[2012] UKFTT 485 (TC)
Value Added Tax - Whether an appeal under s. 80 VAT Act 1994 to recover an amount invoiced as VAT ( decided in 2006 by both parties not to have constituted actual VAT in respect of an actual supply) should be dismissed on the ground that the transaction had been abusive, because it would offend neutrality to allow the appeal when HMRC had been out of time in an earlier appeal to reclaim the corresponding amount initially allowed as input tax to the counter-party in the original transaction - Application of paragraph 5(2) Schedule 11 VAT Act, 1994 - Whether the doctrine of abuse applied in relation to transactions assumed initially to be zero-rated, and thus not expressly dealt with by the Sixth Directive - Whether the feature that the transaction was undertaken but that no supply was in retrospect treated as having been made itself undermined any possible contention by HMRC that there had been an abuse - Whether the BUPA drugs and prostheses appeal could equally have been decided on an abuse approach - Appeal allowed
TAX CHAMBER
-and-
THE COMMISSIONERS FOR HER MAJESTY’S Respondents
GILL HUNTER
Roderick Cordara QC and David Scorey, counsel, on behalf of the Appellant
Owain Thomas, counsel, on behalf of the Respondents
©CROWN COPYRIGHT 2012
DECISION
1. This appeal was the matching appeal to an appeal heard in July 2008 by the VAT and Duties Tribunal, involving the parent company, namely St. Martin’s Healthcare Limited (“Healthcare”) of the present Appellant, St. Martin’s Medical Services Limited (“Services”). The Tribunal Judge in the present Appeal was the Chairman in Healthcare’s earlier appeal.
2. The transaction that occasioned both Healthcare’s appeal and Services’ present appeal was an artificial transaction effected in December 1997, designed to prepay for drugs, and designed therefore to defer for some period the feature that a proposed change in the law to come into effect on 1 January 1998 would result in drug supplies to in-patients by Healthcare ceasing to be zero-rated, albeit that supplies to out-patients were to remain zero-rated.
3. The scheme involved Healthcare prepaying £10 million to Services in return for Services’ commitment to deliver drugs for in-patient use in later periods. Services accounted for £1,489,361 in respect of that assumed supply; Healthcare claimed, and in fact received in cash from HMRC, an input deduction for a similar amount, and the parties hoped that when Services later purchased drugs from third-party suppliers (Glaxo for instance), the VAT in respect of those supplies would be reclaimable by Services, since the matching supply would have occurred back in 1997. The benefit of the scheme consisted essentially in the fact that if Services recovered the VAT that it had initially paid on contracting to make the future deliveries, when it was invoiced by Glaxo etc. on buying drugs (so ending up in a neutral position), then Healthcare would have the enduring benefit of its initial refund of the £1,489,361. Had Healthcare instead received “supplies” only after 1 January 1998 it would have recovered virtually none of the VAT charged on the supplies to it.
4. In late December 2000, HMRC challenged the scheme, claimed that the initial prepayment for future deliveries had in fact occasioned no supply for VAT purposes at all, and sought to assess and recover from Healthcare the input tax that had initially been refunded to Healthcare. Simultaneously HMRC drew Services’ attention to the fact that Services could correspondingly make a Voluntary Disclosure claim to recover the amount that, on HMRC’s analysis, Services had wrongly paid as VAT. Services duly made this claim.
5. Pending the outcome of the challenge on the fairly similar scheme effected by BUPA Hospitals Limited, the matching assessment and Voluntary Disclosure claim were left in abeyance. The 2006 decision of the ECJ in the case of BUPA Hospitals Limited v. HM C&E [2006] STC 967 against the taxpayer resulted however, in Healthcare conceding that its substantive grounds for disputing HMRC’s assessment would fail, and thus those substantive grounds were all withdrawn.
6. Whilst Healthcare thus withdrew its substantive contentions, Healthcare still pursued its appeal on the single ground that HMRC had failed in late December 2000 to make its assessment on Healthcare in time. Services’ Voluntary Disclosure had plainly been made in time, so that the implication of Healthcare’s appeal was that if it prevailed on the “out of time” point, and if Services then pursued its Voluntary Disclosure claim for repayment, HMRC would fail to recover the input tax refund initially made to Healthcare, but would still face Services’ seemingly valid, and “in time”, claim to recover the corresponding amount that it had initially paid. Healthcare won its appeal in 2008. HMRC indicated at the time of that appeal that if Services continued to pursue its Voluntary Disclosure claim, HMRC would be likely to resist that claim on the ground that it was abusive, and that Services would be unjustly enriched were that claim to be allowed. Services has pursued its claim, and this appeal must thus determine whether that Appeal is valid, or whether HMRC’s abuse contention means that it should be dismissed. In 2009, HMRC added an additional argument, geared to the charge under paragraph 5(2) Schedule 11 VAT Act, 1994 on any person issuing a VAT invoice, to its contentions, and moreover suggested that we had no jurisdiction to hear the argument on that further point.
7. Our decision is that notwithstanding the resultant lack of “neutrality”, and indeed notwithstanding that on the “out of time” point the group ends up in a marginally better position than it would have been in had the scheme succeeded in full, the Appeal should be allowed.
8. The text of this Decision up to the end of paragraph 122 deals with the contentions that were explored fully during the hearing. We add that at the very end of the hearing, it was contended on behalf of the Appellant that, were we to be inclined to decide the Appeal against the Appellant on the “abuse” and “neutrality” ground, there was a further way of approaching and considering the “neutrality” contention that we should address, notwithstanding that this contention had nowhere been referred to in the Notice of Appeal or the Skeleton Argument. Since the point was only discussed briefly, and since the limited attention given to it detracted from the Respondents being able to make full submissions in relation to it, we asked the parties to send us written submissions on the point. In the event the parties have done that, and the Respondents have made further submissions on the “abuse” point, they have also somewhat rephrased their contentions in relation to the Moorbury case, and made additional submissions in relation to the reversal of any charge under paragraph 5(2) Schedule 11. We deal with all these points fully, and in a distinct section at the end of this Decision, commencing at paragraph 122.
9. No evidence was given during the hearing, since there were simply two main legal disputes. The first was the issue of whether HMRC was right in its contention that on any version of an “abuse” ground, Services’ appeal should be dismissed on account of the way in which it sought to undermine “neutrality”. The second issue was that Services had been liable for VAT or some equivalent liability to VAT under paragraph 5(2) Schedule 11 VAT Act 1994, simply by having issued a VAT inclusive invoice to Healthcare, and that there was either no entitlement to recover amounts charged under paragraph 5(2) in any circumstances, or that at the very least the right to recover amounts paid under paragraph 5(2) was far more restricted than simply claims to recover wrongly paid output VAT. It was additionally contended that the Tribunal had no jurisdiction to hear that latter point.
10. The earlier Healthcare appeal had revolved almost entirely around evidence. The limitation rules then in force enabled HMRC to raise an assessment within a 2-year period of the end of Healthcare’s VAT period in which the input tax was claimed, or in the one further year if HMRC could establish that new evidence or information had only come to their notice during that third year that enabled them to issue an assessment. Virtually everything there depended on the level of information that had been obtained by HMRC during the 2-year period, and on whether anything remotely material had been revealed to HMRC during the third year that justified, for the first time, the making of the assessment.
11. The evidence in the Healthcare appeal was all summarised in the VAT and Duties Tribunal decision, [2008] UK VAT V20778 (20 August 2008). Insofar as that evidence remains relevant, we will refer to it in summarising the background and the facts now, in more detail.
12. Healthcare’s scheme had involved arranging for a hitherto dormant subsidiary, Services, to contract with Healthcare to supply £10 million worth of drugs for in-patient use to Healthcare, for which Healthcare paid Services the £10 million on 23 December 1997, i.e. in advance of the law change.
13. There is no present dispute that the scheme was implemented hastily and badly. In view of contentions on behalf of the Respondents in this Appeal about the artificial nature of the scheme, we should mention that the directors of the two companies were largely “common”; it is not even clear, whilst an invoice was issued for the purported supply, that the contract for the future supply between Services and Healthcare was actually entered into. If it was, it appears that the intention was that when drugs were later delivered by Services to Healthcare, drugs acquired from a supplier (to use the example often used during the hearing, say drugs purchased by Services from Glaxo for a VAT-inclusive price of £100) would be treated, for some rather odd reason, as reducing by only £80 Services’ contractual commitment to supply £10 million worth of drugs to Healthcare. [We should perhaps clarify that our interpretation of Services’ obligation to deliver drugs in the future referred (notwithstanding that the delivery would not constitute a supply for VAT purposes) to drugs whose “VAT-inclusive price” would have equalled the required target figure. Accordingly we took the 20% discount figure to mean that a delivery of drugs with what would normally have been a VAT-inclusive price of £10 million would have discharged only 80% of the commitment.]
