Ulster Metal Refiners Ltd v Revenue & Customs [2015] UKFTT 255 (TC)

FTT-Tax
Ulster Metal Refiners Ltd v Revenue & Customs
[2015] UKFTT 255 (TC) · 2015-06-02
[81]“81. It will be recalled that the rationale in Kittel for refusing repayment where the purchaser knows that he was taking part in a transaction connected with fraudulent evasion of VAT was that he "aids the perpetrators of the fraud and becomes their accomplice". For my part I have no difficulty in seeing how the purchaser who is not in privity of contract with the importer aids the perpetrators of the fraud. He supplies liquidity into the supply chain, both rewarding the perpetrator of the fraud for the specific chain in question, and ensuring that the supply chains remain in place for future transactions. By being ready, despite knowledge of the evasion of VAT, to make purchases, the purchaser makes himself an accomplice in that evasion .”35. The defaulter is usually the original importer but any company in the chain or connected chains might dishonestly fail to account for output tax. See Christopher Clarke J in Red 12 Trading Limited v HMRC [2009] EWHC 2563 (Ch) at [84].36. We should also say something about a form of MTIC fraud known as contra trading. The term contra trader is now well understood and has been considered in many cases by the First-tier Tribunal, the Upper Tribunal and the Court of Appeal. We note the description adopted by the Chancellor at paragraph 4 of his judgment in Blue Sphere Global [2009] EWHC 1150 (Ch) as to how contra trading is used to conceal the existence of MTIC fraud. Tax losses do not appear in the clean chains. Rather they appear in transaction chains in which a contra trader acts as a broker dispatching or exporting goods out of the UK, sometimes known as the “dirty chains”. A contra trader offsets input tax it would otherwise re-claim following a broker transaction in the dirty chain against output tax it would pay following an acquisition of goods in the clean chain. The effect of contra trading is that the contra trader is less likely to attract the attention of HMRC because any repayment claim is small. The broker which does make a large repayment claim is at first sight in a clean chain unconnected with fraud.37. The Chancellor in Blue Sphere Global considered and rejected an argument that the connection between transactions in the clean chain and tax losses in the contra trader’s dirty chains was “ unreal and is inconsistent with the principles of legal certainty, fiscal neutrality, proportionality and freedom of movement” . That broad submission was dealt with at paragraphs 44 to 46 of the judgment:[44]“ 44. There is force in the argument of counsel for BSG but I do not accept it. The nature of any particular necessary connection depends on its context, for example electrical, familial, physical or logical. The relevant context in this case is the scheme for charging and recovering VAT in the member states of the EU. The process of off-setting inputs against outputs in a particular period and accounting for the difference to the relevant revenue authority can connect two or more transactions or chains of transaction in which there is one common party whether or not the commodity sold is the same. If there is a connection in that sense it matters not which transaction or chain came first. Such a connection is entirely consistent with the dicta in Optigen and Kittel because such connection does not alter the nature of the individual transactions. Nor does it offend against any principle of legal certainty, fiscal neutrality, proportionality or freedom of movement because, by itself, it has no effect. 45. Given that the clean and dirty chains can be regarded as connected with one another, by the same token the clean chain is connected with any fraudulent evasion of VAT in the dirty chain because, in a case of contra-trading, the right to reclaim enjoyed by C (Infinity) in the dirty chain, which is the counterpart of the obligation of A to account for input tax paid by B, is transferred to E (BSG) in the clean chain. Such a transfer is apt, for the reasons given by the Tribunal in Olympia (paragraph 4 quoted in paragraph 4 above), to conceal the fraud committed by A in the dirty chain in its failure to account for the input tax received from B. 46. Plainly not all persons involved in either chain, although connected, should be liable for any tax loss. The control mechanism lies in the need for either direct participation in the fraud or sufficient knowledge of it ...”38. The test in Kittel applies to a transaction in a clean chain which is part of a fraudulent scheme involving contra trading. This was confirmed by the Court of Appeal in Fonecomp v Commissioners for HM Revenue & Customs [2015] EWCA Civ 39 which was released after final submissions were given in the present appeal.39. Neither party has sought to make any further submissions following the Court of Appeal’s decision. No doubt that is because it has not altered their understanding based on the Court of Appeal’s previous decision in Mobilx and decisions of the Upper Tribunal which followed Mobilx.40. In Fonecomp the trader submitted that the principle in Kittel only applies where the fraud occurs in the same chain of supply. That submission was rejected by the Court of Appeal. In doing so Arden LJ (with whom LJJ McFarlane and Burnett agreed) cited with approval a passage from the decision of the FTT in Universal Enterprises (EU) Ltd v Revenue & Customs Comrs [2014] STC 1515 : “ The argument that a trader in a clean chain cannot be affected by anything which happens in a dirty chain is in my judgment wholly misconceived. Mr Young argued that there is nothing inherently wrong with contra-trading, a statement which, put in that way, is true: a trader who both imports and exports may legitimately organise his sales and purchases so that, at the end of a VAT period, he has little to pay, or a repayment claim. If he does so for reasons of cash flow, his conduct is unexceptionable. But that is not the reason for the contra-trading seen in cases of this kind. As has been said many times, not least by the then Chancellor in Blue Sphere Global Ltd v Revenue and Customs Commissioners [2009] STC 2239 , its purpose is to conceal the fraud in the dirty chain and to make it harder to combat. The appellants' argument necessarily treats "clean" as synonymous with "innocent", but a clean chain in cases of this kind—that is, one in which each of the traders accounts correctly for VAT—is not innocent; it is an integral part of the fraudulent scheme. Even if I entertained any doubt (which I do not) that as a matter of EU law there is sufficient connection between a trader in the clean chain and the default in the dirty chain, there remains an insuperable connection with the fraudulent purpose of the clean chain. ”41. In relation to the nature of the connection and the requisite knowledge Arden LJ also stated as follows:[43]“ 43. … Under the jurisprudence of the CJEU it is for the national court to determine if there was a connection on the facts, and this question is to be determined on the objective evidence and without reference to the trader's knowledge. 44. Furthermore in my judgment, there is nothing in Kittel which would lead to the conclusion that HMRC has to show that the transaction provides tangible assistance in carrying out the fraud. If it did, it would be difficult to prove a connection with a fraudulent transaction upstream of the transaction for which the trader seeks a repayment. Furthermore, contrary to the submission of Mr Lasok, there is no warrant for reading in a requirement that, in a contra-trading case, the connection can be established only by inclusion of details of the transaction in question in a VAT return submitted by (in this case) Klick. ”[48]“ 48. As the UT rightly held, these negative findings did not matter in the light of the other findings of the FTT, which are summarised in the UT's summary. Lack of knowledge of the specific mechanics of a VAT fraud affords no basis for any argument that the decision of either tribunal was wrong in law: what is required is simply participation with knowledge in a transaction "connected with fraudulent evasion of VAT" (Kittel, [61], set out in paragraph 5 above) . ”[49]“ 49. The fraud may involve a complex web of transactions. As Briggs J held in Megtian Ltd v Revenue and Customs Comrs [2010] STC 840 : [37] In my judgment, there are likely to be many cases in which a participant in a sophisticated fraud is shown to have actual or blind-eye knowledge that the transaction in which he is participating is connected with that fraud, without knowing, for example, whether his chain is a clean or dirty chain, whether contra-trading is necessarily involved at all, or whether the fraud has at its heart merely a dishonest intention to abscond without paying tax, or that intention plus one or more multifarious means of achieving a cover-up while the absconding takes place.[50]Mr Lasok further relies on the holding of Moses LJ in Mobilx that the words "should have known" (referring to the deemed knowledge of the trader) meant "has any means of knowing": per Moses LJ at [51]. Mr Lasok further argues that Fonecomp could not have found out about the fraud even if it made inquiries because the fraud did not relate to the chain of transactions with which it was concerned.[51]However, in my judgment, the holding of Moses LJ does not mean that the trader has to have the means of knowing how the fraud that actually took place occurred. He has simply to know, or have the means of knowing, that fraud has occurred, or will occur, at some point in some transaction to which his transaction is connected. The participant does not need to know how the fraud was carried out in order to have this knowledge. This is apparent from [56] and [61] of Kittel cited above. Paragraph 61 of Kittel formulates the requirement of knowledge as knowledge on the part of the trader that "by his purchase he was participating in a transaction connected with fraudulent evasion of VAT". It follows that the trader does not need to know the specific details of the fraud. ”42. Where a connection with fraud is established the focus of the Tribunal is on the control mechanism described by the Chancellor at [46] of Blue Sphere Global , namely whether the Appellant knew or should have known of the connection with fraud.43. The Court of Appeal in Mobilx considered in detail the “knowledge” element of the Kittel principle. It stated in terms at [59]:
