CCA Distribution Ltd v Revenue & Customs (VAT - MTIC fraud) [2020] UKFTT 222 (TC) (14 May 2020)

DECISION

Introduction

1. More than twelve years ago, on 13 July 2007, HMRC notified the appellant that they would not repay the appellant the input tax incurred on its transactions in the VAT periods 04/06 and 05/06, which amounted to £6,320,368. On 13 August 2007, HMRC notified the appellant that they would not repay the appellant the input tax incurred on its transactions in the VAT period 06/06, which amounted to £3,553,886.54. The basis of these two decisions was that HMRC considered that all the appellant’s deals were connected to fraud and the appellant, acting by its director Mr Ashley Trees, knew this or ought to have known this.

2. The appellant made timely appeals against these decisions, on 2 August 2007 and on 10 September 2007, and at some point they were consolidated into a single appeal under the above reference number. For reasons we do not know, it took nearly five years for the consolidated appeal to be heard in the Tribunal. In the meantime, on 21 August 2009, the appellant, which had not traded after June 2006, entered into administration. The appeal was heard between 15 March and 4 April 2012 and the Tribunal’s decision was released on 22 April 2013. By the chairman’s casting vote, the appeal was upheld.

3. HMRC appealed the FTT’s decision to the Upper Tribunal. On 24 September 2015, the Upper Tribunal allowed HMRC’s appeal, set aside this Tribunal’s decision, and remitted the case to this Tribunal to be re-heard by a differently constituted panel. The appellant appealed the Upper Tribunal’s decision to the Court of Appeal. On 23 November 2017, the Court of Appeal upheld the decision of the Upper Tribunal. Directions issued in this Tribunal in 2018 enabled the case to be set down for its second FTT hearing in 2019, some seven years after the original hearing.

4. This Tribunal, therefore, was charged with a complete re-hearing of the appeal. We make no reference to the original findings of this Tribunal as its decision has been entirely set aside. We have approached this decision entirely afresh. Nevertheless, the earlier hearing is relevant, not only because it explains the very long delay between the transactions at issue in this appeal and this decision, but also because there was a dispute between the parties over the extent to which evidence given in the first hearing would be relevant in this the second FTT hearing. We deal with that below but start this decision with a description of MTIC fraud, which is the type of fraud HMRC allege was at the root of the disputed transactions, and without an understanding of which this decision will make no sense.

The Fraud

5. As we understood HMRC’s case, they alleged that all transactions the subject of this appeal were transactions orchestrated by fraudsters for the purposes of missing trader inter-community (‘MTIC’) fraud. And we do not think the appellant seriously challenged that: it disputed whether one of the three contra-traders was involved in the fraud but so far as we understood, it did not dispute the existence of the underlying MTIC fraud. And in any event, we find it proved beyond doubt as we explain below.

6. But because some of the submissions were based on the nature of the fraud and on assumptions as to how the fraudster would choose to operate we need to explain the fraud we have found proved.

‘Acquisition’ fraud

7. But first we describe a simpler VAT fraud, as the appellant’s submissions were in part based on this. The simpler fraud is referred to as ‘acquisition’ fraud. In its simplest form, it is really very unsophisticated. A fraudster sells goods within the UK and absconds (intentionally) without paying to HMRC the VAT due on the sale. This fraud only operates in a genuine market for the goods: the fraudster needs to identify a genuine buyer for the goods so that he can sell to the buyer the goods and pocket the VAT. It’s referred to as ‘acquisition’ fraud because, in order to make the fraud lucrative, it is important to the fraudster that it obtains the goods free of VAT. So, to obtain them free of VAT, the fraudster exploits EU VAT law which provides that cross-border intra EU sales of goods are VAT free (zero-rated). So the fraudster imports them free of VAT from the EU (‘acquires’ in EU speak) and the fraud is referred to as acquisition fraud. We note that it is possible for essentially the same fraud to be committed without an acquisition from the EU, for instance, if the goods can be obtained VAT free within the UK ( eg by buying them second hand from unregistered persons) they can be sold on and the VAT retained.

