Deutsche Morgan Grenfell Group Plc v Inland Revenue & Anor [2003] EWHC 1779 (Ch)

Case No HC000 4650 CH[2003] EWHC 1779 (Ch)
IN THE HIGH COURT OF JUSTICE
CHANCERY DIVISION
Venue Royal Courts of Justice, Strand, London, WC2A 2LLDate 18 July 2003
Before
THE HONOURABLE MR JUSTICE PARK
Between
Deutsche Morgan Grenfell Group PlcClaimant(1) The Commissioners of Inland RevenueDefendants(2) HM Attorney GeneralDefendant
Laurence Rabinowitz QC and Francis Fitzpatrick (instructed by Slaughter & May) for ClaimantIan Glick QC and Bruce Carr (instructed by The Solicitor of Inland Revenue) for DefendantsHearing Hearing dates : 12.05.03 to 15.05.03
Approved Judgment
[1]Abbreviations, dramatis personae, glossary, etc. ACT Advance corporation tax CJEC The Court of Justice of the European Communities DBAG Deutsche Bank AG, a German company and the parent company, direct or indirect of DBI and DMG. DBI DB Investments (GB) Limited DMG Deutsche Morgan Grenfell Group Plc, the claimant company ICTA The Income and Corporation Taxes Act 1988 in its form before amendments which took effect in 1999. Kleinwort Benson The House of Lords decision in Kleinwort Benson v Lincoln City Council [1999] 2 AC 349. Metallgesellschaft/Hoechst The combined cases in the CJEC of Metallgesellschaft Ltd and others v Commissioners of Inland Revenue and the Attorney General (Case C-397/98) and(1) Hoechst AG(2) Hoechst United Kingdom Ltd v Commissioners of Inland Revenue and the Attorney General (Case C-410/98). The judgment of the CJEC was delivered on 8 March 2001 and is reported at [2001] STC 452. MCT Mainstream corporation tax, as to which see paragraph 6 of the judgment. Pirelli case, the Pirelli Cable Holding NV and others v Commissioners of Inland Revenue [2003] STC 250.

Overview

[2]At the times relevant for this case United Kingdom tax law provided that, where subsidiary companies paid dividends to their parents, ‘group income’ elections could be made which enabled the dividends to be paid without the paying subsidiary having to pay ACT to the UK Revenue. Until the decision of the CJEC in Metallgesellschaft/Hoechst such elections could be made only between United Kingdom resident companies: section 247(1) of ICTA included the words:
‘both being bodies corporate resident in the United Kingdom’
. However, in Metallgesellschaft/Hoechst the CJEC held that it was contrary to article 52 (now article 43) of the EC Treaty for United Kingdom law to deny the right to make group income elections where the parent company was resident in a Member State other than the United Kingdom. The CJEC also held that, as respects payments of ACT which had already been made in the past, Community law conferred a right of compensation or restitution which the aggrieved companies were entitled to pursue in the United Kingdom courts.[3]Many United Kingdom subsidiaries of Member State parents, including DMG, a subsidiary of a German parent, commenced actions seeking compensation or restitution accordingly. The actions are now all organised within a Group Litigation Order (as to which see the Civil Procedure Rules, rules 19.11 to 19.15), and a series of test cases is being heard to determine various questions of principle to which they, or some of them, give rise. The first such case was the Pirelli case, which concerned an issue which arose where the parent company was resident in a Member State which had a particular kind of Double Taxation Agreement with the United Kingdom. The second such case is this one, in which the test claimant is DMG.[4]The special feature of this case, and of others within the Group Litigation Order for which this is the test case, is that questions of limitation of actions arise: some of the payments of ACT which DMG made to the United Kingdom Revenue were made more than six years before DMG commenced its action claiming compensation or restitution. The Revenue say that DMG’s claims in respect of those ACT payments are barred by the normal six years time limit laid down by section 2 of the Limitation Act 1980. DMG argues that it can circumvent this effect of the 1980 Act by a combination of two principles of law: first, that English law recognises an action for restitution of money paid under a mistake of law (see the decision of the House of Lords in the Kleinwort Benson case identified in the Table of Abbreviations, etc, at the beginning of this judgment); second, that under section 32(1)(c) of the 1980 Act, the limitation period where an action is brought for relief from the consequences of a mistake begins to run only when the claimant discovered the mistake or could with reasonable diligence have discovered it. DMG has pleaded its case in ways which include a claim for restitution on the ground that it made the payments of ACT under a mistake of law. It says that it did not discover the mistake until the decision of the CJEC in Metallgesellschaft/Hoechst. That decision was given on 8 March 2001, by which date DMG had already commenced its action. Alternatively, DMG says that the earliest time at which it could be said to have discovered its mistake was when the Advocate General delivered his opinion in Metallgesellschaft/Hoechst. That was on 12 September 2000, well within six years before DMG commenced its action. DMG argues that therefore it is not time-barred from recovering the relief which it claims, notwithstanding that the ACT payments (or some of them) were made more than six years before the action was commenced.[5]It seems to me that three principal questions arise: i. Does a claim for restitution of a payment made under a mistake of law apply in the case of a payment of tax to a Revenue authority? In my view the answer is: yes. ii. Did DMG make the relevant payments of ACT under a mistake of law so as to be entitled to claim restitution in accordance with the principles laid down by the House of Lords in Kleinwort Benson? In my view the answer is again: yes. iii. Under section 32(1)(c) of the Limitation Act 1980, when did the limitation period begin to run? Did it begin to run only at the time of the CJEC’s decision in Metallgesellschaft/Hoechst (or possibly at the time of the Advocate General’s opinon)? Or did it begin to run at the earlier time in mid-1995 when DMG first learned that the Hoechst group of companies was arguing that the rule that only United Kingdom resident companies could make group income elections was contrary to Community law? In my opinion the period began to run only at the time of the CJEC’s decision. It follows that DMG’s claims succeed. The relevant United Kingdom tax law[6]At the times with which this case is concerned the general rule of United Kingdom tax law was that, if a company resident in the United Kingdom paid a dividend, it had to pay ACT to the Revenue: ICTA s.14. The rate of ACT varied over the years. At the times with