BAILII [Home] [Databases] [World Law] [Multidatabase Search] [Help] [Feedback]

First-tier Tribunal (Tax)


You are here: BAILII >> Databases >> First-tier Tribunal (Tax) >> Philips Electronics UK Ltd v Revenue & Customs [2009] UKFTT 226 (TC) (18 August 2009)
URL: http://www.bailii.org/uk/cases/UKFTT/TC/2009/TC00176.html
Cite as: [2010] 1 CMLR 16, [2009] STI 2839, [2009] SFTD 629, [2009] UKFTT 226 (TC), 12 ITL Rep 194

[New search] [View without highlighting] [Printable RTF version] [Help]



INCOME TAX/CORPORATION TAX
European law
    [2009] UKFTT 226 (TC)
    Appeal number: SC 3124/08
    EUROPEAN LAW – group relief (consortium claim) – whether claimant company not exercising Community rights can seek to disapply UK law affecting a Dutch surrendering company with a UK branch – yes – whether TA 1988 s 406(2) and 403D(1)(c) contain restrictions on the right of establishment – yes – whether such restrictions can be justified – no – whether (if it could be justified) s 403D(1)(c) would be proportional – no
    FIRST-TIER TRIBUNAL
    TAX
    PHILIPS ELECTRONICS UK LIMITED First referrer
    - and -
    THE COMMISSIONERS FOR HER MAJESTY'S
    REVENUE AND CUSTOMS [corporation tax] Second referrer
    TRIBUNAL: JOHN F AVERY JONES CBE
    ROGER BERNER (TRIBUNAL JUDGES)
    Sitting in public in London on 27 and 28 July 2009
    David Milne QC and Daniel Jowell, counsel, instructed by McGrigors LLP and Grant Thornton UK LLP, for the Appellant
    David Ewart QC and David Yates, counsel, instructed by the General Counsel and Solicitor to HM Revenue and Customs for the Respondents
    © CROWN COPYRIGHT 2009

     
    DECISION
  1. This is a joint referral pursuant to para 31A of Sch 18 to the Finance Act 1998 by Philips Electronics UK Limited ("the Taxpayer") and the Commissioners of HM Revenue and Customs ("HMRC"). The Taxpayer was represented by Mr David Milne QC and Mr Daniel Jowell, and HMRC by Mr David Ewart QC and Mr David Yates.
  2. The issue is the compatibility with EU law of two provisions of UK law in respect of a consortium claim for group relief ("consortium relief") within s 402(3) of the Income and Corporation Taxes Act 1988 ("the Taxes Act").
  3. The joint referral was in the following terms:
  4. "1. Whether, in light particularly of any applicable principles of EU law, section 406(2) ICTA 1988, or any other provision, applies so as to prevent consortium relief from being available on the basis that the relevant link companies (as that term is defined in section 406(1) ICTA 1988), being Koninklijke Philips Electronics NV and Philips GmbH, although resident at all times in an EU Member State, were not at any relevant time within the charge to UK corporation tax, and
    2. Whether, in light particularly of any applicable principles of EU law, section 403D ICTA 1988 (entitled 'Relief for or in respect of non-resident companies'), or any other provision, applies so as to prevent consortium relief from being available by reference to the prospect of those losses being utilised (as more fully spelt out in section 403D ICTA 1988, or such other relevant provision), in any period, for the purposes of Dutch corporate income tax or any other non-UK tax."
  5. We had an agreed statement of facts as follows:
  6. The references in square brackets below are to the Agreed Bundle of Documents for the hearing before the First-tier Tribunal (Tax Chamber) held on 27 to 28 July 2009.
    (1) The Taxpayer is a company which is incorporated in England and Wales and resident for tax purposes in the UK and had an accounting period for all relevant years ending 31 December. The Taxpayer has made various consortium relief claims for the years 2001-2004 in relation to the losses of the UK branch of a Netherlands company, LG Philips Displays Netherlands BV ("LG.PD Netherlands"). LG.PD Netherlands was incorporated in the Netherlands on 22 April 2001 and is resident in the Netherlands for tax purposes. The consortium relief claims made were as follows:
    (a) For the period 1 July 2001 to 31 December 2001: claims were originally made on 4 February 2003 with loss surrenders dated in January 2003. Further claims dated 1 August 2008 were made with loss surrenders dated 13 August 2008. Further claims and corresponding surrenders were made on 9 July 2009.
    (b) For the year ended 31 December 2002: claims were originally made on 28 January 2004 with loss surrenders dated in December 2003/January 2004. Further claims dated 1 August 2008 were made with loss surrenders dated 13 August 2008. Further claims and corresponding surrenders were made on 9 July 2009;
    (c) For year ended 31 December 2003: claims were originally made on 17 December 2004 with loss surrenders dated in December 2004. Further claims dated 1 August 2008 were made with loss surrenders dated 13 August 2008. Further claims with corresponding surrenders were made on 9 July 2009;
    (d) For the year ended 31 December 2004: claims were made on 12 December 2005 with loss surrenders dated in November/December 2005. Further claims dated 1 August 2008 were made with loss surrenders dated 13 August 2008. Further claims with corresponding surrenders were again made on 9 July 2009.
    (2) The amounts claimed in the various claims are as follows:
      1st Claim (various dates) 2nd Claim (all 13 August 2008) 3rd Claim (all 9 July 2009)
    2001 £3,980,103
    [B5/Tab 54/1504]
    £3,980,103
    [B5/Tab 73/1552]
    £3,980,103
    [B5/Tab 75A/1587H&J]
    2002 £27,082,654
    [B5/Tab 57/1516]
    £25,308,909
    [B5/Tab 73/1558]
    £25,308,909
    [B5/Tab 75A/1587L&O]
    2003 £29,888,028
    [B5/Tab 62/1527&30]
    £31,661,773
    [B5/Tab 73/1564]
    £31,661,773
    [B5/Tab 75A/1587S-T]
    2004 £3,149,348
    [B5/Tab 67/1538C&E]
    £3,149,348
    [B5/Tab 73/1570]
    £3,149,348
    [B5/Tab 75A/1587X-Y]
    (3) The Taxpayer is part of the Philips Group, the ultimate parent of which is Koninklijke Philips Electronics NV ("KPE NV"). KPE NV was at all material times a Netherlands registered company and resident in the Netherlands for tax purposes.
    (4) KPE NV entered into a joint venture structure with a group headed by LG Electronics Inc. (a South Korean company). The relevant joint venture agreement was signed on 11 June 2001 [B1/Tab 1] and the JV structure was subsequently established on 26 June 2001. The JV structure is itself headed by LG Philips Displays Holding BV ("LG.PD Holding") in which KPE NV and its subsidiaries hold a 50% plus 1 share interest, the remaining 50% being held by LG Electronics Inc. and its subsidiaries. LG.PD Holding is incorporated in the Netherlands and is resident in the Netherlands for tax purposes.
    (5) LG.PD Netherlands is a wholly-owned subsidiary of LG.PD Holding, and formed part of the larger joint venture structure. Part of the UK operation of LG.PD Netherlands itself derived from the Taxpayer as part of a general contribution of assets from various parts of the Philips and LG Electronics groups.
    (6) Some of the constituent parts of the JV structure also formed from 1 January 2002 a fiscale eenheid or "Fiscal Unity" under Netherlands tax law which in turn permitted consolidation of profits and losses between the various members of the Fiscal Unity.
    (7) The position can be summarised in the following diagram:

    Image 1

    (8) As the above indicates, KPE NV and Philips GmbH (a German resident entity) were the relevant link companies for any claim made for consortium relief by the Philips Group. After 25 May 2004, KPE NV held the entire 50% plus one share directly (but nothing turns on this).
    (9) The overall JV structure contained many different subsidiaries and branches in various countries. An overall structure chart of the structure prior to its partial liquidation can be found in Annex 1 of the First Trustee Report [B3/Tab 43/929].
    (10) As mentioned above, some of the members of JV structure also formed a "Fiscal Unity". This was formed with effect from 1 January 2002 and initially contained the following companies [A1/Tab 9/58]]:
    (a) LG.PD Holding;
    (b) LG.PD Netherlands;
    (c) LG.PD China BV;
    (d) LG.PD Investment BV ("LG.PD Investment").
    (11) On 29 July 2004, two further companies LG Philips Displays Barcelona I BV and LG Philips Displays Barcelona II BV ("LG.PD Barcelona I & II"), joined the Fiscal Unity.
    (12) LG.PD China BV was subsequently renamed LG Philips Displays International BV[1] ("LG.PD International") and for convenience will be referred to as LG.PD International.
    (13) LG.PD Netherlands was therefore a member of the Fiscal Unity for every year in which consortium relief is claimed by the Taxpayer with the exception of the period 1 July 2001-31 December 2001.
    (14) The UK trading operation of LG.PD Netherlands proved to be loss making in every year from commencement. The losses according to UK tax principles were:
    (a) 1 July 2001-31 December 2001: loss of £7,960,206;
    (b) y/e 31 December 2002: loss of £50,777,681;
    (c) y/e 31 December 2003 loss of £94,701,489;
    (d) y/e 31 December 2004 loss of £43,879,185.
    (15) In relation to the Fiscal Unity as a whole, the following figures are said to represent the position under Netherlands tax law[2]:
     € 2002 2003 2004
    Holding (139,554,946) (127,337,656) 86,478,932
    Investment 364,063 48,776,042 891,521
    Netherlands (total) (104,275,214) (213,447,194) (143,982,131)
    Netherlands (UK) (61,662,214) (168,877,923) (39,189,828)
    International (198,616) (1,142,867) (1,877,284)
    Barcelona I 0 0 (110,904)
    Barcelona II 0 0 (110,892)
    Total (243,664,713) (293,151,675 (58,710,758)
    (16) On 11 February 2003, the Taxpayer and LG.PD Netherlands entered into an agreement under which the Taxpayer would pay LG.PD Netherlands in respect of the surrender of losses of the UK branch of LG.PD Netherlands to the Taxpayer. A copy of this agreement is found at [B1/Tab 5]. In summary, the Taxpayer agreed to pay LG.PD Netherlands 50% of UK corporation tax saved by utilising the losses of LG.PD Netherlands' UK branch.
    (17) By March 2005 all of the UK Branch's operations had been closed down with the exception of the Blackburn factory (which was originally held by the Taxpayer).
    (18) In the period prior to the commencement of bankruptcy proceedings in the Netherlands in January 2006, a plan was developed to consolidate the viable parts of the JV structure into new incorporated subsidiaries within a subgroup headed by LG.PD International.
    (19) Pursuant to this plan, Caray Limited was incorporated on 18 January 2006 as the vehicle to acquire the trade and assets of the UK Branch of LG.PD Netherlands. Caray Limited was a directly held 100% subsidiary of LG.PD International.
    (20) On 26 January 2006, LG.PD Netherlands and Caray Limited entered into four separate agreements:
    (a) An assignment of the business and assets [B1/Tab 4];
    (b) A global assignment of intellectual property of the Blackburn factory [B1/Tab 9];
    (c) Sale and purchase of two leaseholds of buildings at Blackburn [B1/Tab 7]; and
    (d) An assignment of registered rights [B1/Tab 8].
    (21) On 30 January 2006, Caray Limited was renamed LG Philips Displays Blackburn Ltd. On 20 March 2007, the company was subsequently renamed Blackburn Microtech Solutions Ltd. The company is referred to here as "LG.PD Blackburn".
    (22) On 27 January 2006, LG.PD Netherlands entered bankruptcy.
    (23) On 30 January 2006, LG.PD Holding entered bankruptcy.
    (24) On 28 February 2006, LG.PD Investment entered bankruptcy.
    (25) On 9 January 2009, LG.PD International entered bankruptcy.
    (26) The above companies are now all in a state of bankruptcy. The Netherlands bankruptcy procedure has not ended and none of the companies have been dissolved.
    (27) Notwithstanding the bankruptcy of LG.PD International, LG.PD Blackburn continues to trade, but has incurred losses in all years for which results are available.
    Use of losses of the UK Branch of LG.PD Netherlands
    (28) The contents of the Joint Experts' report are correct [A2/Tab10/260]. In summary the position in relation to the losses arising from the UK Branch as computed for the purposes of Netherlands law is as follows.
    (a) For the period 1 July 2001 to 31 December 2001, LG.PD Netherlands was not part of the Fiscal Unity and hence UK Branch losses arising in this period can only be used by LG.PD Netherlands under Netherlands law. The losses arising in this period are available for carry forward against future profits of LG.PD Netherlands only and they will cease to be available for use under Netherlands law on 31 December 2011 unless LG.PD Netherlands is dissolved at an earlier date in which case the losses will expire.
    (b) For the period 1 January 2002 to 31 December 2002, the losses of the UK Branch are available for use by the Fiscal Unity. It is the Joint Experts' view that €91,993 of LG.PD Investment's profits of that year were set against these losses on a pro-rata basis. The remaining losses arising in this period are available for set off against any future profits of any member of the Fiscal Unity that was also a member at the time the losses were incurred. Remaining losses will cease to be available for use under Netherlands law on 31 December 2011 unless LG.PD Netherlands and LG.PD Holding are dissolved at an earlier date in which case the losses will then expire;
    (c) For the period 1 January 2003 to 31 December 2003, the losses of the UK Branch are available for use by the Fiscal Unity. It is the Joint Experts' view that €24,090,462 of LG.PD Investment's profits were set against these losses on a pro-rata basis. If an ex officio amendment is made to the nil determination for that year the remaining losses arising in this period are available for set off against any future profits of any member of the Fiscal Unity that was also a member at the time the losses were incurred. Remaining losses will cease to be available for use under Netherlands law on 31 December 2012 unless, LG.PD Netherlands and LG.PD Holding are dissolved at an earlier date in which case the losses will then expire.
    (d) For the period 1 January 2004 to 31 December 2004, the losses of the UK Branch are available for use by the Fiscal Unity in future years. It is the Joint Experts' view that €23,439,243 of LG.PD Investment's and LG.PD Holding's profits were set against these losses on a pro-rata basis. If an ex officio amendment is made to the nil determination for that year the remaining losses arising in this period are available for set off against any future profits of any member of the Fiscal Unity that was also a member at the time the losses were incurred. Remaining losses will cease to be available for use under Netherlands law on 31 December 2013 unless LG.PD Netherlands and LG.PD Holding are dissolved at an earlier date in which case the losses will then expire.
    (29) In relation to years 2002, 2003 and 2004, the foregoing summary is to be qualified by the statements made in paragraph 90 of the Joint Expert's report.
    (30) The overall picture of the use of losses can therefore be summarised in the following diagram. This diagram also shows the transfer of the Blackburn business from LG.PD Netherlands to LG.PD Blackburn.