MTIC fraud

8. MTIC fraud is superficially similar to acquisition fraud but in fundamentals very different. Both frauds involve VAT free cross-border sales but MTIC fraud is centred on the law that makes a sale of goods from the UK to continental EU zero rated. Traders exporting (‘despatching’) goods to continental EU have to pay VAT to their UK suppliers, but do not charge it to their EU customer. But, as long as they comply with VAT law, such as having proper proof of despatch, they are then entitled to recover from HMRC the VAT paid to their suppliers.

9. MTIC fraud is designed to create a situation where such tax repayments are due, or at least appear to be due, from HMRC. It relies on engineering purchases and sales of goods; unlike acquisition fraud, it does not take place within a genuine market at all.

10. An MTIC fraud therefore involves engineered sales of goods from the UK to another EU country in order to trigger such repayment. For the fraud to be lucrative, however, it would be essential that the purchase of the goods that are to be sold cross-border did not involve a payment of VAT to HMRC as that would defeat the object of obtaining a VAT repayment from HMRC. So goods to be sold to continental EU would also be purchased VAT free from continental EU: while this would generate a VAT liability to be paid to HMRC, the acquirer (ie the entity which imports from continental EU) would simply default on it. In this way an MTIC fraud superficially resembles an acquisition fraud: both involve a default by the acquirer.

11. As MTIC fraud requires the UK acquirer to default on its VAT liability, while the UK despatcher (ie exporter to EU) is needed to make the VAT reclaim from HMRC that was the object of the fraud, they could not be the same person. The acquirer was called the ‘defaulter’ by HMRC; the despatcher the ‘broker’. CCA was accepted to occupy the position of the broker in the series of MTIC transactions we describe below.

The carousel element

12. It is easy to see that the identical goods could be used repeatedly in the same MTIC fraud as MTIC fraud involves an entirely artificial market. It would be cheaper for the fraudster to use the same goods time and time again. The same goods could be sold from the UK to continental EU, from continental EU back to the UK, and then sent round again, each time generating another VAT repayment from HMRC, and each time involving a default by an acquirer. This was why MTIC fraud was often also called ‘carousel’ fraud: the same goods went round and round, actually or purportedly crossing and re-crossing the Channel many times. As none of the participants in the orchestrated transactions had any interest in the goods the subject of them, it would not matter if the goods were in poor condition from their many journeys nor even if they met their contract description.

Control of the broker?

13. There were two different ways an MTIC fraudster who controlled an artificial supply chain could in theory make its money from the fraud, with any variation in-between the two methods being possible. One method was for the broker to be a VAT-registered entity controlled and funded by the fraudster: the fraudster’s profit was quite simply the VAT refund extracted from HMRC.

14. But the fraudster would not necessarily need to control and fund the broker. The fraudster could instead simply identify an independent, VAT registered entity, probably one already in business with a good VAT compliance record, which was willing, and had sufficient funds, to buy the goods acquired by the defaulter and sell them into continental EU. The net price of the goods offered to the broker would be lower than the net price of the goods at which an EU customer, in the artificial supply chain, would offer to buy them: on paper the broker would make a profit.

15. However, the broker would actually pay more for the goods than it would receive because it would buy the goods plus VAT (as it was a UK to UK deal) and sell them to the EU customer free of VAT as explained above. The broker would then reclaim from HMRC the VAT it had paid its supplier : only when it received the reclaim, would it be able to realise its paper profit. This would also be true of any business involved in genuine transactions which bought goods in the UK and sold them into continental EU, where the profit margin was less than the VAT.

16. The fraudster’s profit, however, was the fact that it received more in cash for the goods than it caused the entity acting as the EU customer to pay the broker for the goods. The defaulter, which directly or through a chain of companies, sold the goods to the broker would receive the cash with the liability to account for the VAT element to HMRC but, as we have said, it would intentionally default on that liability. This was fraud.

