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

United Kingdom VAT & Duties Tribunals Decisions


You are here: BAILII >> Databases >> United Kingdom VAT & Duties Tribunals Decisions >> Honeyfone Ltd v Revenue & Customs [2008] UKVAT V20667 (02 May 2008)
URL: http://www.bailii.org/uk/cases/UKVAT/2008/V20667.html
Cite as: [2008] STI 1688, [2008] UKVAT V20667, [2008] BVC 2394

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


Honeyfone Ltd v Revenue & Customs [2008] UKVAT V20667 (02 May 2008)
    20667
    VAT – Input Tax – MTIC Fraud – Whether trader knew or should have known of fraud – Whether HMRC's practice contrary to Article 29 EC Treaty and if so whether any effect on Tribunal's decision – Whether trader should be denied whole of relevant input tax or that corresponding to fraud only

    LONDON TRIBUNAL CENTRE

    HONEYFONE LIMITED Appellant

    THE COMMISSIONERS FOR HER MAJESTY'S REVENUE & CUSTOMS Respondents

    Tribunal: CHARLES HELLIER (Chairman)

    RANBIR SURI J.P.

    Sitting in public in London on 21 to 24 January and 11 to 15 February 2008

    Eleni Mitrophanous, counsel, instructed by BDO Stoy Hayward LLP for the Appellant

    Christopher Foulkes and David Bedenham, counsel, instructed by the Solicitor to HM Revenue and Customs, for the Respondents

    © CROWN COPYRIGHT 2008

     
    DECISION
  1. Heneyfone Ltd deals in mobile phones. Among its transactions in mobile phones in March and April 2006 were a number of exports. It claimed repayment of its input VAT in relation to its purchase of the phones it exported. Those claims were denied by the Respondents in a letter of 31 January 2007.
  2. The Respondents maintain that each of the export transactions formed part of a chain of transactions which started with a 'defaulting trader' who had fraudulently defaulted in its obligation to pay VAT in respect of its supply of the phones, and that the phones then passed along a chain of 'buffer traders' each of whom bought and sold on the phones until they reached the Appellant, who exported them. The Respondents say that Honeyfone knew or should have known that by each of its purchases it was participating in a transaction connected with fraudulent evasion of VAT, and, as a result of the principles enunciated by the ECJ in Axel Kittel v Belgium C-440/04 Belgian State v Recolta Recycling SPRL [2006] ECR 1-6161, Honeyfone is therefore denied its right to deduct and reclaim input tax on its purchase of the phones. Against that result Honeyfone appeals.
  3. These types of transactions, in which a defaulting trader (who has generally imported goods) sells to Buffer A who sells to Buffer B who may then sell to Buffer C etc) who sells to Exporter C, are a species of 'MTIC' fraud – Missing Trader Intra Community fraud. They are described in many other judgments and decisions and there is no need to describe them in detail here (see for example paragraph 6 of R (Just Fabulous (UK) Ltd) v HMRC [2007] EWHC 521 (Admin), and paragraphs 5-7 of Calltell Telecom Ltd v HMRC [2007] UK VAT 20266 which we gratefully adopt), but the following crude example involving 3 UK parties only may help to elucidate the discussion which follows:
  4. (i) Defaulting trader, DF, acquires phones from French exporter for £115
    (ii) DF sells to "Buffer" A for £105 plus £18.375 VAT
    (iii) A sells to "Exporter" E for £110 plus £19.25 VAT
    (iv) E sells to French importer (who may or may not be the original French exporter) for £115 with no VAT.
  5. After these transactions, DF has 8.375 cash in hand and owes VAT of 18.375, A has 5.875 cash in hand and pays the difference between his output tax and input tax (19.25 – 18.375 = 0.875) to the Commissioners leaving him with 5 cash profit in hand, and E is, until he has received repayment of his input VAT, £14.25 short, but when he received it is £5 cash profit in hand. Together DF, A and E make 18.375 – which is what DF owes the Commissioners. If DF defaults in payment, his default represents the combined cash profit of A and E and the cash of £8.375 which DF runs away with. The chains may of course be longer and the profit taken at each stage smaller, but in this crude example the aggregate profit (treating DF's 8.375 as a profit for those purposes) is equal to the VAT on which DF defaults.
  6. We heard oral evidence from two of the Respondents' officers, Fu Sang Lam and Gregory Rodney, and from Ali Shiralizadeh one of the Appellant's directors and Sushil Arora, an employee and shareholder of the Appellant. All provided witness statements. Among other documentary evidence we also decided to admit part of the transcript of the evidence given by Roderick Stone, another of the Respondents' officers, to the tribunal at the hearing of the appeal of Olympia Technology Ltd. Mr Stone is something of a veteran witness in cases of this sort having also given evidence to the High Court in Just Fabulous and other cases dealing with similar issues.
  7. On the first day of the hearing the Respondents asked us to admit a second witness statement from Mr Rodney with additional exhibits. The statement concerned the allegations of fraud by the defaulting traders and Mr Arora's past history. We decided to admit it: its content was relevant, it had been produced late but its admission would not cause unfairness to the Appellant in the circumstances. On the second day Miss Mitrophanous asked that we admit in evidence part o the transcript of the evidence given by the Respondents' office, Mr Roderick Stone, to the tribunal in Olympia (see below). Mr Foulkes did not object and we decided to admit it but to have regard to the fact that we did not see Mr Stone or hear all his evidence. In the afternoon of the second day we admitted some additional documentary evidence provided by Mr Rodney at the Respondents' request. It related to relevant matters. On the afternoon of the fourth day of the hearing (but pursuant to an advertisement on the third day), and after we had heard the evidence of Mr Lam and Mr Rodney, the Respondents sought to adduce further documentary evidence relating to the defaults by the defaulting traders. We looked at the evidence and decided not to admit it. That was for the following reasons:-
  8. (i) whilst generally the tribunal would lean towards admitting relevant evidence because the better the evidence available to it, the fairer and more just its decision can be, then comes a time when, in the interests of the fair and orderly management of the hearing, enough is enough. When that time may come will depend on the nature of the new evidence and the circumstances of its production;
    (ii) the evidence related to matters which the Respondents knew from the start they would have to prove; it was not something which arose out of a need to counter a statement made by a witness, or a sudden turn in the proceedings. It was more in the nature of evidence which the Respondents could have provided earlier and forgot until the cross-examination of this witnesses started. There is a general interest in this sort of evidence being disclosed at an early stage;
    (iii) the late admission of new evidence would extend the hearing further: the original estimate had been 3 days; we were preparing to sit for another 5 days. It can be unjust and unfair to prolong a hearing beyond certain expectations;
    (iv) the evidence itself did not seem to be of critical significance: there was not a clear prejudice which arise to either party's case from admitting it; and
    (v) on balance the prejudice to the timing and presentation of the Appellant's case from admitting the evidence outweighed the prejudice to the Respondents in refusing it.
    The Structure of this Decision
  9. We divide this decision into the following parts:
  10. The EU Discrimination Argument
  11. Under this heading we discuss an argument put forward by Miss Mitrophanous that the Commissioner's approach to its decision in relation to Honeyfone was unlawful because it was a manifestation of a practice which discriminated against traders exporting from the UK to the EU.
    We make findings of fact and then reach our conclusions on the argument.
    We released this part of our decision to the parties at an earlier stage because our conclusion involved matters of EU law and we wished to give the parties the opportunity of arguing for a reference to the ECJ if they wished.
  12. The Legal Issues
  13. We set out here our interpretation of the other law relevant to the potential denial of Honeyfone's input tax on the grounds that its transactions were connected with VAT fraud and that it knew or should have known that they were so connected.
  14. Background Facts and the Deal Chains
  15. Here we make general findings of fact and specific findings in relation to the alleged chains of transactions which culminated in Honeyfone's exports.
  16. The Existence of Fraud
  17. Here we make findings of fact in relation to the presence of fraud in the deal chains.
  18. Knew or should have known
  19. We make further findings of fact and then consider the extent to which Honeyfone knew or should have known of any fraud.
  20. All the input tax or that lost only?
  21. Here we consider, if the defaulting trader defaulted on £18.375 (see example in paragraph 5 above), and Honeyfone's input tax is £19.25, and in circumstances where Honeyfone knew or should have known of the fraud, whether the whole £19.25 is to be denied or only the £18.375 tax lost to the Exchequer.
  22. Conclusion
  23. The EU Discrimination Argument
  24. We make the following findings of fact in relation to this argument. These findings are drawn principally from the evidence of Mr Lam and Mr Rodney but are supported by the findings in Calltell and the transcript of Mr Stone's evidence.
  25. Findings of Fact
    (1) In early 2006 the Respondents' management, as part of their efforts to tackle the tax loss caused by MTIC fraud, developed criteria for undertaking 'extended verification' on certain repayment claims. If two or more of the following criteria applied: if a trader's claim exceeded a certain amount, if the trade was in a certain trade sector (such as mobile phones) of if the trader had a connection with an MTIC fraud or person involved in such a case, then the repayment claim would not be paid after the making of fairly routine checks, but would be subject to extended verification. That was the process of tracing the goods exported back through the supplier, to the supplier's supplies and so on. The object was to determine whether or not the goods could be traced back to a defaulting trader.
    (2) If the extended verification revealed that goods could be traced back to a defaulting trader the Commissioners consider whether to deny the exporter's input tax claim on Kittel grounds: they consider whether the trader knew or should have known of that it was participating in transactions connected to VAT fraud. The greater the number of a trader's export transactions which trace back to defaulting traders, the more likely they are to conclude that the trader knew or should have known of the fraud.
    (3) If all the export transaction goods trace back to one or more defaulting trader the Commissioners are likely to conclude that the exporter did have means of knowledge, or knew, of the fraud.
    (4) Having established a chain back from exporter to defaulting trader, the Commissioners do not seek to deny an input tax deduction to the buffer (UK to UK) traders in the chain.
    (5) There is no practice of extended verification of the input tax claims of any buffer.
    Background
  26. For present purposes it is sufficient to summarise the decision in Kittel in the following terms, although at a later stage we shall examine it in more detail: where a supply is made to a taxable person who knew or should have known that, by his purchase, he was participating in a transaction connected with fraudulent evasion of VAT "it is for the national court to refuse" that person's right to deduct the input tax on his purchase.
  27. Thus if Buffer A who buys from Defaulting Trader knows that Defaulting Trader intends fraudulently to evade paying the £18.375 VAT in the example above, then Buffer A may be denied the right to deduct the £18.375 input tax on his purchase from Defaulting Trader.
  28. Buffer A has sold the goods on to E, and has been liable to account for output tax of £19.25 on that sale. Ordinarily he would have a right to deduct his £18.375 input tax and account to HMRC for the balance of £0.875 only. But if his right to input tax credit is denied he becomes liable to account for the full £19.25.
  29. Accordingly if HMRC apply the Kittel to Buffer A (who will have made a VAT return showing only £0.875 due, and paid that amount) they will need to assess Buffer A for the difference and take steps to recover it if Buffer A is unable or unwilling to pay.
  30. It seems to us that it follows from the judgment of the ECJ in Optigen Ltd, Falcrum Electronics Ltd and Bond House Ltd v Commissioners of Customs & Excise [2006] STC 419 that Buffer A's knowledge of, or acting as an accomplice in, DF's fraud does not affect the VAT which is due on its sale to E. In that case the ECJ considered whether transactions which were not themselves vitiated by VAT fraud but formed part of a chain in which another transaction was vitiated by such fraud constituted taxable supplies in the course of economic activity. The Court held that 'the mere fact that conduct amounts to an offence is not sufficient to justify exemption from VAT' (paragraph 49) and that the transactions described constitute supplies "where they fulfil the objective criteria on which the definitions of those terms are based", regardless of the trader's intention "and/or the fraudulent nature of another transaction in the chain".
  31. But Exporter E offers the Respondents an easier target: they do not have to seek a payment but merely deny repayment. If Exporter E knows or should have known that he is participating in a transaction connected to fraud then Exporter E's right to deduct his input tax of £19.25 may be denied. Because Exporter E has exported the goods there will be no output tax on his sale: he has exercised his right to deduct by reclaiming £19.25 from the Commissioners. He makes a VAT return submitting that claim. But because of his state of knowledge that claim may be denied.
  32. This opens the possibility that the Respondents, were they both to assess Buffer A and also deny Exporter E its repayment, would make an overall profit: they would have lost £18.75 in DF's default, and recovered an assessment of £19.25 from A without have to repay anything to B: they would thus have received £19.25 net from transactions which should have yielded no net VAT take because they ended with exported goods. In Just Fabulous this possibility was referred to as Multiple Recovery (see paragraphs 33-35, 41-47). In that case the Commissioners made clear that there was no intention to achieve multiple recovery, and in this case Mr Foulkes indicated that the Commissioners undertook that they would not seek double recovery.
  33. The Appellant's Discrimination Argument
  34. Miss Mitrophanous says that the Commissioners' practice (which is to apply the derogation from the right to deduct input tax sanctioned by Kittel only or mostly against exporters) is contrary to EU law.
  35. She says that the evidence showed that the Commissioners 'focus' on exporters and that they do not deny buffer traders their input tax claim. Mr Foulkes says that action is in fact also taken against buffers but does not deny that a high proportion of input tax denials have been against brokers/exporters. Our findings of fact on the evidence before us are that the practice of extended verification is applied wholly or mainly to exporters rather than buffers. It is exporters therefore who are more at risk of 'attack' than buffers. As Burton J noted in Just Fabulous 'a refusal to refund inputs is a very straightforward remedy for the Revenue'.
  36. Miss Mitrophanous says that this practice unlawfully discriminates against exporters. It makes it more difficult to be an exporter rather than a seller on the UK domestic market. We agree.
  37. She notes that: Article 29 of the EU treaty provides that "quantitative restrictions on exports and all measures having equivalent effect shall be abolished between Member States"; the ECJ have held that this provision is directly applicable and so confers rights on individuals which domestic courts must enforce; and that in Grilli C-12/02 the court noted that Article 29 prohibits national measures which:
  38. "have as their specific object or effect the restrictions of patterns of export and thereby the establishment of a difference in treatment between the domestic trade of a Member State and its export trade, in such a way as to provide a particular advantage for … the domestic market of the state in question …" (paragraph 41)

    and at paragraphs 47 and 48

    "For that purpose the national court will be required, inter alia, to consider whether the national rules at issue … may be justified on grounds of public order or security. If so, it will have to establish whether they are necessary in order to achieve [that] object … and do not constitute the arbitrary discrimination on a disguised restriction of trade between Member States."

