BAILII is celebrating 24 years of free online access to the law! Would you consider making a contribution?

No donation is too small. If every visitor before 31 December gives just Β£1, it will have a significant impact on BAILII's ability to continue providing free access to the law.
Thank you very much for your support!



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 >> Dragon Futures Ltd v Revenue and Customs [2005] UKVAT V19186 (28 July 2005)
URL: http://www.bailii.org/uk/cases/UKVAT/2005/V19186.html
Cite as: [2005] UKVAT V19186

[New search] [Printable RTF version] [Help]


Dragon Futures Ltd v Her Majesty's Revenue and Customs [2005] UKVAT V19186 (28 July 2005)
    19186
    Value added tax – recovery of input tax on sales to European Union member states – whether sales formed parts of circular chains of transactions – whether sales were part of an economic activity – findings of fact

    LONDON TRIBUNAL CENTRE

    DRAGON FUTURES LIMITED Appellant

    - and -

    HER MAJESTY'S REVENUE AND CUSTOMS Respondents

    Tribunal: Dr David Williams (Chairman)

    John N Brown, CBE FCA CTA (Member)

    Cyril R Shaw, FCA (Member)

    Sitting in public in London on 10, 11, 12, 13 14, 17, 18 January, 3, 23, March, 29 April 2005.

    Andrew Thornhill QC, Penny Hamilton and James Henderson of counsel, instructed by McGrigors, solicitors, for the Appellant

    Phillipa Whipple and Sarabjit Singh of counsel, instructed by the Acting Solicitor for H.M. Revenue and Customs, for the Respondents

    © CROWN COPYRIGHT 2005
    DECISION

    CONTENTS OF THIS DECISION

    Introduction 1
    The decisions under appeal 5
    Related court decisions 8
    The scope of this decision 13
    The evidence presented 24
    The remit of the tribunal 34
    The questions for the tribunal
    Issue (a) defaulting traders 38
    Issue (b) circularity 46
    Circularity of goods 47
    Circularity of money 58
    Ring fences and ringmasters 66
    Issue (c) no economic activity 71
    The tribunal's approach to the evidence 76
    Evidence from witnesses 81
    Documentary evidence 96
    The forms of documentary evidence 101
    Burden and standard of proof 113
    The parties' general submissions 116
    General comments 121
    Selected individual deals 134
    Deal 532 135
    Deal 548 152
    Deal 572 173
    Deal 618 183
    Other deals in dispute 198
    Deals 514(a) and (b) 199
    Deals 515 and 519 200
    Deals 520 and 524 204
    Deals 521(a), (b), and (c) 211
    Deal 525 220
    Deal 526(a) 223
    Deal 527 229
    Deal 531 232
    Deal 532 233
    Deal 535 234
    Deal 538 238
    Deal 540 240
    Deal 541 243
    Deal 542 245
    Deal 548 249
    Deal 554/555 250
    Deals 570 and 571 253
    Deal 572 258
    Deals 573 and 574 259
    Deals 584 and 585 263
    Deals 591, 592, 593, 597 267
    Deal 604 276
    Deals 618, 619 and 620 278
    Deal 634 280
    Deal 638 283
    The Schedule and general conclusions 291
    Directions 296
    Schedule p1 – p5
    Introduction
  1. The Appellant, Dragon Futures Ltd ("Dragon") is a private limited company established in 2001. It conducted only minimal activities in the accounting years to 31 March 2003. It started the activities relevant to these appeals towards the end of 2003. At the relevant times it had two shareholders and directors, Carl Boraiko and David Griffiths, both of whom provided evidence to the tribunal. Its annual return for 2003 described its activities as SIC code 6523 – other financial intermediation. Between October 2003 and late March 2004 it conducted some 150 transactions. These involved buying mobile telephones or similar goods (all subject to value added tax at the standard rate on sale in the United Kingdom) in the United Kingdom and selling them elsewhere in the European Union Dragon regards these transactions, each of which it termed a "deal", as trading transactions on which it was entitled to a refund of the input value added tax ("input VAT") from the Respondents for the VAT it had paid for purchasing the goods.
  2. The tribunal uses the term "deal" for each of these transactions. As explained below, the appeals concern a significant number of deals involving Dragon, each of which became, for practical purposes, a separate appeal. The tribunal has attached a Schedule to this decision listing all recorded deals by Dragon and its findings in the appeals to date. A fuller explanation of the Schedule is set out below.
  3. During the course of the hearings, the Respondents, the former Commissioners of Customs and Excise, were merged into and became part of Her Majesty's Revenue and Customs. For simplicity in this decision, the tribunal refers throughout to the Respondents as "HMRC" regardless of the relevant date of reference.
  4. HMRC considered that, wittingly or otherwise, Dragon was involved in carousel fraud in most of its deals. It was taking part in arrangements whereby goods went round in circles, in and out of various European Union ("EU") tax jurisdictions, with the only sources of profit being value added tax ("VAT") rebates paid out by the United Kingdom or other EU governments. HMRC did not contend that Dragon knowlingly handled goods more than once, or that it was the trader through which deals were linked. But they identified the existence of a trader that in some way had defaulted with regard to its obligations under the British VAT system at an earlier stage in the chains of transactions that led to most of the deals undertaken by Dragon. HMRC also established to their own satisfaction that goods of the kind purchased and sold by Dragon in its deals had at some stage "up" the chain of most transactions been sold on by a dealer in a chain that led to Dragon purchasing the goods but also at a later stage that dealer purchased the same kind and quantity of goods in a later transaction in the same chain of transactions, thus making the deals part of a circular series of deals. HMRC also took the view that in most of the deals Dragon was not engaged in economic activities. In the view of HMRC, Dragon was therefore not entitled to refunds of input VAT. HMRC refused to refund input tax on many of the transactions. It did this by taking a series of decisions relating to individual transactions. HMRC allowed the input tax on some deals, and had not taken its own view on some of the other deals at the time of these appeals. All refusals, when made, were appealed. The Schedule lists the current position for each deal.
  5. The decisions under appeal
  6. Although listed as a short series of appeals brought together by directions of the tribunal, the tribunal has for the reasons below had to consider each of the deals separately. This decision must therefore give at least some separate consideration to each of the deals. It is therefore first necessary to identify the deals now under consideration.
  7. Dragon presented its case as an appellant that was scrupulous in keeping a record of each deal. They were known by serial numbers starting with 501 (on 15 October 2003) and ending with 645 (on 16 April 2004). It later transpired that some of the deals listed under a single number had to be subdivided and others had to be combined. For example, deal 554/555 is one deal while deal 526(a) and 526(b) are separate deals. The result is less tidy than at first appears. Nonetheless, the tribunal uses these deal numbers, as modified in this way at the hearings, in its directions and decisions, including this decision.
  8. HMRC refused the input VAT to Dragon on most, but not all, of its deals. In total this amounts to over £15 million of input VAT withheld from Dragon. But it took HMRC some time to form a firm view of the deals, and it therefore processed the deals in a series of decisions. Most were adverse to Dragon. Dragon strongly disagreed with those decisions. Each time HMRC took a decision disallowing input tax on one or more deals, Dragon appealed. These were collated by the tribunal as they were made, and the conjoined appeals, as the evolving list of separately appealed decisions were termed, were initially heard together.
  9. Related court proceedings
  10. Dragon considered that the actions taken by HMRC in deciding to refuse its claims for repayment of input VAT were not in accordance with law. It therefore took proceedings in the Administrative Court to secure appealable decisions. Its case came before Mr Justice Burton, who agreed with some of Dragon's concerns. Anan order was made by the Court on 29 September 2004. This provided a timetable for the first stages of these appeals. It also recorded agreement that no point would be taken about late filing of appeals. Following this, the tribunal received no argument on issues of that kind and therefore spends no time on them. Following the order, HMRC accelerated the process of deciding the input tax claims, and the tribunal arranged an earlier hearing date. However, the tribunal records that not all decisions were made in sufficient time for proper incorporation into the current hearings, and full consideration of some of the decisions was adjourned by direction of the tribunal on the application of Dragon and with the agreement of HMRC. The tribunal has also taken account of the action before Mr Justice Burton in three further ways. First, it has sought to keep this litigation moving, following the lead given in that hearing. Second, the parties agreed that the evidence filed before the Administrative Court was to be accepted in full as evidence before the tribunal. Third, it felt it right to allow both parties to produce late evidence that was not available to that party when the proceedings were initially processed for appeal following the Administrative Court order. In some cases, as the detailed discussion below shows, that late evidence was critical to the view taken by the tribunal on individual deals.
  11. The hearings took place under the shadow of a second related court case, the hearings of the European Court of Justice on reference from the High Court of England and Wales in Optigen and Bond House, the two lead carousel fraud cases, discussed below. At the time of the final hearings of these appeals, the tribunal had the benefit of the opinion of the Advocate General in those cases, as noted below. But the decision of the European Court was still awaited. During the hearings, the tribunal directed that the appeals relating to some of the deals in dispute should be postponed to await the decision of the European Court. This decision therefore covers only those deals that are not postponed. Further, as also detailed below, the tribunal limited the scope of the initial hearing of the other appeals to questions of fact pending the European Court judgment. This is therefore a provisional decision in respect of those deals also, and may require revisiting once the decision of the European Court has been given.
  12. The hearings were also affected by a third court action. The tribunal only has limited details of this third case, which was a criminal trial taking place before Mr Justice Crane in Southwark Crown Court. That case involved, the tribunal understands, individuals in or formerly in the employ of a freight forwarding company, Hawk Precision Logistics Ltd, based near London Heathrow (referred to as "Hawk" in the hearings and this decision). The relevance to these appeals is that the freight forwarding company was one of those used by Dragon. Dragon made a number of applications to the tribunal about the extent of disclosure by HMRC of evidence relating to the freight forwarders, and relevant to deals under consideration in this decision, to have the appeals involving documentation supplied from Hawk allowed under rule 19(4) of the Value Added Tax Tribunal Rules 1986 without further consideration. This was because, in Dragon's view, HMRC had failed to disclose to Dragon and to the tribunal all that it required to disclose about Hawk of relevance to these appeals and in so doing put itself in breach of a direction of the tribunal. The direction was one of full disclosure directed by the tribunal on 24 November 2004 in advance of and reparation for these hearings. The parties sought, and obtained, release of evidence from the evidence produced at the criminal trial from Mr Justice Crane, and submitted it to the tribunal.
  13. After hearing the applications in private, the tribunal decided that the Appellant was justified in making the main application and that the Respondents were in breach of the direction. Miss Whipple, on behalf of her clients, fairly accepted that further information should have been disclosed. But the tribunal noted that the direction was widely drawn. It decided, in the exercise of its discretion under Rule 19, that none of the appeals should be allowed without further consideration but that all the evidence produced by the parties, and the submissions made by the parties, should be reconsidered on the basis that they were treated as forming part of the full hearing of the appeals. Accordingly, all evidence produced was admitted as evidence to the appeals. The tribunal also heard the Appellant's application for costs of the applications, but left the determination of the application over for later consideration. In considering the applications, the tribunal noted in particular that some of the evidence produced from the papers before the trial included express references to Dragon that were in documents of a class not produced or identified to the tribunal, as well as evidence about individuals whose names appeared on the Hawk documentation. The tribunal indicated that it would take not only the new evidence but also the circumstances of the production of these documents into account in assessing the evidence on individual deals and that it would treat with caution evidence produced from Hawk files.
  14. The tribunal gave reasons to the parties for its determination of those directions. Dragon's solicitors raised comments about those reasons, with copies to the Respondents, who did not respond. Those comments were considered by the final private meeting of the tribunal (in the absence of all parties) about its decision, and the points made were noted and taken into account by the tribunal in this decision. For the avoidance of doubt, in reaching this decision the tribunal considered all the evidence and submissions put before it in the applications as part of the totality of the appeals.
  15. The scope of this decision
  16. Case management directions for the oral hearing of these appeals were given in a series of directions, the most important of which was on 24 November 2004. It was an aspect of these directions that was the subject of the Rule 19(4) application. At that time it was assumed that the tribunal would be asked to look at all those deals undertaken by Dragon for which HMRC had refused an input VAT rebate, in so far as that was appropriate, ahead of the decisions of the European Court of Justice and following the order of Burton J in the Administrative Court. It was also assumed at that stage that the tribunal would confine itself at the hearings to looking at a number of the deals as examples of the kinds of transactions into which Dragon had entered. This did not prove possible. Nor did it prove possible, as noted below, even to identify specific deals as examples of the issues between the parties. Each deal had to be considered separately and in some detail.
  17. The tribunal has noted above that it considered a number of applications for directions arising because of court actions related to these appeals, and that these have resulted in both the expansion and restriction of the scope of the appeals as they have been heard. A further complication must also be noted at this stage. HMRC sought the assistance of several foreign tax authorities in assembling information about the chains of transactions in which Dragon was involved. HMRC were also interested in other stages of those transactions, but this is not directly relevant to the tribunal. Inevitably, some of that information took time to arrive, and some arrived as the appeal was being considered. Such evidence as arrived during the proceedings was accepted by the tribunal, but not all the evidence sought by either party was available to the tribunal. Further, Dragon raised a number of points about that evidence that HMRC was not able fully to answer within the timeframe of the hearings. Dragon also objected to late service of some of the evidence. As noted above, this timetable followed the order of the Administrative Court, the timetable in and the considerations behind which order the tribunal itself followed. The result was that some of the appeals proceeded to hearing before all the evidence sought by HMRC was available. The tribunal decided to proceed on the evidence to the extent it was available, and to take a relatively liberal approach to admitting late evidence. In particular, the tribunal was offered evidence from a number of foreign tax authorities. It is fully aware of the extent to which those before it had limited control over the time in which, and the form in which, one national tax authority responds to the request for information from another national tax authority. The tribunal has not taken into account any evidence, if there is any, received since the final date of hearing.
  18. As a result of various directions given by the tribunal, further consideration of a number of the deals under appeal stand adjourned, with liberty to both parties to apply for further directions on any adjourned matter. The deals standing adjourned are listed in the Schedule below. The tribunal has not heard detailed argument on any of these deals, nor given detailed consideration to the evidence of those deals. As with the deals on which decisions have been taken, however, the tribunal was supplied with information about each of the deals by Dragon and of the decision if any taken by HMRC on that deal. The tribunal has noted each deal, to that extent, as part of the context for reaching decisions on the deals currently under active consideration.
  19. As the appeals proceeded and the further evidence was admitted, the parties were able to reassess the bases of their cases. But the parties were unable at any stage to form a common view about any deals being prototypical or typical so that the tribunal could take a decision with regard to examples of deals only. Having heard and seen the evidence, the tribunal considers that the parties were right about this, although that it is a counterintuitive result that the tribunal only reached after hearing much of the evidence. While it might be expected that there would be recurring patterns that the tribunal could identify, those patterns proved to be of limited value in reaching a balanced decision on the questions to be answered about each individual deal. As a result, each of the deals has led to separate consideration by the tribunal and each therefore in reality constitutes a separate appeal.
  20. This presents the tribunal with an evidential problem that it considers below. It was originally asked to consider each deal as one of a series of deals taking place in a short space of time and often involving common parties. It is now asked to examine some only of those deals isolated from the evidence about others of the deals and in particular isolated from the immediate temporal context in which some of the deals took place. The tribunal considers it important to recall in connection with its consideration of each deal the context within which that deal took place even though the full evidence about other deals occurring at the same time is not formally before it for decision. The state of the evidence about some deals was also such that the tribunal had to take some views about deals not directly before it in order to make sense of the evidence of the deals that were before it. Any such views are, however, subject to fuller consideration at a later stage and are not to be taken as final.
  21. The tribunal encouraged the parties to identify common ground, and both were able to do so. In a decision dated 19 January 2005, this tribunal allowed Dragon's appeal in respect of deals 514a, 514b, and 527a. In each case HMRC conceded that it could not establish that the deals had the characteristics of a carousel fraud. HMRC further conceded that it could not make its case in deal 531 during the oral hearings, because it could not fully identify a defaulting trader, and in a further decision dated 29 April 2005 the tribunal allowed Dragon's appeal in respect of that deal in similar terms. Both decisions were given with consent and without full reasons.
  22. During the course of the hearings, Dragon conceded that circularity of goods was shown in respect of deals 538 and 574, that is, those deals did have the characteristic that goods were shown to have gone round a circle. In other words, the Appellant accepted that the evidence showed that the goods that Dragon bought and sold were at some stage sold to a trader from whom they were at some earlier stage purchased in those deals, but no others. But Dragon did not concede that this prevented Dragon claiming the relevant input VAT. The tribunal accept this concession and find circularity in deals 538 and 574 without further consideration. It makes findings on the other aspects of those deals below.
  23. The tribunal was invited by the parties to draw conclusions from these concessions. Having considered the matter, it does not draw any general consideration of strong evidential value from these concessions. It records only that together they establish that circularity of goods was present in some deals involving Dragon and that HMRC to that extent had reasonable grounds to investigate Dragon deals more generally. But at the same time the concessions by HMRC remind the tribunal that the case must be made by HMRC, on which the burden of proof rests, in each deal. The tribunal also has in mind the evidence about each of those individual deals when reviewing the context of the other deals to be decided.
  24. As a result of the various directions made by the tribunal, decisions on many of the deals were postponed or adjourned for later decision. This decision makes a final decision on the issues of fact for decision in the following deals only:
  25. 515, 519, 520, 521(a), 521(b), 521(c), 524, 525, 526(a), 527(b), 532, 535, 538, 540, 541, 542, 548, 554/555, 570, 571, 572, 573 (subject to a partial concession by HMRC), 574, 584, 585, 591, 592, 593, 595, 597, 604, 618, 619, 620, 634, and 638.

    The decision in each of those deals is set out in the Schedule.

  26. The deals standing adjourned, subject to liberty to apply, and not fully heard are listed in the Schedule as "part heard". The tribunal draws attention to one part deal in that list, deal 539(b). During its full consideration of the pattern of deals at the conclusion of its considerations of the case after the hearings, it appeared to the tribunal that that part deal had dropped "below the radar" of the tribunal and parties (to use a phrase used in evidence on a number of occasions). To avoid doubt, the tribunal regard this as a part-heard appeal.
  27. The deals not in either of the above lists or mentioned in the preceding paragraphs are deals not currently involved in the appeals before the tribunal. In some cases this indicates that no appealable decision has yet been made. In others it indicates that HMRC have not challenged the claim for repayment of input VAT. As the matters are outside the jurisdiction of the tribunal on these appeals, the tribunal makes no findings about them in detail, although it notes the broad nature of each deal from the schedules of Dragon deals provided by the parties. The Schedule lists all deals recorded in the evidence presented by the parties to the tribunal, and for the sake of completeness notes deal numbers on which it has no information and which appear not to have been used by Dragon. The parties gave the tribunal no evidence about any of these deals and the tribunal assumes, in the absence of any other evidence, that these were either unused numbers or uncompleted deals.
  28. The evidence presented
  29. Following the direction of the tribunal in November 2004, the burden of proof was on HMRC to show reason why the input VAT should not be repaid to Dragon in respect of the deals. HMRC therefore presented the case, with Dragon responding.
  30. HMRC offered statements of evidence from several of its officers about relevant policy issues and the collation of various of the documents in the papers. The tribunal was offered full evidence of provenance of the documents produced by HMRC. The tribunal likewise accepts that evidence as establishing where the various documents produced by HMRC were obtained. The main issue taken by Dragon with the evidence produced by HMRC was that it was not a full disclosure. The tribunal returns to that issue below.
  31. The tribunal was also given statements of evidence to substantiate the identities of the individual companies named in the list of defaulting traders below. Dragon did not take issue with that list and, subject to its separate decision for deal 531, the tribunal accepts it as proved..
  32. Oral evidence was given over several days by Peter Richard Birchfield, an officer of HMRC with oversight of the investigations that resulted in the decisions under appeal. He took the tribunal through the detailed documentary and other evidence available in respect of individual deals under consideration. Oral evidence was also given by David George Atkin, Dragon's VAT officer and one of the officers involved in that investigation in support of the evidence given by Mr Birchfield. Oral evidence was also given by Michael Waddington. This related to a specific point of evidence about a conversation with Dragon's witness Mr Simmonite.
  33. The tribunal found the evidence of Mr Atkin to be particularly helpful. He clearly had direct involvement with the Appellant and also with the paperwork put before the tribunal, and the tribunal considered that both his oral evidence and the schedules of written evidence that he produced were based on thorough and conscientious efforts on his part and of those working with him, presented to the tribunal very openly. By contrast, the evidence of Mr Birchfield seemed for the most part to be at one or more removes from the actual processes of investigation and sifting of the information, and was to some extent evidence of the approaches taken by more senior members of HMRC to the evidence (including approaches to the evidence that it was thought necessary to produce to the tribunal to establish the case for HMRC) rather than evidence of the deals themselves. However, the approach to the evidence is a matter on which the tribunal forms its own view. On some factual issues critical to the decisions of the tribunal it was therefore the evidence of Mr Atkin, rather than that of Mr Birchfield, that was of particular help. Nonetheless, the tribunal records the thanks it gave Mr Birchfield at the hearing for spending some days in the witness box steering the tribunal through the many volumes of paper in front of it.
  34. Written and oral evidence for Dragon was given by David Thomas John Griffiths, now resident in Dubai, one of the two shareholders and directors of Dragon Futures Ltd at the relevant times, and its trader in many of the deals. Mr Griffiths is recorded in the documentation as being the dealer in many of the deals under consideration. It was therefore, as the tribunal understands it, his decision to go forward with individual deals on the terms agreed. While the tribunal has relied heavily on the documentation produced to it of the individual deals, it took fully into account that the initial arrangements of most, if not all, of the deals took place by telephone, and that often involved Mr Griffiths personally. The tribunal was therefore disappointed that Mr Griffiths seemed to remember so little about the details of individual deals, even the unusual ones, although some involved considerable amounts of money, and considerable potential profits, and that it was in some cases only a few months since he helped agree those deals. The tribunal notes below that some of the deals were transacted in quick succession to each other. Undoubtedly, therefore, they would have been the result of quick decision-making. It also accepts Mr Griffiths' evidence that the detailed paperwork was prepared in the usual way by those in the back office rather than him. Nonetheless, the tribunal had hoped to learn considerably more about Dragon and its deals than it was able to glean from this evidence, and it found his inability to answer specific detailed questions raised, rather than answered, questions.
  35. Written evidence was given by the other shareholder and director, Carl John Boraiko. This was largely in the form of evidence already given to the Administrative Court. It shows Mr Boaiko's role as responsible for providing the finances for Dragon's operations. Written and oral evidence was also given by Brian Steven Simmonite, a director of KPMG's tax practice and an adviser to the Appellants. Written evidence was also received from Daniel Crawford Millar, who became involved in Dragon's activities in March 2003. His principal role is described as helping to devise the operating systems and procedures. His evidence contains details of the standard procedures created for Dragon. The tribunal comments further on the oral evidence below.
  36. The tribunal was presented with many volumes of documentation about individual deals, including in some cases several hundred documents for a single deal. By the end of the hearings, the papers ran to over 30 volumes of documents with at last 13,000 pages of primary documentary evidence. They came from a variety of sources available to the Respondents (including information from several foreign tax authorities), from the Appellant, from the Appellant's professional advisers, and from other proceedings in which the Respondents were involved. As noted above, both parties sought to add to this evidence on a continuing basis, sometimes by agreement and sometimes by application to the tribunal for a direction. In particular, as a result of hearings of applications by the Appellant under rule 19(4) of the Value Added Tax Tribunal Rules 1986, the tribunal directed that evidence derived from a criminal trial involving the Respondents and released to the tribunal by Mr Justice Crane be added to the evidence.
  37. The tribunal records the following general comments about the nature and quality of the documentary evidence before it. Both as regards the added documentation and as regards the original documentation, the evidence presented to the tribunal was only partly marshalled. The tribunal bore in mind, in examining the documentary evidence, the following problems in giving full weight to individual documents:
  38. (1) There was no clear chronology to most of the papers as presented to the tribunal. Indeed, little regard appears to have been paid to chronology in some parts of the papers despite time clearly being of the essence in many of the deals.
    (2) Little attempt was made to discriminate between contractual documents indicating conclusion of an agreement or specific contractual actions as against file copies held by the sending party of such documents without other indications that the document had been received, or even sent. As the tribunal observes below, a key document in all of the deals is the invoice or instruction issued by one party to another party as held by that other party. Little attempt was made in the evidence presented to the tribunal to identify those key documents from the others.
    (3) Many documents had written comments on them that could have been of relevance to the findings of the tribunal. No attempt was made ahead of the hearings by either party to explain how or by whom those documents came to be marked. Some of the comments could be given no provenance by either party even when the tribunal specifically asked about them. As a result, the tribunal has ignored unprovenanced handwritten comments on documents in assessing the evidence.
    (4) At times the massive copying operation led to documents failing to be true copies (for example, because the fax details or some other header or footer had only partly been copied, or by obscuring the fact that a copying error was in the receipt of a fax rather than photocopying the received fax). In a few deals, this has led to the loss of timing details that would have assisted the tribunal reach a clearer view about a particular arrangement.
  39. The tribunal does not seek to attribute blame for this. Nor did it seek a comparison of such documents with the originals, in part because the bundles were agreed. But it was left with the impression that neither party had either fully analysed or fully synthesised the documentation presented to the tribunal. Yet the tribunal was asked to spend some time on many individual documents. As it was forced to observe at one point, there was a recurring problem in handling this evidence of distinguishing the forest from the twigs. Nonetheless, the tribunal was presented by the Appellant with a list of individual questions and challenges to the documentary evidence produced by the Respondent for each deal. And the Respondent produced specific answers to each of those points. In forming its own view on the deals where the conclusions were less clear-cut, the tribunal considered each of these points and the answers made to them. But it does not further burden this overlong decision by reference to those exchanges unless they are of critical significance to the tribunal's view.
  40. The remit of the tribunal
  41. The tribunal heard these appeals while the relevant law was under consideration in the European Court of Justice. The tribunal therefore directed, with the agreement of both parties, that it would consider the factual issues before considering the applicable law. In a direction made on 23 November 2004, the tribunal directed that the case be set down to consider, for each deal, whether the following were or were not established:
  42. (a) that there was

    (i) a missing trader, that is a trader who sells goods and incurs a liability to VAT but then disappears without discharging that liability; or
    (ii) a defaulting trader, that is a trader who sells goods, incurs a liability to VAT but then does not discharge that liability; or

    (iii) a trader with a hijacked VAT number;

    (b) that the purchase and sale by the Appellant formed part of a chain of transactions that was circular in nature; and
    (c) that the chain of transactions was carried out in furtherance of an activity which was not an economic activity.

