statutory regime itself, and in my view it would be, if anything, even more difficult to spell a demand out of the thin cap provisions. As I have already pointed out, they do not impose any charge to tax, but merely provide for the recharacterisation of interest in certain circumstances. Furthermore, no question of any demand having been made could arise in cases where the additional tax was not paid but offset by some relief. Accordingly, this relatively narrow ground alone is in my judgment enough to show that the Woolwich cause of action is not, by itself, sufficient to satisfy all of the claimants’ San Giorgio claims.

I wish finally to say a little about the nature of the mistake made by the claimants. As in FII Chancery, I would identify the mistake at a fairly general level as a mistake about the unlawfulness of the thin cap provisions: compare FII Chancery at paragraphs 261 to 262. The evidence given on behalf of all the test claimants established that they were operating under a mistake of law of this nature at the material times. In short, they mistakenly believed that the thin cap provisions as applied by the Revenue were lawful, and that they were obliged to make the relevant payments of tax.

Are any claims which are not San Giorgio claims restitutionary claims under English law?

It is common ground that, if the English law of restitution did not provide an adequate and effective remedy for San Giorgio claims, it would be necessary to develop it so that it did provide such a remedy. It is also common ground, as I have already explained, that the English law of restitution satisfies this test, although there is disagreement about which restitutionary causes of action are required for the purpose. I move on now to the converse question: do the claimants have any other claims, which are not San Giorgio claims, and as a matter of Community law would give rise only to a Factortame claim for damages, but which nevertheless are good restitutionary claims as a matter of English domestic law? If there are any other claims of such a nature, the principle of effectiveness requires that the claimants should be entitled to pursue them within the framework of the domestic law of restitution, whether or not the conditions for a Factortame damages claim would also be satisfied.

A good example of a type of claim which would potentially fall into this category, if my conclusions on the scope of the San Giorgio claims are correct, is a claim based on the use of group relief to set off against an unlawful liability to corporation tax. Can the company which surrendered, or the taxpayer company which utilised, the relief make a mistake-based restitutionary claim to recover the benefit of the relief?

This question gives rise to interesting and difficult questions at the frontiers of the English law of restitution. In FII Chancery I did my best to grapple with the similar questions which arose in that case, and I concluded, in short, that claims of this type, which involve the making of separate decisions by third parties, and which have the result that the otherwise unlawful tax is in fact not paid to the Revenue, sound only in damages: see in particular paragraph 270 of my judgment. Mr Cavender, who argued this part of the case for the claimants as he had done in FII Chancery, subjected my reasoning on this question to some searching criticism, but did not do so at any great length because he recognised that I had already considered the matter on one occasion, and the question was in any event soon going to be considered by the Court of Appeal.

My writing and handing down of this judgment was unfortunately delayed for several months by the need for an adjourned hearing on the issue of sufficiently serious breach, and the position which has now been reached is that the Court of Appeal will have heard argument on the appeals by both sides in FII Chancery (in early October 2009) before this judgment is delivered. Since it will probably be too late for any further thoughts of mine on this subject to be taken into account by the Court of Appeal, and since their judgment, when it is delivered, will inevitably supersede my own, I will confine myself to a few observations.

In the first place, Mr Cavender pointed out that much may depend on the level at which the relevant mistake of law is identified in a mistake-based restitutionary claim. If, as I held in FII Chancery, the mistake should be identified at a fairly general level, as a mistake about the lawfulness of the domestic tax regime which infringes Article 43 (see paragraphs 261 to 262), questions of causation and remoteness will arise in relation to each head of claim that the claimants seek to recover from the Revenue. If, on the other hand, the mistake is identified at a level corresponding to each type of claim, for example as a mistake about the lawfulness of the particular liability to corporation tax which was extinguished by a particular surrender of group relief, there is clearly a direct and immediate connection between the mistake and the benefit to the Revenue which the claimant seeks to reverse. Mr Cavender was disposed to accept, in the light of my judgment in FII Chancery, that a “but for” test of causation was only appropriate for the purpose of identifying any actual payment that had been made to the Revenue as a result of the mistake, and that a stricter test of causation might be needed to identify any associated benefits conferred on the Revenue that should be disgorged. He invited me, in effect, to get round this difficulty by recognising the relevant mistake at a microscopic rather than a macroscopic level; and he submitted that there was no conceptual difficulty in doing so, because the mistake in cases of this nature is to a large extent a legal construct, as I myself said in FII Chancery at paragraph 261.

I do not find myself attracted by this submission, which seems to me in substance an attempt to reintroduce a “but for” test of causation through the back door. In my view it is correct in principle to identify the mistake which grounds the right to restitution at a general level, for the reasons which I briefly indicated in FII Chancery at paragraph 261. More precisely, I would identify the mistake in the present case as the failure to appreciate that the UK thin cap provisions infringed Article 43 because their scope was not confined to cases of abusive tax avoidance. The use of a particular relief to set off against an unlawful tax charge is an example of a consequence of the mistake so identified; it is not a situation in which a separate and specific mistake is made.

Secondly, Mr Cavender pointed out that, although a mistake-based restitutionary claim is not dependent on any fault by the defendant, the same is not true of a Woolwich claim, which is founded on the unlawful levying of tax. He submitted, accordingly, that the objection to a but-for test of causation in the context of a mistake-based claim, which I relied upon in FII Chancery at paragraph 263, cannot apply to all types of restitutionary claim. He suggested that it would be undesirable, and wrong in principle, for different types of restitutionary claim to have different rules of causation, especially as the trend may well be towards the formulation of a single unifying principle of unjust enrichment which will embrace all of the separate restitutionary causes of action which English law now recognises.

I acknowledge that this point has some force, but am inclined to think that the answer to it is that the Woolwich cause of action is, in the context of the law of restitution, relatively unusual in involving a requirement of fault on the part of the defendant. As a matter of general principle, I adhere to the view that a but-for test of causation, save in the limited sense which I recognised in FII Chancery, does not sit happily with the simple fundamental concept of unjust enrichment. If consequential loss identified by a but-for test of causation is to be recoverable at all, it should in my view be recovered by application of a relevant fault-based system of rules (such as the law of tort, contract or breach of statutory duty), and not by what I would regard as an unwarranted extension of the law of restitution.

In all the circumstances, I do not consider that the claimants have any claims, apart from what I would classify as their San Giorgio claims, which are restitutionary claims as a matter of English law.

Is a defence of change of position available in principle to defeat any restitutionary claims which are not San Giorgio claims?

This question raises important issues of principle, but they were barely touched upon in oral argument because I have already dealt with them fully in FII Chancery (at paragraphs 303 to 348) and there is no material difference between the two cases in relation to the potential availability of the defence. It is accepted by the Revenue in the present case, as it was in FII Chancery, that a defence of change of position cannot be relied upon in answer to a San Giorgio claim. Accordingly, if I am right in my view that the claimants have no restitutionary claims under English law which are not also San Giorgio claims, there will in fact be no scope for the defence to operate in the present case. Even if that is wrong, and there is a category of restitutionary claims which are not San Giorgio claims, I have nothing to add at this stage to the detailed discussion of the topic in my earlier judgment.

Introduction

I now come on to the claims for compensation or damages, i.e. the claims which are not San Giorgio claims as a matter of Community law. It is common ground that in order to establish liability under this head three conditions need to be satisfied, which the ECJ has laid down and repeated in a series of cases since Factortame. In paragraph 115 of its judgment in the present case, the ECJ stated the three conditions in familiar terms:

“115.

While it has not gone so far as to rule out the possibility of a State being liable in less restrictive conditions on the basis of national law, the Court has held that there are three conditions under which a Member State will be liable to make reparation for loss and damage caused to individuals as a result of breaches of Community law for which it can be held responsible, namely that the rule of law infringed must be intended to confer rights on individuals, that the breach must be sufficiently serious, and that there must be a direct causal link between the breach of the obligation resting on the State and the loss or damage sustained by the injured parties …”

The Court went on to say in paragraph 116 that it is in principle for the national courts to apply the criteria for establishing the liability of member states for damage caused to individuals by breaches of Community law, in accordance with the guidelines laid down by the Court for the application of those criteria.

As in FII Chancery, there is no dispute between the parties about the first and third conditions. The ECJ itself has held that the first condition is plainly satisfied, because Article 43 confers rights on individuals: see paragraph 117 of the judgment. In relation to the third condition, the Court repeated in paragraph 122 its usual guidance 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. The parties are agreed that a “but-for” test of causation is appropriate for this purpose.

The dispute is whether, and if so from what date, the UK was in sufficiently serious breach of Community law by introducing and maintaining in force the thin cap provisions. Question 8 in the order for reference asked the ECJ, in the same way as Question 9 in the order for reference in FII had asked, 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:

“… in particular as to whether, given the state of the case-law on the interpretation of the relevant Community provisions, the breach was excusable?”

The guidance given by the Court in response to this request was as follows:

“118.

As regards the second condition, it should be pointed out, first, that a breach of Community law will be sufficiently serious where, in the exercise of its legislative power, a Member State has manifestly and gravely disregarded the limits on its discretion. Secondly, where, at the time when it committed the infringement, the Member State in question had only considerably reduced, or even no, discretion, the mere infringement of Community law may be sufficient to establish the existence of a sufficiently serious breach (Test Claimants in the FII Group Litigation, paragraph 212 and the case-law cited there).

119.

In order to determine whether a breach of Community law is sufficiently serious, it is necessary to take account of all the factors which characterise the situation brought before the national court. Those factors include, in particular, the clarity and precision of the rule infringed, whether the infringement and the damage caused were intentional or involuntary, whether any error of law was excusable or inexcusable, and the fact that the position taken by a Community institution may have contributed to the adoption or maintenance of national measures or practices contrary to Community law (Test Claimants in the FII Group Litigation, paragraph 213 and the case-law cited there).

120.

On any view, a breach of Community law will clearly be sufficiently serious if it has persisted despite a judgment finding the infringement in question to be established, or a preliminary ruling or settled case-law of the Court on the matter from which it is clear that the conduct in question constituted an infringement (Test Claimants in the FII Group Litigation, paragraph 214 and the case-law cited there).

121.

In the present case, in order to determine whether a breach of Article 43 EC committed by the Member State concerned was sufficiently serious, the national court must take into account the fact that, in a field such as direct taxation, the consequences arising from the freedoms of movement guaranteed by the Treaty have been only gradually made clear, in particular by the principles identified by the Court since delivering judgment in Case 270/83 Commission v France. Until delivery of the judgment in Lankhorst-Hohorst, the problem raised by the current reference for a preliminary ruling had not, as such, been addressed in the Court’s case-law.”

Paragraphs 118 to 120 of this guidance are in standard form and repeat the well-established jurisprudence of the Court in this area. Paragraph 121 is of more interest, however, because it is directed to the present case and must therefore be taken as the Court’s answer to the request for specific guidance in Question 8 in the order for reference. The national court is directed to take into account:

The clear implication, in my judgment, is that at least until the Court delivered its judgment in Lankhorst-Hohorst (on 12 December 2002) any infringement of Article 43 in the present case (and it must be remembered that the ECJ did not itself hold that there was an infringement) is not to be regarded as an infringement of a rule of Community law which was clear and precise in its application to the UK thin cap provisions. For the same reason, any error of law on the part of the UK in continuing to operate, and amending, the thin cap provisions is more readily to be regarded as excusable. However, in my judgment the guidance goes no further than that. The national court is still required to take account of all the relevant factors, including but not limited to those enumerated in paragraph 119, before forming a value judgment on the question. Furthermore, the words “as such” in the final sentence of paragraph 121 provide a reminder that developing trends in the case law of the Court before Lankhorst-Hohorst may be relevant, even if no specific challenge to the UK thin cap provisions had yet been mounted.

It is instructive at this point to compare what the ECJ said with the view expressed by the Advocate General in paragraph 108 of his opinion:

“108.

I would add that, in the FII case, which concerned the UK’s tax treatment of incoming dividends, I expressed serious doubts whether the Brasserie du Pêcheur conditions – and in particular the requirement of a sufficiently serious breach – was fulfilled for the aspects of the UK’s system which breached Community law. I have even stronger doubts on this point in the present case. The application of Article 43 EC to national thin cap legislation was confirmed by the Court only in 2002 with its Lankhorst-Hohorst judgment, and even following this judgment the scope of such application has not been totally clear. Moreover, the UK altered its legislation on numerous occasions, making the application of its rules more transparent and seemingly, in the case of the 2004 changes, keeping compatibility with Community law in mind. This does not seem to me sufficient to constitute a manifest and grave disregard of the limits on its discretion within the meaning of the Court’s case-law.”

It seems to me that the Advocate General was making essentially the same point as the Court about the uncertainty of the position under Community law before Lankhorst-Hohorst, but he went further than the Court in expressing what appears to be a strong prima facie view that any breach of Community law which may have been involved was not sufficiently serious. The weight that can be attached to such a view is, however, limited, given that the issue of sufficiently serious breach is exclusively one for the national court.