    She notes that Article 22 of the Sixth Directive which grants the right to deduct input tax permits (at 22(8)) Member States to impose obligations for the prevention of fraud but that this is:-

    "Subject to the requirement of equal treatment for domestic transactions and transactions carried out between Member States …"

    And that in Commissioners of Customs and Excise v Federation of Technological Industries[2006] STI 1483 the ECJ held that legislation such as that in s.77A VATA under which persons could be made jointly and severally liable for payment of VAT was permitted by the directive, but that the imposition must be subject to the general principles of community law (objectively justifiable, rational, proportionate, and legally certain).

  39. Miss Mitrophanous says, relying on Adriaan De Peijper C-12/02 that:-
  40. (i) the prohibition against discriminating against exporters is not limited to national rules but includes administrative practices such as those of the Respondents;
    (ii) although Article 30 permits restrictions on exports on public policy grounds, restrictions may not be justified where their object may be met by measures which restrict intra community trade less; and
    (iii) a "concern to lighten the administration's burden or reduce public expenditure" is not normally sufficient justification.

    The mode of application of the Kittel derogation by the Respondents she says clearly offends these principles, and also, she says, is contrary to the principle of legal certainty.

    Discussion
  41. It seems to us that there are three ways of viewing the Appellant's discrimination argument: either (i) that the no discrimination principle is a constraint on the Commissioners' actions in that any decision made by the Commissioners in contravention of that principle must be set aside (in the remainder of this decision we shall term such a decision one which is "wrongly" made); (ii) that it acts as a limitation of the Kittel derogation by providing a codicil thereto which reinstates the taxpayer's right to his input tax deduction if the Commissioners act wrongly; or (iii) that it acts on a constraint in the way in which the national courts must apply the Kittel derogation (bearing in mind the terms of the dispositive part of the decision "it is for the national courts to refuse … entitlement" and the absence of any domestic legislation having such an effect).
  42. (i) a wrong decision must be set aside
  43. This formulation raises the question as to what jurisdiction this tribunal has to set aside the Commissioners' decision. Section 83 VATA sets out the matters in respect of which an appeal may lie to a tribunal. In relation to the making of an assessment section 83(p) provides for two types of appeal: an appeal against the amount of the assessment, and an appeal against the assessment itself. That latter possibility permits questions to be raised as to whether the assessment was properly made, and an improperly made assessment may be set aside in limited circumstances (see C&E Commissioners v Pegasus Birds Ltd [2004] STC 1509). However, in relation to a claim to input tax the statutory provision is section 83(c); "the amount of any input tax which may be credited to any person". That formulation does not clearly encompass a jurisdiction for this tribunal under which it may set aside a decision not to repay input tax because it was wrongly made. The absence of words parallel to those in section 83(p) suggest that section 83 does not confer any jurisdiction upon the tribunal to set aside such a decision, and that the only route for a taxpayer aggrieved by such a decision to contest the decision is via judicial review in the High Court.
  44. Neither does it seem to us that such rights as a taxpayer has to challenge a wrong decision can be read into section 83 to give efficacy to community legislation in the same way as the Court of Appeal in IDT Cash Services Ireland Ltd v HMRC [2006] STC 1252 read words into Schedule 10A VATA to defeat the proposition that certain telephonic services should escape VAT. In that case the words were read in to give direct effect to the Sixth Directive principle of no double taxation and no non-taxation. The additional words were needed to secure the application of that provision. There is no such need to read additional words into section 83 because there is already a UK forum, the High Court, in which those principles can be enforced.
  45. In response Miss Mitrophanous says that if a decision is wrong it cannot be applied by the tribunal. But the tribunal is not applying the Commissioners' decision but hearing an appeal in relation to the amount of input tax which may be credited. If the Commissioners' decision is wrong and a nullity then perhaps the hearing and our decision may be nullities as well; but we have no jurisdiction to say that because the Commissioners' decision is a nullity the input tax should therefore be refunded: instead we can only address the question of what is the proper amount of input tax to which the taxpayer is entitled.
  46. (ii) A Kittel codicil
  47. One starts with the directly enforceable EU right which the taxpayer has to deduct attributable to input tax. That right is to an amount which, on an appeal, this tribunal may determine.
  48. Then one recognises that that right is limited by the 'Kittel derogation': if a person knew or should have known of a connection with fraud "it is for the national court to refuse" the entitlement. That limitation this tribunal can (and must) apply as well: it determines whether, on the facts before it, the words apply, and if they do, determines the right to deduct accordingly.
  49. If the terms of a Kittel codicil are to be capable of application by the tribunal they must be sufficiently certain and clear to enable the tribunal to decide whether to reinstate the right to deduct.
  50. If such terms relate to the conduct of the Respondents in challenging a particular taxpayer's right to input tax then they would involve treating the Kittel derogation, not as an objective denial of a right, but as a denial at the option of the Commissioners: in other words as if the ECJ had said "but the Commissioners may deny the input tax credit if the taxpayer knew …" in other words treating Kittel as conferring powers on the Commissioners rather than limiting a right of the taxpayer. That is plainly wrong. Thus the codicil cannot be based on that presumption.
  51. Neither can the terms of any codicil be based on the Respondents' conduct generally. They cannot say "but that right shall be reinstated if the Respondents act so as to discriminate against exporters". If they did then they would reinstate the right for all taxpayers for all periods in which the Commissioners had been acting wrongly. It cannot have been the intention of the ECJ in Kittel that the derogation was to be escaped by the most culpable taxpayer just because the Commissioners adopted a general course of action which was wrong.
  52. And any formulation which involved reinstating the right if the Commissioners' attempt to get the tribunal to refuse the right was part of a wrong course of conduct, suffers from the criticism that if one scrupulous taxpayer has applied Kittel and treated himself as not having a right to input tax (because of his knowledge), he is at a disadvantage to a taxpayer who was less meticulous and waited for the Commissioners to do so in pursuant of their wrong course of action.
  53. We see Kittel as an interpretation of the Sixth Directive to be read in the light of the general principle that objective tests relating to a particular transaction determine the VAT effects of that transaction. Thus, it limits a taxpayer's right where objectively in relation to that transaction he is an accomplice. It is not possible to read into this any proviso relating to the general conduct of the tax authority.
  54. We therefore cannot formulate any limitation on Kittel reflecting the principles Miss Mitrophanous expounds. It seems to us that those principles apply in relation to the acts of the Commissioners, not the rights of the taxpayer and may thus not be applied in the determination of the amount of input tax credit.
  55. (iii) a constraint on the tribunal's application of Kittel
  56. This formulation overlaps to a significant extent with that in (ii) and to that extent suffers from the same problem; but it also encompasses the possibility of a restraint on the actions of the tribunal akin to that placed by the EU Treaty on the actions of the Commissioners. It could be stated thus: the tribunal should not deny the right to input tax in such a way that its practice amounted to a discrimination against exporters.
  57. It seems to us, however, that where appeals are brought to the tribunal and they are determined by the tribunal in accordance with the law there could be no such discrimination.
  58. Conclusion
  59. We therefore conclude:
  60. (i) that we have no jurisdiction to declare the Commissioner's decision void or illegal;
    (ii) that insofar as it determines the right or otherwise of a taxpayer to deduct input VAT, the principle enunciated by the ECJ in Kittel is not and cannot be subject to a caveat or exemption which reinstates the right to VAT if the Commissioners have acted in such a way as to discriminate against exporters; and
    (iii) that no provision of EU law requires this tribunal, when determining an appeal under section 89(a) to apply the limitation of the right to input tax directed by the ECJ in Kittel differently according to whether or not the Commissioners have acted in such a way as to discriminate against exporters.
  61. These conclusions in particular (ii) and (iii) involve propositions of EU law. We address at the end of this decision whether it would be appropriate to refer a question to the ECJ.
  62. The Legal Issues
  63. We start with two paragraphs from the ECJ's judgment in Kittel: the first relating to the Recolta case where the national court had found that the company was innocent of any fraud ("neither knew nor had any suspicion"):
  64. "5.1 In the light of the foregoing, it is apparent that traders who take every precaution which could reasonably be required of them to ensure that their transactions are not connected with fraud, …, must be able to rely on the legality of those transactions without the risk of losing their right to deduct the input VAT …";

    the second relating to Kittel itself, where the national court had concluded that the company had knowingly participated in a VAT carousel fraud:-

    "61. By contrast [to the case where a person did not know and could not have known of fraud], where it ascertained, having regard to objective factors, that the supply is to a taxable person who knew or should have known that, by his purchase, he was participating in a transaction connected with fraudulent evasion of VAT, it is for the national court to refuse that taxable person entitlement to the right to deduct."
  65. There were four issues debated before us in relation to these passages:-
  66. (i) how "should have known" should be interpreted;
    (ii) whether the knowledge required was knowledge of some fraud or of a particular fraud;
    (iii) whether the goods which are the subject of the input tax claim transaction had to be the same goods which were the subject of the fraud; and
    (iv) the burden of proof of each element of the test.

    We consider these in turn.

    (i) should have known
  67. After some initial submissions by the parties we indicated that we were minded to adopt the following test:
  68. A person should have known that there was fraud if a reasonable trader in the Appellant's position possessed of the information it knew would have taken any additional steps and whether, having taken these steps (if any), it would have come to the conclusion that there was fraud in the chain.
    The Respondents' arguments
  69. Mr Foulkes would prefer us to apply a test which referred to a trader's positive duty to take precautions: he says that the "should have known test" may best be considered as a test for a form of recklessness in which the trader fails in that positive duty. He refers us to paragraph 51 in Kittel "traders who take every precaution which could reasonably be required", to Calltell, where at paragraph 46 the tribunal referred to a positive duty to take precautions (although in the end they based their decision on knowledge rather than means of knowledge), and to the tribunal's findings in Dragon Features Ltd v HMRC VATD 19831 to which we shall return later.
  70. It seems to us that paragraph 51 of Kittel provides a safe harbour for a trader outside the walls of which he may be sunk, but it does not go so far as to say he will be sunk or to impose a positive duty. Outside the harbour he will be sunk only if he sails into a storm of which he knew or should have known.
  71. Mr Foulkes says that a higher threshold – an interpretation that "should have known" was satisfied only if his knowledge and reasonable enquiries gave conclusive proof of fraud, or only if the trader deliberately turned a blind eye – would defeat the purpose of stopping traders becoming "accomplices" to fraud (see paragraph 57 of Kittel: a person who knows or should have known aids the perpetrators of fraud and becomes their accomplice).
  72. The Appellant's arguments
  73. Miss Mitrophanous says the test we outlined leaves room for a trader to be denied his input tax simply on the basis of negligence – the failure to take reasonable steps. That she says is wrong.
  74. She says: that there is nothing in Kittel which suggests that a failure to take a reasonable step can lead to denial of input tax deductions: that "should have known" suggests having a "basis for knowing" not "should have tried to find out more"; and that the test cannot be more stringent on the taxpayer than that in section 77A VATA in which joint and several liability for VAT may be imposed where the trader "had reasonable grounds to suspect". Mere (honest) negligence cannot be enough; there must be something approaching dishonesty: turning a blind eye or something which makes the trader an "accomplice". The ECJ judgment refers to fraud or abuse not negligence. She cites the ECJ's finding in Collée C-146/05 that the absence of good faith is not relevant unless a Member State was at risk of losing tax revenues which was not the case here.
  75. Discussion
  76. In Optigen Ltd, Fulcrum Ltd and Bond House Systems Ltd v CCE C-354/03, C355/03 and C3484/03, the ECJ said that a right to input tax could not be denied where the trader had "no knowledge and no means of knowledge" of the VAT fraud. The ECJ in Kittel drew on Optigen to formulate the conclusions in paragraphs 51 and 61 of its judgment set out above. It seems to us that "taking every precaution" is almost a restatement of "no means of knowledge", and that "should have known" must be a form of "means of knowledge" and is coloured by that phrase.
  77. The Advocate General in Optigen advanced the opinion that Member States could deny input VAT where the trader "knew or ought to have known" that VAT would go unpaid.
  78. It seems to us that in these contexts what a trader "should have known" may include what he ought to have known or had the means of knowing. Those phases indicate to us that it therefore may include what he could have found out if he had made further enquiries.
  79. In Dragon Futures the tribunal (at paragraph 73) said that the test must be applied within the principle of certainty, proportionality and neutrality, and said that the trader "must make a proportionate response to information" and "take proportionate steps to increase knowledge" (paragraph 74). This formulation appears to impose a duty, but in our view the tribunal was applying a test: what would be a proportionate response? What a trader could have found out is to be limited by what is proportionate.
  80. It does not seem to us that the s.77A test is relevant to the application of the EU wide formulation of Kittel. And we see nothing in Kittel which points to a requirement that the trader be in some way dishonest before he can be said to ought to have known: the sense of paragraphs 56 and 57 is that the relevant knowledge effectively makes the taxpayers accomplices, not that they must be dishonest accomplices before it can be said that they have the relevant knowledge.
  81. Since the hearing the tribunal has published its decision in Olympia Technology Ltd VATD [2057D]. There the tribunal said this (at paragraph 16):-
  82. "We consider that on its ordinary wording "ought to have known" is a factual test comprising two limbs. First, one should start with all the facts (a) actually known to the person and ask whether in the light of those facts a reasonable businessman would have known that the transaction in question was connected with fraud. Secondly, it would include (b) those facts that would have been known to the person if he had taken some action to discover them that the reasonable businessman would have taken in the circumstances (which is not necessarily the same as every precaution reasonably required), but which the person did not."