14. It was certainly clear that Services made a virtually immediate loan back to Healthcare of £9 million.
15. As Mr. Thomas, the Respondents’ counsel, emphasised, there were other artificial and non-commercial features. Initially, Services had no drugs licence. The scheme would require 180 third-party suppliers to arrange in future to supply drugs to Services and not Healthcare, at least where drugs were to be for in-patient use. Supplies of drugs by Healthcare to out-patients remained zero-rated, so that in order not to waste the saving potential of the scheme, it was preferable for drugs for out-patient supply not to be delivered in discharge of the prepayment commitment. How and whether and when the contracts with the 180 suppliers were varied, and how the in-patient/out-patient complexity was to be dealt with was never clear.
16. In short, it is now fair to say that the scheme was a tax-driven scheme that was a positive nuisance from a commercial point of view. There was no contention, and more clearly absolutely no justification for a contention, that the scheme was designed to achieve any commercial objective.
The early treatment of the scheme by HMRC
17. HMRC, and in particular the efficient HMRC officer who regularly dealt with Healthcare, had been expecting Healthcare and other hospital groups to effect such a prepayment scheme even before it was implemented. It therefore came as no surprise when KPMG wrote to HMRC in early 1998 indicating that the scheme had been implemented, mentioning how reliance was placed on the decision in favour of the taxpayers in Faith Construction v. HM C&E [1989] STC 539 (a similar prepayment scheme involving building works and “loans back”), and relying also on the feature that prepayment schemes had been implemented and accepted by HMRC when supplies of electricity were first subjected to VAT instead of being zero-rated.
18. Whilst the case officer realised that the scheme was artificial, and she reported this to more senior colleagues, the decision was taken to refund the input tax deduction claimed by Healthcare. The position, thus, by early 1998 was that Services had accounted for VAT of £1,489,361 in respect of the assumed supply, and HMRC had refunded a similar amount in cash to Healthcare.
19. Healthcare and Services were slow to operate the intended post-January 1998 delivery arrangements, largely because of the problems alluded to in paragraphs 13 and 15 above. When, very close to the end of 1998, Services commenced purchasing, and on-delivering drugs to Healthcare, Services reclaimed the input VAT included in supplies from the drug companies, and initially this was duly refunded in cash.
20. HMRC’s desire to challenge the Healthcare scheme was entrusted, not surprisingly, to a high-level team. Whilst the original and the replacement case officers dealing directly with Healthcare reported to this team and attended the team meetings, any decisions as to whether and how to challenge the scheme were made by the high-level team as a whole. The team was considering challenges in relation to schemes by other hospital groups and presumably other quite different schemes, and progress was slow.
21. HMRC’s deliberations in relation to a challenge on the scheme culminated in a letter dated 19 December 2000, which was quoted in full in the earlier decision.
22. It is sufficient to say now that the letter claimed that there had been no supply for VAT purposes when the contract to supply was entered into and payment made and received. The letter went on to volunteer several possible justifications for this contention, including a Ramsay challenge and various articulations of an EU doctrine of abuse.
23. The 19 December letter said that a formal assessment to recover the £1,489,361 that HMRC had initially refunded to Healthcare would be made immediately, as it was. It also mentioned that Services could make a Voluntary Disclosure on the basis that if HMRC was right to say that there had been no supply in 1997, then Services had wrongly accounted for £1,489,361 in VAT which it could now recover.
24. The letter then suggested that if the VAT treatment of both sides of the original transaction was reversed, as suggested, then the further step would be to treat drug supplies from Glaxo etc to Services as if they had been made directly to Healthcare, whereupon Healthcare would only recover the very small percentage of input tax referable to supplies to out-patients, and would be unable to recover that referable to exempt supplies to in-patients.
25. Services duly made its Voluntary Disclosure on 1 February 2001.
26. Services’ timing point for its claim was that the claim could validly be made at any time prior to 31 March 2001. Services had only been registered for VAT purposes in early 1998, backdated to 16 December 1997, with the result that its first VAT period did not close until 31 March 1998, and the 3-year period for making the Voluntary Disclosure claim did not therefore end until 31 March 2001.
27. Shortly after Services’ Voluntary Disclosure claim was made, HMRC issued a document that evidenced that the £1,489,361 was owed to Services.
28. Since it was realised that the outcome of the various claims was likely to be dependent on the outcome of the broadly similar BUPA Hospitals case which was already being litigated, matters were left in abeyance, with Healthcare thus retaining its original input deduction, Services not yet receiving its repayment, and no interest being charged against Healthcare on the assessment, since HMRC essentially had the cash in hand, by not having refunded Services’ payment.
29. We have already mentioned in paragraph 5 above that when HMRC won the BUPA Hospitals case before the ECJ, the immediate significance of this was that on the substantive grounds, Healthcare’s appeal collapsed, but was still pursued simply on the basis that HMRC’s assessment was out of time.
30. There was very little significance at the time to the fact that the BUPA Hospitals decision of the ECJ was not based on any of the grounds initially expected, or the arguments mentioned in HMRC’s letter of 19 December 1997, or indeed as a direct response to the reference actually made by the High Court to the ECJ. There is, however, some present relevance, or possible relevance, to the fact that the ECJ departed from the Opinion of the Advocate-General (that was in fact contained in the case of Halifax PLC v. HM C&E (Case C-255/02) where the Advocate-General gave his opinion on both cases). The ECJ decided that even where a particular point (the “abuse” point) had been referred to it by the court in the Member State, it was perfectly in order to deal with a reference on a different basis if the ECJ considered that the correct approach in a reference was to address some different, and possibly antecedent, point.
31. On this approach, the ECJ decided that since even on the somewhat clearer and more detailed contract in the BUPA Hospitals case, it was inherently impossible to indicate in advance what goods would actually be supplied, and the future deliveries could always be varied or even completely cancelled, the right analysis was that there had, on the “black-letter” law analysis, actually been no supply for VAT purposes at all under the original contract.
32. The Halifax “abuse” doctrine only applied if a transaction had secured a result that produced a tax advantage that was in conflict with the fundamental principles of the Sixth Directive, and the essential aim of the transactions had been to secure that advantage. Where this was demonstrated, the transaction then had to be redefined in a manner to which we will refer in due course. The immediate point, however, was that since the BUPA Hospitals transaction had failed on the “black-letter” law analysis to achieve the result intended, the first requirement for the abuse doctrine to be engaged (securing a tax advantage in conflict with the Sixth Directive) had simply not been satisfied.
33. The point just mentioned was of no immediate relevance to Healthcare when, as already indicated, Healthcare continued to pursue its appeal against HMRC’s assessment solely on the “out of time” point. It does, however, have some potential relevance to this Appeal, which we will deal with below.
34. Healthcare’s appeal was decided in favour of Healthcare, and HMRC did not appeal.
35. The decision of the VAT and Duties Tribunal in the Healthcare appeal was based on the fact that by virtue of information volunteered to Healthcare’s case officer, and further responses to questions, HMRC had enough information by the end of the second year to make assessments, and more relevantly HMRC discovered nothing new during the third year to justify an assessment during that third year.
36. The earlier decision referred to the fact that the high level HMRC officers, who assumed full conduct of the challenge of Healthcare’s scheme, were dilatory and negligent in their slow dealing with that scheme. The reason for the delay was in fact entirely understandable. The Healthcare case was perhaps a relatively minor one amongst numerous artificial cases that HMRC officers dealing with both direct and indirect taxes were dealing with at the time, and it is understandable that more attention was given to wider policy and strategic issues as to how numerous examples of aggressive planning should be tackled. Whilst that explains why people lost sight of the timing points relevant in the Healthcare appeal, it changed none of the facts that:
37. The only facts that thus remain relevant in relation to the earlier appeal are that the appeal was allowed because HMRC’s assessment had been out of time, and secondly some considerable fault or incompetence on the part of HMRC (albeit understandable for the reason just given) occasioned the failure to assess in time.
38. HMRC had intimated prior to and during the Healthcare appeal that if Healthcare had abandoned the appeal, or if HMRC were to win the Healthcare appeal, then HMRC would have correspondingly and readily refunded to Services the amount that Services had wrongly accounted for as if it was VAT back in 1998.
39. It was thus self-evident on the Healthcare side of the equation that since Healthcare and Services could have achieved neutrality without even troubling to appeal, it followed that if Healthcare won the appeal, Services would almost certainly pursue its Voluntary Disclosure claim to recover the £1,489,361 wrongly paid at the outset. For otherwise it would have been completely pointless to have fought the Healthcare appeal.
40. HMRC intimated that if Services did persist in its claim, HMRC would be likely to resist that claim, and assert that the Services’ appeal should be dismissed on the ground that it was abusive or that it involved unjust enrichment to Services.