17. In this type of MTIC fraud, with an independent broker, the fraudster would control to some extent every entity in the supply chain, such as the defaulter, the broker’s supplier and the broker’s customer, but would not control the broker. Therefore, the broker, although a crucial part of any MTIC fraud, did not necessarily have to know anything about the fraud: the broker simply had to be someone willing and financially able to buy and sell the goods. The broker might be under the impression it was trading on a genuine market. It might be an established company which was simply duped into participating in an engineered supply chain.

18. It was the appellant’s case that, in so far as MTIC fraud was proved in this appeal, CCA had been duped into participating as a broker in an MTIC supply chain. As HMRC did not allege fraud against CCA, we understood HMRC to accept that CCA was independent of the fraudster; it was HMRC’s case that CCA nevertheless had participated in the supply chains knowing that they were connected to fraud.

Is the VAT refund important to the fraudster?

19. It is usually said that the input tax reclaim made by the broker is central to MTIC fraud; that is obviously so where the broker is an entity controlled and funded by the fraudster. A circular series of transactions, together with a circular flow of money, is put in place in order to generate the appearance of a genuine trade so that HMRC make a VAT refund in respect of the despatch of the goods to the EU. The object of such a fraud is obviously the VAT refund by HMRC as that is the only money that enters the chain.

20. But, as explained above, where the broker is independent of the fraudster, the fraudster’s profit is more subtle. He realises his profit because the broker pays more in cash for buying goods than it receives from selling them. So, in this type of MTIC fraud, is it really true to say that the broker’s VAT reclaim is central to the fraud? The fraudster has his profit even if the broker is later refused its VAT reclaim by HMRC. This is the position CCA takes. It says that the fraudster has achieved its profit objective when CCA paid it for the goods: the fraudster would not care if CCA was then refused its refund by HMRC.

21. However, we do not agree that it follows that the fraudster has no interest in whether the broker’s subsequent VAT reclaim succeeds. On the contrary, MTIC fraud must be complicated to set up, involving many parties and money movements and the creation of a great deal of paperwork, the movement of actual goods (see §12) and the need to identify a broker who is prepared to participate: the fraudster is likely to want to repeat the fraud as many times as possible to more than recoup the set up costs, so it saves time and effort if he can repeat the fraud with the same parties (and the same goods - see §12). So, it is a logical inference that the fraudster will want the broker’s input VAT claim to succeed so that the broker would agree to enter into further transactions. It’s a logical assumption that a fraudster would do what it could to ensure that the broker recovered its input tax. Logically this is so even where the broker was entirely independent of the fraudster and using its own cash in the transactions, and even if the broker was entirely unaware of the connection to fraud. Therefore, as a matter of logic we cannot accept the appellant’s position that the fraudster was not interested in whether CCA’s VAT claims succeeded.

Distancing the broker from the fraud

22. So, it follows that a fraudster would, logically, distance the broker from the VAT default by the acquirer so that the connection to the default was less obvious and HMRC less likely to deny the input tax refund to the broker. The sort of MITC fraud HMRC alleged to have taken place in this appeal was ‘contra-trading’. HMRC’s position was that contra-trading MTIC fraud was merely a method of distancing the broker from the default and done to support the broker’s VAT reclaim.

23. The appellant did not challenge HMRC’s case that contra-trading took place (save with respect to allegations that Infinity was a party to the fraud, which we discuss below). But the appellant did make submissions that the ‘contra-trading’ nature of the fraud meant that in law CCA (as the broker) was not connected to the fraud and as a matter of fact must have been innocent of the knowledge of it. For this reason, it is necessary to explain contra-trading in order to understand the submissions.

24. The easiest way to understand contra-trading is to consider a normal MTIC VAT fraud without contra-trading. Theoretically, as described above, an MTIC fraud only required 3 entities: a defaulter, a broker and a third entity registered for VAT somewhere in continental EU. The goods would be acquired from the EU customer by the defaulter, the defaulter would sell to the broker, the broker would sell them back to the EU customer. The goods could keep circulating, being bought and sold by these three parties, as long as HMRC kept introducing funds into the supply chain by making VAT repayments to the broker.