    We are grateful for this formulation and respectfully agree with the tribunal's reasons for it. We would add, as the tribunal suggested in Dragon Features (at paragraph 75), that the knowledge of the fraud has to be on the balance of probabilities, not beyond reasonable doubt.

  83. As the tribunal in Olympia acknowledged, the test they adopt leaves open a further question. That is whether the reasonable businessman is either (a) one having the skill and experience of the taxpayer, or (b) one having the general knowledge skill and experience that may reasonably be expected of a person carrying out the taxpayer's functions.
  84. In Olympia the tribunal heard no argument on this question and on balance preferred the test in (a). We had the advantage of Mr Foulkes' and Miss Mitrophanous' views on the subject. Mr Foulkes stressed the objective nature of the Kittel test: having regard to objective factors pointed to the objectively reasonable trader. Miss Mitrophanous favoured an approach which looked at the actual skill and knowledge of the trader. The aim of the test she said was to deter traders from becoming involved in fraud: a failure to take into account their personal characteristics would be a punishment or a penalty. There is no reason to punish a stupid trader merely because he is stupid.
  85. We agree with Miss Mitrophanous. The prevention of fraud is not compromised by addressing actual skill and knowledge. The objectivity required is satisfied by ignoring what the trader actually thought but considering what he was aware of and his actual knowledge and skill. We adopt the same approach as the tribunal in Olympia.
  86. (ii) knowledge of a particular fraud, or of fraud generally
  87. Miss Mitrophanous says that the question is whether the Appellant had knowledge of the fraud by the defaulting trader; it was not sufficient that Appellant had grounds for suspecting fraud X or some fraud Y where there is no evidence of fraud X or fraud Y.
  88. Mr Foulkes says that this is setting the test too high. A trader who knowingly participates in an MITC fraud may only be told from whom he is to buy and to whom he is to sell: he knows he is engaging in fraud but may have no idea of the identity of the defaulter. Kittel requires knowledge of participation in a transaction connected to fraudulent evasion of VAT, not knowledge of the precise fraud.
  89. It seems to us that the knowledge (or means of knowledge) required is knowledge that there is some VAT fraud connected with the transactions which in some way those transactions assist, rather than knowledge of a particular fraud. We reach that conclusion for the following reasons:-
  90. (i) the language of Court's judgment in Kittel "participating in a transaction connected to fraud" suggests a more loosely framed test than participating in a transaction "connected to the fraud of the defaulting trader". We note in this regard the difference between the Recolta formulation in paragraph 52, "fraud committed by the seller" and the wider Kittel formulation in paragraph 51: "fraudulent evasion of VAT". In Kittel the Court was dealing with a trader at one remove only from the defaulter but it chose to frame its decision more widely, and consistently with the object of defeating fraud. The decision is not based on fairness to the taxpayer but on objectivity and preventing fraud; and
    (ii) Just Fabulous involved "contra trading". The broker such as that described at the start of this decision who, in a simple chain, would make a VAT reclaim also imported other goods and sold them to a second exporter who exported them. The broker's input and output VAT broadly cancelled out and the second exporter made a claim for repayment of VAT in respect of the export of goods in respect of which there had been no failure to account for VAT. The issue was whether HMRC could refuse the second exporter's claim (on the assumption that he had relevant knowledge) even though "the VAT claimed relates to different goods from those in the defaulter chain" (paragraph 13).

    Burton J accepted the Commissioners' argument that the words used by the ECJ "are untrammelled by any reference to the need for establishing that the taxable person must be a member of the defaulter chain, or that he must have been dealing in the same goods as had been the subject of the defaulter chain." (paragraph 50 and 43). It seems to us that Burton J's acceptance of a wide interpretation of the ECJ's words suggest that the same approach should be applied to the question of knowledge: the words are untrammelled by any reference to knowledge of a particular fraud.