“ The test in Kittel is simple and should not be over-refined. It embraces not only those who know of the connection but those who ‘should have known’. ” 44. The respondents must satisfy us that the appellant knew, or should have known that the transactions were connected with fraud. They do not need to establish knowledge of a particular fraud or the fraudulent intent of specific individuals. 45. In any particular case there may be specific details of the fraud that the broker knew about or should have known about. HMRC do not need to establish knowledge of such details. However they must establish that the appellant knew or should have known that there was a connection with fraud. It is not sufficient for HMRC to establish that the broker knew or should have known that its transactions were likely to be connected with fraud (see Mobilx at [60]). 46. Each case must be dealt with by reference to its own facts and on the basis of the test outlined and explained in Mobilx and Fonecomp . 47. The respondents’ case on knowledge is based on drawing inferences from a wide range of facts in order to establish that the Appellant must have known that the Relevant Transactions were connected with fraud (see the same approach recorded at [66] and [67] of the Judgment of Floyd J in Calltel Telecom Limited v HMRC [2009] EWHC 1081 (Ch) ). 48. In the alternative the respondents maintain that in all the circumstances the appellant should have known that the Relevant Transactions were connected with fraud. 49. The meaning of “should have known” is considered at [50] – [52] of the judgment in Mobilx . The Court of Appeal’s conclusions at [52] were that: “If a taxpayer has the means at his disposal of knowing that by his purchase he is participating in a transaction connected with fraudulent evasion of VAT he loses his right to deduct, not as a penalty for negligence, but because the objective criteria for the scope of that right are not met. ” 50. The Court of Appeal gave valuable guidance as to how the “should have known” test should be applied. The guidance is first articulated at [59], where, having observed that the test in Kittel “... is simple and should not be over-refined ,”
Moses LJ stated as follows: “If a trader should have known that the only reasonable explanation for the transaction in which he was involved was that it was connected with fraud and it turns out that the transaction was connected with fraudulent evasion of VAT then he should have known of that fact.” 51. Similar guidance appears elsewhere in the judgment. We approach our task on the basis that the respondents have to satisfy us that the evidence, looked at objectively, demonstrates that the connection with fraud was “ the only reasonable explanation ”, or the only “ reasonable possibility ” or the “ only realistic possibility ” to explain the circumstances in which the appellant entered into the transaction. 52. It is not sufficient for HMRC to establish that the trader should have known that he was running a risk that the purchase was connected with fraud (see Mobilx at [55]). In Davis & Dann v HMRC [2013] UKUT 374 (TCC) the Upper Tribunal at [38] described this as a “ high hurdle ”: “ This test presents a high hurdle for HMRC which we think is most easily appreciated by noting that it is not enough that the circumstances of the taxpayer’s transactions might reasonably lead him to suspect a connection with fraud; nor is it enough that the taxpayer should have known that it was more likely than not that his purchase was connected to fraud. In other words, he can appreciate that everything may not be right about the transaction but that is not enough. He should have known that the transactions in which he was involved were connected to fraud: he should have known that they were so connected because that is the only reasonable explanation that can be given in the circumstances of the transactions. ” 53. In S & I Electrical Plc v Commissioners for HM Revenue & Customs [2015] UKUT 162 (TCC) , which was released after submissions in the present case, the Upper Tribunal described the test in relation to “should have known” in the following terms: “ 34. … the question of whether the “only reasonable explanation” test is satisfied is to be answered at the time of the transactions in question and based upon the matters within the knowledge or which ought to have been within the knowledge of S&I at that time. This is clear from the extract from the judgment of Moses LJ in Mobilx to which we have referred. The decision as to whether in fact, there is a connection with fraud, is made at a different time and may be decided by reference to different evidence. … 64. … the FTT’s task was to apply the impersonal standard of the reasonable businessman to the facts which it found, on the basis of the evidence which it heard, as to the circumstances in which S&I carried out the transactions in issue. Would the reasonable businessman have concluded that S&I ought to have known that the only reasonable explanation for the transactions was that they were connected with fraud? ” 54. As well as clarifying what is meant by the concept of “should have known”, the Court of Appeal in Mobilx also offers some clear and helpful guidance as to how tribunals should approach their fact-finding task. 55. In addressing the question of whether a trader should have known of the connection with fraud, the Tribunal must have regard to all the surrounding circumstances. The relevance of the “surrounding circumstances” is apparent at [59] and [60] of the Mobilx Judgment and at [84] where Moses LJ adopts paragraphs [109] – [111] of the judgment of Christopher Clarke J in Red 12 v HMRC [2009] EWHC 2563 . At [111] of Red 12 Christopher Clarke J said, “... in determining what it was that the taxpayer knew or ought to have known the tribunal is entitled to look at the totality of the deals effected by the taxpayer (and their characteristics), and at what the taxpayer did or omitted to do, and what it could have done, together with the surrounding circumstances in respect of all of them.” 56. As well as having regard to all the surrounding circumstances, the trader and consequently the tribunal must have regard to the inferences that can properly be drawn from the primary facts. This is pointed out at [61] of the judgment in Mobilx : “If he [the trader] chooses to ignore the obvious inferences from the facts and circumstances in which he has been trading, he will not be entitled to deduct.” 57. In relation to the significance of due diligence, the Court of Appeal in Mobilx said this at [75]:[75]“ The ultimate question is not whether the trader exercised due diligence but rather whether he should have known that the only reasonable explanation for the circumstances in which his transaction took place was that it was connected to fraudulent evasion of VAT. ” 58. Then at [82] it said:[82]“ …Tribunals should not unduly focus on the question whether a trader has acted with due diligence. Even if a trader has asked appropriate questions, he is not entitled to ignore the circumstances in which his transactions take place if the only reasonable explanation for them is that his transactions have been or will be connected to fraud. The danger in focussing on the question of due diligence is that it may deflect a Tribunal from asking the essential question posed in Kittel, namely, whether the trader should have known that by his purchase he was taking part in a transaction connected with fraudulent evasion of VAT. The circumstances may well establish that he was. ” 59. We have taken into account the principles and guidance set out above in reaching our decision on this appeal. (3) Detailed Findings of Fact (3.1) General Circumstances 60. The respondents accept that the burden of proof is on them to establish fraud, that the Relevant Transactions were connected with fraud and that the appellant knew or should have known of any such connection. Both parties accept that the standard of proof by reference to which we must make our findings of fact is the balance of probabilities. 61. In making our findings of fact we must consider the detailed evidence before us and what inferences we can properly draw from the underlying evidence. We pay particular regard to the parties’ closing submissions on the evidence. We also bear in mind that in the context of fraud documents should not necessarily be taken at face value. 62. We have had regard to a considerable amount of oral and documentary evidence. We do not refer to all the evidence in this decision, but only to such of it as is necessary to understand the findings of fact we have made in relation to contentious issues. 63. Based on our analysis of the law set out above, the factual issues which we have to resolve on this appeal may be summarised as follows:(1) The identity of traders in the deal chains leading to Irwin and to the Relevant Transactions.(2) Whether the Relevant Transactions were connected with VAT fraud?(3) If so, did the appellant know or should it have known of the connection with fraud? 64. We consider each of these issues in later sections of this decision. Before doing so we make various findings of fact in relation to Irwin, Swift, Paradox, Leonsbeg, Texpor, Oriel, Swan, Magee and PCB. 65. Irwin was controlled by an individual called Mr Fearghal Keenan, albeit he was not a director. Mr Donaldson’s understanding was that Mr Keenan was not a director because he was going through a messy divorce. We have no reason to doubt that was the case. Irwin had an office at Unit 55 Armagh Business Centre and used a warehouse at the premises of a company called Derry Transport in Kilmore. 66. Mr Donaldson had known Mr Keenan since about 2006. Mr Keenan was then running businesses called Diamond Marketing and Navan Wholesale which were wholesalers of soft drinks. The appellant traded with both Diamond Marketing and Navan Wholesale. 67. The appellant traded with Diamond Marketing from 2006 until Diamond Marketing ceased trading in 2008 or 2009. Mr Keenan was a director of Diamond Marketing and Mr Gary Chambers was an employee. 68. The appellant traded with Navan Wholesale in 2009. Mr Keenan was a director of Navan Wholesale and Gary Chambers had been a director in 2006 and 2007. 69. The nature of Irwin’s trade in 2010 and 2011 is highly suspicious and we are satisfied that Irwin was involved in VAT fraud. Indeed that much was common ground between the parties. We make that finding notwithstanding that no enquiry into Irwin was ever completed by HMRC. An enquiry was started in November 2011 but Irwin ceased trading and de-registered for VAT in December 2011. 70. On any view Irwin’s deals in soft drinks were clearly indicative of a scheme intended to defraud HMRC. The documentation obtained by HMRC from Irwin suggests, at face value, that Irwin was purchasing lorry loads of various soft drinks from hijackers such as Landmark. A few days later it was purchasing identical loads of soft drinks from Oriel. The documentation suggests that both loads were then separately sold, generally back to back, to either the appellant or Euromark. Euromark was a RoI VAT registered trader. We refer to the purported deals to Euromark in more detail below. 