25. In reality, however, HMRC were likely to spot that the broker was purchasing from an entity repeatedly defaulting on its VAT liabilities. HMRC might cancel the defaulter’s VAT registration and quite possibly refuse to refund the broker.

26. So, as we have said, it would make sense for the fraudster to distance the broker from the default in order to make it harder for HMRC to see the connection between the VAT reclaim by the broker and the default by the acquirer. So, the fraudster would often interpose between the defaulter and broker a series of UK-VAT registered companies, over which it would be able to exercise sufficient control for the purposes of the fraud. The defaulter would sell the goods to the first of these, the first would sell to the second, and so on, and the last would sell to the broker. These companies would typically make a very small profit such that their net VAT liability to HMRC was very small: they would correctly keep their VAT books and account for VAT on this very small margin. This VAT liability would make a small dent in the fraudster’s profit but the companies’ value was that they would provide distance between the defaulter and broker. If HMRC checked out the broker’s supplier, they would discover an entity which properly accounted for VAT on its margin. It was only if HMRC carried out an ‘extended’ verification, checking the broker’s supplier’s supplier and so on backwards through the supply chain, that the connected default might be discovered.

27. These companies, interposed by the fraudster, were described as ‘buffers’ once HMRC discovered their purpose. They have some relevance to the factual matrix of this appeal (we will revert to this later, for example in §56). But for now, we move on to discuss contra-trading.

Contra-trading

28. As we have said, HMRC’s position was that contra-trading was simply a more sophisticated method employed by fraudsters to distance the broker from the default, making it more likely that the broker would be paid the VAT refund by HMRC.

29. The appellant did not see contra-trading simply as a different method of distancing the broker from the defaulter: Mr Pickup described MTIC fraud with contra-trading as an entirely different kind of fraud.

30. To understand this, it is necessary to describe what happens in a contra-trade. A simple contra-trade involves two chains of goods (or purported goods) instead of one. The broker’s supply chain would, as normal in an MTIC, involve goods acquired from the EU by a UK company controlled by the fraudster. In a normal MTIC chain, this company would default on the VAT liability arising from its acquisition. In a contra-trade, however, the acquirer would not default. On the contrary, it would appear to be VAT compliant because, while it would not pay HMRC the VAT owing, it would submit VAT returns. Its VAT returns would show that it had off-set its VAT liability by incurring an equal entitlement to be repaid VAT by HMRC.

31. It would generate its (apparent) entitlement to repayment by acting as a broker in a second supply chain; in other words, it would sell as much in value to an EU customer as it would acquire from an EU supplier. Of course, this second supply chain would start with an acquisition from the EU: the acquirer in this second chain would default on that liability. If it did not do so, as we have already said, the fraud would not be lucrative. So, the supply chain with the contra-trader acting as broker would be described by HMRC as the ‘dirty’ chain as it was the chain with the default.

32. The effect of contra-trading would be that there would be no default in the (first) broker’s supply chain. The fraudster would have nothing to fear from extended verification of the supply chain. The broker’s supply chain would therefore be referred to by HMRC as the ‘clean’ chain.

33. Mr Pickup suggested, however, that contra-trading was not just a sophisticated mechanism for distancing the broker from the defaulter: he suggested that it meant that the broker was no longer integral to the fraud; he suggested that the broker’s repayment claim was no longer integral to the fraud as (he said) the fraudster made its profit irrespective of whether the broker in the clean chain recovered its VAT.

34. While he did not explain this, the only way he could be correct would be if the contra-trader (like the broker in the clean chain) would pay more in cash for the goods in the dirty chain than it sold them for. This would make the need for a broker redundant, thus, the appellant suggested, CCA as broker was quite unnecessary to the fraud and therefore was most unlikely to have known of it.