  91. Thus, in our view, if a trader knows or should know that his transaction is connected to some fraud even though he does not know or have means of knowing the details of that fraud, the test is satisfied …
  92. (iii) whether the goods exported had to be the same as those involved in the fraud
  93. It seems to us that we are bound by Burton J's acceptance of HMRC's arguments in Just Fabulous described above: what is required is that there is knowledge of a connection to fraud not that the goods being sold were involved in fraud.
  94. Thus if A sells to B, who sells to C, who sells to Exporter, 4,000 Nokia phones, and Exporter knows that A is being fraudulent in relation to its sale, it matters not that B may also have bought 4,000 Nokia phones from X and it is those phones which are passed to C while the phones bought from A go to Z. If Z also knew of A's fraud then it might be that Z's input that might also be irrecoverable. Burton J acknowledged difficulties of that type in Just Fabulous but was not deterred by them from concluding that input tax could be denied to a trader who knew of the fraud: some solution to the double recovery would be found by the tribunal or the Court when the issue arose.
  95. (iv) the burden of proof
  96. It as common ground that the burden of proving that the Appellant's transactions were connected with fraud lay with the Respondents.
  97. As regards the burden of proof in relation to knowledge or means of knowledge the Respondents accept the conclusion of the tribunal in Calltell (at paragraph 70) that it was incumbent on the Commissioners to raise a sufficient case to demand an answer, but say that once an answerable case has been raised the burden falls on the Appellant to show it did not have the requisite knowledge or means of knowledge.
  98. Miss Mitrophanous argues that the Appellant should bear the full burden of proof: she says:
  99. (i) it is a general principle of Community law that those seeking to rely on a derogation justify it – and the Commissioners are seeking to rely on a derogation from the right to deduct;
    (ii) it is extremely difficult to prove a negative; and
    (iii) that HMRC accept in relation to s.77A VATA that they bear the burden of showing that a trader knew or had reasonable grounds to suspect VAT default; the Commissioners need to explain why they adopt a different approach in relation to Kittel.
  100. It seems to us that the burden of justifying the applicability of a derogation may lie with the party seeking it so far as concern the determination of the legal criteria relevant to the application, but in the determination of facts relevant to those criteria there is no indication in EU law that the normal rules would not apply. And it seems to us irrelevant to our determination whether the Commissioners have taken a different stance in relation to other legislation.
  101. In Joseph Constantine Steamship Hire Ltd v Imperial Smelting Corporation Ltd a case not cited to the tribunal, Lord Russell said that "the proving of a negative, a task always difficult and often impossible, would be a most exceptional burden to impose on a litigant." On the other hand the general rule in tax cases is that the onus of proof is on the Appellant since generally he is the one in possession of the relevant information. It seems to us that this issue is most fairly addressed by requiring that the Respondents raise a sufficient case suggesting knowledge or means of knowledge which demands an answer. Without that the appellant is in the almost impossible position noted by Lord Russell, but once such a case has been raised then there is a burden which falls upon the appellant to provide evidence to dispute it.
  102. In any event deciding a case purely on the burden of proof is not a course which should be adopted unless owing to the unsatisfactory nature of the evidence that is the only just course to take (see Chadwick LJ in Wood v Holden [2006] STC 443 at paragraph [33]). We have been able to decide the issues without the need to fall back on the question of burden of proof so the issue does not arise.
  103. It was accepted on both sides that the standard of proof was the normal civil standard which is one which takes into account the need for cogent evidence of fraud.
  104. Finally we note that, in relation to whether there was fraud, the question is whether the Appellant had the requisite knowledge at the time of its transaction that, at that time, fraud was involved. If the potential fraudster did not at that time intend to defraud then the test is not satisfied.
  105. Miss Mitrophanous raised a question which we deal with at the end of this decision as to whether, if fraud and knowledge of connection to fraud is found to be established, the Appellant should be deprived of all its input VAT or only that corresponding to the tax defaulted on.
  106. Background Facts and the Deal chains
  107. Background Facts
  108. Honeyfone Ltd was incorporated on 10 January 2001. Its directors are Mr Shiralizadeh and his father Mr Asghar Shiralizadeh. Its shareholders are Mr Shiralizadeh's family (himself, his mother, his father and his siblings) and Mr Arora.
  109. Mr Asghar Shiralizadeh had ran a manufacturing business when Mr Shiralizadeh was at university and had involved his son in the running of that business. After university Mr Shiralizadeh set up a company, Moderntronics Ltd, through which he traded by sourcing electrical equipment from the UK and Europe and selling it in the Middle East.
  110. Through these activities Mr Shiralizadeh met Mr Arora who had some 15 years experience in the telephone industry. They decided to join forces and set up Honeyfone to trade in mobile phones. Mr Arora was given 30% of the shares in the company but was not a director.
  111. Initially Honeyfone operated through retail outlets as well as in wholesale distribution. The retail side was not particularly profitable however and the retail outlets ceased operation in 2004. Thereafter its business consisted in buying relatively large quantities of mobile phones and selling them to others at a profit.
  112. It conducted this business mainly in such a way that it did not hold stocks of mobile phones: Mr Shiralizadeh would phone their contacts to discover who wished to acquire mobile phones and who had mobile phones to sell. Honeyfone would negotiate prices and match quantities and types of phones so that it sold on to its customers all the phones it acquired and did so at a profit. The acquisition and sale would usually be arranged so that they took place virtually simultaneously. So, generally, there was never a significant period when Honeyfone was possessed of a stock of phones it had bought but not sold.
  113. Honeyfone sold both to UK customers and exported to non-UK customers. When it exported it would treat itself on its monthly VAT return as entitled to the repayment of its input VAT on the phones purchased. This in many periods resulted in Honeyfone's VAT return being a claim for repayment of VAT. Until the periods 03/06 and 04/06 the Commissioners, after some inspection of Honeyfone's records, made payment of the VAT claimed.
  114. But for 03/06 and 04/06 they refused, denying Honeyfone the input VAT relating to 29 export sales in those two months. These 29 sales were not Honeyfone's only sales in those months but were its only export sales. It is against that denial, formally communicated in a letter from the Respondents of 31 January 2007, that the Appellant appeals.
  115. The deal chains
  116. The Respondents alleged that each of the 29 export transactions could be traced back along a chain to a defaulting trader – a trader who had imported the goods and fraudulently failed to pay VAT relating to their sale; and that in each chain there were a number of buffers through which phones were bought and sold.
  117. The table set out in the Annex to this decision sets out particulars of each of those chains showing the Honeyfone invoice number and date of sale, the phones sold and purchased ('N' = Nokia, and 'S' = Samsung), the defaulting trader, the number of buffers between it and Honeyfone, Honeyfone's customer and the freight forwarder involved. It also shows, on the information to the tribunal, how many of the transactions in each chain were settled through First Curaço International Bank (FCIB).
  118. The Respondents obtained the information to compile these deal chains in the following way. Pauls Freight Services Ltd is a freight forwarder whose business it is to accept goods imported into the UK on behalf of its clients, to hold goods for its clients and to despatch goods by export or otherwise on its clients' instructions. Honeyfone uses Pauls Freight for the phones in which it deals, as do its suppliers. If A, a client of Pauls Freight sells goods which A is holding at Pauls Freight to B, A will send a 'release note' to Pauls Freight indicating that the goods should now be held for B. If goods arrive at Pauls Freight for A and are released by A to B, B to C, and C to D, then Pauls Freight keep together in an envelope the release notes relating to each of those transfers. They also keep in the same envelope the 'CMR' or import documentation or a copy thereof relating to the import of the goods by A (if they have been imported). If D exports the goods acquired from C, then Pauls Freight will also keep CMRs for the export referenced to the pack containing the sets of release notes. The release notes identify the number and type of the phones being released and to whom they are to be released. They do not identify distinguishing numbers of individual phones.
  119. The Respondents, having identified Honeyfone exports they wish to trace back, go to Pauls Freight and ask for the pack relating to the export. They are given it and then seek from A, B and C the invoices and other paperwork relating to the sales which gave rise to the relevant release notes. Thus they were able to present to the tribunal copies of release notes, sales invoices, purchase orders and cash movements relating to the sales of the phones from A to B, B to C and C to D and CMRs which related to import by A and export by D.
  120. There are some features of that evidence on which we should comment:-
  121. (i) For each of the deal chains there were release notes for the specified quantity of the specified phones from the 'defaulting trader' through each buffer to Honeyfone and on to its customer;
    (ii) for each deal chain with the exception of chain 24 – see (v) below - there were corresponding invoices (and some additional documentation) for the sales from the first Buffer onwards, but in only 10 out of the 28 chains were there copy invoices in relation to the transfers from "Defaulting Trader" to the first buffer;
    (iii) neither the invoices nor the release notes nor any other documentation contains any specific identification of the individual phones being sold (e.g. by serial number) : all that is clear is that there is a sale or transfer of e.g. 4000 Nokia 8800;
    (iv) the import documentation – the CMRs – retained by Pauls Freight does not refer to the specific nature of the goods, rather it refers e.g. to "4 Pallets Electr Equipt" and gives a weight e.g. "1060kg". The link to the goods transferred is established by relying on Pauls Freight's record keeping: that it maintains the CMRs (or copies thereof) relevant to the chains in the same envelopes as the release notes or referenced to them. It is possible that one CMR may also relate to goods other than those the subject of the release note chain. The combined weights of the import CMRs within any pack may differ from the combined weight of the related export CMRs;
    Miss Mitrophanous notes the following oddity in the Paul's Freight documentation provided to the tribunal. The documentation includes copies of CMRs evidencing the import of phones into the UK and copies of CMRs evidencing Honeyfone's export of the phones. But whilst the export CMRs detail the number and type of the phones exported the import CMRs merely indicate something like "4 pallets Electr.Goods", and the combined recorded weights of the import CMRs said to relate to a particular deal differs from the weight recorded for the phones exported. The following table exhibits those differences for deals (1) to (10):
    Deal Chain Total weight of exhibited Total weight of exhibited
    Number import CMRs (Kg) export CMRs (Kg)
    (1) 4160 3263
    (2) 1050 1047
    (3) 3250 2309
    (4) 2097 1515
    (5) 3051 2216
    (6) 2150 1513
    (7) 3300 1471
    (8) 2100 1349
    (9) 3200 1848
    (10) 4383 2369
    And generally the weights on the import CMRs were in round tens but the export CMRs more exact.
    All this called into question the accuracy of the audit trail documents maintained by Pauls Freight. It was suggested by Mr Rodney that the import CMRs might reflect larger and mixed consignments broken up by Pauls Freight in accordance with instructions received or even imports for more than one recipient.
    We do not find that it cast any doubt on the establishment of the deal chains through Pauls Freight's documentation: the chains of release notes did not record weights and were easy to follow; they also led accurately to the export CMRs. But it does call into question whether the goods which were exported were imported by the "defaulting trader" rather than by someone else, and whether the import CMRs can be relied upon to deduce that each consignment originated from the Prologis warehouse in France (that being the place of origin recorded on those CMRs).
    On balance however we conclude that the import CMRs are evidence of import from Prologis' warehouse by the defaulting trader. That is because: we can see no reason why Pauls Freight would have maintained the CMRs with the release notes (and in some circumstances appended manuscript numbers to the release notes referring to manuscript numbers on the CMRs) unless they were linked; we accept that potentially a delivery might need to be split; and the difference between the round sum weight on the import CMRs and the more exact weight on the export CMRs may well be explainable by the procedures of the different parties (Prologis and Pauls Freight) who produced them.
    (v) in relation to deal chain 24 there was an inconsistency between the invoice documentation and the release notes:
    Invoices release notes
    Computec Computec
    First Associates Anderson Cellular
    Global Roaming MP3 Com
    Team Mobile Team Mobile
    Honeyfone Honeyfone
    This discrepancy is difficult to understand or explain. Oddly the invoices and release notes all refer to the same number of the same type of phones, and each set of documentation clearly showed the transfer or sale, as the case may be, from one company to the next in the chain.
  122. The discrepancy in relation to chain 24 did not cause us seriously to doubt the evidence in relation to the other chains, but was sufficient for us to conclude that it was not more likely than not that Honeyfone's transaction in this chain traced back to a defaulting trader. We therefore find that it is more likely than not that in relation to each of the deal chains in the table other than chain deals that (a) there were sales of the relevant number of the specified type of phone from the "defaulting trader" (we make findings later as to whether the trader did default) to the first buffer and onward by sales to the other buffers and thence to Honeyfone, and by Honeyfone to its specified customer; (b) that the sales in each chain all took place on the date of Honeyfone's invoice; and (c) that the defaulting trader had imported the phones.
  123. The lack of information on the invoices and release notes relating to the particular specific identification of each phone sold permits the possibility that if, for example, 4000 phones were sold by buffers, A, B and C: by A to B, and then by B to C, that B had additionally acquired 4000 phones from X and that those were the phones sold to C rather than those acquired from A. The method by which the Respondents collected their documentation meant that this possibility was not addressed. We considered at paragraphs 57 and 58 whether, if C's phones were those acquired from X, this would affect whether or not the Kittel principle applied but we conclude on the evidence before us (the evidence of the contrivance of the chains, the reliance of the parties on Pauls Freight, and the short time for which the goods were in the country) that it is more likely than not that the phones described in the deal chains in the table as sold by defaulting trader were precisely those phones sold by Honeyfone.
  124. Orchestration or contrivance in the chains
  125. In looking at, and the evidence surrounding, the deal chains we were struck by several features:-
  126. (i) the fact that all the contracts in a chain took place from import to export on the same date;
    (ii) the large number of mobile phones exported by Honeyfone: 30,000 odd in March and 35,000 odd in April;
    (iii) the large amounts of money involved: a single deal would be for some £1m;
    (iv) that there were three buffer companies in 20 chains and two buffers in eight chains. The same buffer companies turned up again and again in the chains-normally in the same place but with occasional permutations;
    (v) the small margins made by each of the buffer companies – generally about 25p to 50p per phone on each sale; and, by contrast, the larger margin (about ten times as much: in the order of £5 or so per phone) made by the exporter, Honeyfone. It was as if the exporter was being paid extra for taking the VAT risk. But the amount of profit at each stage was a fairly fixed tariff;
    (vi) the fact that in March each chain led back to C&E Enterprises, and in April, 13 out of 14 chains led back to Computec. It was as if someone had said "this month's importer shall be X";
    (vii) that all but one of "this month importer's" purchases had been from the same company, European Communications;
    (viii) the fact that some companies appeared in one chain as Honeyfone's supplier, and in other chains as that supplier's supplier:
    Acting as Acting as Honeyfone's
    Supplier Name Honeyfone's supplier supplier's supplier
    Zain 8
    MP3 Com 2 10
    Global Roaming 6 7
    Team Mobile 11
    We find it odd that if there is a free market where each trader deals on the phone with others and is in frequent contact, that this would happen
  127. We find these features striking and wholly compelling evidence that the chains were orchestrated. We were left with no reasonable doubt that some person or persons had decided how the phones should move into and through the UK and arranged or indicated the appropriate profits each participant should make. We can see no reason other than the promulgation of a VAT fraud for such orchestration.
  128. But we noted nothing in the documentary evidence for the chains which indicated expressly that Honeyfone knew of the composition of the chains or their details.
  129. Payments
  130. The documentation also contained evidence of some payments being made. That documentation provided evidence of payments made through First Curaço International Bank (FCIB). There was no evidence of the use of any other bank. The last column in the table shows the number of steps in the chain for which there was evidence before us of such payment. Payments for 31 March invoices appeared to have been made in the early days of April (3 or 4 April but sometimes later). Part payment was made in many cases followed by the balance. A similar practice prevailed for April sales. There was however no evidence that any payment on any deal chain had been made to C&E Enterprises.
  131. The Existence of Fraud: Were the "defaulting traders" in default and were they acting fraudulently at the relevant time?