The leading English authority on sufficiently serious breach remains the decision of the House of Lords in R v Secretary of State for Transport, ex parte Factortame Ltd (No. 5) [2000] 1 AC 524 (“Factortame No. 5”), which I reviewed at some length in FII Chancery at paragraphs 363 to 375. I will not repeat that analysis, but will add to my citation (in paragraph 375 of FII Chancery) of the list of potentially relevant factors identified by Lord Clyde at 554-5 two further factors upon which Mr Aaronson QC placed considerable reliance on behalf of the claimants:

“5.

It is also relevant to look at the state of mind of the infringer, and in particular whether the infringer was acting intentionally or involuntarily. A deliberate intention to infringe would obviously weigh heavily in the scales of seriousness. An inadvertent breach might be relatively less serious on that account. Liability may still be established without any intentional infringement. More broadly, the purpose of the infringer should be considered. If the purpose was to advance the interests of the Community a breach committed with that end in view might be seen as less serious than one committed with the purpose of serving merely national interests.

6.

The behaviour of the infringer after it has become evident that an infringement has occurred may also be of importance. At the one extreme the immediate taking of steps to undo what has been done and correct any error which has been committed may operate to mitigate the seriousness of the breach. At the other extreme a persistence in the breach, the retention of measures or practices which are contrary to Community law, especially where they are known so to be, will add to the seriousness of what has been done …”

As I said in Chalke at paragraph 182, I also derive assistance from the discussion of the subject by the Court of Appeal in Byrne v Motor Insurers’ Bureau [2008] EWCA Civ 574, [2009] QB 66 (“Byrne”), which was not reported until after I had handed down my judgment in FII Chancery. The only reasoned judgment in Byrne was given by Carnwath LJ, with whom Keene and Waller LJJ agreed. In paragraph 45 of his judgment, Carnwath LJ said that the “sufficiently serious” criterion laid down by the ECJ is not a hard-edged test, and it requires a value judgment by the national court, taking account of the various factors summarised by the ECJ.

Disclosure of documents by the Revenue

Because the scope of the enquiry by the national court on the issue of sufficiently serious breach is potentially so wide, and because in particular issues of subjective intention as well as objective excusability are involved, the question of disclosure of documents may give rise to difficulties. This is especially so where, as in the present case, the claims cover a period of several years and the relevant legislation was amended on more than one occasion. On the one hand, it is imperative that relevant policy documents, even if they go to the heart of the government machine, should be disclosed without reluctance, however embarrassing they may be, and however difficult it may be to uncover them in physical or electronic archives, possibly many years after they were created. On the other hand, it is necessary to keep a sense of proportion, and not to place an unreasonable burden of disclosure on government departments which have to keep a host of questions of a more or less general nature under review, and where the implications of developments in Community law may be considered in a variety of different contexts. Furthermore, government departments, like private litigants, are of course fully entitled to claim legal professional and litigation privilege, subject to a possible argument that Community law may in certain circumstances override that right, and subject also to the proviso that privilege may be waived, either deliberately (as seems to have happened in Factortame (No. 5)) or inadvertently.

Even making every allowance for the difficulty and sensitivity of the exercise, I have to say that the history of disclosure by the Revenue on this issue in the present case has not been a happy one.

The original order for standard disclosure in relation to the then test claims was made by Park J on 14 January 2004, before the reference to the ECJ. When the matter was restored for further directions after the ECJ had given judgment, Rimer J ordered on 6 July 2007 that the parties were to provide further standard disclosure by list of any documents not previously disclosed relating to the determination of the GLO liability issues by 1 February 2008, and inspection of such documents by 15 February. A further order in similar terms was made by myself at a case management conference on 4 February 2008, when the liability issues were set down for trial in a window between 3 November 2008 and 30 January 2009. My order provided for further standard disclosure by list of any documents not previously disclosed by 30 May 2008, with inspection within 14 days thereafter.

Notwithstanding these orders, and despite numerous requests by the claimants for specific disclosure of all internal Revenue and Treasury documents concerning the UK’s state of knowledge of the possible incompatibility of the thin cap provisions with Community law from 1973 to date, nothing of any substance was disclosed until 18 December 2008, just before Christmas and a few weeks before the start of the trial, when five lever arch files of documents were belatedly disclosed. No explanation or excuse was given for the late disclosure of this considerable volume of material. The documents which the Revenue said they were willing to disclose, and in respect of which no claim to privilege was asserted, were 258 in number and ran to 1,852 pages. They covered the period from February 1994 to August 2003. Certain passages in these documents, however, were redacted, apparently on grounds of irrelevance or reference to privileged material, although the basis upon which the redaction had been made was often far from clear, and there were some instances of the same document appearing twice in both redacted and unredacted form.

Although this deplorably late and careless disclosure caused the claimants obvious difficulties in preparing for trial, they did not ask for an adjournment and the trial began as scheduled on 20 January 2009. On Day 4 Mr Aaronson cross-examined Mr Brooks and Mr Black in relation to some of the documents, and on Day 6 (27 January) he began his submissions on the issue of sufficiently serious breach. One of the questions which he had explored in cross-examination with Mr Brooks and Mr Black, and on which he proceeded to make submissions, was the impact of the decision of the ECJ in a case in 1994 called Halliburton Services BV v Staatssecretaris van Financien (Case C-1/93, [1994] ECR I-01137, “Halliburton”). At least one of the disclosed documents appeared to refer to the substance of some advice that had been given by the Office of the Solicitor of Inland Revenue in relation to the possible implications of Halliburton, and I expressed the view that privilege had probably been inadvertently waived in relation to that advice, and possibly also in relation to the further question of the potential susceptibility to challenge of the 1995 amending legislation, but not necessarily any further. Mr Ewart for the Revenue indicated that he was content for the time being to proceed on that basis, and Mr Aaronson continued with his submissions.

Later on the same day, the question of waiver of privilege arose again in relation to a redaction in another document. At that stage Mr Ewart squarely submitted that the Revenue had not waived privilege, either deliberately or accidentally, in relation to advice on the 1995 legislation, and it was agreed that I would hear further argument on the question the following day.

I duly heard submissions on the question on the morning of Day 7, and I ruled on it at the beginning of Day 8 (29 January). I held that there had been an inadvertent waiver of privilege by the Revenue, which extended at least to legal advice taken, and related communications between the Revenue and their lawyers, on the question of the compatibility with Community law of the proposed new 1995 legislation. In the light of that ruling, which the Revenue did not ask for permission to appeal, it soon became clear that the existing disclosure would need to be reviewed and supplemented, and that an adjournment for that purpose would be necessary. It was accordingly agreed that this part of the trial would be resumed after the disclosure exercise had been completed, but that the remainder of the trial could continue in the meantime.

Unfortunately the completion of the disclosure exercise proved to be far from straightforward. To begin with, there was some dispute about the precise extent of the waiver, but this was resolved by early March when the wording of my order of 29 January was finally agreed. The order defined the issue in respect of which privilege had been waived as “the vulnerability to challenge under Community law of the changes to the thin capitalisation provisions introduced by section 87 and Schedule 29 of the Finance Act 1995”. It directed the Revenue to conduct a further search for documents relating to the issue, including documents which would otherwise have been privileged, and to provide specific disclosure by list of all such documents. The adjourned hearing was directed to take place no earlier than 20 April 2009, with a two day time estimate, and the claimants were given permission to serve amended particulars of claim particularising their allegations that the infringement of Community law was not involuntary and that the error of law was not excusable, such amended particulars to be served no later than 14 days before the hearing.

The Revenue made their further disclosure on 17 March 2009, but the claimants were not satisfied with it. On 5 May the Revenue supplied three further documents, and gave reasons for not producing others. They also gave reasons for certain redactions which had been made in the further disclosure, namely legal professional privilege, taxpayer confidentiality and relevance. The claimants were still dissatisfied, and issued a further application for specific disclosure on 11 May which was listed to be heard at the same time as the resumed main hearing on 21 May. Following further correspondence, it was then agreed, and ordered by consent, that the hearing would be postponed in order to allow time for further searches by the Revenue. In the event, the Revenue made further disclosure of documents on 12, 16 and 18 June. The postponed hearing finally took place on 2 and 3 July 2009, some five months after the initial adjournment at the end of January. Meanwhile, the claimants had served amended particulars of sufficiently serious breach on 15 May which took account of most, but not all, of the additional disclosure. For the purposes of the adjourned hearing, the Revenue’s completed disclosure was arranged in ten chronological bundles, with the pages helpfully colour-coded to indicate the date on which each document was disclosed.

Evidence

In the light of the documentary record and the oral evidence of Mr Brooks and Mr Black, I find the following facts.

By early 1994 the thin cap legislation was under review in International Division, as part of a wider review of the taxation of interest. It will be remembered that the basic legislative provisions in force at this time were ICTA section 209(2)(d) and (e), and interest payments made by a UK-resident company to another group member outside the UK were always treated as a distribution, even where the interest represented a reasonable commercial return on the loan. However, this draconian general rule was overridden in many cases by provisions in DTCs which introduced an arm’s length test and, in effect, allowed a deduction for interest which satisfied the test: see paragraphs 20 to 26 above. One of the matters under consideration by International Division was the compatibility of this thin cap regime with Community law, and the need to ensure as far as possible that any new thin cap legislation would be “Euro-proof”. Another concern was to ensure that there would be no conflict with non-discrimination articles in DTCs.

A senior official in International Division with particular responsibility for EU matters was Mr P A Michael. He had also been involved in a previous review of the thin cap regime in 1989/90. On 3 March 1994 he sent an internal memorandum to Mr E Jukes, also of International Division, who had prepared a discussion paper on possible new thin cap legislation. Mr Michael pointed out that there is “a big difference between on the one hand targeting international tax avoidance within groups of companies and on the other hand hitting groups which for reasons which may have nothing to do with tax are by their standards behaving perfectly normally”. He said this was a point about which considerable concern had been expressed in the context of the previous thin cap exercise, and continued:

“I believe Sue Walton has mentioned to you separately the need to bear in mind the EC dimension. If we are going to work up a scheme along the lines outlined in the papers we shall need to make sure that we do not run a significant risk of covert discrimination. The real problem case is the one to which I have referred to above … In particular, I feel uneasy about imposing UK standards when the facts strongly indicate (or for that matter prove) that a disproportionate amount of the group’s total indebtedness is not being dumped in the UK. This too needs a good deal more thought.”

On 9 March 1994 another senior official in International Division, Mr R E Haigh, replied to Mr Michael’s memorandum (which, as was evidently the general practice, had been copied to a long list of recipients). He reassured Mr Michael that “the EC dimension is not being and will not be overlooked”. He also commented on the concern which Mr Michael had expressed, in a passage which is interesting for its recognition that thin cap rules may apply in cases where there is no tax avoidance motive:

“7.

Returning to the comments in your second paragraph I was concerned at the implication that any thin cap rule we might develop should operate only in the case of international tax avoidance – so that we might for instance be obliged to have some sort of avoidance motive test. In many instances thin cap does indeed represent a device to mitigate tax but I would accept that it does not always – there may indeed be business reasons why debt finance has been raised rather than equity in any particular case, or there may be a mixture of motives … But the point remains that such a business choice erodes the UK tax base.”

On 11 March Mr Michael replied to Mr R E Haigh, saying he personally doubted whether new thin cap legislation would be defensible if it applied in cases where there is no avoidance involved, but “at the end of the day you might well be proved right”. He then discussed various proposals which had been mooted in 1989, and added some thoughtful observations on “the EC dimension”:

“7.

So far as the EC dimension is concerned what can be said with certainty is that if the [ECJ] is in the mood it will have no qualms about trampling all over and rejecting OECD principles. For evidence of that you need look no further than Commission v France and Commerzbank … I suppose it must be the case that thin capitalisation legislation will always be vulnerable to challenge … But it is slightly unfortunate that our debt/equity ratios are out of line with our European partners and enshrining those ratios in legislation would set up a high-profile target. The counter argument I assume would be that the UK company tax regime must be considered as a whole, is built around UK ratios, the arm’s length principle [is] internationally recognised and preserves the “cohesion” of the system both domestically and at the international interface. Whether that argument would succeed for the Court I do not know.”

On 12 April 1994 the ECJ delivered its judgment in Halliburton. The case concerned two subsidiaries of a US holding company, one resident in Germany and the other resident in the Netherlands. The German company had a branch in the Netherlands, which it subsequently sold to the Dutch company as part of a group reorganisation. The sale included some immovable property on which the Dutch tax authorities levied land transfer tax. Had both companies been resident in the Netherlands, the transfer would have been exempt from tax. The claim for tax was resisted on the grounds (among others) that it was inconsistent with Articles 52 and 58 of the EEC Treaty (freedom of establishment). In holding that the challenge by the Dutch company succeeded, the ECJ laid down two important principles. First, repeating what it had already held in Commerzbank (Case C-330/91, The Queen v Inland Revenue Commissioners, ex parte Commerzbank, [1993] ECR I-4017, [1994] QB 219), the Court held (paragraph 15) that:

“the rules regarding equality of treatment forbid not only overt discrimination by reason of nationality or, in the case of a company, its seat, but all covert forms of discrimination which, by the application of other criteria of differentiation, lead in fact to the same result”.