    Dragon submitted, and HMRC did not dispute, that there was a relevant zero-rated sale of goods in every contested deal, and the tribunal therefore makes no specific findings of fact on these issues.

  43. The basis for this framework of analysis lies in two tribunal decisions: Bond House Systems Ltd v Customs and Excise Commissioners [2003] V&DR 210 ("Bond House"), and Optigen Ltd and Fulcrum Trading Ltd v Customs and Excise Commissioners [2003] VAT Decision 18123 ("Optigen"). Both were referred by the High Court to the European Court of Justice for decisions whether the views of the tribunals were in accordance with European law. See for the reference in Bond House,the report at [2004] V&DR 125. The opinion of the Advocate General that the decisions of the tribunals were not in accordance with the First and Sixth VAT Directive had been published at the time of the hearings, but the decision of the European Court is still awaited. Following the agreed direction, and respecting the references made, this tribunal forms no views on the European law and on the correctness in law of the other tribunal decisions, but follows them in its analysis of the facts.
  44. Bond House was a decision based on consideration of evidence that was considerably more straightforward than the evidence in these appeals. The tribunal took the view in that case that, in establishing circularity of goods and the non-economic nature of any activity, the burden of proof was on HMRC and that the standard of proof was that of the balance of probabilities. The directions for these hearings adopted that approach and placed the duty of opening the appeals on HMRC.
  45. In Optigen the tribunal, on agreed facts, further analysed the issues involved in establishing a carousel fraud. The parties addressed the tribunal on these decisions and also drew its attention to the opinion of the Advocate General of the European Court of Justice in Optigen. In so far as it is relevant to the tribunal's factual task, the tribunal sets out its understanding of the relevant law below. In doing so, it bases its thinking on the more fully developed reasoning in Optigen rather than that in the earlier decision, but without repeating that reasoning. However, two points require further consideration below.
  46. The questions for the tribunal
    Issue (a): defaulting traders
  47. The tribunal can deal with this issue swiftly. It does so against the background of the role of the trader noted in the descriptions from Bond House cited below. In all deals other than those in which it conceded the point, HMRC gave evidence of the identity of the defaulting trader linked to each challenged deal made by Dragon. The Appellant produced no evidence to contest these findings. The tribunal is satisfied that in each of the deals dealt with by this decision there is a defaulting trader. There was no use of hijacked VAT numbers. The tribunal was not asked to consider, and does not make any finding about, the extent to which, or when, Dragon was or should have been aware that any particular trader was a defaulting trader.
  48. The tribunal find that the following traders are for the purposes of these appeals defaulting traders:
  49. Smart Telecom Ltd ("Smart")

    Greenday Wholesale Ltd ("Greenday")

    Furylong Ltd ("Furlyong, or F")

    Swift Online Ltd ("Swift Online")

    Phone City Megabyte Cafι

    It makes no findings on further alleged defaulting traders involved in other deals as yet part heard. The defaulting trader identified with a particular deal is listed in the Schedule of individual deals at the end of this decision.

    40HMRC gave evidence, accepted in full by the tribunal, of the way that the defaulting traders ceased to appear in the Dragon deals, to be replaced by other defaulting traders. Smart Telecom Ltd appeared only in disputed deals 515 and 519 in December 2003. (As with others of the defaulting traders, it appeared in other deals not yet fully considered. However, the tribunal regards its findings on the issue of the identity of the defaulting traders, and the way in which one replaced another, as findings that would apply, subject to other evidence, to the consideration of all Dragon's deals).
  50. Smart was the subject of legal action commenced on 19 January 2004, and it ceased to be involved in these transactions. Greenday Wholesale Ltd featured only in disputed deals 502 and 521. It was wound up on 5 April 2004. After that the defaulting trader in the next deals was Furylong Ltd (save for deal 531 which involved another alleged defaulting trader, that that is a deal in which the appeal was conceded by HMRC during the hearings). The tribunal, on close examination of the papers, noted a number of deals where it appeared that Furylong took over the role of Greenday in particular transactions at the time Greenday was wound up in some way of which the tribunal has no evidence. However, it does note that this was apparently accepted by those with whom Greenday had been dealing at the time. This suggests at the least that those dealing with Greenday were immediately aware of what had happened to it, and were content to accept whatever arrangement took place that allowed Furylong to inherit Greenday's contractual obligations. This is relevant to deals 520 to 524.
  51. Furlyong Ltd was deregistered by HMRC with effect from 4 April 2004 on 8 April 2004. The Dragon deals up to 604 involve Furylong Ltd, that deal being logged by Dragon on 2 March 2004. The disputed Dragon deals from deal 612 the following day feature Swift Online Ltd as the defaulting trader. The tribunal notes that most of the deals between those two deal numbers are either not disputed by HMRC or involve a sale to Dragon by a trader other than DD Trading Ltd, the seller in both deal 604 and deal 612. Action was projected by HMRC against Swift Online Ltd on 10 March but it was involved in deals to the end of March. The tribunal again notes the quick way in which Swift Online Ltd replaced Furylong Ltd in the chains when action was taken against the latter company.
  52. Phone City Megabyte Cafe, in deal 638 on 1 April 2004, was deregistered
  53. on 5 April. The tribunal notes the comments of HMRC's officer about the actual circumstances of this defaulter when visited by the officer. It illustrates a more general point, to which the tribunal returns below, of the geographical locations of, and circumstances of, some of the defaulting traders and other companies involved in these chains.

  54. It was argued for Dragon that the existence of a defaulting trader should not lead directly on to issues about circularity. The existence of a defaulting trader might be evidence of what was termed acquisition fraud rather than fraud involving circularity. It might be the defaulting trader alone that was fraudulent. In other words, it should not be concluded that because a particular trader proved to be a defaulter that there was any further fraud in the system, or that the tribunal should expect to find others also involved in the fraud. The general evidence from HMRC was that the main form of fraud in this field involves circularity rather than default by an individual company, and the tribunal accepts this. The tribunal also accepts the evidence of HMRC about the way in which one defaulting trader replaced another in the chains of transactions of which Dragon deals later formed part. It indicates its view below about the role of the buffer companies later in the chains involving in particular Furylong Ltd and Swift Online Ltd. While this does not indicate circularity involving a company in another EU member state, there is clear evidence that the role of the defaulting company in the chains was not limited to that company alone but was associated with other later counterparties. There is also evidence of third party payments being made by those buffer companies to companies "up" the chains from the defaulting companies, so missing out the defaulting companies, which does link to companies outside the United Kingdom. To that extent, on the balance of probabilities, the tribunal started with the view that any fraud or default was not that of the defaulting company alone and did not relate solely to the acquisition.
  55. But the tribunal draws no final conclusions about circularity or economic activity from the finding that there is a defaulting trader in each of the deals. These must be shown separately. However, the tribunal considers that the presence of a defaulting trader in circumstances such as that shown in these deals does indicate that HMRC was acting reasonably in taking its investigations into these deals further.
  56. Issue (b): circularity
  57. HMRC based their arguments in each deal on one or both of two propositions: that there was general circularity of goods established on the evidence and/or that general circularity of goods could be inferred from what was termed "circularity of money". In other words, in some deals HMRC contended that it could be shown on the available evidence about the financial transactions involved in the chains of transactions that it was more likely than not that in each of the deals in issue goods were returned by sale to at least one person who had previously owned those goods.
  58. Circularity of goods
  59. The most basic form of chain is one where it can be shown that specific goods go round in a circle. This is the prototypical carousel. The tribunal adopts the description of a carousel fraud from that given in Bond House as a description of a carousel fraud:
  60. 15 In its simplest form a carousel fraud works in this way. A VAT-registered trader ("A") on one European Union Member State sells taxable goods to a VAT-registered trader ("B") in another Member State. A's sale to B is zero-rated in A's Member State. B should declare the purchase and pay acquisition tax in its own Member State, and, upon the premise that it is intending to use those goods in order to make an onward taxable supply, then claim credit for the same amount as input tax. B then sells goods to another VAT-registered trader ("C") in its own Member State, charging and receiving VAT on the consideration. However, it fails to account to the tax authorities for that VAT and effectively disappears, it becomes what the Respondents refer to as a "missing trader". Nevertheless, as the time of making its sale to C, which is still registered for VAT and before the Commissioners are aware that it is or might become a missing trader and have been able to intervene (for example, by de-registering B) it provides a VAT invoice to C, which claims the VAT it has paid to B as input tax. C (to whom the Commissioners refer to as a "broker") then sells the goods to a registered trader in another Member State: the hallmark of the simplest fraud is that this purchaser is A, and it is this circularity that gives rise to the term "carousel fraud". C has an input tax claim but, because its sale to A is zero-rated in C's own Member State, is not required to account for any output tax. The result, if the fraud is successful, is that B has received, but not accounted for, the VAT that the tax authorities must pay to C. In this situation B is certainly fraudulent and A probably so; C may well be entirely ignorant of what is happening, and of the use which is being made of its participation.

  61. This is expressed in British VAT terminology – others would refer to exemption with credit rather than zero-rating. But it also uses the terminology used in these appeals. Here the Respondents contend that Dragon is the "broker". In place of B there is usually a chain of companies. The first is the company that is the first British buyer of the goods. This will be, in the case put by the Respondents, the defaulting trader. But between the defaulting trader and the broker will be at least one other company, the "buffer". The Respondents further contend that a purpose of a buffer is to prevent money being in the hands of the defaulting trader, so that the Respondents could not recover funds from the first buyer.
  62. In all the deals put to the tribunal with Dragon as the broker, there were at least two companies – usually more – in place of B in the above description. Perhaps for similar reasons, there were other companies between A and B and/or between A and C, but in other Member States. This might mean that the deal involved the chain of sales and purchases involving several Member States. The tribunal was told, but did not explore, that there were common factors in the Member States subjected to fraud in this way. For example, the tribunal was told that the fraud works more easily in Britain than in some other states because in Britain the input tax could normally be reclaimed more quickly by the broker, as compared with the payment of the VAT to the collecting authority by the defaulting trader and buffers, than in some other states. More generally, the tribunal accepts that in circumstances where there is fraud in a chain of transactions, the slower the reclaim procedure then the lower the potential advantage to someone intending fraud because in part there is a higher risk that any fraud is discovered.
  63. In what was termed in Bond House as the simplest fraud, there would be evidence that goods had gone round in the circle A-B-C-A. Do those goods have to be identical? Does it have to be shown that the same specific goods that went A-B also went C-A? Conversely, is it enough in resisting the allegation of circularity that, as the Appellant contended in part, it can show that at no time were the goods on any C-A sale the same specific goods as the goods on the A-B sale? Again, the tribunal adopts the Bond House vocabulary:
  64. "23 In their efforts to detect and prevent such frauds, the Commissioners have developed three concepts, which they term "specific circularity", "general circularity" and the "ring fence". These concepts were developed in detail in some of the statements produced to us, and in an annex to the Respondents' statement of case. However many parties may be involved, the ringmasters' need to control the carousel requires that the totality of the transactions remain within what the Commissioners term a "ring fence". Within the ring fence, there may be instances of the same goods reaching the same trader more than once; the Commissioners refer to that as "specific circularity". Where there is an identifiable series of linked transactions, but it cannot be shown that the identical goods reached the same trader more than once, they say there is "general circularity". That arises particularly where consignments are split and combined."
  65. In these deals, the Respondents contend only general circularity. The Appellant answers the contention of circularity robustly in another way. All mobile telephones – the goods in most of the deals – have IMEI (international mobile equipment identification) numbers. That is, each of them has a unique international identification number. Further, it is a number in bar code form that can be scanned easily into data recorders. The appellant's case is that every one of the telephones in every one of the deals handled by Dragon had its IMEI number monitored as standard practice. And the clear evidence, the Appellant contended, was that Dragon did not deal in the same telephones twice. The submission was that Dragon itself, therefore, could not be the link in the circularity of any chain.
  66. The tribunal had produced to it the lengthy lists of monitoring of the telephones, together with evidence from William Ian Tuffy, an officer of Revenue and Customs confirming those lists, and from Peter Birchfield about the significance of the lists. Mr Tuffy gave evidence that he had checked a file of IMEI numbers against a monitoring list of IMEI numbers maintained by HMRC. Mr Tuffy's evidence was that of the 43,201 numbers checked, 39,057 appeared in the table once , 4,008 twice , and 136 three times . On this basis Dragon pointed out that a few of the telephones only had been through the monitoring of HMRC on a previous occasion.
  67. An identity of IMEI numbers would establish an identity of goods. But the Respondents did not point from this evidence to specific circularity of goods in any specific deal. The tribunal accepts Dragon's submissions that this evidence shows that the Respondents could not establish specific circularity of goods from this information, and that the Appellant took positive steps not to deal in the same telephones twice. HMRC's evidence was that nonetheless a significant number of the individual telephones traded through Dragon had been "circulating" through the UK in different transaction chains prior to their purchase by Dragon. The tribunal notes from the figures in evidence that this was true for about 10 per cent of the telephones involved in Dragon deals. It agrees with HMRC that this shows that circularity was not merely a potential risk but also a continuing reality in the market in which Dragon was dealing. It also notes Mr Birchfield's view that: "it is also highly unlikely that as our recording of IMEI numbers is either on a random basis at freight forwarders or from data supplied by other broker traders that the same telephones would not have reappeared at Dragon without someone ensuring that this was the case." The tribunal has no firm evidence to conclude that this is "highly unlikely" but does have it in mind as at least a possibility. The tribunal comments below in connection with individual deals about the placing and removal of Customs seals on various consignments. The tribunal would anticipate that these might also be the deals in which random IMEI numbers were monitored by HMRC officers or their foreign equivalents. But it was given no specific evidence by HMRC about those individual searches or about the way in which IMEI numbers would be taken in such circumstances.
  68. Subject to this point, the tribunal is concerned in these deals, using the above terms, with general circularity of goods only. And it makes the finding that the Respondents have not established that Dragon was engaged, whether knowingly or otherwise, in specific circularity of goods in any deal.
  69. HMRC contended instead that there was general circularity in the same way as contended in Bond House and Optigen. To establish general circularity, HMRC contended that goods of the same kind could be traced by evidence of a series of transactions in each of the deals from the ownership and sale by a trader outside the United Kingdom back to a purchase by and ownership of those goods for a second time by that trader. The test, as put in the reference to the European Court of Justice in Optigen is to show that "the same goods are circulating within the ring fence of the carousel through linked transactions." The tribunal emphasises that by "the same goods" it takes this to mean by reference to the nature and quantity of the goods, not the specific identity of individual items.
  70. Dragon's challenge to the application of this approach to the facts was essentially one of carefully noting the details and challenging all the weak links. Save for the three deals in which Dragon conceded that circularity (but not non-commerciality) was shown, it contended that HMRC had failed to establish the circularity on reliable evidence.
  71. The primary basis for establishing circularity was therefore by documents showing the transfer of specific kinds and quantities of goods through a series of transactions that involved the same kind and quantity of goods travelling around the circle. Where HMRC could produce evidence to support its submission of circularity by documentary evidence of the supply chain, the transaction was referred to as one of "circularity of goods" in the papers before the tribunal. The tribunal examines below the evidence before it in those deals in which it was contended by HMRC that general circularity of goods was shown directly in this way.
  72. Circularity of money
  73. In some deals, HMRC was unable to show on the evidence a chain of transactions in goods that directly established that goods originating with a VAT-registered trader in another EU member state went back to that trader through a chain of transactions involving Dragon. But it was contended that the same result could be shown by what was termed "circularity of money". In some deals the contention was that financial transactions of which evidence was produced showed, when read with the details of transfers that were in evidence, that on the balance of probabilities there was general circularity. For example, in some deals HMRC could show by documentary evidence the transactions by which goods came into the UK from elsewhere in the EU but with the sender unidentified. The chain showed goods being transferred on from that company through others, including Dragon, to a VAT-registered trader elsewhere in the EU. The evidence of payments linked to those transactions showed at the same time that one of the UK companies that had been in the chain was responsible for making a third party payment to the VAT-registered trader to which the goods were sent after being subject to a deal involving Dragon. This, it was contended, showed circularity because it showed a payment being made from the UK to the non-UK trader who received the goods in a later part of the deal. The assumption was that the payment was for an earlier stage in the deal.
  74. The Appellant was strongly critical of this approach to the evidence, and submitted that it did not establish any form of circularity. At the general level, it was a case simply of third party payments being made. As such that proved nothing. There might be a number of explanations why payments were made to third parties rather than directly between buyers and sellers at particular stages in a chain of transactions. Dragon asserted that such payments were usual in trades of the sort in which Dragon was involved, and its witnesses repeated that assertion. The payments might or might not be related to that chain of goods. They might, for example, be payment of a different debt to the third party. Or they might not be payments of any specific debt at all. In other cases they might be payments down the line rather than up the line. That is, they might be payments ahead to the ultimate foreign buyer the other side of the Dragon deal. Dragon also made the point that the evidence of a request for a third party payment was not always supported by evidence that the payment had been made.
  75. The tribunal agrees with the Appellant's arguments against looking at the financial transactions as relevant only to a limited extent. In its view the label "circularity of money" is not helpful unless the contention is that funds can be shown by evidence to go round the carousel in the same series of deals as the goods (although presumably in the opposite direction). In the simplest series of transactions, the tribunal would expect to find that a transfer of goods from A to B then to C then to A to be matched with a series of transfers of money from B to A, then from C to B, and then from A to C, if only so that it could be shown that VAT invoices presented each stage by one party to the next party had actually been paid.
  76. The tribunal did not have that level of detail before it. Rather, there were, as the Appellant contended, a number of deals where there was evidence of third party payments that looked at in isolation could either reinforce the evidence of a circular series of transactions or alternatively offer secondary evidence of a circular series where the primary documentary evidence was not available (or, perhaps, there was no such documentation).
  77. The tribunal does not agree with the Appellant's general contention that evidence of payments is inadequate as a basis for establishing circularity in a case and should therefore be discounted. It notes that Dragon itself did not take part in any of the third-party payment arrangements. And it considers that the inherent legal and financial risks of third party payments (including fraud between those involved) would lead parties that were genuinely operating at arm's length to take great care when using such methods of payment. It looked to see whether there was such evidence of care and awareness of risk in the documentation about third–party payments put to it. It finds that in some deals the documentation did not suggest the level of safeguard that it would expect to find when large sums of money were moved to third parties in genuine arm's length circumstances, especially when those movements occurred across national and currency frontiers. It agrees that the evidence is of less weight than direct evidence of the movement of goods. But it is relevant evidence and it can assist in establishing a link as contended by HMRC when taken in the fuller context of the total available evidence about a particular series of transactions, and with only the proper weight given to the financial details.
  78. For example, it might be shown (as was contended by HMRC was the case in some of the deals) that:
  79. (a) the series of transactions was from the UK first buyer through three or four buffer companies to Dragon then to a buyer in another EU state,
    (b) one of the UK buffer companies instructed another of the UK buffer companies to make a payment of the largest part of the sums due to it on that deal to a third party rather than directly to it,
    (c) the third party was the same as the buyer in the other EU state later in the series of transactions,
    (d) the payment was made to that third party so as to satisfy obligations within the series of transactions to the buffer company that asked the other buffer company for the third party payment,
    (e) it could not be shown from documentary evidence that the buyer in the other EU state was the seller to the first UK buyer directly or indirectly and
    (f) the evidence was such as to suggest that the payment was properly linked to the series of transactions in that way and not to some other transaction.

    In such cases, if the details – invoice numbers, amounts, times of transactions and payments, and so forth – were consistent then the tribunal would accept in principle that it could be established on the balance of probabilities that the payment to the third party was evidence that it had been in the chain at an earlier stage to the date on which it received the goods after Dragon's involvement.