71. The following summary of some of the deal documentation illustrates the contrived nature of Irwin’s deals. 72. Deal 21 as alleged by HMRC and supported by documentation obtained from Irwin shows goods sold by Landmark to Irwin on 24 February 2011 and by Irwin to the appellant on 28 February 2011. The goods were a specific mix of cases of soft drinks as follows: 720 Coke 500ml 288 Diet Coke 500ml 288 Fanta Orange 500ml 288 Sprite 500ml 144 Fanta Lemon 500ml 73. Irwin’s records also show what purportedly amounts to a separate deal whereby goods of the same description and quantities were sold by Oriel to Irwin on 28 February 2011 and by Irwin to Euromark on 2 March 2011. 74. Transactions from hijacked traders such as Landmark in the other deals often predated the transactions in identical goods from Oriel, at least according to the deal documentation. 75. It is not credible that Irwin would consistently purchase identical loads of soft drinks at or about the same time from different suppliers for onward sale to different customers. Both parties agreed that to be the case. 76. Irwin’s VAT returns from February 2008 to September 2011 show outputs of some £65m and inputs of some £62m, including intra-community dispatches and acquisitions. The output tax due roughly balanced the input tax claimed on a quarterly basis and in the same period of 3½ years the VAT accounted for was only £39,302. 77. It can be seen therefore that the documentation and Irwin’s VAT returns would be consistent with a contra trader. On the face of it Irwin acquired goods from RoI traders, purchased goods from UK defaulters, dispatched goods to RoI traders and supplied goods to UK traders. The net effect was a small quarterly VAT liability. However, as appears later in this decision, neither party suggested that Irwin was a contra trader. 78. Mr Keenan told HMRC officers that he paid UK suppliers such as Landmark in cash which he gave directly to the delivery drivers. We are satisfied that he was lying. It is not credible for example that on 21 March 2011 he would have paid a total of approximately £110,000 in cash for what appear to be 6 separate loads of soft drinks. 79. We note that there is no such suspicion over Irwin’s purchases from Oriel. It appears that following a visit to Irwin in 2011 HMRC officers were satisfied that payment to Oriel was made by telegraphic transfer direct to Oriel’s bank account. 80. Mr Keenan and Irwin had a longstanding connection with Gary Chambers. Gary Chambers had been involved in the soft drinks business for many years. He was on friendly terms with William Kirk of Oriel. He worked as a driver for Diamond Marketing from November 2004 to April 2008. He was a director and company secretary of Irwin from June 2005 to October 2007, at the same time being described as a “sales executive”. He was a director of Navan from May 2006 to October 2007. 81. Gary Chambers employment with Irwin continued until at least April 2010. Irwin hired a lorry from a firm called City West Transport to make its deliveries. City West apparently stipulated that only Gary Chambers could drive the lorry. In his interviews with HMRC, Mr Keenan was contradictory and evasive as to who paid Gary Chambers. In particular whether it was Irwin or City West. PAYE records show that no income was declared for Gary Chambers between 2008 and 2010, a fact which must have been known to Mr Keenan and to Gary Chambers himself. 82. Gary Chambers worked for the appellant from April 2010 until the early summer of 2014. According to Mr Donaldson, Gary Chambers had left Irwin because he was not happy working with Mr Keenan and they had frequent arguments. Evidence from Irwin suggested that Gary Chambers was also employed by Irwin until November 2010. In particular there are 45 Irwin delivery notes purportedly signed by Gary Chambers on behalf of Irwin for deliveries in October and November 2010. 83. The evidence in relation to work done by Gary Chambers for Irwin after April 2010 derives solely from Irwin and Mr Keenan. Given our findings in relation to Irwin we are not satisfied that Gary Chambers was working for Irwin as a delivery driver after April 2010. 84. Gary Chambers’ employment by the appellant was on the face of it as a delivery driver. He certainly had an HGV licence. It was common ground that he was a full time employee on a salary of just £10,000 per annum which works out at £5-6 per hour. Mr Donaldson described Mr Chambers as “ not simply a driver … a good asset to the business ”. However his earnings were very modest even if he was simply an HGV driver. 85. Swift is a company incorporated and registered for VAT in the RoI. It traded and continues to trade in soft drinks. Swift is run by an individual called Paul Martin. The response to a mutual assistance request by HMRC to the Irish Revenue Commissioners in relation to Swift states that “ they have never been accused of involvement in fraudulent activity ”. 86. Mr Keenan had a connection with Paul Martin of Swift. We are satisfied that there was a close relationship between the two men. It arose because Mr Keenan’s fiancée is or was Ms Ashling Mullen. Ms Mullen’s father was a longstanding family friend of Paul Martin’s father. There was no reason why Irwin should not have sold goods directly to Swift rather than to the appellant and none was suggested. Irwin through Mr Keenan knew the identity of the appellant’s customers because until at least October 2010 Irwin delivered goods to the appellant’s customers on behalf of the appellant. Goods purchased by the appellant from Irwin were delivered to the appellant at Derry Transport. Irwin would then deliver directly to the appellant’s customers including Paradox, Swift and Leonsbeg. From June 2010 onwards Paradox was the appellant’s largest customer. 87. Mr Donaldson had a meeting with Mrs Arnold of HMRC on 22 October 2010 at which the risk of MTIC fraud was discussed together with the need to carry out checks on trading partners. Delivery of goods by Irwin to the appellant’s customers was also discussed. Following this meeting Mr Donaldson decided that the appellant should itself deliver goods to its customers. Thereafter the appellant rented a lorry from City West Transport which was driven by Gary Chambers. In about March or April 2011 the appellant stopped using that lorry and used its own smaller lorry, also driven by Gary Chambers. 88. Mr Taylor who appeared for HMRC submitted that Swift was knowingly involved in fraudulent activity. He relied in particular on evidence to the effect that: (1) Swift was involved in other deal chains which included Irwin, although not the appellant, which traced back to Landmark; (2) Swift sold to a large number of missing traders; and (3) Swift made loans to Irwin. 89. There were certainly significant loans made by Swift to companies controlled by Mr Keenan. Beyond the existence of such loans and the existence of a close relationship between Mr Keenan and Paul Martin there was no evidence as to the circumstances in which the loans were made. We are not satisfied on the evidence before us that Swift was complicit in any fraud. 90. Paradox was a company incorporated and registered for VAT in the RoI which traded in soft drinks. It was incorporated on 11 December 2009 and run by an individual called Mark Murray. Paradox ceased trading in August 2012. 91. There was very little evidence as to the nature of Paradox’s trading beyond the deals it entered into with the appellant. We do not have sufficient evidence to reach any conclusion as to the legitimacy of Paradox’s trading generally. We are not satisfied on the evidence before us that Paradox was complicit in any fraud. 92. Leonsbeg is a company incorporated and registered for VAT in the RoI which trades in soft drinks. It is based in the RoI at Carrickmacross and is run by an individual called Peter Martin who is the brother of Paul Martin, with whom Mr Keenan had a close relationship. There was very little evidence as to the nature of Leonsbeg’s trading beyond the deals it entered into with the appellant. We do not have sufficient evidence to reach any conclusion as to the legitimacy of Leonsbeg’s trading generally. We are not satisfied on the evidence before us that Leonsbeg was complicit in any fraud. 93. There was little evidence before us in relation to Texpor. It was associated with L Connaughton & Sons Ltd, a legitimate wholesaler. We have no reason to consider that it was complicit in any fraud. 94. Oriel was the trading name of William Kirk who was a large trader in soft drinks based in Dundalk in the RoI. Mr Kirk ceased trading in April 2013 and his son took over the business through the vehicle of a limited company. Mr Donaldson knew Mr Kirk and there was a small amount of direct trading between Oriel and the appellant. 95. There is no evidence to suggest that Oriel was anything other than a legitimate trader. It had a trading relationship with Coca Cola HBC Northern Ireland Limited (“Coca Cola”) and purchased directly from Coca Cola. It also supplied large volumes of soft drinks to Irwin. 96. We are satisfied that Coca Cola gave certain customers what were described as overriders. Mr Keenan had told HMRC officers that Oriel had the benefit of an overrider from Coca Cola which enabled it to offer competitive prices to Irwin. An overrider is essentially a volume discount. There is a standard wholesale price for sales by Coca Cola to its customers but subsequently a volume discount is received by larger preferred customers. We treat with care any evidence derived from Mr Keenan, however in this regards there was further material to support what Mr Keenan told HMRC. 97. Coca Cola has a manufacturing plant and offices in Northern Ireland near Lisburn and supplies wholesalers in both Northern Ireland and the RoI. In several of the appellant’s transactions soft drinks were supplied by Coca Cola to Henderson Foodservice who supplied the appellant which itself supplied Paradox. These transactions were not challenged by HMRC. They appear from the documentation to show Henderson selling to the appellant at a loss. The existence of an overrider from Coca Cola explains how they are able to do that, because they subsequently have the benefit of a volume discount. 98. There was very little evidence before us in relation to Swan. We had invoices from Swan which show it to be a company trading under the name Swan Wholesale Distributors based at Kernanstown Industrial Estate, Carlow in the RoI. It trades in soft drinks. There is no evidence to suggest that it was anything other than a legitimate trader. 