which the present case is concerned it was normally 25% of the dividend. In the usual case the amount so paid could be set off against the company’s normal liability to pay corporation tax on its profits, sometimes referred to as mainstream corporation tax or (as in this judgment) MCT. Where the ACT was so set off the payment of it did not cause the company to suffer an absolute loss of the money, because the set-off reduced on a pound for pound basis a later payment of MCT which the company would otherwise have had to make anyway. However, it did have the effect that the company had to pay some of its corporation tax bill early, so the ACT meant that the company suffered a timing disadvantage and the United Kingdom Revenue received a corresponding timing advantage. (There could be cases where a subsidiary had paid ACT and could not set it off against MCT. In such cases the ACT payment did represent an absolute loss of the money and not just a timing disadvantage. However, in the case of DMG all the relevant ACT payments were eventually set off against MCT, so that the disadvantage for which DMG seeks relief is a timing disadvantage only.)[7]There was an exception to the rule that a company which paid a dividend had to pay ACT. If it was a subsidiary of a United Kingdom holding company the two companies could jointly make a group income election under section 247 of ICTA. The effect was that the subsidiary did not have to pay the ACT after all. It did still have to pay its full MCT when the due date came round, but the group income election removed the timing disadvantage. However, if the dividend-paying company was a subsidiary of a non-United Kingdom holding company the two companies could not make a group income election, so the subsidiary had to pay ACT and suffer the timing disadvantage, in circumstances where a subsidiary of a United Kingdom holding company, if it and its parent had made a group income election, did not. That was the feature of United Kingdom law which, in Metallgesellschaft/Hoechst, the CJEC held to have been contrary to EC law and also to entitle the companies which had suffered from it to claim compensation or restitution. Where, as in this case and as in the cases of the Metallgesellschaft and the Hoechst groups, the ACT had been set off against MCT, the compensation or restitution would be an amount calculated by reference to interest over the period from the payment of the ACT until it was set off against MCT. The Limitation Act 1980[8]Sections 2 and 5 of this Act provide that an action founded on tort (section 2) or on simple contract (section 5) may not be brought after the expiration of six years from the date when the cause of action accrued. The parties agree that a claim for compensation or restitution of the nature which the CJEC said in Metallgesellschaft/Hoechst should be available would almost certainly be a claim in tort or contract and subject to a six years time limit. An important issues is : when does the six years period start to run. The parties also agree that, if section 2 or section 5 applies and the period runs only from the date when the cause of action accrued, some of DMG’s causes of action accrued more than six years before it brought its action and would therefore be time-barred. However section 32 opens the possibility that, if one of the detailed circumstances described in the section exists, the limitation period, though it is still six years, may commence to run at a date later than the date when the cause of action accrued. So far as relevant the section provides as follows: 32 Postponement of limitation period in case of fraud, concealment or mistake (1) … where in the case of any action for which a period of limitation is prescribed by this Act either – (a) …; or (b) …, or (c) the action is for relief from the consequences of a mistake, the period of limitation shall not begin to run until the plaintiff has discovered the … mistake … or could with reasonable diligence have discovered it. The facts[9]At the times when the ACT payments relevant to this case were made the group structure was a little more complicated than a parent company with a 100% direct subsidiary, but nothing turns on the extra complications. The German holding company was DBAG. It was not the simple 100% owner of DMG. Instead it owned 14% of the shares in DMG directly. It owned all of the shares in DBI, an interposed United Kingdom holding company, and DBI owned the other 86% of the shares in DMG. I mention for completeness that in November 1997 the structure was simplified and DBI, the 100% direct subsidiary of DBAG, became the 100% (as opposed to merely the 86%) direct owner of DMG. However, the dividends and the associated ACT payments which are relevant to this case were paid before November 1997, so it is the earlier structure which is in point.[10]According to the terms of ICTA s.247, group income elections could not have been made between DMG and DBAG or between DBI and DBAG, because DBAG was not a company resident in the United Kingdom. A group income election could have been made between DMG and DBI (because both companies were resident in the United Kingdom), but no such election was made. The reason was that, although an election would have enabled dividends to flow from DMG to DBI without ACT, onward dividends from DBI to DBAG would have had to be paid subject to a liability on the part of DBI to pay ACT: or at least that was how the matter appeared on the terms of the domestic United Kingdom legislation. In the circumstances DMG paid ACT by reference to all its dividends – the 86% of them paid to DBI as well as the 14% of them paid to DBAG – and DBI did not pay ACT when it paid onward dividends to DBAG. DBI did not have to pay ACT on those onward dividends because of the rules about franked investment income in ICTA ss.238 and 241.[11]DMG has pleaded in this case, and the Revenue admit, that, if section 247 had permitted group income elections to be made between a United Kingdom subsidiary and a parent company in another Member State, elections would have been made between DMG and DBI, between DMG and DBAG, and between DBI and DBAG. The corollary is that, because section 247 appeared clearly not to permit group income elections to which a parent company in another Member State was a party, the companies did not attempt to make group income elections. There is no doubt that, if they had attempted to make them, the Revenue would have rejected the elections and pointed to the clear terms of the United Kingdom statute in justification of the rejection.