    Image 2

  7. This referral concerns consortium relief. It is not necessary to go into the details of the relief. As the structure chart above shows the Taxpayer claimant company is a subsidiary of KPE NV and the surrendering company is LG.PD Netherlands in respect of its UK branch. That company is a consortium company owned directly or indirectly as to 50% plus one share by KPE NV. KPE NV and Philips GmbH are link companies. If all companies were UK resident LG.PD Netherlands would be entitled to surrender approximately 50% of its losses to the Taxpayer. It is common ground that as a matter of UK domestic law it cannot surrender such losses, but only because of two provisions of UK law which the Taxpayer contends infringe EU law. Those provisions (as they were applicable for the accounting periods in question) are:
  8. (1) Section 406(2) of the Taxes Act 1988:
    (2) Subject to subsections (3) and (4) below, where the link company could (disregarding any deficiency of profits) make a consortium claim in respect of the loss or other amount eligible for relief of a relevant accounting period of a consortium company, a group member may make any consortium claim which could be made by the link company; and the fraction which is the relevant fraction for the purposes of section 403C where a group member is the claimant company shall be the same as it would be if the link company were the claimant company.
    This limits the consortium claim by a group company to circumstances where the link companies could themselves have made consortium claims if they had taxable profits. Here the link companies, KPE NV and Philips GmbH, being non-resident and not carrying on business in the UK, could not make any consortium claim.
    (2) Section 403D(1)(c) of the Taxes Act 1988:
    (1) In determining for the purposes of this Chapter the amounts for any accounting period of the losses and other amounts available for surrender by way of group relief by a non-resident company, no loss or other amount shall be treated as so available except in so far as—
    (a)     …
    (c)     no part of—
    (i)     the loss or other amount, or
    (ii)     any amount brought into account in computing it,
    corresponds to, or is represented in, any amount which, for the purposes of any foreign tax, is (in any period) deductible from or otherwise allowable against non-UK profits of the company or any other person.
    (6) So much of the law of any territory outside the United Kingdom as for the purposes of any foreign tax makes the deductibility of any amount dependent on whether or not it is deductible for tax purposes in the United Kingdom shall be disregarded for the purposes of this section.
    This prevents losses of a UK branch of a non-resident company being surrendered if any part of those losses corresponds to amounts deductible for foreign tax. The provision is extremely strict. It prevents relief of any losses if (a) any part of the losses, however small, is deductible abroad; (b) in any period in the future, and (c) even if the losses are not allowed abroad because of a rule that prevents losses being allowed if they are deductible in the UK.
  9. The following issues were raised:
  10. (1) Whether the Taxpayer can raise EU law in these proceedings.
    (2) Whether the above provisions of UK law are a restriction on the freedom of establishment.
    (3) Whether, if they are, that restriction is justified.
    (4) Whether the restriction is proportionate.
    Issue 1: the standing of the Taxpayer to raise the EU issue
  11. Mr Ewart for HMRC contends that the Taxpayer is not in a position to ask the Tribunal to disapply domestic law in its favour (assuming for the moment that domestic law does not comply with EU law). The Taxpayer's argument is that the freedom of LG.PD Netherlands (or its holding company, LG.PD Holding, or KPE NV), rather than its own freedom, to establish a UK branch is restricted by s 406(2), preventing a consortium claim entirely, and by s 403D(1)(c), preventing double use of losses. The Taxpayer has not exercised any freedom of establishment and so cannot claim to have any Community right that requires protection. He relies on ICI v Colmer [1999] STC 1089 in the House of Lords following the decision of the ECJ in which Lord Nolan said at 1094j:
  12. "It remains to consider the question of disapplication in accordance with the provisions of s 2(1) and (4) of the European Communities Act 1972 (the 1972 Act). Explaining the effect of the section in R v Secretary of State for Transport, ex p Factortame Ltd [1990] 2 AC 85, Lord Bridge of Harwich said (at 140):
    'By virtue of section 2(4) of the Act of 1972 Part II of the Act of 1988 [the Merchant Shipping Act] is to be construed and take effect subject to directly enforceable Community rights and those rights are, by section 2(1) of the Act of 1972, to be "recognised and available in law, and ... enforced, allowed and followed accordingly; ..." This has precisely the same effect as if a section were incorporated in Part II of the Act of 1988 which in terms enacted that the provisions with respect to registration of British fishing vessels were to be without prejudice to the directly enforceable Community rights of nationals of any member state of the E.E.C.'
    So, in the present case, the effect of s 2 of the 1972 Act is the same as if a subsection were incorporated in s 258 of the 1970 Act which in terms enacted that the definition of 'holding company' was to be without prejudice to the directly enforceable Community rights of companies established in the Community. As the concluding paragraphs of the judgment of the Court of Justice make plain, this in no way affects the application of the definition to companies established outside the Community (cf in this connection the comments of Lord Keith of Kinkel on the effect of the Factortame decision in R v Secretary of State for Employment, ex p Equal Opportunities Commission [1995] 1 AC 1 at 27)."
  13. Mr Ewart contends that the Taxpayer cannot claim "directly enforceable Community rights of companies established in the Community" because it has not exercised any Community rights.
  14. Mr Milne contends that it cannot matter in EU law who challenges the application of the legislation provided the person bringing the challenge is affected by it. He points out that the UK company in Factortame, Case C-213/89, [1990] ECR I-2433 and Case C-221/89, [1991] ECR I-3905 was entitled to claim Community rights. Similarly in Halliburton, Case C-1/93, [1994] ECR I-1137 it was the Dutch transferee company of the assets of the former German company that successfully claimed the relief from transfer tax on Dutch real property charged on the transferee that applied for transfers between two Dutch companies. He made an application, which he did not press, for KPE NV, LG.PD Holding and LG.PD Netherlands to be joined as a party if this were necessary to enable the point to be argued in these proceedings.
  15. Mr Ewart replies that "it" in Mr Milne's contention at the end of the first sentence of the previous paragraph is the domestic legislation whereas one should be concentrating on the Community rights. He gives an analogy of the three-year cap in VAT. A person making a claim after the introduction of the cap is affected by it but has no Community rights; it is only those who had an existing right to claim before the legislation was introduced without any transitional provision who have Community rights. Mr Ewart also contends that Factortame is not relevant because the UK company was the person complaining that its right of establishment was being restricted. He said that one did not know what the position was in the Netherlands in Halliburton (by which he may have meant Germany as the case was a reference from the Netherlands and it was the Dutch transferee company that was claiming the relief). He contends that joining a party with Community rights would make no difference.
  16. Mr Ewart told us that similar issue had been raised in the Thincap GLO proceedings before Henderson J but no decision was expected until the autumn. We considered that it would be better if we decided the issue rather than hold up our decision as this is a threshold issue.
  17. In normal cases it will be the person exercising directly enforceable Community rights who will complain if the exercise of those rights is restricted. But the nature of the legislation with which we are concerned means that this is not a normal case. Section 403D(1)(c) relates to the surrender of losses by the UK branch of a non-resident company to a claimant who is taxable in the UK. Necessarily the person primarily exercising Community rights is the company with the UK branch with the loss. But the person whose tax liability is directly affected by a restriction on that person's right to surrender the loss is the claimant company (the branch may be affected by not receiving a payment for group relief). If domestic law is contrary to EU law the person who suffers is the claimant company. While the claimant company is not itself exercising Community rights it is the person most directly affected by such exercise by another person. This is very different in nature from the situation in Colmer in which the definition of "holding company" had to be disapplied in relation to companies established in the EU but not those established in non-member states. The House of Lords was making a distinction between subsidiaries affected by EU law and those which were not, rather than a distinction between companies exercising Community rights and UK companies which were affected by the exercise of such rights by others. It is also different from persons not affected by the transition being bound by the three-year cap, which is a case of domestic legislation having full effect because EU case law only affected the transitional cases. We agree with Mr Ewart that Factortame is not an example of a claim by a person other than the one claiming directly enforceable Community rights because the issue was whether restrictions on the nationality and residence of shareholders and directors of the UK incorporated company that brought the proceedings were compatible with EU law. Halliburton is certainly an example of the ECJ allowing the transferee company that had not itself exercised any Community rights claiming such rights in the ECJ, although we consider that not too much can be read into that case because the point was not argued, and both the transferor and transferee companies were subsidiaries of a US company that was treated as if it were a Dutch company under the ownership non-discrimination provision of the US-Netherlands tax treaty. But it was clearly only the German transferor company that had in the past exercised Community rights in setting up the Dutch branch; and presumably the US company had set up the Dutch subsidiary.
  18. One of us as a Special Commissioner dealt with a similar issue in Jansen Nielsen Pilkes v Tomlinson [2004] STC (SCD) 226 in which a UK subsidiary of a Netherlands company was claiming that the existence of the parent company and other EU companies in the group affected its claim to small companies relief which was a restriction on the freedom of establishment of the parent company. The appellant had applied to join the EU parent company, which had been refused. The Special Commissioner said:
  19. "13. I should mention that the appellant did try to join the parent company in this appeal which was refused by my colleague Dr Brice at a preliminary hearing. Since the parent company is not a UK taxpayer it is difficult to see what purpose could be served by joining it. If the parent company suffers a restriction on its freedom of establishment (or discrimination) because of the appellant paying an increased amount of tax, that can be dealt with in these proceedings."

    As in this case, it was the tax of the UK company that was affected by the alleged restriction on another company's Community rights. It was held that the parent company's Community rights were not affected and so the point is obiter (and in any event not binding on us) but we propose to follow it on the basis that it gives effect to the principle of effectiveness. It would be more difficult for a company to claim that its Community rights were affected if it had to start proceedings for damages in the High Court on an issue that was essentially one arising in a tax appeal. The approach of the House of Lords in Autologic Holdings plc and others v Her Majesty's Commissioners of Inland Revenue (Re Claimants under Loss Relief Group Litigation Order) [2005] STC 1357 points to the Tribunal having jurisdiction to decide an EU law issue when there are proceedings properly brought before it. What is odd is that HMRC should agree to join in making the referral, which is entirely on the compatibility of two provisions of consortium relief with EU law, and then argue that the Taxpayer could not raise the issue. It would be preferable in such a case for such jurisdictional issues to be raised as a separate issue in the referral.