99. Mr Paul Magee lived at Hannahstown, Belfast. He was employed full time as a social care worker at a residential care home. He commenced trading in soft drinks in July 2010, initially with a load of alcohol and thereafter only soft drinks. He had only one supplier which was Mark Cartel. When interviewed by HMRC officers in April 2011 he said that he found new customers by cold calling over the phone and in person. Mr Donaldson said that Mr Magee arrived at Mr Donaldson’s office and introduced himself. That was in or about October 2010. 100. Mr Boyle of PCB lived in Newry and worked full time for the Environment Agency as a dog warden. He commenced trading in soft drinks in April 2010 as he had been told it was “ an easy way of making money ”. One of his suppliers was Mark Cartel. When interviewed by HMRC officers in August 2011 Mr Boyle stated that he had cold called the appellant having seen a sign for M1 Confectioners. That was at or about the end of February 2011. We heard evidence that the appellant and M1 Confectioners were on an industrial estate and it was unlikely someone would happen upon M1 Confectioners by chance. We do not read anything into that. It may or may not be correct. 101. Mr Donaldson carried out certain due diligence or commercial checks on Magee and PCB. We deal with this aspect of the evidence in more detail below when we come to consider our findings of fact relevant to knowledge and means of knowledge. (3.2) The Deal Chains 102. Under this heading we consider the evidence and make findings of fact in relation to the deal chains of which the Relevant Transactions were a part. 103. HMRC rely on statements made by Mr Keenan and Maura Cox, Irwin’s bookkeeper, during the course of VAT visits by HMRC to Irwin. Both said that all supplies purchased by Irwin from UK suppliers such as Landmark, Linkup and Eurolink were sold to UK customers such as the appellant. All supplies from RoI traders such as Oriel were sold to RoI customers such as Euromark. 104. In the light of our findings in relation to Irwin and Mr Keenan above we can place no reliance on unsubstantiated statements from Mr Keenan. Mr Taylor submitted that there was no reason for us not to accept the statements of Ms Cox. However she was not called as a witness and all we know about her is that she was the bookkeeper of a company which was involved in significant VAT fraud. In those circumstances we cannot accept the untested and unsubstantiated statements of Ms Cox. 105. There is no reliable evidence of any movement of goods relating to the Relevant Transactions from UK suppliers to Irwin. There are 6 delivery notes which purport to evidence the delivery of soft drinks from Landmark to Irwin’s warehouse at Derry Transport although not in relation to the Relevant Transactions. 106. Derry Transport is a haulage business mainly transporting chilled goods. It rents warehouse space to Irwin, although there are no Irwin employees based there. The delivery notes are purportedly signed by employees of Derry Transport. Mr Keenan also told HMRC officers that Mr Derry who owns Derry Transport is a cousin of his. We are not satisfied that the delivery notes evidence genuine deliveries of goods from Landmark to Irwin. 107. It is significant that if the deal chains are as alleged by HMRC then Irwin sells at a profit to the appellant and also at a profit on what were parallel sales to Euromark. In contrast, if the deal chains are as alleged by the appellant then in many deals where Irwin’s specific purchase invoices can be identified it sells at a loss to the appellant. If the transactions between Irwin and Euromark actually took place then Irwin achieved a higher profit on those deals which was in excess of and compensated for the loss on sales to the appellant. 108. The point can be illustrated by reference to the documentation for Deal 1. Transactions between Oriel and Irwin and between Irwin and Euromark were in Euros. There was no dispute about the appropriate exchange rate at the time. 109. Looking at the deal chains alleged by HMRC, Oriel sold to Irwin at €12.80 and Irwin sold to Euromark at €13.65. The sterling equivalent profit was 72p per case. Linkup sold to Irwin at £9.79 and Irwin sold to the appellant at £10.29. The profit was 50p. Overall therefore there was a profit of £1.22 per case and £2,109 on the transactions as a whole. 110. Looking at the deal chains alleged by the appellant, Oriel sold to Irwin at €12.80 and Irwin sold to the appellant at £10.29. The sterling equivalent loss was 60p. Linkup sold to Irwin at £9.79 and Irwin sold to Euromark at €13.65. The sterling equivalent profit was £1.82p per case. Overall there was the same profit of £1.22 per case and £2,109 as a whole, but it begs the question why Irwin would sell to the appellant at a loss. 111. The same points arise in relation to deal chains where the appellant alleges that they can be traced back to Swan. 112. Mr Jenkins for the appellant suggested that the loss could be explained by overriders, but there was no evidence at all that Oriel or Swan gave Irwin any overrider or volume discount. Mr Kirk has not given evidence, and there is no statement or explanation from Mr Kirk which describes any volume discount. We are not satisfied that this is an explanation for the losses suffered by Irwin on sales to the appellant. 113. Mr Jenkins also suggested that Irwin might have been content to maintain its trade with the appellant to give a cloak of respectability to other fraudulent transactions, even at the expense of making a loss on such trade. We do not consider that is a likely explanation. 114. On the other hand it is also notable that the deal chains alleged by HMRC involve a substantial difference between the price Euromark was apparently willing to pay, equivalent to £11.61 per case, and that paid by the appellant of £10.29. The evidence before us which we accept is that this was a competitive market with tight but consistent margins. It seems unlikely to us that Euromark would consistently pay so much more than the appellant managed to sell the goods for, mainly to Swift and Paradox. 115. We have no evidence which sheds any light on Euromark and whether it is a legitimate trader. There is also no evidence at all to support delivery of goods to Euromark or that any payment was made by Irwin to Landmark and the other hijackers. We have already found that Mr Keenan was lying when he told HMRC that he paid Landmark drivers in cash. 116. The only documents provided by Irwin to support purchases from Landmark and the other hijacked VAT numbers and sales to Euromark are sales invoices. There are no purchase orders, goods received notes, any evidence of transport or any evidence of payment. 117. Mr Jenkins submitted that on the evidence before us we can be satisfied that no goods passed between the hijackers of UK VAT numbers and Irwin, or between Irwin and Euromark. He submitted that Mr Keenan had dishonestly created a paper trail for transactions which did not take place. 118. One obvious reason why Irwin might wish to create a paper trail would be to offset the input tax credit it would have on the apparent purchase from Landmark and the other UK suppliers, with the apparent sale to Euromark being zero rated, against the output tax which would fall due on its sales to the appellant. If that is the explanation then Irwin was operating in a manner similar to a contra trader. Not necessarily for the purpose of concealing defaulters from HMRC on an enquiry into the appellant’s input tax reclaim, but possibly for the rather more blunt purpose of defaulting on its own output tax liability. 119. That explanation would still leave Irwin making a loss on sales to the appellant. In Deal 1 the loss was 60p per case. However Irwin was not accounting for VAT of £2.05 per case to HMRC and therefore overall it would still make a “profit” of £1.45 per case. It would also leave Irwin making a loss when it could have dealt directly with Swift. In doing so Irwin would have been purchasing from a RoI trader and selling to a RoI trader and would have had no VAT liability. 120. That is one explanation, but we accept Mr Jenkins’ submission that there may be all sorts of possibilities and fraud can be a “web of deceit”. The true nature of the fraud may be different. However, whatever the true nature of the fraud we are satisfied that Irwin never intended to account to HMRC for output tax on its supplies to the appellant. 121. The Swan sales to Irwin are generally for less common soft drinks, including Lucozade. Oriel did not supply Lucozade. HMRC have analysed Irwin’s purchases of Lucozade from Swan which shows that Irwin had no purchases of Lucozade from Swan in May 2011 but that Irwin had sales to the appellant and others in that month. There were however invoices for Lucozade from the UK hijackers in the volumes sold by Irwin to the appellant. 122. We are not satisfied that this leads to a conclusion that the Lucozade must have been supplied originally by a UK hijacker. In particular it does not take into account the possibility that the Lucozade was supplied by Irwin from stock. 123. In the light of all the evidence we think it more likely than not that the deals purportedly involving purchases by Irwin from the UK hijackers and sales by Irwin to Euromark did not actually take place. We accept Mr Jenkins’ submission that these were not genuine transactions, but simply a paper trail created by Irwin. The suppliers to Irwin in the Relevant Transactions were Oriel and Swan. 124. In reaching that conclusion we recognise that there appear to be some small discrepancies in the deal documentation, For example in relation to timing, in Deal 2 the goods were invoiced, delivered and paid for by Paradox on 11 January 2011. The invoice from Oriel to Irwin was dated 12 January 2011. We do not consider such apparent discrepancies to be significant. 125. In addition to the document trail, Mr Donaldson gave evidence which he said supported his case that the goods he purchased from Irwin originally came from Oriel and Swan. He also gave evidence to support the appellant’s case that Oriel had purchased the soft drinks it supplied from Coca Cola. 126. In his witness statements Mr Donaldson said that: (1) Gary Chambers had worked for Irwin for 6 years prior to joining the appellant and was certain that Irwin was supplied by Oriel and Swan and not by UK traders. (2) When Gary Chambers collected the goods from Derry Transport the incoming lorry always arrived with Mr Kirk’s son Aiden as the driver. (3) Gary Chambers would always phone Oriel direct to find out what time the goods were expected to arrive at Derry Transport. Phone records from Gary Chambers demonstrated calls from his phone to Oriel coinciding with some of the purchases from Irwin(4) Mr Donaldson and Gary Chambers saw Swan’s name on the side of pallets purchased from Irwin(5) On one occasion Gary Chambers had asked a delivery driver where the goods were from and had been told they were from Swan Wholesale. 