  132. C&E Enterprises Ltd
  133. Mr Rodney told us that:
  134. (i) on 31 March 2006, following a discovery that C&E Enterprises had imported significant numbers of mobile phones and sold them on to UK companies, officers of the Respondents visited the address which appears on that company's invoice and release notes to deliver a notice requiring a VAT return to be made up to 31 March 2006 and returned to the Respondents by 3 April 2006;
    (ii) when the officers arrived at that address they found it to be a residential address. No one was contactable. They delivered the notice;
    (iii) on 3 April the officers returned to collect the VAT return. No one was contactable;
    (iv) no one from C&E Enterprises has contacted HMRC since then;
    (v) between 12 April 2006 and 10 May 2007, ten assessments were made on C&E Enterprises. None have been paid, disputed, or acknowledged.
    (vi) the aggregate amount assessed and unpaid amounts to some £74m.
    (vii) in July 2006 HMRC instigated civil recovery proceedings against C&E Enterprises.
  135. Save in one respect there was no real challenge to this evidence. That challenge was in relation to a form which an officer of HMRC was required to sign to instigate civil recovery procedure against C&E Enterprises. That form appeared to have been completed on 17 July 2006 (but there was no evidence of signature) and included confirmations that "there has been contact with the trader in the last 3 months", and "there are no indications that the trader is missing". Whereas the first of these might be taken to refer to letters sent by HMRC, the latter casts a shadow on whether C&E Enterprises were truly not responding.
  136. On balance we conclude that the civil recovery form was carelessly completed and that the absence of a signature is telling. We accept Mr Rodney's evidence and conclude that C&E Enterprises did default on its assessed VAT liabilities in respect of the 03/06 period and has not been in contact with HMRC since March 2006.
  137. From the reports appended to Mr Rodney's evidence we find that for the preceding three quarters it had rendered nil returns. The majority (11 out of the 14 chains beginning with a C& E Enterprises transaction in 03/06) were transacted on 31 March 2006, the very day HMRC's officers went to its address and delivered their notice. The failure to deal with that notice or to contact HMRC in relation to it suggests to us that on that day at the very least C&E Enterprises did not intend to account for the VAT on those sales. The very large, sudden, VAT liability and the residential address point in the same direction, but the shadow cast by the civil recovery form remains.
  138. It would have been helpful to have had evidence of what other (pre-March) correspondence there had been with C&E Enterprises, how the documents which the Commissioners' correspondence indicates were received from the company and which triggered the visit were received, and whether any further visits had been made to the house. The lack of such evidence is troublesome, and what evidence we heard was second hand.
  139. This evidence was by no means enough to find fraud beyond a high measure of reasonable doubt, but it was just enough for us to conclude on balance that C&E Enterprises intended to evade the payment of its VAT in relation to the transactions in the deal chains.
  140. Computec Solutions Ltd
  141. We find that after 2 years in which Computec made nil returns, HMRC became concerned in May 2006 that it had made very significant supplies in April 2006. A visit was made on 5 May 2006 with a notice requiring a VAT return made up to 5 May 2006 to be delivered by 8 May. The officer went to its address and spoke to a receptionist. He was told the trader had a presence at the office but was not in at present. The notice was left at the address. No reply or communication has since been received from the company. Assessments totalling more than £100m have been made on the company since 5 May 2006. None have been paid.
  142. In October 2006 the directors of Computec appear to have instructed an insolvency practitioner to convene a meeting of creditors, but later that month the Insolvency Practitioner wrote to creditors (including HMRC) indicating that the resolution to wind up was not passed 'because the director failed to attend the meetings."
  143. Miss Mitrophanous says that the actions of directors who instruct an insolvency practitioner to put the company into liquidation are not the actions of people whose original (April) intention was not to pay VAT. The insolvency practitioner's letter had even indicated that his firm had "been authorised fees in advance by the Directors in the sum of £2,500 plus VAT". Neither, she says, does the presence of a receptionist at the premises where the company have a presence betoken a missing trader. She says that there is no indication in the evidence before the tribunal that the officers did not return and meet the trader.
  144. We find that Computec did intend to evade its VAT liability in relation to its April sales and formed that intention at the time of the relevant sales. The presence and the receptionist, and the instruction of the insolvency practitioner suggest that the individuals were not missing, but it is not clear that the funds for the practitioner's fees were authorised to be paid by the company or the directors personally. What tips the balance for us is that after six quarters of nil returns, the company made, in March and April, supplies generating at least £20m of VAT liability and took no steps to return any of that liability to, or otherwise to contact, HMRC. That does not look like the behaviour of a trader who gets into cashflow difficulties and holds his head in the sand, but like the actions of someone who never intended to account for the VAT on a sudden massive accumulation of sales and who had lain dormant for several periods ready to commit a fraud. And that would be our conclusion even if it were clear that the individuals concerned had not 'gone missing'.
  145. Worldwide Enterprises Ltd
  146. Mr Rodney's evidence was based on a report produced by another officer in January 2008 and a report of a visit on 25 April 2006. He says that this company had been deregistered and that it was then established that supplies were being made and invoices raised using the deregistered number. He says that mail was being returned from the principal place of business and a visit by officers to the directors' home addresses found they had been demolished.
  147. It is difficult to draw clear conclusions from the January 2008 report:-
  148. (i) deregistration. The report indicates that deregistration was requested on 19 April 2006. But the reports indicate that a visit was arranged 6 days after that because there was evidence that the taxpayer was acquiring goods: that suggests to us the carrying on of a business. On 15 May a request was made to defer deregistration because other supplies (presumably those listed in the report as made on 12, 13, 18, 19 and 20 April) had been made. It is unclear whether at the time of the 19 April request the Commissioners knew of the April supplies, but by 15 May it appears that they did. The Commissioners may deregister a person only if they are satisfied he has ceased to make supplies or has ceased to be registrable because the value of his supplies falls below the relevant thresholds. There is no indication that they were so satisfied – and given the circumstances of the 25 April visit, the balance of evidence is that they could not have been so satisfied. The report indicated that the taxpayer was deregistered on 18 May 2006 but there is no indication in the report that this was communicated to the taxpayer. If deregistration was on 18 May rather than 19 April then knowledge of the taxpayer's activities seems even more likely to have been in the Commissioners' hands and valid deregistration even less likely. Even if the taxpayer was correctly deregistered from 19 April 2006 there was no evidence before us that the taxpayer was or could have been aware of that. As a result we cannot find that its mere use of its VAT registration number on invoices made after that date was fraudulent in any way.
    (ii) addresses. The report indicates a succession of addresses: the first was 4 Randal Street; Mail was returned from that address in July 2005 but on 25 August the taxpayer wrote giving a new address 546 Chorley Old Road. The 02/06 return is said to have been 'returned' (presumably gone away/not there) from that address but the report then indicates a letter was sent to the trader asking for confirmation of its place of business. Where the letter was sent is not clear but apparently the trader replied giving the original 4 Randal Street address.
    The report indicates that a letter warning of a VAT debt of £36,925,961.30 was sent to both addresses on 15 June 2006. Further letters were sent to 546 Chorley Old Road one of which was returned with an indication that the company had left the premises in 2004.
    Mr Rodney exhibits a letter warning that proceedings would be taken if the £37m debt was not paid which was sent to 4 Randal Street on 15 June 2006. In his witness statement Mr Rodney says that this letter was returned to HMRC by the current residents indicating that the company had not been at that address for 3 years. This appears to conflict with the report for, as noted above, there it is said that on 15 June 2006 a warning letter sent to 546 Chorley Road was returned with the same message. We suspect that Mr Rodney is muddled and that he had intended to refer to 546 Chorley Old Road rather than 4 Randal Street.
    On 25 April officers visited the home addresses of company officials. They were found to have been demolished.
    (iii) On 19 July a letter was written to the company indicating that a debt arose to the Crown under Schedule 11 paragraph 5 VATA of £36,591,462 because it had been issuing VAT invoices when it was not entitled to (presumably because its VAT number had been cancelled). But the calculations exhibited to that letter relate to invoices said to be dated April 2006 – a date before the report indicates that deregistration took place. If that is right then the £36,591,462 was not properly sought under Schedule 11 paragraph 5.
    The report also notes that returns had been submitted for the 05/05, 08/05 and 11/05 periods (with a liability, a repayment and a nil result respectively). The debt claimed of £37m or so was based on VAT input tax claimed by customers.
    Discussion : Worldwide
  149. This evidence is not very much on which to base a conclusion that Worldwide intended not to pay VAT on sales made in April 2006: it is not clear that an assessment was made for 2006; it is not clear whether the relevant letters were sent to the right addresses; the demolition of the residential addresses of the directors does not appear to bear upon the existence of the business addresses. The very most that can be said is that it appears that letters were sent on 15 June 2007 to both addresses warning of proceedings for a debt of about £37m those letters were not answered, and no subsequent communication has been had from the company.
  150. On that evidence we do not find that there was a fraudulent intent in April 2006 not to pay VAT on sales forming part of the April chains.
  151. Other Pointers
  152. Mr Foulkes says that evidence of the confirmed nature of the deal chains suggest fraud. We agree that for reasons already indicated and to be drawn together below, that the deal chains are suspicious, but that is not enough in our view to make it more likely that Worldwide was fraudulent.
  153. Miss Mitrophanous says that there are common elements of fraud which are missing from the facts: there is no evidence of the diversion of payment to third parties, there is no evidence that the defaulting traders cut their prices, or even that the price paid by any of the traders was transparently low. These may be indications of fraud but their absence does not seem to us to make it significantly less likely that C&E Enterprises or Computec were fraudulent. The evidence of their default and the surrounding circumstances points to fraud sufficiently clearly not to need bolstering by other factors.
  154. We conclude that:-
  155. (1) C&E Enterprises Ltd were fraudulent in relation to their March 2006 VAT, and tax for that period has been lost.
    (2) Computec was fraudulent in relation to its April 2006 VAT, and tax for that period has been lost.
    (3) We are not persuaded that Worldwide Enterprises was fraudulent in relation to its April 2006 VAT.
  156. Knew or Should Have Known
  157. Although Mr Shiralizadeh was the only active director of Honeyfone it was clear to us that he ran the business with Mr Arora. In considering the knowledge of Honeyfone we have treated it as having the knowledge and skills of those who acted on its behalf namely Mr Shiralizadeh and Mr Arora.
  158. In relation to this aspect we find the following facts:-
  159. (1) In March and April 2006 the Appellant was well aware that there was VAT fraud in the mobile phone market, and that such fraud could involve the non payment of VAT by a missing trader. The Appellant was not aware of the extent of the fraud.
    (2) Prior to March 2006 the Appellant was aware of the contents of HMRC's Notice 726 on Joint and Several Liability under section 77A VATA. That notice contained advice on checks a trader might undertake "to establish the legitimacy of your supplier to avoid being caught up in a supply chain where VAT would go unpaid." (paragraph 4.5).
    (3) The Appellant was aware that, since mobile phones were not manufactured in the UK, the phones which were in the UK which it bought from other UK traders would have been imported by an earlier acquirer and that there was a chain of supply from that supplier to it.
    (4) The Appellant was aware of the progress of the Bondhouse case to the ECJ and understood that in the autumn of 2005 or early 2006 there had been a victory. It was aware that Bondhouse had resulted in or involved the denial of input tax. After January 2006 that cloud lifted and the Appellant came to the view that "we had clear guidelines: as long as we made sure that our suppliers, our customers, are legitimate businesses, then we should not really worry about what is happening behind or after the chains." Prior to January 2006 the Appellant's concerns about the possibility of a Bondhouse attack had contributed to its limiting the level of its business. In our view such concerns also limited the business Honeyfone was offered: more deals or more valuable deals were available thereafter.
    (5) In 2001/2 Honeyfone's monthly turnover was about £1.6m; in 2003/4 £7m and 2004/5 £17m. In March 2006 it was over £12m, and in April 2006 about £17m. Mr Shiralizadeh said that he regarded himself as being able at this time to take his business "to the next level" because the market was active and people in the market were trusting him. We accept that Mr Shiralizadeh and Honeyfone had a proper commercial motive in wanting to make more profit. We do not accept that the reason for Honeyfone's increased turnover was only the result of more hard work by Mr Shiralizadeh and Mr Arora: we believe it also arose because the market became more active and because Honeyfone became less concerned.
    (6) Each of the deal chains involved goods being shipped from the Prologic warehouse in France to Pauls Freight in the UK where they would be released by consecutive release notes along the chain to Honeyfone and then, on the authority of Honeyfone's release note, released to Honeyfone's purchaser and shipped back to Prologic. Mr Shiralizadeh told us (and we accept) that "Pauls Freight [was] in a position to know where the goods have actually been imported from, and who the chains – who the companies in the chains are." We conclude that the Appellant was aware that Pauls Freight would know the identity of each party in the chain from import to export. We also concluded from Mr Shiralizadeh's evidence that he was aware that there would be other traders in the chain who would be receiving and releasing via Pauls Freight. We did not conclude from Mr Shiralizadeh's evidence that he was told by Pauls Freight that the phones Honeyfone was exporting to Prologic had come from there.
    (7) All of Honeyfone's export deals in March and April were concluded at or very close to the end of the month. From Honeyfone's perspective there was a cashflow advantage in concluding the transactions then because it meant that there was less time between becoming obliged to pay the VAT inclusive price to their supplier and the time it got its VAT back. Its customer would pay only the net amount and it would be out of pocket by the amount of the VAT (less its profit) until repayment (save to the extent that it did not always make immediate payment in full). But among those with whom Honeyfone dealt and knew of there was also a general trend to do deals at the end of the month. The momentum built up at the end of the month; deals might be discussed early in the month, might begin to solidify as the month went on, and would then set in the last few days. Mr Shiralizadeh said that his suppliers "know that I am an export customer so they should just leave Honeyfone to the latter part of the month." Mr Shiralizadeh offered us a further explanation for his suppliers' concentration on the last days of the month: he suggested that they became more anxious to meet their monthly budgets or targets towards the end of the month and did not wish to hold stock over the month end; as a result they became more anxious to trade and the pricing became keener and dealings were done. We do not think that likely: Honeyfone's suppliers were small companies unlikely to have been driven by monthly targets; if they held stock they would be keen to sell it sooner rather than later and selling earlier in the month would be more advantageous than selling at the month end. If there were larger suppliers in the UK they would have been motivated by the cashflow advantages of early sales too. We conclude that the drive to deal at the month end came only from the exporters like Honeyfone and their VAT recovery timing.
    (8) The increase in Honeyfone's turnover in March and April 2006 required additional sources of funding. Generally it paid its suppliers on or very soon after the date of sale and received payment from its customer at about the same time (see (10) below). Because its non-UK customers paid without VAT Honeyfone had to fund the difference between its VAT inclusive purchase price and its VAT exclusive sale price until it got its VAT repayment. That called for additional funding of some £2m in April and £2.8m in May 2006. A small part of this was met by retaining profits which would otherwise have been dividend to shareholders, the rest by an extension of credit to it from its suppliers. It currently remains owing some £2m-£3m to Team Mobile, one of its suppliers.
    (9) In March and April all Honeyfone's export deals were back to back: it agreed contracts for the sale and purchase of a number of phones at the same time.
    (10) In general Honeyfone was paid by its export customers on or a few days after the day of sale. On receipt of this payment it would make a payment generally of the same amount to its supplier as part payment of its purchase price (since its purchase price included VAT it was higher than its sale price). It would them make balancing payments as and when funds became available – e.g. through VAT reclaim payments.