41. In their letter of 11 May 2009 HMRC confirmed that they did indeed reject Services’ claim, and they added an additional ground for challenging Services’ appeal. They contended that not only was it abusive to claim the refund but that as the amount paid by Services in early 1998 was not VAT in respect of a supply, but an amount wrongly invoiced as VAT, and thus properly chargeable under paragraph 5(2) Schedule 11 VAT Act, 1994, there were additional grounds for resisting the refund.
42. In their Statement of Case, issued on 19 November 2009, HMRC not only repeated the same abuse and paragraph 5(2) points, but for the first time they asserted that because paragraph 5(2) Schedule 11 to VAT Act 1994 did not fall within the appealable matters under section 83 VAT Act 1994, the Tribunal had no jurisdiction to hear the Appeal in relation to that matter.
43. The various contentions in relation to the “abuse” ground for challenging the present Appeal all rely on the proper application of ECJ case law. They are all therefore encompassed in the contentions of the parties (fully summarised below), and require no summary or quotation of any particular legal provisions.
44. The paragraph 5(2) Schedule 11 point does require some explanation.
45. Paragraph 5(2) and 5(3) of Schedule 11 to the VAT Act 1994 provide as follows:
“(2) Where an invoice shows a supply of goods or services as taking place with VAT chargeable on it, there shall be recoverable from the person who issued the invoice an amount equal to that which is shown on the invoice as VAT or, if VAT is not separately shown, to so much of the total amount shown as payable as is to be taken as representing VAT on the supply.
(3) Sub-paragraph (2) above applies whether or not:-
(a) the invoice is a VAT invoice issued in pursuance of paragraph 2(1) above; or
(b) the supply shown on the invoice actually takes or has taken place, or the amount shown as VAT, or any amount of VAT, is or was chargeable on the supply; or
(c) the person issuing the invoice is a taxable person;
and any sum recoverable from a person under the sub-paragraph shall, if it is in any case VAT be recoverable as such and shall otherwise be recoverable as a debt due to the Crown.”
46. It is worth also mentioning the provision of the Sixth Directive, which paragraph 5 is addressing. Article 21.1.c of the Directive provides that:
“The following shall be liable to pay value added tax:
1. under the internal system:
(a) …...
(b) ……
(c) any person who mentions the value added tax on an invoice or other document serving as an invoice.”
Article 203 of the Principal VAT Directive (2006/112) reads:
“VAT shall be payable by any person who enters the VAT on an invoice.”
47. These provisions constitute an important part of the VAT legislation. Were the provisions not in force, it would be possible, in countless situations, for people to abuse the VAT system in fraudulent ways. For instance if a person who is not a taxable person (in other words someone who is neither registered nor registerable) sells an asset to a registered trader, who buys it for £10,000, it costs the trader £10,000 and, special schemes apart, if the trader sells it for £15,000, the registered trader must obviously account for VAT in respect of the VAT-inclusive price of £15,000, having no benefit of an input deduction. If on the other hand the initial seller is fraudulent, he might observe that if he issues a fraudulent and bogus VAT invoice, and purports to sell the asset for, say £12,000 inclusive of VAT, the registered trader might be ready to pay this price because on a VAT-inclusive basis, it assumes that it now has an input deduction for the VAT element of the VAT-inclusive purchase price, so that its net VAT liability on selling for the same VAT-inclusive price of £15,000 will be considerably less. The registered trader may indeed be entirely innocent, and may simply be fooled by the bogus invoice. The initial seller may simply have perpetrated a fraud, knowing that by issuing the fraudulent VAT invoice, he is likely to receive a better price.
48. As a mechanism to defeat this type of arrangement, which may be either fraudulent or equally entirely honest and mistaken, paragraph 5(2) Schedule 11 renders the person who has provided the VAT invoice or who has sold on a VAT- inclusive basis liable for the VAT. We will deal in due course with whether we have jurisdiction to hear an appeal on this ground; whether a claim by the payer for refund can be pursued under section 80, and the EU guidance as to when wrongly paid VAT should indeed be repaid by the taxing authorities.
The contentions on the part of the parties
49. Since this Appeal essentially raised a challenge by the Respondents in relation to what was, their challenge apart, a fairly straightforward appeal seeking to sustain a claim to recover money wrongly paid on the original assumption that Services had made a supply for VAT purposes when it contracted to make later deliveries of £10 million worth of drugs, and since the Appellant’s Voluntary Disclosure was indisputably made “in time”, it probably makes better sense to summarise the Respondents’ contentions, and their challenge to the claim, first, and then to summarise the Appellant’s contentions subsequently. We will summarise the contentions in relation to all points in dispute, albeit that when giving our decision, we will deal separately with the abuse, the paragraph 5(2) and the jurisdiction points.
50. It was contended on behalf of the Respondents that:
The contentions on the part of the Appellant
51. It was contended on behalf of the Appellant that:
The decision on the “abuse” point
52. We will seek to deal with all the points that have been raised in argument in relation to the abuse contention. While we will deal with them, some appear to us to have little or no relevance to the basis of our decision, so that we will deal with them relatively shortly.
53. In order to indicate the essential thread of our decision, we might summarise our decision immediately, albeit that we will explain it in more detail below.
54. We are largely uninfluenced by the various arguments about “all-permeating artificiality”, the doctrine of “abuse” being irrelevant outside the sphere of the Sixth Directive matters, and there inherently being no abuse if there is analysed to have been no supply.
55. The reality seems to us to be that Healthcare and Services implemented a wholly artificial and non-commercial tax avoidance scheme, the object of which was to extend the benefit of zero-rating to transactions that should not have been so treated. However it is described, it was relatively likely to be seen to be objectionable by HMRC, and there was a realistic possibility that the scheme would fail. Indeed, as it was actually implemented, it was actually quite difficult to see that it had much chance of succeeding.
56. The scheme, however, failed to achieve its objective, and there was no doubt that as a substantive matter, the correct result would have been for HMRC to recover the input deduction initially conceded to Healthcare, and for Services to recover the amount initially, and wrongly, accounted for as VAT. None of the parties disputed that.
57. Timing points and limitation points are perfectly legitimate procedural measures, designed to give some finality to disputes, and there has very understandably been no contention in this case that either of the timing points, applicable to Healthcare’s or Services’ appeals was improper or incapable of being complied with by the relevant parties.
58. Where timing points, legitimately taken (which means “taken in time”) block some potentially sound substantive contention (rather than a hopeless claim), it will inevitably be the result of the timing point being taken that some fundamentally right substantive point will be blocked, and therefore the result will be in conflict with what would otherwise have been the coherent, the neutral, or the fundamentally correct, result. That is the obvious effect of timing points where they block coherent claims.
59. The fact that Healthcare and Services were originally seeking to effect an artificial scheme, and to enjoy thereby unintended advantages, does not have any continuing relevance once the scheme had manifestly failed, and once all steps were then being taken, and indeed all taken at the instigation of HMRC, to reverse the scheme. The fact in this case that thereafter one assessment was held to be time-barred, and that the other company’s claim was made in time resulted entirely from the neglect and incompetence of HMRC in failing to serve their assessment on Healthcare in time. It is fanciful to say that Healthcare and Services planned in advance to effect their scheme, on a “fall-back” basis, by exploiting timing differences. The mismatch in these timing points is not remotely occasioned by any of the original scheme intentions of the parties. The suggestion that the end result remains abusive because some original and indelible stain of the original planning has had any influence on the ultimate result is simply unfounded when the mismatch results solely from HMRC’s own unprovoked and single-handed failure. Indeed now to reject a perfectly sound claim, on the altar of supposedly inviolable “neutrality” when the plain effect of that would be to overturn the earlier appeal that HMRC rightly lost would itself be an abuse.
60. We attach little significance to Mr. Cordara’s initial contention to the effect that when the scheme was implemented, the parties regarded the steps as legitimate tax planning and nothing either abusive, or to re-use the more general and less technical expression, offensive. For what it is worth, we accept that the parties explained the steps fully to HMRC at the very outset, all in an effort (successful as it turned out) to speed up Healthcare’s large input recovery claim, and not even in response to questions from HMRC. Moreover the parties had some hope that the steps in the scheme would be seen to be acceptable in the light of the Court of Appeal decision in Faith Construction. The parties may have noted that there was more of a difficulty in prepaying for some future supply that simply could not be identified at the point of prepayment, but they might equally have regarded that as a difference of detail in relation to something, i.e. deliberate prepayments to defer some impending VAT change, that had generally been considered to be effective.
61. Whatever they thought at the time, we accept that they were entirely straightforward about the transactions, and we are not particularly influenced by their expectations.
62. Mr. Thomas, for the Respondents, placed very great emphasis on the fact that the two companies were associated, that they had common directors, and that they acted entirely as participants in a scheme. He also stressed that, whilst each taxpayer’s claims must be looked at individually, we should address the nature of the scheme by looking at all the realistic economic circumstances and at the two companies together.