  20. Here the person most affected by any restriction on the freedom of establishment is the Taxpayer which will have to pay more tax if it cannot use the losses because of a breach of the directly enforceable Community rights of another company established in the Community. Accordingly, the Taxpayer has standing to raise the issue in these proceedings.
  21. Issue 2: Whether the two provisions of UK law amount to a restriction
    Section 406(2)
  22. Mr Milne contends in outline:
  23. (1) The Taxpayer is placed in a less favourable position because other companies in its group, the link companies, are resident and carrying on business in another EU state. Such a restriction makes it less attractive for companies established in another member state to establish a branch or subsidiary in the UK.
    (2) This is a difference in treatment between resident subsidiaries according to the seat of their parent company, see Oy AA, Case C-231/05, [2008] STC 991:
    "39. A difference in treatment between resident subsidiary companies according to the seat of their parent company constitutes an obstacle to the freedom of establishment if it makes it less attractive for companies established in other member states to exercise that freedom and they may, in consequence, refrain from acquiring, creating or maintaining a subsidiary in the state which adopts that measure (see Lankhorst-Hohorst GmbH v Finanzamt Steinfurt (Case C-324/00) [2003] STC 607, [2002] ECR I-11779, para 32 and the Test Claimants in the Thin Cap Group Litigation case (para 61)).
    41. The United Kingdom government further argues that, since the intra-group financial transfer was not taxed in the United Kingdom, which Oy AA, however, denies, and since the losses of AA Ltd could be carried forward to other financial years so as to be set off against profits subsequently made, the delay which that parent company suffered before being able to equalise profits and losses had only an indirect and uncertain effect on its decision whether to establish a subsidiary in Finland.
    42. However, for legislation to be regarded as a restriction on the freedom of establishment, it is sufficient that it be capable of restricting the exercise of that freedom in a member state by companies established in another member state, without there being any need to establish that the legislation in question has actually had the effect of leading some of those companies to refrain from acquiring, creating or maintaining a subsidiary in the first member state (see the Test Claimants in the Thin Cap Group Litigation case (para 62)).
    43. It follows that the difference in treatment to which resident subsidiaries are subjected, under a system such as that at issue in the main proceedings, by reason of the place of the corporate seat of their parent company constitutes a restriction on the freedom of establishment."
    (3) The position is covered by the decision of the ECJ in Papillon, Case C-418/07, [2009] STC 542. There a French parent and sub-subsidiary (the intermediate subsidiary being a Dutch company) wanted to establish tax integration, which would have been possible if the Dutch company had been French and had been included in the tax integration, but no claim to include the Dutch company was capable of being made (or was made). The ECJ held this to be a restriction:
    "20. As was noted in paras 3 and 4 of this judgment, the parent company of the group can, if it is to benefit from the tax integration regime, have an indirect holding in another group company only if this is done through a company which is itself a member of the integrated group and is accordingly liable to corporation tax in France.
    21. Thus, a parent company which has its registered office in France and which holds its French sub-subsidiaries through a subsidiary established in another member state cannot benefit from the tax integration regime. By contrast, a French parent company is able to achieve tax integration with its French sub-subsidiaries where the intermediate subsidiary is established in France.
    22. As the Advocate General stated at para 30 of her opinion, the effect is that the provisions of the CGI at issue in the main proceedings create a difference in treatment since the ability to elect for the tax integration regime is dependent on whether the parent company holds its indirect shares through a subsidiary established in France or in another member state.
    23. The French government contends, however, that that difference in treatment can be explained by the fact that those two situations are not objectively comparable.
    24. In a situation such as that which arises in the main proceedings, a subsidiary which is established in a member state other than the French Republic is, by virtue of that fact, not subject to corporation tax in France, unlike the position which arises when the subsidiary has its registered office in that state.
    25. That argument cannot be accepted.
    26. Acceptance of the proposition that the member state may freely apply a different treatment solely by reason of the fact that a company's registered office is situated in another member state would deprive the rules relating to the freedom of establishment of all meaning (see, to that effect, EC Commission v French Republic (Case 270/83) [1986] ECR 273, [1987] 1 CMLR 401, para 18, and Metallgesellschaft Ltd v IRC and A-G; Hoechst AG v IRC and A-G (Joined cases C-397/98 and C-410/98) [2001] STC 452, [2001] Ch 620, [2001] ECR I-1727, para 42).
    27. In order to establish whether discrimination exists, the comparability of a Community situation with one which is purely domestic must be examined by taking into account the objective pursued by the national provisions at issue (see, to that effect, Metallgesellschaft, para 60, and Proceedings brought by Oy AA (Case C-231/05) [2008] STC 991, [2007] ECR I-6373, para 38).
    28. In the main proceedings, the provisions of the CGI at issue aim to treat, as far as possible, a group constituted by a parent company with its subsidiaries and its sub-subsidiaries in the same way as an undertaking with a number of permanent establishments, by allowing the results of each company to be consolidated.
    29. That objective can be attained both in the situation of a parent company which is resident in a member state and holds sub-subsidiaries also resident in that state through a subsidiary which is itself resident, and in the situation of a parent company which is resident in the same member state and holds sub-subsidiaries also resident in that state, but through a subsidiary established in another member state.
    30. Having regard to the objectives of the CGI at issue in the main proceedings, those situations are thus objectively comparable.
    31. Consequently, the tax regime at issue in the main proceedings gives rise to unequal treatment based on the place of the registered office of the subsidiary through which the resident parent company holds its resident sub-subsidiaries.
    32. In as much as, from a taxation perspective, they put Community situations at a disadvantage compared with purely domestic situations, the provisions of the CGI at issue in the main proceedings thus constitute a restriction which is, in principle, prohibited by the provisions of the Treaty relating to the freedom of establishment."
    (We shall deal with the justification and proportionality arguments raised in that case later in this decision.) The effect of the intermediate Dutch company on loss relief between two French group companies is directly comparable to the effect of the link companies here.
  24. Mr Ewart did not deal with this aspect, concentrating on s 403D(1)(c) on the basis that there was no need to decide the issue on s 406(2) if the Taxpayer failed under s 403D(1)(c). He did, however, say that a reference to the ECJ might be appropriate if that were the only issue but he was not proposing a reference at this stage.
  25. We agree with Mr Milne that s 406(2) is a clear case of a restriction and any doubt about it was resolved by the ECJ in Papillon which is identical in all material respects to the situation of the link companies. In Papillon relief for losses between two French companies was restricted by the existence of an intermediate Dutch company; here group relief between a UK branch and a UK company is restricted by the existence of non-UK link companies.
  26. Section 403D(1)(c)
  27. Mr Milne contends in outline:
  28. (1) The effect of s 403D(1)(c) is that because losses made in a UK branch rather than a UK subsidiary cannot be surrendered this discriminates against non-UK companies setting up UK branches rather than subsidiaries.
    (2) The comparison between a branch and a subsidiary is appropriate when one is dealing with the source (or host) state, as in Avoir fiscal (Commission v French Republic, case 270/83), Saint Gobain, Case C-307/97, Futura, Case C-250/95 and Oy AA, Case C-231/05, rather than the residence state (where, for example, in Marks & Spencer, Case C-446/03, the correct comparison was not between a foreign branch and a foreign subsidiary). This distinction is clearly made in Advocate General Geelhoed's Opinion in Test Claimants in Class IV of the ACT Group Litigation, Case C-374/04, [2007] STC 404 where he says (the footnotes have been renumbered to avoid discontinuous numbering):
    "48. A third restrictive consequence of the fact that direct tax systems are national is the necessity to divide tax jurisdiction over the income of cross-border economic operators (dislocation of tax base). As with disparities, these restrictions should be distinguished from discrimination, as they result not from the rules of just one tax jurisdiction, but from the co-existence of two separate tax jurisdictions (ie, no one tax jurisdiction is to blame for the tax disadvantage). However, unlike disparities, they would continue to exist even if national tax systems were exactly the same in design and content.
    49. The nature of this type of restriction can be explained quite simply. Clearly, the co-existence of national tax systems means that, in order to deal with cross-border economic operators, it is necessary to decide how these systems interact. In particular, a state must choose a criterion by which it decides which part of an economic operator's income falls within their tax jurisdiction. In the present state of international tax law, one of the most important methods of dividing tax jurisdiction is based on the distinction between 'home state' (taxation of residents) and 'source state' (taxation of non-residents) taxation.[3]
    50. In the case of home state taxation, the state of residence of the taxpayer has in principle tax jurisdiction over its total income ('worldwide' taxation). A principal rationale behind this is that the place where the taxpayer uses most facilities (eg, public services, social security, infrastructure etc) is his state of residence. In contrast, in the case of source state taxation, the state of the non-resident has tax jurisdiction only over that part of the non-resident's income that is earned within the source state's territory ('territoriality' taxation). A principal rationale for this is that it is the source state which provides the 'economic opportunity' to earn this income.
    51. As a consequence of this way of dividing tax jurisdiction, an economic operator earning foreign-source income may, in the absence of priority rules between the relevant states, be subject to juridical double taxation. Under international tax law, the generally accepted rule of taxation priority is that of 'source country entitlement': that is, priority of taxation right over source country income lies with the source state. Insofar as juridical double taxation is to be relieved, therefore, this is in principle a matter for the home state, which can choose whether and how it wishes to provide such relief.[4] For example, a state may choose to relieve juridical double taxation unilaterally or by DTC; and using an exemption or credit method.[5] Clearly, therefore, the distinction between residents (home state, worldwide taxation) and non-residents (source state, territorial taxation) is crucial to the current division of tax jurisdiction between states, as reflected in international tax law.
    55. To repeat, where a restriction on freedom of establishment results purely from the co-existence of national tax administrations, disparities between national tax systems, or the division of tax jurisdiction between two tax systems (a quasi-restriction), this should not fall within the scope of art 43 EC. In contrast, 'true' restrictions, that is to say, restrictions to free movement of establishment going beyond those resulting inevitably from the existence of national tax systems, fall under the art 43 EC prohibition unless justified. In the terminology used above, in order to fall under art 43 EC, disadvantageous tax treatment should follow from discrimination resulting from the rules of one jurisdiction, not disparity or division of tax jurisdiction between (two or more) member states' tax systems.
    56. As I recalled above, the court has held that discrimination consists in the 'application of different rules to comparable situations or in the application of the same rule to different situations'.[6]
    57. In my view, it follows as a consequence of the method of dividing tax jurisdiction adopted by member states—that is, the distinction between worldwide (home state) and territorial (source state) tax jurisdiction—that the concept of discrimination applies in different ways to states acting in home state and source state capacity. Quite simply, as the nature of the tax jurisdiction being exercised in each case differs fundamentally, an economic operator subject to home state jurisdiction cannot per se be considered to be in a comparable situation to an economic operator subject to source state jurisdiction, and vice versa. As a result, art 43 EC imposes two different categories of obligation on a state, depending upon the jurisdictional capacity in which it is acting in a particular case.
    i) Article 43 EC home state obligations
    58. The central obligation imposed on states exercising home state jurisdiction is, in essence, to treat foreign-source income of its residents consistently with the way it has divided its tax base. Insofar as it has divided its tax base to include this foreign-source income—ie, by treating it as taxable income—it must not discriminate between foreign-source and domestic income. This principle has been illustrated in the court's case law. Thus, in the case of corporate income taxation for example, the court has held that:
    – Insofar as it chooses to relieve economic double taxation on its residents' dividends, a home state, with tax jurisdiction over the worldwide income of its residents, must provide the same relief for incoming foreign-source dividends as for domestic dividends, and must take foreign corporation tax paid into account for this purpose.[7]
    – Similarly, where a home state offers the possibility to offset domestic losses against domestic prior or future profits, this possibility cannot be denied on the sole ground that the relevant company also earns foreign-source income.[8]
    – In addition, where a state makes group income relief from the obligation to pay ACT available to domestic subsidiaries distributing profits to domestic parents, it must extend this possibility to domestic subsidiaries distributing profits to foreign parents, which subsidiaries would otherwise be subject to ACT.[9]
    59. Conversely, in Marks & Spencer, the court held that, in principle, insofar as a member state does not exercise tax jurisdiction over a non-resident subsidiary of a resident parent company, then it does not have to give loss relief.[10] Put otherwise, if a home state has divided its tax base so that it does not exercise tax jurisdiction over a foreign subsidiary of one of its corporate residents, it is in principle consistent for that state to refuse to take into account deductions relating to that foreign-source income in assessing its resident's tax.
    ii) Article 43 EC source state obligations
    66. As source states have tax jurisdiction only over the income that is earned by the non-resident within the source state's jurisdiction, they are subject to a more limited obligation under art 43 EC. In essence, this can be expressed as an obligation to treat all non-residents in a comparable way to residents (non-discrimination), insofar as these non-residents fall within their tax jurisdiction—ie, subject to the difference in the extent of their tax jurisdiction over residents and non-residents.
    67. In the case of corporate income taxation, this obligation has been applied to mean, for example, that:
    – Tax benefits accorded to resident companies—including those granted pursuant to DTCs[11]—must be accorded in the same way to branches (permanent establishments) of non-resident companies if these branches are otherwise subject to corporation tax in the same way as resident companies.[12] Thus, for example, branches of non-resident companies are entitled to the same imputation credits for dividends received if they are taxed on these dividends in the same way as resident companies.[13]
    – In addition, insofar as a source state exercises tax jurisdiction over a foreign branch, it cannot impose a higher corporate tax rate on this branch than applied to its own resident companies.[14]
    – Nor can a source state subject only outgoing interest repayments on a loan to a resident subsidiary from a controlling non-resident shareholder to, eg, minimum capitalisation requirements (thin capitalisation rules), while not subjecting domestic repayments to any such requirement, unless such a requirement is justified.[15]
    – Conversely, the court has held that the fact that, for the purpose of calculating the basis of assessment to tax for non-resident taxpayers, a source state only takes into account profits and losses arising from their activities in that state—and not, for example, losses arising in their home state—is in no way prohibited by the Treaty.[16]
    – Further, source states should not impose disproportionately heavy administrative or accounting requirements on foreign companies active within their territory (ie, requirements that go beyond what results from the fact that tax administrations are national), including where compliance with these requirements is necessary to avail of a tax benefit relating to source state income.[17]
    69. A further application of the source state non-discrimination obligation is that, insofar as a source state chooses to relieve domestic economic double taxation for its residents (for example, in taxation of dividends), it must extend this relief to non-residents to the extent that similar domestic double economic taxation results from the exercise of its tax jurisdiction over these non-residents (for example, where the source state subjects company profits first to corporation tax and then to income tax upon distribution). This follows from the principle that tax benefits granted by the source state to non-residents should equal those granted to residents insofar as the source state otherwise exercises equal tax jurisdiction over both groups."[18]
    (3) We are dealing here with source state tax obligations for which the principle is, as the Advocate General says in para 66 "…an obligation to treat all non-residents in a comparable way to residents (non-discrimination), insofar as these non-residents fall within their tax jurisdiction." Applied to the present case, the UK branch is treated worse than a UK subsidiary which is not subject to the restriction in s 403D(1)(c). As he says at para 51, "Under international tax law, the generally accepted rule of taxation priority is that of 'source country entitlement': that is, priority of taxation right over source country income lies with the source state." If prevention of double relief for losses is justified, which we deal with below, that is a matter for the residence state, just as it is a matter for the residence state to relieve double taxation of the income of the branch.
    (4) The same distinction was made by Advocate General Maduro in Marks & Spencer, Case C-446/03, [2006] STC 237:
    "47. It is clearly apparent that in all those cases [Avoir fiscal, Saint-Gobain and Royal Bank of Scotland] discrimination in the choice of form of establishment is inextricably bound up with discrimination as to the choice of place of residence. That is owing to the fact that the state concerned chose to place the different forms of establishment on the same footing for the purposes of taxation in its territory. If in such a case a difference of treatment is none the less established it is because it in fact conceals a case of discrimination on the ground of nationality as against the companies operating those establishments.