  80. This is, however, a matter of considering evidence of any unusual financial arrangements as part of the total evidence available and applying the probability test to the evidence as a whole. In so doing, the Appellant is right to indicate that there may be several other explanations for what happened, and that "circularity of money" was not shown by a single third party payment. But at the same time the tribunal bears in mind that third party payments can give rise to significant additional business risk, and that this may conversely indicate that there are other arrangements in place between those involved in the payments that remove or dilute that risk.
  81. The tribunal therefore rejects the approach of HMRC in categorising deals as shown to be "circularity of goods", "circularity of money" or both. It does not adopt the same classification of the deals as HMRC. In each case it must ask the identical question: is it probable in the light of all the evidence that the same party appears both "up" and "down" the chain from Dragon? That is the test it applied in deciding whether circularity was shown.
  82. Ring fences and ringmasters
  83. Linked with these issues are those of identifying a "ring fence" and a "ringmaster". Dragon contended that the evidence did not show any ring fence or ringmaster, and that in the absence of those features of the transactions the deals could not be shown to be circular or non-economic. The language comes from past decisions. The tribunal did not find those concepts assisted its exploration of the evidence in these appeals greatly, and indeed found the mixed metaphors used served to confuse more than clarify. Whether "carousel" is a reference to a circular conveyor belt for baggage at a transport terminal or what used to be called a merry-go-round, carousels do not have ringmasters or ring fences. The image of the circus ring invokes horses or other animals being put their paces in a circus ring rather than the artificial horses or suitcases on a carousel. This is not mere banter. The tribunal must look through the attractive but – in its view – essentially unhelpful phraseology to the underlying issues. HMRC at no time accused Dragon of being complicit in any fraud. It did not contend that Dragon was either a "ringmaster" or subject knowingly to a ringmaster. Nor was it contended that Dragon was aware that it had been ring-fenced. The Appellant contended that there were no ring fences, that the Respondents could not show the presence of a ringmaster, and that all the transactions were in the open, unfenced and unmarshalled market place.
  84. Behind these phrases were serious issues of another kind – business risks. Conducting a carousel transaction in cases such as the deals before the tribunal involves moving goods of considerable value across international frontiers in exchange for transfers of money of equally considerable value. Whenever that happens there are a number of risks. Not least is the risk that goods will be passed on without a matching payment back being secured. It may be assumed therefore that those involved will take action to minimise the risks.
  85. As the tribunal sees it, the action that attracts the epithet of "ring fence" in this context is the need to prevent leakage of value from a circular transaction. If the goods do go round in a circle, genuine value added must be injected at some point for the transactions to have commercial content. The allegation, if a ring is truly circular, is that the value is obtained – and obtained only - by national tax authorities contributing more in VAT refunds to the transactions forming the ring than is received from the members of the chain or circle contributing in the other direction. The converse to this is that this is the only injection of new funds, so value must be prevented from leaving the circle if the maximum advantage is to be obtained from the VAT refund. In addition, an innocent third party is only likely to enter such a ring if there is profit or value to be realised by that third party. Equally, each other person in the chain will want a reward. From where can the rewards be realised? If the transactions are truly circular, the total value added at all stages of the transactions will equal the net VAT gained from the national tax authorities. So it is necessary to minimise not only the risks but also the outlay in acquiring goods to be taken round the carousel to offset the limited scope for profit. It follows that those organising the carousel should stay in control of those responsible for the transactions.
  86. On that basis, the tribunal does not see the necessity for a "ring fence" to establish circularity unless there is some specific reason to ensure specific circularity of goods. In these cases, there was no specific circularity. If goods being circulated are, in effect, non-specific goods specified only to a specific number, value, and generic description, there is limited need for a "ring fence". There is need only to control the business risks. The "ringmaster" is a metaphor for what in direct tax cases has been referred to as the circumstances or individuals establishing a preordained series of transactions. Again, there is need to control the business risks. But there are more ways of controlling those risks than just a "ring fence" and "ringmaster" approach. In the view of the tribunal, it needs only certain well-placed performers who can operate in a coordinated way either side of a gain of VAT into a transaction from a public authority to ensure that sufficient value is captured to add value to the entire chain of transactions.
  87. The approach taken by the tribunal is therefore to look at the business reality of the chains of transactions of which the Dragon deals form part. In so doing, it looks at the reality of the business risks to Dragon and the other parties. That links to the third of the questions asked of the tribunal.
  88. Issue(c): no economic activity
  89. The question for the tribunal in each deal is: did the chain of transactions take place as part of an activity that was not an economic activity.
  90. This question reflects the fundamental nature of value added tax as a tax on added value from the activities of taxable persons "acting as such". As defined in Article 4 of the Sixth Directive this includes any person carrying out any economic activity. Again, this issue was discussed thoroughly in the Bond ouseHou House decision and further developed in Optigen. It was put this way in Bond House:
  91. 154 Once it is found … that Bond House's transactions did not contribute in any proper sense to the process of production or distribution, but merely played a part in the circulation of chips round a circle which had nothing to do with their distribution to the final consumer, just as repeatedly circumnavigating a roundabout does not contribute to the traveller's journey, it seems to us an inevitable conclusion that they were not transactions within the contemplation of the Sixth Directive; they were not "economic activities" …

    The tribunal does not consider it necessary to set out the analysis that led to this conclusion, or that in Optigen, at length, although it is of course aware of the views of the Advocate General in the case on this issue. For the purposes of this decision it is necessary only to look at the factual question: was there economic substance to the deals undertaken by Dragon when looked at in the context of the chains of transactions of which they were part? In considering this, the tribunal follows the approach in Bond House and Optigen of looking at the whole sequence of events in each case.

  92. This must be answered, as with the previous question, in the light of all the evidence.
  93. While considering the individual deals, the tribunal formed the view that for the purposes of analysing these deals, there was a link between the existence of circularity and the question whether a chain of transactions was "economic activity". It considered first that the absence of circularity in a deal, or at the margin the absence of circularity from a significant part of the goods involved in a deal, meant that the approach taken by the tribunal to the assessment of business risk in a deal did not apply. In the absence of circularity, it might be that there were other risks. In addition, there was the opportunity of value to be added to the transactions otherwise than from those within the circle. The tribunal therefore took the view that, on the balance of probabilities, if it could not be shown that there was circularity, then the probability was that it could not be shown that there was no economic activity. As, strictly, the tribunal does not have to make a finding on economic activity where no circularity is found, and it heard no argument on this conclusion, it records no finding of fact about economic activity in those deals where it finds no circularity of goods.
  94. Conversely, the tribunal takes the view that the existence of circularity is a strong indicator that the chain may not be an economic activity. If there is in reality no business risk to a series of deals because they are coordinated in such a way that the usual risks of the market are absent, and also there is no scope for value to be added to the chain by those involved in it save by the injection of public funds derived from input VAT refunds paid to one or more counterparties to the chain, and there is no "consumer" in view in the series of transactions, then there must a question whether those activities can be said meaningfully to be commercial or economic in the context of value added tax.
  95. The tribunal's approach to the evidence
  96. the tribunal has set out its approach to these deals by reference to the individual questions asked. Bringing the focus of the three questions asked of the tribunal together, the tribunal is concerned to establish from the evidence whether or not the nature of the separate chains of transactions of which each disputed Dragon deal was part were or were not chains of transactions such that there was general circularity of goods in a context where there was no economic reality to the transactions.
  97. Each series of transactions can be examined by reference to a number of different criteria to establish its circularity and economic reality. The tribunal's approach to its task can be illustrated by reference to an example. Say that it is alleged by HMRC that a particular purchase and sale of mobile telephones by Dragon was part of a circular chain of transactions with no economic reality to it. The chain starts with a trader ("T") elsewhere in the EU that sells, through a foreign buffer company ("F1") also elsewhere in the EU, to the British first buyer ("B1"). B1 sells through a series of other British buffers ("B2, B3, B4"). B4 sells to Dragon. Dragon sells to another foreign buffer ("F2") who sells on to T.
  98. If the chain of transactions is genuine and each step takes place in the market place, then each deal will be at arm's length. And the goods have to be moved from T's ownership and possession to that of F1 and then to the UK to the ownership and possession of B1. They then move to B2, B3, B4 and then to Dragon. Dragon must have ownership and possession at the relevant time so that it can make the deal to pass the goods to F2, from whence they go back to T. For this to take place in conditions that entitle the parties to value added tax rebates:
  99. - title must pass along the chain so that each party is in a position to

    sell the goods on to the next party;

    - the goods must actually be moved into and out of the United Kingdom

    at the relevant times;

    - VAT invoices must be presented by the sellers to the buyers at the relevant times;

    - payment must be made on those invoices;

  100. In addition, for business risks to be minimised and for the transactions to make commercial sense:
  101. - each party must ensure that it has title to relevant goods before it passes title on to the next party;

    - each party must ensure that it receives or has a guarantee of payment

    before releasing title to the goods;

    - each party will take steps to minimise its trading risk, for example by ensuring it has an onward buyer before it purchases the goods for onward sale and by proper insurance arrangements;

    - each party will wish to make a profit out of the general activity if not

    from each individual transaction.

  102. The tribunal considers it important to take account of how Dragon saw itself as involved in the chains of transactions. There were certain common features to all its deals, whether disputed by HMRC or not, and the tribunal was told of the systems established by Dragon to control each deal.
  103. Evidence from witnesses
  104. The tribunal received written evidence of the business approach of Dragon from two witnesses who gave oral evidence and in witness statements from two others actively involved in the business. Daniel Millar, one of the original members of the group that established the business, set out in his witness statements the systems for internal and external due diligence established for checking counterparties to the various deals undertaken. This included standard paperwork and checking procedures, particularly with regard to VAT registrations. Mr Millar accepted however that these procedures had not been applied to dealings with a foreign purchaser called Mobile World. The evidence was given to the Administrative Court as part of its proceedings and is accepted by the tribunal.
  105. Mr Millar's evidence includes the verification procedures established for DD Trading Ltd ("DD"), the seller to Dragon in several of the deals now in question. Mr Griffiths confirmed in oral evidence that the approach in this case was first from Dragon to DD . The tribunal cannot help but notice that the documentation from DD Trading attached to his statement, like most of the DD Trading Ltd documentation, bears a fax header indicating that it was sent from a fax machine of Teleos plc. The tribunal assumes this to be the same Teleos plc as has been engaged in legal actions against HMRC, including that decided by the Court of Appeal at [2005] EWCA Civ 200, and a business clearly actively involved in this commercial area. This would in the tribunal's view raise a number of due diligence questions. Dragon appears to be aware of Teleos plc at that time as it asked HMRC to check its VAT number at the same time as it asked for a check on DD Trading. But the tribunal sees no comment in the evidence of the checks performed about why DD Trading should be using someone else's fax machine even when making its introductory communications with Dragon. Nor was any evidence given to the tribunal by Dragon about its interest in Teleos plc. The tribunal draws no conclusions from this save that it would have expected this issue to have been explored by Dragon. It might have helped Dragon take the view that DD was a legitimate commercial concern genuinely engaged in the same market place as Dragon, and it might have suggested that there were market opportunities for Dragon to deal with Teleos plc directly.
  106. There are also details in that evidence of the checks undertaken against DRT Vertreibs, including checks with both the German and Danish VAT authorities. These show that DRT identified Dragon as a possible counterparty as a result of noting a web site, and the terms of an agreement between the parties similar to that entered into by DD Trading. Further details about DRT (both Danish and German offices) are set out in the witness statements of Mr Simmonite.
  107. The tribunal was not offered, and has not attempted, a legal analysis of the paperwork set out in this statement and clearly used and to be used by Dragon in its trading. The tribunal notes that the counterparties in some of the deals also use standard forms and conditions. Nor is it readily clear which set of forms apply to any particular deal as there appears to be much scope in the way in which the paperwork was handled in individual deals for what lawyers term the "battle of the forms" - and also potential "battles of jurisdictions". It is also not clear whether the oral agreements to buy and sell were in law the operative agreements, with the paperwork acting as confirmation rather than as the contractual documentation. But the tribunal accepts that a serious attempt was made by Dragon at the initial stage to establish a deal-making regime that included both checks on counterparties and a legal framework within which to handle the deals. This includes the requirement that there be a purchase order from a customer to Dragon and a purchase order from Dragon to the supplier, faxed confirmation of instructions to inspect goods for Dragon and faxed instructions to the freight forwarder.
  108. The evidence from Mr Boraiko explains the market and financial backgrounds to the operations established by Dragon. This confirms Dragon's awareness that there was fraud in the market in the kinds of goods in which Dragon decided to deal and its decision to seek both advice from HMRC and from professional advisers about the issues and precautions it should take. It also confirms its ambitions to become a major player in the field. Mr Boraiko gave a number of statements in evidence to the Administrative Court, and the tribunal accepts that evidence save in so far as it indicates otherwise in this decision.
  109. Mr Griffiths gave oral evidence. He was Dragon's trader, and came from a background of trading and broking finance. He explained that Dragon has sought to enter a buoyant market in goods that were steady products with an aim to make profit on size of deals rather than being innovative. He considered the market at that time to be an active global market. But he was aware of the risks of VAT fraud. Dragon had sought to work with its solicitors and business advisers as well as with HMRC to reduce its exposure to the fraud. Dragon had by the beginning of 2004 access to major funding, and could use that to give it a market position. But this was high cost capital and the pricing of sales was aimed at a 5% to 3.5% margin to cover the cost of capital. He aimed to arrange back to back purchases and sales in each case. He did not look beyond the immediate seller to, and purchaser from, Dragon when making the deals.
  110. With regard to the actual transactions, Mr Griffths confirmed that once oral deals were struck, paperwork was handled by the back office. The aim was to get things done on the same day if possible, but they had to watch time differentials and banking cut off times and use fax machines. Dragon worked on a ship on hold basis.
  111. The documentation that mattered was the purchase order and the pro forma invoice.

  112. As commented above, the tribunal did not find Mr Griffiths to be a helpful witness on the details of any deals, and the level of detail he could recall was disappointing. The general tenor of his evidence was that the appeals related to past events and that many of the details of those past events, even though relatively recent, either had originally been for others or were matters he could not fully recall. In particular, he was unable to give much specific evidence to confirm that the systems of which the tribunal was given details by Mr Millar were in fact carried out in individual deals once trading started. The paper work, he told the tribunal, was a back office issue. He was also unable, under cross-examination, to comment on why most of the deals in Nokia 7250i telephones, all of which were purchased by him or others in Dragon in the UK, were in the two-pin variety of chargers when the UK used three-pin chargers. He assumed end users for the telephones but was not concerned with more than the sale on by Dragon. He resisted suggestions about whether Dragon should have cut others out of the chain, or itself be cut out, on the basis that he only had the suppliers he had. He also was not involved in arranging inspection of the goods, and could not help with details about an inspection report that suggested in appropriate language guides for some goods, or others where parts were missing. Nor could he remember any goods being rejected after inspection. He similarly was unable to comment on most other specific points put to him in cross-examination about individual deals.
  113. Mr Simmonite, a director in KPMG's tax practice and formerly an officer of Customs and Excise, gave oral evidence of his role in helping Dragon from September 2003 as part of the KPMG support to Dragon. He confirmed KPMG's involvement in setting up due diligence systems for Dragon and in checking individual counterparties, and he outlined ongoing advice to Dragon. KPMG had also carried out its own investigations into the integrity of deals undertaken by Dragon once it became clear that HMRC were questioning some of the deals. He considered that Dragon had done everything within its powers to avoid being involved in fraudulent trading.
  114. Mr Simmonite was critical in his evidence of the records kept by two freight forwarders, Hawk and Interken Freighters (UK) Ltd ("Interken"). The papers on some of the files were very messy and difficult to deal with. He contended in particular that he had been informed by Mr Waddington, an officer with HMRC, that Interken records were not reliable, but qualified this in oral evidence to refer only to papers relating to a particular matter. When questioned about third party payments, he agreed that some payments did suggest third party payments linked to specific sums in transactions, but he did not agree that the evidence was that the payments were payments back up the chain rather than forward down it. Nor did he agree that it was likely that such payments would be payments back. The tribunal found Mr Simmonite's evidence careful and relevant to assessing the way in which it was initially considered that Dragon should go about its role in the market.
  115. The main oral evidence for HMRC was given by Mr Birchfield. More specific oral evidence was given by Mr Atkin of HMRC, the Dragon VAT officer. The tribunal has commented above on its views of the evidence of these officers. In his statement to the Administrative Court Mr Atkin stated that in each of the 123 deals underlying Dragon's repayment claims in December 2003, February 2004 and March 2004, the transaction chain in the UK started with a defaulting trader. The tribunal has indicated above that it accepts that there were defaulting traders in all the transactions before it. He also contended that onward sale for all those transactions was to seven direct EU customers only. Of those seven, two were the German and Danish branches of DRT. The only others relevant to the deals now being considered by the tribunal are:
  116. Mobile World Gmbh (Germany)

    Sunico AS (Denmark)

    S R Communications (Germany)

    Arkridge Trading (Ireland)

    The other direct customer was Tocado Holdings BV. For reasons not relevant to this decision, the tribunal accepted an application to adjourn further consideration of all deals involving Tocado Holdings BV, and therefore has not looked at these deals.

  117. Specific third party payments were identified by the HMRC witnesses in all but 35 of the 123 transactions at the time the statement was made, and enquiries were continuing into the others. But none of the HMRC witnesses was able to help the tribunal about the provenance or meaning of the annotations on many of the third party payment faxes or the references to third parties in connection with those payments that were not specifically involved in any of the parts of individual deal chains evidenced to the tribunal.
  118. In oral evidence, Mr Atkin updated his evidence, in particular with regard to the evidence obtained from requests for help from foreign tax authorities. No evidence had yet been received about the companies referred to as Mobile World or Portland Better, although there was information about DRT. However, this had come in after decisions had been taken, and was not part of the thinking of HMRC in taking those decisions.
  119. Mr Atkin lso explained how he had put together information obtained from the Interken files as a check on transactions involving Interken as a freight forwarder, and equally how he had assembled information about third party payments. He agreed that Dragon had kept him informed of its activities and that in his view it appeared that it was making meticulous efforts to meet the requirements of HMRC on due diligence. He was happy with the quality of Dragon record-keeping, although he had not undertaken a full VAT check of the documents at that stage. He had had only one meeting with Mr Simmonite, although he was comparing notes with KPMG's investigations when he looked at each deal. But he did not rely on KPMG information as he considered he should form his own view. His evidence related to specific deals is taken into account in connection with those deals.
  120. The tribunal found consistency in the approaches of the written evidence of Mr Millar and the evidence of Mr Simmonite and Mr Atkin, about the awareness of Dragon that it was entering a field where there was a significant known risk of fraudulent trading, about the need for careful systems to minimise risk of direct involvement in that fraud, about the need to have and keep full documentation, and about keeping HMRC informed. But none of those individuals was actually involved in making out the deals.
  121. Documentary evidence
  122. The central evidence in these appeals is therefore that of the operative documents by which the deals took place or were confirmed. In particular, a key document in any deal is the sales invoice (or similar documentation) issued by the seller to the buyer and provided by the buyer. And in any chain, the key lies in a series of such documents. The tribunal takes this view for several reasons. In high value transactions it would expect all parties trading genuinely to keep careful records of what was agreed and also about whom had title over what. By contrast, documentation that was less than rigorous could suggest that the deals were between associated parties rather than at arm's length. Dragon was aware that there was fraud in the market, and knew that it was expected to keep its records in good order to protect itself. And Dragon's own evidence was that it was ensuring prompt documentary backup to its transactions.
  123. All the deals were fast-moving transactions in high-value goods in which Dragon held no stock. Although the way in which the business was conducted reflected some aspects of futures markets, this was not futures trading. It was trading in actual goods by matched deals. Dragon could only sell when it had title from those selling to it. By doing that, it minimised its exposure and business risk. It therefore had to ensure, as did the parties elsewhere in the transaction chains, that it had title to sell specific goods under specific contracts before it sold them and also that it had an insurable interest in goods being transferred by freighters under its instructions.
  124. Linked with this, it is clear that much of the business was conducted by exchanges of faxes. This enabled these document-based transactions to be conducted at the relevant speed. And much of the documentation was designed for faxed exchanges and replies. But for that reason the tribunal would expect to see the relevant documents showing that they had been faxed and, following standard practice, also showing by whom, to whom and when they were faxed.
  125. These appeals concern value added tax liabilities and entitlements between registered traders. Therefore all transactions are required by law to be evidenced by invoices in the proper form. In addition, all those involved would be aware that any party claiming input tax for goods moved to another European Union state was required to produce documents to show that the goods had left one member state for another when it said they had done so. The additional element in these transactions was that the goods were handled throughout by third parties and not by the buyers and sellers themselves. Not only did Dragon not hold any stock, it did not handle any stock either.
  126. The tribunal therefore consider it right to look to the documents produced to tell the story of each deal in its entirety. If both parties have been able to produce all the relevant documentation (and assuming that the documentation existed), the tribunal expects to see some of the documents in multiple copies held by the buyer, the seller, and those responsible for handling the goods.
  127. The forms of documentary evidence
  128. Evidence before the tribunal came from several sources. These included the following forms of documentation.
  129. (1) Deal documents supplied by the parties
  130. These documents are, in the tribunal's view, the best evidence where available. It particularly looked to a deal document originating with one party but in the hands of the other party. If there were copies held by both parties, then this was clearly strong corroboration. If the documents had been faxed from one party to the other, and the unfaxed copies held by the send matched the faxed copy held by the recipient, then the tribunal felt confident in taking that transaction as confirmed. The tribunal was supplied with the documentary records of Dragon – and it accepts that it has had full disclosure from Dragon. It was also supplied with documents assembled by KPMG while acting as adviser to Dragon. This included information about others in the chains in which Dragon found itself, and the tribunal accepts that there would be obvious reservations in third parties producing some of that information in the absence of an order or direction for production. And it was supplied with information uplifted by HMRC from other traders, as well as the information obtained by HMRC separately mentioned below.

  131. The tribunal has noted some general concerns about the quality of some of these documents and the ways in which they were presented to the tribunal. While it bears those concerns in mind, nonetheless the tribunal considers that the provenance of these documents is accepted. Members of the tribunal individually undertook examination of the documentary evidence on a number of the deals. From these analyses of the evidence, the tribunal formed the general view that (apart from obvious miscopying) the documents should individually be accepted as accurate copies unless when read with other documents discrepancies emerged. In particular, the tribunal found that – bar a few cases – the fax headings on documents should be assumed to be genuine and that numbers, times and dates shown should be taken at face value. But it also found that there were documents of which it should have been given fax details on the copies but where faulty copying had accidentally obscured the evidence. It does not find any such obscuring to be deliberate, and therefore merely notes the missing evidence.
  132. The tribunal was unable to find adequate explanation for the pencilled annotations on many of the documents. Some individual comments were identified to one party or the other, but many were in handwriting not recognised by any witness nor any of those representing either party. Particularly puzzling were references to third party payments to unidentified recipients of which there was no other evidence. The tribunal took the view that these annotations should be ignored unless for some reason they called the original content of a document into question, in which case it would be safer to ignore the document as a whole.
  133. (2) Freight forwarder files
  134. This evidence was particularly controversial. Indeed, Dragon applied to get deals involving Hawk documentation struck out on the grounds that HMRC were aware of problems with evidence from Hawk and had failed to comply with the tribunal's directions in informing the tribunal and Dragon of all it knew about the Hawk documentation. The tribunal has recorded that it found that HMRC, although not the Dragon team, were in breach of its direction in respect of the information given about Hawk. It did not strike out the deals. But it did admit all additional evidence offered to it about Hawk, and has treated the submissions and evidence about the application and about Hawk as part of the main hearing of the appeals. It recorded in its directions that it saw no reason to reject the evidence from the Hawk freight forwarder files simply because it came from the Hawk files. But it does treat with particular caution the documents generated by Hawk itself, such as the covers of Hawk files. And it notes Dragon's allegations that documents may have been removed from files or wrongfully placed in them.