63. We entirely accept these points. They are indeed perfectly obvious. We would go further than Mr. Thomas and say that the scheme was implemented by associated companies not only with virtually no individual thought to the steps undertaken by each company but in an utterly chaotic way, and in circumstances whereunder not only would there be no commercial objectives achieved apart from the tax saving, but where the resultant structure would be virtually unworkable commercially speaking. One extraordinary detail that we have never understood is why, having received a prepayment for drugs, estimated to be deliverable over a 3-year period, and having made a loan back initially at 7% interest, Services was going to deliver drugs at a 20% discount, against its VAT-inclusive purchase prices from Glaxo etc. This might have been explicable if the cash saving of the scheme was going to arise (were the scheme successful) in the hands of Services, such that most of the discount was designed to pass the saving to Healthcare. So far as we can see, had the scheme succeeded, the benefit would anyway have been in the hands of Healthcare, and the 20% discount would actually have meant that Services would have been unable to pay drugs suppliers for drugs towards the end of the 3-year period because its trading pattern would have rendered it insolvent.
64. We thus entirely accept that the scheme was wholly artificial, and thrown together in haste. We also accept that in looking at economic effect, we do of course pay regard to the overall position of both parties.
65. One suggestion made by the Respondents that we do not accept, however, is that the VAT return periods of Services (the first of which ended on 31 March 1998, rather than 31 December 1997) were fixed with even the faintest thought that three years later, HMRC’s assessment on Healthcare might have been “out of time”, when made on 19 December 2000, whereas Services would then have until 31 March 2001 in which to make an “in time” Voluntary Disclosure.
66. Services’ VAT period ended on 31 March 1998 for the simple reason that whilst Services’ VAT registration was back-dated to 16 December 1997, Services was not actually registered until some date in early 1998 and it was entirely natural that its first VAT period would end on 31 March 1998. In any event the idea that Services might have contemplated that the scheme would fail, but that HMRC would also be out of time in its effort to recover input tax initially refunded to Healthcare three years later is fanciful. Moreover it was not the 3-month gap that has led to one assessment being out of time in this case, and the other being in time. The imbalance results from the manifest failure of HMRC to assess Healthcare within the 2-year period, such that the gap is of 1 year and 3 months, and nothing to do with the 3-month mismatch between the original VAT periods.
67. In the Appeal involving Healthcare, the VAT and Duties Tribunal noted that it was possible that during the year 2000, the advisers to the two companies might have noted that there was a possibility that HMRC might have missed the boat for making a valid assessment on Healthcare, albeit that Services would have until 31 March 2001 in which to make its own valid Voluntary Disclosure claim, should the scheme be faulted. This possibility, which was never put higher than that, resulted from the fact that during 2000 PWC refused to answer any further questions from HMRC. This may have been because they perceived that HMRC had had adequate information to challenge the scheme by the end of the 2-year period, and it was risky to answer any further questions and then be exposed to the possibility that HMRC might say with some credibility that they had learnt some crucial further fact.
68. We put that no higher than a possibility. Its present significance is that we say that this thought cannot conceivably have occurred to anyone at the outset. It would only have occurred to PWC and the parties when they had made absolutely forthright disclosure of everything to HMRC within the 2-year period, and crucially when HMRC had then failed to assess within the 2-year period. The fault for that lay, of course, with nobody other than HMRC.
69. Our starting point, therefore, is that this scheme was thoroughly artificial, it was thrown together in haste, the parties might have been optimistic that it would succeed and be treated as only a small extension of the points established in Faith Construction, but viewed objectively (not that this is particularly material) the aims of the scheme could be said to be “objectionable” or at least likely to be so regarded by HMRC, and the mismatch in VAT periods was wholly irrelevant.
70. The BUPA Hospitals decision was of triple significance. Most obviously it resulted in the Healthcare parties accepting that their original technical or substantive contentions for sustaining the scheme became hopeless, such that the Healthcare appeal was only pursued on the procedural timing point. The second relevance of the BUPA Hospitals decision was the feature that the ECJ decided the case on the “black-letter” basis that, quite apart from any doctrine of abuse, there had simply been no supplies. This occasions Mr. Cordara’s claim that if there was no transaction, there can have been no abuse, so that nothing can be abusive. We re-phrase that point marginally in that there was clearly still a transaction, albeit one not analysed to occasion a supply of goods for VAT purposes. Thirdly there is Mr. Cordara’s other contention that the EU doctrine of abuse applies only to steps that achieve an abusive result in relation to something required by the Sixth Directive, such that a scheme relating to a domestic law zero-rating provision is unaffected by the abuse argument. There was even the fourth point, namely Mr. Cordara’s observation that the ECJ might have decided the case on the “black-letter” basis in order to avoid answering the difficult question posed by his “zero-rating” contention just mentioned.
71. We firstly reject that last point just touched on by Mr. Cordara. Once the ECJ observed that there had been no supply on the “black-letter” basis, that conclusion definitely came first, and the feature that that undermined any abusive result meant that the first pre-condition for the application of the Halifax abuse test (a tax avoidance scheme that resulted in an abusive result from the standpoint of the Sixth Directive) was wholly absent. We therefore dismiss Mr. Cordara’s suggestion that the ECJ might have decided the BUPA Hospitals case on the “black-letter” basis, simply as the preferred one of two equally cogent bases for deciding the case.
72. Turning to the next issue of whether the feature that there was deemed to be no supply, and therefore nothing to be struck down as “abusive”, and nothing to be “re-defined”, means that there can have been no abuse, we are not convinced by Mr. Cordara’s contention. Whilst there is presently no clear authority on the degree to which the wider EU notion of abuse might apply to VAT, beyond the doctrine expounded in the Halifax case, we accept that there is such a wider doctrine, and that it has been invoked in various non-VAT areas. In the present case, it seems indisputable that the parties were aiming to have made a supply, which, had they succeeded, might well have been struck down by the “abuse” doctrine of Halifax, exactly as the Advocate-General contemplated. Thus we stop short of saying that the mere fact that the supply was nullified on the “black-letter” approach eliminates all question of the application of any wider abuse doctrine. The reason why we say that this is now immaterial is not that the “black-letter” conclusion eliminates all question of abuse, but that once the transaction failed, the eventual resultant lack of neutrality was occasioned not by what the parties once hoped to achieve, but simply by HMRC’s single-handed failure to assess Healthcare in time, all in the context of a re-analysis of the transactions advanced by HMRC themselves, and fundamentally accepted by both Healthcare and Services.
73. We are also unconvinced by Mr. Cordara’s other argument, namely that any abuse doctrine was irrelevant because the scheme all related to a zero-rated transaction. In this regard, we actually note that what the scheme was trying to do was to avoid the proper application of the January 1998 law change, which actually rendered future drug supplies for in-patient use exempt, rather than zero-rated in any event. Furthermore zero-rated status was still something sanctioned and permitted by the Sixth Directive such that the notion of being able to abuse something permitted by the Sixth Directive, but not something directly implementing that Directive seems too fine, and unmeritorious, a point.
74. Beyond the point just made, the Respondents contended for a much wider doctrine of abuse than that actually provided for in the Halifax decision. We make just two observations in relation to that contention. The first is that the Respondents could certainly point to no authority that established this position. Notably, when the High Court had referred much this type of question to the ECJ in the BUPA Hospitals case it had not been answered. Other authorities were outside the sphere of VAT and generally involved fraud. HMRC had sought to raise this argument in the recent Tribunal case of Birmingham Hippodrome v. HMRC [2011] UKFTT 117 (TC), decided by Judge Kempster and Helen Folorunso, to which we will refer below. That case is to be the subject of an appeal to the Upper Tribunal, but for present purposes it rejected any such wider doctrine of abuse, at least in relation to the facts in that case. For our part, we do not decide that there is definitely no such wider doctrine. Our decision is based on the point that we now re-summarise.
75. Having now dealt briefly with various arguments that we consider to have been of only marginal relevance, we say that the reason why we reject the notion that any element of abuse, or any aim to achieve an artificial result (however described) led to the non-neutral result that the Respondents now seek to sweep away in this case is that any such argument is unrealistic. The scheme failed; the parties accepted that the coherent position was exactly as HMRC indicated in their letter of 19 December 1997; the VAT consequences of the transaction were then to be reversed in the manner dictated entirely by HMRC, and there was one reason and one reason only (i.e. HMRC’s own failing to assess in time) that now occasions the lack of neutrality.