    48. In the present case foreign subsidiaries and branches are indeed governed by different tax regimes. However, that difference in treatment is not due solely to the fact that they are subject to different tax obligations but to the United Kingdom system of corporate taxation. Under that system the difference in tax treatment is determined by the legal form of the secondary establishment. Groups of companies are not entitled to consolidation for tax purposes which applies to the income of permanent establishments. In that connection, although the group relief system modifies the rule of separate taxation of group companies, it cannot have the effect of assimilating the situation of subsidiaries to that of branches. Under that regime the transfer of losses is treated in a specific way. There is no consolidated joint taxation. That is because subsidiaries are always treated as independent legal and fiscal entities. Accordingly, the difference in treatment of those two categories of establishment does not merely comprise loss of a specific benefit as a result of the option being made in favour of the establishment of foreign subsidiaries. It stems from a difference in the tax regimes applicable to the different types of establishment."
    The source state comparison between a branch and a subsidiary arises because the source state chooses to place them on the same footing; the residence state comparison is not between foreign branches and foreign subsidiaries because they are placed in different tax regimes.
    (5) A similar argument to HMRC's here was raised in Oy AA, Case C-231/05, [2008] STC 991 and rejected by the ECJ:
    "39. A difference in treatment between resident subsidiary companies according to the seat of their parent company constitutes an obstacle to the freedom of establishment if it makes it less attractive for companies established in other member states to exercise that freedom and they may, in consequence, refrain from acquiring, creating or maintaining a subsidiary in the state which adopts that measure (see Lankhorst-Hohorst GmbH v Finanzamt Steinfurt (Case C-324/00) [2003] STC 607, [2002] ECR I-11779, para 32 and the Test Claimants in the Thin Cap Group Litigation case (para 61)).
    40. That conclusion cannot be called into question by the argument of the United Kingdom government that the parent company could have attained the objective pursued by creating a branch in Finland rather than a subsidiary. The second sentence of the first paragraph of art 43 EC expressly leaves traders free to choose the appropriate legal form in which to pursue their activities in another member state and that freedom of choice must not be limited by discriminatory tax provisions (see Commission v France (para 22) and CLT-UFA SA v Finanzamt Köln-West (Case C-253/03) [2007] STC 1303, [2006] ECR I-1831, para 14)."
    Here, it is no answer for HMRC to say that if a subsidiary rather than a branch had been chosen s 403D(1)(c) would not have been applicable.
  29. Mr Ewart contends in outline:
  30. (1) Class IV ACT is not relevant because there the UK was not claiming to tax dividends paid by a UK company either to a resident company or to a non-resident company and so there was no difference in treatment. Much of what the Advocate General said was wide-ranging and not relevant to the case.
    (2) The distinction is not between the source and residence state but whether the parties are objectively in the same situation. In relation to losses a branch and a subsidiary are not in the same situation because of the potential for double use of losses.
    (3) One must concentrate on the restriction in question. There is no blanket restriction on the use of losses, but only a restriction on group relief. In relation to that the situation of a branch and a subsidiary are objectively different because in the branch case the losses are capable of being used twice.
    (4) There is no principle of primacy of source taxation in EU law as there is under the OECD Model tax convention, as is shown by Damseaux, Case C-128/08, which concerned whether Belgium had to give relief for French withholding tax (reduced in amount by the tax treaty) on French dividends in which the ECJ said:
    31      As stated in paragraph 15 above, in the present case, in accordance with the attribution of powers of taxation agreed on by the French Republic and the Kingdom of Belgium, dividends distributed by a company established in France to a Belgian resident are liable to be taxed in both Member States.
    32      In a situation where both the Member State in which the dividends are paid and the Member State in which the shareholder resides are liable to tax those dividends, to consider that it is necessarily for the Member State of residence to prevent that double taxation would amount to granting a priority with respect to the taxation of that type of income to the Member State in which the dividends are paid.
    33      Although such an attribution of powers would comply, in particular, with the rules of international legal practice as reflected in the model tax convention on income and on capital drawn up by the Organisation for Economic Cooperation and Development (OECD), in particular Article 23B thereof, it is not in dispute that Community law, in its current state and in a situation such as that at issue in the main proceedings, does not lay down any general criteria for the attribution of areas of competence between the Member States in relation to the elimination of double taxation within the Community (see Kerckhaert and Morres, paragraph 22, and Columbus Container Services, paragraph 45).
    34      Consequently, if a Member State cannot rely on a bilateral convention in order to avoid the obligations imposed on it by the Treaty (see Case C-170/05 Denkavit Internationaal and Denkavit France [2006] ECR I-11949, paragraph 53, and Amurta, paragraph 55), the fact that both the Member State in which the dividends are paid and the Member State in which the shareholder resides are liable to tax those dividends does not mean that the Member State of residence is obliged, under Community law, to prevent the disadvantages which could arise from the exercise of competence thus attributed by the two Member States.
    35      In those circumstances and to the extent that solely the France-Belgium Convention is the subject of the first question of the referring court, the answer to that question is that, in so far as Community law, in its current state and in a situation such as that at issue in the main proceedings, does not lay down any general criteria for the attribution of areas of competence between the Member States in relation to the elimination of double taxation within the Community, Article 56 EC does not preclude a bilateral tax convention, such as that at issue in the main proceedings, under which dividends distributed by a company established in one Member State to a shareholder residing in another Member State are liable to be taxed in both Member States, and which does not provide that the Member State in which the shareholder resides is unconditionally obliged to prevent the resulting juridical double taxation.
    Since both France, as source state, and Belgium, as residence state, were entitled to tax the dividends Community law does not require elimination of the double taxation.
    Reasons for our decision
  31. We prefer Mr Milne's approach. While much of the analysis of the Advocate General in Test Claimants in Class IV of the ACT Group Litigation may not have been relevant to the ultimate decision, it is a useful distillation of the EU cases and explains clearly what is the correct comparison in particular circumstances. Mr Ewart is correct in saying that the reason for the distinction is that in each case the question is whether the situations are objectively the same, but this will lead one to the same result as the Advocate General's analysis so long as the comparison is made at the right level of generality.
  32. The difference between the parties is the level of generality to be adopted in making the comparison: Mr Milne says it is tax treatment of branches generally (under which they are taxed on the branch profits in the same way as subsidiaries), and Mr Ewart says it is group relief for losses of branches (which are objectively different from subsidiaries because the loss is capable of being used in both states). We consider that Mr Milne's comparison is the correct one. If one narrows down the comparison to the legal provision under consideration there will always be a difference and so on that basis it could always be argued that there is no restriction; the time for looking only at the legal provision under consideration is in connection with justification. The ECJ compares the situation of branches and subsidiaries in cases where this is the right comparison at a high level of generality, as, for example, in Saint-Gobain in which relief for the foreign tax by the head office in France was not considered, and Royal Bank of Scotland, in which the absence of dividend withholding tax on transfer of the branch profits was not considered. Mr Ewart's argument would have more force, although we still do not consider that it would be lead to the result that there was no restriction, if there were limitations on the use of losses of branches in all cases (including the carry-forward or backwards of losses) because of the potential for double use of losses, rather than a limitation only on group relief. We tax a UK branch in exactly the same way as a UK subsidiary (or a separate UK company) so far as the branch profits or losses made in the UK are concerned, except in relation to group relief under s 403D(1)(c). Because we treat them in the same way generally the correct comparison is between the UK branch and a UK company. On this basis, the limitation on group relief is something that affects non-resident companies only and is therefore a restriction.
  33. We do not consider that Damseaux is relevant because the ECJ was looking at the situation in the residence state only. It is authority for saying that the residence state is not obliged by tax treaty to completely eliminate double taxation when there is overlapping taxation in accordance with EU law (worldwide taxation in the residence state, and source taxation in the source state), a proposition that the ECJ has adopted consistently. It does not deny that the source state has the primary right to tax and it is for the residence state to deal with any double taxation if it wishes to do so: "Article 56 EC does not preclude a bilateral tax convention…which does not provide that the Member State in which the shareholder resides is unconditionally obliged to prevent the resulting juridical double taxation." If there were no source state priority the issue would not be whether the residence state had to eliminate the double taxation but whether either state was obliged to do so. More relevantly the ECJ adopted the principle of the primary use of the losses in the permanent establishment state, Luxembourg, in Lidl, Case C-414/06, in deciding that the head office state, Germany, was justified in not giving relief for the losses since it did not tax the profits of the permanent establishment. The analogy here is that if the double use of losses is to be restricted it is for the residence state to do so if that is in accordance with the balanced exercise of taxing power, which it was in the case of Germany.
  34. Issue 3: whether the restrictions are justified
    Section 406(2)
  35. A restriction on a fundamental freedom may nevertheless be lawful if it is "justified by overriding reasons of public interest" (see, for example, Lidl Belgium GmbH & Co KG v Finamzamt Heilbronn (Case c-414/06) [2008] STC 3229 at p 3427, para 27).
  36. In the context of s 406(2) we consider here the possible justification based on fiscal coherence of the tax system. This justification has been rarely applied by the ECJ, and essentially requires that the advantage in question is directly linked to an offsetting levy or deduction (see, for example, Bachmann v Belgium (Case C-204/90) [1994] STC 855; Manninen (Case C-319/02) [2004] STC 1444; and Finanzamt Offenbach am Main-Land v Keller Holding GmbH (Case 471/04) [2007] STC 962).
  37. A recent case in which fiscal coherence was held to be a justification was Papillon, to which we have referred above. As we have described, that case concerned the refusal by the French tax authorities to extend the benefit of the French group integration regime to a French parent company and its French sub-subsidiary which was held through a Netherlands subsidiary. A feature of the French regime was that, in a domestic situation, the possibility of double relief for losses, once by consolidating the results of the French parent and a French subsidiary, and again by allowing a provision by the parent for depreciation of its holding in the subsidiary, was prevented by neutralising the depreciation provision. That neutralisation was not possible in the case of the Netherlands subsidiary because, as a non-French resident company, it was not itself subject to the integration regime. Since any depreciation in value of the Dutch subsidiary would reflect the depreciation in value of the French sub-subsidiary there would be double relief for losses in France if the consolidation of the sub-subsidiary were allowed. The French government accordingly argued that the restriction under its law was necessary to maintain the coherence of its system of tax integration.
  38. In giving her opinion, Advocate General Kokott referred to the restrictive nature of the justification and to the need to establish a link between the advantage concerned and the offsetting of that advantage by a particular tax levy. Coherence must first and foremost be adjudged in the light of the aim and logic of the tax regime at issue. The Advocate General then considered the particular sections of the French tax integration regime, including the neutralisation of certain transactions, and concluded (at paras 57 to 60, pp 555 to 556):
  39. "57. At this point it must be noted that the purpose and the system of the French legislation on 'tax integration' requires that certain intra-group transactions be neutralised before the consolidation of results. However, is there a direct link between the tax advantage of group taxation and the offsetting of that tax advantage by a tax deduction within the meaning of the case law of the Court of Justice?
    58. In the Bachmann judgment, the Court of Justice found a direct link between the tax deductibility of insurance premiums and the taxation of sums payable by the insurers under insurance contracts. Since Mr Bachmann had concluded an insurance contract in Germany pursuant to which payments made by the insurers would not be subject to taxation in Belgium, the refusal of the Belgian tax authorities to allow the deduction of his premiums paid pursuant to the insurance contracts was justified.
    59. In the cases decided so far, in which the coherence of the tax system was the decisive justification, the disadvantage in question was a compensatory levy, in other words a tax. In the present case the disadvantage lies in the neutralisation of certain intra-group transactions. In the strict sense, this cannot be viewed as the imposition of tax, because neutralisation measures do not constitute the levying of a tax. Nevertheless, they are to be seen as a tax disadvantage, because certain measures, namely the provision for the loss of share value, in contrast to the individual assessment of a company, are not taken into consideration. There is also a direct link between the provisions on neutralisation and the tax integration, because a consolidation of the results of the various companies will only be accorded if it can at the same time be assured that the objective pursued is not distorted by the use of losses more than once.
    60. Thus, a measure, which seeks to avoid the use of losses more than once, may be justified in order to maintain the coherence of the tax system."
  40. In its judgment in Papillon the ECJ adopted the same approach as the Advocate General, holding that the necessary direct link existed under the tax integration regime between the tax advantages and the neutralisation of intra-group transactions, and that this link would be eliminated, and affect the coherence of that regime if the restriction were not to apply. The ECJ concluded:
  41. "51      Consequently, in refusing to extend the benefit of the tax integration regime to a resident parent company wishing to include its resident sub-subsidiaries in that regime where it holds those sub-subsidiaries through a non-resident subsidiary, the provisions of the CGI at issue in the main proceedings have the effect of ensuring the coherence of that regime."
  42. Mr Milne argued that, looking at the UK's domestic system, there was no risk that UK losses would be used more than once in the UK and that accordingly the coherence of the tax system could not be used as a justification for the UK restrictions. He referred to s 411(1) of the Taxes Act which provides that:
  43. "Relief shall not be given more than once in respect of the same amount, whether by giving group relief and by giving some other relief (in any accounting period) to the surrendering company, or by giving group relief more than once."
  44. Mr Ewart's and Mr Yates' skeleton argument asserted that the restriction imposed by s 406(2) could be justified on the grounds of maintaining the coherence of the UK's tax system, but this was not expanded upon in argument before us. We agree with Mr Milne on this issue. The position under the UK's group relief rules is quite unlike that of the French tax integration system. There can be no justification based on fiscal coherence for a provision that precludes relief for UK losses in a case where, unlike as in France, there is no elimination of a link between the tax advantage of the loss relief and the neutralisation of a further advantage: s 411 applies irrespective of whether the structure includes a non-UK resident company. The coherence of the UK's group relief system is preserved, and the restriction created by s 406(2) is neither necessary to maintain coherence nor, as a result, justified.
  45. Section 403D(1)(c)
  46. Mr Ewart submitted that s 403D(1)(c) is justified by the three factors described in Marks & Spencer, at paras 42 to 51 of the ECJ's judgment (at p266-267):
  47. "42. … the United Kingdom and the other member states which submitted observations put forward three factors to justify the restriction.
    43. First, in tax matters profits and losses are two sides of the same coin and must be treated symmetrically in the same tax system in order to protect a balanced allocation of the power to impose taxes between the different member states concerned. Second, if the losses were taken into consideration in the parent company's member state they might well be taken into account twice. Third, and last, if the losses were not taken into account in the member state in which the subsidiary is established there would be a risk of tax avoidance.
    44. As regards the first justification, it must be borne in mind that the reduction in tax revenue cannot be regarded as an overriding reason in the public interest which may be relied on to justify a measure which is in principle contrary to a fundamental freedom (see, in particular, Proceedings brought by Manninen (Case C-319/02, [2004] STC 1444, para 49 and the case law cited).
    45. None the less, as the United Kingdom rightly observes, the preservation of the allocation of the power to impose taxes between member states might make it necessary to apply to the economic activities of companies established in one of those states only the tax rules of that state in respect of both profits and losses.
    46. In effect, to give companies the option to have their losses taken into account in the member state in which they are established or in another member state would significantly jeopardise a balanced allocation of the power to impose taxes between member states, as the taxable basis would be increased in the first state and reduced in the second to the extent of the losses transferred.
    47. As regards the second justification, relating to the danger that losses would be used twice, it must be accepted that member states must be able to prevent that from occurring.
    48. Such a danger does in fact exist if group relief is extended to the losses of non-resident subsidiaries. It is avoided by a rule which precludes relief in respect of those losses.
    49. As regards, last, the third justification, relating to the risk of tax avoidance, it must be accepted that the possibility of transferring the losses incurred by a non-resident company to a resident company entails the risk that within a group of companies losses will be transferred to companies established in the member states which apply the highest rates of taxation and in which the tax value of the losses is therefore the highest.
    50. To exclude group relief for losses incurred by non-resident subsidiaries prevents such practices, which may be inspired by the realisation that the rates of taxation applied in the various member states vary significantly.
    51. In the light of those three justifications, taken together, it must be observed that restrictive provisions such as those at issue in the main proceedings pursue legitimate objectives which are compatible with the Treaty and constitute overriding reasons in the public interest and that they are apt to ensure the attainment of those objectives."