  135. The new evidence was taken from papers then available to the criminal trial taking place before Mr Justice Crane at the time of the oral hearings in these appeals. This suggests that HMRC did have evidence of direct to some of the deals in issue in these appeals. That evidence was of two kinds. First there were documents mentioning movement of Dragon goods that were not disclosed by HMRC. Examples are the daily stock sheets for 21, 25 and 27 November 2003, all showing goods being held related to Dragon deals. The tribunal found this evidence interesting, because it is not immediately clear, when those dates are compared with the dates of the deals recorded in the Schedule, to which of the recorded deals these movements of stock referred. There is separately more general evidence of doubts about the reliability of both Hawk documentation and also named Hawk officials. There was also evidence that it was thought that Hawk was aware of the fraud in the market in which it was involved and statements that it was thought that individuals had actively assisted organisers, for example by suppression of documents.
  136. The tribunal must therefore proceed on the basis that there may be further documentation that would assist its findings of fact about deals involving Hawk which it has not seen for one reason or another. And it must proceed on the basis that there may be false documents presented to it. But much of the specific evidence related to periods earlier than those of relevance to these appeals, and the tribunal confirms its earlier view that the Hawk freight forwarder files can be viewed as part of the total evidence.
  137. Turning to the other main forwarder used in the deals, Mr Simmmonite gave evidence that Mr Waddington had told him that HMRC did not trust Interken evidence. As noted above, in oral evidence he confined this to doubts about specific transactions. Mr Waddington did not recall making any general comment about Interken. He accepted that he had taken action to close part of the Interken premises on one occasion, and he explained the details (which are not specifically relevant to these appeals). The tribunal finds no strong reason to pay any less attention to the Interken files than to the Hawk files, while also bearing in mind that for the following reasons this evidence should also be treated with care.
  138. The tribunal also heard evidence that the freight forwarder files were often in a muddle, and that – for whatever reason - documents that might have been expected to be there were not. The evidence in individual deal files put to the tribunal confirmed this evidence. But the tribunal does not accept that these problems should lead to the exclusion of evidence simply because it is obtained from a freight forwarder or a file containing irrelevant documents. In each case, the tribunal looked for corroboration (for example the copy of a document on the freight forwarder file found on the file of the sender). And it also looked at timings, as most of the documents to and from the freight forwarders were sent by fax.
  139. (3) Information from foreign tax authorities
  140. HMRC sought information from several foreign tax authorities about the foreign traders in the chains, including in particular the parties with whom Dragon dealt directly. Some of the information sought by HMRC had not been received at the time of the oral hearings. Some was received belatedly (and admitted late), but did not form part of the evidence considered by HMRC. The evidence also varied in the extent to which it was supported by original documentation or explanation. Detailed information was received from the Danish, German and Irish authorities and this has been taken into account in looking at individual deals. Some of that information is in detailed form and provides information both about onward sales from the foreign purchaser from Dragon, and also about third party payments. More generally, the tribunal notes specific comments of the German tax authorities on S R Communications, and that it warned S R Communications to change aspects of its activities to comply with the VAT requirements on it at times prior to those when Dragon was dealing with it.

  141. (4) Financial and other third party evidence
  142. Evidence was produced by both parties in the form of schedules and other documents to trace payments into and through third party bank accounts. Quite rightly, the methodology of and underlying evidence behind these schedules was tested in evidence. The tribunal had some trouble in following some of the schedules, but in general found them helpful in dealing with individual deals.

  143. More generally, the tribunal is satisfied that there is sufficient evidence in some of the deals for it to be established that third party payments were not only directed but also made, and that these payments appear in some cases to be payments that meet the tests the tribunal itself outlined above as being payments that indirectly suggest, on the balance of probabilities, that circularity of goods occurred.
  144. Burden and standard of proof
  145. During the hearings, the parties returned to the issue of the burden and standard of proof. Dragon resolutely took the approach that it was for HMRC to establish any proposition. It also argued that the evidence required to meet the agreed standard of proof must be particularly persuasive. This argument was based in part on the tribunal decision in Hindforce v Customs and Excise Commissioners (2004) (VAT Decision LON/04/241) where the tribunal accepted that the standard of proof on HMRC was a high standard. It was also based on the view that the case for HMRC involved the accusation that someone – though not Dragon – was engaged in fraudulent or negligent conduct and that this should not be assumed lightly. HMRC made it clear that they did not allege fraud against Dragon at all. They also resisted the attempt to shift or gloss the basic standard of proof. But they accepted that the burden of proof was on them.
  146. The tribunal's own view is that the burden of proof must be on HMRC to establish matters where HMRC, but not Dragon, have power to investigate and enquire. The most obvious example of that is where HMRC ask, through the proper national authorities, for assistance from another tax authority in the European Union. But, equally, HMRC can ask for, or uplift, information from third parties such as freight forwarders, that Dragon has no power to obtain. That was the basis on which the tribunal issued directions in November 2004, based on the consideration of the issue in Bond House. However, the tribunal comments that it sees no reason in law why the burden of proof should be so firmly on HMRC where the matter is one within the particular knowledge of Dragon. At the same time, there being no allegation that Dragon is fraudulent, the key aspects of the transactions that call their commerciality into question, would normally be expected to be in parts of the transactions in which Dragon was not directly involved. The tribunal has also recorded in connection with receipt of the evidence that it accepts that it has had full disclosure from Dragon and those advising Dragon.
  147. Unless the facts are agreed or the tribunal indicates a higher degree of acceptance of the evidence, the decision in these appeals is based on the balance of probabilities. The tribunal sees no reason to gloss the direction of or shift from the usual standard in these matters. It notes that in some of the submissions HMRC itself appears at times to be working against a higher standard. For example, the tribunal has taken the broad view that it is more likely than not that the fax markings on most of the documents are accurate as to date, time and telephone numbers recorded. It notes that HMRC withheld reliance on some of these details in constructing their cases because no direct evidence of the accuracy of the fax machines could be produced. If the tribunal was acting on the criminal standard of evidence, then this would need further consideration. The tribunal also takes the view in these appeals that it can take account of all the documentation produced before it as sent between particular traders in forming a general view as to the probability of accuracy of individual documents. Again, in some cases HMRC presented its case as based only on some of the evidence, while the tribunal has felt it can look wider. But where there is a doubt in a specific case, this is recorded and conclusions drawn.
  148. The parties' general submissions on the evidence
  149. In opening its case, counsel for Dragon submitted that there were recurring flaws in the HMRC approach to the evidence. That submission was made both in general terms and also by reference to individual documents (and gaps in the documentation) in individual deals. The criticisms of the freight forwarder evidence have already been noted. It was accepted on behalf of Dragon that some of the inconsistencies were removed during the hearings and also that some omissions were made good. The tribunal also note that HMRC withdrew some of its contentions on a number of deals during the hearing and consented to the appeals being allowed. But the tribunal accepts the submissions of the Appellant that there are gaps in the full chain of evidence in many of the deals and that it should be alert to the fact that HMRC has made assumptions in some of these deals that it should itself test against the standard of probabilities.
  150. In closing the case, it was submitted for Dragon that, save for the two deals in which circularity was no longer contested, HMRC had failed to show either circularity or absence of economic purpose in any of the deals before the tribunal. It further contended that the evidence showed that Dragon had throughout acted with integrity and thoroughness and it invited the tribunal to make findings that this was so.
  151. For HMRC it was contended in opening, and maintained throughout in argument, that there was no accusation of fraud against Dragon or any named person within it. Further, the direction under which the tribunal was hearing the case did not call for any findings about whether Dragon or individuals within it were, or were not, acting with integrity. Miss Whipple stated that on behalf of her clients she would be making no submissions on the matter, nor leading any evidence about it, nor seeking to challenge any of the Dragon evidence on that basis. The tribunal indicated that it would confine its proceedings to the set questions on the basis that no fraud was alleged. Despite the invitation from Dragon that the tribunal should draw conclusions about its integrity and honesty in approaching the deals, the tribunal makes no findings of fact about this aspect of the approach of Dragon to the deals.
  152. Consistently with that, the tribunal also refused to entertain assertions made during the hearing about questions of the good faith of those producing evidence for HMRC. It recorded that no such assertions were maintained against counsel for HMRC or those instructing counsel, and it refused to hear such issues about any other individual involved in HMRC without proper notice. Those issues are also therefore not the subject of any finding by the tribunal.
  153. More generally, the tribunal is of the view that this is a case where the documentation has to "do the talking", with conclusions being based primarily on whether the documentary evidence shows, or does not show, circularity or economic activity. And that requires a separate view to be taken about each deal.
  154. General comments
  155. From the above evidence, Dragon entered the mobile phone and similar high value electronic goods markets in September 2003. Its business model was designed and intended to allow those involved to realise higher profit margins than in the markets in which its members had previously operated but using skills from those markets. Dragon would also use its available sources of finance to inject capital into transactions in high value goods treated as commodities. The business model was intended to provide a trading context where Dragon had minimum risk exposure so that it could safely both service the borrowed capital and make adequate profits. It approached the market with caution, taking into account professional advice it obtained and in consultation with HMRC. It was not disputed that Dragon aimed to do this not only in full compliance with the law but with the aim of avoiding being caught by the fraud of others. Those involved in Dragon were fully aware that they had chosen to operate in a market where national tax authorities were alert to the existence of significant tax fraud. Further, Dragon was warned directly about the risks of trading with individual potential counterparties.
  156. Dragon did not avoid entanglement, wittingly or otherwise, with transactions that involved – on the balance of probabilities – fraud or VAT evasion elsewhere. As Dragon did not dispute, there were defaulting traders in all the deals in dispute before the tribunal save where HMRC conceded otherwise. As noted, Mr Atkin's evidence was that there were defaulters in at least 123 of the 150 or so deals that Dragon undertook. And there may be defaulters in others not yet decided. That is a high proportion of the total deals undertaken by Dragon. That of itself made it reasonable for HMRC to examine all Dragon's deals in detail.
  157. The evidence of the IMEI numbers was put to the tribunal by Dragon as evidence of the extent to which it went to ensure that it was not involved in circular trading. It also shows that even with full safeguards in place Dragon still dealt with previously circulated telephones to a significant extent (some 10 per cent of telephones traded). Dragon accepted that it had never cancelled a deal because of the results of its inspection regime. And, as the tribunal discusses below, it is by no means clear that in individual deals there were checks in place to ensure that the IMEI scans were of the telephones actually released to and by Dragon.
  158. It is clear from several deals that there were goods handled that had been subject at some stage to having Customs stamps placed on the packaging, and that there was considerable sensitivity among traders to the placing of Customs stamps on goods, and there is evidence that Dragon was aware of this. Why? For example, in deal 548 there were customs stamps put on the goods on the way in by HMRC, as reported to Dragon by A1 Inspections. (No direct evidence was given to the tribunal by HMRC about this). Dragon sold to DRT Vertriebs, whose terms of trade demanded unmarked goods. The inspection undertaken for DRT Vertriebs did not note any marks. What happened to them? And why were those terms of trade made and accepted? Customs marks were also removed from the goods in several other deals, and in other deals again the goods were not in the original Nokia packaging.
  159. As that point illustrates, the tribunal saw evidence in a number of deals that those making the trades for Dragon did not abide by the careful processes it established to avoid dealing with fraudsters, known tax evaders or those with questionable records.
  160. For example, it was accepted in evidence by both Mr Millar and Mr Griffiths that Dragon did not follow its own pre-vetting procedures when dealing with Mobile World GmbH ("Mobile World"), even though it first traded with Mobile World (based in Germany) in early December. Why not? The tribunal was told that this was because it was one of the biggest dealers in the market place. Why did that justify Dragon dropping its guard when it dealt, as Mr Atkin's evidence and that of Mr Simmonite both confirm, with so few separate customers? Was Dragon aware that when it stopped dealing with Mobile World directly it nonetheless continued to deal with it indirectly via Arkridge Trading Ltd ("Arkridge Trading") (based in Ireland)? Mr Boraiko gave evidence that on 9 February 2004 he had a conversation with a member of HMRC in which it was confirmed that it was believed an earlier deal involving Mobile World (deal 519) was circular. Dragon decided not to deal with Mobile World. Dragon's last direct deal with Mobile World was on 6 February (deal 525). Its first indirect deal with Mobile World, via Arkridge Trading, was the following day, 10 February (deal 532), one of the first deals it did with that trader. The information obtained from the Irish tax authorities shows that
  161. Arkridge Trading had only two customers for the goods it purchased from Dragon: Mobile World and S R Communications. And Dragon was dealing directly with S R Communications at the same time as it was dealing with Arkridge. It also knew that Arkridge was buying for a third party for delivery to the German city in which Mobile World operated. Did it not look further?

  162. There is another recurring inconsistency between the business plans of
  163. Dragon and its advisers and what actually happened. Deal 548 is an illustration. The business plan included a 100 per cent IMEI and other inspections of goods on each deal. The tribunal was told that the industry average was 10 per cent inspections. Undoubtedly inspections occurred. But what was inspected? Both the model documentation and the documents used on the basis of the model, and the inspection reports put in evidence, lack the vital linking detail in making full sense of this model. For it to work, specific goods had to be identified for inspection and those goods – and only those goods – had to be the ones sold to and by Dragon. But most of the deals are not specific deals. Rather, it appears to be enough that goods of the required description, quantity and quality were released to and by Dragon. Some of the instructions to the freight forwarders, for example, direct the release "from stock" rather than of specific goods. There is nothing in the paperwork showing clearly (by, for example, pallet or box number) that the goods inspected were also the goods sold on. And it is clear that the warehouses were at the relevant times holding large stocks of some of the goods traded, and that they were holding much larger stocks for the suppliers than were being sold on down the chain to Dragon. The tribunal cannot on the evidence before it exclude the possibility that in some of the deals the goods inspected for Dragon were not identical with the goods delivered by Dragon to its customer.

  164. The tribunal also note in one or two cases that documents in freight forwarder files said to have schedules of pallet numbers or seals attached do not have those schedules attached in the documents before the tribunal. While most such schedules are present, and this is by no means a common point, it does raise questions, in the light of the further evidence about a freight forwarder introduced on the Appellant's application, about whether in those cases the documents were deliberately removed fromthe papers and, if so, why. And that must raise the possibility that this might link in with any substitution of goods between inspection and delivery. However, the tribunal saw no positive evidence of this (nor would expect to do so).
  165. There were also timing inconsistencies in a number of deals. The business model was for back-to-back purchases and sales of identified and inspected goods with deals being followed promptly by payment and then delivery (or delivery on hold pending payment). Following Dragon's business model, it was to make no forward sales, uses of options, sales on credit, unmatched purchases or speculative sales matched later by purchases. And, although it described itself on its notepaper as "Import/Export Specialists" there is no evidence from either party of any deal that involved Dragon importing anything or selling anything save by sale to a trader elsewhere in the European Union. (Strictly, none of these deals involved exports either, as all the deals involved transfers of goods within the European Union, but the term was used very loosely in most of the documentation seen by the tribunal and may therefore also be used wrongly on occasion in this decision.)
  166. Timing was to be of the essence of the deal model. But the evidence before the tribunal indicates that this tidy, risk-minimising model was not observed as strictly as the model required. Dragon contended that these timing defects put the question of fax times and dates in question, and that it was the dates not the existence of the deals that was unreliable. The tribunal accepts that in some cases it is clear that dates and times are inconsistent and some must be wrong. Mr Birchfield accepted as much in his evidence of some documents. But the business model used by Dragon and others made heavy use of the fax machine to ensure a combination of accurate documentation and rapid action. The counterparties themselves needed accurate settings on their fax machines to ensure that deals had actually taken place. To that extent, the dates and timings were themselves part of the business model, and they would be equally important to any other genuine trader, or a trader seeking to appear genuine. But evidence shows that even Dragon itself did not keep to those timings. Again, deal 548 is an example. From the documentation it appears that Dragon ordered an inspection of goods before they were appropriated by the supplier to the contract. That should have been known to the inspectors if not to Dragon. The tribunal accepts that if the goods were nominally shipped on hold or appropriated to Dragon, it need not wait until the goods were released before organising the inspection. But how could it organise an inspection of goods not formally identified for its purchase?
  167. The tribunal therefore has a number of concerns about the extent to which the Dragon deals in question matched the intended approach of Dragon to the market.
  168. The tribunal was offered a full comment on the documentation of each deal by both parties. And it heard extensive evidence form Mr Birchfield and Mr Atkin about individual deals. Combined with the tribunal's own concerns, this evidence was such that, it having been established that all the disputed deals involved defaulting traders, the tribunal was unable to take a broad decision that Dragon was not involved in circular trades or that its activities were, as assessed against the test in Bond House and Optigen, not economic activities. Nor was it able to say about the deals as a whole, or any significant groupings of them, that HMRC had in argument removed all the doubts raised by Dragon's detailed analysis of the evidence and by cross-examination. It is therefore faced with the need to decide each deal separately. To reduce the burden this presented, the tribunal decided to test the evidence further by detailed reference to some of the deals. For the purposes of this decision, to avoid both inordinate delay and inordinate length of decision, the members of the tribunal, and then the tribunal as a whole, fully analysed all the evidence about some of the deals. Some of these were then selected for full presentation in the decision. All the other deals were then examined with the benefit of the conclusions drawn from that detailed analysis. As will be seen below, the result was an inconsistent picture. This confirmed the tribunal's view that the parties were right not to agree to "typical" examples of deals. The tribunal sets out what it found in selected individual deals in this decision in detail before turning more briefly to the others.
  169. Finally, the tribunal collated its findings into the attached Schedule of individual deals. This put each deal back into its context in the full series of Dragon deals. And it gave the tribunal a final opportunity to look at each deal in the context of all the evidence before it. The tribunal turns to that Schedule after its comments on individual deals.
  170. Selected individual deals
  171. 134 The following individual deals are presented in date order.
  172. Deal 532
  173. 135 This deal is contended by HMRC to show both by reference to documents about the sales of the goods themselves and by reference to payments made that there is general circularity of goods. As HMRC contended that there was both direct and corroborative evidence of circularity, it was one of the strongest cases presented by HMRC save, in the view of the Appellant, for one feature. It also involved Hawk documentation, that is, documentation produced from the freight forwarder files of Hawk Precision Logistics Ltd's warehouse facilities near London Heathrow. This was called into question by the Appellant in its applications during the hearings. If the appellant's doubts were right, then this would suggest a general weakness in the HMRC case, at least where Hawk was the freight forwarder. The deal has also been mentioned above because it was the first deal of a closely timed series of deals involving an Irish purchaser, Arkridge Trading Ltd ("Arkridge"), that declared that all its purchases were for a third party elsewhere in the European Union and so a triangular deal for VAT purposes. Evidence was given of the deal by Mr Birchfield. Save for that, there was little mention of this deal at the oral hearing but, as with all deals, the tribunal had detailed comments from both parties about the documentary evidence in their final submissions. The tribunal does not attempt, in this or any other deal, to note every one of those many detailed comments separately. But its sets its findings and considerations about this deal in more specific detail to indicate its approach.
  174. 136 Deal 532 is logged by Dragon as a purchase and sale of 1,000 and 2,500 Nokia 7250i mobile telephones ("7250i"). They were purchased at a unit price of £166.50, and sold at two prices, £173 and £174. This was consistent with all the deals that the tribunal saw. In each the pricing was of each telephone or other item, with prices usually being rounded to the nearest £1 but sometimes to 25pp. Pricing was always in £sterling. The defaulting trader was Furylong Ltd ("F"). The freight forwarder was Hawk. The input VAT claimed by Dragon and refused by HMRC was £29,137.50 for the sale of 1,000 units and £72,843.75 for the sale of 2,500 units.
  175. 137 To put this deal in context, Dragon logged 6 deals (528 to 533) on 10 February (a Tuesday). Five were with Arkridge, and four of those deals, and the other deal, all involved the same kind of mobile telephone, 7250i. All were purchased from DD, at a unit price of £166.00 or £166.50. They were sold on at unit prices of £173, £174 or £175.
  176. 138 There is clear evidence on file that 7 pallets of the telephones were transported for Mobile World to the UK to be held by Hawk on the night of 8/9 February 2004 (Sunday/Monday). The 7 pallets were a rather unusual load size, (loads of 2 or 4 pallets being more typical), each consisting of a standard 500 telephones in 50 boxes of 10.
  177. 139 At 15.11 (Spanish time?) on 10 February Tradewinds S/A, a company based in Spain, ("Tradewinds") instructed Hawk to allocate and release to F the 3,500 telephones supplied to it by what it terms its "main supplier", Mobile World. At 16.37 that day (German time?) Mobile World instructed that the goods be delivered on hold to Tradewinds. These documents are both on the Hawk file, and the tribunal sees no reason to reject them as otherwise than genuine. The Tradewinds comment about Mobile World is, if anything, a disclosure against interest. On 10 February (untimed) F invoiced the sale of 3,500 7250i to TM at a unit price of £163.50. The tribunal notes that this invoice has a number on it in addition to the usual invoice numbers, which is the number used as the PMN reference number later in the series of deals. There is no obvious explanation for this number appearing on this invoice other than that someone at F knew the PMN reference number and, perhaps carelessly, inserted it.
  178. 140 The same kind and quantity of goods were sold from F to TM (at £163.50) to PMN (at £164.50) to DD (at £165.50) and then to Dragon (at £166.50). The key transaction timings noted by the fax imprints on the documentation suggest, however, that things did not proceed in a very tidy fashion. The following all happened on 10 February. At 11.46 TM faxed PMN instructing it to make a third party payment to Mobile World, with the balance to TM. PMN promptly complied. At 15.11 (or 14.11 UK time?), as noted, Tradewinds released the goods (identified as from Mobile World) to F. The release by F to TM appears to take place at 17.30 (there is no other evidence on file indicating appropriation or release), but the release by TM to PMN was at 15.20, to be followed 30 minutes later by its purchase order. At 16.04 PMN released telephones (identified expressly as TM telephones) to DD. Dragon told HMRC at 15.51 it was buying from DD and selling to Arkridge. At 15.55 Arkridge faxed a purchase order to Dragon, indicating that this was a triangulation deal and that the goods were destined for Germany. At 16.56 the goods were inspected for Dragon. At that time, no goods had been properly appropriated to this deal for inspection.
  179. 141 At 17.51 DD told Hawk to ship on hold to Dragon's order but not to release as the goods had not been released by DD's supplier. This indicates that DD knew that Dragon was sending the goods abroad, so was aware of some onward details. Why did DD say the goods had not been released to it when they had been released (even, belatedly, by F at that time)? The release to DD also appears to be ahead of full payment. The goods left the United Kingdom at 22.00 that evening.
  180. 142 Arkridge sold the goods on to Mobile World at £175 and £176 (a £1 mark up) and they were delivered to Frankfurt. Like nearly all the transactions elsewhere in the European Union in these deals, all transactions took place in £ sterling even where the currency of both counterparties to a deal was the Euro. A fax from Mobile World to Hawk on 11 February at 12.49 releases what the tribunal find to be the same or directly equivalent 3,500 7250i to Tradewinds. That raises further questions, but they need not be pursued here.
  181. 143 Also on 10 February (though the fax date and time are not copied fully) TM faxed PMN to pay £631,000 to Mobile World for the 2,500 and 1,000 Nokia 7250i. The unit prices are stated as £180 and £181 per unit (stated to be the unit price plus VAT – but at what rate? Was this a rate other than the UK rate? The tribunal heard no evidence on this). The fax followed a standard form adopted between TM and PMN that the tribunal saw many times. There are three copies of this fax in the papers, one stamped "paid", and the tribunal accepts that this instruction was made and subject to action by PMN. Bank details confirm payment by PMN to Mobile World for the 3,500 7250i at the unit price stated on 10 February (confirmed at 16.26) and 11 February (confirmed at 11.59), the payments totalling £631,000. Confirmation that DD paid PMN followed at 13.59 and 15.23 on 11 February. So PMN was asked to pay, and paid, Mobile World before receiving funds from DD. There was no evidence of payment to Tradewinds or to F. The tribunal accepts that this is clear evidence both of a payment "up" the chain – that is, to a previous trader – and that this suggests that both F and Tradewinds were deliberately cut out of the payments. Together with the evidence in the other deals examined, that would suggest that those involved were expressly avoiding putting money into Furylong Ltd. That also suggests that the arrangements from Mobile World to at least TM were preordained, confirming other suggestions of links in the evidence.
  182. 144 Several other aspects of the paperwork followed behind the transactions. For example, Dragon agreed the sale to Arkridge before receiving a reply from HMRC about its query about the VAT number, and it agreed the purchase from DD before receiving its required supplier declaration back signed from DD. According to the available documentation, DD only accepted Dragon's purchase at 13.11 on 11 February, which perhaps explains why it allowed the goods to be shipped on hold only.
  183. 145 The tribunal has little hesitation in concluding that HMRC have established circularity of goods in this deal, by reference to contractual and shipping documentation and supported by the detail of the third party payment. There is confirmatory evidence of the onward movement of the goods to Mobile World from the Irish tax authorities, which the tribunal also accepts. It is clear in the tribunal's view that the 3,500 7250i originated from Mobile World, and proceeded, by the foreign buffer of Tradewinds through the British chain F-TM-PMN-DD to Dragon, and thence by the foreign buffer Arkridge back to Mobile World and on again to Tradewinds. In that context, the tribunal is also satisfied that the third party payment was a payment made to Mobile World for TM in respect of these goods, confirming the link.
  184. 146 The evidence includes several indications that parties, including Dragon, were aware of more than the immediate transactions in the chain. The tribunal is satisfied that there are clear links between all the counterparties from Mobile World to PMN. In addition, the consistent £1 margins, the slack documentation, third party payments and other payments out of transaction order, and erratic timing in the F-TM-PMN-DD chain indicate strongly that, as in other deals (see deal 548 below), the deals between those companies were not at arm's length. Consequently, the tribunal is satisfied that the "ring" or non-market chain also includes DD. In this case DD knew that Dragon was sending the goods abroad, and Dragon knew that Germany, not Ireland, was the destination of the goods it sold to Arkridge (as Arkridge told Dragon, and Dragon informed Hawk of this). And, of course, the freight forwarder knew about the chains on both the inward and outward legs. On that basis it would seem probable that the only party not privy to at least some part of the circularity was Dragon.
  185. 147 The pricing policy is also one the tribunal noted repeatedly in other deals. There is a steady unit-based mark up of the goods as they pass from counterparty to counterparty. This steady mark up usually also involved the price at which the goods are sold on to Dragon. This suggests either that the goods were sold to Dragon on a "take it or leave it" basis in terms of the price, or that the counterparties fed the price agreed with Dragon's back up the chain. For example, in this deal each of the British counterparties before Dragon made the same £1 mark up. That could only happen if either Dragon purchased from DD at DD's price, or if, once the price at which DD sold to Dragon was known, the unit price of the sale to Dragon was factored back up the chain. The documentation makes the latter unlikely in this deal (and in most other deals) . Indeed it does not appear chronologically possible in this deal, for all the untidiness.
  186. 148 The prices were always in sterling, were always unit-based and rounded, often to the nearest £1, and were never fixed by reference to the value of a deal as a whole. The conclusion that the tribunal draws from this is that on the balance of probabilities Dragon was buying goods in this and similar deals at an offer price presented to it by the supplier with little or no scope for negotiation. That would also explain why Dragon bought similar products on a single day at variant prices, and why prices for identical quantities of identical kinds of goods sometimes varied significantly within a short period but without any general trend, although it does not explain why Dragon was able to sell them on at a range of differing prices. The picture that emerges is less one of the open market place than of Dragon being offered its place in a chain, even if it was not aware of the full implications of this.
  187. 149 The tribunal is satisfied that the evidence of knowledge beyond the immediate purchaser and seller, combined with the obvious absence of an arm's length relationship between Mobile World and the British chain, plus the open actions of both Arkridge and Tradewinds as intermediaries show not only that there was circularity but that the deal was a clear carousel with no real economic content. It has been seen that the prices charged by individual counterparties was under some sort of control or system. But the steady mark up policy demands that value be added to the chain at some point or points. Where else could value be derived in this series of deals but from anticipated VAT refunds from the national tax authorities? How else could the transaction costs of moving the goods from Germany to Britain and back (nominally through a legal series of transactions also involving Spain and Ireland) be funded and justified in a commercial market? The goods were, as in other deals, manufactured in Hungary and of central European specification, with 2 pin plugs and with no obvious UK consumer market. And in this case the evidence suggests that the goods, having been sent back to Mobile World, were promptly put back in circulation again.
  188. 150 The absence of economic activity in this deal is further confirmed by the evidence of Mr Atkin that the profit made on the circuit was equal to the VAT refund to F, £100,143.75. He schedules this as follows:
  189. Recipient of profit £
  190. TM 3,500 (£1 mark up)