    (11) From 2001 to November 2005 Honeyfone's main banking arrangements were with Lloyds Bank plc. But in November 2005 Lloyds wrote to Honeyfone giving a month's notice that they were withdrawing their services and would not provide references. Mr Shiralizadeh said that Lloyds provided no reasons for their decision but that he became aware that other people in the telecoms section had had their banking facilities withdrawn by Lloyds both before and after Honeyfone's were withdrawn. Honeyfone approached Barclays who said they were no longer able to open accounts for the telecoms sector. Honeyfone were able to open bank accounts with other banks, but, with the exception of FCIB (and possibly Habib Allied International Bank) these banks did not offer 24 hour transfer facilities. On the evidence before us Honeyfone received and paid through FCIB on most of the deals in the deal chains and other parties in the chain also used FCIB.
    (12) Honeyfone did not insure the goods it bought and sold in the back to back export transactions. For March those goods had a total value of some £12m, and for April some £17m. Honeyfone maintained £10,260 of insurance for stock in hand and £10,000 for goods in transit. Mr Shiralizadeh told us that Honeyfone had taken a risk – a calculated risk not to insure: he accepted that if one consignment of phones was lost or damaged that he could be put out of business but (i) given that the phones were held and dispatched by Pauls Freight he felt he would have some claim on Pauls Freight if there were a loss, (ii) insurance in the mobile phone industry had become difficult and complex to arrange, and (iii) premiums were high and would have absorbed up to 25% of Honeyfone's profit on a deal. He therefore decided to take the risk. We conclude that the cost of the premiums, Honeyfone's transient interest in the phones, some expectation that loss was unlikely at Pauls Freight and hope that no loss would arise in transit to their customer convinced Honeyfone to take a risk. Taking such a risk is, to our minds, indicative of a less cautious general approach to business and is consistent with a less cautious approach to the evaluation of whether or not its transactions were connected with VAT fraud than that which a reasonable businessman might exhibit.
    (13) Each mobile phone has an International Mobile Equipment identity number (an IMEI number). The recording of the IMEI numbers of a consignment of phones can assist negotiation with a customer or a supplier if a phone is alleged to be faulty or damaged, and can assist a trader to determine whether the phones he is dealing in have passed through his hands before: if they have it may suggest that they are involved in some carousel or fraud.
    Honeyfone did not record the IMEI numbers of the phones it dealt in. Each phone was packed in a box, between 5 and 10 boxed phones were packed in a larger box, and up to 200 larger boxes were put on a pallet. The whole was shrink wrapped. The determination of IMEI numbers would have been awkward and time consuming.
    (14) Honeyfone did not undertake its own inspection of the goods released to it at Pauls freight before releasing them to its customer, but generally, after its release of the phones, it received a report from Pauls Freight which confirmed the numbers and specifications, and the pallet weight of the goods exported. It was apparent that Honeyfone relied upon this and its trust in its supplier to rebut any claim from its customer that the agreed goods had not been delivered. This however involved an element of risk: it would release the goods to its customer before knowing that they were in order, and, if its customer paid before receipt by Honeyfone of the confirmation (a payment Honeyfone would generally immediately transmit to its supplier) Honeyfone could find itself facing a claim from its supplier without funds immediately available to satisfy that claim.
    (15) All the phones exported by Honeyfone in March and April were of a non-UK specification: inter alia they came with chargers that had two pin mains plugs rather than three.
    (16) There were instances among the March and April deals of the transactions of purchase and sale along the chain being completed several days before the phones arrived in the UK at Pauls Freight. In such cases the traders in the chain prior to Honeyfone formally released the goods (by instruction to Pauls Freight) to the next trader in the chain on the deal date i.e. before the goods had arrived at Pauls Freight. On occasion Honeyfone also instructed Pauls Freight to release the goods to its customer before their date of arrival.
    (17) There were examples of phones being exported on the same day as they arrived.
    (18) There was no evidence that the price at which Honeyfone dealt was significantly below the normal market price.
    (19) The Appellant also engaged in sales to UK customers in March and April 2006. The total value of these sales was significantly smaller than its export sales (only about £1m in March) and the size of the individual deals was also smaller (no more than £300k). Mr Shiralizadeh told us, and we accept, that these deals were also back to back: there were matching purchases and sales, and that payment for such transactions were often made by cheque rather than by bank account transfer.
    Miss Mitrophanous suggested that the domestic deals were smaller because Honeyfone did not have the additional money from the VAT repayment in the export deals. That did not make sense: Honeyfone would have received the VAT as part of its sales price.
    (20) Mr Shiralizadeh, in the manner of many a professional lawyer, had drawn up terms and conditions of business for its sales by cobbling together other companies' precedents. Like many printed terms they favoured the promoter of the script.
    (21) In August 2003 the Commissioners published Notice 726. In March and April 2006 Honeyfone was aware of the contents of this notice. It related to potential joint and several liability under section 77A VATA for unpaid VAT in a supply chain. It explained that there was widespread MTIC fraud and made reference to telephones. It suggested things a trader might do to avoid becoming involved in VAT fraud via its supply chain. Paragraph 8.1 and 8.2 gave some specific examples:-
    "8. Dealing with other businesses – How to ensure the integrity
    of your supply chain
    8.1 Checks you can undertake to help ensure the integrity of
    your supply chain
    The following are examples of checks you make wish to undertake to help establish the integrity of your supply chain.
    (1) Undertaking reasonable commercial checks to consider the legitimacy of customers or suppliers. For example:
    (2) Undertaking reasonable checks to ensure the commercial viability of the transaction. For example:
    (3) Undertaking reasonable checks to ensure the goods will be as described by your supplier. For example:
    We recommend that sufficient checks be carried out in each of the above categories to ensure that you are not caught in a fraudulent supply chain.
    8.2 Checks carried out by existing businesses
    The following are examples of specific checks carried out by existing businesses. These may also help you to decide what checks you should carry out, but this list is not exhaustive and you should decide what checks you need to carry out before dealing with a supplier or customer:
    (a) payments would be made to a third party; and
    (b) that in the case of import, the supplier and their bank shared the same country of residence.
    Paperwork in addition to invoices may be received in relation to the supplies you purchase and sell. We believe that this documentation should be kept as evidence of a transaction's legitimacy. The following are examples of additional paperwork that some businesses retain:
    Again this is not an exhaustive list, but does show some of the more common subsidiary documentation."
    Honeyfone undertook many if not most of these checks. Our impression however was that they were undertaken to keep HMRC happy with Honeyfone's VAT reclaims (or to avoid joint and several liability under section 77A) rather than as part of any serious investigation into its supply chain. We reach that conclusion for the following reasons:-
    (i) the evidence of both Mr Shiralizadeh and Mr Arora was that they believed in basing business on trust: trust developed by speaking on the phone, instinct and previous dealings. It was clear from their evidence that these factors were more important to them than documentary and third party evidence.
    (ii) although Honeyfone used a third party credit agency "Riskdisc" and sometimes Dun and Bradstreet to obtain references, it was clear that they used the information obtained merely as confirmation of the existence of this counterparty, and they treated comments on the reports warning of credit risk or the recent incorporation of their supplier as not giving rise to concern because Honeyfone was not extending credit – although such information might have given rise to concern about how such companies could deal in suck large sums and extend credit to Honeyfone (on the occasions it paid its suppliers only in part at the time of the deal) and
    (iii) although written references were obtained from other companies the references were occasionally in consistent and generally in standard form. It appeared to us that little consideration had been given to the strength of the reference given.
    On the other hand Honeyfone ensured it made payments only to the supplier's bank account, had met some of the individuals involved, obtained declarations from suppliers (in fairly standard form) that they would pay VAT and did not recognise VAT fraud in the supply to them, and checked their suppliers' and customers' VAT numbers with HMRC's officers at Redhill.
    (22) In their conversation with those whom it dealt Mr Shiralizadeh and Mr Arora became aware of the interest of others in VAT repayments. One of the questions asked of them was: "Are you still getting your VAT repayments: have there been any problems?"
    (23) Honeyfone's VAT repayment claims for the seven months to 30 April 2006 were:
    10/05 £243k
    11/05 £128k
    12/05 £199k
    01/06 £276k
    02/06 £869k
    03/06 £2,086k
    04/06 £2,906k
    Mr Lam was responsible for authorising payment of the chains up to 03/06. Payment for 10/05, 11/05, and 12/05 was made without making further enquiries. Mr Lam however sought further documentary evidence before authorising repayment for 01/06 and 02/06. The documentary evidence sought and provided related principally to proof of export and due diligence documents in relation to its customers.
    Honeyfone kept and presented to HMRC documentation in respect of its transactions which it had been requested to maintain (in a letter of 1 December 2005) to "assist HMRC in verifying who has legal title to the goods, that a transaction has taken place, and that the transaction relates to the goods physically exported."
    In the 01/06 period Honeyfone had increased its turnover somewhat and had had only three export customers, yet Mr Lam had not passed adverse comment. In 02/06 there was a threefold jump in monthly turnover from £375k to £808k, but on the documentation and evidence of due diligence presented to him Mr Lam authorised payment.
    These repayments in our view gave Honeyfone the impression that they were doing the right things to get repayment. They did not however, and in our view could not, give Honeyfone any confidence that in fact their transactions were not connected with fraud.
    (24) The due diligence undertaken by Honeyfone on its supplies revealed to Honeyfone that many of them were fairly recently formed companies. The table below sets out the suppliers in March and April and the date of formation of the supplier as obtained from Honeyfone's due diligence enquiries:
    Supplier Incorporation Month of supply
    Team Mobile UK 16/12/97 April
    Global Roaming Ltd 9/9/02 April & March
    MP3 Dot Com Ltd 15/3/05 April
    Euro Asia Telecom Ltd 19/9/05 April
    Universal Distribution 6/11/03 March
    Zain Communications Ltd 19/1/04 March
    TM Global 9/6/99 March
    In early 2006 these were all relatively young companies.
    (25) Honeyfone had only three export customers in March and April 2006
    Customer Date of Formation
    Celcom Trading 24 August 2005
    Freshnet International 2005
    M S Enterprise 2004
    Honeyfone obtained confirmation from HMRC's Redhill office that M S Enterprise and Celcom Trading had VAT registrations in their member States on 27 February 2006, and a similar confirmation for Freshnet International on 15 February 2006. The confirmations contained a disclaimer indicating that they did not mean that input tax claims would not need to be verified.
    Celcom Trading was deregistered by the Dutch VAT authorities on 6 May 2006. M S Enterprises was, possibly on the basis of information from HMRC, investigated by the French authorities.
    (26) Each of Honeyfone's three export customers in March and April 2006 namely M S Enterprise, Celcom and Freshnet International had a sole director who was recorded in Honeyfone's due diligence documents as having a home address in the UK. Two were recorded as having British nationality.
    Knowledge of fraud
  160. The Respondents rely upon a number of factors as indicating that Honeyfone knew that there was fraud in each deal chain. We consider them in turn:-
  161. (i) Mr Foulkes says that the deal chains were contrived and that suggests that the Appellant was a knowing participant. Although it might be possible for a party unwittingly to be guided towards particular deals Mr Shiralizadeh had been clear that he had not been.
    We have already noted the features of the deal chains which suggest that they were engineered. We find the pricing differential enjoyed by the exporter particularly suspicious – it has the smell of an extra payment for taking on a risk of non-recovery of input VAT. It also points away from a chain which is created by creating demand at one point in the market and supply at another and leaving the market to fill the space between the two by negotiated deals. The length (3 buffers) of most of the chains points the same way. It seems to us clear that the chains were engineered by at least one mind.
    But that mind (or those minds) could have whispered in the ears of Honeyfone's customers and suppliers without involving Honeyfone: Honeyfone could have been placed in the chain without receiving a nudge or a wink. And a skilled orchestrator may well have wanted Honeyfone to remain ignorant. However, it seems to us unlikely that this could have been achieved in so many deals without giving Honeyfone at least some suspicion that it was being at the lowest manipulated into position. To that extent we do not accept Mr Shiralizadeh's evidence that they were not aware of being guided.
    (ii) The Respondents say that whilst there may be a small legitimate wholesale market generally for packages of phones sold for £19,999 to £50,000, 95% of the wholesale market was in early 2006 driven by and its transactions were connected to VAT fraud. They rely upon the transcript of part of the evidence given by Mr Roderick Stone in the Olympia case to support this.
    We did not hear Mr Stone in person. We are happy to adopt the approach of the tribunal in Olympia to it: they approached the use of the evidence with caution. Even though they did not doubt the statistics they doubted that their extent had been publicly revealed or understood at the relevant time. What is at issue is whether Honeyfone knew that its transactions were linked to fraud rather than whether they were. Given that we have found that 26 out of 28 of Honeyfone's export deals in March and April 2006 were connected to fraud and that Honeyfone knew that there was some fraud in the market this takes us no further.
    (iii) Mr Foulkes says that the very existence of a chain of traders conducting a rapid succession of deals cannot be a feature of a genuine market. The Appellant knew there were chains and that they completed rapidly: it most therefore have known it was not participating in a genuine market.
    He says that the chains cannot be features of a genuine market since, where all the traders have access to the market, the only function of each trader in the chain is to increase the price before the goods are eventually exported: it beggars belief that the exporters could not have found the importer and buy the phones directly from him at a lower price.
    Mr Shiralizadeh said that one of the reasons for the market was to split large numbers of phones into digestible chunks: the manufacturers and distributors dealt only in large numbers of phones; the phone dealers split these into a smaller and yet smaller parcels until they became acceptable to a retailer. Some phone dealers had the capacity to deal (and were trusted to deal) in large numbers of phones, other lesser dealers could deal only in smaller parcels. He explained that of the 16 deals done on 31 March, eight had been split deals in which a purchase from one supplier, Zain, had been onsold to two different customers; and one of the nine deals done on 28 April had been similarly split. In each case however Honeyfone's supplier had invoiced the sale to Honeyfone in two tranches, one corresponding to each onward sale by Honeyfone. We were told Honeyfone asked its supplier to provide two invoices to aid its administration.
    The following table shows the chains of supply involved in these split deals in March. The columns in the table show the successive suppliers in the chains:
    Deal No (5) (11) (4) (11) (4) (5) (7) (8)
    Phones N 90 N 90 6680 6680 0600 0600 8800 8880
    X 1550 X550 X700 X500 X1552 X1410 X1455 X1330
    Sold to MS Ent. Fr Callc. Fr Callc MS Ent Callc MS Ent
    Bought
    from: Zain Zain Zain Zain
    Who bought
    from MP3 MP3 MP3 MP3
    Who bought
    from StyleZ First Ass. StyleZ First Ass StyleZ First Ass. StyleZ First Ass
    Who bought
    From C&E C&E C&E C&E
    Thus, for example, the first column shows that C&E Enterprises sold 1550 N90s to Stylez and 550 to First Associates; they onsold to MP3.Com; it onsold to Zain Communications, but provided two invoices – one for 1550 and the other for 550; Zain sold to Honeyfone, providing two invoices similarly, and Honeyfone to M S Enterprises and Freshnet. The last four columns show that the phones for the split all came from the same chain – but in each case each supplier produced split invoices to its customer.
    We find these chains very suspicious. We cannot see why the all earlier suppliers in the chain would also split their invoices, we are surprised that the sales by C&E Enterprises to Stylez and First Associates in columns 1 to 4 are of numbers of phones which exactly match the onward sales by Honeyfone – after having been reunited in MP3, Zain and Honeyfone, and the common members of each chain are startling.
    These concerns and the nature of the chains point to contrivance of the chains, but they also cast considerable doubt on Mr Shiralizadeh's evidence that his sales were really split deals in which he bought a larger number of phones and sold them in two tranches: it is particularly difficult to accept that in relation to the first four columns where the Stylez/First Associates split exactly mirrors the M S Enterprise/Freshnet split.
    We conclude; (i) there were no split deals entered into by Honeyfone in the period; (ii) either Mr Shiralizadeh was not telling the truth or he was very skilfully manipulated in relation to those transactions; (iii) this is further clear evidence of contrivance of the chains and of Honeyfone's being manipulated or involved in that contrivance; and (iv) that some suspicion that all was not pure must have reached Honeyfone.
    It is not inconceivable that Honeyfone genuinely believed that the market in which it operated was genuine for the reasons Mr Shiralizadeh gave, but these issues cause us not to be able to accept that evidence.