[12]Over the years several dividends flowed through from DMG to DBAG, either directly as to 14% of them or through DBI as to 86% of them. In the case of all of them DMG paid ACT. As respects some of the ACT payments no issue arises in the present case (because no limitation arguments arise), but as respects three dividends and associated ACT payments there are limitation issues in dispute which I must decide. Before I give the dates of the dividends and ACT payments it is useful to record that DMG issued its claim form, commencing the present action, on 13 October 2000, served its particulars of claim on 9 February 2001, and amended the particulars so as to give specific details of the ACT payments which this case is about on other dates. One amendment, giving particulars of an ACT payment made in 1993, was made on 16 August 2001. Another amendment, giving particulars of two ACT payments made in 1995 and 1996 was made on a date which must be taken to have been 19 August 2002. The periods which ran for six years up to those dates began on 14 October 1994, 10 February 1995, 17 August 1995, and 20 August 1996.[13]The three ACT payments which are relevant for this judgment were made on 14 October 1993, 15 February 1995, and 14 January 1996. At a later point I shall go into more detail on these matters, but it will be apparent from the dates which I have just given that a limitation issue certainly arises as respects the ACT paid on 14 October 1993, and at least might arise also as respects the ACT paid on 15 January 1995 and on 14 January 1996. If the section 2 limitation period (as opposed to the section 32 limitation period) started to run on the date of payment of the ACT, six years from the payment of the ACT on 14 October 1993 had expired before DMG had even issued its claim form. The position is more complicated as regards the two subsequent payments of ACT (on 15 January 1995 and 14 January 1996), because it gets involved with questions of whether what mattered was the service of the particulars of claim or rather the making of the amendments to the particulars of claim. I will explain about that in a later section of this judgment. Does English law recognise a claim in restitution to recover taxes paid under a mistake of law?[14]This question subdivides into two questions. The first is whether English law recognises a claim in restitution to recover money paid under a mistake of law. If the answer to the first question is yes, the second question arises: does it make any difference if the money paid was paid to a Revenue authority in discharge of what was, or at least was thought to be, a liability for taxes?[15]The answer to the first question is now clear, as a result of the decision of the House of Lords in Kleinwort Benson. English law does recognise a claim in restitution to recover money paid under a mistake of law. The Kleinwort Benson case is so celebrated that it would be excessive for me to describe it in any detail. It is sufficient to say that a bank (Kleinwort Benson) had made payments to a local authority (Lincoln in the particular case dealt with by Their Lordships) under swap agreements which the bank thought, in common with virtually everybody who had a view on the point at the time, were legally enforceable. An intervening decision of the House of Lords (Hazell v Hammersmith and Fulham LBC [1992] 2 AC 1) showed that the bank and virtually everybody else were mistaken about the law: the swap agreements were not legally enforceable after all. Kleinwort Benson sued to recover the payments which it had made. It had limitation problems, but it sought to overcome them by relying on a right to recover payments made under a mistake of law and also by relying on section 32 of the Limitation Act 1980.[16]The majority in the House of Lords (Lords Goff, Hoffman and Hope) held that Kleinwort Benson was entitled to succeed. Under English law money paid under a mistake of law was recoverable; earlier cases to the contrary were overruled. And section 32(1)(c) did operate to postpone the start of the six years limitation period until the time when the mistake of law was discovered or discoverable with reasonable diligence. The majority also took the view that Kleinwort Benson had indeed made the payments to Lincoln under a mistake. That was one respect in which the minority differed. Lords Browne-Wilkinson and Lloyd considered that, because the bank had made the payments in accordance with what it and everyone else had believed to be the law at the time, it had not made them under a mistake. The contrary view of the majority on that issue is summarised in the following passage from the speech of Lord Goff (at p.379): To me, it is plain that the money was indeed paid over under a mistake, the mistake being a mistake of law. The payer believed, when he paid the money, that he was bound in law to pay it. He is now told that, on the law as held to be applicable at the date of the payment, he was not bound to pay it. Plainly, therefore, he paid the money under a mistake of law, and, accordingly, subject to any applicable defences, he is entitled to recover it. That passage does conclude with a suggestion that there might be applicable defences, but Lord Goff went on to consider whether there indeed were any, and concluded that there were not, at least so far as the instant case was concerned. He did think it possible that there ought to be some defences, but he did not consider that the House of Lords could introduce them as a matter of case law.[17]I will consider later how the foregoing aspects of the House of Lords’ decision in Kleinwort Benson apply to the facts of the present case. For the moment I wish to address the second question which I identified in paragraph 14 above: does it make any difference if the money was paid to a Revenue authority in discharge of what was, or at least was thought to be, a liability for taxes? In principle it is hard to see why that should make any difference, but the question must be considered with some care because of three paragraphs in Lord Goff’s speech (at pages 381 and 382 in a speech which extends from page 365 to page 389). I believe that I ought to quote them in full: At this point it is, in my opinion, appropriate to draw a distinction between, on the one hand, payments of taxes and other similar charges and, on the other hand, payments made under ordinary private transactions. The former category of cases was considered by your Lordships' House in Woolwich Equitable Building Society v. Inland Revenue Commissioners [1993] A.C. 70, in which it was held that at common law taxes exacted ultra vires were recoverable as of right, without the need to invoke a mistake of law by the payer. Moreover reference was made, in the course of the hearing, to the various statutory provisions (usefully summarised in the Law Commission's Consultation Paper (Law Com. No. 120) at pp. 74-84) which regulate the repayment of overpaid tax. For present purposes it is of interest that, in the case of some taxes (including income and corporation