    Mr Ewart referred in particular to paras 47 and 48. He argued that the statements in these paragraphs made it clear that Member States must be able to prevent the occurrence of double losses of the nature that might arise absent s 403D(1)(c). He submitted that where a measure is targeted at preventing losses being used twice, it must be correct that such a measure may be justified on those grounds.

  48. Mr Milne referred us first to the principle of territoriality. This principle, he said, was relevant both to the question whether there exists a legitimate overriding public interest objective in this case, and also how such public interest objective is properly to be formulated and understood. The core principle is that the state in which the profits are made, or losses are incurred ("the source state") has primary taxing competence in respect of such profits or losses. This, Mr Milne submitted, was the effect of the ECJ cases of Futura, Marks & Spencer, Oy AA and Lidl. He argued that it follows that once a source state decides to tax branches of foreign companies in its territory on the same basis as it taxes the profits of its own resident companies, it is accepting that there is no distinction between such companies and branches and it must treat them the same, above all in granting relief for losses. This is a case for symmetry which, whilst it may not be everything, is nevertheless extremely important.
  49. Mr Milne argued that context is also a vital consideration. As the Court in Marks & Spencer made clear (para 38 at p 266):
  50. "In each specific situation, it is necessary to consider whether the fact that a tax advantage is available solely to resident taxpayers is based on relevant objective elements apt to justify the difference in treatment."
  51. In Marks & Spencer, as we have seen, the restrictive provisions (on group relief for losses of non-UK subsidiaries) were held to pursue legitimate objectives that were both compatible with the EU Treaty and constituted overriding reasons in the public interest. This was in the light of the three justifications taken together (emphasis supplied). In considering the application of these tests in this case we need therefore to determine whether each of the tests is capable of operating independently (which Mr Ewart essentially contends) or if (as Mr Milne argues) the justification must in all cases be based on the balanced allocation of taxing rights, with or without one of the other categories (prevention of double use of losses, and tax avoidance) referred to in Marks & Spencer.
  52. The case of Proceedings brought by Oy AA (Case C-231/05) [2008] STC 991 was a source state case, like that of the Taxpayer. In that case a tax deduction was available in Finland on an intra-group financial transfer (a form of subvention payment) made by a subsidiary to its parent, but only if both parent and subsidiary were established in Finland. The group in question was headed by a UK parent company that had itself no establishment in Finland. It was held that the restriction was justified by a combination of the balanced allocation of taxing powers and the need to prevent tax avoidance. Advocate General Kokott said at paras 46 to 65 (pp 1003-1005) (we have renumbered the footnotes as before, and if the citation of a case is cross-referenced to a footnote that we have not reproduced we have added the citation in the footnotes below):
  53. "46. In the Marks & Spencer case the court recognised three grounds of justification (éléments justificatifs) and concluded from them—
    'taken together … that restrictive provisions such as those at issue in the main proceedings pursue legitimate objectives which are compatible with the Treaty and constitute overriding reasons in the public interest and that they are apt to ensure the attainment of those objectives.'[19]
    47. The three justifications it recognised were:
    —preservation of the allocation of the power to impose taxes between member states;
    —the danger that losses would be used twice; and
    —the risk of tax avoidance.
    48. The formulation cited above already makes it clear that all three elements are closely linked to one another and cannot be viewed in isolation. In this connection preserving the allocation of the power to impose taxes is at the heart of these elements.
    49. As already mentioned above, at the current stage of development of Community law power to impose direct taxes lies with the member states.[20] In the absence of harmonisation at Community law level it is likewise a matter for the member states to lay down criteria for allocating their powers to impose taxes by the conclusion of double taxation conventions or by unilateral measures.[21]
    50. For the purposes of the allocation of powers of taxation, it is not unreasonable for the member states to find inspiration in international practice and the model conventions drawn up by the Organisation for Economic Cooperation and Development (OECD).[22]
    51. By taxing resident companies on their worldwide profits and non-resident companies solely on the profits from their activities in that state, the parent company's member state is acting in accordance with the principle of territoriality enshrined in international tax law and recognised by Community law.[23]
    52. In the Marks & Spencer case the court found that the allocation of power to impose taxes had been jeopardised as follows:
    '[T]o give companies the option to have their losses taken into account in the member state in which they are established or in another member state would significantly jeopardise a balanced allocation of the power to impose taxes between member states, as the taxable basis would be increased in the first state and reduced in the second to the extent of the losses transferred.'[24]
    53. That finding applies mutatis mutandis to an extension of the rules in the Law on Intra-group Financial Transfers to cross-border transfers. It would likewise undermine the allocation of powers to impose taxes according to the principle of territoriality if taxpayers had a free choice as to the member state in which their profits should be taxed, by extracting a company's profits from its tax basis and adding them to the tax basis of a group company established in a different member state.
    54. The second element of justification recognised in the Marks & Spencer case, namely preventing the danger that losses are used twice, is closely connected to the allocation of the power to impose taxes.
    55. The allocation of power to impose taxes on the basis of elements of territoriality (an undertaking's residence or source of income within the territory) serves to confer on a state a primary right to tax certain income. This, taken together with the rules to prevent double taxation, creates an international system of tax competence. In principle, and even if not without lacunae in particular cases, this system is intended to ensure that all income is taxed, and taxed only once.[25]
    56. If it were possible for losses to be set off against profits more than once, some profits would not be taxed at all, notwithstanding that in fact there were no corresponding deductible losses. This would infringe the principle of once-only taxation, which is fundamental to the international allocation of the power to impose taxes.
    57. Recognising cross-border intra-group transfers could also lead to double non-taxation of income if the transfers could be deducted from the taxable profits of the transferor company notwithstanding that they were not taxable in the country in which the recipient company had its seat.
    58. According to the observations submitted by the United Kingdom government, that might actually happen in the circumstances of the main proceedings. That is because under United Kingdom law intra-group transfers are not taxable income, and accordingly could not be taxed in the hands of the recipient company, AA Ltd, although that is disputed by Oy AA.
    59. Irrespective of the actual treatment of intra-group transfers to British companies, it should be held that a rule which generally provides that only intra-group transfers between Finnish companies are deductible for tax purposes is, in theory, apt to prevent losses from being used twice. This is because so far as purely internal transfers are concerned the member state can ensure that the transfers are subject to tax. The further question remains to be considered as to whether the measure in this form goes no further than is necessary to achieve its purpose.[26]
    60. The risk of tax avoidance as the third element of justification is also closely linked to the other two elements of justification. One might regard intra-group transfers to companies resident in member states in which such payments are not taxable in itself as tax avoidance. To that extent this justification may be considered together with the second justification.
    61. Moreover, in the Marks & Spencer case the court regarded as tax avoidance the 'escape' by undertakings to the state in which losses had their highest value and thus gave the largest reduction in tax liability, by transferring losses from low taxation to high taxation jurisdictions. It is likewise to be regarded as tax avoidance in this sense for income to be deliberately transferred, by means of intra-group transfers, to companies resident in low taxation jurisdictions.
    62. Strictly speaking, prevention of this form of 'tax avoidance' is not a separate ground of justification which can justify a restriction on freedom of establishment. The fact that undertakings seek to profit from the differences between national tax systems is a legitimate form of economic conduct, and is indeed inevitable in an internal market in which taxation of corporations is not harmonised. Accordingly, an undertaking cannot simply be prevented from moving its seat to another member state which offers more favourable conditions of taxation.[27]
    63. Restrictions on the fundamental freedoms can be justified only if such 'tax optimisation' also undermines the allocation of powers to impose taxes between member states.
    64. Recognising the allocation of powers to impose taxes according to the principle of territoriality as a ground of justification is not inconsistent with the principle that restrictions on the fundamental freedoms cannot be justified by reference to the purpose of preventing a reduction in tax revenue.[28] This principle means simply that fundamental freedoms cannot be restricted on account of purely fiscal considerations. By contrast, the present case concerns the fundamental interest of granting the member states a right at all to impose taxes according to the principle of territoriality.
    65. In summary, it should be held that restricting the deductibility of intra-group transfers to transfers to Finnish companies is apt to safeguard the allocation of powers to impose taxes between member states, to exclude the possibility that income which is transferred is not taxed, and to combat tax avoidance. It ensures that profits earned by group companies in Finland are subject to taxation there according to the principle of territoriality."
  54. Mr Milne argued that it was not permissible for a Member State to rely on any one of the three justifications in isolation, and that the real test must be based on the allocation of taxing powers. However, he recognised that the views of Advocate General Kokott in para 48 were not universally shared. Mr Milne also referred us to the ECJ judgment in Oy AA, from which we take in particular the following paragraphs (at p1013):
  55. "51. As is apparent from para 51 of the judgment in the Marks & Spencer case, the need to safeguard the balanced allocation of the power to impose taxes between the member states was accepted by the court in conjunction with two other grounds of justification, based on the risks of the double use of losses and of tax avoidance (see also Rewe Zentralfinanz eG v Finanzamt Koln-Mitte (Case C-347/04) [2007] ECR I-2647, [2007] 2 CMLR 1111, para 41).
    52. It should also be remembered that, in the absence of any unifying or harmonising Community measures, member states retain the power to define, by treaty or unilaterally, the criteria for allocating their powers of taxation (see Gilly v Directeur des Services Fiscaux du Bas-Rhin (Case C-336/96) [1998] STC 1014, [1998] ECR I-2793, 1 ITLR 29, paras 24, 30, N v Inspecteur van de Belastingdienst Oos/kantoor Almelo (Case C-470/04) [2006] ECR I-7409, [2006] 3 CMLR 49, para 44, Kerckhaert v Belgium (Case C-513/04) [2007] STC 1349, [2007] 1 WLR 1685, paras 22, 23 and the Test Claimants in the Thin Cap Group Litigation case (para 49)).
    53. Concerning, first, the need to safeguard a balanced allocation of the power to tax between member states, it should be pointed out that that need cannot justify a member state systematically refusing to grant a tax advantage to a resident subsidiary, on the ground that the income of the parent company, having its establishment in another member state, is not capable of being taxed in the first member state (see, to that effect, the Rewe Zentralfinanz case (para 43)).
    54. That element of justification may be allowed, however, where the system in question is designed to prevent conduct capable of jeopardising the right of the member states to exercise their taxing powers in relation to activities carried on in their territory (see the Rewe Zentralfinanz case (para 42)).
  56. Lidl was a residence state case where under German law losses from a Luxembourg permanent establishment were disallowed in Germany on the footing that corresponding income from that permanent establishment was not subject to tax in Germany under the double tax convention between Germany and Luxembourg. In her Opinion, after referring to Marks & Spencer, Advocate General Sharpston continued (paras 10 to 18 at pp 3233-3235) (footnotes renumbered):
  57. "10. In that case the court prefaced its discussion of the three justifications by stating that, in order to ascertain whether restricting group relief to losses incurred by resident companies was justified, it was 'necessary to consider what the consequences would be if an advantage such as [group relief] were to be extended unconditionally'.[29] It was in that context that the court considered, and accepted (subject to two provisos reflecting the need for proportionality), the three justifications put forward by the United Kingdom and the other member states which had submitted observations. I see no reason for not adopting the same approach when assessing whether a measure limiting relief for losses incurred by a permanent establishment of a resident company to those losses incurred by its resident permanent establishments is compatible with the Treaty. From the point of view of the company, the ability to deduct losses of a foreign subsidiary by way of group relief is clearly analogous to the ability to deduct losses of a foreign permanent establishment. Indeed, the purpose of group relief is to avoid penalising companies which, rather than establishing branches, decide to expand their activities by setting up subsidiaries.[30]
    11. Whether any or all[31] of the three justifications may apply in a given situation involving a permanent establishment rather than a subsidiary is a different issue.
    12. With regard to the first justification, it seems to me that the preservation of the allocation of the power to impose taxes between member states might mean that the economic activities of the permanent establishment in one member state of a company established in another should be subject only to the tax rules of the first state in respect of both profits and losses.[32] The member state concerned in Marks & Spencer (the United Kingdom) did not assert taxing rights over foreign subsidiaries of resident companies. In the present case, the member state concerned would, in the absence of the DTC, have the right to tax a cross-border permanent establishment but has waived that right by exempting profits of such an establishment from tax. The effect is the same. As the court stated in Marks & Spencer, 'to give companies the option to have their losses taken into account in the member state in which they are established or in another member state would significantly jeopardise a balanced allocation of the power to impose taxes between member states, as the taxable basis would be increased in the first state and reduced in the second to the extent of the losses transferred'.[33] Similarly in the present case it would entail a breach of the symmetry between taxation and reliefs agreed by Germany and Luxembourg and reflected in the DTC if Germany were to grant relief for losses incurred by a Luxembourg permanent establishment of a German company where it had waived the right to tax profits made by such an establishment.
    13. With regard to the second justification, the court in Marks & Spencer stated that member states must be able to prevent the danger that losses might be used twice, that that danger exists if group relief is extended to the losses of non-resident subsidiaries and that it is avoided by a rule which precludes relief in respect of those losses.[34]
    14. Again, I do not see why that principle should not apply in a case such as the present. There is clearly scope for the same loss to be used twice. If anything that risk is greater in the context of losses made by a permanent establishment which remains in business than in the context (as in Marks & Spencer) of losses made by subsidiaries which had been sold to third parties or ceased trading. In the former context, if the losses could be used in the member state where the company was resident there is a risk that they would be claimed again in the member state of the permanent establishment once the permanent establishment became profitable, without the member state where the company was resident being able to recoup the benefit obtained. Indeed it appears from the reply from the representative of Lidl to a question put by the Judge Rapporteur at the hearing in the present case that the loss which prompted the reference has now been set off against profits made in Luxembourg. [35] As the court stated in Marks & Spencer, member states must be able to tackle the risk that losses will be used twice, and that risk is avoided by a rule which precludes relief in respect of those losses.[36]
    15. With regard to the third justification, relating to the risk of tax avoidance, the court stated in Marks & Spencer that the possibility of transferring the losses incurred by a non-resident company to a resident company entails the risk that within a group of companies losses will be transferred to companies established in the member states which apply the highest rates of taxation and in which the tax value of the losses is therefore the highest. In contrast to the situation concerning subsidiaries, where it is conceivable that the transfer of losses will be systematically organised within groups of companies and losses directed solely to companies of the group established in member states with higher rates of taxation,[37] in the case of permanent establishments there is clearly no scope for such 'jurisdiction shopping' if a company in one member state is allowed to deduct losses made by a permanent establishment in another member state. That is because, as the Commission submits, measures enabling the 'transfer' of losses from a permanent establishment, as opposed to a subsidiary, are neither necessary nor theoretically conceivable: losses made by a permanent establishment are losses of the taxpayer 'parent' company. Losses made by national permanent establishments can accordingly be directly and immediately deducted. The only difference between that domestic situation and the present case lies in the allocation of the power to impose taxes: if the permanent establishment is resident in another member state with which there is a double taxation convention such as that at issue in the present case, that state will have the exclusive right to tax the income of the permanent establishment in question. That does not affect the fact that a loss accrues to the taxpayer company, directly reducing its taxable income in its state of residence. The other state's right to tax simply creates a second connecting factor for the same loss and, therefore, the potential for double deduction. In cases such as the present, therefore, it seems to me that the third Marks & Spencer justification adds nothing to the first and second grounds.
    16. I accordingly agree with, variously, Finland, France, Germany, Greece, the Netherlands, Sweden, the United Kingdom and the Commission that national legislation restricting the possibility for a company to deduct losses made by a permanent establishment in another member state may be justified by, first, the need to preserve the balanced allocation of the power to impose taxes between the different member states concerned and, second, the risk that losses might be taken into account twice.
    17. The Bundesfinanzhof notes in the order for reference that it is uncertain whether the court intended the three justifications in Marks & Spencer to be cumulative, so that all three must be present.
    18. Admittedly the court in Marks & Spencer referred to the 'three justifications, taken together', which supports that view. In N v Inspecteur van de Belastingdienst Oos/kantoor Almelo (Case C-470/04) [2008] STC 436, [2006] ECR I-7409, para 42, however, the court stated that 'preserving the allocation of the power to tax between member states is a legitimate objective' without reference to other grounds. In AA, the court accepted that legislation of a member state which precludes a subsidiary resident in that member state from deducting a financial transfer to its parent company from its taxable income unless that parent company is established in the same member state could in principle be justified on the basis of two of the three Marks & Spencer grounds, namely the first and the third.[38] And in Amurta the court, having stated that the second and third Marks & Spencer justifications had not been pleaded, none the less went on to consider (and reject) an argument based on the need to safeguard the balanced allocation between the member states of the power to tax.[39] It seems, therefore, that the three justifications do not necessarily all have to be applicable in a given case. In my view the legislation at issue in the present case could in principle be justified by reference to the first and second Marks & Spencer grounds, namely the objectives of preserving the balanced allocation of the power to impose taxes and of avoiding the danger that losses would be used twice."
  58. In its judgment in Lidl, the ECJ considered the three justifications described in Marks & Spencer and said (at paras 27 to 42, pp 3243-3244):
  59. "27. It is clear from the court's case law that a restriction on the freedom of establishment is permissible only if it is justified by overriding reasons in the public interest. It is further necessary, in such a case, that its application be appropriate to ensuring the attainment of the objective in question and not go beyond what is necessary to attain it (see Marks & Spencer plc v Halsey (Inspector of Taxes) (Case C-446/03) [2006] STC 237, [2005] ECR I-10837, para 35; Cadbury Schweppes and Cadbury Schweppes Overseas (para 47); and Test Claimants in the Thin Cap Group Litigation (para 64)).
    28. In that connection, the national court points out in particular that, in accordance with the provisions of the convention, the income generated by the permanent establishment in Luxembourg is not taxed in the member state of residence of the company to which it belongs.
    29. In their observations to the court, the German, Greek, French, Netherlands, Finnish, Swedish and United Kingdom governments essentially submit that tax rules which, for the purposes of determining the tax base of a resident company, restrict the possibility for that company to deduct losses incurred by a permanent establishment belonging to it which is situated in another member state can, as a rule, be justified.
    30. Those governments take the view that the justification for such rules in the light of Community law may be based, first, on the need to preserve the allocation of the power to impose taxes between the member states concerned and, secondly, on the need to prevent the danger that losses may be taken into account twice.
    31. As regards the first of these justifications, it should be noted that the preservation of the allocation of the power to impose taxes between member states may make it necessary to apply to the economic activities of companies established in one of those states only the tax rules of that state in respect of both profits and losses (see Marks & Spencer (para 45), and Proceedings brought by Oy AA (Case C-231/05) [2008] STC 991, [2007] ECR I-6373, para 54).
    32. To give companies the right to elect to have their losses taken into account in the member state in which they are established or in another member state would seriously undermine a balanced allocation of the power to impose taxes between the member states, since the tax base would be increased in the first state, and reduced in the second, by the amount of the losses surrendered (see Marks & Spencer (para 46), and Oy AA (para 55)).
    33. With respect to the relevance of the first of these justifications in the light of the facts in the main proceedings, it should be pointed out that the member state in which the registered office of the company to which the permanent establishment belongs is situated would, in the absence of a double taxation convention, have the right to tax the profits generated by such an entity. Consequently, the objective of preserving the allocation of the power to impose taxes between the two member states concerned, which is reflected in the provisions of the convention, is capable of justifying the tax regime at issue in the main proceedings, since it safeguards symmetry between the right to tax profits and the right to deduct losses.
    34. In circumstances such as those of the main proceedings, to accept that the losses of a non-resident permanent establishment might be deducted from the taxable income of the principal company would result in allowing that company to choose freely the member state in which those losses could be deducted (see, to that effect, Oy AA (para 56)).
    35. As regards the second justification put forward in the observations submitted to the court, which is based on the danger that losses might be taken into account twice, the court has accepted that the member states must be able to prevent such a danger (see Marks & Spencer (para 47), and Rewe Zentralfinanz eG v Finanzamt Koln-Mitte (Case C-347/04) [2008] STC 2785, [2007] ECR I-2647, para 47).
    36. In this connection, it must be pointed out that, in circumstances such as those which underlie the main proceedings, there is clearly a danger that the same losses will be used twice (see Marks & Spencer (para 48)). It is possible that a company might deduct, in the member state in which its seat is situated, losses incurred by a permanent establishment belonging to it and situated in another member state and that, despite such offsetting, the same losses might be taken into account subsequently in the member state in which the permanent establishment is situated, when that establishment generates profits, thereby preventing the member state in which the principal company has its seat from taxing that profit.
    37. Consequently, the two justifications put forward must each be considered as being capable of justifying a restriction on the freedom of establishment arising from the tax treatment by the member state in which the seat of a company is located of losses incurred by a permanent establishment belonging to that company and situated in another member state.
    38. The national court asks, however, whether the justifications set out in paras 44 to 50 of the judgment in Marks & Spencer, which also include the need to prevent the risk of tax avoidance, must be understood as being cumulative or whether the existence of only one of those factors is sufficient for the tax regime at issue in the main proceedings to be treated, in principle, as being justified.
    39. It must be pointed out in that regard that, in para 51 of the judgment in Marks & Spencer, the court held that the three justifications taken together, which underlay the legislation at issue in the main proceedings, pursued legitimate objectives compatible with the Treaty and thus constituted overriding reasons in the public interest
    40. However, bearing in mind the wide variety of situations in which a member state may put forward such reasons, it cannot be necessary for all the justifications referred to in para 51 of the Marks & Spencer judgment to be present in order for national tax rules which restrict the freedom of establishment laid down in art 43 EC to be capable, in principle, of being justified.
    41. Thus, in the judgment in Oy AA, the court acknowledged in particular that the national tax legislation at issue could, in principle, be justified on the basis of two of the three justifications referred to in para 51 of the judgment in Marks & Spencer, namely the need to safeguard the allocation of the power to tax between the member states and the need to prevent tax avoidance, taken together (see Oy AA (para 60)).
    42. Likewise, the tax regime at issue in the main proceedings can, in principle, be justified in the light of two of the factors referred to in para 51 of the judgment in Marks & Spencer, namely the need to safeguard the allocation of the power to tax between the member states and the need to prevent the danger that the same losses will be taken into account twice."
  60. In Rewe Zentralfinanz eG (as universal legal successor of ITS Reisen GmbH) v Finanzamt Köln-Mitte (Case C-347/04) [2008] STC 2785, a German parent company had established a holding company in the Netherlands with subsidiaries in the UK and Spain. The German tax authorities disallowed write downs in the German parent company to the value of the Netherlands holding company. This was on the basis that such write downs could be allowed in the case of a subsidiary established in another Member State only where, broadly, the negative income stemming from the write downs was offset by positive income from that Member State. Mr Milne, after reminding us that this was a residence state case, took us first to the opinion of Advocate General Maduro (paras 24 to 38, pp 2792-6):
  61. "24. The German government puts forward a number of legal arguments in support of the contested measure, based on considerations at once political (the need to ensure the balanced allocation of the power to impose taxes), ethical (the risk of tax evasion through losses being taken into account twice and the risk of tax avoidance), administrative (the need to ensure effective supervision), systemic (the need to ensure the uniformity of the tax system) and economic (the risk of budgetary losses). It claims to find support for these arguments in the judgment in Marks & Spencer [2006] STC 237, [2005] ECR I-10837. The analysis of this reasoning should therefore begin with reference to that case law.
    1. The balanced allocation of the power to impose taxes between member states