127. In his evidence in chief Mr Donaldson expanded on what he had said in his witness statements. He said that there were occasions when he and Gary Chambers had seen pallets of Coca Cola with the name Oriel written across the pallet. If that was true we would have expected Mr Donaldson to have said as much in his witness statement in the same context as he had described Swan’s name being written on the pallets. 128. Mr Donaldson’s case was that he knew, all along, that his goods originated from Coca Cola in Belfast via Oriel and Swan. If that was true then we would expect him to have offered that explanation at the outset of the enquiry. However in a chain of correspondence with HMRC dating back to July 2011 no such explanation was offered. During the course of that correspondence Mrs Arnold was notifying the appellant which of its deals led back to UK tax losses. A simple answer, which was not given by Mr Donaldson, was that unless Irwin was the defaulter the appellant’s supplies could not trace back to UK tax losses because they traced back to Oriel or Swan in the RoI who in turn purchased from the manufacturers. 129. It is notable that in his correspondence Mr Donaldson did not offer that simple answer, nor did he point the finger at Irwin. The first time Mr Donaldson contended that the goods were supplied to Oriel and Swan from Coca Cola was in his second witness statement dated 8 September 2014. It is notable that in none of his dealings with HMRC prior to the hearing of this appeal did Mr Donaldson point the finger at Irwin. 130. Mr Donaldson volunteered for the first time in his oral evidence in chief that the appellant’s customers wanted “Irish Coke” which was manufactured in Belfast. He said Irish Coke was a bit sweeter than that produced in England. He claimed that he wasn’t sure why his customers asked for Irish Coke, but that they always said when ordering “ make sure it’s Irish ”. As an experienced trader in soft drinks Mr Donaldson would have known exactly why customers would express a preference for Irish Coke. 131. There are some references to Irish Coke in invoices from Swan to Oriel. Whether or not there is any difference between Irish Coke and Coca Cola manufactured elsewhere, if this was a real factor in his dealings with customers and suppliers Mr Donaldson would have identified it well before he came to give oral evidence. 132. Later in his evidence Mr Donaldson sought to bolster the appellant’s case by claiming that in his discussion with Mr Magee he had asked Mr Magee whether it was “ English or Irish drinks that he had? ”. According to Mr Donaldson, Mr Magee had said “ well, you know the English doesn’t sell very well ”. This was a convenient late addition to Mr Donaldson’s evidence. 133. It is clear that if the appellant’s case in relation to the deal chains is right then Gary Chambers would have been able to provide highly relevant evidence. Mr Donaldson was asked why Gary Chambers was not giving evidence on behalf of the appellant. Mr Donaldson said he had asked Gary Chambers to give evidence but he had refused claiming that he would be too nervous. In the context of the issues in this case that is an unconvincing explanation. 134. Mr Donaldson’s oral evidence aside, there was no evidence that Mr Chambers was asked to give evidence or that he refused and if so for what reason. There does not appear to be any reason why a witness summons should not have been requested. We do not accept that the reason Mr Chambers did not give evidence was because he was too nervous. The explanation must lie elsewhere. Indeed Mr Jenkins himself said that Mr Chambers “ may have had his own agenda ”. The most likely explanation in the light of the evidence as to his connections with Mr Keenan is that Gary Chambers was himself aware of Irwin’s fraud. He was plainly trusted by Mr Keenan. 135. In the light of our findings in relation to the involvement of Gary Chambers we cannot accept anything he might have told Mr Donaldson without reliable corroboration. 136. In relation to Mr Donaldson’s evidence on the matters we have just described we are satisfied that his evidence to us was not truthful. 137. Annex 1 to this decision is a summary of our findings in relation to the deal chains for all the Relevant Transactions, and in particular the identity of the supplier to Irwin and whether supplies could be traced back to Coca Cola. We are satisfied that all the Relevant Transactions where Irwin was the supplier trace back to Oriel or Swan and in 10 deals from Oriel to Coca Cola. 138. The supplier in Deal 27 was Henderson Foodservice which in turn was supplied by Coca Cola. It is not clear from the evidence why this has been included as a Relevant Transaction and we shall assume that input tax credit has not been denied in relation to this deal. (3.3) Connection with Fraud 139. In the light of our findings of fact it is clear that the appellant’s purchases from PCB and Magee were connected with fraud. The appellant did not dispute the deal chains and we are satisfied that in each of the 12 deal chains M J Cartel was a fraudulent defaulter. That is sufficient connection with fraud. 140. In relation to supplies from Irwin, Mr Taylor suggested that the appellant was caught up in a carousel fraud where the goods effectively move between a circle of traders. In the present context we could only make such a finding if we were satisfied that Swift, Paradox, Leonsbeg and Texpor were complicit in the fraud. For the reasons given above we are not satisfied that they were complicit or that this was a form of carousel fraud. MTIC fraud can be present in the absence of a carousel. 141. In Blue Sphere Global the Chancellor at [44] (cited above) described the necessary connection with fraud in what were essentially accounting terms. On that basis, the fact that a contra trader offsets input tax incurred in tax loss chains against output tax on acquisition chains involving UK traders who ultimately make intra community supplies would be a sufficient connection. 142. The Chancellor’s description of the necessary connection has never been doubted in any of the subsequent authorities concerned with MTIC fraud. The Court of Appeal held in Fonecomp at [24] that “ Kittel itself is expressed in terms sufficiently wide to cover contra trading ”. 143. We were not invited to treat Irwin as a contra trader and we do not do so. The purpose of Irwin’s transactions with the appellant and setting up the paper trail were not necessarily those of a contra trader as described by Floyd J in Calltel. Those purposes might primarily include hiding the existence a defaulter, but also providing liquidity for the supply chain, rewarding the perpetrator of the fraud and/or ensuring that supply chains remain in place for future transactions. 144. We have found as a fact that Irwin never intended to account to HMRC for output tax on its supplies to the appellant. The existence of a dishonest intent by Irwin to default on payment of output tax due on a supply to the appellant is a more direct connection with fraud than a default in a separate supply chain linked by a contra trader. That is so whether or not as part of the fraud it was necessary for Irwin to generate a VAT liability on a sale to a UK trader. The Court of Appeal in Fonecomp said at [44] that “ there is nothing in Kittel which would lead to the conclusion that HMRC has to show that the transaction provides tangible assistance in carrying out the fraud ”. 145. Even if the fraud did not necessarily involve a default by Irwin on payment of the output tax generated by its supplies to the appellant, the appellant accepts that the fraud involved creating a paper trail of fictitious transactions mirroring Irwin’s supplies to the appellant. It seems to us that in itself would be a sufficient connection with fraud. 146. Mr Taylor conceded that if the deal chains trace back to Coca Cola then there could not be any fraud in those deal chains. However the fraud we have found was not precisely the fraud Mr Taylor invited us to find. HMRC’s case was that the tax loss arose with the hijacked traders. We have found that the tax loss was Irwin’s failure to account for VAT on its supplies to the appellant. Fraud is, as Mr Jenkins put it, a web of deceit. It is not the detail of the fraud which is relevant. It is the existence of fraud, in this case the VAT fraud of Irwin itself. 147. We are satisfied therefore that on the facts as found there is a sufficient connection between the Relevant Transactions and the fraud of Irwin. Whatever the purpose of Irwin’s transactions with the appellant, given the connection with fraud the control mechanism which determines whether the appellant is entitled to an input tax refund is whether the appellant knew or should have known of the connection with fraud. (3.4) Knowledge and Means of Knowledge 148. The respondents must satisfy us that the appellant knew of a connection with fraud at the time it entered into the Relevant Transactions. Alternatively that it should have known of such a connection at that time. There is no direct evidence that the appellant knew or should have known of any specific connection with fraud. Much of the evidence is material from which the respondents invite us to infer that the appellant either knew or should have known of a connection with fraud. 149. The appellant through Mr Donaldson denies that it knew or should have known of the connection with fraud. On the facts of the present case it is what Mr Donaldson knew or should have known that is relevant. It was not suggested that knowledge on the part of the appellant’s employee, Gary Chambers, would be sufficient. In the following paragraphs we record our findings of fact relevant to the issue of what Mr Donaldson knew or should have known. 150. There is no single circumstance from which we can infer that the appellant knew or should have known of the connection with fraud. It is a matter of considering all the circumstances and then whether, on balance, we are satisfied that the appellant knew or should have known of the connection. It is important when considering what the appellant knew or should have known to do so in context. We also acknowledge in the light of our findings in relation to the nature of the fraud that there is no obvious reason why Mr Keenan would need to make Mr Donaldson aware of his fraud. 151. Mr Donaldson was aware of the risk of MTIC fraud when dealing in soft drinks. It was specifically drawn to his attention at the meeting with Mrs Arnold in October 2010. 