tax), no relief is given "in respect of an error or mistake as to the basis on which the liability . . . ought to have been computed where the return was in fact made on the basis of or in accordance with the practice generally prevailing at the time when the return was made:" see the proviso to section 33(2) of the Taxes Management Act 1970. Two observations may be made about the present situation. (I of course have it in mind that this is the subject of proposals for legislative reform contained in the Law Commission's Report (Law Com. No. 227), but your Lordships are concerned with the law as it stands at present.) The first observation is that, in our law of restitution, we now find two separate and distinct regimes in respect of the repayment of money paid under a mistake of law. These are(1) cases concerned with repayment of taxes and other similar charges which, when exacted ultra vires, are recoverable as of right at common law on the principle in Woolwich, and otherwise are the subject of statutory regimes regulating recovery; and(2) other cases, which may broadly be described as concerned with repayment of money paid under private transactions, and which are governed by the common law. The second observation is that, in cases concerned with overpaid taxes, a case can be made in favour of a principle that payments made in accordance with a prevailing practice, or indeed under a settled understanding of the law, should be irrecoverable. If such a situation should arise with regard to overpayment of tax, it is possible that a large number of taxpayers may be affected; there is an element of public interest which may militate against repayment of tax paid in such circumstances; and, since ex hypothesi all citizens will have been treated alike, exclusion of recovery on public policy grounds may be more readily justifiable. In the present case, however, we are concerned with payments made under private law transactions. It so happens that a significant number of payments were in fact made under interest rate swap agreements with local authorities before it was appreciated that they were void; but the number is by no means as great as might conceivably occur in the case of taxes overpaid in accordance with a prevailing practice, or under a settled understanding of the law. Moreover the element of public interest is lacking. In cases such as these I find it difficult to understand why the payer should not be entitled to recover the money paid by him under a mistake of law, even if everybody concerned thought at the time that interest rate swap agreements with local authorities were valid.[18]Is Lord Goff saying in those paragraphs that the cause of action for restitution of money paid under a mistake of law, the existence of which he, Lord Hoffman and Lord Hope are affirming, does not apply to money mistakenly paid in taxes? In my judgment the answer is that he is not. In support of my view I make the following points. i. There is certainly no principle that, for public policy reasons or for other reasons, there can never be restitutionary claims for recovery of taxes wrongly paid. See Woolwich Equitable Building Society v IRC [1993] AC 70, recognising the existence of a restitutionary claim in respect of tax paid under protest in response to what turned out to have been an unlawful statutory demand. ii. If Lord Goff intended to say that, although money paid under a mistake of law was generally recoverable, money paid in taxes under a mistake of tax law was not, he would not have said it at the point in his speech where the three paragraphs which I have quoted appear. The general structure of the speech was to begin with an analysis of whether a cause of action for restitution of payments made under a mistake of law existed at all, and then, having concluded that it did, to go on to consider whether there were any defences to such a cause of action which might apply in particular cases. If His Lordship wished to say that a cause of action for payments of taxes made under a mistake of law did not exist he would surely have dealt with the point in the first part of his speech. However, the paragraphs which I have quoted appear in the second part of the speech, and in particular in a section where Lord Goff is considering whether the House of Lords should introduce a ‘settled law defence’ along the lines of a proposal which had been made by the Law Commission. The Commission had considered that legislation would be necessary to create a settled law defence, and in a draft Bill had included a provision expressed as follows: ‘An act done in accordance with a settled view of the law shall not be regarded as founding a mistake claim by reason only that a subsequent decision of a court or tribunal departs from that view.’ iii. In the context I do not believe that Lord Goff was saying that there could never be a restitutionary claim for tax paid by mistake. Rather he was observing that, although there could be such a claim, the courts might at some future time have to consider whether there was also a settled law defence. In the Kleinwort Benson case itself he went on to decide that (at common law, and leaving out of account possible future legislative changes) there was no such defence where the mistaken payments had been made under private law. He did not wish to be understood as deciding the same thing where the payments were tax payments. However, he was not deciding that there was a settled law defence in the case of tax payments. He was leaving the matter open. iv. In the present case the submission put to me by Mr Glick QC on behalf of the Revenue was that a cause of action for the restitution of tax payments made under a mistake of law did not exist at all. He expressly said that, if I considered that the cause of action did exist, he did not invite me to decide at High Court level that a settled law defence might be available. He reserved the right to advance such an argument in a higher court, and I gave permission for an amendment to the Revenue’s defence which would lay the basis for such an argument if the Revenue wish to advance it hereafter. v. In the earlier part of Lord Goff’s speech, where he discussed whether there should be a cause of action in restitution for money paid under a mistake of law, he reviewed a number of authorities in other jurisdictions which had held that such a cause of action did exist. Among them was the South African case of Willis Faber Enthoven (Pty.) Ltd v Receiver of Revenue, 1992 (4) SA 202. That was a tax case in which a restitutionary claim for the recovery of tax paid under a mistake of law succeeded. The impression which I glean from the speech is that Lord Goff approved of the Willis Faber decision and saw no objection to a similar result forming part of United Kingdom law. vi. In none of the other speeches in the House of Lords, either of the majority or of the minority, is there any suggestion that there could be a difference between payments of tax made under a mistake of law and payments of different kinds made under a mistake of law.