76. It is worth noting that HMRC fought the Healthcare appeal simply on the ground that it claimed (virtually without foundation) to have discovered new relevant information during the third year, such that the assessment made prior to the end of year 3 was valid. No suggestion was made that it was abusive for the then appellant to raise a time-bar point, notwithstanding that it was perfectly obvious that if Healthcare won the appeal, then Services would continue to assert its claim, with the result that neutrality would be undermined. In the light of the decision that we have reached on the abuse point in this Appeal, we would have dismissed any abuse contention in the Healthcare appeal, but were our decision wrong, it would follow that the point could have been argued in the earlier appeal. The point would indeed have been slightly less objectionable in that case. We find it particularly offensive in this appeal, because it amounts to an attempt to reverse HMRC’s deserved fate of having lost the previous appeal, when there is actually no defect whatsoever in the claim that Services is now pursuing.
77. We certainly do not say that HMRC’s present conduct of this appeal has itself been abusive, because we accept that HMRC is merely trying to achieve a coherent result, and HMRC honestly believes that there is some “abuse” doctrine of wider application that they seek to establish. We do however consider that HMRC’s present claims risk being quite as objectionable as the failed suggestion that any original objectionable aspirations of Healthcare and Services have now occasioned, in any causal sense, the result faced by, and offensive to, HMRC.
78. We asked the Respondents’ counsel in this case whether:
79. No response was given to the first question, though manifestly it would seem that if there was no fraud, there had been full disclosure, and HMRC had failed, particularly through incompetence, to assess in time, the time-bar point would prevail.
80. The answer given to us in response to the second question was that in the case of Moorbury v. HMRC [2009] UK FTT 180 (TC) and [2010] STC 2715, HMRC had sought to block neutrality by raising a time-bar point against the appellant, but the Tribunal and the Upper Tribunal had ruled in favour of the Appellant. Implicitly it was suggested therefore that the Moorbury case was a precedent (surprisingly one that nobody had quoted to us in advance) that indicated that time-bar points should be swept away in the interests of achieving neutrality.
81. Having now read the decisions of the First-Tier and Upper Tribunals fully, it emerges that they are actually authority for precisely the opposite point.
82. The Moorbury case was somewhat similar to the facts of the more famous Halifax case. Various group companies had therefore been involved in an effort to facilitate the recovery of input tax on building construction, where the principal company, whose supplies were largely exempt supplies, would have recovered only a small proportion of the input tax. In constructing a redefinition of the steps in the scheme, the effect of which was to undermine the scheme, HMRC had drawn to the attention of Moorbury that it should make a claim to recover wrongly paid output tax, which HMRC was entirely ready to repay. For some unexplained reason, Moorbury did not make the claim under section 80 VAT Act, and the time for making the claim then expired. Notwithstanding that this failure led to the group as a whole being effectively penalised by the re-definition and Moorbury’s own failure to make its claim, HMRC contended that the unbalanced result should prevail because Moorbury could easily have made its claim, and the fact that it did not do so and that the claim became time-barred was Moorbury’s fault.
83. HMRC lost the case. This, however, was definitely not on the ground that the time-bar point should be re-opened on the altar of neutrality. Judge Bishopp made it perfectly clear that on the time-bar point, EU law required Moorbury to have had a remedy (i.e. an opportunity to make a claim to achieve the coherent result), and that claim had to be capable of being made in a workable manner. All that had been respected, however, and it was Moorbury’s fault for not making the section 80 claim in due time (as it had been invited to do) that blocked the claim so that the time-bar point should not be re-visited and re-opened. In other words, HMRC won the time-bar point.
84. HMRC lost the Moorbury case on the quite different principle that in accordance with the Halifax decision, the ECJ had said that when an abusive transaction was undermined by the abuse doctrine and had to be “redefined” that redefinition had to ensure that no penalty was occasioned. HMRC had to effect a redefinition that looked at the net position of all parties, undermined the abuse (whatever it was) but then left the parties in the realistic net position that should prevail simply to reverse the abuse but occasion no penalty. The Tribunals decided that this required HMRC, on its own initiative, to achieve this neutral end result, regardless of the claims by the taxpayers. It was therefore on this ground that credit had to be given in the redefinition for the amount properly repayable to Moorbury, and not because Moorbury could re-open a time-bar point.
85. We draw three conclusions from this case, which we consider to be quite extraordinary in the context of the arguments advanced by HMRC in this Appeal. These conclusions are that:
86. Another recent case, the Birmingham Hippodrome case, to which we have already referred, also appears to support the decision that we have reached in this Appeal. We appreciate that there is to be an appeal to the Upper Tribunal in this case, but that will certainly not affect one of the points for which we refer to the case, and we find it difficult to conceive that it will affect the other.
87. In this case a theatre charity had wrongly assumed that its services were liable to VAT, and, when this was realised to be wrong, it then had the opportunity to reclaim the wrongly paid “net” VAT (”net” in the sense that it could only claim the net excess of wrongly assumed output tax over the input tax that it had claimed). HMRC had made it clear that claimants seeking to recover wrongly paid VAT could not “cherry-pick” so as to make claims for the years where outputs had exceeded inputs, and refrain from making claims for years when there had been excess inputs. Without any cherry-picking, the Appellant had made claims for all the years for which claims could be made. Claims for two other years, during which the theatre had been closed for refurbishment such that substantial inputs had been recovered when no supplies were being made, were years that were inevitably time-barred. No claims could have been made for them, and it was not through cherry-picking that no claims were made for those years.
88. In some way HMRC contended that there was some form of abuse in the way that the claim was made. Secondly, reliance was placed on section 81(3A) VAT Act 1994, a recently introduced provision that provided that where sums were reclaimed by a taxpayer on a Fleming type claim, there could be set against refunds certain counter-balancing liabilities, even where, but for the obligation to make the refunds, the liabilities would be time barred.
89. The significance of the first “abuse” point is that it was conceded on the part of HMRC that HMRC was endeavouring to rely on EU law to sustain some wider form of abuse notion than the doctrine based on the Halifax case. We find it slightly difficult to understand what was abusive about simply making claims for all the years for which claims could be made, but whether that be right or not, the Tribunal decided that the claimed new doctrine of “wider abuse” had not been sustained. Whether it is eventually established or not, we simply take the view that lack of neutrality resulting from a culpable failure by HMRC cannot be said to occasion abuse by the Appellant, when the Appellant seeks, at worst opportunistically, to exploit that failure.
90. The other point (which is immaterial in this case because section 81(3A) only applies in relation to time-barred liabilities of the very taxpayer making the claims) was all dependent on whether the sub-section applied only as between claims and liabilities in the same accounting period, or whether it could re-open liabilities in an otherwise closed period, and net them off against claims for other periods. The Tribunal decided that the latter was correct. Whether this was right or not is of no significance to us. The feature that is significant in this case is that it took a statutory insertion into section 83 to enable time-barred liabilities to be revived for the purposes of netting them off against recovery claims. What HMRC has sought to do in this case is to block their refund liability, and in effect revive their own blocked right to recover Healthcare’s input deduction. There is firstly no statutory modification equivalent to section 81(3A) that sanctions this. Secondly it would seem to us to be particularly offensive for there to be such a provision. If, through its own fault, HMRC has failed to make an assessment within the time periods, and Moorbury and common sense (the feature that the time-barring of any claim will often lead to an inappropriate substantive result) indicate that the time-bar point remains a valid objection to further adjustment, it would seem wrong for any equivalent to section 81(3A) to apply in the present circumstances. At any rate, at present, there is no equivalent.
91. Our conclusions in relation to the “abuse” points are thus as follows:
Paragraph 5(2)
92. HMRC’s claim in relation to liability under paragraph 5(2) Schedule 11 VAT Act 1994 is that where a person has issued a VAT-inclusive invoice to a purchaser or a customer, that fact alone, regardless of whether there was properly a taxable supply of goods or services, renders the person issuing the invoice liable either to VAT or to some equivalent liability, and the ability thereafter to reclaim the amount owed or paid is either non-existent or more circumscribed than the ability to recover wrongly paid output tax in the normal situation.
93. Before addressing the paragraph 5(2) issue in detail it is worth observing, just as various cases before the ECJ have observed, that the points now in issue are both difficult and unsatisfactory, largely because there appears to be a glaring lacuna in the domestic VAT provisions. One might say that there is a similar lacuna in the Directive provisions, save for the observation of the ECJ that issues in relation to reversing the charge to VAT on anyone rendered liable to VAT by issuing a VAT-inclusive invoice, or one that specifically refers to the charge to VAT, is a matter for Member States to address, subject to some general guidance.
94. The starting point to this topic is the undoubted feature that both the Sixth Directive, and paragraph 5(2) of Schedule 11 to the VAT Act 1994 render a person who has issued a VAT-inclusive invoice liable to pay either VAT, or what we will term “VAT equivalent”. Whilst paragraph 5(3) refers to this charge applying “whether or not” a person is a taxable person, whether or not there is really a supply, and whether the supply is properly chargeable with VAT, the provision is principally directed at the case where for one reason or another there is no substantive liability for VAT, and very likely not even thought to be a substantive liability to VAT. As explained in paragraphs 47 and 48 above, the whole purpose of the provision is to stop tax evasion, resulting from purchasers claiming input deductions for VAT, for which they have been invoiced, in circumstances where the supplier was probably not substantively liable for VAT, and has very likely not accounted for it either to the exchequer.
95. Prior to addressing the circumstances where a person, seemingly rendered liable under paragraph 5(2) and (3) for VAT or “equivalent VAT” can either defend a claim to pay under the provisions, or recover an amount already paid, it is worth considering two preliminary points.
96. First there is the question that is potentially of more than pedantic importance which is whether the charge under the two sub-paragraphs is actually to VAT, or just something equivalent to VAT. This question is posed on the basis that “VAT proper” is something chargeable in respect of taxable supplies of goods or services by registered or registerable traders”, whereas in many cases the charge under the two sub-paragraphs will clearly not be of “VAT proper”. The confusing aspect of this question is that whilst the EU provisions quoted in paragraph 46 above clearly indicate that the charge in question is of VAT, the domestic provision hedges its bets in the last three lines of paragraph 5(3), and if anything suggests that where there is no proper charge to VAT, then the liability is simply that of a distinct debt to the Crown but not VAT as such. We will revert to the possible significance of this point below.
The improperly wide application of paragraph 5(2)
97. The next preliminary point to note is that on account of the very wide drafting of the two relevant sub-paragraphs, they are not confined in their application simply to the type of abuse at which they are obviously principally aimed, i.e. to the type of fraud mentioned in paragraph 47 above. Thus to take the everyday example of a taxable trader who assumes quite honestly that a particular supply is a standard-rated taxable supply, that trader will inevitably issue a VAT invoice (as it believes it is required to do), and it will account for the VAT in its return. If it turns out (within the time limits for making adjustments) that the trader was wrong, and that the supply was exempt, the trader obviously seeks to recover wrongly accounted for output tax under section 80(1) VAT Act, which provides that “Where a person has accounted to the Commissioners for VAT for a prescribed accounting period and … has brought into account as output tax an amount that was not output tax due, the Commissioners shall be liable to credit the person with that amount.” Whilst this seems clear enough, the confusing point is that that person will inevitably have issued a VAT invoice, or a VAT-inclusive invoice, so as to be simultaneously liable for VAT or “equivalent VAT” under paragraph 5(2). The question that then arises is what happens if looking at the “VAT proper” analysis, it is obvious that the amount accounted for as output tax should be refunded by HMRC, but (as HMRC now contend) the rules for refunding VAT or “equivalent VAT” paid under paragraph 5(2) are quite different and more restrictive?
98. The practical answer to this question is of course that in an honest case, where a wrong classification has been made by a taxable person, and within the time-limits, that person seeks to reverse the wrong payment of output tax, HMRC will never raise or mention paragraph 5(2). The claim will be made and accepted under section 80(1). And this will be the situation not just in relation to the first example chosen by us (i.e. that mentioned in the previous paragraph) but in relation to virtually every claim to recover wrongly-paid output tax. After all, when taxable persons wrongly account for output tax, they will almost inevitably issue VAT or VAT-inclusive invoices.
99. Whilst the practical approach assumed in the previous paragraph will usually be followed, HMRC now assert that Services was liable for VAT or “equivalent VAT” under paragraph 5(2), simply through having issued the VAT-inclusive invoice, and they contend that the circumstances where an amount paid over, or payable, under that paragraph can be reclaimed are much more restrictive, than where the claim is simply for overpaid output tax, disregarding paragraph 5(2).
Liability to refund amount paid under paragraph 5(2)
100. At one point, HMRC appeared to contend that there was no liability to refund an amount due under paragraph 5(2) at all. This contention was at least consistent with the clear fact that there is no statutory or regulatory provision that actually deals expressly with the circumstances where a person has a right to recover tax or an amount charged under paragraph 5(2). It is also consistent with guidance from the ECJ to the effect that it is for Member States to regulate the circumstances where liability under Article 21.1.c. should be reversed, albeit that Member States should comply with certain guidelines. Where however this particular Member State appears to have enacted no rules whatsoever, this may initially seem to support the extraordinary proposition that there is no liability to refund anything charged under paragraph 5(2).
101. One point that is absolutely obvious is that liability to refund VAT or “equivalent VAT” charged and charged only under paragraph 5(2) cannot be governed by the normal rules for ascertaining whether there was a liability for “VAT proper”. After all, the whole purpose of the liability under paragraph 5(2) is to occasion a liability to pay “VAT equivalent” to stop the type of abuse referred to in paragraph 47 above where VAT-inclusive invoices have been issued where indeed there was no liability for VAT proper. It obviously follows that if the liability could be reversed just by showing that the transaction occasioning the issue of the VAT-inclusive invoice was not properly liable to VAT, then the charge under paragraph 5(2) would be totally undermined. Some much more targeted rule for dealing with refunds of this liability are needed, and they appear to be non-existent.
102. We were referred, during the hearing, to the three ECJ cases of Genius Holdings [1989] ECR 4227, Case C-454/98 Schmeink & Cofreth and Strobel [2000] STC 810 and Case C-566/07 Stadeco [2009] STC 1622.
103. The Genius case essentially governed the question of whether the purchaser or the recipient of a service who had received a VAT-inclusive invoice could claim an input deduction when it emerged that the supplier had simply invoiced the supply as a VAT-able supply when it was not. The answer was that there was no right in this situation to deduct input tax. The case also led to the supposition that whilst in some circumstances the person issuing the invoice might recover tax charged under Article 21.1.c, that was probably dependent on the person issuing the invoice having issued the invoice “in good faith”.
104. The Schmeink and Strobel cases both involved the improper issue of invoices, though in circumstances where the issuer “repented”, declared its fraud to the authorities, and as a result no input deduction was even sought. It was implicit that the “good faith” test of Genius would not have been satisfied. It was also implicit that the decision in Genius (technically denying the purchaser an input deduction) did not eliminate the liability under Article 21.1.c because it was entirely possible that for one of various reasons the Member States tax authority might not be able to deny or reverse the input deduction. Accordingly, to prevent evasion the issuer of the invoice did have to retain the liability to account for the tax under Article 21.1.c. Nevertheless, the decision was that even when there had not been good faith, the fact that the circumstances had all been revealed so that no input deduction was in the event claimed, meant that refusing the issuer of the invoice a refund of the tax under Article 21 would be “disproportionate” so that the tax had to be refunded. This conclusion was so even though in the particular circumstances German law did not directly provide for that refund.
105. The Stadeco case was of less direct relevance. A cross-frontier supply of services had wrongly been assumed to be taxable, and had been so invoiced. The acquirer was a state organisation unable to claim any sort of input deductions in either the state of the supplier, or its jurisdiction, for any sort of VAT. When it was discovered that VAT should not have been charged, the supplier requested a repayment and received it, on representing that it would re-invoice the acquirer on a VAT-exclusive basis, and refund the relevant overcharged tax (refunded by the first state) to the acquirer. It subsequently transpired that although the tax had duly been repaid, the acquirer had not been re-invoiced, and had not received the promised refund. The first state thus sought to recover the VAT that it had initially refunded, on the basis that it was entitled to make its refund dependent on the recipient of the refund not being “unjustly enriched”. There was therefore no issue of VAT neutrality in this case because the acquirer could never make any sort of input tax claim. It was held that the first state could make its refund dependent on precluding “unjust enrichment”, and therefore the taxpayer had to refund to the exchequer the amount that it had initially received back itself.
106. The relevant principles that we derive from these ECJ cases are as follows:
· It is certainly quite unacceptable for a Member State to charge tax or “equivalent VAT” under its domestic version of Article 21.1.c. and to refuse refunds of that in all circumstances. For instance if, to take an extreme case, the invoice was issued as an honest mistake, the acquirer claimed or retained no input tax deduction, and the issuer re-invoiced the acquirer without VAT, any tax once charged had to be refunded.
· The liability to refund in at least some circumstances had to be accepted even though domestic VAT law might make no provision for such refunds.
· It is equally unacceptable for the domestic taxing authority to have a discretion as to when to refund and when not to refund. In view of the observations in paragraphs 98 to 100, which strongly suggest that HMRC would not usually take the point presently being argued, but are doing so solely as a “late in the day” effort to supplement their “abuse” case, HMRC seem to be on weak ground.
· “Good faith” does not as such assure a person who issues an invoice of a refund of any tax charged. The “good faith” point emerged the other way round, in that it was once thought that absence of good faith would block a refund, even where the acquirer had obtained no input deduction, such that there had been no breach of neutrality. It later emerged that when the buyer had abandoned or refunded any input tax deduction, and the supplier had managed, as in Strobel, to reverse the initial issue of the VAT-inclusive invoice, a refund was due, notwithstanding the original absence of good faith.