    25. In the judgment in Marks & Spencer, the court (at para 46) accepted for the first time that the principle of the balanced allocation of the power to impose taxes between member states should be taken into account for the purpose of determining whether tax legislation is compatible with the fundamental freedoms. However, it was also careful to specify and delimit the conditions under which that principle applies.
    26. In the first place, such a requirement is relevant only at the stage of justifying the restrictive measure at issue. It cannot be relied on, as the German government does in the present case, at the stage when a restriction on freedom of establishment is analysed. In the second place, the justification in question has a meaning, in the Community context, that calls for careful delimitation.
    27. In this regard, the German government appears to acknowledge that that requirement could enable certain forms of discrimination to be excluded from the scope of the freedoms of movement. Its argument is based on a rule of symmetry between the right to tax a company's profits and the duty to take that company's losses into account. Since, in tax matters, profits and losses are two sides of the same coin, the German tax authorities should not have, when assessing the liability of a parent company resident in German territory to tax, to take account of losses incurred in connection with the activity of a subsidiary established in another member state, as they are not entitled to tax the profits of that subsidiary. Only such a rule on allocation would secure compliance with the sovereignty accorded to member states in matters of taxation and with the rules of international tax law.
    28. This definition of the requirement of the balanced allocation of the power to impose taxes cannot be accepted. Considered in that light, it does not differ materially from a purely economic justification. Such an interpretation would enable a member state to refuse as a matter of course to grant a tax advantage to an undertaking on the ground that it had developed a cross-border economic activity which was unlikely to generate tax revenue in that state. Put in that form, such a justification was, moreover, expressly rejected by the court in the judgment in Marks & Spencer. The court held, with regard to that specific justification, that it must be borne in mind that the reduction in tax revenue cannot be regarded as an overriding reason in the public interest which may be relied on to justify a measure which is in principle contrary to a fundamental freedom.[40]
    29. The principle that member states are free to determine the organisation of their tax system and to allocate their powers of taxation amongst themselves, must of course be taken into account.[41] However, it cannot be denied that the fundamental freedoms impose certain constraints on the member states in the exercise of their powers in this area. Those constraints essentially require compliance with the obligation not to place taxpayers exercising a cross-border activity at a disadvantage when compared with national taxpayers, even if to do so entails a loss of tax revenue for the state concerned.
    30. This view was upheld by the court inter alia in the judgment in Bosal [2003] STC 1483, [2003] ECR I-9409. In that case, the court held that the Treaty precludes a national provision which makes the deductibility of expenditure related to a Netherlands parent company's shareholding in a subsidiary established in another member state subject to the condition that such costs be indirectly instrumental in making profits which are taxable in the Netherlands. That view has been challenged on the ground that it is contrary to the principle of the fair allocation of the member states' power to impose taxes. As the expenditure incurred by the Netherlands parent company was economically connected with the profits made by its subsidiary established in another member state, it would have been legally more consistent to regard those costs as foreign expenditure that could be taken into account only in the state in which the profits were made.[42] However, such an analysis does not take sufficient account of the situation of a Community national operating in the wider context of the internal market. That situation must not be considered differently according to the territory in which the tax is imposed in each case; it must be assessed globally. From that point of view, it is clear that a difference in the taxation of parent companies according to whether or not they have foreign subsidiaries cannot be justified on the ground that they have transferred economic resources to a part of the territory of the European Union in which the state in question cannot exercise its power to impose taxes. To take any other view, in the absence of common rules on the subject, would be tantamount to depriving the fundamental freedoms enshrined in the Treaty of all practical effect.
    31. Moreover, if the argument of symmetry advanced by the German government were to be accepted in the area of taxation, there is no apparent reason why it should not be extended to the other areas covered by the rules on the freedoms of movement. Just as the principle of the allocation of the power to impose taxes could be invoked, it would then be possible to rely generally on a principle of allocation of the power to legislate. On that principle, a member state would be entitled to refuse to take into account cross-border economic situations that might call into question its freedom to legislate. Thus, for example, goods lawfully produced in accordance with conditions imposed by another member state could be refused entry to a national market on the ground that the goods in question did not meet the legal conditions obtaining in that market. The free movement of goods would then be reduced to a purely formal rule of non-discrimination, consisting of according equal treatment only to goods subject to the rules of the state concerned. Such a result would be completely contrary to the settled case law of the court on the subject.[43]
    32. That cannot therefore be the scope that should be accorded in the Community context to the legitimate requirement of a balanced allocation of the power to impose taxes. While the court accepted, in the judgment in Marks & Spencer [2006] STC 237, [2005] ECR I-10837, a justification based on that requirement, it was only in relation to the risk of abuse or fraud that might arise, in certain cases, from inadequate co-ordination of the member states' powers to impose taxes. In the absence of harmonisation of tax legislation, there is reason to fear that the exercise of free movement might give rise to a veritable 'trade in losses' at Community level. As the court noted in that judgment (at para 46) 'to give companies the option to have their losses taken into account in the member state in which they are established or in another member state would significantly jeopardise a balanced allocation of the power to impose taxes between member states, as the taxable basis would be increased in the first state and reduced in the second to the extent of the losses transferred'. Economic operators would then be free to surrender their losses to companies established in member states with the highest rates of taxation, where the tax value of the losses would accordingly be greatest. Such a situation might call into question the neutrality which Community law is required to maintain with regard to national tax systems.[44]
    33. In accordance with this principle of neutrality, the right of establishment may not be used by economic operators to gain advantages that are not connected with the exercise of the freedoms of movement. That would, however, be the case if a transfer of activity in the Community were determined solely by tax considerations, irrespective of any intention to seek real establishment and to integrate with the economy of the host society, with the sole aim of circumventing national laws or artificially exploiting differences between those laws.[45] Where a risk of abuse of this kind arises, it may be necessary, as the court held in Marks & Spencer [2006] STC 237, [2005] ECR I-10837, para 45, to apply to the economic activities of companies established in one of those states only the tax rules of that state in respect of both profits and losses. That, in my view, is the true meaning of the requirement of the allocation of the power to impose taxes in the Community context.
    34. It must also be shown that such a risk exists. That is why the court held, in the same judgment (at para 51), that the justification based on the preservation of the allocation of the power to impose taxes between member states cannot be separated from two other justifications relating, first, to the risk of losses being taken into account twice and, secondly, to the risk of tax avoidance. It is only in the light of these three justifications, 'taken together', that the court held that the restrictive provisions at issue could be justified.
    35. It must therefore be determined whether, as the German government contends, there is a risk of losses being taken into account twice or a risk of tax avoidance in the present case.
    2. The risk of losses being taken into account twice
    36. The German government argues that, like the legislation at issue in Marks & Spencer, the legislation at issue is necessary to prevent a company benefiting from multiple tax advantages in the form of losses incurred abroad being taken into account twice.
    37. That argument is irrelevant in the context of the present case. The losses at issue in this case are not, as in Marks & Spencer, losses incurred abroad by independent subsidiaries and subsequently surrendered against the profits of the parent company. They are losses incurred by the parent company as a result of a fall in the value of its shares in foreign subsidiaries. They are not to be confused with losses incurred by the subsidiaries themselves. These two kinds of losses are treated differently for tax purposes. It cannot therefore be held that there is a risk of the same losses being taken into account twice because a parent company is allowed to make such a deduction.
    38. Even if it were to be accepted that there is an economic link between these two kinds of losses, as the German government contends, so that taking the losses of the subsidiaries and those of the parent company into account separately could be described as 'taking the losses into account twice', it does not appear that in the present case taking the losses into account twice would have a specific connection with a transfer of activity to another member state. The alleged 'double advantage' is not confined to companies with cross-border activities. The fact that a parent company with subsidiaries in Germany may offset write-downs to the book value of its shares in those subsidiaries against its taxable profits does not preclude the subsidiaries from taking their own losses into account for tax purposes in that state. There is consequently no connection between losses being taken into account twice and the allocation of the power to impose taxes between member states, and the fact that losses may be taken into account twice cannot justify a restriction on freedom of establishment."
  62. The Court in Rewe Zentralfinanz declined to uphold the symmetry argument of the German government that, in the context of the tax treatment of a German resident company, the German tax authorities should not have to take account of losses stemming from the activity of a subsidiary which is established in another Member State, where such authorities had no right to tax the profits of that subsidiary. The Court continued (at paras 41 to 43, p 2807-8):
  63. "41. As the Advocate General stated at para 32 of his opinion, it is necessary to define the scope to be accorded to the legitimate requirement of the balanced allocation of the power to impose taxes between the member states. In particular, it must be noted that such a justification was accepted by the court in the judgment in Marks & Spencer only in conjunction with two other grounds, based on the taking into account of tax losses twice and on tax avoidance (see, to that effect, Marks & Spencer [2006] STC 237, [2005] ECR I-10837, paras 43 and 51).
    42. It must be acknowledged in that regard that there are courses of action which are capable of jeopardising the right of the member states to exercise their taxing powers in relation to activities carried on in their territory and thus of undermining a balanced allocation of the power to impose taxes between the member states (see Marks & Spencer, para 46) and which may justify a restriction on freedom of establishment (see Cadbury Schweppes [2006] STC 1908, [2007] Ch 30, paras 55 and 56). The court has thus held that the fact of giving companies the right to elect to have their losses taken into account in the member state in which they are established or in another member state would seriously undermine a balanced allocation of the power to impose taxes between the member states, since the tax base would be increased in the first state, and reduced in the second, by the amount of the losses surrendered.
    43. However, a difference in tax treatment between resident parent companies according to whether or not they have subsidiaries abroad cannot be justified merely by the fact that they have decided to carry on economic activities in another member state, in which the state concerned cannot exercise its taxing powers. Accordingly, an argument based on the balanced allocation of the power to impose taxes between the member states cannot in itself justify a member state systematically refusing to grant a tax advantage to a resident parent company, on the ground that that company has developed a cross-border economic activity which does not have the immediate result of generating tax revenues for that state."
  64. Mr Ewart, in arguing that the restriction imposed by s 403D(1)(c) could be justified by reference to the prevention of double use of losses, referred in particular to the judgment in Lidl. In outline, he argued that:
  65. (1) In Lidl the two justifications put forward were the balanced allocation of taxing rights and the prevention of the risk that losses may be used twice. The Court held that each of those justifications could be considered as capable of justifying a restriction.
    (2) The Court in Lidl did not settle the question, posed by the national court in that case, whether the three justifications set out in Marks & Spencer were cumulative or whether one only of those factors was sufficient. However, it made it clear that it was not necessary for all three factors to be present.
    (3) It was necessary in Lidl, as in Marks & Spencer, to justify the restriction by reference to more than one of the factors, as those cases concerned blanket restrictions on the use of losses. By contrast, what is involved in this case is a targeted provision. In such a case one is concerned only with the justification relevant to the targeted provision, and it is not necessary to look at other factors, in the same way that in Cadbury Schweppes[46] and Vodafone[47] it was not necessary to look at justifications other than tax avoidance.
  66. Mr Milne argued that:
  67. (1) On a proper analysis there is no "overriding reason of public interest" to justify the restriction imposed by the operation of s 403D(1)(c). The objectives of the UK rules on group relief are to give businesses the ability to structure their activities in a way that best suits their commercial aims and to recognise the fact that a group of companies is an economic unit. These are objectives that can be attained (indeed, can be better attained) without the restriction in s 403D(1)(c).
    (2) The only objective identified by the ECJ that is conceivably relevant to s 403D(1)(c) is the elimination of the prospect of the double use of losses. There is no risk of UK losses being used twice in the UK (because of s 411). The only theoretical possibility is that UK losses might be used both in the UK and in another Member State. The prevention of the double use of losses has never been upheld, on its own, as a justification for a restrictive measure. The common factor in all cases has been the allocation of taxation powers, described by Advocate General Kokott in Oy AA at para 47 as "at the heart of these elements". Furthermore, in Rewe Zentralfinanz, although the court in that case (following the Advocate General's Opinion) rejected the proposition that the losses of the parent company in question (constituted by a fall in the value of the shares held by the parent company in the foreign subsidiaries) were the same as the losses in the subsidiaries themselves (so there was no risk of the same loss being used twice), nevertheless Advocate General Maduro's Opinion at para 38 provides support for the proposition that prevention of double losses can only be considered in the context of balanced allocation.
    (3) In this case, the relevant losses have been incurred in the UK and the relevant profits (in respect of which the losses are proposed to be offset) have been made in the UK. Accordingly, pursuant to the principle of territoriality and of "source state taxation" (as referred to by Advocate General Geelhoed in Test Claimants in Class IV of the ACT Group Litigation, paras 50-51 of his Opinion), the UK has primary taxing competence. In particular, losses made in the UK fall to be deducted or relieved, in the first instance, from UK taxed profits. There is therefore no overriding reason of public interest for the UK to seek to prevent losses incurred in its own territory from being deducted against profits also made in its own territory simply because the losses are made by an undertaking incorporated in another Member State and those losses might potentially be used in that other Member State. Rather, it is for that other Member State to take steps, should it see fit to do so, to prevent the double use of such losses.
    (4) In considering this question it is important to have regard to the context. That includes the fact that, in the case of the taxation of permanent establishments in the UK, there is no general objective pursued by the UK in preventing the double use of losses. There is nothing to prevent the losses of a UK branch being used in the parent jurisdiction, and at the same time being carried forward (and, to a limited extent, being carried back) against profits of the UK branch.
  68. We prefer the submissions of Mr Milne to those of Mr Ewart. From a review of the ECJ cases, we conclude that, according to the present jurisprudence, a Member State cannot rely solely on the prevention of the use of losses twice to justify a restriction, even one that is targeted at such double use. It is clear now, according to Lidl, that it is not necessary for all three of the factors referred to in Marks & Spencer, and in that case taken together, to be present in order that a restrictive measure may be justified, but we do not regard Lidl as authority for the proposition that the double use of losses factor alone may amount to a justification. In Lidl itself the decision was based on a combination of the allocation of the power to tax and the need to prevent the danger that the same losses would be taken into account twice (see Lidl judgment, para 42). It was an essential reason for the decision that under the double taxation agreement the head office state, Germany, had given up the right to tax the profits of the permanent establishment in Luxembourg, and so it was justified in not giving relief for losses of the permanent establishment.
  69. We do not accept Mr Ewart's argument that in Marks & Spencer and Lidl the Court was constrained to rely on more than one justification, including balanced allocation, only because the restrictions in question were blanket restrictions, and that this case is different because s 403D(1)(c) is a targeted provision. We do not consider that Cadbury Schweppes or Vodafone support Mr Ewart's argument. As the ECJ in Rewe Zentralfinanz made clear (at para 42), Cadbury Schweppes is itself an example of a restriction on freedom of establishment being justified on the ground that related to courses of action that undermine the balanced allocation of taxing powers. In Cadbury Schweppes [2006] STC 1908 the ECJ held (paras 55-56, p 1939):
  70. "55. It follows that, 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.
    56. Like the practices referred to in para 49 of Marks & Spencer, which involve arranging transfers of losses, within a group of companies, to companies established in the member states which apply the highest rates of taxation and in which the tax value of those losses is therefore the highest, 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 (see Marks & Spencer, para 46)."
  71. The common theme throughout the cases is the balanced allocation of taxing rights. This is not surprising, because it is a general concept of a different quality to the narrower factors of tax avoidance and double use of losses. Since Marks & Spencer, the ECJ has consistently held that restrictions may be justified if they are aimed at preventing a serious undermining of a balanced allocation of the power to impose taxes between the Member States (see the judgments in Marks & Spencer (para 46), Rewe Zentralfinanz (para 42) and Lidl (para 55). This is a recognition of the need to safeguard that balanced allocation, and it can thus be concluded from the cases that balanced allocation is the fundamental basis on which Member States' claims for justification must be judged. There is nothing to suggest that a restriction on relief for losses can be justified solely on account of prevention of the use of those losses twice without regard to the effect on the balanced allocation of taxing rights. The test is conjunctive, as is made clear in Rewe Zentralfinanz (para 41). In the same way as the scope of the legitimate requirement for balanced allocation must be defined by reference to one of the other two grounds, namely the prevention of the use of losses twice or tax avoidance, so it follows, in our view, that the use of losses twice cannot on its own be a justification. In Cadbury Schweppes it was held that the abusive practice identified had to be one that undermined the right of the Member States to exercise their tax jurisdiction in relation to activities carried out in their territory and thus to jeopardise a balanced allocation between the Member States of the power to impose taxes. Likewise, in our view, a restriction under the law of a Member State on the use of losses can only be justified on the ground that it prevents the use of those losses twice if it jeopardises the balanced allocation of taxing rights between that Member State and another Member State or States.
  72. We consider that it is right to view the restriction imposed by s 403D(1)(c) in the context of the source state taxation that applies to the profits of permanent establishments in the UK of companies in other Member States. That is not to say that we base our conclusion on the need for symmetry between profits and losses. Symmetry in the sense that a source state should offer non-residents equivalent tax benefits to those offered to residents to the extent that the source state exercises equal tax jurisdiction over both (see Advocate General Geelhoed in Test Claimants in Class IV of the Act Group Litigation, para 69) is an accepted principle. But the ECJ has rejected arguments on symmetry raised by the Member States themselves (for example, in Marks & Spencer, in Oy AA and in Rewe Zentralfinanz) and we do not consider that the need for symmetry can override a restriction that can be justified by reference to the factors identified by the Court. Nevertheless, where the balance of allocation of taxing rights is in favour of one Member State (as in our judgment it is here, in the UK), it is relevant to have regard to that fact in determining whether a restriction on a corresponding relief can be justified. It is not just a case, as Mr Ewart contends, that if a provision is aimed at preventing double use of losses then, ex hypothesi, it must be justified. The proper analysis, in our view, is whether, in its context, and having regard to the balanced allocation of taxing rights in favour of the source state, there is an overriding reason of public interest in the restriction sought.
  73. We do not consider in this case that there is such an overriding reason. We are concerned here with taxation in the source state. The primary right of the source state to tax profits and accordingly the primary obligation to relieve losses is an established principle of international taxation that was recognised by the ECJ in Lidl as justification for the head office state to refuse to give relief for the losses in the permanent establishment state. Here the UK, as source state, does give relief for losses by carry forward (or backwards) and it cannot justify not giving relief for losses of a branch in one particular circumstance, group relief, when a subsidiary is permitted to relieve losses in this way. In the case of a permanent establishment (or branch) the balance of the allocation between Member States of the power to impose taxes lies with the source state. The use of losses of a branch cannot therefore in our view jeopardise the balanced allocation. The restriction imposed by s 403D(1)(c) is accordingly not justified.
  74. On the basis of our finding as to the nature of the balanced allocation between the source state and the residence state, the analysis of the position in the source state is one of principle, and does not depend on how the residence state chooses to provide relief for losses of the permanent establishment in the source state, or whether it provides relief for such losses at all. Any concern about the use of losses in another Member State is a matter for that state. Our conclusion is therefore unaffected by the tax position in the Netherlands.
  75. There is no need for us to consider the position in the residence state but we set our thoughts in order to explain why we consider that the position in the source state is unaffected by the position in the residence state whilst emphasising that this has not been the subject of any argument before us. Where the residence state defines its taxing rights to include the profits of the permanent establishment, in an ordinary case for a tax credit residence state, and losses of the permanent establishment are carried forward in the source state, the permanent establishment's future profits will generally become taxable in the residence state in full (because the loss carry forward will eliminate taxation in the source state up to the amount of the loss, resulting in no credit relief for those profits when taxed in the residence state). It seems to us that the residence state would clearly not be justified in refusing to give relief for the losses even though the losses are also relieved in the source state. Such relief for losses (subject to effective clawback when profits are subsequently taxed) would be inherent in the taxing powers that the residence state has adopted. However, where, as here, the source state permits surrender of losses away from the permanent establishment, the result in the residence state is likely to be that future profits of the permanent establishment will become taxable subject to credit for source state tax on those profits (which will not, in these circumstances, be reduced by the carry forward of the surrendered losses). In Lidl, the residence state defined its taxing rights to exempt from taxation the profits of the permanent establishment in the source state. It was held that, subject to any question of proportionality, the residence state would be justified in not giving relief for the losses of the permanent establishment, on the grounds of the need to safeguard the allocation of the power to tax between the Member States and the need to prevent the danger that the same losses will be taken into account twice. A similar analysis would arguably apply in a case where profits of the permanent establishment were not exempt in the residence state, but the taxing rights of that state over those profits were in practice rendered less effective by the application of foreign tax credit enhanced on account of the surrender of losses out of the permanent establishment in the source state. Indeed, in Krankenheim (Case C-157/07) [2009] STC 138, on a reference from a residence state, the ECJ held that a restriction in the residence state in the form of reintegration of losses of a permanent establishment in the source state previously allowed up to the amount of profits made in that permanent establishment was justified on the basis of coherence of the residence state's tax system, even in a case where, because there was a possibility of use of the losses in the residence state, there was no carried forward relief for the losses in the source state, and thus no risk of double use of losses. But all these matters are ones for the residence state alone. The particular application of the residence state tax system to the profits and losses of the permanent establishment and the way in which the residence state chooses to address (or not to address) any perceived tax distortion under its own domestic law does not affect the position of the source state so far as concerns the fundamental balanced allocation as between the source state and the residence state of the right to tax the permanent establishment's profits, and cannot therefore, in our view, alter the position as regards the justification of a restriction on the right to use losses in the source state. In particular, it does not make a difference whether the residence state allows loss relief without restriction, allows it only on the basis of the no possibilities test set out in Marks & Spencer (as in Lidl itself) or allows it subject to less restrictive measures such as loss recapture (as in Krankenheim). None of these alternative approaches would affect the balanced allocation so far as the source state's right to tax profits of the permanent establishment was concerned and in none of the circumstances described would the use of the losses in the source state seriously undermine that balanced allocation.
  76. Issue 4: Proportionality
  77. We have held that each of s 406(2) and 403D(1)(c) impose restrictions on the freedom of establishment and that those restrictions cannot be justified on the basis that there is an overriding reason of public interest. That is sufficient to dispose of this joint referral, but in case we are wrong in relation to s 403D(1)(c) (and that instead a provision to prevent the use of losses twice can be justified in the circumstances of this case), and having regard to the arguments on proportionality in relation to s 403D(1)(c) adduced by the parties, we consider whether s 403D(1)(c), if justified, would be proportional.
  78. As described by the ECJ in Papillon (at para 52, p 563), it is necessary that national legislation does not go beyond what is necessary to attain its objective, that is to say, that the same objective cannot also be attained by measures which are less restrictive of the freedom of establishment. According to Advocate General Kokott in that case (at para 62, p 556):
  79. "In particular, there should be no more moderate means available to prevent the use of the loss more than once."