[19]There is one last point to make on this aspect of the case. I ought to mention the judgment of Brooke LJ in Eagerpath Ltd v Edwards (2001) 73 T.C. 427. The case was about a technical issue of tax appeal machinery involving an ‘error or mistake claim’ under section 33 of the Taxes Management Act 1970. However, there are some observations of Brooke LJ which assume that a result of the Kleinwort Benson decision in the House of Lords is that a taxpayer who has paid too much tax because of a mistake of law has a cause of action at common law entitling him to recover the overpayment. However, the observations were clearly obiter. I am informed that the issue of whether the Kleinwort Benson decision extended to tax payments had not been raised in argument. I think it is unlikely that Brooke LJ considered the three paragraphs from Lord Goff’s speech which I have quoted. In the circumstances I do not attach much weight to his observations, but they are at least a tentative opinion of a Court of Appeal judge that tax payments made by mistake can be recovered. Did DMG make its payments of ACT under a mistake of law?[20]In my opinion the answer is: yes. However, the question is not entirely straightforward. In my view it is more complicated than the equivalent question which arose in the Kleinwort Benson case, and it requires careful analysis. In that case Lord Goff said, in a passage which I have already quoted:
‘The payer believed, when he paid the money, that he was bound to pay it. He is now told that, on the law as held to be applicable at the date of the payment, he was not bound to pay it.’
In this case it cannot be put as simply as that. There are two different reasons why that is so.[21]The first reason is this. The relevant United Kingdom tax provisions did not did not take the simple form of providing that, if a subsidiary paid a dividend to another company which was in the same group, it did not have to pay ACT. They provided:(1) that if the companies were in a group relationship, they could make a group income election;(2) that if there was a group income election in force at the time of payment of a dividend, ACT was not payable (unless the companies specifically gave notice to the contrary as regards a particular dividend);(3) that a group income election did not take effect for three months or, if earlier, until the inspector notified the companies that he accepted it; and(4) that an election would be of no effect if, in those three months, the inspector notified the companies concerned that he was not satisfied that it was validly made. See generally ICTA s.247(2) and (3), s.248(2). On the last point, if the inspector gave notice that he was not satisfied that an election was validly made the companies could appeal, but unless and until the appeal was determined in the companies’ favour, there was no group income election in force. It followed that as a matter of law ACT was properly payable in respect of dividends paid in the meantime, and even a successful appeal against the refusal of a group income election would not change that position. The dividends paid while the appeal was pending continued to give rise to ACT liabilities, and the success of the appeal would only be relevant to future dividends paid after the appeal decision. See as to this observations of the CJEC in paragraph 104 of its judgment in Metallgesellschaft/Hoechst.[22]The point which emerges from the foregoing is that, on a close analysis, the mistake of law which DMG made was not that it paid ACT which was not payable: the ACT was as a matter of law payable when DMG paid it, and the decision of the CJEC in Metallgesellschaft/Hoechst does not mean that it was not payable. Rather the mistake of law which DMG made was that it did not realise that it and DBI could have made group income elections with DBAG, which would have had the effect of preventing the ACT from being payable.[23]The second reason why this case is not exactly the same as one covered by the passage from Lord Goff’s speech in Kleinwort Benson is this. His Lordship he describes a situation where :
‘The payer believed, when he paid the money, that he was bound in law to pay it.’
In the present case it is not as clear cut as that, at least as regards the last of the three ACT payments to which the case relates. The special feature which affects this case is that Hoechst commenced its action against the Revenue in April 1995, and DMG, through its Head of Taxation Mr Thomason, knew about it at the latest by early July 1995. I will consider the implications of this in subsequent paragraphs, but I record now that Mr Glick on behalf of the Revenue seeks to make much of it. In the simplest form his argument is that, if DMG paid ACT at a time when it knew that the United Kingdom statute law was being challenged as contrary to Community law, it cannot say that it paid the ACT under a mistake. The argument is more complicated as regards the first and second ACT payments, both of which were paid before Hoechst commenced its challenge. However, it will be recalled that under section 32(1) of Limitation Act 1980 the limitation period begins to run when the claimant discovers his mistake (or could with reasonable diligence have discovered it). Mr Glick argues that, even if DMG was suffering from a mistake when it made the first two ACT payments, it discovered its mistake by July 1995 when its Head of Taxation learned of the argument which Hoechst was putting forward.[24]In my view, however, DMG did pay the ACT, including the third amount of ACT paid on 14 January 1996, under a mistake of law. I do not accept Mr Glick’s submissions about the effect of DMG’s knowing of Hoechst’s argument from mid-1995 onwards. In this respect I do not think that my reasoning precisely follows the arguments of counsel. To some extent the arguments explored the question of what degree of doubt there had to be about a proposition before the person concerned can be said to be suffering from a mistake or not. As my reasons, explained in the ensuing paragraphs, will show, I do not see the issue in quite that way.