Secondly, the Court held that the right of establishment of the German company had been infringed, even though the company which suffered the charge to tax was the Dutch company. The argument of the Dutch government that Community law was not engaged, because the situation was purely internal to the Netherlands, was rejected, on the ground that the imposition of the tax would be reflected in a lower purchase price received by the vendor, and therefore placed the vendor in a less favourable position than if it had carried on its business in the Netherlands through a Dutch subsidiary (paragraphs 18 to 20).

I comment that both of these principles had potential implications for the UK’s thin cap regime. The first principle showed that a system which automatically recharacterised all interest paid to a non-resident EU parent company as a distribution, whereas in a purely domestic context only interest which represented more than a reasonable commercial return was so recharacterised, prima facie involved indirect discrimination on the grounds of nationality, because the residence criterion would in practice almost inevitably apply when an EU resident parent lent money to a UK subsidiary (the only exception being where the loan was made through a UK branch or agency of the parent). Just as importantly, the second principle suggested, albeit in a very different fiscal context, that the UK might be unable to defend a challenge to the thin cap regime on the basis that the company which suffered the additional charge to tax was the UK subsidiary, not the foreign parent. There was clearly a risk that the ECJ would say that the parent company’s right of establishment was engaged, and it was not just a purely internal UK tax matter.

Interested taxpayers soon began to raise arguments based on Halliburton, which were referred to International Division. The first case appears to have involved a UK subsidiary of a Canadian company which issued a promissory note at a deep discount to a fellow subsidiary in Ireland. On the footing that the discount represented a payment of interest on a security issued by the UK company, within the meaning of ICTA section 209(2)(e)(iv), the question arose whether the recharacterisation of the interest as a distribution was compatible with Community law. Presumably because of the corporate structure, with a non-EU holding company, freedom of establishment was not in issue; but the free movement of capital (then contained in Article 67) was, together with the prohibition on discrimination in Article 7. In a memorandum dated 2 June 1994 Mr Michael commented that the argument based on Article 7 was more difficult from the Revenue’s point of view, “since section 209(2)(e)(iv) certainly smacks of covert discrimination on grounds of nationality”. He said that the Revenue’s argument would have to be that the provision was nonetheless objectively justified because it preserved the cohesion of the tax system, by preventing foreign parent companies shifting profits out of UK subsidiaries to avoid the UK corporation tax charge. He continued:

“The problem with this however is that looked at in isolation section 209(2)(e)(iv) is a “blunderbuss” provision and might therefore fail the proportionality test. At that point I suspect things start to get a bit tricky and less clear cut.”

He recommended that the papers should be referred to Mr W J Durrans in the Solicitor’s Office for advice.

On 30 June a request for advice was duly submitted to Mr Durrans. On the same date Mr R E Haigh asked Mr Durrans if they could meet to discuss the matter informally before he provided his advice. He said he wished to talk about the possible ramifications of Halliburton for the provisions of section 209(2)(e)(iv) more generally, and expressly asked whether the provisions might be susceptible to challenge under any other Articles of the Treaty:

“In particular I wonder whether it could not be said that Section 209(2)(e)(iv) offends the freedom of establishment provisions of Chapter 2 – on the basis that a non-resident company will find it more expensive to establish a subsidiary in the UK, financed with debt, the interest on which does not attract relief for the subsidiary in arriving at its UK tax bill, than a UK competitor company establishing an equivalent UK subsidiary, complete with its UK tax relief for interest charges.

4.

Section 209(2)(e)(iv) is the basis for the UK’s thin capitalisation defences. If it is susceptible to challenge then we would have to move quickly to put something else in its place. I would like to be able to establish what the likely risk of a successful challenge is generally …”

This was evidently not the only request for advice arising out of Halliburton which Mr Durrans had received, and on 12 July 1994 he produced a paper giving an “overview” of the decision and stating some general conclusions. I will not quote the conclusions, which follow a careful review of the judgment of the Court and the opinion of Advocate General Lenz, but they clearly drew attention to the implications of Halliburton in the context of freedom of establishment and pointed out that the defence which the Dutch government had advanced could no longer safely be relied upon. He also pointed out that the ECJ did not appear to treat group companies any differently from companies not in a group, and every company had the right to freedom of establishment even if it was in fact acting under the direction of its parent. However, he did offer one crumb of comfort:

“22.

To end on a slightly more optimistic note, Halliburton was concerned with a transaction which was internal to the Netherlands … There was, therefore, no Bachmann cohesion of the tax system argument available to the Dutch. It seems to me that where the transaction is a cross-border transaction it may still be possible to argue that the different treatment is justified by the need to ensure the cohesion of the UK tax system.”

The informal meeting which Mr Haigh had requested with Mr Durrans seems to have taken place a little later, but before 22 July when Mr Haigh sent a memorandum to the director of International Division, Mr Spence, to inform him about recent developments relating to thin capitalisation and the interest review. This is an important document, because it leaves little room for doubt about Mr Durrans’ views on the vulnerability of the UK thin cap rules:

“Mr Durrans’ view is that we are vulnerable to challenge and that our chances of resisting the challenge are considerably less than slim. Very briefly, it is likely that the provisions are contrary to European law on grounds that they are discriminatory and because, viewed as a piece of anti-avoidance legislation, they are out of proportion to the abuse which they seek to counter.”

Mr Haigh continued:

“Loss of section 209(2)(e)(iv) to combat thin capitalisation even if only for EC transactions would be extremely costly. The UK would quickly become a dumping ground for increasing amounts of intra-group debt, routed via the EC where necessary. There would be a considerable drain on the UK tax take. I have no doubt whatsoever that, on the basis of Mr Durrans’ advice, we have no alternative but to seek a legislative solution.”

Mr Haigh went on to identify a number of options for addressing the problem, although he said none of them was particularly appealing. One option would be to work up a complete package of new legislation for inclusion in the Finance Act 1995, but this would leave no time for the consultation process that would be desirable, and expected by the outside world, for major new legislation in such a controversial area. The second option would be to take limited legislative action in the Finance Act 1995 “to amend the existing provisions to make them EC proof”. Mr Haigh commented:

“That in itself might not be without difficulties but could be accomplished in the time available and a lack of consultation would be more defensible – and could be explained. However, it would be far from an ideal solution. We acknowledge that, present EC pressures apart, the legislation is by no means perfect so that tinkering at the edges would represent an impoverished course of action, in particular coming so close after the Interest Review. It would be no more than a patch-up job.”

The third option would be to defer legislation until the Finance Act 1996, which would enable the Revenue to work up the new provisions on the basis of full consultation and as a natural continuation of the Interest Review. Mr Haigh said that this would be by far the most preferable of the legislative options, but it left the difficulty of how to deal with the interim period.

Mr Haigh continued:

“9.

We are not entirely without arguments on which to place a defence of the application of Section 209(2)(e)(iv) and there are some possible bargaining counters we could use in order to negotiate agreements with those companies raising the matter, so one option is to resist any challenge to the application of section 209 by deploying these arguments/counters pending the introduction of legislation in 1996. If we could do so, that would be a satisfactory approach from our viewpoint. But Mr Durrans has made it clear that he believes such arguments would ultimately fail and that they could not be pressed to litigation.

…

10.

It could instead be accepted that, following Halliburton, Section 209(2)(e)(iv) is in breach of European law and that it cannot be applied, beyond the time of the Halliburton decision, to affected transactions. Effectively that would largely close down the thin capitalisation shop until remedial legislation began to bite. Even if this were only for one year – and it could be for much longer … - the damage would be serious. For 1993/94 the yield … was £204 million from adjustments to profits of £650 million and a fair proportion of this related to the combating of thin capitalisation cases. That would represent a mere fraction of the potential tax loss should we openly acknowledge that we had lost Section 209(2)(e)(iv), leaving the field clear for the tax planners. In such case, announcement by the Chancellor to signal the intention to legislate for thin capitalisation in 1996 might help to put a brake on the process but it could be a pretty ineffectual brake.”

Mr Haigh went on to say that his preference would be for limited remedial action in the Finance Act 1995, coupled with an announcement of a full continuing review. He also said that the new Financial Secretary to the Treasury (Sir George Young MP) should be informed without delay.

Four days later, on 26 July 1994, a policy memorandum was sent by Mr Michael to the Chancellor of the Exchequer (Mr Kenneth Clarke MP) and copied to a long list of recipients, including the other Treasury minsters. It was headed “EC: Direct Tax: European Court of Justice”, and described the “issue” as being “Early warning of problems ahead with the Court of Justice applying the principles of the Treaty to national tax provisions”. The memorandum is too long to quote in full, but the summary at the beginning gives the flavour:

“1.

As we have mentioned to Ministers in the past, the European Court has held in recent years that provisions of direct tax systems at national level (which continue to remain unharmonised) are subject to the non-discrimination rules of the Rome Treaty. So far, we and other Member States have survived largely (albeit not completely) unscathed. To that extent the present position whilst sub-optimal is just about tolerable. But there is no guarantee that things will stay this way and the evidence now rapidly accumulating strongly suggests that they will not. There is a growing risk of a major setback in Exchequer and political terms.

2.

In so far as there is any comprehensive solution to this problem the only option worth serious consideration is a Treaty revision with the objective of securing tax carve-outs in the relevant Articles. At this distance, the prognosis for achieving that is highly uncertain … In practice, much may turn on future developments including the judgment in what will become a leading case, Schumackers (involving Germany), expected in early 1995. However, even at this stage our concerns are shared by senior officials in France and Germany and we have agreed to meet early next year to go over the issues.

…

3.

Short of amending the Treaty there are very real limits to what can be done at national level to reduce the potential for damage. But there are some practical steps that can be taken which would involve building on the strategy we have developed over the last few years. This has consisted of –

(i)

making full use of our right to make observations to the Court in individual cases (and encouraging others to do likewise);

(ii)

ensuring that we do not put new provisions on the statute book with obvious vulnerabilities; and

(iii)

where practicable and consistent with Parliamentary sensitivities, using legislative opportunities as and when they arise to revisit existing legislation.”

In paragraph 11, under the heading “Exchequer consequences”, Mr Michael said it was “well within the bounds of possibility” that the actual cost of any future setbacks could run into “hundreds of millions if not billions of pounds” if the ECJ continued to erode the residence basis of taxation. He added that, in the field of tax, decisions of the Court “will almost invariably have at least some retrospective effect”.

In the course of argument at the resumed hearing in July Mr Aaronson described this document, in terms which I would respectfully endorse, as “a remarkably comprehensive, analytical and accurate assessment”.

The next task was to draft a detailed minute to the Financial Secretary setting out the legislative options for dealing with the thin capitalisation problem and making appropriate recommendations. The documentary record shows that this draft went through a number of versions over the course of the summer, and was finally sent on 19 August 1994. In the meantime, the Financial Secretary had read Mr Michael’s memorandum to the Chancellor and on 28 July he made some brief comments which were passed on to Mr Michael. On 1 August Mr Michael responded to these comments. In answer to a request for further details of the challenges which were anticipated, he said this:

“The vulnerabilities of ourselves and indeed other Member States run to the length of a ball of string. Our external customers have not given notice of where they intend to concentrate their attention next. For internal purposes we have drawn up a list of UK vulnerabilities. We shall let you have a copy as soon as it is in a more user-friendly form.”

Despite repeated searches, I was told that it has not been possible to locate a copy of this list, but there can be no doubt that it must have included the thin cap provisions.

On 6 August 1994 Mr Haigh wrote to Mr Durrans asking for his urgent advice on two questions:

Mr Durrans gave his advice on 11 August. He confirmed that the proposed amendments, leaving aside the section 212 exception, would result in a non-discriminatory piece of thin cap legislation. He continued:

“The difficulty with the notion of adding to these amendments a provision excluding intra-group payments between UK residents is, of course, that the additional provision reintroduces discrimination. Using the device in s.212 ICTA 1988 of disapplying s.209(2)(e)(iv) where the lender is within the charge to corporation tax, whether in relation to the loan income or any income, would avoid direct discrimination contrary to Article 52 but not, I think, indirect discrimination. This is because although on its face the exclusion would apply without distinction to all companies wherever resident it would essentially favour UK resident companies.”

Mr Durrans went on to consider whether any such indirect discrimination could be justified on objective grounds consistent with the Treaty, and discussed the defence of fiscal cohesion which the ECJ had applied in Bachmann. He said it was “not altogether easy to anticipate in relation to particular circumstances whether the ECJ would accept an argument of objective justification”, partly because of the paucity of the case law and partly because of general uncertainty whether the Court would continue to follow a line which it had previously taken. However, he considered that there would be “a reasonable prospect” of defending an amended section 209(2)(e)(iv) linked to an exclusion for interest where the lender was within the charge to corporation tax. He then set out at some length his reasons for coming to that conclusion. Mr Aaronson described Mr Durrans’ advice as “very measured and accurate”, and again I would respectfully agree.