    This formulation of the proportionality test in Papillon follows that set out in the ECJ judgments in Marks & Spencer (para 53), Oy AA (para 61) and Lidl (para 44).

  80. In Marks & Spencer the ECJ was considering the application of the UK's group relief rules to losses of subsidiaries established in other Member States. Having decided that the restriction on the surrender of such losses was justified, the Court went on to determine that, in certain circumstances, the restrictive measure went beyond what was necessary to attain the essential part of the objectives pursued. The Court said (at paras 55-56, p 268):
  81. "55. In that regard, the court considers that the restrictive measure at issue in the main proceedings goes beyond what is necessary to attain the essential part of the objectives pursued where:
    —the non-resident subsidiary has exhausted the possibilities available in its state of residence of having the losses taken into account for the accounting period concerned by the claim for relief and also for previous accounting periods, if necessary by transferring those losses to a third party or by offsetting the losses against the profits made by the subsidiary in previous periods, and
    —there is no possibility for the foreign subsidiary's losses to be taken into account in its state of residence for future periods either by the subsidiary itself or by a third party, in particular where the subsidiary has been sold to that third party.
    56. Where, in one member state, the resident parent company demonstrates to the tax authorities that those conditions are fulfilled, it is contrary to arts 43 EC and 48 EC to preclude the possibility for the parent company to deduct from its taxable profits in that member state the losses incurred by its non-resident subsidiary."
  82. Mr Ewart's primary submission is that s 403D(1)(c) is compatible with the "no possibilities" test as propounded in Marks & Spencer, and that accordingly it is compatible with Community law. Mr Milne, on the other hand, argues that s 403D(1)(c) goes well beyond what would be appropriate and proportionate to the attaining of the objective of preventing the double use of losses. He pointed to what he referred to as "draconian" features of that provision, which, he said, rendered the provision unnecessarily restrictive in two important respects:
  83. (1) There is no need to stipulate that if even any part of the loss for a relevant year has been used then none of the losses at all are available for group relief. This "all or nothing" test plainly goes beyond what is necessary to achieve the objective. A less restrictive alternative is simply to preclude the use of that part of the losses that have already been used.
    (2) It would suffice to achieve any objective of no double use of losses that s 403D(1)(c) precluded the use of such losses as have actually been set off against profits overseas.
  84. The Tribunal in Marks and Spencer [2009] SFTD 1 considered the issue whether the no possibilities test applied to each euro of foreign losses (as the taxpayer contended) or to losses in general so that if there is a possibility of some losses being used, no losses can qualify under the no possibilities test (as HMRC contended). Here s 403D(1)(c) appears to provide that if a loss computed under foreign tax principles corresponds to any pound of UK losses in the UK branch computed on UK tax principles, then no losses are allowable. The Tribunal concluded:
  85. "We consider that the answer is determined by the proper application of the no-possibilities test in accordance with the guidance given by Park J and the Court of Appeal. At one extreme if the company is still trading (or, as in Park J's example at [37], has ceased trading) but losses can still be carried forward, whether or not it is possible to use them depends entirely on the likelihood of making profits, so that none of the losses qualify. At the other extreme if, as happened here, the company is in liquidation and cannot start any new business but is holding some cash pending distribution which earns interest against which the losses can be set, we consider that one must split the losses into those that on the balance of probabilities can be set against the interest, and those that cannot. This is on the ground that the estimated income during the liquidation is one of 'the objective facts of the company's situation at the relevant time.' The no-possibilities test is applicable to the latter. The fact that the liquidator might make a profitable investment that used more of the losses is, we consider, in the fanciful category in circumstances in which the liquidator's duty is merely to pay the liabilities and distribute the assets."

    For the same reasons as the Tribunal in that case we regard it as disproportionate that if one pound of losses corresponds to foreign allowable losses then no losses are allowable in the UK. The same principles as apply to the no possibilities test should be applied here.