[25]At the times of the first and second dividends and of the two ACT payments to which they gave rise DMG knew nothing about the argument that, by virtue of EC law, it and DBI were entitled to make group income elections with DBAG despite the express provision in section 247 that they were not. Because of the CJEC decision in Metallgesellschaft/Hoechst DMG knows now that the true state of the law then was that the companies were entitled to make group income elections; the Revenue knows that now as well; indeed we all do. But we did not know that then. If the true state of the law had been known at the time the companies would have made elections and DMG’s dividends would have been paid under them as group income, not attracting a liability on DMG’s part to pay ACT. At this stage in the analysis it must be assumed that, if the true state of the law had been understood, both DMG and the Revenue would have understood it. In other words we know now that DMG, DBI and DBAG were mistaken in thinking that DMG and DBI could not make group income elections with DBAG; if they had not been mistaken in that respect they would have made elections; if they had made elections all of the DMG dividends which in fact attracted ACT would not have attracted ACT. Therefore the ACT which DMG in fact paid was paid under a mistake of law: if the mistake had not existed the ACT would not have been paid. It is, however, important to appreciate that the mistake was not directly a mistake about whether there was a liability to pay ACT. It was directly a mistake about whether group income elections could be made. The liabilities to pay ACT arose as secondary consequences of that primary mistake.[26]I move on now to consider whether it makes any difference that by July 1995 DMG knew that Hoechst was challenging the provision in section 247 that, for a group income election to be possible, both companies had to be resident in the United Kingdom, and that the basis of the challenge was that the provision was unlawful because it contravened articles of the EC Treaty. I will deal first with the third ACT payment of just under £3.5m, which was paid by DMG on 14 January 1996. The payment related to two dividends, both paid on 17 July 1995: one paid direct to DBAG on its 14% shareholding, and the other paid to DBI on its 86% shareholding. The evidence shows that the DBAG group, including DMG, knew about the Hoechst argument before 17 July 1995, but made no attempt to get group income elections in place. So in the context of the third ACT payment the question is whether, given that DMG knew about the argument, DMG can still say that it paid the ACT under a mistake. In relation to the first and second payments the detailed facts are different: the ACT payments were made in October 1993 and February 1995 in respect of dividends paid in July 1993 and July 1994: at none of those times did DMG know about the Hoechst argument. For those ACT payments the question is: if DMG made the payments under a mistake of law (which in my opinion it did), did it discover the mistake in 1995 when it discovered about Hoechst’s argument?[27]Before I move on I wish to record the evidence which Mr Thomason (the Head of Taxation at DMG) gave about his state of mind when he learned of the Hoechst argument. In his witness statement he said this: [A]t all times prior to the determination of the European Court in the Hoechst case, I believed that the UK statute denying the ability to make a group income election was the law and I was bound to act in accordance with this law. … It did not occur to me that I could ignore the law as it stood for the simple reason that the law is the law. Just because another taxpayer challenged the law that did not mean that I could or should ignore it. Mr Thomason was cross-examined, but I do not believe that the foregoing passage was challenged or affected by his answers on other points. He added the general point (obvious but plainly relevant) that it was not clear in 1995 what would be the outcome of the Hoechst case (or, as it turned out, of the Metallgesellschaft/Hoechst conjoinedcases). He also said that there were more arguments being advanced by Hoechst in 1995 than the one which eventually succeeded in the CJEC. For example, relief was being claimed by Hoechst AG, the German parent company, as well as by the United Kingdom subsidiary which had paid the dividends and the associated ACT. In addition arguments were being advanced in reliance, not on the EC Treaty, but on the non-discrimination article in the Double Taxation Agreement between the United Kingdom and Germany.[28]I now proceed with my own analysis of the position. It is obvious that, when DMG learned of the argument which Hoechst was advancing, it had come to know something – to discover something – which it had not known before. That is true, but DMG did not come to know, or to discover, that it and DBI could make group income elections with DBAG after all. All that it discovered was that another company was arguing that it could make a group income election with its German parent company in comparable circumstances, and that the Revenue were disputing the argument. I examine the position as matters actually stood before DMG paid the dividend which it was planning to pay and did pay in July 1995. If at that time DMG and DBI had submitted group income elections with DBAG, it is in my opinion certain that the inspector would have rejected the elections within the three month period which was available to him under section 248(2). Short of DMG abandoning the intention to pay dividends altogether (which I assume would have been unacceptable to the group as a whole) DMG would have had to pay the ACT. Further, such ACT would have been correctly paid even if DMG had appealed against the rejection of the elections and a later decision (like the actual decision in Metallgesellschaft/Hoechst) showed that the inspector ought to have accepted them. (See as to this paragraphs 21 and 22 above.)[29]Concentrating for the moment on the first and second ACT payments (the payments associated with the dividends which DMG paid in July 1993 and July 1994), I repeat that the mistake which DMG made at those times was that it did not realise that it and DBI could have made valid group income elections with DBAG. In 1995, when DMG learned of the case which Hoechst had begun against the United Kingdom Revenue, it did not discover that it (DMG) and DBI could have made group income elections with DBAG after all. All that it discovered was that there was a possible argument that they could have done that. So it did discover something, but not something which made any difference to the matters which are critical in this case. Suppose that back in 1993, when DMG was planning to pay the first of the dividends, it had already known what it actually discovered in 1995: that there was an argument that group income elections could be made with parent companies in Member States other than the United Kingdom, but that the Revenue did not agree with the argument. Would DMG, DBI and DBAG have submitted group income elections then? I do not think that they would. Still less do I think that, if they had submitted elections then, DMG would not have been liable to pay ACT when it paid the dividends. I have no doubt that the Revenue would have rejected the elections, and that DMG would have been liable to pay the ACT and would have paid it.