On 11 August Mr Spence circulated an internal memorandum in which he described Mr Durrans’ analysis as “extremely helpful”. The conclusion he derived from it was that the Financial Secretary should be given a firm recommendation for legislation in the forthcoming Finance Bill:

“The residual risk that our Euro-proofing might still get knocked flat by ECJ is something we will have to live with … The risk of our losing in ECJ is pretty marginal – clear enough from Bill Durrans’ note, and confirmed to me on the telephone. If the ECJ do hit us, it will be on the basis that the only way of conforming to EU law will be for UK to impose on domestic taxpayers a nonsensical tax charge, with significant compliance burdens. If the threat comes, the Germans and others will have to join in with us in representations to ECJ to protect their own position. (Because our new formula is, in effect, modelled on what Germany already has to keep its domestic taxpayers out of the frame.) If ECJ were nevertheless perverse enough to find against us, it would be such a preposterous result that it would add momentum to the campaign to get an ECJ change in the Treaty of Rome …”

Mr Spence then gave instructions for the submission to Ministers to be revised, with “the main focus on the immediate Euro-proofing issue”, and for draft instructions to Parliamentary Counsel to be finalised.

On 19 August 1994 the submission to the Financial Secretary was signed off on behalf of Mr Spence and sent to him, with copies going to the private offices of the Chancellor and other Treasury ministers and the usual long list of other recipients. The submission reviewed the decision of the ECJ in Halliburton and its implications for the UK thin cap legislation. Mr Spence’s advice was stark:

“9.

… It is now clear that our vulnerability is real, and acute. It arises because our thin capitalisation legislation expressly applies only to payments made to non-residents … Our legal advice, in the light of Halliburton, is that our legislation would be held to be discriminatory and contrary to Community law, if we were taken to Court.

10.

The final nail in the coffin of our defences arises from an element of over-kill in the thin capitalisation rules. Where they apply, they knock out all of the interest, not just the amount that exceeds what would be involved in an arm’s length financing arrangement. So the legislation can apply even when companies’ financing decisions are not driven primarily by tax considerations. On this count the provisions would be held to have an effect disproportionate to the abuse they counter, on top of being “discriminatory”.

11.

The over-kill effect is usually limited through our double taxation treaties with other Member States. But the ECJ has specifically held that alleviating provisions in double taxation treaties will not be taken into account in considering the basic legislation. In any event, some of our other double taxation treaties (eg with Ireland and Italy) have no such alleviation.”

Mr Spence’s comment in paragraph 11 about alleviating provisions in DTCs needs to be noted. The clear view at the time, based on statements by the Court in earlier case law including Case 270/83 Commission v France, was that the compatibility of the UK thin cap legislation with Community law had to be considered without reference to the DTCs which in most cases introduced an arm’s length test. On that footing, section 209(2)(e)(iv) as it stood was pretty clearly indefensible. However, in its judgment in the present case the ECJ considered the pre-1995 legislation in conjunction with the relevant DTCs and looked at the overall effect of the provisions. Whether it would have done the same in 1994 is an open question, but two points may be made. First, the jurisprudence of the Court is far from immutable, and as well as building on earlier foundations the Court does sometimes, either explicitly or tacitly, depart from principles which it has previously enunciated. Secondly, it can now be seen that the deeply pessimistic views expressed in 1994 about the vulnerability of the UK legislation were in one very important respect based on a premise which may have been false. If the ECJ would then have been willing to consider the UK legislation in conjunction with the relevant DTCs, the problem of over-kill would have been eliminated in the case of most, if not all, member states, and a defence based on the need to counter tax avoidance would have been potentially available, even if it might not ultimately have succeeded, as well as a defence based on fiscal cohesion as it was then understood.

Mr Spence went on in his submission to deal with the tax revenue at risk and the inadvisability of delaying remedial legislation. He recommended an announcement on Budget Day that payments of interest made under thin cap arrangements from then onwards would be looked at under new, “Euro-proofed” rules. The immediate legislation which he recommended took the form which Mr Durrans had cautiously endorsed. Finally, Mr Spence discussed the case for more general reform of the thin cap regime.

At the beginning of September the Financial Secretary gave his approval to the proposed legislative changes, and detailed instructions to Parliamentary Counsel were worked up. As one would expect, points of detail were then discussed between Parliamentary Counsel and International Division. On 5 October Mr Spence sent a further memorandum to the Financial Secretary, in which among other matters he drew attention to the residual risk that the solution adopted would still be open to challenge in the ECJ. He said:

“It needs some rather subtle drafting to produce an end-product which both guards against ECJ challenges and still leaves intra-UK loans unaffected. The solution we have got, modelled on the German solution, should do the trick. But it leaves a residual risk of an ECJ challenge. This clearly has to be accepted. But we thought you should be aware of it.”

When he gave his approval to the proposed legislation on 2 September, the Financial Secretary asked Mr Spence for a draft minute to send to the Chancellor and, if appropriate, a draft press release. Mr Spence then prepared a draft note to the Chancellor, and on 10 October 1994 the Financial Secretary sent a note to the Chancellor which largely reproduced Mr Spence’s draft, but also included a final paragraph dealing with the wider repercussions of Halliburton. The note said in terms that the effect of Halliburton was that the UK thin cap legislation “would be held to be contrary to Community law if we were taken to Court”, and that unless immediate remedial action were taken tax planners were likely to exploit the breach in the UK’s defences on a large scale, and £150 million tax revenue would be at risk.

Paragraphs 4 and 5 of the note read as follows:

“4.

I am clear that we should take immediate remedial action to Euro-proof the anti-avoidance provisions, so that they remain effective despite the Halliburton ruling. The solution proposed broadly reproduces the status quo. In particular, it would still ensure that intra-UK loans in practice are kept out of the thin capitalisation net. There is a residual risk of an ECJ challenge in this. But it is much better to accept this small risk than engage in the policy nonsense of exposing intra-UK loans to unnecessary restrictions, and the compliance burdens that would go with them. The legislation would have two incidental effects. One is that the immunity currently enjoyed by Japanese and German parent companies because of the terms of the current tax treaties will disappear. This will remove an anomaly and should not be controversial.

5.

Another effect of the revised rules will be to take out an element of over-kill in the present provisions. Under the current rules, the whole of the loan is caught if the thin capitalisation legislation applies. Our tax treaties override this to produce a more sensible result: only the excess of the loan (over and above an arm’s length standard) is caught. The new legislation will generalise the tax treaty treatment, even where no tax treaty is in force. This relaxation should be presentationally helpful.”

In the final paragraph of his note dealing with the wider repercussions of Halliburton, the Financial Secretary said that the Euro-proofing legislation on thin capitalisation was “only the first step in what threatens to be a long and extensive Euro-proofing exercise”. He said that the Revenue were “doing a lot of detailed work on the extent of our vulnerabilities”, on which he promised to let the Chancellor have a further note. He concluded:

“This year’s thin capitalisation legislation means replacing a few lines of the tax code by 3 pages of complex legislation, with some fancy footwork to avoid the nonsensical result of bringing intra-UK loans into the thin capitalisation net. But this, I fear, is only a minor example of the problems we will face in dealing with the impact of the Halliburton decision, and other ECJ decisions in the future.”

Meanwhile an initial draft of the budget press release was prepared, and on 10 October it was circulated for comment to a number of colleagues in International Division by Mr Ian Hardie. Mr Hardie was the leader of a small team which had been established in International Division to deal with the remedial legislation. The other members of the team were Mr Black, Mr Brooks and Ms Lyn McDonald. According to paragraph 1 of the draft press release:

“Following a recent European Court of Justice decision claims have been made that the rules relating to certain payments between connected companies is [sic] discriminatory. Although the Inland Revenue has not accepted that this is so, the Government wishes to amend the legislation in a way designed to avoid any such suggestion.”

Paragraph 3 of the annexed Notes for Editors said:

“Following a recent European Court of Justice decision involving a Netherlands company (Halliburton Services BV) it has been suggested that these provisions may be discriminatory and would be struck down by the ECJ. Although the Inland Revenue has not accepted that this analysis is correct, it is hoped that the new provisions will put the matter beyond doubt.”

On the same day Mr Michael responded with some comments on the draft. He said he had little doubt that Mr Spence would have his own ideas on presentation, but he personally would prefer something on the following lines:

“A decision of the Court of Justice has cast doubt over some aspects of the distribution legislation as it applies to certain payments of interest and other similar sums between connected companies.

This has created uncertainty for the Revenue and taxpayers alike.

The proposed change is intended to clarify the position in a way which will result in no change of practice for the majority of companies etc.”

He said he believed that this formula would also find favour with the Treasury, and if his colleagues agreed the draft press release would need “some slight amendment”.

On 14 October Mr Hardie sent a copy of the first draft of the press release to the press office, asking for their assistance on certain points. He also included Mr Durrans in the list of recipients, and asked him specifically to indicate whether he was content with their efforts “to simplify the description of the ECJ impact and the purpose of the new legislation”.

Down to this stage no serious criticism could in my view be made of the proposed press handling of the remedial legislation. It is true that it sought to put the best complexion it could on a deeply worrying problem, but the government should not in my view be criticised for adopting a fairly bland public approach to the implications of Halliburton and not advertising its concerns to the tax planning industry. However, matters then took a different turn following a meeting between Mr Spence, Mr Michael and the Financial Secretary. On 19 October Mr Spence sent a memorandum to Mr Hardie from which I need to quote substantial extracts:

“BUDGET PRESENTATION OF THE ECJ THREAT – THIN CAP ETC

1.

FST’s decision at a meeting Peter Michael and I had with him yesterday was that everything pointed to low-key, low profile presentation of thin cap and other ECJ vulnerabilities at Budget time – and for as long as possible afterwards. This for a multitude of good reasons like:-

(a)

discouraging the tax planning market from thinking we are seriously worried we will lose in the ECJ if they take us to court on our vulnerable areas

(b)

similarly, damping down speculation that there will be a raft of Finance Bill Euro-proofing legislation in future years

(c)

not provoking the question as to whether remedial action is on our ECJ agenda

(d)

avoiding shock/horror politicking from the Euro-sceptical tendency in Parliament.

2.

In other words, the public stance for as long as possible is that we are looking at a boring technical issue. And a brave face to anyone who suggests they have got us bang to rights on ECJ vulnerability. “We are not convinced; as far as we are concerned UK law has the effect we have intended it to have unless and until ECJ rules otherwise; anybody, of course, has the right to take us to the courts (and through to ECJ) if they want to”.

3.

Line on Thin Cap at the Budget

(a)

Nothing in the Budget Speech.

(b)

Press notice presentation on the lines that: speculation about the effect of the ECJ judgment in relation to thin capitalisation; legislation introduced to remove this uncertainty; taken the opportunity to modify impact of rules in one significant respect [knocking out blunderbuss effect]; operation and effectiveness of legislation will be kept under review [general reform angle].

…

5.

Press Office briefing – As above, with a general boring technical issue flavour.

6.

Action points. In the first instance, could you revise the Press Notice to fit with this.

…

NB We agreed with Financial Secretary that the Press Office lead should come from the Revenue, not the Treasury, to reinforce the message that we were looking at a boring technical issue, not a political one …

7.

We will need some careful Press Office briefing – because the handling line is, self-evidently, going to be difficult. And, of course, you and lead players in [International Division] will need to be on hand to deal with follow-up questions.”

On 21 October Mr Hardie responded to Mr Spence’s memorandum, enclosing the current draft of the press release for comments. He said he hoped Mr Spence would find that the current draft was already cast in the form of the outline in paragraph 3(b) of his memorandum, and asked Mr Spence and the copy recipients to let him have any further proposed drafting amendments by 26 October. The current draft enclosed with this minute was the initial draft with some minor amendments. No amendments had been made in the light of Mr Spence’s memorandum, and Mr Hardie evidently hoped that the current draft would be acceptable.

Over the next few days a number of comments on the draft were submitted to Mr Hardie, and on 25 October he circulated a revised draft. In his covering note he said that the main changes were the removal of any overt reference to “discrimination”, and a reference to the mitigation of the “blunderbuss” effect as being a response to representations. Paragraph 1 of the draft press release now read as follows:

“A recent European Court of Justice decision has prompted speculation about the application of the existing rules relating to certain payments between connected companies. This has led to some uncertainty for taxpayers. The Government therefore wishes to amend the legislation in order to clarify the position.”

The corresponding paragraph in the Notes for Editors now read:

“There has been some speculation about the possible relevance for these provisions of a recent [ECJ] decision involving a Netherlands company (Halliburton Services BV). The Government intends the new provision to remove any possible doubt this may have caused. At the same time the opportunity has been taken to respond to representations about the harshness of the current rules in certain circumstances.”

On 1 November the Financial Secretary sent a minute to the Chancellor advocating the same policy stance as he had adopted at his meeting with Mr Spence and Mr Michael in October. He referred to the ever-increasing risks of challenge to the UK direct tax system in the ECJ, and strongly recommended that the issue should be made a priority for the UK agenda for the 1996 Inter Governmental Conference “since Treaty revision is the only long-term solution”. Under the heading “Damage limitation” he then said this:

“6.

An IGC change, even if successful, would not take effect until the new Treaty comes into force (probably early 1998). The scale of the damage we will experience until then is largely outside our control since it depends on the number of successful challenges in the Courts and the extent to which ECJ rulings expose our vulnerabilities. However, we must be prepared for the possibility of significant legislative changes and tax loss.

7.

To minimise this, I recommend that we:

- play down the seriousness of the issue (to avoid provoking legal challenges);

- resist cases strongly in the Courts, and through to the ECJ whenever we reasonably can and encourage other countries to do the same;

- when necessary, introduce “Europroofing legislation” which removes the current discrepancies between residents and non-residents.