    (iv) The Respondents say that the back to back chains in which all transactions take place within a short space of time show (a) everything is pre-arranged, and (b) that the Appellant's explanation of the way the deals get at the end of the month cannot be correct: it must have known that these were continued chains connected to fraud.
    We agree, for reasons explained elsewhere, that the back to back timing at the end of the month points to a continued chain and raises a suspicion of connection to fraud. We did not however find the witnesses' description of how deals were negotiated towards the end of the month and gelled in the last few days unbelievable.
    (v) Mr Foulkes says that a legitimate market is led by the demand of the eventual users. That demand he says will not peak at the end of each month although of course it may fluctuate. The Appellants "market" he says was (on Mr Shiralizadeh's evidence) substantially led by suppliers, and governed by VAT repayment timing: it could not be a genuine market. Although he did not say so expressly in his closing remarks he asks us implicitly to infer that the Appellant must therefore have known he was dealing in transactions connected to fraud.
    We also find the concentration of the market export transactions on the closing days of each month disturbing. If there is someone who holds a stock of phones he will have paid for them. The longer he holds them the longer he has to meet the cost of financing that stock. He will wish to sell as soon as he can: his financing cost (say the 7% interest which his bank charges) will be on 100% of the cost of the phones. The exporter who buys and sells on the same day will have to finance the VAT input tax cost (17½% of the cost) until repayment. He wants to delay his transaction until the end of the month. But his finance cost is less per day than that of the supplier. Would not the ordinary result be some compromise whereunder the goods were bought earlier than the end of the month?
    These features of the transactions in which Honeyfone was engaged excite suspicion in our minds that there was something funny about the market in which Honeyfone dealt. They indicate that obtaining a VAT repayment in respect of exports was a particularly relevant aspect of the timing and execution of deals. In the context of knowledge that there was some VAT fraud in the mobile telephone market they raise, and would have raised, in our minds real concerns that there might have been VAT fraud somewhere in the chains which led to Honeyfone. However they do not on their own indicate to us that it was more likely that not that Honeyfone knew that there was such fraud.
    (vi) There was one deal chain in which Honeyfone bought phones from Global Roaming, and another deal chain (in which Honeyfone's transaction took place on the same day) in which Honeyfone bought phones from TM Global who bought from Global Roaming. Mr Shiralizadeh said this could happen where he was offered stock by Global but had not found a customer; he might then ring round to find a customer but by the time he had found one Global may have agreed to sell to T M Global. Mr Foulkes regards this as inconsistent with a market in which T M Global would have agreed to purchase only if it had found a customer and the deals gelled only at the end of the month. He also notes that in a market in which Honeyfone had some years experience that it could not source phones at prices less than offered by its suppliers.
    In a marketplace where deals gel very close to the month end, and in which no one wishes to be left holding the parcel when the music stops we can just about see how the two deals may have happened. But it is very odd that when negotiating one deal with Global Roaming no further discussion took place about the phones which Global Roaming was about to sell to T M Global for onward transmission to Honeyfone. It is possible that in the frenzy of finalising 11 deals on 31 March that this was overlooked or that by then Global Roaming felt honour bound to sell to T M Global (who might in turn have been honour bound to sell to Honeyfone). But we have great difficulty in imagining the way this could happen. This confirms us in our view that in fact the chains were orchestrated and that at the very least Honeyfone was a puppet. It casts some doubt on Mr Shiralizadeh's and Mr Arora's account of the operation of their market, and in that, our view is a pointer to a conclusion that they knew that there was something dodgy in their market.
    (vii) The Respondents rely upon the Appellant's failure properly to inspect the goods they bought (both the limited nature of the inspection and the fact that the request was received after the sale), their failure to record IMEI numbers and their failure to insure the phones, as evidence that they knew that thee were transactions connected to VAT fraud.
    We cannot see how these failures (apart perhaps from IMEI number recording) point clearly to knowledge of fraud. If it were alleged that there were no phones at all these facts could be relevant to that allegation. But it is not. By its failure to inspect and insure the Appellant exposed itself to potential commercial disadvantage. But how can that indicate that it knew someone else was involved in VAT fraud? The most it might indicate is that it had been told that it would get its money for its sale whether or not the goods were actually delivered and accordingly that it was participating in a scam. The Respondents suggest that the profit of MITC fraud would be sufficient to compensate for any loss along the chain. But as likely is the Appellant's explanation that it took a calculated risk: after all it knew the phones were at Pauls Freight and that they would be at Honeyfone's risk for only a short moment of time; and if the goods it received were faulty it would be in the same position whether or not it made extensive spot checks before selling them on. These failures point to a practice wherein ownership was only ever transient, to a recognition of a chain, but not clearly to knowledge of fraud.
    Mr Foulkes says failure to record IMEI members suggests that Honeyfone knew its customer was not going to complain: the phones were merely a tool of the fraud. It seems to us that an expectation based on experience and the safe packing of large numbers of phones may be an equally likely reason for a failure to inspect.
    The failure to determine IMEI numbers is different. It points in our view to a lack of care as to whether or not these phones could be involved in fraud. That lack of care could itself be something present when it was known that fraud was involved. It therefore may point to knowledge.
    (viii) Mr Foulkes says that the fact that in each chain the phones start at Prologis (in France) and end back there a day or so later, and the fact that each chain has one of three defaulters at the end of it indicated contrived chains. And that can only be explained by each participant knowing of the fraud or acting on the direction of the organiser.
    We agree that these features point to orchestration of the chain, but they leave open the clear possibility that Honeyfone was an unwitting puppet. We cannot say that they point to it being likely that Honeyfone knew of the fraud (but see our comment at (i) above).
    (ix) Although sales and purchases were generally completed by same day or almost same day delivery or release and invoicing, payment was not always made in full by Honeyfone on the day of sale. It appeared that part payments were made when monies were received from customers and that balancing payments after as part of a large sum were made later. The balancing amounts could be in six figures. Mr Shiralizadeh's evidence was that his suppliers were used to dealing with Honeyfone and trusted it to make payment when it could. The receipt of the VAT reclaim would assist the making of the balancing payments.
    There was also an example of payment being made by Honeyfone's customer to Honeyfone before the goods were released to the customer, and by Honeyfone to its supplier before the goods were released by the supplier: so that both the customer and Honeyfone were, for a short period, at risk. The transactions were thus not without risk to the credit or performance of other parties in the chain. Mr Shiralizadeh said that the willingness to accept that risk was a result of trust and also that "we are all relying on Pauls Freight as a warehouse … to take care of our interests everybody's interests."
    The Respondents suggest that the granting by its suppliers in this way of credit without terms or agreement suggests that the transactions were not at arm's length.
    In 2006 Honeyfone had been trading for more than five years. We think it is possible that in the conversations which Honeyfone had with its supplier that some measure of its age and experience was conveyed to its suppliers. It is possible that Mr Shiralizadeh and Mr Arora treated their suppliers' willingness to extend credit to them as reflecting the established nature of Honeyfone. But its suppliers were generally younger companies and it is surprising that they had the funds to extend credit to Honeyfone unless they too took credit from their suppliers. Overall we agree with Mr Foulkes that the granting of quite substantial credit is odd and suggests something smelly in the chain. But the fact that Honeyfone was granted credit and occasionally took some credit risk itself does not point squarely to Honeyfone knowing that there was fraud; for even if there was fraud these risks were still in reality taken and could not be washed clean by the fraud.
    (x) The Respondents point to the almost twelve-fold increase in the turnover of Honeyfone in March and April following the Bondhouse decision. That increase required extra funds – Honeyfone could not pay its suppliers the full amount it owed them until it got the VAT back unless additional funding was obtained. Some additional funding was obtained by the retention of profits but the balance some £2-3m came from taking extended credit from its suppliers.
    Mr Foulkes says that it is apparent that those in the chains concluded that VAT payments could not be denied even though there was fraud and thinking they were safe were happy to increase their credit and activity knowing that they were not involved in the fraud itself.
    We do not see the increase in activity as a pointer that Honeyfone knew about fraud in the chains. Rather, it points to Honeyfone (and others) being careless as to whether there was fraud. So long as they were not themselves fraudulent and didn't know about a fraud in the chain, they felt they were safe.
    However, the increase in the availability of deals to be done after the Bomdhouse victory might have suggested that there was a possibility that the deals were potentially tainted with fraud. If a fraud in the chain was prior to the Bondhouse victory a concern but not thereafter that could give rise to a suspicion that there was indeed fraud.
    (xi) All the export transactions related to non-UK specification goods held physically in the UK. The Respondents say it makes no sense to import the phones just to export them.
    We agree that it is odd that non-UK specification phones needed to be imported into the UK in order to be traded. They could more easily have been bought and sold while remaining at depositories abroad. This points again to contrivance in, and suspicion of, the chains. It is also of note that all of the Appellant's deals involved such imports: one would have expected at least some of the deals to relate to offered and bought phones located abroad. Thus one would also have expected Honeyfone to be suspicious about the nature of its market unless it knew it was connected with MTIC fraud.
    Mr Shiralizadeh says that the UK operated as a "hub" into which phones were imported and that he felt more secure dealing with phones held through a freight forwarder he knew and trusted. We do not accept that evidence. Even if some non-UK specification phones were released directly into the UK by distributors we would expect the vast majority to be released into the market for which they were intended, and if traded by phone dealers, traded while remaining abroad. What is suspicious is not that Mr Shiralizadeh preferred a UK freight forwarder but that the phones were always put in UK by those who sold to Honeyfone.
    We find that to be a possible pointer towards knowledge of fraud and a clear pointer to a reasonable suspicion that something was wrong if all the phones the subject of those deals were to be delivered in the UK. We do not see the failure of Notice 726 to mention this feature as relevant.
    (xii) Mr Foulkes says the fact that Honeyfone's three customers in this period had UK residing directors is an extraordinary feature which suggests knowledge of fraud.
    We do not agree. It is odd that all three had directors residing in the UK but we cannot see how that coincidence points to knowledge of fraud anymore than all the customers having French directors would. The coincidence would however have caused some concern to a reasonable businessman in their position
    (xiii) Two of the Appellant's customers had been deregistered or investigated by their local VAT authorities.
    We see this as a pointer to orchestration on contrivance but we cannot see that it points to Honeyfone's knowledge of fraud.
    (xiv) Mr Foulkes points to the common use of FCIB by the participants in the chains including Honeyfone. He says that the inference to be drawn is that this bank offered a particular service: a willingness to ask no questions about huge sums passing through small companies' accounts.
    Mr Shiralizadeh's evidence was that other banks had withdrawn 24 hour banking and that Honeyfone (and presumably others as well) gravitated towards FCIB because it still offered the service. We accept that evidence but we note that the withdrawal by its former bankers of their services should have further alerted Honeyfone to the dangers of fraud in its market. We also find the use of the same bank by all the participants suspicious, and think that Honeyfone should have found the use of the same bank by all its customers and suppliers, at least unusual.
    (xv) The Respondents say that the due diligence conducted by the Appellant was window dressing. It was done merely to satisfy HMRC and get its VAT back and did not reflect a serious attempt to ensure it was not connected to fraud.
    We accept, for reasons we shall explain later, that the investigations undertaken by the Appellant were in particular sense inadequate. But we do not see that as a pointer to knowledge of fraud, merely to a measure of carelessness as to whether or not there was fraud. Fulfilling HMRC's guidelines is quite understandable if they are relevant to getting your VAT back.
    (xvi) Mr Foulkes says that it is very odd that Honeyfone had only three export customers in this period. Mr Shiralizadeh say that two months is merely a snapshot of his business. We agree: we think little can be inferred from two months sales.
  162. In summary the Respondents say that the fact that there was a contrived chain and the Appellant's lack of apparent suspicion and investigation point to knowledge of fraud.
  163. Standing back and considering what we heard and saw in the round we cannot unanimously conclude that it is more likely than not that the Appellant knew that there was fraud in the chains. There are pointers towards knowledge but, or one of us, there were not quite enough or of quite enough weight to be able to reach that conclusion.
  164. Means of knowledge
  165. Our impression of Mr Shiralizadeh was that he was someone who was keen to succeed and to justify the trust placed in him by his family. He revered his father and took to heart his father's approach of trusting others. These traits in our view made him ignore the features of Honeyfone's business which, in another man, would have excited suspicion and grave concern. Mr Shiralizadeh gave us the impression that he took the market in which he operated at face value: he assumed that because it worked in a particular way there was genuine commercial logic for that, he arbitraged, he did the deals and didn't need to think about where the phones came from or whither they went.
  166. Mr Arora was older and had greater experience than Mr Shiralizadeh: Mr Shiralizadeh was not inexperienced but Mr Arora had almost 20 years experience in the telephone industry by 2006. Mr Arora indicated that he too believed in trust.
  167. The features of Honeyfone's business and transactions were known to Honeyfone and which in our view together would excite serious concern in the mind of a reasonable businessman having the combined skill and knowledge of Mr Shiralizadeh and Mr Arora were these:
  168. (i) there was some VAT fraud in the mobile phone market (726 and FCIB);
    (ii) NatWest and then Lloyds bank withdrawing their services; Barclays refusing services.
    (iii) the concentration of export deals at the end of the month;
    (iv) the recognition that the concentration at the end of the month was associated with VAT repayments;
    (v) the large amounts of credit granted by fairly young companies to Honeyfone;
    (vi) the number of export deals which were in relation to non-UK specification phones which had been imported into the UK, but that none were in relation to such phones held outside the UK;
    (vii) the sudden increase in possible trade after Bondhouse and a lurking question as to whether the perceived lack of need to look beyond your immediate neighbours in a chain meant in fact that this was fraud somewhere;
    (viii) that there was a chain of deals prior to Honeyfone's purchase and that some of the deals in that chain were substantially contemporaneous;
    (ix) that the Appellants export deals were larger individually and in aggregate than its domestic deals;
    (x) the youth of its supplying companies;
    (xi) all the Appellants' suppliers used the same freight forwarder;
    (xii) the coincidence that all three of its export customers in March and April had UK directors;
    (xiii) the fact that one of the questions asked of Honeyfone was whether it was getting its VAT back. Someone who was getting their VAT back would be a good person to make more exports: a good person to 'place' in other chains.
  169. Those factors would, in our view, have led a reasonable businessman possessed of the skill and experience of Mr Shiralizadeh and Mr Arora to have suspected VAT fraud in connection with Honeyfone's transactions. We have also indicated that we believe that it is unlikely that the chains were orchestrated without giving Honeyfone some suspicion that it was being at the least nudged into its place in each chain. We noted above indications that Honeyfone may have been somewhat careless as to whether there was VAT fraud (approach to due diligence, IMEI numbers). In our view such a reasonable businessman would not have relied mainly and simply upon trusting his suppliers and would not have had his concerns about the possible existence of fraud abated by the Bondhouse case; he would either have concluded that there was VAT fraud in the chains or would have made a thorough investigation with the aim of determining whether or not there was fraud in the chain.
  170. Miss Mitrophanous says that we should have regard to HMRC's notice 726 and the prior conduct of the Appellant's VAT officers:
  171. (i) Notice 726 she said made it clear that it was transparently low prices which were the key to identifying fraud and that a trader should have an awareness of its trading partners. The Appellant had not done deals at low prices and had conducted due diligence on his suppliers.