    PMN 3,500 (£1 mark up)

    DD 3,500 (£1 mark up)

    Dragon 25,250 (£8.50/9.50 mark up)

    Arkridge 3,500 (£1 mark up)

    Mobile World 17,500 (£5 mark up)

    F and others 42,006.25 (the payment by PMN

    to TM less the TM profit)

    less from TM -612.50 (VAT on TM's mark up)

  191. Total 100,143.75
  192. Although the tribunal is not entirely persuaded by the entirety of this Schedule, it agrees with HMRC that enough of the arithmetic of this schedule is too neat to be explained by a coincidence of market forces. It makes sense only as a shareout of the input VAT refund, and sharing a VAT refund is not, in the view of the tribunal, an economic activity. And it makes sense only if the business risks are highly controlled.
  193. 151 More generally, the tribunal concludes that this suggests that individual deals should be approached on the basis that third party payments can evidence circularity of goods; that Hawk documentation (that is, documentation found on Hawk files) is not unreliable simply because of its source; that Dragon's deals are to be judged by what it actually did and not by the plan it intended to follow; and that there is a clear association between the other counterparties (both in the United Kingdom and elsewhere) that is also seen in a considerable number of the Dragon deals. That association is, at least in some deals, of a kind that could be qualified by the epithets "ringmaster" and "ring fence". It also confirms that the evidence before the tribunal does establish, at least in some cases, that HMRC was right to withhold the input VAT refund.
  194. Deal 548
  195. 152 The tribunal examined this transaction in detail because it is one of those relying only on indirect evidence of circulation of goods by what HMRC termed "circulation of money." It was therefore, in the view of Dragon, one where HMRC had completely failed to establish its case. But it also illustrates (as does deal 532) a transaction chain common to a considerable number of deals undertaken at that time with Furylong Ltd (F) as the defaulting trader and three "buffer" companies between the defaulter and Dragon: T M Mobiles (TM), Phone Mad Network (PMN), and DD Trading (DD). The tribunal notes that Furylong and these three companies always appear in deal chains in the order just stated, and never in any other order, and with price differentials between them highly controlled. It is worth emphasising that Dragon never dealt directly in a deal with either TM Mobiles or Phone Mad Networks, although they were routinely counterparties at earlier stages in the chains leading to Dragon deals. Finally, this deal involves another freight forwarder, Interken Freighters (UK) Ltd ("Interken"), based near London Heathrow, that was also involved in a significant number of the deals.
  196. 153 Deal 548 is booked by Dragon as the purchase and sale of 5,000 7250i telephones on Tuesday 17 February 2004 in two lots of 2,500 units. The goods were handled in the United Kingdom by Interken as freight forwarder. The defaulting trader is Furylong Ltd. The input VAT claimed by Dragon, but refused by HMRC, is £145,250.
  197. 154 Dragon entered two other deals that day of which the tribunal has details. Deal 552 was the purchase from DD, and sale to S R Communications, of 11,575 7250i units at the same purchase and sale prices as this deal. The other deal concerned other telephones. There were seven further deals later in the week of which four involved 7250i phones, all purchased from DD at unit prices varying from £147.50 to £166.50 with sale prices (to Arkridge Trading and S R Communications) ranging from £155 to £174.
  198. 155 The evidence of this deal was presented in the main deal bundles. There is additional evidence not taken into account when HMRC made its decision to refuse payment, but accepted late by the tribunal, in the supplementary bundles. This consists of material from an Interken file. Evidence was given by Mr Birchfield and Mr Atkin. Mr Griffiths gave no specific evidence about this deal. Not all the documentary evidence presented to the tribunal as relevant to this deal is consistent with the narrative that the tribunal accepts – a point accepted by HMRC in its final submissions. There is also inconsistency in dates and times on some documents, but the tribunal accepts most of the dates and times at face value.
  199. 156 The goods in question in this deal were, as in most other deals involving 7250i telephones, reported to Dragon by A1 Inspections to be Central European 2-pin goods. In other words, they did not have any obvious market in Britain. The goods inspected for and reported to DRT Vertriebs (a company located both in Denmark and in Germany) two days later did, however, include some British variants, although it is not clear how they became involved in the deal. But the tribunal agrees with the contention of HMRC that the nature of the goods, which were manufactured in Hungary, suggests that they are not goods imported into Britain for the British domestic market. There is no evidence of any retail element in any of the sales by Dragon or any of the other counterparties in these chains either in Britain or elsewhere.
  200. 157 The British stages of this deal start at 17.13 on 17 February when F faxed TM with a sales invoice for 5,000 7250i units at £163 giving a TM reference. There is an untimed matching purchase order on file, with a partially completed supplier declaration by F to TM. There is a faxed matching release note from F to Interken (timing unreadable) for TM of that date. Other parts of the chain of transactions appear to have been put in place ahead of this. At what appears to be 10.19 that day TM sent a sales invoice to PMN for the 5,000 7250i at £164 (a £1 mark up – a regular mark up level between these companies) and at 15.52 PMN faxed a matching purchase order to TM. Interken received a release note from TM to PMN that day (fax details unreadable on copy in tribunal papers). A copy of the TM invoice indicates payment on 20 February.
  201. 158 PMN invoiced DD that day for 5,000 7250i at £165 (again a £1 mark up) (timing not available, but file copies of invoice indicate both that it was sent and that it was paid). Bank evidence shows that PMN received confirmation that DD paid for the goods at 12.35 on 18 February. PMN sent a fax to Interken releasing the goods to DD at 13.25. But the allocation note allocating the good to DD (and identifying them as goods purchased from TM) was sent at 15.56 that day.
  202. 159 DD invoiced Dragon for the sale of 5,000 7250i at £166 (again a £1 mark up) on 17 February (no timing available). It also sent a customer delivery note to Interken for Dragon that day (copy in papers unsigned). The purchase order from Dragon to DD was sent at 8.12 on 18 February (verified fax timing by Dragon). The pro forma invoice from DD to Dragon was sent at 11.01 on 18 February but bears the date 17 February. With it was sent a supplier declaration signed for DD in Dragon's standard form and a copy of Dragon's purchase order to DD signed as accepted. The release note for the goods was sent to Interken by DD at 14.00 on 18 February.
  203. 160 Dragon asked A1 Inspections to inspect several sets of goods, including these, at 16.48 on 17 February. The request was for some reason repeated at 17.57. What appears to be the reply by A1 Inspections is dated 16 February but was faxed to Dragon at 17.00 on 17 February. It confirms all stock in excellent condition. But – as noted generally above – there are no specific stock identifiers in Dragon's request or the A1 reply. Dragon asked HMRC to confirm the VAT number of DRT Vertriebs as a potential customer at 16.19. It also made a check with the Europa VAT listing facilities that day (untimed). HMRC did not reply to Dragon until 13.52 on 18 February. Dragon made a further request to HMRC citing a purchase from DD and sale to DRT Vertriebs "yesterday late in the day" at 13.59 on 18 February.
  204. 161 At 9.29 (Danish time?) on 18 February DRT Vertriebs sent two revised purchase orders to Dragon for delivery of 2,500 7250i to DRT Vertriebs Germany and 2,500 to DRT Vertriebs Denmark at the unit price of £173.50. A reply to the German order was faxed at 9.41 (British time). The reply to the Danish order is dated 18 February and signed but has no timing. A sales invoice for deal 548/1 was faxed to DRT Vertriebs Germany at 15.00 on 18 February, referring to the DRT Vertriebs order of 17 February. At 15.31 on 18 February Dragon received banking confirmation of a payment to it of the full sum for one of the two deals. There is confirmation to DRT Vertriebs from its bankers for payment to Dragon that day for the German deal. The Dragon invoice for deal 548/2 is stated to be made at 15.00 on 19 February (typed, not faxed, timing detail that may not be accurate). The payment against this invoice was confirmed both to Dragon and to DRT Vertriebs on 19 February (at 11.48 British time and 13.41 Danish time). Dragon instructed Interken to release both sets of goods to DRT Vertriebs at 14.02 on 19 February and Interken confirmed release at 14.19.
  205. The actual movement of the goods from the UK preceded these arrangements. Instructions to Interken to ship the goods on hold in two separate deliveries to Germany and Denmark were faxed by Dragon at 22.08 and 22.10 on 17 February. The instructions were acknowledged by Interken. There were also instructions to insure the shipped goods. The Interken files contain documents by way of CMR freight notes and shipping tickets. (A CMR note is a certificate in agreed form, signed by the relevant parties, to confirm with full details the international shipment of goods between two locations.) The tribunal finds that these show that the goods left the United Kingdom on the night of 17/18 February in accordance with the Dragon sales to DRT Vertriebs and were delivered on 18 February. There is a Kuhne & Nagel inspection report on one of the deliveries on 18 February showing 90 of the 2,500 units supplied by Dragon were UK variants (a point not noted on the A1 Inspection report). A document dated 20 February suggests the onward sale of the goods sent to Denmark to Tradewinds S/A (the same Spanish company as in deal 532) at £179 a unit. That is consistent with evidence from the German tax authorities that DRT Vertriebs only sold goods on to favoured customers in other EU states, of which Tradewinds was one.
  206. This evidence suggests that the documentation for the transactions was not keeping pace with the counterparties' intentions, although the underlying legal transactions, movement of goods and payments to and from Dragon all took place. It also suggests that the course of transactions from F to DD was following a set form (each making a £1 mark up as a turn on the deal) that was, on the balance of probability, not taking place at arm's length but in a way that suggested that the traders were associated in some way. This is confirmed by reference to other deals involving those companies that follow the same pattern.
  207. The evidence further suggests that the inspection evidence could not be linked clearly to the goods actually moved for Dragon to DRT Vertriebs even when measured only by the standard of probabilities. This calls into question whether the safeguards said to be in place in this deal were effective. There is also evidence that the original sale by Dragon to DRT Vertriebs was changed after the deal was struck so as to turn it into two deals, so diverting the goods from one of DRT Vertriebs' two branches to both of them. There is no obvious reason for that if they were promptly sold on to a trader in another state, and the tribunal has no evidence of any reason.
  208. There is additional evidence of the deal from the Interken file admitted late by the tribunal. The inward inspection report of the goods produced from the relevant Interken file and dated 13 February lists 10 packs of 500 7250i of which 2 were "virgin stock", 5 "good stock" and 3 mixed stock. These were sent from S R Communications in Germany. The documentation also suggests that HMRC intercepted these goods on the way in and inspected one box from one pallet at Dover. (One CMR is noted "one box from one pallet opened for Customs examination at Eastern Docks, Dover, Officer 1087", and other documents are also noted by reference to the customs inspection). This is confirmed by the A1 inspection report to Dragon on 16 or 17 February which notes that "31 outer cartons were stamped on pallet F". It records all goods as of excellent condition and all English apart from some goods in Pallet D.
  209. The existence of the Customs stamps on the goods does not tie in with the terms on which those goods were sold on in the UK. PMN following its normal terms sold the goods as "all stock is clean with no stamps". And DRT Vertriebs' purchase order, again in usual terms, stated "boxes with marks, stamps or signs will be rejected." Responding to that, Dragon's purchase order acceptance 548/1 expressly states in the description "no stamps". (The tribunal notes that the usual Dragon wording did not include any reference to stamps, so it may be assumed that this wording was added deliberately). The inspection by Kuhne & Nagel for DRT Vertriebs on 18 February notes 90 UK variants but no Customs stamps. The presence of the reports of the Customs stamps in the inspection reports of both 13 and 17 February suggests that the goods inspected for Dragon were the goods brought to the UK directly or indirectly from S R Communications in Germany. This is not, however, confirmed by the evidence available either directly from F or in the form of the schedule prepared following F's liquidation. It would also appear that the Customs stamps were not removed before Dragon inspected them (as would be suggested by PMN's terms), as they were there when the goods were inspected for Dragon. But it seems that they were removed before the goods went to DRT Vertriebs, and that Dragon knew this (as evidenced by the unusual terms of trade). Who removed them, and why? The tribunal has no evidence of that.
  210. The tribunal notes that on this evidence F bought the telephones for a unit price of £163 directly or indirectly from S R Communications and that the same goods were sold on three days later by DRT Vertriebs to Tradewinds for £179, having changed ownership at least six times in the intervening period. This is despite the fact that conducting each of the transactions included several transaction costs: inspection, warehousing, insurance and two or three sets of freight and shipping costs. This is a significant increase in price, at noticeably more than a marginal transaction cost, for commodities that do not appear to be destined for a consumer market. Looking at this series of transactions in isolation must give rise to the question from where the added value was coming. The only potential sources of value of which there is any evidence lies in the claims for VAT refunds from the national tax authorities involved.
  211. The critical element in this evidence is that on detailed examination there is no clear contractual documentation indicating the source of the goods purchased by F. HMRC nonetheless contend that the source, directly or indirectly, was DRT Vertriebs. There are two lines of evidence on which HMRC invited the tribunal to reach the conclusion that it had shown circularity of goods.
  212. The first item of evidence is a document faxed by TM to PMN at what appears to be 15.50 on 20 February. This document is in the papers in three different versions, with different markings on each version. It follows the standard form seen in deal 532 and other deals. The multiple copying of a standard form strongly suggests that these are copies of a genuine document on which action was taken. (The tribunal has ignored, for reasons stated above, the unprovenanced handwritten comments on the documentation as being of no evidential weight). The document is a third party payment request to PMN to pay £809,200 to DRT Vertriebs and £90,186 to Senater (another company that appears frequently in the evidence on other deals), with a balance of £63,494 payable to TM. The total, £962,880, is fractionally less than the amount owed by PMN to TM (5,000 units at £163 plus 17.5 per cent VAT).
  213. One of the copies of this document is stamped "paid". HMRC could not produce other evidence that these sums were actually paid. In other deals (see deal 532) the tribunal was given evidence by Mr Atkin that there were banking records showing the third party payments being made (see deal 532) and that the electonic records of Interken could provide evidence beyond that in the documents (see deal 618). There is no corroboration in either of these forms in this deal. There is some evidence of payment in that on 20 February DRT Vertriebs received £325,000 as a balance of a payment by PMN to it for 5,000 7250i. This is the same day as the day the third party order was made to PMN. There is documentary evidence in the deal papers that PMN paid £100,000 twice, followed by a payment of £429,820 to DRT Vertriebs on 13 February for 3499 7250i, and £180,000 to DRT Vertriebs for 1,000 7250i on 17 February, but these appear to be irrelevant to this deal. Nor is there any specific evidence that the payment was made "up" the chain rather than "down" it – that is, that the payment if made was to DRT Vertriebs for an earlier transaction rather than because of the purchase from Dragon. However, the tribunal cannot see any obvious reason why such a payment request would have been made for a later stage in the chain. The request, even if not paid, does suggest that the relationship between TM and PMN, and between those companies and the third party payees, seems on the balance of probabilities to be otherwise than fully at arms length.
  214. The tribunal has examined the records in considerable detail. And there are a number of elements of the evidence that suggest that the deals that actually took place raised a number of questions about whether they were truly market transactions carried out entirely in a commercial setting. But was there circularity? The tribunal concludes from this evidence that it is not satisfied on the balance of probabilities that HMRC has shown that there is evidence that establishes that the goods sold by Dragon to DRT Vertriebs were the same goods (or the direct equivalent of goods) sold by DRT Vertriebs directly or indirectly to SR Communications for supply to the UK. It does not accept that the payment request is of itself sufficient to establish the link in the absence both of any evidence of the goods, or title to the goods, moving in that way, or of sound evidence that the third party payment was in fact made in a timeous way that would confirm the nature of the request. Nor is the fact that other elements of the deals, such as the evidence about the Customs seals disappearing, relevant to this question.. It notes several features of this deal that suggest that the transactions were not fully arm's length transactions and that there is no obvious source to explain the significant mark-up of prices during the deals other than input VAT refunds. These confirm to a significant extent the general impressions drawn by the tribunal from deal 532. But after a careful examination of all the evidence available to it, the tribunal considers the link back to DRT Vertriebs to be too uncertain in this case, and that circularity is not shown.
  215. (1) 172 In the absence of evidence that circularity probably occurred, the tribunal concludes that there is no definite basis to find that these transactions are without economic content. And as the issue is no longer relevant to this deal, the tribunal makes no finding on it.
    Deal 572
  216. The tribunal looked at this deal in some detail in connection with the application under rule 19 of the Value Added Tax Tribunal Rules 1986. It did so as an example of a deal in which Hawk documentation was said to be essential to the submissions by HMRC and where the Appellant invited the tribunal to exclude all such evidence. The tribunal instead admitted additional evidence put to it by the Appellant as to the reliability of the Hawk documentation. But it commented in its reasons for that general view that it did not consider the evidence in this deal to suggest exclusion. And it noted that caution would be appropriate with some of the freight forwarder documentation or the absence of documentation. It did not form a view at that stage on the questions now to be answered.
  217. Deal 572 was logged by Dragon as the purchase and sale of two quantities of 7250i units: 3600 and 2500, on 26 February 2004. The unit price of both purchases was £159.50 and the unit sale price was £167. The defaulting trader was Furylong Ltd (F). The input VAT claimed on the two sales was £100.485 and £69,781.25. As noted, the goods were handled by Hawk, and additional evidence from the Hawk files was admitted late.
  218. 175 To put this deal in its context, it was one of 8 deals logged by Dragon as made on Friday 26 February 2004 and of which the tribunal was informed (deals 572 – 579) with a further 5 deals the following day (deals 580 – 1, 583-5). Most of the deals, like this one, involved the 7250i mobile telephone. All the 7250i units were purchased over those two days from the same trader (DD) at prices varying from £163 to £156 a unit and were sold on to S R Communications, Sunico A/S or Arkridge at unit prices varying from £170 to £160. There is no obvious reason why the prices varied in this way, as the prices at which the goods were sold to Dragon appear so far as the tribunal can see to be the result of cumulative pre-programmed mark ups along the prior chain of transactions and not the result of negotiations by Dragon.
  219. 176 HMRC contended that the evidence of this deal showed circularity of goods. The submission is that the goods both started and finished with DRT Vertriebs. The papers show the tribunal a familiar series of transactions moving three quantities of 7250i (2,000, 2,000 and 2,100) from F to TM at £156.50, then to PMN at £157.50, then to DD at £158.50, then as one sale of 6,100 to Dragon at £159.50 on 26 February. Dragon sold the goods to DRT Vertreibs in separate orders for 3,600 and 2,500 at £167 on 27 February (although the deal was logged the previous day). The goods for one order went to Denmark and the other to Germany on 26 February. The "Atkin Schedule" shows an onward sale of some of the units by DRT Vertreibs to Tradewinds at £170.
  220. 177 Details in the Hawk file show the release of two batches of 2,000 7250i shipped to Hawk in London to Senater on 26 February, having been loaded in Germany on 25 February. The chain from F shows the same order of traders and closely controlled margins but untidy detail of paperwork and timings seen in other transactions and strongly suggests that the movement of the goods, although technically in three separate series of sales and purchases, was a coordinated operation in respect at least of the two quantities of 2,000 7250i until the goods were in the hands of DD.
  221. 178 The evidence of circularity produced by HMRC for this deal is in the warehouse file MD 32416 from Hawk. The front cover of that file, dated 26/2/04 shows that the file is about the movement of 4,000 7250i that went along the chain: "Senater/ Furylong/ TM Mobile/ Phonemad/DD Trading/Dragon Futures. The file also notes a release for the phones "ex Senater" to TM and "ex Phone Mad" to Dragon. This is, however, evidence to be treated with caution. The file contains a series of release notes releasing a block of 6,100 7250i, save for a release note, dated, 16 February at 15.24, releasing 6,300 7250i to Furylong that day. The other release notes all also bear that date, but not all the timings are readable. The tribunal takes into account that some of the dates, perhaps including that on the Hawk file cover, may have been completed after the event, but it finds no significant problem with the chronology evidenced on the face of the Hawk documentation as compared with the other documentation in the evidence of this deal as a whole.
  222. 179 There is also a separate file, MD 32417, for 2,300 7250i of the same date showing goods shipped from DRT Vertriebs and passing from Senater to F to TM to PMN to DD to Dragon. There is a DRT Vertriebs transaction number consecutive to the two shippings of 2,000 7250i. There appears to be a degree of muddle between the papers on this file and the papers on the previous file indicating some unexplained goods movements or perhaps errors. There is clear documentation that DRT Vertriebs moving 2,300 7250i to Hawk "on hold" on countersigned papers dated 26 February. There are countersigned instructions also of that date to allow Senater to inspect and also to move the goods on hold. And there are shipping and inspection papers indicating the arrival of the goods on 26 February at Hawk. But the Senater instructions are for 6,100 units in both this file and the previous file. There is also documentation in the file suggestion a chain of sales and purchases of a quantity of 4,500 units separately from the 6,100 units from F up to and including a release of 4,500 units to Dragon faxed by DD to Hawk on 27 February at 11.00 together with a customer delivery note signed at that time by someone on behalf of Dragon. This is clearly separate from the release and customer delivery acceptance at the same time for the 6,100 units that are the subject of this deal. These papers would therefore appear to relate to deal 573. But there is nothing to indicate what happened to the extra 200 7250i sent to Hawk by DRT Vertreibs/Senater but not passed on in this deal.
  223. 180 The tribunal concludes that there is obviously some degree of muddle in the papers held in the two Hawk files. But that taking the documents as a whole across the two files and noting the details of other deals being carried on at the same time, the documentation appears genuine as to both details and timings, and the Hawk generated documents appear consistent with the other documents. This therefore suggests that the 6,100 7250i held by Hawk and sold to DRT Vertriebs in two linked deals by Dragon were either the same 7250i as the bulk of those brought in to Hawk from DRT Vertriebs on the order of Senater as three amounts of 7250i the previous day, or direct equivalents. That indicates circularity. The "missing" 200 7250i and the muddled paperwork serve to confuse but do not, in the tribunal's view, alter its conclusion on the balance of probabilities that the papers show circularity of goods from DRT Vertriebs to DRT Vertriebs.
  224. 181 The papers also contain third party payment instructions from TM to PMN in the standard form noted before, the instructions being made on 27 February for 2000 7250i at £174 (not a price quoted in any of these deals, but perhaps a lower unit price plus VAT), with the balance to TM. The reference numbers on the instructions separately identify these to the deliveries of 2,100, 2,000, and 2,000 7250i covered in this deal. There are multiple copies of the instructions, some stamped paid. There is banking evidence that PMN paid Senater sums slightly exceeding the totals required on 27 February. As Senater is shown by other documentation to be a link, if not the link, between DRT Vertriebs and the sale to F, this is additional evidence of circularity. In the view of the tribunal this also confirms, despite its finding in the previous deal, that the evidence of a request for a third party payments, and the making of that payment, may show circularity where more direct evidence of the movement of goods or of the title to goods is incomplete.
  225. 182 Is deal 572 (and with it the associated deals) an economic activity as judged by the Bond House and Optigen criteria? The tribunal considers not. The chain of transactions shows all the features of risk-free, pre-priced, pre-ordained movements of goods not so much round in a circle as backwards and forwards. The only source of gain evidenced is the claims that could be made, and the payments that could be avoided, to the national VAT authorities. The pattern, seen in several series of transactions, is that the goods move between two European Union states while title is passed between entities in three, four, or even more European Union states. But neither the goods nor the legal title leave the European Union at any stage (for example through the use of direct tax low-tax regime territories). The goods were, again, Hungarian-manufactured goods with central European specification and two pin chargers that could not be readily sold on the UK market. Nor is there any evidence of any onward sale into the consumer market. They were inspected for Dragon by A1 who noted Customs marks removed from the goods and also noted that some of the goods had French/Italian manuals, not English (or German or Spanish) . There appears to be no note of any query about this from Dragon. DRT Vertriebs' terms made it clear that any Customs marks had to be removed, and they had been. The tribunal can see no probable explanation for the series of transactions of which this deal forms part other than as a series of legal transactions and movements of goods designed to profit from national VAT authorities but no other source.
  226. Deal 618
  227. The tribunal looked in detail at this deal as one of the later deals undertaken by Dragon, at a time when it knew earlier deals were being examined by HMRC, and when it had move on to trade with different counterparties. It is also one in which HMRC rely on evidence in an Interken bond check file, an element of the evidence not examined in deals 532 or 548. The tribunal noted that some relevant (and repeated) evidence for this deal was in the papers for deal 619 and later numbered deals undertaken at the same time. This is because the 7250i telephones in this deal were brought into the United Kingdom as part of a much bigger consignment, and Interken appears to have kept the papers for all the onward deals relating to these goods on the same file. The deal again involved 7250i.
  228. Deal 618 was logged by Dragon on 3 March as a purchase and sale of 5,000 7520i, purchased at £157 and sold at £163.50. The defaulting trader is Swift Online Ltd ("S"). The input VAT claimed and refused was £137,375.00. HMRC contend that the deal involved circulation of goods.
  229. This deal is one of a significant number of deals concluded by Dragon at this time. That week 11 deals of which the tribunal was informed were logged on Monday 1 March, 8 on 2 March, 11 including this on 3 March, 7 on 4 March, and 11 on 5 March. At that point Dragon suddenly stopped dealing in mobile telephones. Most of these final deals concerned the purchase and sale of 7250i telephones at purchase prices varying from £157 to £141 and sale prices ranging from £164 to £148. That this might lead to some confusion in the papers is made more likely because some of the deals have very similar details. For example, both deal 618 and deal 619 are for the sale of 5,000 7250i units at the same unit price, and the ship on hold instructions for both deals were sent by Dragon to Interken in the same fax transmission, both technically preceded in that same transmission by a similar ship on hold instruction for deal 520 (which was for 3275 7250i), all destined to the same purchaser. Instructions to Interken from Dragon about deals 621, 622 and 623 were also sent in the same fax transmission (at 18.13 on 4 March).
  230. Dragon told HMRC on 3 March that it intended to buy 5,00 units from Beechson plc and sell them to S R Communications. It also told HMRC of other deals.
  231. This deal links S R Communications with Beechson plc. The tribunal observes what the German tax authorities comment as "interesting" information sent to HMRC at HMRC's request that the public records show that Beechson lent money to S R Communications at the time it was established to assist its foundation. The loan had been repaid. But the tribunal was given no other evidence of any links between Beechson and S R Communications.
  232. The documents put in evidence show that Quantitron (a Spanish company) sold 7250i units to S on 4 March at a unit price of £154.50, the goods being stated as held at Interken's warehouse. Quantitron's sale invoice states that the goods remain the property of Metropolitana De Telefonia SL until payment is made in full. Metropolitana does not otherwise feature in this deal (although it does in others). This would suggest that Quantitron was acting only as a buffer. Documents exchanged that day confirm the onward sale from S to Masterpiece Technology Ltd at £155.50, and the further onward sale to Beechson plc ("B") at £156. The latter was confirmed by fax at 13.41. A Dragon purchase order went to B at 16.08 at a unit price of £157. Dragon ordered an inspection by A1 Inspections that day (no timings available and nor are other details, giving rise again to concerns about the inspection). The S R Communications purchase order to Dragon was dated that day but faxed to Dragon at 12.49 on 5 March. Dragon's instructions to ship the goods on hold to S R Communications is timed at 18.13 on 4 March (see above) and confirmed at 19.36. The Dragon invoice to S R Communications at £163.50 was sent on 8 March with the (notional?) timing of 15.00. Dragon released the goods on 9 March at 14.40. The documentation is, as seen above, again somewhat untidy, but it would appear from this that the deal with S R Communications was not confirmed until 5 March and the goods remained the property of Dragon until 9 March.
  233. Information from the German tax authorities indicates that most of the goods were promptly sold on to Senater (another company that appears in a number of these deals) at £168.50. Presumably this had to wait until after 9 March to be executed.
  234. The Interken freight forwarder file in which these deal papers were held was opened on 4 March, noting a shipping from, and a charge to, S R Communications. The agent is stated as RTR and no specific commodities are named on the file front cover. A fax from S R Communications sent on 6 March at 14.55 and also dated 06.03.2004, informs Interken of the pickup for 04.03.2004 of 20,390 pieces including 12,500 7250i from RTR. Dragon queried these dates. The tribunal is also a little puzzled, but the explanation may be that, as with faxes seen on several other deals, the paperwork was sent in late, or later copied with some other documents to which it was linked (this fax was transmitted as page 6 of 7, the others not on file). There is certainly correspondence with the front of the Interken file as to the shipper, the agent, and the delivery date. And, as noted below, replies to that fax bear the date 4 March.
  235. The file contains a number of CMR notes (badly copied, so that the key dates are missing from some copies and the document numbers from others) and other messages about transportation. Two are faxes from RTR dated 4 March (but with fax timings missing) giving – as requested on the fax from Interken of 4/6 March – details of the vehicle numbers and drivers bringing the goods in. Some of those vehicle and driver details are clearly on some of the CMR forms in the file (others cannot be read or have the details missing). Some of those forms show the goods leaving Frankfurt on the evening of 4 March and being receipted by Interken in the UK on 5 March. There are CMR documents to account for the movement of 38 (possibly, though it is not fully readable, 42) pallets of the 42 of which RTR informed Interken. That would also be about the right number for moving 20,390 pieces in pallets of 500 units (the usual size – pallets weigh 200–300 kg each). There are also some CMR documents on file showing goods being sent to Frankfurt to S R Communications for Dragon. These show the goods being signed out on 5 March (no timings). Two of the more readable CMR documents clearly show that one of the drivers who brought goods in on 4/5 March (and whose name and vehicle was notified to Interken by RTR in one of the faxes) also took goods out for Dragon on 5 March.
  236. The tribunal looked at this evidence particularly closely as Dragon criticised it strongly in its submissions. Close examination reveals considerable problems with the standard of copying of some documents, in addition to absence of documents that were obviously missing (such as the pallet seal details originally scheduled to the faxes to Interken but not with the papers on file – were these removed form the file?), other documents in the file (not detailed here) being clearly irrelevant to any of the deals taking place at that time, and some dates on documents not making sense. But working to the standard of probabilities rather than the higher standard of the criminal courts, the tribunal is satisfied - without calling for a check of copies against original documentation or disclosure of any other documents - that there is sufficient evidence here to show that it is more probable that not that the 5000 7250i units in this deal (or direct equivalents from stock) were brought in from S R Communications on 4/5 March and despatched back to S R Communications on 5/6 March. The tribunal does not accept the submission from Dragon that taken together the dates on these files do not work out or that the goods appear to have left before they arrived.
  237. There is further confirmatory evidence available. As a result of deals 618 to 623, Interken billed Dragon on 12 March for shipping to Frankfurt for S R Communications on 4 March and following: 13,275 7250is, 500 6600s, and 2799 7201s. Those details are in line with the details of the multiple inward and outward shippings evidenced in the file. While they do not agree in complete detail, they do suggest that Dragon was responsible for the outward sale of most of the telephones brought in from S R Communications in the inward delivery to Interken.
  238. Information about the inward and outward movement of goods through Interken was also presented to the tribunal in the form of a "bond check file" containing extracts of information collated from Interken files held in electonic form. Evidence was given about the compilation of this file by Mr Atkin. He explained that it was necessary on the grounds of confidentiality to extract the information provided from the original records. The tribunal accept this evidence of provenance. The extracted information shows Dragon deals 618 to 623 together. It confirms the inward figures of goods from S R Communications noted in the freight forwarder documentary file. And it links it to the Dragon "exports" noted. In the tribunal's view this serves to corroborate the information in document form on the Interken file and the Interken account and therefore to confirm both. Equally, as the tribunal drew its conclusions above without this evidence, it confirms that evidential weight can be given – in context – to the Interken bond check file.
  239. On that evidential basis, the tribunal finds that on the balance of probabilities HMRC has shown that the goods sold by Dragon to S R Communications in deal 618 and sent to S R Communications on hold on 5 March were part of the batch of goods sent to Interken from S R Communications the day before and which arrived with Interken earlier that day (or at least their direct equivalents). It also finds that although the deal was logged by Dragon as taking place on 3 March, nothing of significance took place on that date. The key dates for the deal were 4 March (release of the goods to Dragon and instruction by Dragon for delivery on hold), 5 March (delivery on hold) and 9 March (release to the buyer).
  240. Was Dragon engaged in an economic activity in deal 618? Dragon was one of several traders who handled these telephones in quick succession at ever increasing prices. The goods came from Germany from S R Communications on 4 March, but were sold by a Spanish company, apparently acting for another Spanish company that had title to the goods, to a British defaulting trader. The price was £154.50 a unit. The goods went to the first British buffer at a mark up of 50p a unit and to the second at a mark up of £1 a unit. Dragon then added a mark up of £6 a unit and, relying on the German tax authority information, the tribunal notes that S R Communications then added a further mark up of £5 a unit. So the price increased from what can reasonably be assumed to be £154 or less a unit when Quantitron received them directly or indirectly from S R Communications to £168 a unit when the same business, S R Communications, sold them on the next time. And the only intervening acts were that they were transported from a German warehouse on one night to a London warehouse and then back to the German warehouse the following night (at least in part by the same vehicles on both legs) and that at some point an input VAT refund was claimed from a tax authority and that at some other point a trader did not pay VAT to the same tax authority. Further, the goods were handled in the United Kingdom in very quick succession by four traders none of whom had any intention, so far as the tribunal can see, of doing anything other than passing the goods down a chain that was provisionally (if not definitively) in place before the goods started into the chain. They were, as the tribunal has observed before, goods with no obvious British market. The only reasonable explanation for their transport to the United Kingdom in the context of these deals was so that they could be transported out again with appropriate VAT consequences. Further, this was happening in several similar high value back-to-back matched deals all taking place at the same time. The tribunal can see no profitable, or potentially profitable, activity taking place other than that of moving the goods through VAT frontiers. It does not regard that as an economic activity as discussed in Bond House and Optigen.
  241. More generally, the tribunal considers that this deal illustrates that the goods in some of the chains of transactions in which Dragon was involved (and particularly the deals made during its final trading period along with deal 618) were moving not so much round in a rapid circle that might fairly compared with a carousel but back and forth while the title to the goods went round a longer circuit that did not have any obvious commercial basis. As in the other deals noted in detail, again there is evidence that the paperwork and payments followed behind both the oral agreements and the movements of goods without that apparently causing any undue concern to any of the counterparties.
  242. Other deals in dispute
  243. The tribunal now turns more briefly to the other deals that formed part of the hearings in these appeals.
  244. Deal 514(a) and (b)
  245. HMRC conceded that it could not show circularity of goods in these cases and the tribunal allowed the appeal on these deals without reasons.
  246. Deals 515 and 519
  247. In both cases the defaulting trader was Smart Telecom Ltd ("Smart"). Both involved Hawk as the freight forwarder.