8.

This strategy should also minimise the political problems of ECJ interference by helping to keep it off the agenda. At present, ECJ vulnerability is a live issue in technical tax circles, but is not a political issue.”

The Financial Secretary went on to say that, in line with the strategy described above, he suggested that the Euro-proofing legislation on thin capitalisation in that year’s Finance Bill “should be presented as a minor technical issue”, and any questions about wider implications “should be given similarly low-key treatment”.

On 2 November the Chancellor minuted his agreement with the damage limitation strategy proposed by the Financial Secretary, and also agreed that the Euro-proofing legislation should be presented “purely as a minor technical issue”.

On 7 November Mr Spence sent a further draft of the press notice to the Financial Secretary for approval, and copied it to a long list of recipients. In his covering note he said:

“As agreed, this gives very low key treatment to the ECJ angle. We have buried this as well as we can in the thickets of technical explanation about the changes (in fact relatively minor) that are being made to the rules. The technical commentators will, of course, register the true significance of the ECJ point. But with a bit of luck (and some suitably low-key briefing) this should postpone the point at which the issue goes live politically.”

The revised draft formed the basis of the budget press release which was eventually issued on Budget Day, 29 November 1994. In the part of the press release headed “Details”, the Chancellor’s proposals were presented as follows:

“1.

At present payments of interest and similar sums to certain non-resident connected companies are, for the purposes of United Kingdom domestic law, treated in the same way as distributions. This means that the paying company is not entitled to deduct the payment in calculating its profits for tax purposes and is obliged to make a payment of [ACT] at, or around, the time of the payment. These rules have enabled the Inland Revenue to challenge “thin capitalisation”. “Thin capitalisation” is the process of financing subsidiaries with greater amounts of debt in comparison with equity than would be normal in an arm’s length funding arrangement. Such a funding structure may be intended to reduce or eliminate the company’s liability to corporation tax by loading it with “excessive” interest payments.

2.

This domestic treatment is modified by the terms of agreements between the governments of the United Kingdom and certain overseas countries for the elimination of double taxation. The precise wording of such agreements varies but, for the majority of companies, the effect is to limit the amount of a payment to be treated as if it were a distribution to only the part which exceeds what would have been paid between unconnected companies.

3.

The proposed measures will amend the rules for identifying the connected companies to which the legislation applies. They will deal with relevant payments between such companies on a consistent basis no matter where in the world the companies are resident and will follow the approach of most of the United Kingdom’s recent double taxation agreements by only treating the “excessive” part of such payments as if they were distributions.

4.

The new provisions will leave most companies in the same position as at present – after taking account of the interaction of current domestic law and the relevant existing double taxation agreement. Where there is no such agreement or where it varies from the current norm, a minority will find their position altered. To assist companies (and their advisers) in establishing whether and how the new measures will affect them, a draft of the Finance Bill clause has been issued today. A copy is annexed to this Press Release.

5.

The new measures respond to representations about the harshness of the current rules in a few cases. The Government has also taken the opportunity to remove any uncertainty for taxpayers which recent speculation about the application of the existing rules following a European Court of Justice decision may have caused.

6.

The Chancellor intends the new provisions to apply generally to payments made on or after today.

…

7.

The Government intends to keep this legislation – and its effectiveness in achieving its purpose – under review.”

The Notes for Editors included the following paragraph:

“5.

As a further point, there has been some speculation about the possible relevance for the existing provisions of a recent European Court of Justice decision involving a Netherlands company (Halliburton Services BV). The Government intends the new provision to remove any possible doubt this may have caused.”

Mr Black was given the responsibility of dealing with press enquiries, together with Mr Brooks and Ms McDonald. On 30 November he circulated within International Division a short briefing paper which had been given to the press office. In his covering letter he said:

“Hopefully this will give you a rough idea of what we are doing and how we are handling questions. The general line we want to give is that this measure will not alter the position for most companies.”

In cross-examination Mr Black said that the briefing note would have been prepared by someone in Mr Hardie’s team of four, and although he could not recall he assumed that they would all have been closely involved with it. One of the questions in the briefing paper was:

“What are the reasons for the “uncertainty” to which the press release refers?”

The answer to that question was this:

“Some people have suggested that our existing approach is discriminatory under the Treaty of Rome. Although we do not agree, we have taken the chance in this legislation to remove some of the aspects of the existing provision (particularly the distinction between resident and non-resident holders of securities) which led people to think they might have a case.”

Mr Aaronson pressed Mr Black with this passage, and put it to him that the words “we do not agree” were a downright lie. He put the same point, equally forcefully, to Mr Brooks. While neither of them had any positive recollection of the matter, which is not surprising 15 years after the event, neither of them was able to offer any plausible explanation for those words which did not involve an intention to mislead. I find that the statement “we do not agree” was untrue, to the knowledge of the Revenue and the Treasury, and if those words (or their equivalent) were repeated in response to actual enquiries made by members of the public, they would have been deliberately misled.

A question which arose in December 1994 was whether the remedial legislation should extend to corporate charities and pension funds which were resident in the United Kingdom but were exempt from corporation tax and therefore not within the charge to corporation tax. The Revenue would in principle have liked the legislation to apply to such charities and pension funds, because it would provide two substantial examples of cases where the new rules could apply to UK residents and thus bolster the Revenue’s case if a challenge were mounted on the basis of indirect discrimination. However, it soon became clear that there were strong policy reasons why charities should not be adversely affected, and there was also concern about picking a fight with the pensions industry. In a memorandum dated 7 December 1994 to Mr Hardie, Mr Spence said that “for Euro respectability purposes” they wanted to have some UK residents who were within the scope of the charge, but also wanted to ensure that no UK residents would actually have to pay any tax under the charge. Politically, however, the desire was “to avoid a fuss from charities etc”, in particular “because a fuss might blow the gaff on the subtleties of our Euro-proofing”. In the event, the outcome of this debate was that a specific exception for charities was enacted, but no steps were taken to prevent the potential application of the new rules to UK pension funds.

In January 1995 Mr Hardie’s team prepared drafts of the speaking note, defensive note and background note which were intended to constitute Part II of the notes on clauses for the relevant part of the Finance Bill. The draft defensive notes included the question “Is this legislation being forced on the UK by a European Court decision?”, to which the response was:

“22.

No. Some taxpayers have contended that the present law is discriminatory and have referred to a European Court of Justice decision which dealt with a discrimination issue, but not one involving thin capitalisation. These contentions have been rejected by the Inland Revenue and have not been pursued in the Courts. As the Press Release accompanying announcement of the new clause explained, we simply decided to use the opportunity presented by the change in law to remove any uncertainty which may remain in the minds of taxpayers.”

This response was in my judgment at best disingenuous, and at worst positively misleading. Even though the ECJ decision in Halliburton did not literally force the UK to act, the truth was that the perceived implications of that decision were the driving force behind the remedial legislation. The true purpose of the exercise was to protect tax revenues, and to “Euro-proof” legislation which the Revenue had been advised, and themselves believed, almost certainly infringed Community law. The suggestion that advantage was being taken of a convenient opportunity to remove any lingering uncertainty in the minds of taxpayers was, on any view, a travesty of the true position. The fact that this misleading emphasis was not accidental is brought out by paragraph 36 in the draft background note which was also enclosed with Mr Hardie’s letter of 10 January 1995. Paragraph 36 was placed in square brackets, and when a revised version of the Notes on Clauses was sent to the Financial Secretary on 20 February in preparation for the Committee Stage debate on the following day it was sidelined “not for use in debate”. Paragraph 36 said:

“Although it has not been conceded that the recent ECJ decision causes any problems for the UK, legal advice received on the subject figured prominently in the decision to amend the present law. It was decided that this aspect should be played down and the measure portrayed as a boring technical one.”

On 24 March 1995 the Financial Secretary asked to be briefed on what the financial consequences of amendments taken during the Committee Stage would have been, on the alternative hypotheses that they had and had not been passed. Mr Hardie responded on 28 March, saying it was difficult to answer the question because the narrow financial consequences were quite different from the potential financial consequences which might have flowed from not tabling the Government amendment to relieve charities. He explained this point in the following terms:

“The amendments to Clause 75 were designed to exclude financial arrangements between charities and their trading subsidiaries from its scope. At most a few million pounds, which it had never been intended to catch, was being “given up” as a result of the amendment.

The wider consequences of not announcing and introducing this amendment would, however, have been considerable and, potentially, financially costly.

Following three scare stories in national newspapers, the Charities Lobby was beginning to build up a head of steam at the time when the announcement of the intention to amend the legislation was made …

What can be said with greater certainty is that the continued focus on the inclusion of UK residents within the scope of the clause would, inevitably, have risked drawing attention to its “Euro-proofing” aspects in a way which Sir George had decided he wished to avoid. An undue focus on the European dimension might, in itself, have added to the difficulty of getting the clause past Standing Committee. It would also without doubt have alerted many taxpayers and their advisers to the possibility of contending that the legislation being replaced was discriminatory and therefore could not be applied. If such arguments had gained wide currency and been accepted by the Courts we estimated that approximately £25 million tax would be lost.

If, in addition to alerting taxpayers and losing before the Courts the new clause had not passed through Parliament at all, we estimated that a figure well in excess of £150 million would be lost in a year – before taking account of the likely behavioural effects which would have led to further exploitation of the resultant weakness in our defences.”

At this point the documentary record relating to the 1995 amendments comes to an end, and the date of the next document disclosed is 4 March 1997, some two years later. The documents from 1997 are relatively few in number, and relate to the proposed introduction of the 1998 transfer pricing rules. They show that the problem of indirect discrimination was still being debated within International Division, and that the decision of the ECJ in the case of Wielockx (Case C-80/94, GHEJ Wielockx v Inspecteur der Directe Belastingen, [1995] ECR I-2493, [1995] STC 876) was thought to pose problems for a defence of “objective justification”.

There is then an even longer gap in the Revenue’s disclosure, from July 1997 until February 2001, when a draft strategy paper on international direct tax was circulated by Revenue Policy, International (as International Division had by then become).

On 21 December 2001 a meeting took place between representatives of the Chartered Institute of Taxation and the Revenue to discuss UK tax legislation and discrimination under the EU Treaty. One of the issues discussed was the Bachmann defence, which now seemed to be confined to situations where the same taxpayer suffered a tax advantage and the related tax disadvantage. In response to that point, one of the Revenue representatives, Mr Addison, said that there were other defences that were available. He believed that defences could be mounted based on the general shape of the tax system, and that something new might come out of the pending case of Lankhorst-Hohorst. The categories of justification were not closed; there was also a defence based upon the necessity for effective fiscal supervision. The meeting was informed that one of Mr Addison’s jobs was to scrutinise Finance Bill proposals for their EC vulnerabilities so that those issues were “fed into the development of policy at an appropriately early stage”. Mr Addison said that there had in fact been a review of the vulnerability of UK legislation in the 1980s, although some of the provisions that had been thought to be vulnerable had since been removed.

The opinion of Advocate General Mischo in Lankhorst-Hohorst was delivered on 26 September 2002. The implications were discussed within the Revenue, and on 29 October a report was prepared for the Paymaster General dealing with recent decisions of the ECJ in the business tax area, and their implications for the UK. The report identified a worrying, and accelerating, trend: the ECJ had struck down a number of provisions in the tax systems of member states for non-compliance with fundamental freedoms, and of the 17 cases which it had heard it had decided all but two against the member state concerned. Furthermore, 11 of the 17 cases had been decided in the last four years, compared with only six in the period between 1986 and 1998. If the opinions of the Advocates General in Lankhorst-Hohorst and another recent case, Bosal, were confirmed by the ECJ, the tally of adverse decisions would rise to 17. The report identified the next steps as being to establish the views of other member states; to draw up a list of UK measures that were potentially vulnerable, with a possible menu of options for dealing with them; and to consider possible options for handling the wider issues surrounding the role of the ECJ in the direct tax area.

On 12 December 2002 the ECJ delivered judgment in Lankhorst-Hohorst, and it was widely circulated within the Revenue and the Treasury. In a covering note, Mr Addison said there was “no disguising the fact that this is very bad news”. The only solution he could think of was to apply the same rules to UK parents and subsidiaries. With his usual insight, Mr Durrans commented:

“On a quick read the Court seems to me to be leaving the door open, albeit ever so slightly, for thin cap regimes to be capable of being objectively justifiable where they are closely targeted on abusive (wholly artificial?) arrangements. Whether a Member State could ever get such a closely targeted regime through the eye of the objective justification needle on the facts of a particular case is, of course, another matter.”

On 20 December 2002 a policy update was sent to the Paymaster General by Revenue Policy, International. It recognised the strong likelihood that the UK’s thin cap rules would be caught in the same way as Germany’s, in which case the choice was between scrapping the UK’s rules or extending them so that they also applied to UK-owned companies. The first of those options was unacceptable, so urgent consideration would be given to how best to implement the second option, with a view to reporting in good time for a Budget 2003 announcement. The author of the paper, Mr Mike Williams, commented:

“There isn’t really any intrinsic merit in this (that is, apart from its enabling us to ECJ-proof the application of the thin capitalisation rules to foreign-owned companies), so we should have to minimise the additional compliance burden on domestically-owned companies while ensuring we did enough to satisfy the ECJ.”