152. We heard detailed evidence as to the nature of the transactions themselves. 153. Back to back transactions of the type entered into by the appellant are consistent with fraud. However we are satisfied that legitimate trade in soft drinks is also carried out in that way. Further, not all the Relevant Transactions were carried out back to back. It was common ground that on occasion the appellant would purchase goods which were taken into stock and held at a warehouse at the appellant’s premises. On a small number of occasions sales by the appellant were made up of purchases from more than one supplier. 154. The appellant made very similar profit margins on all deals in the two periods. In period 03/11 for all transactions other than Lucozade the purchase price was £10.29 per case and the sales price was £10.59. In period 06/11 the purchase price was £10.59 and the sales price was £10.89. The profit margin was therefore 30p per case which equated to a mark up of 2.91% and 2.83% in the respective periods. 155. For deals involving Lucozade, the purchase price was £9.45 per case and the selling price £9.75 per case, again producing a 30p per case margin equating to a mark up of 3.17%. 156. The appellant also conducted a significant scrap metal business. When it transported scrap metal, the loads were insured through the haulage company. When it dealt with soft drinks the loads were not insured. 157. There was evidence before us as to the due diligence checks carried out by Mr Donaldson into his trading partners. 158. Mr Donaldson’s evidence was that following his meeting with Mrs Arnold in October 2010 his awareness of the risk of trading in soft drinks increased and he therefore increased his due diligence. In his first witness statement he said:
“ In all of the [Relevant Transactions] … we had previously traded with our suppliers and customers, had conducted extensive due diligence and had a long-standing personal relationship with the key figures within the company. ” 159. We have specifically identified the due diligence on Magee and PCB and not the appellant’s other trading partners because that is how the parties, in particular HMRC, presented the evidence. Our findings in relation to due diligence on Magee and PCB contrast markedly with Mr Donaldson’s description. 160. Mr Donaldson’s due diligence on Mr Magee simply established that Mr Magee was VAT registered and that he was who he said he was. Mr Donaldson claimed that he had obtained a credit report on Mr Magee but it was not produced and we are not satisfied that he did. In cross examination Mr Donaldson accepted that as far as due diligence on Mr Magee was concerned “ there’s not a lot ”
. That is all the more surprising because the introduction to Mr Magee was at or about the time of Mrs Arnold’s visit which according to Mr Donaldson led to increased due diligence. 161. Mr Donaldson said in chief that his due diligence on PCB was “good”. In fact it was far from good and simply established that PCB was incorporated, registered for VAT and that Mr Boyle was who he said he was. Documents lodged at Companies House showed that Mr Boyle’s occupation was “enforcement officer”. It turned out he was a dog warden although Mr Donaldson was not aware of that fact because he did not ask him about his occupation. Mr Donaldson did obtain a credit report. It contained very little information because PCB was a newly established company which had not filed any accounts. As at 27 January 2011 it showed that PCB was good for credit of up to £1,000. Mr Donaldson accepted that he only obtained the credit report because following a meeting with Mrs Arnold in October 2010 he understood that HMRC wanted him to obtain such information. 162. Mr Donaldson did at least verify the VAT registration of his trading partners in 2010 and 2011 through a facility that HMRC had in Wigan. 163. We are driven to the conclusion that Mr Donaldson’s description of due diligence in his first witness statement was simply not true. We consider that Mr Donaldson was seeking to exaggerate and mislead in his evidence rather than seeking to assist the tribunal with an honest and candid account. 164. We have identified a number of areas where Mr Donaldson’s evidence has been unreliable. Mr Taylor also relied on a number of other areas which he submitted brought into question Mr Donaldson’s credibility as a witness. 165. Mr Taylor relied upon material relevant to M1 Confectioners which he submitted evidenced Mr Donaldson being at best misleading and at worst dishonest in his dealings with HMRC. 166. In September 2010 Mr Donaldson was the sole shareholder in and director of M1 Confectioners Ltd (“M1”). On 30 September 2010 Mr Donaldson’s accountant, Opus Accountancy Services Ltd, emailed HMRC in connection with certain output tax the appellant had not accounted for on a supply to M1. The email stated:
“ As Ulster Metal Refiners Ltd were in dispute with M1 Confectioners, invoices were not issued to M1. As invoices were never issued Mr Donaldson made the error in not recording them in the VAT Returns. ” 167. Mr Donaldson’s explanation was that he had set up M1 with his accountant Mr Jon Dickinson of Opus. They each held 50% of the shares. Mr Donaldson said that he had put assets, money and time into M1 but Mr Dickinson had not put anything in at all other than its opening capital. The appellant made supplies to M1 but M1 could not make payment for the supplies. Mr Donaldson did not think it appropriate for the appellant to issue VAT invoices until M1 could afford to pay. 168. We accept that there was a dispute, but it was not between the appellant and M1. It was between Mr Donaldson and Mr Dickinson. Mr Donaldson was asked in cross examination whether it would have been honest to state in the email that he controlled both businesses. He did not answer that question directly, but maintained that HMRC were aware that he controlled both businesses. 169. It may well be that HMRC had information available to them which showed that Mr Donaldson controlled both businesses. We have no reason to think that the email was not written on Mr Donaldson’s instructions and with his knowledge. We consider that the email was a less than candid account of the relationship between the appellant and M1. In that sense it is consistent with other features of Mr Donaldson’s evidence which we regard as being designed to mislead in relation to significant matters. 170. Mr Donaldson explained how he dealt with transactions where the appellant was purchasing from a RoI trader and supplying to a RoI trader. One example was a purchase from Oriel on 30 July 2012 where the goods were sold back to back to Leonsbeg. Mr Donaldson sent a lorry from Belfast to collect the goods from Dundalk, brought them into Northern Ireland at Newry and them took them back to Carrickmacross in the RoI. The lorry was in Northern Ireland for just 15 minutes. 171. Mr Taylor suggested that this was not an honest way of doing business. We disagree. As long as the various traders are able to satisfactorily evidence the movement of goods, the supplier is entitled to zero rate the supply. The benefit to a trader such as the appellant is that it does not need to register for VAT in the RoI as well as the UK. It runs the risk together with its trading partners that movement of the goods may be closely scrutinised by HMRC and the Irish Commissioners. 172. It was put to Mr Donaldson that his relationship with Mr Kirk of Oriel should have enabled him to cut Irwin out of the deal chain. Mr Donaldson said that on one occasion in 2010 he had approached Mr Kirk of Oriel to see if he could purchase directly from Oriel the goods that Irwin was supplying to the appellant. Mr Kirk effectively told him that he would not do that because of the greater volumes being purchased by Irwin. When Irwin ceased trading in December 2011 the appellant did trade directly with Oriel in 2012. There is a commercial logic in Mr Donaldson’s account and in the light of our finding that Oriel was a legitimate trader we accept Mr Donaldson’s evidence in this regard. 173. HMRC’s extended verification of the appellant’s input tax claim for period 03/11 appears to have begun with a letter dated 18 April 2011 and a visit on 10 May 2011. On 11 July 2011 HMRC wrote to the appellant stating that the return for period 06/11 was to be the subject of an extended verification. Mr Taylor invited us to draw adverse inferences in relation to Mr Donaldson’s credibility from his responses to HMRC’s enquiries. He submitted that the responses were not those of an honest trader. 174. Mr Donaldson said that he had contacted Mr Kirk and visited him together with his adviser, Mr Ahmed of CTM Ltd. He had first contacted Mr Ahmed on 21 September 2011, shortly after Mrs Arnold’s letter. They then visited Mr Kirk together on 26 April 2012. Mr Kirk had confirmed to them that Oriel had supplied Irwin with the appellant’s goods. Mr Kirk provided copies of some invoices from Coca Cola to Oriel which matched some of the goods supplied by Irwin to the appellant. 175. Mr Donaldson claimed that Mr Kirk had told him that he would help in any way he could but be would not provide a witness statement or give evidence. Given the involvement of Mr Ahmed who was present at the tribunal hearing we are content to accept Mr Donaldson’s evidence in this regard. 176. Mr Donaldson said that the last time he had seen Mr Keenan was around July or August 2012. For the first time in cross examination Mr Donaldson stated in response to questioning that when the extended verification was under way he had tried to contact Mr Keenan by phone and letter. Initially Mr Donaldson said he had never received a reply to his letter. In fact a copy of a reply was later produced. More significantly, Mr Donaldson also claimed for the first time that he had tried to obtain money from Mr Keenan by way of compensation for his losses. 177. There is corroboration of Mr Donaldson’s evidence that he wrote to Mr Keenan and received a reply. Following his evidence a copy of Mr Keenan’s reply dated 8 August 2011 was produced referring to a letter dated 2 August 2011 from Mr Donaldson. It expressed Mr Keenan’s surprise and offered any assistance the appellant might need. We are satisfied that Mr Donaldson was mistaken in his evidence that he didn’t receive a reply. We are not satisfied however that Mr Donaldson ever sought to obtain compensation from Mr Irwin. Nor did he make any serious attempts following the correspondence in August 2011 to press Mr Keenan for an explanation. 178. In similar vein, Mr Donaldson claimed for the first time in cross examination that he had visited Mr Magee’s home in Hannahstown, Belfast. He said that he tried to phone Mr Magee 5 or 6 times and visited his house up to 5 or 6 times but was unable to make contact. He claimed that he never left a message and visited mid-morning when one might expect Mr Magee to be at work. 179. Mr Donaldson also claimed to have contacted Mr Boyle by telephone and told him about the tax losses. Mr Boyle had allegedly responded to the effect that HMRC had contacted him and he had given them the identity of his supplier. He also claimed to have made further calls without response and to have visited Mr Boyle’s home in Armagh, an 80 mile round trip, on 3 or 4 occasions. 180. We are surprised Mr Donaldson had not mentioned in his witness statements his attempts to contact Mr Keenan, Mr Magee and PCB. It would of course be a natural reaction for someone caught up in a fraud to seek explanations and recompense from traders who had sold the goods. Especially in the case of Mr Keenan where the only possibility of a UK tax loss was, on Mr Donaldson’s case, if Irwin had not accounted for VAT. Mr Donaldson’s evidence of attempting to contact Mr Keenan, to seek compensation, and Mr Magee and Mr Boyle was not credible. We do not accept Mr Donaldson’s account of his attempts to contact those individuals. He gave that account because he realised when cross-examined that the reaction of an honest trader would be to seek a full explanation and recompense from the suppliers. 181. Mr Keenan had a close relationship with Paul Martin, of which Mr Donaldson was aware. Mr Donaldson was also aware that Mr Keenan knew that Swift and Leonsbeg were two of his customers. There was no reason we can see that would have prevented Mr Keenan dealing directly with Swift or Leonsbeg which was run by Paul Martin’s brother. No reason was suggested. If Irwin had dealt with Swift and Leonsbeg directly it would have realised an additional margin of 3%. Each load of soft drinks was generally sold by the appellant for a profit of £500 per load in a trade which was characterised by small margins. There was no reason Irwin should not have taken that profit. 182. Mr Keenan also knew in the latter part of 2010 that the appellant’s main customer was Paradox. Again, there was no obvious reason why Mr Keenan should not have dealt directly with Paradox. 183. Mr Donaldson was aware, on his case, that the deal chains involved supplies from Coca Cola to Oriel to Irwin to the appellant and on to its customers. He told us that his customers were wholesalers but would also deal with retailers. Mr Donaldson explained the number of wholesalers and the length of the deal chains by reference to overriders. They give the opportunity to firms like Henderson to undercut other wholesalers who might buy direct from Coca Cola without an overrider. Dealing with retailers is also more labour intensive and requires administering more customer accounts and the logistics of deliveries. 184. We can see that in the legitimate deal chains where for example the appellant was supplied by Henderson Foodservice the deal chains comprised Coca Cola to Henderson to the appellant and on to its customers. There was therefore only one more link in the chains connected with fraud, but that link was Irwin. (4) Decision on Knowledge (4.1) Actual Knowledge 185. We have considered all our findings of fact including the context in which the Relevant Transactions took place. 186. The burden of establishing knowledge is on the Respondents and Mr Taylor placed particular reliance on the following facts and matters: (1) Mr Keenan’s close relationship with Swift and that from Irwin’s perspective there was no need for the appellant to be in the deal chains leading to Swift. (2) Similarly, in relation to Paradox, whilst there was no evidence of a close relationship between Irwin and Paradox, Irwin was aware that Paradox was a significant customer of the appellant. (3) The circumstances in which the appellant employed Gary Chambers given his connection to Irwin, his very modest salary, his intimate involvement in the deliveries and his failure to give evidence. (4) Aspects of the deals which were said to be uncommercial. (5) The circumstances in which the appellant came to deal with Magee and PCB and the lack of any effective due diligence. (6) Various areas of Mr Donaldson’s evidence which lacked credibility. 187. Mr Jenkins submitted that there was nothing in the circumstances of the deals to suggest that they were in any way uncommercial or from which Mr Donaldson might suspect that there was a connection with fraud. Nor was Mr Donaldson’s evidence dishonest or unreliable. In any event, suspicion of a connection with fraud is not sufficient. HMRC must establish that Mr Donaldson knew of a connection with fraud which is a high hurdle. 188. We have made our findings of fact in relation to all the matters relied on by Mr Taylor. We consider the submissions of Mr Taylor and Mr Jenkins against the background of those findings and our findings in relation to Mr Donaldson’s credibility. 189. We have found that there was a connection with fraud in all the Relevant Transactions, apart from Deal 27. Where Irwin was the supplier, the fraud was that of Irwin. Where Magee and PCB were the suppliers, the fraud was at least that of the defaulting trader Mark Cartel. We acknowledge that there was not necessarily any reason why Mr Keenan, Mr Magee or Mr Boyle would wish to make Mr Donaldson aware of fraud. We have not found that this was a carousel fraud. 190. We can deal with certain matters which we do not consider to be probative of knowledge. We are not satisfied in light of the evidence as a whole that the appellant’s rapid rise in turnover in soft drinks in 2010 and 2011 is indicative of knowledge. Similarly the existence of back to back transactions, the lack of insurance and consistent profit margins earned by the appellant. Those factors were all present in the appellant’s deals with suppliers such as Henderson which HMRC accept were not connected with fraud. 191. The fact that Irwin’s supplies to the Appellant were made at a loss is a factor we have taken into account in making findings as to the deal chains. It is not a factor which Mr Taylor urged on us in relation to knowledge and we do not take it into account in this aspect of our decision. 192. We have made findings in relation to the length of the deal chains. We do not consider that the addition of one more trader in the deal chains would ordinarily be indicative of knowledge on the part of Mr Donaldson. However Mr Donaldson did know that the additional trader was Irwin about which he had other knowledge indicative of fraud. 193. It is significant that in relation to the Irwin deals, Mr Donaldson was aware of the close connection between Irwin, Swift and Leonsbeg. He was also aware that Mr Keenan knew the identity of his customers generally, if not in relation to specific deals. We consider that Mr Donaldson must have realised that Mr Keenan was essentially giving up the profits available on a sale to Swift and Leonsbeg without any good reason. That is a significant factor indicating that the deals were in some way connected with fraud. 194. We acknowledge that the same connection did not arise in relation to deals with Paradox, although an honest businessman would have then have been suspicious about any deal with Irwin. Particularly where Irwin knew that the appellant’s main customer was Paradox and there was no obvious reason why it should not deal directly with Paradox. 195. It is telling that Mr Donaldson did not at any stage prior to the appeal point the finger at Irwin or seek recompense from Irwin. He had been told in July 2011 that his deals with Irwin had been traced to a UK tax loss. Knowing that to be the case, an experienced trader such as Mr Donaldson must have known that the tax loss had to be with Irwin. Whilst he wrote to Mr Keenan in August 2011 he did not make any serious attempt to press him for an explanation or to seek recompense. 196. We have found that Gary Chambers was trusted by Mr Keenan and was aware of Irwin’s fraud. The close relationships between Mr Donaldson, Mr Keenan and Gary Chambers might have been explained by Mr Chambers. The failure of the appellant to call Gary Chambers or to give a cogent reason for not doing so is telling. It may be that there were grounds to suspect that Gary Chambers was involved in the fraud, but that is not the reason Mr Donaldson gave. 197. Mr Donaldson dealt with Magee and PCB knowing that they were each new to the trade, had other full time jobs, and were able to source goods cheaper than the appellant. Even though this was a small marketplace with consistent but small margins and where most traders knew one another and shared the same contacts. Far from carrying out “ extensive commercial checks ”
Mr Donaldson did not carry out commercial checks of any substance whatsoever. 198. It is not credible that an honest businessman with Mr Donaldson’s experience, knowing there was a risk of fraud in the soft drinks trade and having been warned of it by Mrs Arnold, would have contemplated entering into deals with Magee and PCB. 199. These factors which are indicative of knowledge on the part of Mr Donaldson become more significant when viewed against the credibility of Mr Donaldson’s evidence generally. 200. We have found that Mr Donaldson gave a less than candid account to HMRC in relation to the appellant’s trade with M1. We have also found that in significant respects Mr Donaldson was not truthful in his evidence. In particular: (1) The oral evidence he adduced to support the appellant’s case that the goods traced back to Coca Cola. In particular his claim of a crucial distinction between Irish Coca Cola and other Coca Cola and his claim that Oriel was written across the pallets. (2) The unconvincing explanation as to why Gary Chambers did not give evidence. (3) The oral evidence as to his dealings with Mr Keenan, Mr Magee and Mr Boyle after it became apparent that HMRC were alleging UK tax losses in the deal chains. 201. In the context of knowledge we have considered the Relevant Transactions as a whole. We have also considered whether we should reach any different conclusion in relation to the separate supplies from Irwin, Magee and PCB, in relation to Irwin’s supplies which trace back to Coca Cola or in relation to supplies by the appellant to Swift, Paradox, Leonsbeg or Texpor. 202. The evidence is such that we are satisfied that the only credible explanation is that Mr Donaldson was aware of the connection with fraud in all the Relevant Transactions, save Deal 27. In relation to Deal 27 it does not appear that there was a connection with fraud. (4.2) Should the Appellant have Known of the Connection with Fraud? 