  248. Deal 515 was logged by Dragon on 3 December 2003 as the purchase of 2,000 7250i telephones from Sound Solutions for sale to Mobile World, with input tax claimed of £59,850.00. The chain was stated by HMRC as Smart – West Coast Services – Phone Mad Networks (PMN) – Sound Solutions – Dragon – Mobile World – Tradewinds, and the tribunal accepts that this is shown on the evidence. HMRC contended that there was circularity by reference to a third party payment from PMN to Tradewinds. Dragon contended that this was a payment down the chain not up it. The original decision by HMRC was taken without access to Hawk files, but Hawk documentation was admitted late by the tribunal. This suggests an inward delivery of 5,000 7250i from DRT Vertriebs via Tradewinds of which these 2,000 formed part, 3,000 being sold on by another chain. The tribunal sees no reason to reject the Hawk evidence. Taking that evidence with the evidence of the third party payment request and payment, the tribunal is satisfied that there is circularity through the link of Tradewinds. It also notes that the papers suggest links between DRT Vertriebs, Mobile World, and Tradewinds. Applying the same criteria as those examined in more detail in the deals set out above and deal 519, it concludes that on the balance of probabilities there was no economic activity in this deal. But it notes that in so concluding it is relying in part on late admitted evidence.
  249. Deal 519 was logged by Dragon on 10 December as the purchase and sale of 4,000 Samsung V200 mobile telephones. HMRC contended that the chain was Smart – West Coast Services – PMN – DD Trading (DD) – Dragon – Mobile World. HMRC contended that circularity was shown by a third party payment by PMN at the order of West Coast to Mobile World. Dragon contended that no evidence of payment actually being made was shown. At the time of the original decision, no Hawk files were available. These were admitted to the evidence late by the tribunal. The Hawk documentation suggests the chain for the 4,000 V220 is Mobile World (sent to Hawk in 2 batches of 2,000) – A-One Trading BV (Netherlands company) – Smart – West Coast – PMN – DD - Dragon. Taken together with the other evidence, the tribunal is satisfied that this shows circularity of goods with the common link being Mobile World. But that decision, as with the decision on deal 515, is based on all the evidence available to the tribunal, and not merely that on which HMRC made its own decision.
  250. There is again a pattern of controlled mark ups, of goods moving swiftly to and from Germany and the Hawk warehouse, untidy paperwork, and clear coordination in parts of the chain such that the tribunal finds, putting this deal in its context, that it is satisfied that this is also a chain of transactions that does not constitute an economic activity. In particular, this is another chain where there is no evidence that any of the counterparties either intended or tried to make any onward sale into the consumer marketplace. There was thus no source of profit outside the VAT consequences of the movements of the goods.
  251. Deals 520 and 524
    (1) 204 Deals 520 and 524 are logged by Dragon on 5 February 2004 as purchases from DD of 2,000 7250i telephones at £165 a unit and later sold to DRT Vertriebs at £178 a unit in deal 520 and a further 3,999 7250i sold at £174 a unit in deal 524. The freight forwarder was Interken, and HMRC offered evidence of this from the Interken bond check file. The identified deal 520 chain is contended to be DRT Vertriebs – Greenday (the defaulting company) – TM – PMN – DD – Dragon –DRT Vertreibs. The identified deal 524 is contended to be DRT Vertriebs – Tradewinds – Furylong (the defaulting trader) – TM – PMN – DD – Dragon – DRT Vertriebs – Senater. HMRC contended both direct and indirect evidence of circularity of goods, as there was a third party payment.
    (2)
    (3) 205 The tribunal notes that the evidence suggests that the defaulting trader in deal 524 should be Greenday, or both Greenday and Furylong, and that deal 524 appears to be the last deal conducted by Dragon that involved Greenday rather than the first that involved Furylong. But on either analysis the chain went through one or both of these traders and there is clearly a defaulting trader present.
    (4)
    (5) 206 There was evidence admitted late from the freight forwarder files for deal 524 but not deal 520. That file appears also to contain details relating to deal 525 (see below), in which Interken was also involved. It suggest that the goods came in to Interken from DRT Vertriebs but with Tradewinds interposed as a buffer with DRT Vertreibs' full knowledge. It also suggests that the 4,000 units were allocated and released not to Furylong but to Greenday, and that the 4,000 units became 3,999 on release from PMN "stocks" (or "stock" – both terms are used). This raises questions, discussed more generally above, about when the "stock" that was inspected for a particular deal was identified to that transaction consistently both at the time of the deal and on release of the goods. The terms of the release in this transaction suggest that PMN was releasing an amount of stock that was at that stage unidentified save by the generic description of the goods and the quantity. There is in particular no indication of batch, pallet, IMEI or earlier transaction numbers.There is a faxed release to Interken dated 6 February at 18.09 of 4,000 7250i units from Greenday to TM. The tribunal also notes that the documentations suggests that no specific goods were released by PMN to DD. But there is a clear link in the flow-through of the quality and quantity of 7250i units transacted from DRT Vertriebs/Tradewinds to Dragon. The link on from Dragon to DRT Vertriebs is also clear, and circularity is shown.
  252. The evidence of third party payment in deal 520 is in the form of a faxed instruction (fax timing partly obscured in copying but suggesting 13.48) in standard form from TM to PMN on 5 February for payment to DRT Vertreibs, and there is evidence of payment by PMN to DRT Vertriebs that day (timestamped at 15.55). The tribunal take that time-stamping as reliable, and so confirming the date of the request, if the two are linked. The unit price, as Dragon noted, was said to be £180, while the sale to TM by PMN was £163 plus VAT. This price difference was queried by Dragon. Other documents have suggested that the mark up is related to VAT, possibly not at the UK rate, and Mr Birchfield contended this in evidence. At the same time, in connection with deal 524, he conceded that the dating on the third party payments faxes from TM to PMN was not always reliable. The tribunal has commented on the dating. The pricing issue does not in the tribunal's view mean that no weight can be put on the fact that money – provided it is shown by adequate evidence to be linked to this deal - was clearly paid across to DRT Vertreibs, sidestepping the defaulting trader in a way common to all the chains involving Dragon deals that were examined by the tribunal. The tribunal is satisfied in this case that the request was made specifically with regard to this deal, and that the payment was similarly so made.
  253. Mr Millar gave written evidence about deal 520. He stated that this was an example of how Dragon kept its documentation in connection with deals. The tribunal finds that the deal papers used by Mr Millar in his evidence are the same as those in the deal papers produced in the deal bundles. It also notes that the terms of the Dragon purchase order in this deal are in more detail than similar orders in other deals. For example there is a requirement that the goods be "Sim unlocked and never locked previously. Flash stock is not accepted… Boxes with marks, stamps or signs will be rejected." The same wording was used in deal 524, but not in some later deals, in which there was marked stock transacted without any recorded objection by Dragon.
  254. While circularity is clearly shown in deal 524, the tribunal hesitated about deal 520. But it finds that, weighed on the balance of probabilities, the series of transactions was such that the third party payment avoiding the defaulting trader followed the same pattern and format seen repeatedly, and that on balance the evidence is that the third party payment was promptly made in a context that suggested the absence of an arm's length reason for the third party payment, with evidence that this was a payment to a linked party to the chain. Was there another explanation of the payment more probable than this, or such as to make this explanation less than probable? There was no good reason shown why that payment in this chain should be "down" the chain for later transactions rather than "up" it, or otherwise calling the probability of this being a payment "up" the chain into question. The tribunal therefore finds that there is sufficient evidence to satisfy it that HMRC has shown circularity through DRT Vertriebs in this deal also.
  255. (6) 210 The tribunal finds both deals to be within chains that amounted to activities that were not economic activities by the criteria identified and discussed above. These were again goods moving into and out of the United Kingdom on a controlled circular series of legal transactions cloaking a straightforward movement of goods to Interken from DRT Vertreibs' warehouses and then from Interken to DRT Vertreibs' warehouses in a short time.
    Deal 521 (a), (b), and (c)
  256. These are logged by Dragon on 5 February 2004 as linked deals involving the purchase from DD Trading of 1,000 Nokia 6600 telephones, 2,000 Samsung V200 telephones, and 1048 Samsung E700 telephones, all being sold on to Sunico A/S. The input tax claimed was £42,437.50, £61,950.00 and £44,474.50. The freight forwarder was Hawk. There was a standard chain for all three deals running Greenday (the defaulting trader) on 4 February 2004 – TM – PMN – DD – Dragon – Sunico, with an onward sale of some of the units to Tradewinds. HMRC contend evidence of circularity by reference to a third party payment from PMN to Sunico and also to Senater. Dragon dispute that the evidence shows circularity.
  257. The tribunal is fully satisfied that there was a coordinated series of deals from Greenday to DD such that these deals should be assessed together. They were also purchased by Sunico as a single deal (fax to Dragon on 5 February). The units purchased by Sunico was sold on by Sunico to Tradewinds in exactly the same quantities (including the unusual order of 1048 E700 units) the following day with marked up prices, this being confirmed on the freight forwarder files by documents from Tradewinds and Sunico.
  258. The warehouse files show the V200 units coming from Tradewinds. The files also show that all the V200 were Customs stamped at the time they were consigned to Sunico but were later reported as repacked by the shippers but with some items missing. There is no evidence that Sunico rejected the deal. The description of the goods as unstamped in the paperwork, previously noted on Dragon orders, is missing in the documentation for this deal. The V200 were also reported to Dragon by Hawk as Customs stamped on 5 February, and with 420 pieces loose and not boxed. This was in breach of the terms offered by DD to Dragon in its pro forma invoice (and also the terms offered by PMN to DD), and was not subject to any relevant conditions (such as those evidenced by Mr Millar) in the Dragon purchase order.
  259. The E700 units are on separate Hawk files. These suggest that 598 of the E700 units arrived separately from S R Communications and Senater and were released to Dragon via PMN and DD. 450 E700 units came to Greenday via Machatel Develops S.L. ("Machatel") (a Spanish company) although the goods physically came from Denmark. Machatel also appears to be the source of the 6600.
  260. A document dated 9 February on the deal file for deal 525 is deal file indicates Furylong releasing the 1048 E700 to TM from "Inter Cam" along with 2,000 V200, 1,000 6600 and 7,000 7250i. This appears to suggest that Furylong "inherited" Greenday's stock at Interken (although it did not know its name!) on or before 9 February (the day on which Greenday disappears from the deal chains in these appeals). The tribunal noted no other evidence on this, but it does suggest that by some mechanism Furylong replaced Greenday in the supply chains at that time and that TM and the freight forwarder would have known this and accepted it.
  261. Attempting to unwind this composite deal to and by Dragon is in the view of the tribunal particularly difficult. It finds evidence of circularity with regard to the V200, confirmed by the third party payments and also by the coincidence of the Customs marks on both the inward and outward movements of the telephones. But it is unable on the balance of probabilities to accept that HMRC has shown circularity with regard to the other telephones in the deal. These appear to have come together at Hawk from a number of different inward consignments and sources. There is a link which the tribunal finds established with Senater as to 598 of the E700 units, but otherwise the tribunal does not find circularity established for the balance of the E700 or any of the 6600 telephones.
  262. The tribunal makes no finding about economic activity where circularity is not established.
  263. The tribunal is therefore unable to reach a single conclusion about these linked part-deals. In the case of the V200 there is in the view of the tribunal clear evidence that this was part of a circular transaction chain and that the course of transactions was not truly an economic activity. HMRC argued that in some deals at least the goods appeared to be tokens. In the view of the tribunal this argument is made good in the case of the V200 because of the way in which the Customs marks were ignored by the counterparties despite the specific terms of sale in the chain, and the fact that no party rejected the marked goods or the damaged and unpackaged goods at any stage despite clear inspection evidence that several of the parties would have known about, or had certified, the goods as unstamped and undamaged. While the issue of Customs stamping might be of importance to intermeidarieis, the fact that telephones were damaged or unpackaged would suggest that they had potentially lost significant value for a sale to an ultimate consumer. And if the telephones were not being traded towards an end sale to a consumer at some stage, then their only economic value related to the process of transacting itself and there is weight in the HMRC argument.
  264. The E700 present a more confused picture. Circularity has been established for part of the deal only. That presents a new issue. Can the deal as a whole be non-economic if only part of it is shown to be circular? While in principle the tribunal does not rule this out, there must be a serious doubt on the total absence of economic activity where the failure to show circularity applies to a significant part of the value of a deal. Operating again on the balance of probabilities, the tribunal do not find that the transactions involving the E700 units were, as a whole, non-economic. It makes no finding, as it is not necessary to do so, with regard to the 6600 telephones.
  265. Deal 525
  266. Deal 525 was logged by Dragon on 6 February 2004 as the purchase of 2,500 7250i units from DD at £161.50 and the sale to Mobile World at £170. The defaulting trader is said to be Furylong Ltd. As noted in deal 524, the freight forwarder is Interken and some of the evidence about deal 525 is interleaved with that for 524. The tribunal does not draw any significant adverse conclusions from that. Given the similar parties, goods, and dates, these suggest somewhat slack or hurried filing rather than anything more suspicious. The UK chain was contended to be F- TM – PMN – DD – Dragon, with the sale on to Mobile World and the contention that circularity is shown by a third party payment to Mobile World by PMN on the standard form of instruction from TM. The input VAT claimed by Dragon and refused by HMRC was £70,656.25.
  267. There is no direct evidence in the form of goods invoices or movement orders showing that Mobile World supplied the goods to F that then moved down the chain back to Mobile World. Instead, HMRC rely on a third party payment. The evidence in this case is weaker than in some others, but it follows what the tribunal regards as a familiar pattern on documentation in a standard form and which gave rise to actual third party payments to Mobile World related to this deal and, as before, cutting out the defaulting trader(s) in favour of payment to an offshore supplier. There is also evidence in the form of the Interken file documentation, including but not limited to the file cover, suggesting the goods originated with Mobile World. This is again weak evidence, and Mr Birchfield was cross-examined on the third party payment with that in mind. The point about the price used for the third party payment was made to him. His view was that it related to the sums involved "up" the chain with VAT on. The tribunal forms no firm view of this, but does not regard the issue as preventing identificationof a third party payment to a particular deal if there is adequate other evidence. Taking the evidence as a whole, and allowing for the (again familiar) pattern of paperwork not always keeping up with goods, the tribunal find that the pattern of the request for and making of the payment is such in this case to establish circularity linking the chain through Mobile World on both sides of this deal.
  268. Once the circularity is shown, the rest of the detailed evidence of the chain, including the timing issues, regulated mark ups, the presence of a defaulting trader, the usual order of counterparties in the United Kingdom, the swift movement of the goods into and out of the Interken warehouse, the non-market associations that appear to apply between many of the counterparties, the nature of the goods which again have no ready market in the UK, the seeming absence of business risk and of any profit source other than VAT input claims, points clearly to a series of events that do not constitute economic activity.
  269. Deal 526(a)
  270. Deal 526(b) was not decided at the time of the hearings, so is not subject to consideration by the tribunal. Deal 526(a) is logged by Dragon as the purchase of 4,000 V200 telephones on 9 February 2004 from DD at £166.50 with a sale to DRT Vertreibs at £174. The input VAT claimed and refused was £116,550. The defaulting trader is Furylong Ltd (F), and the chain is contended to be DRT Vertreibs – Senater – F – TM – PMN – DD – Dragon - DRT Vertriebs (in two amounts of 2,000 and 1,999) – Senater. In addition to the original evidence on which HMRC refused the refund of VAT to Dragon, additional evidence was admitted late by the tribunal from the German authorities. Much of this relates to the V200 units, but it does show that the 7250i units arrived from Dragon in boxes some of which were damaged and none of which had original Nokia seals – and with one phone missing. It was the usual central European, 2 pin charger, stock.
  271. The papers are somewhat confused in this case by the interleaving of deal 526(b) but as this concerned Samsung V200 phones and not 7250i it was relatively easy to unwind that aspect of the evidence.
  272. The documentation contains a copy of the same "inheritance" release by F as that on which the tribunal commented in deal 525, again appearing to have been accepted without explanation to and without query by the other parties. The other documentation is also in a familiar pattern with the mark ups being added along the chain. The documentation also clearly shows the goods coming to Interken from DRT Vertreibs Copenhagen on 6/7 February and being released via Tradewinds to Greenday on 6 February (a Friday) then being released by F (as already noted) on the following Monday. But there is also a release by Senater specifically to Furylong sent to Interken on 9 February 2004 at 17.39 (Spanish time?). At that point it appears that Tradewinds dropped out of the deals involving the 7250i as did Greenday, although Tradewinds remained involved with the V200s and was involved again later on the onward sale from DRT Vertreibs. There is also last minute (between a fax on 9 February and another on 10 February) rescheduling of the purchase by DRT Vertriebs from one purchase to two after the original single deal had been accepted by Dragon. It nonetheless remained part of a single deal in Dragon's books. Tthe goods nonetheless appear to have arrived in one delivery, with a report of the consignment including repacked, damaged and missing goods noted but leading to no action or comment evidenced in the papers shown to the tribunal, save one. This is that the Dragon invoice for one of the two sales is for 1,999 7250i although it has sold two batches of 2,000. This suggests it was aware of the missing telephone, but that no other amendment was made to the terms of sale and purchase. This appears to be another deal where the comments of HMRC discussed above about the telephones being mere "tokens" may also be appropriate.
  273. There is also evidence in familiar form of third party payment to Senater for the 7250i units in the TM and PMN stages of the deal. This again both confirms the documentation about the physical movement of the goods and also that evidence of the movement of money may be reliable evidence of the movement of goods.
  274. This evidence appears to suggest a number of changes to the various transactions as they went through, including unexplained changes to some of the counterparties. Again judging all the evidence together the tribunal finds that HMRC has satisfied it on the balance of probabilities that the supply of 3,999 units was circular, the link to establish circularity being DRT Vertreibs and in part Senater. It is a trivial point compared with the evidence as a whole, but the tribunal notes that Dragon contends on appeal that it is entitled to a refund for 4,000 7250i. Its own documentation confirms it knew that it sold 3,999. The tribunal also notes more generally that the counterparties seem unconcerned in this case about the actual quality of the goods being moved, taking a more relaxed view of quality than in earlier deals and apparently content with what appear to be repackaged goods.
  275. For the same reasons as given in other deals, but also noting that this is a transaction where the goods appeared to have lost at least some potential value for eventual sales to end users without the counterparties expressing any recorded concern, the tribunal is satisfied by reference to the evidence as a whole that this is another transaction where there is no economic activity (in the Bond House and Optigen sense. Title to the goods went round in a circle that was closely coordinated, while the goods went to and from the UK freight forwarder, with no obvious UK domestic market for the goods and no source of added value outside the VAT consequences of the deals.
  276. Deal 527
  277. Deal 527 involved two deals. HMRC conceded that the deal referred to as deal 527(a) was one in which they could not establish circularity, and the tribunal had decided with consent that the appeal on that deal is allowed.
  278. Deal 527(b), on 9 February 2004, is logged by Dragon as the purchase of 1,000 Nokia 6600 telephones at £240 from DD and the sale of those units to Sunico A/S at £250, with an input VAT reclaim for £42,000. The defaulting trader is F. HMRC contend evidence of circularity both directly and indirectly, the chain being (or including) Sunico – Tradewinds – F – TM – PMN – DD – Dragon - Sunico – Tradewinds. It is further suggested that the goods in this deal are the same goods as in 521(a) going round a second carousel, so that the start of this chain is the end of that chain. The tribunal notes, however, that it was not satisfied about circularity in deal 521(a) but that does not directly affect its view of this deal. Dragon also points out that deal 521 involved Hawk while the forwarders in this deal were Interken.
  279. There is considerable documentation on this deal, including the physical and electronic records of Interken, the parties' records, and evidence of a third party payment both requested and made in the usual pattern omitting the defaulting trader. The tribunal sees no features of this deal that warrant further attention as it is satisfied that HMRC has shown both circularity and an absence of economic activity in this case by reference to the criteria and standard applied by the tribunal.
  280. Deal 531
  281. HMRC conceded that it could not show that this deal, which related to the purchase and sale of 500 7250i units logged by Dragon on 10 February 2004, involved a defaulting trader and the tribunal has allowed this appeal.
  282. Deal 532
  283. The tribunal discusses this deal in detail above.
  284. Deal 535
  285. This was logged on 11 February 2004 as the purchase from DD of 2,000 7250i telephones at £162 and the onward sale to Arkridge at £169. The defaulting trader was Furylong. The input VAT refused was £56,700. The chain contended was Sunico – Tradewinds – F – TM – PMN – DD – Dragon – Arkridge – Mobile World. There are elements of the deal through Arkridge that are similar to deal 532 noted above. Circularity is contended to be shown by HMRC through third party payments to Mobile World. HMRC contend that it is reasonable to assume that Mobile World directly or indirectly introduced the goods to the UK and this was reflected in the payment. No late evidence was offered for this deal.
  286. The tribunal notes that the documentation of this chain shows that PMN informed DD that Customs stamps had been removed from some of the boxes of 7250i. Dragon's purchase order from DD specifically stated "No Customs stamps or stamps removed". The inspection report to Dragon of goods from DD on 11 February notes "stamps removed". Dragon appears to have proceeded with the purchase in full knowledge that DD had not supplied the quality of goods ordered. This may explain why the purchase price from DD was £162 a unit rather than the £166 - £167 Dragon was usually paying in deals at that time. But it is not consistent with the evidence given by Dragon about the integrity of its monitoring systems. At the same time, the evidence suggests identity of goods moving through the chain.
  287. The tribunal finds that evidence of a request for a third party payment to Mobile World is the only clear evidence pointing to circularity. But, having examined the available information. The tribunal take the view that the papers lack sufficient confirmatory evidence to support the link of Mobile World to the inward leg of this deal, and the tribunal is not satisfied on the balance of probabilities that circularity is shown.
  288. The tribunal accordingly makes no finding on economic activity.
  289. Deal 538
  290. Dragon logged deal 538 as the purchase of 2 amounts of 1,000 7250i telephones from DD and their sale to DRT Vertriebs on 11 February 2004. Dragon conceded during the hearing that HMRC had shown circularity of goods in this deal. But it did not concede that there was an absence of economic activity in the chain as a whole.
  291. The accepted chain is therefore DRT Vertriebs – Tradewinds – F – TM – PMN – DD – Dragon – DRT Vertreibs (by two separate deals) – Tradewinds. There is some muddle in the tribunal's papers on this deal, as they appear also to contain papers relating to deal 548 and deal 537 (which is not before the tribunal). Ignoring these papers as not relevant, and also ignoring unprovenanced handwritten comments (in accordance with the view taken by the tribunal generally), the tribunal concludes that there is nothing in this series of coordinated deals to suggest that there was any substantive economic activity behind them. Again they were, so far as the British chain of counterparties were concerned, dealing with goods for a central European market. The mark ups were consistent and artificial, and there was no sense of arm's length dealing. Dragon, as usual, appears to have accepted DD's price, with the £1 mark up for DD and for the earlier counterparties that is a feature of many of these appeals. Again, the legal title to the goods went round in a circle while the goods themselves went from a DRT Vertriebs warehouse to a DRT Vertriebs warehouse via the UK freight forwarder. There is no opportunity for economic activity, and no source of added value, between the sale by DRT and the purchase by DRT save from the relevant national VAT authorities. Nor is there any suggestion that DRT or any other party was buying the units to cover an exposed position resulting from other sales ahead of purchases. The tribunal saw nothing in any of the chains suggesting anything other than the same kind of matched deals that was the standard approach taken by Dragon to its activities. This is not economic activity.
  292. Deal 540
  293. This is logged as the purchase by Dragon of the unusual number of 1096 7250i units from DD at £166 on 11 February with onward sale to DRT Vertriebs at £174. A rebate of £31,838 input VAT was refused by HMRC. The defaulting trader was F, and the freight forwarder was Hawk. There is late admitted evidence was in relation to Arkridge from the Irish tax authorities and also from the freight forwarder. There is also information from the German tax authorities about DRT Vertriebs. As noted below, the tribunal finds the new evidence of critical importance to the case put forward by HMRC in this deal.
  294. The chain contended by HMRC is Sunico – Tradewinds – F – TM – PMN – DD –Dragon – DRT Vertriebs – Tradewinds. The tribunal considers this deal only briefly because the crucial link has to be made good through the onwards sale to Tradewinds. The rest of the chain is, in the view of the tribunal, clearly evidenced. The link through Tradewinds depends on accepting the evidence of the German tax authorities, and the tribunal does so. The evidence in this case is made clearer than might otherwise be the case because of the unusual number of goods in the deal, namely 1,096 units. . That feature however is enough, together with all the other evidence, to satisfy the tribunal of circularity in this deal. This is because the number goods sold, 1, 096, is unique to this deal and allows more weight to be put on evidence about individual steps in the chain of evidence than is the case with rounded transaction numbers.
  295. There is no feature in this deal that suggests that it is otherwise than a series of transactions devoid of genuine economic activity. Again, so far as the British part of the deal, the goods had no obvious British market. There are the usual indicators of both slackness and coordination operating at the same time such as removed Customs stamps, slipping paperwork and third party payments bypassing the defaulting trader, with no obvious intended onward market outside the carousel companies.
  296. Deal 541
  297. This is logged by Dragon on 12 February 2004 as the purchase of 2,000 7250i telephones from DD at £167.50 and the onward sale to Arkridge for £176, entitling Dragon in its view to a refund of £116,550 VAT. The defaulting trader is again Furylong. Dragon conceded circularity of this deal during the hearing but then withdrew the concession. HMRC contended circularity as follows: Mobile World – Tradewinds – F – TM – PMN – DD – Dragon – Arkridge – Mobile World.
  298. The key linking information on the outward transactions "down" the chain came in part from the Irish tax authorities and the tribunal accepts that evidence. It has discussed its general nature above. The inward linking information comes from the freight forwarder files and third party payment. The evidence in this deal is substantially the same in form as that seen in several other deals. The tribunal has examined the weight to be attached to the separate elements in those other deals, and finds nothing in this case to suggest unusual features warranting further discussion. It finds both circularity and an absence of economic activity for the same reasons as indicated in similar deals above. But again it does so in part because of the evidence of the Irish tax authority.
  299. Deal 542
  300. This deal, also logged for 12 February, shows Dragon buying another 4,000 7250i units from DD, this time at £166.50, and selling on to DRT Vertriebs at £174, with a refund claim for £116,550 from HMRC. The defaulting trader is again F.
  301. The alleged chain is from DRT Vertriebs and Tradewinds via the usual UK counterparties at this time of F- TM – PMN – DD – Dragon and then on/back to DRT Vertriebs and Tradewinds, with a third party payment to DRT Vertriebs from PMN linked to this deal, and evidence on an Interken file. It is clear that not all the documents in the deal papers put to the tribunal relate to this deal, with documents in particular relating to a movement of 3,499 7250i at the same time (possibly deal 547 which is not before the tribunal?). Mr Birchfield strongly resisted in cross-examination the suggestion that the third party payment in this case was a forward payment (down the chain), his view being that it was clearly a payment related only to the inward leg of the deal. But he accepted that in some cases the payments might be forward payments and this conclusion was deal-specific. He accepted that the price used in the third party payment was not the actual price at the time, but it might refer to the inward price. He could see no reason why the price would indicate a forward payment.
  302. Dragon was now using Interken to carry out its inspections for it. The tribunal was given no evidence by Dragon of why it used particular inspection agents, but notes that using the freight forwarder as the inspector removes one of the safeguarding elements of the original approach taken in Dragon's deals. Interken gave a detailed report showing that none of the goods were in original boxes with factory seals. But the terms of trade demanding such boxing do not appear on Dragon's orders or those of the other counterparties in this series of deals. This again is a diversion from Dragon's original approach. The timing suggests that Dragon again appears to have accepted DD's price, leaving the usual staged mark ups for each British counterparty.
  303. The tribunal sees this as another in a closely timed series of chains and is satisfied that there is clear evidence taken together from all the sources of evidence both of circularity of goods in this deal and also an absence of any economic activity.
  304. Deal 548
  305. This has been discussed in detail above.
  306. Deal 554/555
  307. Deal 554 is logged by Dragon on 19 February 2004 as the purchase of 2,000 7250i telephones from DD at £147.50 and the onward sale to Arkridge at £255. Deal 555 is a parallel purchase and sale at the same prices of a further 1,40 7250i. The input VAT claimed was £51,625 and £38,460. The defaulting trader is Furylong. The freight forwarder for both deals is Interken. The deals are considered to be one deal in two parts.
  308. There is a clear chain based on the documentary evidence supplemented by the Irish tax authority evidence of 3490 7250i moving from F to TM to PMN to DD to Dragon in a familiar way (with Dragon again appearing to accept the DD price), and an onward sale via Arkridge to Mobile World. There is also clear documentation showing the goods being allocated to F from Tradewinds with a reference to Mobile World as the main supplier. These follow patterns established from other deals, and point clearly to circularity. The slightly unusual feature is the depressed price of the goods as compared with other transactions in 7250i telephones at the same period. The price seems to have gone down sharply part way through this chain with an amended purchase order from TM to F on 19 February stamped as faxed. There is also an amended TM invoice with the same difference noted. Perhaps the answer lies in the details of the inspection report to Dragon. 50 outer cartons of every pallet had been Customs stamped. Further, there was reported to be Arabic stock present in the goods. This conflicts with the terms of trade on a number of the documents. But, aside from the amended terms of trade, there seems to be no rejection or other reaction to the breach. Another possible answer is some accounting change to allow the goods to proceed through the chains with the "right" price on them (to stop ever accumulating profit margins". If so, then someone took a loss on the deal. But the evidence carries no suggestion of any real commercial pricing about this adjustment, and the parties suggested no explanation to the tribunal of this.
  309. The links inward and outward involving Mobile World are clearly established for deal 554/55. The tribunal also finds, for the reasons discussed above, that there is nothing in these transactions to suggest genuine economic activity is taking place.
  310. Deals 570 and 571
  311. Deal 570 is logged as two parallel deals on 24 February 2004. 4,000 and 2,300 7250i units were purchased from DD at £163 and sold to DRT Vertriebs at £170, with a VAT reclaim of £114,100 and £65,607.50. Deal 571 is logged as a further 4,300 7250i units purchased from DD that day at £156, and sold on to Sunico A/S at £163. The defaulting trader is again F and the freight forwarders are both Hawk and Interken. Additional evidence was accepted late from the German tax authorities. HMRC also contend that the Interken bond check file shows 6,600 7250i inward from DRT Vertriebs being offset by the outward sale by Dragon of 6,300 7250i to DRT Vertriebs, and that the physical files adequately show the movement of the goods in deal 571.
  312. The chain in deal 570 is said to be DRT Vertreibs – Tradewinds – F – TM - PMN – DD – Dragon – DRT Vertriebs – Senater. That in deal 571 is stated as Matachel – F – TM – PMN – DD – Dragon- Sunico – Matachel. HMRC also contend that the final stage of deal 571 is the start of deal 573.
  313. The tribunal accepts the evidence, including the bond check file, as establishing circularity through DRT Vertriebs in deal 570. The new evidence from the German tax authority establishes the onward sale to Senater.
  314. The evidence of circularity in deal 571 is less clearcut but is, in the view of the tribunal, also made good on the balance of probabilities, the link being Matachel. It notes the concerns of the Appellant about the confusion in the freight forwarder files. Some such confusion is perhaps not surprising given what was going on at the time in terms of large numbers of goods being moved fast along similar transactions routes. On balance of probabilities, the tribunal again finds that HMRC has made good is case for circularity in deal 571 as well as 570.
  315. The tribunal records that, applying the approach noted above, it finds that on the balance of probabilities these deals had no element of economic activity in them.
  316. Deal 572
    (7) 258 The tribunal has discussed this in detail above.
    Deals 573 and 574
    (8) 259 HMRC conceded in argument that it had not made its case good for 200 of the units in deal 573, but maintained its arguments for the other 4,300 7250i units. Dragon logged the deal as a purchase on 26 February 2004 from DD at £156 a unit and sold them to Sunico at £163.50. The defaulting trader is Furylong and the freight forwarder Hawk. The input VAT refused is £122,850. Deal 574 was logged on the same day involving the same parties and products, but a quantity of 2829 units and slightly lower unit prices, £155.50 and £163.25. In this case, in direct contrast to deal 573, Dragon conceded in argument that circularity of goods was shown, but not an absence of economic activity. The defaulting trader in both cases was Furylong, and Hawk handled both sets of goods. The input VAT claimed and refused was £122,850 for deal 573 and £76,984.16 for deal 574.
  317. The tribunal has in mind its consideration of deal 572 in considering these deals, made on the same day with the same kind of goods from the same supplier and involving the same freight forwarder. But the fact that the Respondents conceded that they could not show circularity as to all the goods in one deal, while the Appellant conceded that they had shown circularity in the other deal, emphasises just how difficult general conclusions are in these appeals.
  318. HMRC contend that, save for the concession, the goods in deal 573 came in from Matachel, then went through the usual UK chain of F- TM – PMN – DD to Dragon with the same feature as in 570, 571 and 572 that the goods were split up after arriving with F to be put back together again on sale to PMN. The Atkin schedule shows the onward link is Sunico then Matachel. The tribunal has accepted the provenance of the Atkin schedule, and accepts this evidence as showing circularity in this deal. Accordingly, save again for the concession, circularity is shown in both deals.
  319. Was there any economic activity in either deal (save as to the extent of the concession by HMRC)? As previously noted, both deals were made in the context of a major series of deals made by Dragon on the same day with goods of the same kind from the same supplier, bought (the tribunal finds) at the supplier's price (even thought that varied from £156 a unit to £163 a unit) and sold on to DRT Vertriebs, Sunico or S R Communications in matched back-to-back and largely risk free deals. All involved goods bought in the UK from the same chain for sale elsewhere in the EU, the goods being ones with no obvious UK market. Dragon booked a mark up of about £7 a unit on these deals – a notional gross profit margin on the day approaching £200,000. But wherein lay the added value such that Dragon - and all the other counterparties in the chain - could claim to realise a healthy profit? There is no evidence before the tribunal that any of these deal in any of the chains recorded a loss or that any of the margins were less than 50p a unit, save possibly for the unexplained adjustment in price noted in one deal above. Again, the tribunal can see no source other than the VAT consequences of the deals. That, as the tribunal has indicated before, is not in its view economic activity. Accordingly, it finds that neither deal 573 nor deal 574 involved economic activity.
  320. Deals 584 and 585
  321. 263 This is another set of very similar deals on the same day. Deal 584 involves the purchase of 3,000 7250i units from DD at £156 and the onward sale logged on 27 February to Sunico A/S at £163, with a refund claim for £81,900. Deal 585 has the same date, product, parties, and prices but the quantity is 2,000 and the refund claim is £54,600. The defaulting trade was Furylong in both deals, and the freight forwarder Hawk.
  322. Deal 584 was contended by HMRC to follow the chain of Matachel Developments SL ("Matachel") – F – Masterpiece Technology Ltd ("Masterpiece") – DD – Dragon – Sunico A/S – Matachel. The same chain is contended for deal 585.
  323. Dragon issued a common invoice to Sunico A/S for deals 584 and 585 on 1 March 204, having issued separate purchase order acceptances for the two deals at an earlier stage.