In a later section of the paper, Mr Williams expressed the view that the Revenue’s long standing “wait and see” approach to potential ECJ decisions “looks as though it’s still a sensible and prudent one”. He gave two reasons. First, it is difficult to act without knowing the precise terms in which the ECJ reaches an adverse judgment. There was always a danger that acting in advance of an ECJ decision might lead the UK to do something that was unnecessary, or might not go far enough, or might inadvertently focus on the wrong area for change. Secondly, as the discussion of the specific issues in the paper illustrated, “the choices for action are in any event generally pretty unpalatable”. He concluded in paragraph 25:

“All that said, we clearly need to guard against complacency here. So we shall keep the position under active review. We are also urgently exploring the scope for reducing the Exchequer cost of any ECJ decisions that go against us.”

The matter continued to be debated within the Revenue, and further meetings took place with the Chartered Institute of Taxation. Eventually, instructions were sent to counsel (Dr Richard Plender QC and Mr Ewart) to advise in consultation on 10 March 2003. One of the questions on which counsel were asked to advise was the prospects of successfully defending the UK thin cap provisions in the light of Lankhorst-Hohorst. According to a note of the consultation, Dr Plender’s advice was “very pessimistic about the implications of Lankhorst for UK thin cap. He was unable to distinguish the UK thin cap rules from the German rules that had been considered by the ECJ, and the only crumb of comfort was that “legislation targeted on artificial arrangements could work””. It appears from subsequent documents in the bundle that counsel estimated the chances of successfully defending the thin cap rules at around 10%.

Discussion

I must now evaluate this material, together with all the other relevant factors, and reach a conclusion on the question whether the UK’s breach of Community law was at any stage sufficiently serious to ground liability in damages. There is no suggestion by the claimants that any sufficiently serious breach occurred before the decision of the ECJ in Halliburton, so I will begin by considering the position in the aftermath of that case and up to the enactment of the 1995 changes to the thin cap regime.

It is ironical that, although Halliburton was clearly perceived by the Revenue at the time to be a very significant decision, it has not in fact played a leading role in the subsequent jurisprudence of the Court, and is probably little known even to specialists. However, there can be no doubt that it alerted the Revenue to the potential vulnerability of the UK thin cap provisions, and it focused attention in particular on two points: first, the width and potency of the principle of indirect (or covert) discrimination; and secondly, the likelihood that a defence based on the fact that the taxpayer company was UK-resident would not succeed. These points, taken in conjunction with the general understanding at the time that provisions in DTCs had to be left out of account, fully justified Mr Durrans’ pessimistic advice that the chances of successfully resisting a challenge to the existing thin cap rules were “considerably less than slim”. The threat to the Exchequer was a real and pressing one, and the urgent need for remedial legislation was rightly appreciated by the Government.

It is crucial at this stage, in my judgment, to draw a distinction between the content of the amending legislation which was enacted in 1995, and the way in which it was presented to the public. With regard to the content of the legislation, I am satisfied that the Government’s intention throughout was to introduce changes which would render the thin cap regime immune to challenge under Community law, or in other words to “Euro-proof” it. I can see no reason to doubt that this was the genuine intention of all concerned, and it was confirmed to be the overriding objective by both Mr Brooks and Mr Black in their oral evidence. Indeed, it would have been a strangely pointless exercise to introduce amending legislation that was deliberately designed to fall short of achieving consistency with Community law, and which would have left tax revenues of £150 million or more at risk.

The Revenue and the Treasury were, however, equally concerned not to bring intra-UK arrangements within the scope of the new regime, for the simple reason that such arrangements do not normally give rise to the mischief that thin cap rules are designed to prevent. The problem of thin capitalisation is inherently a cross-border one, and the Government was naturally unwilling to inflict the burden of compliance with a thin cap regime on UK companies which were financed by their UK parent, and where both of them were subject to UK corporate taxation. In my view the desire of the Government to keep intra-UK arrangements out of the thin cap net cannot possibly be characterised as unreasonable from a Community law perspective, given the acceptance by the ECJ in the present case of the basic mischief which thin cap rules exist to counter (paragraphs 76 and 77 of the judgment). See too the trenchant comments of the Advocate General in paragraph 68 of his opinion:

“Nor am I of the view that, in order to conform with Article 43 EC, Member States should necessarily be obliged to extend thin cap legislation to purely domestic situations where no possible risk of abuse exists. I find it extremely regrettable that the lack of clarity as to the scope of the Article 43 EC justification on abuse grounds has led to a situation where Member States, unclear of the extent to which they may enact prima facie “discriminatory” anti-abuse laws, have felt obliged to “play safe” by extending the scope of their rules to purely domestic situations where no possible risk of abuse exists. Such an extension of legislation to situations falling wholly outwith its rationale, for purely formalistic ends and causing considerable extra administrative burden for domestic companies and tax authorities, is quite pointless and indeed counterproductive for economic efficiency. As such, it is anathema to the internal market.”

Reconciliation of these twin objectives – Euro-proofing of the legislation, and the exclusion of intra-UK arrangements - was always going to be a difficult exercise, because the exclusion would of course run the risk of re-introducing indirect discrimination on the ground of nationality, and it would then be necessary to rely on one or more of the relatively ill-defined and developing justifications available under Community law if infringement of Article 43 was to be avoided. This dilemma was clearly recognised at the time, and the remedial legislation was drafted in a way that maximised the chances of successfully resisting a challenge. Nor was this just wishful thinking. Advice was sought from the Solicitor’s Office, and on 11 August 1994 Mr Durrans advised that there would be a “reasonably good prospect” of persuading the ECJ that the discriminatory aspect of the amended provisions was not objectionable. There would be a residual risk of successful challenge, and views no doubt varied about the precise degree of that risk. But given the sheer pointlessness of applying thin cap rules in a purely domestic context (Mr Spence went further, and described such a result as “nonsensical” or “preposterous”), and given the fact that Germany also had a similar system already in operation, on which the new UK legislation was explicitly modelled, it was in my judgment both reasonable and excusable for the Government to go down this route. Short of making the new legislation apply to intra-UK arrangements, which according to the Advocate General in the present case would have been “anathema to the internal market”, it is hard to see what more could have been done at this stage to minimise the risk of infringing Community law.

I will deal here with an argument which Mr Aaronson made much of in his oral submissions, and which featured prominently in his cross-examination of Mr Black and Mr Brooks. The point was that the 1995 legislative changes were essentially cosmetic and insubstantial, a sticking-plaster applied to a gaping wound, and viewed objectively they left the UK in just as vulnerable a position as before, or at best effected only a marginal improvement in the situation. It is true that from one point of view the changes were relatively minor. The introduction of a statutory arm’s length test reproduced in substance (although with a great deal of additional detail) the position that already existed in practice in most cases, where the general rule in section 209(2)(e)(iv) was overridden by DTCs with an interest article in one of the two standard forms. Similarly, the confinement of the disallowance to interest above an arm’s length threshold reproduced the position under most DTCs, and as the Revenue witnesses made clear this was anyway the general practice that was followed in thin cap investigations. Finally, the exclusion for cases in which both companies were within the charge to corporation tax was in practice likely to have no application to the cross-border financing of UK subsidiaries, as both Mr Black and Mr Brooks readily acknowledged.

There is considerable force in these points, but I am not persuaded that the 1995 legislation can be dismissed as a cynical piece of window-dressing. It is essential to remember at this point that the effect of DTCs was generally thought at the time to be something the ECJ would refuse to take into account in considering an Article 43 challenge. On that footing, it was imperative to remove the “blunderbuss” rule in section 209(2)(e)(iv) and to replace it with a statutory arm’s length test. It was equally necessary to remove the potential for “overkill” by confining any disallowance of interest and recharacterisation as a distribution to interest which exceeded the arm’s length threshold. Quite apart from that, the enactment of a detailed arm’s length test was in itself a substantial undertaking, even if in general terms the intention was to give statutory force to existing practice. The new rules were unfortunately drafted in a style of almost impenetrable complexity, which made them exceedingly difficult to understand unless one knew in advance what they were meant to achieve. Justifiable complaints were made at the time on that score, but they were for the most part complaints about the drafting of the legislation, not that it lacked substance.

The one area where the 1995 legislation may, at least with hindsight, be criticised as purely cosmetic is in its use of the criterion of being within the charge to corporation tax, instead of the simple criterion of UK residence, in order to achieve the practical objective of excluding most, if not all, domestic arrangements from the scope of the thin cap rules. Even at the time, it must have been fairly obvious that a device of this nature was unlikely to circumvent the problem of indirect discrimination, especially in view of the clear and recent statements of principle by the ECJ in Commerzbank and Halliburton. Nevertheless, I do not think that the Government can sensibly be criticised for doing all they could to Euro-proof the new legislation, even if this particular expedient now appears misguided.

I am also satisfied that the references in the documents to “sticking plaster” solutions, and similar phrases, were only meant to refer to the fact that this was urgent remedial legislation, as compared with the much more fundamental review of interest generally, and the whole corporate tax regime, which was then in progress. Furthermore, even in the narrower context of thin capitalisation itself, it was recognised that the 1995 legislation was only the first stage in dealing with the problem, and that further work still needed to be done both in relation to the Community law aspect of the matter and in relation to perceived abusive practices (such as routing inward investment into the UK through a third party, such as a bank, financed with back to back loans) which the existing legislation would not catch.

I now turn to the way in which the 1995 legislation was presented to the public. The documentary record, still incomplete and disclosed in the belated and piecemeal fashion which I have described, does not present an attractive picture. Under political pressure from above, the Revenue sought to present the changes as boring technical measures of a tidying-up nature, rather than as the urgent response to the ECJ judgment in Halliburton that they actually were. On at least two occasions, in my judgment, explanatory and publicity material prepared by the Revenue clearly crossed the border line, blurred though it may sometimes be, between putting the best possible complexion on the matter and being actively and deliberately misleading: see in particular paragraphs 297 and 300 above. It was not suggested, and I do not find, that this was done in bad faith. The question of how to present the legislative changes was undoubtedly a difficult and sensitive one, and nobody could reasonably expect the Revenue to have advertised the vulnerability of the existing thin cap rules when so much tax was potentially at stake, and no specific Community law challenge to the rules had yet been mounted, let alone ruled upon by the ECJ. Nor should one underestimate the difficulties, under a Conservative administration, of securing the passage through Parliament of legislation that was designed to ensure compliance with Community law in the field of direct taxation. It would no doubt be easier to achieve that laudable aim if the amendments were disguised as being minor technical ones, prompted in part by representations made by taxpayers, which also had the incidental advantage of helping to dispel probably groundless concerns about the impact of Community law which had been expressed in some quarters.

It is also pertinent to bear in mind, as Mr Ewart submitted, that the multinational groups which might be expected to challenge the UK thin cap rules were well able to investigate the question for themselves. They had access to, and could obviously afford, the services of accountants, tax lawyers and the tax planning industry. All the material that was needed to form a judgment on the issue was in the public domain: the relevant Treaty provisions, the UK legislation and DTCs, and the decisions of the ECJ. They had every motive to mount a challenge to the legislation, either before or after its amendment, and it is fanciful to suppose that they would have been seriously influenced by what the Revenue or Treasury ministers did or did not say about the purpose of the amendments.

It is a striking fact that no evidence was given on behalf of any of the claimants that a challenge to the UK thin cap rules was contemplated before the judgment of the ECJ in Lankhorst-Hohorst. The witnesses for each test claimant were cross-examined by Mr Baldry for the Revenue on this point, and the picture which emerged was uniform. Before Lankhorst-Hohorst, everybody proceeded on the assumption that the UK rules were valid and compatible with Community law. So, for example, Kathleen Bishop, the tax director for Europe, the Middle East and Africa for the IBM group, said in paragraph 35 of her first witness statement:

“Before hearing about the Lankhorst-Hohorst case, I had no idea that the UK’s thin cap rules might be unlawful and certainly it did not cross my mind to advise the business that they could ignore those rules. I do not recall any conversations with any other IBM employee prior to December 2002 in which the legality of the thin cap rules was questioned.”

To similar effect, Mr Gibson, the chief executive of Standard Bank Offshore Group Ltd, said in paragraph 27 of his statement:

“I cannot remember exactly when I learnt of the possibility that the thin capitalisation rules might be unlawful under European law, but I am sure that it was not until late 2003.”

In these circumstances, although the public presentation of the amendments in my judgment fell significantly short of the high standards of probity that are rightly expected of a Government department, I do not consider that this lapse can turn what would otherwise be an excusable breach into an inexcusable one, although I accept that it is a relevant matter for me to take into account. The critical points, in my judgment, are that the intention of the Government was at all material times to secure compliance with Community law, and the fact that the error of law made by the Government was objectively an excusable one.

I am fortified in reaching this conclusion by a number of other considerations. This is not a case, like Factortame (No. 5), where the Government deliberately enacted provisions which directly discriminated against individuals on grounds of nationality or residence, and chose to run the substantial risk of non-compliance with Community law. Thin cap legislation is a legitimate part of any state’s armoury against tax avoidance, and the method adopted by the UK to deal with it (through the interest article in DTCs) was based on OECD models, as Mr Brooks confirmed in his evidence. No complaint about the UK approach or legislation had been made by the Commission or any other Community organ, nor had any case on the subject come before the ECJ. Against this background, it was in my view perfectly proper for the Government to take immediate steps to Euro-proof the legislation, to continue with the wider general review of interest and corporate taxation, and to await further developments in the case law of the ECJ.