  86. We therefore conclude that, on its own, s 403D(1)(c) is not compatible with the no possibilities test. We consider below how that incompatibility may be addressed, but we must first deal with the submissions of Mr Milne to the effect that s 403D(1)(c) may be interpreted by reference to an alternative test to the no possibilities test.
  87. The first such submission was that s 403D(1)(c) could operate so as to preclude only the use of losses that had already been set off against profits in the non-UK territory. This was on the basis that there is no legitimate need to stipulate that group relief should be denied merely because the same losses might be used in the future in some other jurisdiction. Secondly, Mr Milne argued that even if it were a legitimate objective for the UK to seek to preclude the double use of the same losses in the future in circumstances where the losses are used in the UK and are then used in another Member State, even that wider objective could be achieved by more moderate means than those contained in s 403D(1)(c). For example, it was suggested that (a) a taxpayer could be obliged to adjust its claim for group relief in the event that the same losses are used overseas, and/or (b) a further assessment could be made in that event.
  88. Mr Ewart characterised Mr Milne's arguments in this respect as at odds with Marks & Spencer. We agree. In Marks & Spencer the company itself and the European Commission put forward various suggestions for less restrictive measures. These are described by the Court at para 54 (at p 267):
  89. "Marks & Spencer and the Commission contended that measures less restrictive than a general exclusion from group relief might be envisaged. By way of example, they referred to the possibility of making relief conditional upon the foreign subsidiary's having taken full advantage of the possibilities available in its member state of residence of having the losses taken into account. They also referred to the possibility that group relief might be made conditional on the subsequent profits of the non-resident subsidiary being incorporated in the taxable profits of the company which benefited from group relief up to an amount equal to the losses previously set off."

    These submissions are similar to those made by Mr Milne. The first refers to actual use, and the second to a form of recapture charge. Neither was accepted by the Court which, as we have seen at para 55 of its judgment, adopted the no possibilities tests.

  90. In Lidl, the Advocate General contrasted the factual situation in Marks & Spencer, where losses were made by subsidiaries that had been wound up and sold, with that of an ongoing permanent establishment, such as in Lidl. In her opinion, Marks & Spencer did not fully examine the cash flow disadvantage of having to carry losses forward instead of using them immediately. On this basis the Advocate General would have favoured a deduction of losses combined with recapture of loss relief. However, the Court in Lidl did not follow the Advocate General's Opinion. The Court said (at paras 45-51, p 3224):
  91. "45. Lidl Belgium and the Commission of the European Communities have referred in particular to the possibility that the right of a principal company to deduct the losses incurred by a permanent establishment belonging to it could be made subject to the condition that the company incorporate in its future profits the subsequent profits of the permanent establishment, to the extent of the losses previously offset. In that context, they refer to the system which operated in the Federal Republic of Germany prior to 1999.
    46. It should be pointed out that reference was made in para 54 of the judgment in Marks & Spencer to the possibility of making the benefit of the tax advantage at issue subject to such a condition, together with the possibility of making the benefit of that advantage conditional upon the subsidiary established in a member state other than that in which the principal company has its seat having taken full advantage of the possibilities available in its member state of residence of having the losses taken into account.
    47. In that regard, the court held in para 55 of the judgment in Marks & Spencer that a measure which restricts the freedom of establishment goes beyond what is necessary to attain the objectives pursued where a non-resident subsidiary has exhausted the possibilities for having the losses incurred in the member state where it is situated taken into account for the accounting period concerned and also for previous accounting periods and where there is no possibility for that subsidiary's losses to be taken into account in that state for future periods.
    48. In para 56 of that judgment, the court also stated that where, in one member state, the resident parent company demonstrates to the national tax authorities that those conditions are fulfilled, it is contrary to art 43 EC to preclude the possibility for the parent company to deduct from its taxable profits in that member state the losses incurred by its non-resident subsidiary.
    49. As regards the main proceedings, it must be pointed out that Luxembourg tax legislation provides for the possibility of deducting a taxpayer's losses in future tax years for the purposes of calculating the tax base.
    50. As was confirmed at the hearing before the court, Lidl Belgium has in fact benefited from such an offsetting of the losses incurred by its permanent establishment in 1999 in a subsequent tax year, namely 2003, in which that entity generated profits.
    51. Accordingly, Lidl Belgium has not shown that the conditions laid down in para 55 of the judgment in Marks & Spencer, for establishing the situation in which a measure constituting a restriction on the freedom of establishment for the purposes of art 43 EC goes beyond what is necessary to attain legitimate objectives recognised by Community law, were satisfied."

    The judgment is reflected in the dispositif in which the Court ruled:

    "Article 43 EC does not preclude a situation in which a company established in a member state cannot deduct from its tax base losses relating to a permanent establishment belonging to it and situated in another member state, to the extent that, by virtue of a double taxation convention, the income of that establishment is taxed in the latter member state where those losses can be taken into account in the taxation of the income of that permanent establishment in future accounting periods."
  92. We agree with Mr Ewart. It is clear on the ECJ case law that the test to be applied in the case of a restrictive measure that is justified on the basis of the prevention of the double use of losses is the Marks & Spencer no possibilities test. We do not accept Mr Milne's submission in this regard that the fact that both Marks & Spencer and Lidl were not source state cases (such as we are considering here) but residence state cases, where the Member State in question did not have taxing rights over the profits, can alter the position so as to permit the use of loss recapture as an alternative means of achieving the objective sought by the restrictive measure. In Lidl the use of loss recapture was a feasible alternative measure according to the Advocate General, but nevertheless it was rejected by the Court.
  93. It was common ground before us that if we were to find that s 403D(1)(c) goes beyond what is necessary to attain its objective, and that a less restrictive measure is appropriate, it was open to us to interpret s 403D(1)(c) so as to apply that less restrictive measure. This follows from the Court of Appeal judgment in Vodafone 2 v Revenue and Customs Commissioners [2009] STC 1480 in which it was held that the CFC legislation was susceptible to a conforming interpretation so as to interpret that legislation in a manner which did not restrict the freedom of establishment, as declared by the ECJ in Cadbury Schweppes v IRC. On the basis of our finding that, if (contrary to our view) s 403D(1)(c) is justified, it goes beyond what is necessary to attain its objective, and that a less restrictive measure is available in the form of the no possibilities test, we decide that s 403D(1)(c) should be interpreted so as to be in conformity with the no possibilities test as set out in Marks & Spencer.
  94. Result
  95. Our answers to the questions in the joint referral are accordingly as follows:
  96. 1. Whether, in light particularly of any applicable principles of EU law, section 406(2) ICTA 1988, or any other provision, applies so as to prevent consortium relief from being available on the basis that the relevant link companies (as that term is defined in section 406(1) ICTA 1988), being Koninklijke Philips Electronics NV and Philips GmbH, although resident at all times in an EU Member State, were not at any relevant time within the charge to UK corporation tax
    Answer. Section 406(2) contains a restriction that cannot be justified and so does not apply to prevent consortium relief from being available in these circumstances.
    2. Whether, in light particularly of any applicable principles of EU law, section 403D ICTA 1988 (entitled 'Relief for or in respect of non-resident companies'), or any other provision, applies so as to prevent consortium relief from being available by reference to the prospect of those losses being utilised (as more fully spelt out in section 403D ICTA 1988, or such other relevant provision), in any period, for the purposes of Dutch corporate income tax or any other non-UK tax.
    Answer. Section 403D contains a restriction that cannot be justified and so does not apply to prevent consortium relief from being available in these circumstances. If we are wrong about justification there is a more proportional method of restricting the double use of losses in the form of the no possibilities test adopted in Marks and Spencer and s 403D should be interpreted so as to be in conformity with that test.
    JOHN F AVERY JONES
    ROGER BERNER
    TRIBUNAL JUDGES
    RELEASE DATE: 18 August 2009

Note 1   See Matthijs Overbosch at paragraph 51 [A1/Tab 9/64]    [Back]

Note 2   These figures are derived from the witness statement of Matthijs Overbosch. HMRC reserve their position in relation to the correctness of such figures    [Back]

Note 3   See the OECD Model Double Taxation Convention on Income and on Capital, with Commentaries to the articles, OECD, Paris, 1977, as revised.    [Back]

Note 4   See the OECD model double taxation convention.    [Back]

Note 5   In the case of the exemption method, the taxpayer's state of residence exempts the foreign-source income of its residents, on the basis that this income has already been taxed in the 'source' state (ie, the state in which the income was earned). In the case of the credit method of avoidance of double taxation, however, taxpayers earning foreign-source income are taxed in their state of residence on their worldwide income, including foreign-source income, but may credit the tax paid in the source state against the home state tax attributable to this foreign-source income.    [Back]

Note 6   Royal Bank of Scotland, para 26, and cases cited therein.    [Back]

Note 7   See Manninen (tax credit granted for domestic dividends must also be given for foreign-source dividends); Verkooijen (home state must grant same exemption from individual income tax on foreign-source dividends as grants to domestic dividends), Lenz v Finanzlandesdirektion für Tirol (Case C-315/02) [2004] ECR I-7063, [2004] 3 CMLR 274 (option for income tax treatment available for domestic dividends must be extended to foreign-source dividends). See also EC Commission v France (Case C-334/02) [2007] STC 54, 6 ITLR 642 (benefit of low rate final withholding tax was restricted to proceeds from debt claims paid by resident debtors only; a similar benefit should be granted to beneficiaries of payments by foreign debtors) and the opinion of Advocate General Tizzano of 10 November 2005 in Meilicke v Finanzamt Bonn-Innenstadt (Case C-292/04) (2005) Transcript (Opinion) (Fre) 10 November, (2006) Transcript (Opinion) (Eng) 5 October.    [Back]

Note 8   Algemene Maatschappij voor Investering en Dienstverlening NV (AMID) v Belgium (Case C-141/99) [2003] STC 356, 2 ITLR 201, Mertens v Belgium (Case C-431/01) [2002] ECR I-7073 (domestic company required to offset domestic losses against foreign profits); ICI (domestic loss relief conditional on question whether domestic company had foreign subsidiaries).    [Back]

Note 9   Metallgesellschaft.    [Back]

Note 10   Marks & Spencer, para 46. The court justified this on the basis, inter alia, that companies cannot have an option to choose tax jurisdiction from free movement provisions, as this would jeopardise a balanced allocation of tax jurisdiction.    [Back]

Note 11   Saint Gobain.    [Back]

Note 12   EC Commission v France (Case 270/83) [1986] ECR 273 (Avoir Fiscal), R v IRC, ex p Commerzbank AG (Case C-330/91) [1993] STC 605, [1994] QB 219 (branch of non-resident company entitled to same interest on repayment of overpaid taxes), Futura (branch of non-resident company entitled to same loss carry-over possibilities).    [Back]

Note 13   Avoir Fiscal.    [Back]

Note 14   Royal Bank of Scotland.    [Back]

Note 15   Lankhorst-Hohorst GmbH v Finanzamt Steinfurt (Case C-324/00) [2003] STC 607, 5 ITLR 467, and see the pending reference in Test Claimants in the Thin Cap Group Litigation v IRC (Case C-524/04) [2007] All ER (D) 219 (Mar).    [Back]

Note 16   Futura, para 21.    [Back]

Note 17   Safir v Skattemyndigheten i Dalarnas Län (Case C-118/96) [1998] STC 1043, [1999] QB 451, Futura.    [Back]

Note 18   See cases cited in footnotes [9] and [10] above.    [Back]

Note 19   Marks & Spencer, paragraph 51 (Case C-446/03).    [Back]

Note 20   See above (para 15 [not reproduced].    [Back]

Note 21   See Case C336/96 Gilly [1998] ECR I-2793, paragraphs 24 and 30; Case C385/00 De Groot [2002] ECR I-11819, paragraph 93; Case C376/03 D [2005] ECR I-5821, paragraphs 50 and 51; Case C513/03 van Hilten-van der Heijden [2006] ECR I-1957, paragraph 47; and N (Case C-470/04), paragraphs 43 and 44.    [Back]

Note 22   Futura Participations and Singer (cited above), paragraph 22; van Hilten van der Heijden (cited above, footnote [21]), paragraph 48; and N (cited above, footnote [21]), paragraph 45.    [Back]

Note 23   Marks & Spencer (cited above), paragraph 39. As regards allocation according to the worldwide income and source state principles, see the Opinion of Advocate General Geelhoed in Case C374/04 Test Claimants in Class IV of the ACT Group Litigation [2006] ECR I-11637, points 49 to 51.    [Back]

Note 24   Marks & Spencer (cited above, footnote [19]), paragraph 46.    [Back]

Note 25   Once-only taxation can be achieved by two States acting with reference to one another, for example by the application of the imputation method.    [Back]

Note 26   See below, paragraph 67 [not reproduced].    [Back]

Note 27   As regards the questions of taxation on migration arising in this context, see the recent judgment in N (cited in footnote [21] above).    [Back]

Note 28   See Manninen’s case (Case 319/02), paragraph 49, and Marks & Spencer (cited in footnote [19] above), paragraph 44.    [Back]

Note 29   Paragraphs 40 and 41.    [Back]

Note 30   See point 16 of the Opinion of Advocate General Poiares Maduro in Marks & Spencer.    [Back]

Note 31   The question whether the three justifications must be considered cumulatively is considered below.    [Back]

Note 32   Marks & Spencer, paragraph 45.    [Back]

Note 33   Paragraph 46.    [Back]

Note 34   Paragraph 47 and 48.    [Back]

Note 35   Since the loss was made in 1999 but could not be used in Luxembourg until 2003, Lidl was still disadvantaged by its inability to set the loss off against its profits in Germany at an earlier stage. The issue of cash flow is explored further below, in the context of proportionality.    [Back]

Note 36   Paragraphs 47 and 48.    [Back]

Note 37   Although it may be noted that tax avoidance by such ‘trafficking in losses’ does not appear to have been relevant on the facts in Marks & Spencer, which concerned vertical upwards transfer of losses from subsidiary to parent.    [Back]

Note 38   Proceedings brought by Oy AA (Case C-231/05) [2008] STC 991, [2007] ECR I-6753, paras 51 to 60.    [Back]

Note 39   Amurta SGPS v Inspecteur van de Belastigdienst/Amsterdam (Case C-379/05) [2008] STC 2851, [2007] ECR I-9569, paras 57 to 59..    [Back]

Note 40   Marks & Spencer, paragraph 44, reflecting settled case-law cited, in particular in Proceedings brought by Manninen (Case C-319/02) [2004] STC 1444, [2004] ECR I-7477, para 49.    [Back]

Note 41   See, inter alia, Cura Anlagen GmbH v Auto Service Leasing GmbH (ASL) (Case C-451/99) [2002] ECR I-3193, para 40.    [Back]

Note 42   See, to this effect, the Opinion of Advocate General Geelhoed in Case C-374/04 Test Claimants in Class IV of the ACT Group Litigation v IRC (Case C-374/04) [2007] STC 404, [2007] All ER (EC) 351, paras 62 and 63. See also D Weber ‘The Bosal Holding Case: Analysis and Critique’ EC Tax Review, 2003-2004, p 220, and PJ Wattel ‘Red Herrings in Direct Tax Cases before the ECJ’ Legal Issues of Economic Integration, 2004, No 2, pp 81 to 95, particularly pp 89 and 90.    [Back]

Note 43   See, in this connection, Rewe-Zentral AG v Bundesmonopolverwaltung für Branntwein (Case 120/78) [1979] ECR 649, ‘Cassis de Dijon’.    [Back]

Note 44   See, on this point, my opinion in Marks & Spencer.    [Back]

Note 45   See, to the same effect, the opinion of Advocate General Léger in Cadbury Schweppes plc v IRC (Case C-196/04) [2006] STC 1908, [2007] Ch 30.    [Back]

Note 46   Cadbury Schweppes v IRC (Case C-196/04) [2006] STC 1908    [Back]

Note 47   Vodafone 2 v HMRC (No 2) [2009] STC 1480    [Back]


BAILII: Copyright Policy | Disclaimers | Privacy Policy | Feedback | Donate to BAILII
URL: http://www.bailii.org/uk/cases/UKFTT/TC/2009/TC00176.html