  324. Documents on the Hawk file confirm the inward movement of two batches of 2,500 7250i units on 27 February from Matachel. ("… wich (sic) is still not owner of the goods" is written on the document, but the tribunal has seen no evidence of provenance of that remark and puts no weight on it). The goods came from Copenhagen, but there is no clear evidence of prior ownership. However, the tribunal is satisfied from the documents in the files that Sunico sold these 7250i on to Matachel at a price initially put at £167 a unit and then revised to £168 (the price from Dragon being £163). Circularity is therefore established in respect of both deals.The tribunal therefore does not need to explore further why there appears to be a third party payment from PMN detailed in the papers or other irrelevant documentation. It notes that these goods were again reported on inspection to Dragon to be subject to Customs stamping with no apparent effect on the deal.
  325. For the reasons noted above, the tribunal finds that these deals show no economic activity in the Bond House and Optigen sense. There is evidence showing consistent but controlled mark ups both before and after the Dragon deals. Further, the evidence suggests the prices being altered after the sales had been concluded both up the chain and down the chain from Dragon but not involving Dragon itself. Dragon, however, appears again to be accepting the supplier's price as DD made its usual £1 mark up a unit. The total mark ups for the circuit are £14 a unit on a base price of £154 without any hint that at any stage there was to be onward consumer sale of any of the goods. Wherein lay the source of that added value but from national VAT authorities? The tribunal can see none.
  326. Deals 591, 592, 593, 597
  327. These deals can also be taken together as again involving the same kind of goods, 7250i telephones, logged as purchased on the same day, 1 March 2004, from the same supplier, DD, at what the tribunal takes to be the supplier's price. All involved Furylong as the defaulting trader and Hawk as the freight forwarder. Deal 591 was a purchase of 3,500 7250i at £156.50 with the onward sale to Sunico A/S at £163.50 and a VAT refund claim for £95,856.25. Deal 592 was for 3,400 units purchased at the same price as deal 591 and sold to the same customer at the same marked up price, with a VAT refund claim for £93,117.50. Deal 593 was a purchase of 4,100 7250i units at the lower price of £156.50 and the onward sale to DRT Vertriebs at £164, with a refund claim for £112,288.75. Deal 594 was similar to deal 593, but was adjourned and not fully heard because of late service of evidence. Deal 596, in which the purchase price from DD was much lower, is not before the tribunal as the tribunal was not informed of any appealable decision. Deal 597 was a purchase at £153.50 of 1,985 units and an onward sale to Arkridge at £160 a unit, with a VAT reclaim for £53,322.06.
  328. One major difference between deal 592 and the other deals is that in deal 592 HMRC contends that it can show circularity of goods from the transaction details, while in the other deals it contends that it can show circularity only indirectly through third party payments. Late evidence was, however, admitted from Hawk files about deals 593 and 595. Dragon strongly contended that the evidence in 591 and 593 was particularly weak and relied only on assumption and hypothesis. HMRC rested its case in deals 593, 594 and 595 on the same third party payment. In deal 597 HMRC relied on Irish tax authority evidence to show the link from Arkridge to Mobile World. The tribunal has indicated that it accepts this evidence about Arkridge onward deals.
  329. The chain contended in deal 592 is Matachel – F – Masterpiece – DD – Dragon –Sunico A/S – Matachel and then back to F. There is, it is also contended, a third party payment from Masterpiece to Sunico A/S. HMRC had originally contended an inward movement from DRT Vertriebs but changed this at the hearing. HMRC also contends strong links between deal 591 and this deal including a common third party payment. The chain for deal 591 is the same as that for deal 592. The contended chain for deal 593 starts with F and goes via the same UK buffers to Dragon and from Dragon to DRT Vertriebs and thence (on German tax authority evidence) to Portland Better (a Spanish company). The chain involving deal 595 also follows that route. The chain in deal 597 runs from F to Dragon in the same way, with the goods going outwards to Arkridge and then to Mobile World.
  330. 270 The tribunal accepts that HMRC has shown circularity of goods in deal 592. By reference to the links with that deal, it is also satisfied on the balance of probabilities that the same is shown for deal 591.
  331. Circularity in deals 593 and 595, which again should be considered together (as they are grouped in some Dragon paperwork), depends on less clear evidence, as Dragon submitted. The Hawk file produced for these deals, MD 32553, is in a particularly muddled state with a range of documents involving goods and counterparties that have nothing to do with these deals. Someone has attempted to make sense of some of the documentation by handwritten comments, but in the absence of evidence of provenance, and consistent with its approach in other deals, the tribunal ignores these. Having reviewed the significant documentation for these two deals in the light of the submissions of both parties, the tribunal is unable to conclude that HMRC has shown even on the balance of probabilities the source of the goods that entered the proven chain clearly from F. The tribunal observes that this may be reflected in the fact that HMRC changed its own views during the hearing.
  332. The tribunal has some evidence in the form of covers of and contents of freight forwarder files and evidence in the form of what was termed the "Masterpiece Schedule", evidencing payments reported by the liquidator of Masterpiece as payments shown to have been made by Masterpiece at Furylong's request. But much of the usual Furylong paperwork is absent from these deal bundles. The tribunal puts little weight on the covers of freight forwarder files without supporting evidence, and in this case finds that evidence particularly weak as there are several movements of goods conflated into the file. Further, the payments recorded as made by Furylong on 1 March, contended by HMRC to be the critical date, appears, according the the Masterpiece Schedule, to go to at least five offshore companies (Mobile World, Tocado, Senater, DRT, Sunico and possibly "other") and it is not clear to the tribunal on the evidence what payment went to whom in respect of which deal. The tribunal concludes that circularity is not shown in deals 593 and 595.
  333. It therefore makes no finding about economic activity in deals 593 and 595.
  334. The tribunal fully accepts that the chain of transactions and movement of goods from F (the defaulter) to Mobile World is shown in deal 597. The freight forwarders are Sky High Freight Forwarders Ltd of Cardiff (but with warehouses again near Heathrow). Is circularity shown? There is much less documentation available about this deal (including also the absence of irrelevant information). But the deal went through different shippers to those of the other deals, so the tribunal did not expect to see any information in or from the other deal bundles (as it has in other deals). The only evidence of any weight linking the goods back up the chain to Mobile World in this deal is that of the timing of the deals by Furylong and Masterpiece made clearer by the precise quantity of goods in the deal (1985 units) taken together with the "Masterpiece Schedule" noted above. That schedule shows the Masterpiece PO (purchase order) 900110 of 1 March against the Furylong Sales Invoice #FL480 of that date (both of which are in the deal file) with an amount to be paid of £353,354.81 (the correct amount). The tribunal accepts those details as entirely internally consistent. |It then shows a payment made by Furylong to Mobile World on that date of £332,000.00. HMRC contends that this is a third party payment in respect of that deal, as it has explained in the other deals. The tribunal does not accept this evidence as sufficient to satisfy it. The sums paid to Mobile World in various of the deals at which it has looked do not always equate to the sums from which they are paid. But in other cases the tribunal has seen the third party request and has seen details of the timings and receipts of the payments. These provide internal corroboration. But there is not enough supporting evidence here to show, even to the standard of balance of probabilities, that the payment from Furylong for that deal, and no other deal, went to Mobile World, and no other payee. Accordingly, the tribunal finds that circularity is not established.
  335. The tribunal therefore finds circularity in deals 591 and 592 but not 593, 595 or 597. It makes no finding on economic activity in the latter three deals. Given the clear evidence in deal 592, and the links between that deal and deal 591, and the context of both those deals , the tribunal finds no evidence of economic activity in deals 591 and 592. Its reasons for that conclusion are similar to those explained for other similar deals, namely that the tribunal cannot see wherein lies the opportunity to add commercial or economic value to the chains save from a net input to the chain from the public finances by way of VAT repayments exceeding VAT payments.
  336. Deal 604
  337. Although deal 604 suggests a later deal that the previous deals considered in this decision, and comes before the tribunal stripped for various reasons of the deals next to it in number order, it is a deal logged only one day after the series from 591 to 597, also involving DD, 7250i telephones, Furylong as the defaulting trader and Hawk as the freight forwarder. 1,995 units were purchased at £142.75 a unit – again a very low price – and the sale price was £149 to Arkridge. The input tax claimed was £49,837.59. It was the last of the sales to Arkridge.
  338. HMRC contended evidence of circularity both by reference to the evidence of the transactions in the goods, and their movements, and also by reference to third party payments. The tribunal regards this deal as straightforward as compared with other deals examined above, and is satisfied on the evidence (including that from foreign tax authorities) that there is circularity shown through Mobile World. It similarly has little hesitation in noting that, again, it can see no evidence of any movement of goods to the retail market or other genuine commercial and economic activity in this deal.
  339. Deals 618, 619 and 620
  340. Deal 618 is dealt with in detail above. That discussion also involves deals 619 and 620. All the deals involved 7250i telephones purchased on the same day, logged as 4 March 2004, at very similar prices and sold on to S R Communications at the same price, £163.50. But while deal 618 involved buying 5,000 7250i units from Beechson plc for £157, deal 619 involved a purchase at that price of another 5,000 units from DD and deal 620 involved buying 3,275 units for £156.75 from another company called Topnotch. Dragon first purchased from Topnotch in deal 598 on 2 March and had then entered a number of other deals with that supplier. The defaulting trader was in each case Swift Online (Furylong having disappeared from the scene on 4 March). The freight forwarder was Interken. As noted in the previous examination of these deals, there are strong links between the deals on the Interken files. Dragon claimed a refund of £137,3754 VAT in respect of both deals 618 and 619 and a further £89,837.34 for deal 620.HMRC refused all the claims as it was satisfied in each case that there had been circularity of goods.
  341. The tribunal does not propose to repeat for deals 619 and 620 the full analysis it gave to deal 618. All involve the same defaulter (Swift Online), the same freight forwarder (Interken), the same kind of goods (7250i telephones) and the same link in the chains (S R Communications). The tribunal records that it is satisfied by HMRC that circularity is shown in deals 619 and 620 as well as 618. It also again records, without further discussion here, that it is satisfied that there was no genuine economic activity in any of these deals.
  342. Deal 634