I now move on to consider the period between 1995 and Lankhorst-Hohorst. Mr Aaronson’s submissions in relation to this period had both a positive and a negative aspect. The positive aspect was to analyse the development in the ECJ’s treatment of the Bachmann (fiscal cohesion) defence, with a view to showing that it quickly became clear that it could not be relied upon to justify the UK thin cap regime. The negative aspect was to emphasise the conspicuous gaps in the documentary record to which I have already drawn attention, from which Mr Aaronson asked me to infer that, although the question of compliance of the thin cap regime with Community law should have been kept under regular review, it was not.

I begin with the fiscal cohesion defence. As the claimants say in their skeleton argument on sufficiently serious breach, it is easy to understand why in 1994 Mr Durrans thought there was a reasonable chance of defending the proposed new rules: the cohesion of the tax system would require the tax deduction at the level of the subsidiary to be matched by taxation at the level of the parent in order to avoid extraction of profits in the form of interest. Mr Aaronson began his review of the subsequent case law by referring to the case of Schumacker (Case C-279/93, Finanzamt Köln-Altstadt v Roland Schumacker, [1995] ECR I-225, [1995] STC 306). This was a decision of the Grand Chamber of the ECJ, delivered on 14 February 1995. It will be remembered that Mr Michael had said it was likely to become a leading case in his memorandum to the Chancellor of 26 July 1994, some months before the Advocate General (Léger) gave his opinion on 22 November 1994.

In the event, however, it seems to me that Schumacker cast little fresh light on the fiscal cohesion defence, at any rate in the present context. One reason for this is that it dealt with personal, not corporate, taxation. The taxpayer was a Belgian resident who worked and earned all his money in Germany, and his complaint was that under German law he was taxed on his German income only, with reliefs that were less generous than those available to a German resident who was taxable on his worldwide income. The member states which submitted observations to the Court, including the UK, submitted that this less favourable treatment was justified by the need for consistent application of tax regimes to non-residents, and that there was a link of the type upheld in Bachmann between the grant of reliefs which took account of personal and family circumstances and the right to tax worldwide income. In rejecting this argument, the Court said in paragraph 41 of its judgment:

“In a situation such as that in the main proceedings, the State of residence cannot take account of the taxpayer’s personal and family circumstances because the tax payable there is insufficient to enable it to do so. Where that is the case, the Community principle of equal treatment requires that, in the State of employment, the personal and family circumstances of a foreign non-resident be taken into account in the same way as those of resident nationals and that the same tax benefits should be granted to him.”

It can be seen that this reasoning was dependent on the fact that the taxpayer had insufficient income in Belgium for the reliefs available in his state of residence to be of any use to him, and for all practical purposes he was in the same position as a German resident taxable on his worldwide income. In those exceptional circumstances, the Community principle of equal treatment required him to be treated in the same way as a German resident. If anything, the implication is that the defence might well have succeeded if the taxpayer had been in receipt of enough income in Belgium to make use of his Belgian reliefs.

The next case relied upon by Mr Aaronson was Wielockx, in which the Court delivered judgment on 31 May 1995. This was another case involving a Belgian resident taxpayer who worked and received his entire income in another member state, this time the Netherlands. He complained about the refusal of the Dutch tax authorities to allow him to deduct pension contributions from his taxable income. The Court held that he had suffered discrimination in comparison with a Dutch resident taxpayer, who would have been entitled to such a deduction, and then considered the defence of fiscal cohesion. The Dutch government argued that if the taxpayer were entitled to a deduction for his pension reserve in the Netherlands and thus secured a right to a pension, that pension would not be taxed in the Netherlands by virtue of the DTC between the two states, but would be taxed in the state of residence. Accordingly, so the argument ran, the deduction was properly disallowed in order to preserve the cohesion of the tax system. This argument was rejected by the ECJ:

“24.

As the Advocate General observed in point 54 of his Opinion, the effect of double-taxation conventions which, like the one referred to above, follow the OECD model is that the State taxes all pensions received by residents in its territory, whatever the State in which the contributions were paid, but, conversely, waives the right to tax pensions received abroad even if they derive from contributions paid in its territory which it treated as deductible. Fiscal cohesion has not therefore been established in relation to one and the same person by a strict correlation between the deductibility of contributions and the taxation of pensions but is shifted to another level, that of the reciprocity of the rules applicable in the Contracting States.

25.

Since fiscal cohesion is secured by a bilateral convention concluded with another Member State, that principle may not be invoked to justify the refusal of a deduction such as that in issue.”

Mr Aaronson relied on this passage as establishing the proposition that the defence of fiscal cohesion is available only in very narrow circumstances, where it is established at the level of “one and the same person”. Mr Ewart submitted that this was, at least arguably, too restrictive a reading of the passage, and that the true test was the need for a direct link between the deduction and the charge to tax, which would often, but not necessarily, be at the level of the same taxpayer. He also submitted that the decision turned, to a considerable extent, on the point that fiscal cohesion had anyway been secured by the relevant DTC and left no room for application of the principle in a way that would undermine the agreed allocation of taxing rights under the DTC. I agree with Mr Ewart on both points, and do not accept that Wielockx should have been recognised by the Revenue as making any fiscal cohesion defence of the thin cap regime obviously unsustainable.

Mr Aaronson referred to two further cases decided in 1995 and 1996, but I can say at once that in my judgment they take the matter no further, because in each the Court dealt with the fiscal cohesion defence very briefly and referred only to the need for “a direct link” between the relevant deduction and the charge to tax. Nothing was said about the alleged need for the link to exist at the level of one and the same taxpayer. The cases were: Case C-484/93, Svensson and Gustavsson v Ministre du Logement et de l’Urbanisme, [1995] ECR I-3955, decided on 14 November 1995 (see paragraphs 16 to 18 of the judgment) and Case C-107/94, P. H. Asscher v Staatssecretaris van Financien, [1996] ECR I-3089, decided on 27 June 1996 (see paragraphs 55 to 59 of the judgment).

On behalf of the Revenue, Mr Ewart referred me to two cases decided not long before Lankhorst-Hohorst which show that in 2001 and 2002 the ECJ was still applying the “direct link” test, and had not further developed it. The first case was the well known decision in joined cases C-397/98 and C-410/98 Metallgesellschaft Ltd and others and Hoechst AG and Hoechst (UK) Ltd v Commissioners of Inland Revenue and HM Attorney General, [2001] ECR I-1727, [2001] Ch 620. The relevant part of the judgment is at paragraphs 69 to 73, where the Court rejected a fiscal cohesion defence on the ground that there was no relevant direct link between the UK’s refusal to exempt UK subsidiaries of non-resident parent companies from payment of ACT under a group income election, on the one hand, and the fact that a parent company in another member state which receives dividends from a UK subsidiary was not liable to corporation tax in the UK, on the other hand. As Mr Ewart pointed out, it would have been the shortest of answers to this defence to say that the contrast did not involve the same taxpayer, if that was indeed a requirement for the defence to apply. However, that is not the way in which the Court approached the matter, and it relied instead on the absence of a direct link. The second case, decided on 3 October 2002, was case C-136/00, Rolf Dieter Danner, [2002] ECR I-8147, where the defence was again rejected on the ground that there was “no direct connection” between the deductibility of insurance contributions which was in issue and the taxation of sums payable by insurers. The judgment is also of interest for its explanation, and endorsement, of the second ground relied upon in Wielockx, namely that the principle of fiscal cohesion may not be invoked to justify the refusal of a deduction where the point in issue is dealt with by reciprocal rules under a DTC: see paragraphs 40 to 43 of the judgment. Furthermore, the “direct link” test was applied by the ECJ in the present case (see paragraph 68 of the judgment), and continues to be applied by the Court up to the present day: see, for example, Case C-418/07 Société Papillon v Ministère du budget [2009] STC 542 at paragraph 44, and the discussion of the principle by Advocate General Kokott at paragraphs 49 to 67 of her opinion.

In his submissions in reply, Mr Aaronson changed his tack. He accepted that the “direct link” test was the correct one, but submitted that it should have been obvious to the government that the UK could never hope to satisfy it. I do not agree. In the present case the Revenue argued strenuously that there was a sufficient link to satisfy the test, because any increased charge to tax in the UK as a result of application of the thin cap rules would be matched, through the operation of relevant DTC provisions, by a corresponding advantage afforded to the lending company in its state of residence. The argument was rejected by the ECJ on the ground that it had not been established to its satisfaction that this would in fact always be the case: see paragraphs 55, 56 and 69 of the judgment. I do not regard the argument as such an obviously hopeless one that the UK should have conceded, before Lankhorst-Hohorst, that it stood no chance of success.

In any event, as Mr Ewart pointed out, the jurisprudence of the ECJ on justifications for discriminatory treatment was far from static, and by the late 1990s the justification of combating tax avoidance had clearly emerged as a potential separate defence in its own right: see, for example, Case C-264/96 Imperial Chemicals Industries Plc v Colmer [1998] ECR I-4695, [1999] 1 WLR 108, at paragraph 26. This defence was of course at the heart of the UK’s defence in the present case, and if I am right in my conclusion on the first issue it has ultimately failed on a fairly narrow ground.

Looking at the question objectively, as I must, I do not consider that any of the developments in the case law of the ECJ between 1995 and Lankhorst-Hohorst were of such a character as to make it clear that the UK thin cap provisions were indefensible. In all essentials, I consider that the position remained unchanged from what it had been in 1995. In these circumstances, the question of whether, and to what extent, the Revenue kept the question under review during the period is in my judgment of only marginal relevance. The gaps in the documentary record suggest, and I am prepared to infer, that any consideration of the question during these years was low level, informal and infrequent. There is no indication that further legal advice was sought from Mr Durrans, or from any other source. The Revenue were running a risk in adopting such a passive approach, but in the absence of any clearly fatal developments in the case law I am satisfied that the “wait and see” policy decided upon in 1995 continued to be an objectively reasonable one, at least until the ECJ ruled on a thin cap case. I therefore see no reason to conclude that the breach of Community law, which I have found not to be sufficiently serious in 1995, changed its character and became sufficiently serious at any time between 1995 and the delivery of the judgment in Lankhorst-Hohorst on 12 December 2002.

The position following delivery of the judgment, however, is a different matter. The ECJ had now ruled on the German thin cap provisions which were used as a model for the 1995 amendments, and had held that they breached Article 43. Furthermore, defences of fiscal cohesion and the risk of tax avoidance had been considered and rejected. The writing was now clearly on the wall so far as the UK thin cap provisions were concerned, and although the fiscal cohesion and tax avoidance defences remained arguable, they offered only a slender prospect of success. Leading counsel advised the Revenue in March 2003 that the German provisions could not be relevantly distinguished from the UK regime, and apparently estimated the chances of successfully resisting a challenge to them in the ECJ at around 10%. In my judgment that was an accurate assessment.

The decision of the Court in Lankhorst-Hohorst largely followed the advice and reasoning of the Advocate General, who had delivered his opinion on 26 September 2002, so the judgment did not come as a bolt from the blue. On the contrary, it had been expected, and the mood within International Division was pessimistic. Against that background, it seems to me that the UK’s breach of Article 43 became, objectively, a serious one as soon as the ECJ had delivered its judgment, and that liability to the claimants in damages for maintaining the UK provisions in force should run from 12 December 2002.

I have considered whether the commencement of the period of liability should be somewhat later, to allow time for the Revenue to have taken advice about the judgment and made a public announcement of its future policy. However, I do not consider an enquiry of that nature to be appropriate. The question is rather when the breach, viewed objectively, became sufficiently serious, taking account of “all the factors which characterise the situation”. At least in the context of the present case, and the passive policy which the UK had adopted since 1995, it seems to me that the critical turning point was the ruling in Lankhorst-Hohorst, and that it would be potentially unfair to the claimants to prevent them from recovering damages from the moment when the law was clarified.

For these reasons, I conclude that no sufficiently serious breach is established before 12 December 2002, but that damages are in principle recoverable by the claimants from that date.

Introduction

The issues which arise under this heading are the following (which I have adapted slightly from the Summary of Issues):

I will deal with these issues briefly, because (with the exception of issue (2)) I have already covered them in some detail in FII Chancery, and for that reason very little time was spent on them in oral argument. I would add that the issues are also much less important in the present case than they were in FII Chancery, because only a small proportion of the thin cap claims pre-date the issue of the claim form by more than six years.