    Paragraph 2.5 of Notice 726 describes the rebuttable presumption that a transparently low price indicates that the VAT somewhere in the chain would go unpaid. To that extent it merely paraphrases section 77A(6). But that paragraph continues:

    "These tests … are made without prejudice to any other way of establishing reasonable grounds for suspicion."

    There is nothing in that notice which suggests to us that a trader should ignore or give less weight to other matters which might reasonably suggest fraud. We have set out in paragraph 104(21) above the checks suggested in paragraphs 8.1 and 8.2 of the notice to help to ensure the integrity of the supply chain. But paragraph 4.6 of the Notice puts these in context; it is headed "Can you tell me exactly what checks I should undertake?" And the first word of the paragraph is "No." We read the notice as suggesting things the trader might do, rather than giving comprehensive guidance as to how and when to identify fraud.

  172. Whilst the failure to do something suggested by the Revenue in that notice in circumstances, where doing that thing may have uncovered knowledge of fraud, might indicate that a reasonable businessman who had knowledge of the notice should have known of the fraud, it does not seem to us that the reasonable businessman would necessarily consider that he had laid a suspicion sufficiently to rest merely because he had mechanically followed the suggestions in the notice. If the stink from the transactions was strong enough he would not find the perfume of HMRC's tests enough to cover it up.
  173. We have recorded elsewhere our conclusions as to the approach of the Appellant to the checks suggested in Notice 726: they were viewed as tests to keep the Revenue happy not ways of identifying fraud or allaying suspicion of fraud. They were directed principally to obtaining documentation to confirm the existence of their suppliers and customers, their VAT registration and that they were in business. This approach led them to being less critical in the appraisal of the responses received and information gained than would a reasonable businessman have been in this position. In our view such a person would not have had his suspicions allayed by conducting the mechanical due diligence undertaken by Honeyfone and in some respects would have had it heightened. (Those respects relate to the obvious youth of their trading partner companies and an apparent carelessness in some of the trading references.)
  174. We therefore do not consider that the existence or context of Notice 726 or the actions the Appellant took as a result of it would have abated the suspicions which a reasonable businessman in the Appellant's position would have had of the export transactions it was engaged in, or reduced such a person's need to investigate further.
  175. (ii) in prior periods HMRC's officers had authorised repayment claims fairly readily and had not commented adversely on the Appellant's business or its characteristics.
    We agree that the Appellant could take some comfort from this; but there is in our view a difference between presenting a month's papers to HMRC and awaiting repayment with bated breath, and setting out before HMRC the characteristics of the market which give rise to concern and asking for specific comfort.
  176. What more could Honeyfone reasonably have done? It could not have expected its suppliers to tell it who their suppliers were so that it could itself have investigated the chain. If it had suspected that 95% of the deals in the market were linked to fraud it might have been reasonable to conclude that all its deals were so linked; but it did not have that information: it could only deal with the information it had.
  177. In our view it would have been reasonable (or proportionate) for Honeyfone to have contacted its VAT officers and said "Look I'm concerned about this, can you investigate for me: after all you have contact with my suppliers?"; it could have asked Pauls Freight: "You knew who the first supplier in this chain is (or was), can you give me any idea whether they are substantial and get them to confirm they intend to pay this VAT and "Can you tell me where the phones are imported?"; it could have gone in the person of Mr Shiralizadeh or Mr Arora, to visit its customers and expressed the concerns and ask them to do the same and so on down the line and to report back; it could have asked why the non-UK specification phones were in the UK, when they were brought in and why and sought the help of their suppliers in finding the answers. It could have made some of the enquiries in relation to past deals, but could also have pursued some of these in relation to contemplated deals. Some of these may have produced no information, but we think it likely that at least some of them would have produced something which would have led to a conclusion that it was more likely than not that the relevant importer in relation to each of its April and March deals was intending to defraud: given the contrivance of the chain and its odd features some smell of rot would in our view have reached Honeyfone. A reply need not have been categoric to give rise to that conclusion – an evasive reply, pulling out of a deal or an unwillingness to help might each, in these circumstances have been sufficient grounds for belief that on balance VAT fraud was involved. Making a fuss might have delayed the finalisation of its deals; or it might have meant that its suppliers or customers pulled out, but it seems to us that those would have been risks a reasonable businessman would have borne in order to provide himself with some substantive comfort that his seemingly suspicious transactions were not linked to VAT fraud. Thus in our view Honeyfone had the means of knowing that those of its transactions which were connected to VAT fraud were so connected: if it had conducted those enquiries which in our view a reasonable businessman in its position, possessed of its skill and knowledge, would have conducted, the result would, in our view, have been a conclusion that the relevant transactions were connected to fraud.
  178. We conclude in relation to deals 1 to 14, 16 to 23, and 25 to 28 that Honeyfone should have known that they were transactions connected to the fraudulent evasion of VAT. We wish to make clear that this is an objective factual evaluation: we are not suggesting that Mr Shiralizadeh failed in any duty, need or statutory, but finding that a reasonable businessman in Honeyfone's position would have concluded on the balance of probabilities that those transactions were connected with VAT fraud.
  179. All the Input Tax or the tax lost only
  180. In the example we gave at the start of this decision the defaulting trader had failed to pay £18.375 of input tax and the Exporter was claiming input tax of £19.25. The difference of £9.875 was the tax payable by the Buffer A. If Buffer A has paid the £0.875 to HMRC and HMRC are successful in denying Exporter the whole of the £19.25, then the Commissioners will reap a profit of £0.875.
  181. At paragraph 15 above we referred to the possibility of Multiple Recovery in the context of denial of input tax at each stage in the chain discussed by Burton J in Just Fabulous and noted that Mr Foulkes had indicated in this case that the Commissioners undertook that they would not seek double recovery.
  182. Paragraphs 33-39 and 46 to 54 of Burton J's judgment in Just Fabulous touch on multiple recovery both in the sense discussed in paragraph 15 above, and in relation to the £0.875 discussed above, but give no clear steer (understandably – because the issue was not before him on the judicial review application) as to whether the full £19.25 or the lost £18.375 should be denied: at paragraph 47 he records an argument for HMRC, which he accepts at paragraph 54, and which seems to relate mainly to the paragraph 15 style Multiple Recovery, that if the Revenue sought such recovery "no doubt ways can be found by the Tribunal or the Courts to prevent double or multiple recovery".
  183. Miss Mitrophanous asks us to determine the input tax credit at issue in this appeal by denying, where appropriate, only the tax lost by the Commissioners. Where the traders between the defaulter and Honeyfone had paid VAT on their margin, she says that to deny Honeyfone all its input tax would be a kind of multiple recovery which is unwarranted.
  184. Mr Foulkes refers us to the argument in Just Fabulous that the recovery of more than is lost could amount to a penalty. He reminds us of Burton J's extreme example in which £2m of input VAT was denied where the original default was only £1m, and that Burton J notes that the Revenue did not accept that "recoupment of a sum more than the loss caused by the original defaulter, simply by virtue of the mark-ups on each buffer transaction along the chain, constitutes a penalty either." Mr Foulkes notes that Burton J then says at paragraph 50 that he was "wholly persuaded by the submissions" of the Revenue. The possible increase in the amount of tax claimed beyond the original default in this case falls, Mr Foulkes says, far below the level that Burton J envisaged as possibly giving rise to issues such as penalty and multiple recovery.
  185. In Calltell the tribunal noted Burton J's indication that ensuring that there was no risk of double recovery was essentially a matter for the tribunal and recorded HMRC's assurance that if any of the tax was recovered from another source the Appellant's corresponding claims would be met to that extent (see paragraph 54). But the tribunal went on to consider the difference it saw between the approach in Halifax and that in Kittel. It saw it as significant (paragraph 54) that in Kittel the "court did not indicate that the right to deduct is lost only to the extent of the tax lost … on the contrary the indications are that a trader … forfeits the entire right to deduct regardless of the measure of the tax lost and … that that outcome is desirable as a means of discouraging fraud." The tribunal then went on to consider whether the loss of the entire right in such circumstances was a penalty (which would be proscribed by paragraph 53 of the judgment in Halifax). The tribunal concluded that it was not. The implication therefrom is that HMRC's undertaking was a concession.
  186. In this manner it seems to us the tribunal in Calltell applied the flexibility to find ways to prevent double or multiple recovery alluded to by Burton J in Just Fabulous to find that such recovery did not matter at law.
  187. We do not accept Mr Foulkes' inference that Burton J was saying that multiple recovery was acceptable. Burton J was "wholly satisfied" by submissions which included one that ways could be found by the Tribunal or the Court to prevent multiple recovery.
  188. We agree with the tribunal in Calltell that the language of the Court in Kittel is of the right to deduct being lost rather than lost in part. But the Court was dealing with the position of the purchaser from the fraudulent trader not a trader at several venues. The reasoning of the Court is that denial of that trader's right to deduct is because the knowing trader is a participant in the fraud and is relying on the right for fraudulent ends, and that denial is apt to prevent fraud. But both these criteria would be fulfilled if only the tax lost were denied as input tax: it is only that tax in respect of which a claim is being relied upon for fraudulent ends, and the denial of a deduction for that tax would also be a deterrent. It would not be perhaps as much of a deterrent as denial of all the input tax, but to the extent the denial exceeds the tax lost the denial has more of the flavour of a penalty. The tribunal says at paragraphs 59 and 60:
  189. "59. The question, therefore, is whether the loss of the right to deduct is properly to be regarded as a penalty, either to the extent that the value of the lost right exceeds the tax loss elsewhere, or to the extent that it is not regained when tax originally lost elsewhere is recovered.
  190. If the … they were participants or … accomplices in the perpetration of a fraud. The objective of the fraud is to extract from the Commissioners the very input tax which is the subject of the appeals. If that objective is not achieved, there is no purpose at all behind the transactions … It cannot be said that preventing the participants from achieving that objective is tantamount to the imposition on them of a penalty, just as one would not regard the preventing of a thief from taking the property he intended to steal amounted to the imposition on him of a penalty. Viewed in that way, it is immaterial whether there is any relation between the value of the tax for which, elsewhere in the chain, another trader has failed to account: what the Commissioners are seeking to do is to prevent a theft, rather than to recover a loss. In our judgment there is no basis on which it can be said that they are seeking to impose a penalty on the Appellants, and we do not accept that what was said by the Court of Justice at paragraph 93 of its judgment in Halifax is of any application in this case."
  191. We are not clear, in our factual circumstances, that the objective of the fraud was to extract from the Commissioners "the very input tax which is the subject of the appeal". In this case it seems clearer to us that the objective of the orchestration of the chain was to deny the Commissioners the output tax which the defaulting trader failed to pay. That is the distinctive feature of the fraud. The monies to provide the defaulting importer with its illicit profit could derive equally well from selling the phones at a car boot sale or from their export. Thus, on the facts of this case, we do not find the reasoning in this respect of the tribunal in Calltell helpful.
  192. In our view: (a) the decision of the Court in Kittel leaves open the possibility that the denial should be only of the tax which was evaded; (b) denial of the evaded tax only is a denial of the use of the right to deduct to the extent it is used for fraudulent and: (c) such (limited) denial is apt to discourage further fraud; (d) if and to the extent that the denial extended beyond the tax lost it is possible that the denial could be regarded as a penalty; and (e) that last result is avoided and the neutrality of the Common System of VAT maintained by denying only the tax evaded.
  193. We therefore find in this case that tax actually paid by the buffer traders should not be denied as input tax to Honeyfone.
  194. We are comforted in this decision by the language of the Court in Kittel: that it is for the national Court to deny the input tax, and also by the jurisdiction given to the tribunal under section 83( c) VATA – to hear an appeal in relation to the amount of any input tax credit rather than an appeal against a decision of HMRC. That jurisdiction seems well suited to exercising the flexibility of determination which helped convince Burton J that denial of a trader's input tax would not be a penalty.
  195. Conclusions
  196. We find:
  197. (i) that issues relating to the lawfulness or otherwise of the Commissioners' approach to denying exporters their input tax are irrelevant to our duty to determine the amount of input tax for which credit should be given (see paragraphs 8 to 37);
    (ii) that of the 28 transactions listed in the Annex which were the subject of this appeal, all of them apart from number 24 were connected with a sale by one of three traders: Computec, Worldwide Enterprises and C&E Enterprises (paragraphs 75 to 81);
    (iii) that of the 27 remaining transactions with the exception of transaction number 15 at the start of the chain fraudulently defaulted in the payment of VAT on its sale and intended to evade payment at the time of the transaction (paragraphs 86 to 102);
    (iv) that in the case of each of the 26 transactions which thus traced back to a fraudulently defaulting trader the Appellant should have known that those transactions were connected with the fraudulent evasion of VAT (paragraphs 103 to 118);
    (v) that accordingly there should be a denial of Honeyfone's input VAT credit in respect of all the transactions other than number 15 and 24; and
    (vi) but that such denial should be of the input tax claimed by Honeyfone less the amount of VAT actually paid on each of those transactions by the buffers in the chains between Honeyfone and the defaulting trader (paragraphs 119 to 131).
  198. Since we do not have evidence of the VAT actually paid by the buffers we adjourn this appeal for the parties to agree the relevant amount and in default of agreement to return to the Tribunal.
  199. The appeal is therefore allowed in part.
  200. We direct that the Appellant pay three quarters of the Respondents costs of, incidental to, and consequent upon the appeal on the standard basis to be determined in default of agreement by a Taxing Master of the Supreme Court.
  201. We direct that any application by the Appellant relating to interest on the repayment of input tax due to the Appellant be made to the tribunal with a copy to the Respondents within 21 days of the date of the release of this decision.
  202. Our decision was serve as noted in paragraph 107 above, unanimous.
  203. We indicated at paragraph 36 that our conclusions involved matters of EU law. We heard no argument during the hearing as to whether a reference to the ECJ was appropriate. We provided a preliminary decision to the parties before releasing this decision which set out the contents of paragraphs 1 to 36 above, and we directed that the parties indicate within 14 days of the release of that preliminary decision whether they wished any question to be referred to the ECJ and we set out a timetable for approaching that issue. The Appellant served notice in accordance with those directions that it wished to refer a question the ECJ. This decision is therefore subject to any reference which the tribunal might decide to make. The appeal is therefore formally adjourned for further steps in accordance with these directions.
  204. CHARLES HELLIER
    CHAIRMAN
    RELEASED: 2 May 2008

    LON 2007/0404


BAILII: Copyright Policy | Disclaimers | Privacy Policy | Feedback | Donate to BAILII
URL: http://www.bailii.org/uk/cases/UKVAT/2008/V20667.html