[30]Accordingly, my conclusions in relation to the first two ACT payments are(1) that DMG did pay them under a mistake of law;(2) that, unless a limitation defence applies, DMG is entitled to claim restitution on the basis of a mistake of law;(3) that DMG did not discover its mistake in 1995 when it learned about the argument which Hoechst was advancing;(4) that therefore the limitation period applicable to DMG’s mistake of law claim did not begin to run in or about July 1995; and(5) that no limitation defence applies. In my opinion DMG did not discover its mistake until the decision of the CJEC in Metallgesellschaft/Hoechst was released on 8 March 2001. DMG had already commenced its action claiming relief before then: it did so shortly after the Advocate General’s opinion was delivered. That opinion was in favour of the claimants, and thus it foreshadowed the later decision of the CJEC itself. It might be arguable that DMG discovered its mistake at the time of the Advocate General’s decision. I would not take that view myself: it was not inevitable that the CJEC would take the same view as the Advocate General. But the point does not matter, because, even if the limitation period started run at the time of the Advocate General’s opinion, DMG commenced its action only a month and a day later (12 September 2000 and 13 October 2000).[31]As regards the third ACT payment – the one made on 14 January 1996 – DMG knew about the Hoechst argument when it paid the ACT, and it had also known about the argument when it and DBI made no attempt to make group income elections before it paid the dividends on 17 July 1995. Notwithstanding that, I accept that DMG paid the ACT under a mistake of law, and is in principle entitled to restitution for the timing disadvantage which it thereby suffered. The limitation period would only start to run against DMG on the release of the CJEC decision, by which time it had already commenced its claim for relief.[32]The result is that in my judgment DMG’s claims succeed and that none of them is statute barred. I will, however, briefly indicate what I believe the position would have been if I am wrong and if, once DMG knew of the Hoechst argument in 1995, it was no longer suffering from a relevant mistake. What if (contrary to my view) DMG was no longer mistaken once it learned of the argument being advanced by the Hoechst group?[33]I will deal first with the first two ACT payments, which were made before DMG knew about the Hoechst argument. On the hypothesis which I am now assuming DMG was under a mistake when it made the payments but discovered its mistake on a date between April 1995 (when the Hoechst action was commenced) and mid-July 1995 (by which time there is no doubt that DMG knew about the Hoechst case). So the limitation period started to run on a date in that period. I do not know exactly which date, but for illustration purposes in the discussion which follows I will assume that it was 1 July 1995. For DMG’s claim for restitution to escape being statute barred it would have had to have been commenced before 1 July 2001. DMG’s claim form was issued on 13 October 2000, and the particulars of claim were served on 9 February 2001. On the face of it both were within time, whether the relevant date is taken to be that of the claim form or that of the particulars of claim.[34]However, there is a complication which arises from the detailed contents of DMG’s pleadings. The question is whether the claim form and the particulars of claim as they were originally expressed always claimed relief in respect of the ACT payments made in October 1993 and February 1995, or whether claims in respect of those particular ACT payments were new claims added by amendments made on 16 August 2001 (as respects the October 1993 ACT payment) and on 19 August 2002 (as respects the February 1995 ACT payment). If they were new claims added to the particulars of claim by the amendments, and if (contrary to my view) the limitation period had started running on or about 1 July 1995, the claims would be time barred by the Limitation Act. In my opinion, however, the original particulars of claim, on their proper interpretation, already claimed relief in respect of the ACT payments made in October 1993 and in February 1995.[35]This is a fine point on the small print wording of the particulars of claim, and I will not go into it at length. The critical point is that the particulars claimed relief in respect of all payments of ACT made by reference to ‘the Dividends’. ‘The Dividends’ was a defined term: it meant the dividends which DMG had paid from time to time to DBI and DBAG, thus (as it seems to me) covering all dividends, including those which had led to the payments of ACT in October 1993 and February 1995 (the first and second payments which are in issue in this case). The original particulars of claim did say that the payments of ACT ‘include but are not limited to’ certain payments set out in a schedule. The payments so set out did not include the payments made in October 1993 and February 1995: those payments were only added to the schedule by the amendments made more than six years after 1 July 1995. However, in my view the original particulars of claim claimed relief in respect of all payments of ACT which had been made by DMG in respect of all dividends paid by it to DBI and DBAG. The claim was not limited to the payments specified in the schedule.[36]In this respect the original particulars of claim were different from those which I considered in the recent case of Hoechst United Kingdom Ltd v IRC [2003] EWHC 1002 (CH). The particulars of claim in that case (as I interpreted them) only claimed relief in respect of the specific ACT payments which were identified in a schedule, so that when the schedule was amended to add further payments the amendment had the effect of adding new claims in respect of additional causes of action. In the present case, in contrast, I consider that the amendments did not add new claims, but rather gave further details of claims which had already been made within the causes of action which had already been pleaded. Therefore, even if, contrary to my view, the six years limitation period started to run in July 1995 as respects the first two payments of ACT which are in issue, I still consider that DMG’s claims for relief in respect of those two payments were brought within the period and are not statute barred.