Section 32(1)(c) of the Limitation Act 1980

I see no reason to doubt that the extended limitation period for mistake in section 32(1)(c) is available both for San Giorgio claims and (if they exist, which in my judgment they do not) for any claims which are not San Giorgio claims but which are properly to be classified as restitutionary claims under English law, provided in each case that the claim can be framed in English law as a mistake-based restitutionary claim in accordance with the decisions of the House of Lords in Kleinwort Benson and DMG. For the reasons which I have already given in FII Chancery and the earlier ACT GLO case of Europcar (Europcar UK Ltd v Revenue and Customs Commissioners [2008] EWHC 1363 (Ch), [2008] STC 2751), I do not consider that the extended limitation period is available for Woolwich-based restitutionary claims or for Factortame damages claims, because in those cases the mistake (assuming that there was one) is not an essential ingredient of the cause of action: see FII Chancery at paragraphs 243-4 and 405. I should note, however, that, as in both Europcar and FII Chancery, the claimants (represented in all three cases by substantially the same legal team) have reserved the right to argue in a higher court that section 32(1)(c) can also apply in cases where mistake is not an essential ingredient of the cause of action, but an operative mistake was nevertheless made as a result of which the claim was not begun within the standard six year limitation period.

Where section 32(1)(c) applies, from what date does the limitation period start to run?

In the light of my discussion of the question of sufficiently serious breach, I consider that the extended limitation period started to run when the ECJ delivered its judgment in Lankhorst-Hohorst, that is to say on 12 December 2002. Before that date, I do not consider that any mistake made by the claimants about the lawfulness of the thin cap regime could have been discovered with reasonable diligence.

Section 320 FA 2004 and section 107 FA 2007

I have nothing to add to my discussion of this issue in FII Chancery at paragraphs 406 to 431. If my earlier decision is correct, it is, I think, common ground in the present case that the two sections breach the principle of effectiveness and must be disapplied in relation to San Giorgio claims which are based on mistake. If, however, the Revenue were right in their submissions (which I have of course rejected) that the Woolwich cause of action alone is enough to provide the claimants with an effective remedy for the UK’s breach of their Article 43 rights, and that the principle of effectiveness does not require the availability of any mistake-based cause of action, then it would follow that sections 320 and 107 can operate in accordance with their terms, and the application of section 32(1)(c) would be curtailed accordingly.

Section 33 TMA 1970

Here too I have nothing of substance to add to my discussion of the issue in FII Chancery at paragraphs 432 to 439, although I would also refer to my judgment in Chalke at paragraphs 64 to 66 for a slightly fuller account of the important judgment of the Court of Appeal in Monro, which is now also reported at [2009] Ch 69. In summary, I consider that section 33 (and its corporate equivalent in paragraph 51 of schedule 18 to the Finance Act 1998), which anyway apply only to tax paid under an assessment (including a self-assessment) by reason of some error or mistake in a return, must yield to the Community principle of effectiveness where the claimant has a San Giorgio claim. The same questions therefore arise as under the previous heading as to whether mistake-based claims in English law are needed in order to give effective redress for San Giorgio claims. In my judgment they are, but if I am wrong in that conclusion, I would hold that paragraph 51 of schedule 18 FA 1998 ousts all other domestic remedies for overpaid corporation tax within its scope.

Introduction

I come finally to an esoteric claim advanced by the IBM test claimants which is not based directly upon Community law, but instead relies on the non-discrimination article in the DTC between the USA and the UK. The relevant DTC is the Convention of 31 December 1975, SI 1980 No. 568 (“the US Treaty”).

The factual background to the claim can be briefly stated. The claim relates to the period before the group-wide restructuring which took place on 16 July 2001, when the holding company of IBM’s UK subsidiaries was UK Holdco. In the early 1990s UK Holdco received loans from the group’s Dutch finance subsidiary, IIF. Both UK Holdco and IIF were owned by the ultimate US parent company, IBM Corp, through separate chains of intermediate companies. The contention is that, as a result of the UK thin cap rules, the immediate parent of UK Holdco, EMEA, made three substantial contributions of capital to UK Holdco between December 1992 and December 1993, totalling £550 million, which it would otherwise have made in the form of loans. It is not alleged that any interest paid by UK Holdco to IIF was in fact disallowed and recharacterised as a distribution. The argument is, rather, that following negotiations with the Revenue it was decided to make these capital contributions instead.

Paragraph 24 of IBM’s seventh amended Particulars of Claim pleads that the thin cap provisions and their application by the Revenue “is other or more burdensome than the taxation and connected requirements to which other similar enterprises of the United Kingdom are or may be subjected”, contrary to Article 24(5) of the US Treaty and ICTA section 788(3). The relief claimed includes a declaration that the thin cap provisions, in so far as they restrict the ability of a UK-resident company, the capital of which is wholly or partly owned or controlled, directly or indirectly, by one or more US residents to deduct interest on loan finance by reason of that ownership or control, are contrary to the same provisions and are therefore ineffective. There is also a claim for compensation or damages for loss caused by the Revenue’s alleged failure to comply with the provisions.

I should add for completeness that the pleaded claim includes a similar claim based on the non-discrimination article in the DTC between the UK and the Netherlands, but this contention was not pursued in argument.

The relevant provisions of the US Treaty

Article 24 of the US Treaty is headed “Non-discrimination”, and paragraph (5) provides as follows:

“(5)

Enterprises of a Contracting State, the capital of which is wholly or partly owned or controlled, directly or indirectly, by one or more residents of the other Contracting State, shall not be subjected in the first-mentioned Contracting State to any taxation or any requirement connected therewith which is other or more burdensome than the taxation and connected requirements to which other similar enterprises of the first mentioned State are or may be subjected.”

The paragraph is easier to understand if the appropriate references to the UK and the USA are inserted:

“Enterprises of the UK, the capital of which is wholly or partly owned or controlled, directly or indirectly, by one or more residents of the US, shall not be subjected in the UK to any taxation or any requirement therewith which is other or more burdensome than the taxation and connected requirements to which other similar enterprises of the UK are or may be subjected.”

It is also necessary to have regard to the interest article, Article 11. Article 11(2) lays down the general rule that:

“Interest derived and beneficially owned by a resident of the United States shall be exempt from tax by the United Kingdom.”

However, by virtue of paragraph (5):

“Where, owing to a special relationship between the payer and the person deriving the interest or between both of them and some other person, the amount of the interest paid exceeds for whatever reason the amount which would have been paid in the absence of such relationship, the provisions of this Article shall apply only to the last-mentioned amount. In that case, the excess part of the payment shall remain taxable according to the law of each Contracting State, due regard being had to the other provisions of this Convention.”

It follows that if the relevant loan finance had been provided to UK Holdco by a US-resident company, instead of by IIF, the arm’s length test in Article 11(5) would have applied, and any interest paid in excess of the arm’s length threshold would have remained taxable in the UK, subject to “due regard being had to the other provisions” of the US Treaty.

Discussion

As refined in the course of oral submissions, the argument for IBM may conveniently be broken down into the following steps:

Mr Ewart did not, as I understood his submissions, dispute steps (a) to (c) in the argument, but he took a preliminary point which, if it is sound, would dispose of the claim. The point was essentially this. Article 11(5) shows that the high contracting parties, the US and the UK, agreed upon and incorporated into the US Treaty a thin cap provision which was based on an arm’s length test, and which did not depend for its operation on the presence of a tax-avoidance purpose. Thus if the relevant loan finance had been provided to UK Holdco by a US group company, the arm’s length test would have applied, and in order to avoid taxation of the excess interest in the UK similar injections of capital would still have had to be made. Since the parties made this express thin cap agreement in relation to loans made directly by a US group company to a UK subsidiary, they cannot possibly have intended that a UK subsidiary should be in a better position if the loans were instead routed through an EU resident company.

I am unable to accept this argument. I can well understand that if Article 11(5) and Article 24(5) both prima facie applied to the same factual situation, the specific thin cap agreement in the former should probably be construed as prevailing over the general provisions of the latter. However, the present case is not of that nature, because Article 11 is not engaged at all. The relevant interest was all paid to the Dutch lender, IIF. I therefore do not think that Article 11(5) can be read as implicitly limiting the scope or application of Article 24(5). The purpose of the latter provision, on the face of it, is quite different, namely to place US-owned or controlled UK enterprises (such as UK Holdco) in no different or more burdensome a tax position than if they were UK-owned or controlled. On the hypothesis of UK ownership or control, the UK owner would be entitled to claim the benefit of Article 43 EC in relation to its EU-resident subsidiaries.

Mr Ewart’s next argument was that, even if the UK owner’s right of establishment under Community law was notionally engaged, the only parties affected were the UK owner and IIF. The UK owner could have no right of establishment in relation to UK Holdco, the taxpayer company, because the right is inherently one that applies only to an establishment in a different member state. On this argument, the UK’s breach of Article 43 might in theory provide the UK owner with a notional remedy in relation to IIF, but it could not do so in relation to UK Holdco.

This submission is in my judgment well-founded, and I accept it. The operation of the UK thin cap rules may infringe the right of establishment of an EU resident parent of a UK subsidiary which suffers the UK tax consequences of the application of the rules. In those circumstances both the EU parent and the UK subsidiary may have a good claim under Community law, and as Halliburton shows it would be no defence for the Revenue to say that the adverse tax consequences were all suffered in the UK. The value of the parent’s investment in its UK subsidiary would, at least potentially, be prejudiced. However, that is a very different matter from saying that the operation of the UK thin cap rules adversely affected the notional UK owner’s right of establishment of its Dutch subsidiary IIF. The parallel situation would be if the Dutch thin cap rules (assuming them to exist) prejudiced the UK owner’s right of establishment of IIF in the Netherlands. It cannot in my judgment be assumed that the UK owner’s right of establishment of IIF in the Netherlands was somehow prejudiced merely because, in response to the UK thin cap rules, the group chose to finance UK Holdco directly with equity rather than with further debt, channelled through IIF.

Even if that argument is wrong, Mr Ewart had a further string to his bow. Relying on the decision of the House of Lords in Boake Allen Ltd v Revenue and Customs Commissioners [2007] UKHL 25, [2007] 1 WLR 1386 (on appeal from the decision of the Court of Appeal in NEC Semi-Conductors), he submitted that the question posed by Article 24(5) of the US Treaty is whether there is discrimination against a UK company “on the ground that its capital is “wholly or partly owned or controlled, directly or indirectly” by residents of the US …”: see the speech of Lord Hoffmann (with whom all of their Lordships agreed on this point) at paragraph 16. In Boake Allen the question arose in the context of a claim that the inability of a UK company and its US parent to make a group income election under ICTA section 247 upon payment of a dividend breached Article 24(5). In disagreement with Park J and the Court of Appeal, the House of Lords held that it did not.

Lord Hoffmann explained why he reached this conclusion in paragraph 17:

“In my opinion [section 247] plainly does not [discriminate on the grounds that the capital of the subsidiary is controlled by a non-resident company]. For example, if a US parent were to interpose a UK-resident holding company between itself and its UK-resident subsidiary, the control would remain in the US but there would be no objection to an election by the UK subsidiary and its immediate, UK-resident parent. On the other hand, an individual US shareholder and the company he controls in the UK could not elect, but the reason is not because the company is subject to US control. An individual UK shareholder and his company could not elect either, for the same reason that a non-resident company cannot elect. It is because an individual is not liable to corporation tax. An election is a joint decision by two entities paying and receiving dividends that one rather than the other will be liable for ACT. This is not a concept which can meaningfully be applied when one of the entities is not liable for ACT at all.”

By parity of reasoning, Mr Ewart submits that in the present case the correct question to ask is whether the UK thin cap rules discriminated against UK Holdco on the grounds that the capital of that company was owned or controlled by residents of the USA. The answer to that question is clearly No, because what matters is the residence of the lending company, IIF, not the residence of the company which owns or controls the UK subsidiary. The point may be tested, says Mr Ewart, in the same way as Lord Hoffmann did. If the US parent were to interpose a UK-resident holding company between it and the UK and Dutch subsidiaries, the claim of breach of Article 43, assuming it to be otherwise valid, would still apply, but control would remain in the USA.

I confess that I do not understand the last stage in this argument. The point in Boake Allen was that, if a UK parent were interposed, it would then be possible to make a group income election, even though control of the UK subsidiary would remain in the USA. This showed that the inability to make an election was not caused by the US control of the UK subsidiary. In the present case, by contrast, I cannot see what difference to the application of the UK thin cap rules would be made by the interposition of an intermediate UK parent. I am also unable to agree with Mr Ewart’s basic point that the crucial factor in the application of the thin cap provisions is the residence of the lender. It is true that the discrimination identified by the ECJ depends on drawing a contrast between the treatment of loans by a UK lender and loans by a non-UK lender, but the important point in the present context is that only parent companies resident in another member state can complain of infringement of Article 43. As the ECJ made clear in its answer to Question 2 in the order for reference, Article 43 has no bearing on the application of thin cap rules to third countries: see paragraphs 97 to 100 of the judgment.

The true answer to the question posed by the House of Lords in Boake Allen, in my judgment, is that the UK thin cap rules discriminate between groups where the ultimate ownership and control is to be found in a member state and groups where it is to be found in a third country, because the former can invoke Article 43 and the latter cannot. Since the UK is a member state, and the USA is not, it seems to me that the necessary discrimination for the purposes of Article 24(5) of the US Treaty is made out. However, as I have already said, I consider that IBM’s claim must fail, because the hypothesis required by Article 24(5) does not relevantly engage Article 43 EC.

In the circumstances it is unnecessary for me to consider what relief, if any, would be available to the IBM test claimants if a breach of Article 24(5) were established.

My main conclusions may be summarised as follows: