The Commissioners for HM Revenue and Customs v S and I Electronics PLC: [2012] UKUT 87 (TCC) [2012] UKUT 87 (TCC)

UPPER TRIBUNAL
TAX AND CHANCERY CHAMBER
[2012] UKUT 87 (TCC)Case No FTC/17/2009, FTC/18/2009
Between
The Commissioners for HM Revenue and CustomsAppellantS and I Electronics PLCRespondent
Before
MR JUSTICE NEWEYJUDGE JOHN WALTERS QCDate 12 March 2012Category: Tax
[2]S&I began running a family electrical business in 1980. By 2006, it had three divisions. One of these involved the import of small household electrical items, another importing and exporting consumer electrical goods such as televisions and camcorders. The turnover of these divisions was, however, 20 dwarfed by that of the third division, the business of which consisted predominantly of the purchase and sale of mobile phones. This division had a turnover of some £11.5 million in April 2006 alone. S&I would typically (though not always) buy phones within the United Kingdom and export them. 25 3. Between April and July 2006, S&I entered into 132 transactions for the sale of mobile phones. 99 of the 132 sales were for export.[4]In 2007, HMRC denied S&I’s entitlement to deduct input tax in respect of its purchase of the phones comprised in 90 of the 99 export sales. This was on the 30 basis that, in each case, the phones in question had previously been sold by a person who had fraudulently evaded VAT (or, in one instance, by a “contra trader”) and that S&I knew or should have known that its transactions were connected with fraud. The sums disallowed amounted to about £4.3 million. 35 5. S&I’s appeals from HMRC’s decisions occupied the First-tier Tribunal (“the FTT”) for some 14 days in the autumn of 2008. The FTT summarised in these terms the questions which had to be addressed in relation to each of the 90 batches of phones sold by S&I (paragraph 6 of the decision): 40 “(i) whether the purchase of the phones can be traced back to a person whom HMRC allege was fraudulently evading VAT … ; (ii) whether the person alleged by HMRC to have evaded VAT, had done so in relation to its sale; and (iii) whether S&I knew or should have known that its purchase was 45 connected to fraudulent evasion”. 4[6]In the great majority of cases (but not all of them), the FTT answered the first two of these questions in the affirmative. In other words, it concluded that the bulk of the 90 batches of phones could be traced back to persons who had fraudulently evaded VAT. As regards the third issue, the FTT decided that 5 S&I had not known that its deals in the relevant period were connected to fraud, but that it should have known of the connection. As a result, the FTT held that HMRC had been justified in denying S&I’s entitlement to much of the input tax it claimed. 10 7. On the other hand, the FTT rejected a submission by HMRC that, where a trader knew or ought to have known that a transaction was connected with fraud, input tax fell to be denied in its entirety. The FTT took the view that the tax denied should not exceed that lost in the related fraud. Where, therefore, it was proved that VAT had been fraudulently evaded by an importer, “only the 15 VAT evaded by that fraudulent importer should be denied” (paragraph 78 of the decision).[8]The overall result of the FTT’s decision was that HMRC had to make a repayment to S&I of about £962,500. HMRC were otherwise held to have 20 been justified in denying the input tax claimed by S&I.[9]HMRC and S&I have both appealed against the FTT’s decision. HMRC’s appeal challenges the FTT’s view that the tax denied should be limited to the tax loss. By a cross-appeal, S&I seeks to overturn the FTT’s decision to the 25 extent that it was adverse to it. S&I maintains that no input tax should have been denied.[10]The hearing of the appeal and cross-appeal was postponed pending the outcome of appeals in three other MTIC cases (Mobilx Ltd v Revenue and 30 Customs Commissioners, Blue Sphere Global Ltd v Revenue and Customs Commissioners and Calltel Telecom Ltd v Revenue and Customs Commissioners) which the Court of Appeal heard together in 2010. Judgment on the appeals was handed down on 12 May 2010: see [2010] EWCA Civ 517, [2010] STC 1436. 35 Kittel and Mobilx[11]In Kittel v Belgium; Belgium v Recolta Recycling SPRL (Joined Cases C- 439/04 and C-440/04) [2008] STC 1537, the European Court of Justice (“the 40 ECJ”) ruled that entitlement to the right to deduct input tax can be refused where: “it is ascertained, having regard to objective factors, that the taxable person knew or should have known that, by his purchase, he was 45 participating in a transaction connected with fraudulent evasion of VAT” 5 (see paragraphs 59 and 61).[12]Guidance as to the implications of the Kittel decision was given by the Court of Appeal in the Mobilx case. In the course of his judgment in that case (with 5 which Carnwath LJ and Sir John Chadwick agreed), Moses LJ arrived at the following conclusions: i) the principles enunciated by the ECJ in Kittel fall to be applied by domestic Courts without further legislation to that effect. Moses LJ 10 explained (in paragraph 47): “… the objective criteria which form the basis of concepts used in the Sixth Directive [i.e. EC Council Directive 77/388] form the basis of the concepts which limit the scope of VAT and the right to deduct under ss 15 1, 4 and 24 of the 1994 Act [i.e. the Value Added Tax Act 1994]. Applying the principle in Kittel, the objective criteria are not met where a taxable person knew or should have known that by his purchase he was participating in a transaction connected with fraudulent evasion of VAT. That principle merely requires 20 consideration of whether the objective criteria relevant to those provisions of the VATA 1994 are met. It does not require the introduction of any further domestic legislation”; ii) the right to deduct input tax will be lost if the trader knew or should 25 have known that he was taking part in a transaction connected with fraudulent evasion of VAT, but not if he merely knew or should have known that the transaction was more likely than not to be so connected. Moses LJ said (at paragraph 60): 30 “The true principle to be derived from Kittel does not extend to circumstances in which a taxable person should have known that by his purchase it was more likely than not that his transaction was connected with fraudulent evasion. But a trader may be regarded as a participant where he should have known that the only reasonable explanation for 35 the circumstances in which his purchase took place was that it was a transaction connected with such fraudulent evasion”; iii) fraudulent evasion need not precede a trader’s purchase immediately to be relevant. As to this, Moses LJ said (at paragraph 62): 40 “The principle of legal certainty provides no warrant for restricting the connection, which must be established, to a fraudulent evasion which immediately precedes a trader’s purchase. If the circumstances of that purchase are such that a person knows or should know that his 45 purchase is or will be connected with fraudulent evasion, it cannot matter a jot that that evasion precedes or follows that purchase. That 6 trader’s knowledge brings him within the category of participants. He is a participant whatever the stage at which the evasion occurs”; iv) the input tax denied need not equate to the tax lost as a result of the 5 relevant fraud. Moses LJ said (at paragraph 65): “It is true that there may well be no correlation between the amount of output tax of which the fraudulent trader has defrauded HMRC and the amount of input tax which another trader has been denied. But the 10 principle is concerned with identifying the objective criteria which must be met before the right to deduct input tax arises. Those criteria are not met … where the trader is regarded as a participant in the fraud. No penalty is imposed; his transaction falls outwith the scope of VAT and, accordingly, he is denied the right to deduct input tax by reason of 15 his participation”; v) tribunals “should not unduly focus on the question whether a trader has acted with due diligence” (to quote from paragraph 82). Moses LJ observed (at paragraph 82): 20 “Even if a trader has asked appropriate questions, he is not entitled to ignore the circumstances in which his transactions take place if the only reasonable explanation for them is that his transactions have been or will be connected to fraud”. 25 Should Mobilx be followed?[13]The submissions advanced by Mr Michael Patchett-Joyce on behalf of S&I were in a range of respects inconsistent with the Court of Appeal’s decision in 30 Mobilx. Mr Patchett-Joyce did not shrink from saying that the approach adopted by the Court of Appeal in Mobilx (in which he appeared for all but one of the appellants) was wrong and that we should not follow it. Mr Patchett-Joyce recognised that the Upper Tribunal would normally be bound by decisions of the Court of Appeal, but argued that domestic rules of 35 precedent are inapplicable in the present context. European Union law, Mr Patchett-Joyce pointed out, has primacy over domestic law. Accordingly, so it was submitted, the Upper Tribunal (and also the FTT) must disregard a Court of Appeal decision which fails to reflect European Union law accurately. 4014. Mr Patchett-Joyce advanced similar submissions in another Upper Tribunal case, POWA (Jersey) Ltd v Revenue and Customs Commissioners [2012] UKUT 50 (TCC), Roth J’s decision in which was released on 8 February 2012. Roth J, however, regarded Mobilx as binding on him. In paragraph 39 of his decision, for example, Roth J observed that: 45 “the judgment of the Court of Appeal is clear authority, binding on the Upper Tribunal, that the fact that the trader claiming credit for input 7 tax did not deal directly with a fraudulent trader but was more remote in the chain does not preclude his being denied repayment under the rationale of Kittel”. 515. The normal rule is that a High Court judge should follow a previous decision of another High Court judge unless convinced that it is wrong: see e.g. R (on the application of B) v London Borough of Islington [2010] EWHC 2539 (Admin), at paragraph 31. Judges of the Upper Tribunal (which was described by Laws LJ as “an alter ego of the High Court” in R (Cart) v The Upper 10 Tribunal [2009] EWHC 3052 (Admin), at paragraph 94) should similarly, it seems to us, usually follow other Upper Tribunal decisions unless convinced that they are wrong. Far, however, from being convinced that Roth J was wrong to take the view that Mobilx is binding on the Upper Tribunal, we agree with him. 15[16]The relationship between European law and domestic rules of precedent was the subject of comment in Condé Nast Publications Ltd v Customs and Excise Commissioners [2006] EWCA Civ 976, [2006] STC 1721. It was suggested in that case that a previous decision of the Court of Appeal (Fleming (t/a 20 Bodycraft) v Customs and Excise Comrs (Condé Nast Publications Ltd intervening) [2006] EWCA Civ 70, [2006] STC 864) was inconsistent with European law and, hence, that it should not be followed (see paragraph 34). The Court of Appeal, however, concluded that it was not free to refuse to follow the decision of the majority of the Court of Appeal in Fleming. 25 Chadwick LJ (with whom Arden and Smith LJJ agreed) said this: “[44] I am content to assume that there may be circumstances in which the obligation imposed on courts by s 3(1) of the European Communities Act 1972 would require this court to refuse to follow its 30 own earlier decision as to the meaning and effect of a Community instrument—including, in the present context, the effect of a judgment of the Court of Justice. Those circumstances would, I think, include a case in which the judgment of the Court of Justice under consideration by this court in the earlier case had been the subject of 35 further consideration—and consequent interpretation, explanation or qualification—by the Court of Justice in a later judgment. But, as it seems to me, one constitution in this court should not substitute its own view as to the effect of a judgment of the Court of Justice for the view which has been reached by another constitution in this court in 40 an earlier case on consideration of the same judgment in circumstances in which there has been no opportunity for the Court of Justice to review that judgment. In those circumstances, if persuaded that there are strong grounds for thinking that the earlier decision is wrong (as a matter of Community law) this court may think it right to 45 refer the point to the Court of Justice for a preliminary ruling. Or it may follow the earlier decision and give permission to appeal. But it should not refuse to follow the earlier decision merely because, on the 8 same material and the same arguments, it is satisfied that a different conclusion should have been reached. [45] The need for a disciplined adherence to precedent in a comparable (but not precisely analogous) field was emphasised by 5 Lord Bingham of Cornhill (with whom the other six members of the House expressly agreed on this point) in his speech in Lambeth London Borough Council v Kay; Price v Leeds City Council [2006] UKHL 10 at [40]–[45], [2006] 2 WLR 570 at [40]–[45]. After referring to the observation of Lord Hailsham of St Marylebone LC 10 in Broome v Cassell & Co Ltd [1972] AC 1027 at 1054, that ‘in legal matters, some degree of certainty is at least as valuable a part of justice as perfection’, Lord Bingham said this (see [2006] 2 WLR 570 at [43]): ‘[43] … That degree of certainty is best achieved by 15 adhering, even in the Convention context, to our rules of precedent. It will of course be the duty of judges to review Convention arguments addressed to them, and if they consider a binding precedent to be, or possibly to be, inconsistent with Strasbourg authority, they may express their 20 views and give leave to appeal, as the Court of Appeal did here. Leap-frog appeals may be appropriate. In this way, in my opinion, they discharge their duty under the 1998 Act. But they should follow the binding precedent, as again the Court of Appeal did here.’” 25[17]Mr Patchett-Joyce observed that Chadwick LJ did not have available to him the subsequent decisions of the ECJ in cases such as Skatteverket v Gourmet Classic Ltd (Case C-458/06), Kücükdeveci v Swedex Gmbh & Co KG (Case C- 555/07), and Elchinov v Natsionalna zdravnoosiguritelna kasa (Case C- 30 173/09). We do not think, however, that these cases undermine what Chadwick LJ said in the Condé Nast case.[18]If the Court of Appeal “should not refuse to follow [an] earlier decision merely because, on the same material and the same arguments, it is satisfied that a 35 different conclusion should have been reached”, still less should the Upper Tribunal take it upon itself to decline to follow a Court of Appeal decision in such circumstances. The position might be different if a subsequent decision of the ECJ had cast new light on the matter, but there can be no question of the Upper Tribunal “substitut[ing] its own view as to the effect of a judgment of 40 the [ECJ] for the view which has been reached by [the Court of Appeal] in an earlier case on consideration of the same judgment in circumstances in which there has been no opportunity for the [ECJ] to review that judgment” (to adapt words of Chadwick LJ). 45 19. In the present case, the ECJ has not reviewed the relevant law in any significant way since Mobilx was decided in May 2010. It is therefore 9 incumbent on us to follow the interpretation of the law which the Court of Appeal adopted in Mobilx. The translation of “impliquée dans” 5[20]As already mentioned, the English version of the ECJ’s judgment in Kittel speaks of a taxable person being refused the right to deduct where he knew or should have known that he was participating in a transaction “connected with” fraudulent evasion of VAT (paragraphs 59 and 61 of the judgment). Mr 10 Patchett-Joyce pointed out that the French text uses the words “impliquée dans” in place of “connected with”. Thus, the passage from Kittel quoted in paragraph 11 above reads as follows in the French:
“il est établi, au vu des éléments objectifs, que l’assujetti savait ou 15 aurait dû savoir que, par son acquisition, il participait à une opération impliquée dans une fraude à la TVA”. “Impliquée dans” also appears in paragraphs 2, 28, 51, 52, 56 and 60, where the English version uses “connected with”
. The words “impliquée dans” 20 feature in paragraphs 17 and 27 of the French version of the judgment too. In paragraph 17, “impliqués dans un vaste mecanisme de fraude” is used where the English version reads “involved in a major fraud scheme”. In paragraph 27, “etait impliquée dans une fraude commise par le vendeur” corresponds to “was part of a fraud committed by the seller” in the English version. 25[21]Mr Patchett-Joyce submitted that, since the language of the Kittel case was French, the French version of the judgment is both the original and authentic one. That, he argued, is important because the sense of the words “impliquée dans” is most closely expressed in English by the words “involved in” or “part 30 of” rather than “connected with”, which, it was said, is a much looser expression. A transaction, Mr Patchett-Joyce submitted, is only “impliquée dans une fraude” if it is “involved in”, “part of” or “aimed at” the fraud. MTIC fraud (so it was said) is perpetrated by the missing trader, so: 35 “[t]he transaction that is ‘involved in’ the fraud, or which is ‘part of’ the fraud, or which is ‘aimed at’ a fraud is the transaction between the fraudster and the immediate counterparty of the fraudster to whom the fraudster had sold the goods, and by whom he had been paid a VATinclusive price but, having obtained VAT, fraudulently had failed to 40 account for it”.[22]Mr Malcolm Davis-White QC, who appeared for HMRC with Mr Aidan Robertson QC, argued that S&I should not be allowed to run this argument. The point was not, he said, raised either in the original Notices of Appeal from 45 HMRC’s decisions or before the FTT, and it would be unfair to allow S&I to take it for the first time on appeal to the Upper Tribunal. Had the contention been advanced in the Notices of Appeal or even before the FTT, HMRC might 10 have approached matters differently. Among other things, HMRC could have sought to adduce evidence that immediate counterparties of S&I had been fraudulent. 5 23. In support of his submissions on this aspect, Mr Davis-White took us to Crane v Sky In-Home Ltd [2008] EWCA Civ 978 and BT Pension Scheme Trustees Ltd v British Telecommunications plc [2011] 2071 (Ch). In Crane, where the Court of Appeal considered whether to allow an appellant to raise a new case, Arden LJ said (in paragraph 21) that the court had “to be satisfied the [the 10 respondent] will not be at risk of prejudice if the new point is allowed because it might have adduced other evidence at trial, or otherwise conduct the case differently”. At paragraph 22, she commented:
“the principle that permission to raise a new point should not be given 15 lightly is likely to apply in every case, save where there is a point of law which does not involve any further evidence and which involves little variation in the case which the party has already had to meet”
. In the BT Pension Scheme case, Mann J arrived at the following conclusions 20 as to the circumstances in which a party should be permitted to depart from a conceded position in the context of an appeal (see paragraph 44): “i) The resiling party has the burden of establishing that the previously forgone point should be raised. 25 ii) It will be harder to raise a point which has been expressly conceded. iii) If taking the point would risk causing prejudice to the other party, in the sense that it might have been deprived of the opportunity of dealing with the case differently in the court below, then it is unlikely that resiling will be allowed. The greater the risk, the less likely it is 30 that it will be allowed. iv) There is a low threshold of risk for these purposes …. v) The burden of establishing no risk is on the party who wishes to withdraw the concession, and the other party should have the benefit of any doubt in this area.” 35 24. Mr Patchett-Joyce took issue with the proposition that HMRC might have approached matters differently if the “impliquée dans” point had been taken sooner. It seems to us, however, that it is by no means fanciful to suppose that HMRC could have acted in a different way (in particular, by seeking to assemble evidence that immediate counterparties had been fraudulent) had the 40 issue been raised before the FTT. In our view, there would be a real risk of prejudice to HMRC were we to permit S&I to pursue the “impliquée dans” argument now. 11[25]Mr Patchett-Joyce prayed in aid the primacy of European law. However, we have not been persuaded that there is any principle of European law which entitles an appellant to insist on raising a point for the first time on appeal where doing so could unfairly prejudice the respondent. 5[26]Mr Patchett-Joyce further submitted that it could not be right for the Upper Tribunal to be precluded from applying the true legal principles. However, it is implicit in cases such as Crane v Sky In-Home Ltd and BT Pension Scheme Trustees Ltd v British Telecommunications plc that an appellate Court or 10 Tribunal will not always be able to take account of all the law that might be relevant.[27]In all the circumstances, it seems to us that the point that S&I wishes to pursue is not open to it. It was not taken before the FTT, and it would not be fair for 15 us to give S&I permission to raise it now. Since, however, we have heard full argument on it, we shall nevertheless comment on it.[28]Mr Patchett-Joyce sought support for his contentions on the point in the decisions of the ECJ in R (Teleos plc and others) v Customs and Excise 20 Commissioners (Case C-409/04) [2008] QB 600, Netto Supermarket GmbH & Co OHG v Finanzamt Malchin (Case C-271/06) [2008] STC 3280 and Criminal proceedings against R (Case C-285/09) [2011] STC 138. Mr Patchett-Joyce pointed out that the words “impliquée dans” feature in the French versions of all these judgments. Of Teleos, Mr Patchett-Joyce said that, 25 in using the words “impliquée dans” and translating them as “involved in”, the Court “must have intended to clarify the inconsistent translation of those words in Kittel and resolve the ambiguity, inherent in the English translation of that judgment, in favour [of] the subsequent consistent translation, ‘involved in’.”. The equation of “impliquée dans” with “involved in” was, Mr Patchett- 30 Joyce submitted, continued in Netto, and the construction was put beyond doubt by R.[29]Mr Patchett-Joyce advanced submissions to similar effect to Roth J in the POWA (Jersey) case. Roth J rejected them not only as being inconsistent with 35 Mobilx, but after a review of the ECJ authorities. He concluded that the question whether “impliquée dans” is better rendered as “involved in” as opposed to “connected with” made no substantive difference. He said (in paragraph 34): 40 “irrespective of whether the test should be expressed as ‘connected with fraudulent evasion’ or ‘involved in the fraudulent evasion’, I consider that, if PJL [the appellant] should have known that the transactions in which it was engaged were part of a chain in which one or more earlier transactions were fraudulent, albeit that its immediate 45 supplier was not dishonest, that test [i.e. the test derived from Kittel] is satisfied”. 12[30]We agree with Roth J’s conclusions as to the law. In the light of the submissions we heard on the issues, we would add the following points to those he made: 5 i) Roth J noted that Moses LJ would have been aware in Mobilx of the “impliquée dans” point because it was mentioned by the Chancellor in his judgment in Blue Sphere Global Ltd v Revenue and Customs Commissioners [2009] EWHC 1150 (Ch), [2009] STC 2239. It can be seen from the materials put before us that the point also featured in the 10 skeleton arguments before the Court of Appeal. Mr Patchett-Joyce’s skeleton argument in support of the Calltel appeal made specific reference to Teleos and Netto and the fact that “the English word ‘involved’ (or variant) is used to translate the French term ‘impliquée dans’ (or variant)”; 15 ii) There is, as it seems to us, a further reason why the ECJ’s use of “involved in”/“impliquée dans” in Teleos and Netto does not help with the interpretation of equivalent words in Kittel. Far from it being the case that the ECJ was trying to clarify what it had meant when it used 20 “impliquée dans” in Kittel, the words appear to have been used in a somewhat different sense in the French version of the Teleos and Netto judgments. In Teleos (and also Netto, which followed Teleos) the words “involved in”/“impliquée dans” referred to complicity. Thus, in Teleos, having recorded that HMRC acknowledged that the claimants 25 “were in no way involved in any fraud” (paragraph 16 of the judgment), the Court observed that a regime imposing responsibility on suppliers, “regardless of whether or not they were involved in the fraud”, does not necessarily safeguard the system from evasion and abuse by purchasers (paragraph 58 of the judgment). On each occasion, 30 “involved in” was translated as “impliquée dans”. However, neither HMRC nor the Court was saying anything about whether a transaction had to be one to which a fraudster was the immediate counterparty. What was being said was that the relevant persons were not complicit in the fraud. That is not, as we see it, the sense in which “impliquée 35 dans” was used in Kittel. When, say, the Court referred in Kittel (at paragraphs 59 and 61) to a supply to a person who “savait ou aurait dû savoir que, par son acquisition, il participait à une opération impliquée dans une fraude”, the words “impliquée dans” were intended to say something about the relationship between the “opération” and the 40 fraud, not about whether the person was complicit. Words do not usually have single immutable meanings (see e.g. Phillips and Strattan v Dorintal Insurance Ltd [1987] 1 Lloyd’s Rep. 482). With Kittel, Teleos and Netto, as elsewhere, the meaning of words is influenced by context. 45 13 Other challenges to Mobilx[31]Mr Patchett-Joyce did not challenge the correctness of the Court of Appeal’s decision in Mobilx only by reference to the meaning of “impliquée dans”, as 5 used in Kittel. He also submitted that the Court of Appeal’s approach offended fundamental principles of EU law in a range of ways. Reference was made to principles of legal certainty, fiscal neutrality, proportionality and effectiveness. Human rights jurisprudence was invoked as well. It was argued that Article 1 of the First Protocol to the European Convention on Human Rights indicates 10 that the circumstances in which a trader can be denied input tax must be interpreted narrowly. It was contended too that, in the absence of domestic legislation to give effect to them, the principles espoused by the ECJ in Kittel cannot represent the law in this country. 15 32. We can dispose of all these points shortly. To at least a substantial extent, they involve a re-run of arguments that Mr Patchett-Joyce also advanced, without success, in Mobilx. In paragraphs 45-49 of his judgment, Moses LJ considered and rejected the proposition that the principles enunciated in Kittel could not be applied as part of UK domestic law without specific legislation. At 20 paragraphs 61 and 62, he addressed the impact of the principle of legal certainty. At paragraphs 63-65, he explained why he regarded his approach to the law as compliant with Article 1 of the First Protocol to the European Convention on Human Rights. At paragraph 66, he concluded that it was “not arguable that the principles of fiscal neutrality, legal certainty, free movement 25 of goods and proportionality were infringed by” the ECJ in Kittel. In his skeleton argument in support of the Calltel appeal, Mr Patchett-Joyce referred to the principle of effectiveness in much the same way as in his skeleton argument for the present appeal. In any case, as we have already said, we regard ourselves as bound by the Court of Appeal’s decision in Mobilx. 30 Non-discrimination[33]Another ground of appeal relies on the principle of non-discrimination. One implication of the principle, Mr Patchett-Joyce argued, is that EU law needs to 35 be applied uniformly. That was not the case, it is said, in the present context: HMRC denied input tax to S&I and exporters (referred to as “brokers”) but not to “buffers” in the chains who bought and sold phones within the United Kingdom. 40 34. It was similarly submitted to the FTT that HMRC should be barred from contesting input tax entitlement if their approach was tainted by a discriminatory policy for which there was no objective justification. The FTT, however, decided that HMRC’s conduct was not relevant. Among other things, it concluded that “the effect of the Kittel principle is to limit a trader’s 45 right to repayment of VAT” rather than to confer a discretion on HMRC (paragraph 82), with the result that “HMRC’s action cannot affect the proper amount of tax in this case” and “the actions of the administration in applying 14 that law and in selecting the cases in which they seek to apply it are irrelevant to us” (paragraph 84). “If,” the FTT went on, “S&I say that HMRC’s actions are, in this case, or more generally, contrary to EU law, then that cannot be a matter for this tribunal for we have no general jurisdiction to review their 5 actions: S&I must seek relief in a different forum” (paragraph 84).[35]A non-discrimination argument was also rejected in 4 Distribution Ltd v Commissioners for Her Majesty’s Revenue and Customs [2009] UKFTT 242 (TC). In that case, the FTT (Judge Walters QC and Ms O’Neill) took the view 10 that “circumstances which demonstrate that the Appellant has abused its right to repayment of input tax also demonstrate that it has abused any right not to be discriminated against as a trader supplying to an entity in another Member State” (paragraph 133). The FTT continued as follows: 15 “134. The ECJ said in Kittel at [54] that ‘Community law cannot be relied on for abusive or fraudulent ends’ and cited Kefalas and Others (Case C-367/96) [1998] ECR I-2843 at [20], Diamantis (Case C-373/97) [2000] ECR I-1705 at [33] and Fini H (Case C-32/03) [2005] ECR I-1599 at [32] in support of that proposition. 20 135. In any event, we are not satisfied that the circumstances of this appeal do, or even could, give rise to any right in the Appellant's favour not to be discriminated against as a trader supplying to an entity in another Member State.[136]It seems to us that the Appellant's point that where a Tribunal 25 has found objective knowledge sufficient to deny repayment of input tax as a matter of law, nevertheless that result can, as a matter of law, be reversed by reliance on another Community law principle (equal treatment) is clearly misconceived.” 30 36. More recently, Roth J rejected a non-discrimination argument in the POWA (Jersey) case. At paragraph 60 of his decision, Roth J said: “As to non-discrimination, this appeal concerns the decision by HMRC that the objective criteria determining the right to deduct input tax were 35 not met as regards these claims for repayment by PJL [the appellants]. If that is the case, PJL were not entitled to such repayments, irrespective of the position of anyone else …. Furthermore, whether or not HMRC could have applied a similar approach to the traders who served as buffers in the chains (who would generally not be making a 40 repayment claim to HMRC but simply crediting the input tax against the output tax received) does not affect that conclusion; and whether HMRC should have pursued those traders for an account of the output tax received is a question of policy regarding the effective enforcement of the VAT regime, with no doubt limited resources. Accordingly, I 45 consider that the principle of non-discrimination is not engaged.” 15[37]Essentially the same considerations apply in the present case, and it appears to us that in this case, as in POWA (Jersey), the principle of non-discrimination is not engaged. We thus agree with the FTT on this aspect of the case. 5 Extent of tax loss[38]In the case of a small number of the transactions at issue, S&I challenges findings as to the extent of the related tax loss. As regards deals 18, 66, 70, 71, 74 and 75, the FTT concluded that the relevant phones had probably been 10 imported and, hence, that there had been tax loss equal to the amount of output tax on the earliest identified sale within the United Kingdom. S&I argues that the evidence did not entitle the FTT so to hold.[39]If we are correct in our conclusions on HMRC’s appeal (as to which, see 15 paragraphs 53-57 below), the extent of the tax loss does not matter: there need be no correlation between input tax denied and tax lost. We shall nonetheless address this ground of appeal briefly.[40]Guidance as to the grounds on which factual findings can be challenged on 20 appeal is to be found in Edwards v Bairstow [1956] AC 14. Viscount Simonds there said (at 29) that a finding of fact should be set aside if it appeared that the finding had been made “without any evidence or upon a view of the facts which could not reasonably be entertained”. Lord Radcliffe (at 35) quoted a passage from a judgment of Lord Normand in which the latter had said that an 25 appellate Court could intervene if the lower tribunal had “misunderstood the statutory language” or had “made a finding for which there is no evidence or which is inconsistent with the evidence and contradictory of it”. Lord Radcliffe went on to say this (at 36) about the position where “the facts found are such that no person acting judicially and properly instructed as to the 30 relevant law could have come to the determination under appeal”:
“I do not think that it much matters whether this state of affairs is described as one in which there is no evidence to support the determination or as one in which the evidence is inconsistent with 35 and contradictory of the determination, or as one in which the true and only reasonable conclusion contradicts the determination. Rightly understood, each phrase propounds the same test. For my part, I prefer the last of the three, since I think that it is rather misleading to speak of there being no evidence to support a 40 conclusion when in cases such as these many of the facts are likely to be neutral in themselves, and only to take their colour from the combination of circumstances in which they are found to occur.”
[41]In the present case, we are satisfied that the evidence before the FTT entitled it 45 to find that the phones comprised in deals 18, 66, 70, 71, 74 and 75 had probably been imported. The FTT explained its thinking in sections 8, 29, 33 and 34 of Appendix 2 to its decision. It clearly gave careful consideration to 16 the issues, and its reasoning appears to us to be cogent. We note the following points in particular: i) stock allocation and release notes indicated that Goodluck Employment 5 Services Limited, to which the phones in deal 18 were traced back, had been importing phones from elsewhere in the European Union; ii) as the FTT observed, “third party payment instructions and Crossview’s evanescent appearance” tended to suggest that either 10 Crossview Consortium Limited, to which the phones in deal 66 were traced back, had imported the goods or someone else had. Third party payment instructions were formerly a typical feature of MTIC fraud; iii) the individual behind RS Sales Agency Limited, to which the phones in 15 deals 70, 74 and 75 were traced back, made an enquiry on 17 May 2006 about whether he could import electrical goods using the existing VAT number; iv) as regards deal 71, where phones were traced back to JD Telecoms UK 20 Limited, a hijacked trader, there was evidence that invoices where goods had been acquired from the European Union differed from other invoices. Extent of knowledge 25[42]Only one of S&I’s grounds of appeal appears to us to have real substance. That relates to what S&I ought to have known.[43]The FTT proceeded on the basis that input tax properly fell to be denied if the 30 trader should have known that the transaction into which he was entering was more likely than not to be connected with fraud. Thus, the FTT said this in paragraph 33 of its decision:
“[W]e would, in the absence of further guidance, have applied the 35 following test in relation to the ‘should have known’ question: namely whether a reasonable man with ordinary competence in the position of S&I, and knowing what S&I knew, (a) would have taken any additional steps, and (b) would have come to the conclusion, on the basis of what he knew and had found out, that it was more likely than 40 not that the transaction concerned was connected to fraud.”
In paragraph 36 of the decision, the FTT explained: “In the formulation in para 33 we say ‘more likely than not’: it seems 45 to us that no higher test is indicated”. 17[44]The FTT’s approach was warranted by the authorities as they stood at the time of its decision, but, as Mr Davis-White accepted, Mobilx establishes that it was wrong. As mentioned above, the Court of Appeal held in Mobilx that the right to deduct input tax would be lost if a trader knew or should have known that 5 he was taking part in a transaction connected with fraudulent evasion of VAT, but not if he merely knew or should have known that the transaction was more likely than not to be so connected.[45]In Mobilx itself, the Court of Appeal concluded that the wrong test had been 10 applied at first instance as regards Mobilx Limited (see paragraph 77). The Court of Appeal nevertheless decided that input tax had rightly been denied. An incorrect test having been adopted, the question arose (paragraph 68):
“whether, on the application of the correct test, the true and only 15 reasonable conclusion is that the trader knew or should have known that his transactions were connected with fraud or that there was no reasonable possibility other than they were … connected with fraud”
. On the facts, that question fell to be answered in the affirmative. The case was 20 one in which “Mobilx knew that those transactions which could be traced by HMRC had led back to fraud in the past in a trade where fraud was rife” but “chose not to change the manner in which it conducted its trade but merely continued to trade in the same pattern as before” (paragraph 79). As Moses LJ explained (at paragraph 80), on the basis of the tribunal’s findings: 25 “the true and only reasonable conclusion, is that Mobilx ought to have known that the only realistic possibility, as it continued to trade in that manner, was that its purchases would be connected with fraudulent evasion of VAT and not merely that all its transactions were more 30 likely than not to be connected with fraud. In those circumstances, despite the tribunal's error of law in the test which it applied, that error makes no difference to the true and only reasonable conclusion”.[46]Can it then be said in the present case that the FTT’s findings mean that “the 35 true and only reasonable conclusion is that [S&I] knew or should have known that [its] transactions were connected with fraud or that there was no other reasonable possibility other than that they were connected … with fraud”?[47]Mr Davis-White argued that it can, Mr Patchett-Joyce that it cannot. On 40 balance, we agree with Mr Patchett-Joyce.[48]In the first place, the FTT’s decision contains numerous references to risks, concerns and likelihood. In paragraph 199 of its decision, for example, the FTT said that S&I must have been aware by 2006 that its business was “not 45 immune from the risk that its transactions might be connected to fraud (or at least that a reasonable businessman in possession of the information held by S&I would have been so aware)”. After discussing respects in which S&I’s 18 dealings were said to be uncommercial, the FTT concluded that they “do not point unequivocally to fraud, but they would raise in the mind of a reasonable businessman serious concerns about such a connection” (paragraph 201). With regard to certain suppliers, the FTT expressed the view that a reasonable 5 businessman possessed of the knowledge of S&I would have “come to the conclusion that it was more likely than not that purchases from the identified suppliers would have been connected with the fraud of a person earlier in the chain of supply to it” (paragraph 208). So far as other suppliers were concerned, the FTT considered that a variety of factors would have “given 10 rise, in the mind of a reasonable businessman, to very serious concern, and possibly a conclusion that it was more likely than not, that each of S&I’s April, May and June transactions would have been connected with fraud” (paragraph 214). 15 49. A second point arises from the fact that, when S&I undertook the transactions at issue, it knew that HMRC had checked a number of its earlier transactions and traced a proportion of those selected to a missing trader or defaulter, but that not all of them had been so traced (see paragraph 183 of the FTT’s decision). The relevant letters from HMRC “did not indicate whether the deals 20 in each sample which had not been traced back to a defaulter had been traced back to an importer who had paid VAT” and the FTT did “not believe that a reasonable man would have concluded that it was only the identified deals which were connected to fraud” (paragraph 210(iv) of the decision). Even so, this is not a case where the trader had been informed that every previous 25 transaction (or even every one of those that HMRC had inquired into) was connected with fraud.[50]A third point is that, while the FTT found that the majority of the transactions under appeal were connected with fraud, it did not consider a connection with 30 fraud to have been proved in every case. If S&I was entering into transactions which were not connected with fraud (or may have been doing so), it is harder to say that, with the balance, there was “no other reasonable possibility other than that they were connected … with fraud”. 35 51. In all the circumstances, we do not think the FTT’s findings entitle us to infer that “the true and only reasonable conclusion is that [S&I] knew or should have known that [its] transactions were connected with fraud or that there was no other reasonable possibility other than that they were connected … with fraud”. On the other hand, we do not consider either that we would be justified 40 in deciding the opposite: that it could not reasonably be concluded that S&I should have known that the relevant transactions were (and not merely were likely to be) connected with fraud. The FTT framed its decision by reference to what it took to be the appropriate test, viz. whether S&I should have known that its transactions were more likely than not to be connected with fraud. It is 45 not clear how the FTT would have answered the question, “Should S&I have known that its transactions were connected with fraud?” 19[52]This being so, we can see no alternative but to remit the case to the FTT. HMRC’s appeal 5 53. The question raised by HMRC’s appeal is this: Should a trader who is to be treated as having participated in VAT fraud (because he either knew or should have known that he was participating in a transaction connected with such fraud) be deprived of all the relevant input tax or only so much of it as is equal to the tax lost as a result of the fraud? 10[54]As already explained (in paragraph 7 above), the FTT, following the decision of the VAT & Duties Tribunal in Honeyfone Ltd [2008] UKVAT 20667, concluded that input tax should be denied only to the extent of the tax loss (see paragraphs 77 and 78 of the decision). 15[55]Shortly afterwards, however, Floyd J took a different view in Calltel Telecom Ltd v Revenue and Customs Commissioners (No 2) [2009] STC 2164. Having noted (in paragraph 94) that the traders with which he was concerned were contending for the principle that “the right to repayment of input tax can 20 continue to be exercised, notwithstanding knowledge of the fraud of the importer, to the extent that the claim for repayment exceeds the loss to HMRC”, Floyd J held (in paragraph 96) that “there is no principle which requires HMRC to acknowledge a claim for repayment to the extent that the claim exceeds HMRC’s tax loss”. 25[56]An appeal from Floyd J’s decision was one of those that the Court of Appeal heard with that relating to Mobilx Limited and on which it gave judgment on 12 May 2010. Before the Court of Appeal, as before us, Mr Patchett-Joyce argued that allowing HMRC to withhold sums in excess of the lost VAT 30 would amount to a penalty. As already mentioned (paragraph 12(iv) above), however, the Court of Appeal determined that the input tax denied need not equate to the tax lost as a result of the relevant fraud. It thus agreed with Floyd J, and the appeal from his decision was dismissed (see the judgment of Moses LJ at paragraph 80). 35[57]It is now clear, therefore, that the FTT was mistaken in thinking that input tax should be denied only to the extent of the tax loss. The position is rather that a trader who falls to be treated as a participant in tax fraud loses the right to any input tax [credit], whatever the extent of the tax loss. It follows that, but for 40 S&I’s cross-appeal, we would have upheld HMRC’s appeal and decided that HMRC had been entitled to disallow the relevant input tax claims in their entirety rather than merely to the extent of the proven tax loss. 45 20 Should there be a reference?[58]Mr Patchett-Joyce suggested that, if we were not willing to depart from Mobilx ourselves, we should refer the relevant questions to the ECJ for a preliminary 5 ruling.[59]Mr Patchett-Joyce also argued for a reference in the POWA (Jersey) case. However, Roth J declined to make a reference. In paragraph 40 of his decision, he said this: 10 “I was told that an application for permission to appeal against the judgment in Mobilx to the Supreme Court had been dismissed as inadmissible, and on that basis I was urged to make a reference to the ECJ of this question based on the translation point. But even if that 15 course were open to me, given that, in my judgment, the alternative translation discussed above does not impinge in any way on the rationale and principle as explained in Mobilx, I would see no ground upon which a reference would be justified. I should add that the fact that there has been a subsequent reference to the ECJ made by the 20 Bulgarian court in Case C-285/11 Bonik, OJ 2011 C238/08, which includes a question or questions related to this point and to which my attention was drawn subsequent to the hearing, does not alter my view.” 25 60. In our view, a reference is no more appropriate in the present case than in POWA (Jersey). In the absence at least of further guidance from the ECJ, it seems to us that the Upper Tribunal should take the law to be as explained by the Court of Appeal in Mobilx. We shall not, therefore, make a reference. 30 Conclusion[61]The case will be remitted to the FTT.[62]It seems to us that there is scope for argument as to the basis on which the 35 further hearing before the FTT should be conducted. The present appeal should be re-listed for argument on that question. 40 Mr Justice Newey Judge John Walters QC 45 RELEASE DATE: 12 March 2012 21

Cited in 17 later judgments

and 8 more of the most senior