· The single test that we consider to be the most important to emerge from the various cases is the proposition that if the issuer of the invoice has done everything in time to enable the transaction to be reversed, such that there is no ultimate breach of neutrality, then any tax initially charged should be refunded.
107. HMRC’s contentions in this case are essentially that:
108. Our decision is that HMRC’s contentions are all flawed, and that Services’ claim should prevail. The reasons for this will regrettably re-raise the circumstances leading to the lack of neutrality, and the whole issue of abuse, but are nevertheless as follows.
109. We unhesitatingly reject the suggestion that Services has shown any bad faith. What has happened in this case is that the original scheme failed; all parties conceded that in substance; the transactions were therefore to be reversed in the manner laid down by HMRC, and the worst that can be alleged against Services, in terms of “bad faith” is that it opportunistically sought not to abandon a perfectly valid, in time and legitimate claim (one that it had been invited to make by none other than HMRC), simply because, through neglect and incompetence, HMRC had failed to assess Healthcare, as it could, and as it should, have done in time. There was no bad faith on the part of either taxpayer, Healthcare or Services.
110. We find the implicit discretion that HMRC seek to assume in this case to be exceptionable. Were the case one where paragraph 5(2) was fundamentally and properly in point, i.e. where a VAT-inclusive invoice had been deliberately and fraudulently issued, then HMRC’s stance might not be so unreasonable. This case is one, however, where not only the two taxpayers, but HMRC also initially thought that the initial transaction occasioned a supply. The facts were fully disclosed; Services was registered on the basis that its supply rendered it registerable; HMRC accepted that the transaction was artificial but they accepted the VAT from Services on the basis that it was “VAT proper”, and they refunded it to Healthcare for the same reason. And for some considerable period, they refunded VAT invoiced against Services by third-party suppliers (Glaxo etc) to Services, on the reasoning that Services’ onwards supply had been made in 1997. We are thus squarely in the situation where HMRC is now raising an argument, late in the day, when originally all parties (including HMRC) addressed this case on the basis of whether “VAT proper” was owed or not owed. And now HMRC, exercising a very inappropriate discretion, seek to rely on different arguments whose real compass is more appropriate to the situation that we described in paragraph 47 above.
111. Finally, however, we turn to the last bullet point of paragraph 106 above. It is indisputably the case that if our decision in this case, in favour of Services, stands, then it will indeed result in neutrality being undermined. Since the most important requirement that we glean from the ECJ cases is that tax charged under paragraph 5(2) should only, with certainty, be refunded if Services has done everything to secure ultimate neutrality, how do we justify deciding this appeal in favour of Services, thus undermining neutrality?
112. The answer to that is twofold. Firstly it is very relevant to remember that the situation to which liability under Article 21.1.c is properly aimed is one where the VAT-inclusive invoice will very likely have been provided fraudulently. It is not therefore surprising that there is the attention to ensuring that the person that provided that invoice does everything in due time to preclude some lack of neutrality. For that is the concern to which the liability in question is directed, so that eliminating that concern is understandably paramount.
113. The second point, however, is that it was definitely not through any lack of information, or for instance through concealing the identity of the recipient of the service who might claim the wrong input deduction, that this case may result in a lack of neutrality. The two parties, Healthcare and Services, provided full information to HMRC at the very outset, and they, and HMRC, initially believed that Services’ transaction had occasioned a supply, and thus a liability to VAT proper. The only reason in this case why there will be a lack of neutrality is the culpable failure of HMRC to do what they could and should have done, namely to challenge the scheme before the end of 1999, and assess Healthcare so as to recover the wrongly-conceded input refund made in early 1998. On this basis, the only omission that can be levelled against Services, in terms of the supposed obligation to do everything to ensure ultimate neutrality, would be some asserted failure to remind HMRC towards the end of 1999 that if they proposed to assess Healthcare, they would almost certainly need to do so by the end of 1999, since by that point HMRC would have had all required information to attack the scheme if that is what they intended to do. Alternatively, perhaps, HMRC contend that Services should have dropped a perfectly good claim, one that HMRC had invited Services to make, simply so as to ensure ultimate neutrality, and effectively reverse the feature that HMRC’s own neglect in making their assessment on Healthcare resulted in Healthcare’s appeal against the late assessment being allowed.
114. We consider it unacceptable that:
115. Our decision in relation to the paragraph 5(2) point is that in accordance with the guidance given by the ECJ, persons assessed under paragraph 5(2) must certainly in some situations have a right to recover VAT or “equivalent VAT” charged under paragraph 5(2), and that this case is one such case where recovery is allowed. Services acted entirely in good faith, and it afforded HMRC every opportunity to eliminate any “lack of neutrality”. HMRC’s single-handed failure to achieve neutrality cannot block what is an entirely bona fide, legitimate and “in time” claim. On this basis we dismiss HMRC’s contentions under paragraph 5(2) and re-confirm that Services’ appeal should be allowed.
116. HMRC’s claim was that we have no jurisdiction to hear the appeal in relation to the paragraph 5(2) point because section 83, which lists the cases where the Tribunal has jurisdiction, makes no reference to any claim to recover tax charged under paragraph 5(2).
117. HMRC did, however, concede that we had jurisdiction under section 83(1)(t) to hear an appeal for the recovery of an amount claimed under section 80, and the Appellant’s contention was that their claim was indeed made under section 80 such that we did have jurisdiction.
118. We accept that Services’ claim is indeed properly made under section 80. On the basis on which all parties originally proceeded, all incidentally at the instigation of HMRC, Services’ claim fell squarely within the wording of section 80(1), being for recovery of an amount that was treated as output tax that was not output tax.
119. Reverting now to the point made in paragraph 96 above, it is at least clear that Article 21.1.c treats the liability on the part of the person who issued a VAT invoice, regardless of whether there is a proper charge or not, as a liability to VAT as such. Paragraph 5(3) is less clear on this point, as we conceded in paragraph 96, but since section 80(7) eliminates any liability (outside the provisions of section 80) on the part of HMRC “to repay any amount accounted for or paid to them by way of VAT that was not VAT due to them”, and since the ECJ has made it clear that in some circumstances the person liable under paragraph 5(2) must have a right to recover the amount owed, that right can only be under section 80. Section 80 must thus be read consistently with this requirement. It certainly involves (as we conceded in paragraph 101 above) applying a different test in granting refunds than the test that obviously governs the refund of “VAT proper”. But there is considerable guidance by the ECJ as to the tests that we should apply. Our decision is that Services’ appeal is and can only be made under section 80; that we must deal with that appeal, for which we have jurisdiction, in accordance with the guidance given to us by the ECJ, and that is what we have done in reaching our conclusion given in paragraph 115.
120. Services contended in the alternative that HMRC had raised both the paragraph 5(2) point and its jurisdiction point too late in the day for us to deal with the present appeal by reference to those points. We have assumed and claimed jurisdiction primarily on a different basis, but we do accept that:
121. We accordingly decide that because we consider that the whole paragraph 5(2) contention is not properly appropriate to the facts of this case, and because that contention and the jurisdiction points were only raised after Healthcare had won its appeal, those points were raised too late in the day for us to pay regard to them.
Services’ claim in relation to the input deductions that it claimed in relation to supplies of drugs to it, during the period when HMRC initially refunded that input tax
122. At the end of the hearing, as we mentioned in paragraph 8 above, Mr. Cordara raised a new argument to the effect that if we considered that some feature of abuse led us to consider making adjustments in order to achieve or preserve “neutrality”, we ought still to decide that Services’ initial payment of output tax should be refunded. His contention was that that output tax should be matched with the fact that Services had later suffered input tax, on being invoiced by Glaxo etc, and that to refuse the refund and deny relief for the input tax would be a worse breach of neutrality than the one that the Respondents advanced.
123. This further contention by Mr. Cordara is now of course superfluous, since we have not barred Services’ claim on any abuse ground, but we ought to state Mr. Cordara’s argument in case we are overturned on appeal on the primary point.
124. Before actually summarising Mr. Cordara’s contention, it will probably be clearest to re-summarise some of the facts on various different assumptions.
125. Starting with the original scheme, and the way in which it was hoped that it would operate, the summary appears to be as follows:
126. When the scheme was first challenged by HMRC on 19 December 2000, the contention was that Services had made no supply; Healthcare had no corresponding “input”, and that the easiest approach was to reverse those two transactions and to treat the third party drug companies as supplying directly to Healthcare, whatever the mechanics involving Services.
127. This latter suggestion was not in fact adopted, because from December 1998 (when Services and Healthcare had worked out how to operate the scheme) until December 2002, drugs were purchased by Services, and delivered to Healthcare (even indeed for out-patients).
128. The challenge to the scheme that was eventually sustained to be the correct challenge was not that there was any Halifax abuse and redefinition but that the original transaction had occasioned no supply.
129. It followed that the correct procedure would have been for Services to recover the initially wrongly paid VAT; for Healthcare to refund to HMRC the input deduction initially conceded by HMRC, and for drug purchases from Glaxo etc and deliveries to Healthcare by Services now to be treated as VAT-able purchases and supplies, such that the VAT would “wash through”. Ignoring the discount point again, Services would pay £117.5, and ignoring discount and margin, would receive consideration of £117.5 from Healthcare in the sense of its £10 million commitment being reduced by that amount, obviously furnishing Healthcare with a VAT invoice which would be of little significance to Healthcare, since it would be almost entirely referable to exempt supplies.
130. Turning now to Mr. Cordara’s contention, he points out that what actually happened, following the decision in BUPA Hospitals is that Services concluded that its claim to recover its wrongly-paid output tax would be met, and it therefore refunded input tax that had been refunded to it in respect of its drugs purchases in periods 12/98 and subsequent periods (see the third bullet-point in paragraph 125 above), and abandoned its appeal for further recovery of input tax that had been refused, once HMRC had blocked the claimed repayments.
131. Mr. Cordara’s contention was accordingly that:
“The Commissioners’ approach is valid on the assumption that the output tax is repaid to Services. If this is not the case then, in respect of the supplies of goods and services which were in fact made to Healthcare:
a. Services has wrongly accounted for output tax which has not been repaid; yet
b. Services has incurred input tax which it has not been permitted to deduct.
Thus, Services gets the worst of both worlds and the Commissioners keep Services output tax and prevent deduction of input tax – despite the Commissioners’ case that Services should be in a VAT neutral position”.
132. We note in passing that HMRC did not particularly stress that Services should be in a VAT neutral position, but rather they said that we should consider neutrality from an overall economic standpoint and by considering the position of the two companies together. More relevantly, we do not agree with Mr. Cordora’s further argument.
133. Matters are all confused, we accept, because Services did not account for output tax on what now emerge to have been the first drug supplies that it made, namely the ones that commenced in the period 12/98. The correct position should however have been that the input tax on the Glaxo invoices should simply have washed through Services, on Services making output supplies to Healthcare. The result would have been that (whatever the position in relation to HMRC now being unable to assess Services for non-paid output tax in period in and after 12/98), Services would still have had the benefit of off-setting the input tax against its output liabilities, albeit that it would obviously no longer have received cash repayments. And just as HMRC are now too late to assess for non-paid output tax, so Services is also too late to revive its once-abandoned claim for the input tax, attaching to Glaxo invoices etc.
134. This whole argument is irrelevant on the basis of our decision that Services should recover the whole of the £1,489,362 output tax, and that this claim should not be blocked on any sort of “abuse” contention. Were we wrong on this, however, and some element of abuse were perceived by a higher court, and that abuse had to be eliminated in order to achieve neutrality, we consider that the more material neutrality to consider, and to match against Services’ claim is (as HMRC has always contended) HMRC’s blocked claim to recover the precisely matching input tax, originally claimed by Healthcare in respect of the prepayment. We consider that Services’ input tax attaching to Glaxo invoices was more obviously to be matched with its proper output liabilities on making the on-supplies. This “double matching” at least ties together first the treatment of each side of the transaction that was wrongly treated as involving a supply in late 1997, and secondly it treats Services and Healthcare precisely as they would have been treated if Services had simply on-supplied drugs to Healthcare in all the later periods. That, after all, is the proper treatment in the light of the “black-letter” decision reached in BUPA Hospitals, and tallies with what is now the correct analysis of the various transactions.
135. We do not object to the fact that HMRC dedicated far more attention in their further submissions to the Moorbury issue, than they did to resisting the Appellant’s contention that we have now dealt with. Were we to have been influenced by their further contentions, it would have been appropriate to give the Appellant an opportunity to respond, but this is unnecessary in the light of our earlier conclusions.
136. We had already dealt fully with the Moorbury issue prior to receiving the Respondents’ further comments, and we see no need to revisit that part of our decision. We simply state that, having again read both the First-Tier and Upper Tribunal decision, we remain unhesitatingly of the view that the reason why HMRC lost the Moorbury case was not that they had to allow a taxpayer to make an out of time claim in the interests of neutrality. The basis of both decisions was that in constructing a Halifax redefinition, it was incumbent on HMRC to devise a neutral redefinition that eliminated the abuse but occasioned no penalty, and that this had to be achieved as between various taxpayers and over quite possibly several VAT periods. There was no need for Moorbury to make a claim. HMRC had to achieve that end on their own initiative.
137. Reverting to the question that we once asked, namely whether HMRC would feel bound to allow a taxpayer to make an “out of time” claim in order to achieve a fair result, where in a non-abusive situation a taxpayer was in a hopelessly unfair position due to its own remiss failure, we consider that the plain symmetrical answer is that the time-bar point stands. The decisive factor in this example is that Halifax redefinition is a one-way street that enables HMRC to re-define a transaction to eliminate abuse. There is no taxpayer matching entitlement to be able to re-define transactions that might lead to great unfairness to the taxpayer, and the taxpayer has no entitlement to claim that out of time claims can be made, as of right and whatever the circumstances, when the failure to have made those claims is what may have occasioned manifestly unfair results.
138. It is our decision that the time-bar points are entirely valid, provided that parties have opportunities to make their claims in time, and we consider that this is indeed supported by the decision in Moorbury, not undermined by it.
Achieving ultimate neutrality in order to recover tax or VAT equivalent paid under paragraph 5(2)
139. HMRC have taken the opportunity to advance, or re-advance, arguments in relation to the circumstances where ECJ authority suggests that liability under Article 21.1.c. and under paragraph 5(2) should be reversed and any tax or equivalent VAT initially paid refunded.
140. In one respect, HMRC appear to us to be making a concession in favour of the taxpayer, which seems to us not to be justified, but they then accompany that with a suggestion that we find untenable. By this obscure remark we refer to the fact that HMRC seem to say that tax or anything paid under paragraph 5(2) should be refunded where there is no bad faith. We expressed the “good faith/bad point” differently in the fourth bullet point of paragraph 106, and in a manner that appears to us to be consistent with the authorities, and to make much better sense. Were HMRC to be right, however, we should then make it very clear that we consider that their allegation of bad faith against Services is wholly wrong.
141. Both points dealt with in the previous paragraph are illustrated by the following passage from HMRC’s further contentions:
“Where the correction has not been made and is no longer possible, i.e. the risk of tax loss has not been wholly eliminated then there is a requirement of good faith. That equates to fair dealing and a deliberate decision to maintain the section 80 claim which had been introduced (and originally treated by both sides) as a mechanism for achieving fiscal neutrality for the purpose of seeking a result completely contrary to fiscal neutrality does not amount to good faith.”
142. The worst that can be asserted against Healthcare and Services is that, whilst they would certainly have acknowledged that their respective claims had no substantive merit, in the light of the BUPA Hospitals decision, they always intended to rely on legitimate time-bar points, and chose not to seek to overturn the consequence of HMRC’s own single-handed failure to assess Healthcare in time. We do not consider that to be bad faith, or to get as close to bad faith as the thrust of HMRC’s present contentions.
143. On the more substantive question of whether we consider that the appeal should be dismissed (on our approach to the ECJ tests that we summarised in paragraph 106 above), simply because Services has not been able to achieve “ultimate neutrality”, we have little to add to what we said in paragraphs 107 to 115. To level the blame for failing to achieve neutrality at Services, when the neglect and incompetence of HMRC is what has occasioned the lack of neutrality seems to us not to be how the ECJ expected us to apply the test.
144. The Appellant claimed interest, and indeed interest on a compound basis, were it to succeed in its appeal. We were asked however to leave the question in relation to interest to await the outcome of the compound interest question currently being dealt with by the higher courts. We accept this request, and simply note now that the matter deferred in relation to interest relates not just to the compound interest issue, but to whether any interest should be awarded at all. It was the Respondents’ contention that no interest of any sort was due, but this aspect as well as the compound interest question, are omitted from this decision.
145. This appeal was categorised as a Complex appeal, and the Appellant did not opt out of the costs regime. Accordingly the Appellant requested an award of costs, were it to win this appeal, and we award the Appellant its reasonable costs, computed on the standard basis, any dispute as to costs to be dealt with by a costs judge.
Right of Appeal
146. This document contains full findings of fact and the reasons for our decision. Any party dissatisfied with this decision has a right to apply for permission to appeal against it pursuant to Rule 39 of the Tribunal Procedure (First-tier Tribunal) Tax Chamber Rules 2009. The application must be received by this Tribunal not later than 56 days after this decision is sent to that party. The parties are referred to “Guidance to accompany a Decision from the First-tier Tribunal (Tax Chamber)” which accompanies and forms part of this decision notice.
HOWARD NOWLAN
TRIBUNAL JUDGE
RELEASE DATE: 18 July 2012