  343. Deal 634 is logged as Dragon's final deal in mobile telephones on 5 March 2004. It purchased 5,000 7250i units from DD at £157.25. But this was not done in isolation. Although the contemporary deals were not before the tribunal as they had all been adjourned, the tribunal notes from Dragon's schedule of deals that all the deals numbered from 624 to 634 were logged on that day, and all but deals 630 and 631 involved 7250i telephones. Deals 624, 625, 626, 632, 633 and 634 involved the purchase of those telephones from DD at various prices. The onward sale in deal 634 was to S R Communications at £164, on which a refund claim for £137,593.75 was made. The basis for disallowance of that refund was circularity of goods. S R Communications also purchased 7250i units from Dragon in deals 626, 627, 628, and 629. The defaulting trader in this deal and the others of which the tribunal has details was Swift Online Ltd. The freight forwarder was Interken.
  344. The evidence from Interken used by HMRC to make its case is subject to detailed criticism by Dragon both as to the details in the physical file and as to the bond check file. The tribunal has already indicated that in general terms its accepts the provenance of the bond check file. The tribunal would not put it as strongly as HMRC that there is "no doubt" that the goods went S R Communications – Metropolitana – Swift Online – TM – PMN – DD – Dragon – S R Communications – Senater, but it is satisfied that HMRC has shown on the balance of probabilities that this was the chain and that therefore circularity was shown. In so finding, it accepts the evidence of the Interken bond check file, as it has in other deals, as confirming and confirmed by the other available evidence. There is also some evidence of a familiar form of third party payment from PMN to S R Communcations that also confirms the finding.
  345. This is again a picture of goods moving to and fro while title goes round a circle involving several EU states. The tribunal yet again sees no evidence of any genuine economic activity in any aspect of this chain.
  346. Deal 638
  347. This is a completely different deal to the others before the tribunal in these appeals. It is the only one of the final series of deals undertaken by Dragon after it ended trading in mobile phones and also decided to increase its investment in individual deals that come before the tribunal in these appeals. But its supplier, DD, was familiar, as was its customer, DRT Vertriebs. However, it was using a different freight forwarder, Skye High, and the defaulting trader was also new, Phone City Megabyte Cafe. The deal was for 28,080 units of Matrox P750 64MB Graphics and Video Cards, purchased at a unit cost of £105.50 and sold on at a unit cost of £109.85. The input VAT refund claimed was £518,427.00. It was refused on the basis that the transaction showed circularity of goods by reference to third party payments. But as the case was put to the tribunal, that was shown only as to 90 per cent of the funding.
  348. The tribunal take that as meaning that HMRC cannot show full circularity but has not conceded this. In addition to the main evidence, the tribunal admitted late evidence about this deal obtained from the German tax authorities. But even with this evidence taken into account, Dragon strongly criticises the basis of HMRC's submission and contends that none of the evidence shows circularity and can only be made to do so by making unwarranted assumptions and hypotheses.
  349. The tribunal finds little assistance in considering this deal from its conclusions about the previous deals involving familiar chains and third party payment procedures. While it may be that anyone engaged in active fraud was now being more careful, it still must be shown to the satisfaction of the tribunal that, on the balance of probabilities, there was a common link on the inward and outward legs of the movement of these cards.
  350. The cards came into the UK from Portland Better SL (based in Spain) to Phone City, thence, but without PMN being interposed at any stage, to TM, then to DD, who sold them to Dragon. Dragon's onward sale was to DRT Vertriebs. It was contended that they then went back to Portland Better.
  351. Deal 638 was one of the deals examined by the tribunal during the course of the hearings in some detail, and the tribunal chairman presented the parties with an informal chronology of events during that hearing. The tribunal does not either repeat or adopt this informal document here save that it confirms its view that the links to establish a chain from Phone City to DRT are fully shown. But additional evidence was admitted at the hearings (and not referred to in the informal document prepared). This provided evidence indicating the onward sale to Portland Better by DRT in three separate transactions, with the freight forwarder's confirmation that the goods were moved. There is also evidence of a request for third party payment to Portland Better from Phone City to (somewhat oddly) an unnamed party, and of a third party payment being made to Portland Better for £700,000 by TM and also to DRT of £1,900,000 both on the same day (and shown as timed not long after the third party payment request was faxed). (Handwritten comments on these documents are again ignored. If accepted in evidence these might explain further aspects of the third party payments as contended by HMRC). The total sum owed to Phone City by TM was £3,451,172.40.
  352. Taking the evidence as a whole (and in particular noting the new evidence), the tribunal finds that the chain from Phone City to Portland Better is clearly shown, and the movement of the goods in 26 pallets also clearly shown. The evidence linking this with DRT on the inward leg is in the new evidence added to the papers. This shows the goods being sent to Skye High "on hold by DRT" on 5 April. It also shows a payment made to DRT from Influx Trading Ltd on that date for £598,000, but why this is with the other papers is not clear. What is established to the required standard by this new evidence is that the goods came into the UK from both DRT and Portland Better (though in what order is not clear save that the goods were on hold from DRT on arrival). The third party payment, originally the only basis for this decision by HMRC, can now be regarded as corroboration, and the tribunal is satisfied on the balance of probabilities of circularity of all the goods. This is also confirmed by the clear evidence of the movement of a precise numbers of pallets of goods of the correct description with clear inspection reports at the relevant time.
  353. Circularity is shown. Is there any genuine economic activity here? The products may have changed from previous deals, and so may some of the parties, but there is much of the familiar choreography of movements in place here. The evidence suggests that the goods (or, rather, title to the goods or to direct equivalents of those goods) went from DRT Vertriebs in Germany to Portland Better in Spain, then to the UK, then back to Germany and back to Spain, all in a short time. Some £3 million plus of goods were moved in that way, but with no attempt apparently made anywhere in the chain to seek any third party consumer market. Given the value of the goods, this again appears as a minimum risk transfer of title and movement of goods to no significant end other than the movement of the goods outward from a series of national VAT authority jurisdictions (including that of HMRC). The goods came to Phonecity as a unit price of £104.60 and were sent back to Portland Better (probably the company that imposed that price) at £110.35 a unit. Why would Portland Better want to enter into a deal involving, and how would it fund a loss of, about £6 (all the prices are marked in £sterling) on each of 28,000 units sold and then bought back in so short a period after incurring significant shipping and insurance costs? If it had a genuine market would it not have moved goods of this high capital value to that genuine market without these circulations? The tribunal finds no genuine economic explanation for this chain of transactions..
  354. One final point of evidence is relevant to the findings of the tribunal in this case. There is a witness statement made by Peter Gaston, an officer of Revenue and Customs, in November 2004. They describe a visit to the Phone City Megabyte Cafι principal address and interviews with those . The statement,which was not challenged and which is accepted, described the office used as the principal address and an individual who introduced himself as a director of the company. It is a sufficient summary of that evidence that it seems entirely inconsistent with a legitimate commercial activity in high value specialist electronic equiment that they be handled by individuals who appear to have no knowledge either of the parts or the transactions, from premises as inadequate and run down as those described, and with the finances and financial problems with bailiffs and others.
  355. The Schedule and general conclusions
  356. The tribunal, having conducted the detailed analyses of individual deals as described, then examined each deal and reached provisional conclusions on them. It finally collated its findings on the attached schedule, and used the benefits of its examination of each case and its provisional conclusions to review its findings as a whole.
  357. The tribunal finds that its conclusions that circularity and lack of economic activity are found in many of the deals strengthened by looking at the whole course of dealing that took place through Dragon during the few months under review in these appeals. Because the evidence about defaulting traders was not in dispute, and because it found what it has termed "familiar patterns" recurring, the tribunal has not felt it necessary to look in more detail at the evidence from HMRC officers available about the United Kingdom defaulting traders and buffer companies. It has noted the speed with which defaulters were replaced and the apparent unexplained overlaps on the transfers. And it noted without comment, save for the final deal, descriptions of those defaulters and the circumstances of the defaults. It also noted available evidence about the buffer companies in some deals. More generally, the nature of those defaulting companies and buffer companies, including the weak finances and inadequate premises noted by officers in the unchallenged evidence, confirms the view that the chains of deals it saw were not a series of genuine commercial arm's length deals that represented an economic activity.
  358. The tribunal takes one illustration of this point. There are repeated transactions through the Furylong – TM Mobiles – Phone Mad Network – DD Trading chain. Those companies always appear in the same order. The HMRC evidence was that the premises of Furylong Ltd were a clothing warehouse in Enfield. Furylong dealt on a regular basis with TM Mobiles. TM Mobiles was described in HMRC evidence as a company that had had numerous trading addresses since registering in 2002. It moved from Northampton to Leicester in April 2004. But its VAT invoice address was still in Northampton early 2004. The tribunal heard no explanation why a company that appeared to be able to move so easily should have such disparate geographical trading partners for these deals. The next company in this chain was Phone Mad Networks Ltd, with an address suggesting a small office in a small town between Southampton and Fareham. Finally, DD Trading was at an address in Birmingham where it could use the Teleos plc fax. That is the only reason the tribunal can see for any of these disparate locations of what was clearly a coordinated group of traders. This is emphasised by noting that all the relevant goods passing through these chains went through freight forwarders in the Heathrow area. Why did the title go through offices spread round the country in this way? The tribunal finds that these considerations strengthen its view that the arrangements were not economic activity.
  359. The Schedule confirms the final decisions of fact of the tribunal on each deal, and the procedural position with regard to each deal as at the date of issue of this decision. It is to be read both as part of this decision as a summary both of the individual decisions of the tribunal in accordance with the directions of November 2004 and also as a statement of the current position for procedural reasons with regard to all others of the conjoined appeals.
  360. For the reasons above, the tribunal formally upholds the appeal of the Appellant in the following deals:
  361. Deal 521(a)

    Deal 521(c) in part only

    Deal 535

    Deal 548

    Deal 573 (as to 200 units only)

    Deal 593

    Deal 595

    Deal 597

    It also records that it accepts an application for costs in respect of these appeals, to be considered as below.

    Directions
  362. The tribunal formally adjourns the appeals about deals in which it has found facts but not upheld the Appellant's appeals pending the decision of the European Court of Justice in Optigen and Bond House. Those appeals are to be considered with the other appeals listed as adjourned in the Schedule at the expiry of six months from the release of this decision, or sixty days after the release of the decision of the European Court of Justice, whichever is earlier. All the deals noted in the Schedule as part heard remain adjourned in accordance with the directions of the tribunal.
  363. Any further appeals made by the Appellant as a result of a decision by HMRC on outstanding claims of the Appellant in respect of any of the deals recorded as undecided in the Schedule are to be joined with the adjourned appeals and considered with them. For the avoidance of doubt, deal 539(b) is included in this direction.

  364. The Appellant asked for costs both on the main appeals in so far as successful (including appeals determined in earlier decisions by the tribunal) and on applications which now form part of this decision. The tribunal grants leave to the Appellant to make a specific application for costs as a result of this and earlier decisions, and in so far as matters are not postponed. The parties shall consider together the extent to which there is common ground as to costs (the tribunal having recorded some common ground in one application) and shall inform the tribunal within one month of issue of this decision of that common ground. If full agreement is not reached on the question of costs, either party is then at liberty to apply to the tribunal for a decision on costs. The tribunal will then make any further necessary directions for submissions and a hearing, and the parties, in making application, are directed to submit any agreed draft directions for such a hearing.
  365. Both parties are at liberty to apply for further directions with regard to any other matter currently postponed, and for any necessary clarification of the effects of this decision on individual repayments of VAT to the Appellant.
  366. David Williams

    CHAIRMAN

    Released: 28 July 2005

    LON/2004/1461

    LON/2004/1462

    LON/2004/1826

    LON/2004/1854

    LON/2004/1880


BAILII: Copyright Policy | Disclaimers | Privacy Policy | Feedback | Donate to BAILII
URL: http://www.bailii.org/uk/cases/UKVAT/2005/V19186.html