[37]I turn to the third ACT payment, the one of about £3.5m which was made on 14 January 1996. The result is the same as respects that payment, but the precise route by which the result is reached is different. When the payment was made, and also at the earlier time when the dividend which gave rise to it was paid (on 17 July 1995), DMG already knew about the argument which Hoechst was advancing. In my view, for the reasons which I have explained earlier, that did not change the position that DMG paid the ACT under a mistake of law. However, Mr Glick has submitted that it did, and that DMG did not pay the January 1996 ACT under a mistake. Let me suppose that he is right. Three specific consequences would follow:(1) DMG’s claim for restitution, in so far as it rests on the Kleinwort Benson principle that money paid under a mistake of law is recoverable, would fail.(2) However, DMG would have a good claim for compensation or restitution based on the different ground that it had paid the money pursuant to an unlawful taxation measure (as in Woolwich Equitable Building Society v IRC [1993] AC 70). (3). The limitation period for that cause of action began to run immediately on payment of the ACT, on 14 January 1996. Therefore DMG would need to commence its action before 14 January 2002.[38]As I have already recorded the claim form was issued on 13 October 2000 and the particulars of claim were served on 9 February 2001, both comfortably within the six years period. Therefore DMG’s claim in respect of the last of the three ACT payments would have been in time anyway, as long as the claim as pleaded in the original particulars did claim relief in respect of that ACT payment. In my judgment it did, for the same reasons as those which I explained in paragraph 35 above. The particulars claimed relief in respect of all ACT payments made by DMG, not only in respect of those associated with the dividends which were scheduled to the pleading. When the schedule was amended, effectively on 19 August 2002 and thus too late to be within the six years limitation period which ended on 13 January 2002, the amendment did not plead a new cause of action, but rather gave further factual details of a cause or causes of action which had already been pleaded in the original particulars served on 9 February 2001. Therefore, even if DMG, by reason of knowing about Hoechst’s argument, did not make its January 1996 ACT payment under a mistake, it still brought its claim for relief in respect of that payment within the time permitted by the Limitation Act.[39]There is one other detailed matter to be mentioned in connection with limitation. In my judgment in Hoechst United Kingdom Ltd v IRC,[2003] EWHC 1002 (CH) I suggested the possibility that, in claims like those of Hoechst and of DMG in this case, the limitation period may not have expired all at once on the sixth anniversary of the date when a payment of ACT was made, but progressively over the period from that sixth anniversary date until the later sixth anniversary of the date when the ACT was set off against MCT. The theory behind the suggestion is that the relief is calculated on the basis of notional interest over that period, and thus may be regarded as accruing from day to day. It seems to me to be at least arguable that the limitation period would also run from day to day and would expire from day to day six years later.[40]In the oral argument in this case I drew attention to the possible argument. Counsel had not considered it in preparation for the hearing, so it was agreed that they would not address it in their oral arguments, but might supply me with written submissions on it after the hearing. They did that. Mr Rabinowitz QC and Mr Fitzpatrick on behalf of DMG supported the argument. Mr Glick QC and Mr Carr on behalf of the Revenue said that it was wrong. In the event the question does not arise on the view which I take of this case, so, while I am grateful to counsel for their careful written submissions upon it, I hope that they will forgive me if I do not express a view on it. It is better to leave the point to be decided in a future case where it does matter. Conclusion[41]In the result I conclude that DMG’s claims, in so far as they are formulated as claims for restitution in respect of money paid under a mistake of law, succeed in principle, and that they are not time-barred by the Limitation Act 1980. I wish only to add that this is not a result which I reach with much enthusiasm. When Parliament first enacted the predecessor of section 32(1)(c) of the Limitation Act 1980 (postponing the running of the limitation period in claims based on mistake until the mistake is discovered), it did not have in mind claims for recovery of money based on a mistake of law, particularly a mistake which was shared by almost everyone else at the time. This was one consideration which influenced the dissenting minority in Kleinwort Benson. Lord Browne-Wilkinson said: On every occasion on which a higher court changed the law by judicial decision, all those who had made payments on the basis that the old law was correct (however long ago such payments were made) would have six years to bring a claim to recover money paid under a mistake of law. The three Law Lords in the majority clearly felt that the limitation consequences of their decision would be unacceptable in the long term and would have to be changed. Thus Lord Goff said (at p.389): I recognise that the effect of section 32(1)(c) is that the cause of action in a case such as the present may be extended for an indefinite period of time. … [T]he recognition of the right at common law to recover money on the ground that it was paid under a mistake of law may call for legislative reform to provide for some time limit to the right of recovery in such cases. The Law Commission may think it desirable, as a result of the decision in the present case, to give consideration to this question; indeed they may think it wise to do so as a matter of some urgency. The Law Commission considered the matter in its major report on Limitation of Actions, dated 9 July 2001 (see paragraphs 2.49, 2.90 and 4.76 to 4.79). However, no legislation consequent upon that report has yet been enacted.[42]In the circumstances I believe that I am bound by the decision of the majority in the House of Lords in Kleinwort Benson to hold that a cause of action for recovery of money paid under a mistake of law does exist, and that where such an action is commenced section 32(1)(c) of the Limitation Act 1980 applies with the effect that the limitation period does not begin to run until the mistake is discovered. For the detailed reasons which I have given in this judgment, I consider that DMG did make the three payments of ACT under a mistake of law; that DMG is therefore entitled to restitution; that it makes no difference that the payments were tax payments; and that DMG’s claims for restitution are not time-barred by the 1980 Act. For those reasons I will give judgment in favour of DMG.

Cited in 12 later judgments

and 3 more of the most senior