203. The respondents contend that if the appellant did not know of the connection with fraud then it should have known of that connection. The appellant denies that it should have known of any such connection. 204. The facts and matters relied upon by the respondents in this regard are much the same as those set out above in relation to actual knowledge. 205. In S & I Electrical Plc the Upper Tribunal described the test to be applied by reference to the standard of a reasonable businessman as applied to the facts found. For this purpose the focus is on what a reasonable businessman should have known in the circumstances in which the appellant came to enter into the Relevant Transactions. In particular whether a reasonable businessman should have known that the only reasonable explanation for the transactions was that they were connected with fraud. We must also guard against the unfair application of hindsight, knowing as we now do that the transactions were in fact connected with fraud. 206. In summary, the reasonable businessman would have known in particular that: (1) There was a serious risk of VAT fraud in the soft drinks wholesale trade in Northern Ireland. It was important to carry out effective commercial checks into trading partners. (2) There was a close connection between Irwin, Swift and Leonsbeg and Mr Keenan was essentially giving up the profits available on a sale to Swift and Leonsbeg without any good reason. (3) Irwin was aware the appellant’s largest customer was Paradox and there was no apparent reason why Irwin should not supply Paradox directly, retaining the appellant’s profit for itself. (4) This was a small marketplace with consistent but small margins and where most traders knew one another and shared the same contacts. There could be no criticism of Irwin if it had sought to deal directly with the appellant’s customers (5) Notwithstanding the nature of the market place, Magee and PCB were new to the trade and had other full time jobs, yet they were able to source goods cheaper than the appellant. 207. In the light of all the evidence we are satisfied that Mr Donaldson should have known that the only reasonable explanation of the circumstances in which the appellant was able to enter into the Relevant Transactions is that they were connected with fraud. Conclusion 208. For the reasons given above we are satisfied on the balance of probabilities that the Relevant Transactions, excluding Deal 27, were connected with fraud and that the appellant, through Mr Donaldson, knew or should have known of that connection. 209. In all the circumstances therefore we must dismiss this appeal. 210. This document contains full findings of fact and reasons for the decision. Any party dissatisfied with this decision has a right to apply for permission to appeal against it pursuant to Rule 39 of the Tribunal Procedure (First-tier Tribunal) (Tax Chamber) Rules 2009. The application must be received by this Tribunal not later than 56 days after this decision is sent to that party. The parties are referred to “Guidance to accompany a Decision from the First-tier Tribunal (Tax Chamber)” which accompanies and forms part of this decision notice. JONATHAN CANNAN TRIBUNAL JUDGE RELEASE DATE: 2 JUNE 2015 ANNEX 1 Deal No Likely Supplier to Irwin Likely Traced to Coca Cola Appellant’s Customer General Findings 1 Oriel Paradox Differences in dates are not significant 2 Oriel Paradox 3 Oriel Coca Cola Paradox 4 Oriel Coca Cola Swift 5 Oriel Paradox 6 Oriel Coca Cola Paradox 7 Oriel Coca Cola Paradox 8 Oriel Paradox Irwin’s deal was not back to back, but Oriel was likely to be the supplier 9 n/a Paradox Appellant supplied by Magee 10 Oriel Paradox 11 Oriel Paradox 12 Swan Leonsbeg Goods supplied were Capri Sonne (sic) which Oriel did not trade in. Satisfied likely to have been supplied by Swan 13 Oriel Swift 14 Swan Paradox Swan was likely to be the supplier despite some doubt about delivery dates 15 Swan Swift The quantities on the Irwin invoice to the appellant do not match, but we are satisfied that Irwin was the supplier 16 Oriel Paradox Differences in dates are not significant 17 Oriel Swift 18 Oriel Paradox 19 Oriel Paradox Irwin’s deal was not back to back, but Oriel was likely to be the supplier 20 Oriel Swift No Oriel invoice, but it was likely to be the supplier 21 Oriel Paradox 22 Swan Swift Differences in dates are not significant 23 Oriel Paradox 24 Oriel Coca Cola Swift 25 Oriel Paradox 26 Oriel Swift 27 n/a Paradox Traced to Henderson and Coca Cola rather than Irwin. Not clear the input tax credit has been denied. 28 Swan Paradox Differences in dates are not significant 29 n/a Paradox Appellant supplied by Magee 30 Oriel Paradox No Oriel invoice, but it was likely to be the supplier 31 Oriel Paradox 32 Swan Swift Differences in dates are not significant 33 Oriel Swift No Oriel invoice, but it was likely to be the supplier 34 Swan Paradox Differences in dates are not significant 35 Oriel Coca Cola Paradox Differences in dates are not significant 36 Oriel Coca Cola Texpor Appellant’s deals with trading partners not back to back, but satisfied they were likely to be Oriel and Texpor 37 Swan Paradox Differences in dates are not significant 38 n/a Paradox Appellant supplied by PCB 39 Oriel Coca Cola Paradox 40 Oriel Coca Cola Swift 41 Swan Paradox 42 Swan Swift 43 Oriel Paradox Differences in dates are not significant 44 Oriel Paradox No Oriel invoice, but it was likely to be the supplier 45 Oriel Paradox 46 Swan Swift Irwin’s deal was not back to back, but Swan was likely to be the supplier 47 n/a Paradox Appellant supplied by Magee 48 Oriel Paradox 49 Swan Paradox No Swan invoice, but it was likely to be the supplier 50 Swan Paradox 51 Oriel Coca Cola Paradox 52 Oriel Paradox Irwin’s deal was not fully back to back, but Oriel was likely to be the supplier 53 Swan Paradox No Swan invoice, but it was likely to be the supplier 54 Oriel Paradox Differences in dates are not significant 55 n/a Paradox Appellant supplied by Magee 56 Swan Paradox 57 Swan Paradox Goods supplied were Capri Sonne (sic) which Oriel did not trade in. Satisfied likely to have been supplied by Swan 58 Oriel/Swan Paradox No matching invoice but satisfied likely to have been either Oriel or Swan 59 Swan Paradox No matching invoice but goods supplied were Lucozade. Satisfied likely to have been supplied by Swan 60 Oriel/Swan Paradox No matching invoice but satisfied likely to have been either Oriel or Swan 61 Swan Paradox No matching invoice but goods supplied were Lucozade. Satisfied likely to have been supplied by Swan 62 n/a Leonsbeg Appellant supplied by PCB 63 Oriel/Swan Swift No matching invoice but satisfied likely to have been either Oriel or Swan 64 Oriel/Swan Swift No matching invoice but satisfied likely to have been either Oriel or Swan 65 Swan Paradox Irwin’s deal was not fully back to back, but Swan was likely to be the supplier 66 Swan Paradox Irwin’s deal was not fully back to back, but Swan was likely to be the supplier 67 Oriel Paradox 68 Swan Swift No matching invoice but goods supplied were Lucozade. Satisfied likely to have been supplied by Swan 69 Oriel/Swan Paradox No matching invoice but satisfied likely to have been either Oriel or Swan 70 Oriel/Swan Texpor No matching invoice but satisfied likely to have been from Irwin and from either Oriel or Swan 71 Oriel/Swan Paradox No matching invoice but satisfied likely to have been from Irwin and from either Oriel or Swan 72 Oriel/Swan Swift No matching invoice but satisfied likely to have been from Irwin and from either Oriel or Swan 73 Oriel Paradox 74 Swan Paradox No matching invoice but goods supplied were Lucozade. Satisfied likely to have been supplied by Swan 75 Oriel Paradox 76 n/a Paradox Appellant supplied by PCB 77 Oriel/Swan Paradox No matching invoice but satisfied likely to have been either Oriel or Swan 78 Oriel/Swan Paradox No matching invoice but satisfied likely to have been either Oriel or Swan 79 Swan Paradox Goods supplied were Capri Sonne (sic) which Oriel did not trade in. Satisfied likely to have been supplied by Swan 80 n/a Paradox Appellant supplied by PCB 81 Oriel/Swan Paradox No matching invoice but satisfied likely to have been either Oriel or Swan 82 n/a Paradox Appellant supplied by PCB 83 Swan Paradox No matching invoice but goods supplied were Lucozade. Satisfied likely to have been supplied by Swan 84 Oriel/Swan Swift No matching invoice but satisfied likely to have been either Oriel or Swan 85 Oriel/Swan Paradox No matching invoice but satisfied likely to have been either Oriel or Swan 86 Oriel Paradox Differences in dates are not significant 87 n/a Paradox Appellant supplied by PCB 88 Oriel Leonsbeg 89 Oriel/Swan Paradox No matching invoice but satisfied likely to have been either Oriel or Swan 90 Swan Paradox No matching invoice but goods supplied were Lucozade. Satisfied likely to have been supplied by Swan 91 Oriel/Swan Paradox No matching invoice but satisfied likely to have been either Oriel or Swan 92 Oriel Leonsbeg 93 Oriel/Swan Paradox No matching invoice but satisfied likely to have been either Oriel or Swan 94 Oriel/Swan Swift No matching invoice but satisfied likely to have been from Irwin and from either Oriel or Swan 95 Swan Swift No matching invoice but goods supplied were Lucozade. Satisfied likely to have been supplied by Swan 96 Oriel Paradox Differences in dates are not significant 97 Oriel/Swan Paradox No matching invoice but satisfied likely to have been from Irwin and from either Oriel or Swan 98 n/a Paradox Appellant supplied by PCB 99 Oriel Paradox Differences in dates are not significant 100 Oriel Paradox Differences in dates are not significant 101 Swan Leonsbeg 102 Oriel/Swan Paradox No matching invoice but satisfied likely to have been either Oriel or Swan 103 Oriel/Swan Swift No matching invoice but satisfied likely to have been from Irwin and from either Oriel or Swan 104 Oriel/Swan Paradox No matching invoice but satisfied likely to have been either Oriel or Swan 105 Swan Swift Differences in dates are not significant 106 Oriel Paradox 107 Oriel/Swan Paradox No matching invoice but satisfied likely to have been either Oriel or Swan 108 Paradox No matching invoice but satisfied likely to have been either Oriel or Swan 109 Oriel/Swan Paradox No matching invoice but satisfied likely to have been either Oriel or Swan 110 Oriel/Swan Swift No matching invoice but satisfied likely to have been either Oriel or Swan 111 Oriel Paradox 112 Oriel/Swan Paradox No matching invoice but satisfied likely to have been from Irwin and from either Oriel or Swan 113 n/a Leonsbeg Appellant supplied by PCB 114 Oriel/Swan Paradox No matching invoice but satisfied likely to have been either Oriel or Swan 115 Oriel/Swan Paradox No matching invoice but satisfied likely to have been either Oriel or Swan

Cited in 1 later judgment