35. We do not agree that the broker became unnecessary to the fraud where contra-trading was concerned. Firstly, if the clean chain was not necessary to the fraud, then the fraudster would not engineer it. An engineered chain is a necessary chain. And for the reasons explained below, we find that CCA’s supply chains were orchestrated. Secondly, a moment’s reflection would show that in a contra-trade the fraudster could not extract its profit from the dirty chain alone. Either the contra-trader would be an entity controlled by the fraudster such that extracting any cash from it would not be profit but the fraudster’s own money. Alternatively, if the contra-trader was independent of the fraudster, it would be netting off the cash as well as the VAT. So while the contra-trader in its dirty chains was paying more cash to its supplier than it received from its customer (due to the VAT difference), in the clean chain it was receiving more cash from its customer than it was paying to its EU supplier (due to the VAT difference). The contra-trader would be both cash and VAT neutral: the cash it added to the dirty supply chain came out of the clean supply chain. So, the fraudster could not extract cash from a contra-trader. It needed the broker in the clean supply chain to introduce the cash into the clean supply chain by paying more for the goods it purchased than it received on their sale.

36. In other words, we consider contra-trading was a sophisticated method of distancing the broker from the defaulter by putting the default in a different supply chain to the broker, but fundamentally it was the same fraud. We reject the appellant’s analysis of contra-trading. The broker’s role was as crucial in contra-trading MTIC fraud as in a simpler MTIC fraud. The broker introduced the cash into the chain and for the reasons already given the fraudster would wish the broker to successfully reclaim VAT from HMRC.

37. But it remains the case that the broker could be entirely independent of the fraudster and could be entirely ignorant of the fraud. The broker needed to be someone who was willing to buy and sell the goods.

38. We mention in passing that we have slightly simplified contra-trading to make it easier to understand. In reality, as the evidence in this case shows, a single clean chain would not be off-set by a single identifiable dirty chain. That was unnecessary. All a contra-trader needed to do would be to enter into sufficient dirty chains in one VAT accounting period as equalled in value its clean chains in that same period, so that by the end of the accounting period its net VAT liability to HMRC was nil (or nearly so).

Were contra-traders necessarily a party to the fraud?

39. As we understood it, HMRC’s position was that the logic of the fraud indicated that any contra-trader must, if not actually controlled by the fraudster, be a knowing conspirator in the fraud. The appellant did not agree. The appellant considered that a contra-trader, like a broker, could be duped into participating in engineered supply chains: they would be a part of the fraud but without necessarily being a party to it or knowing of it.

40. His case was that it would be rational for a trader to seek to offset its VAT in the manner of a contra-trader to avoid being out of pocket for a period pending HMRC’s repayment.

41. There were three contra-traders in this appeal: Future, Soul and Infinity. Mr Pickup accepted in closing that the evidence that Future and Soul had behaved fraudulently was over-whelming. He did not accept that was true of Infinity. We consider the facts on this below at §§224-233.

The purpose of contra-trading

42. Another submission which the appellant made on the nature of contra-trading was that, even if its purpose was to distance the broker from the default, the underlying reason for doing so was to keep the broker ignorant of the fraud. So, as we understood Mr Pickup to say, because all the broker deals the subject of this appeal involved contra-trading, putting aside the issue of whether Infinity was a party to the fraud, the Tribunal should more readily assume that CCA was kept in ignorance of the fraud.

43. The appellant’s submissions relied on the description of contra-trading from the Upper Tribunal decision in Fonecomp [2013] UKUT 599 at [7] and its description of its purpose as:

….the dirty input VAT is by this means sought to be transmuted into clean input VAT in the hands of the third trader [ie the broker or someone who sells to the broker]; or at any rate the third trader is sought to be so distanced from the default that he could not know of his connection to it, or HMRC to discover it’

44. In so far as the appellant suggested this description was binding on the FTT, we do not agree. It is not a statement of law. It is a description of how a fraud has been structured and why the Tribunal supposed it has been structured in that way.

45. And while we largely agree with it, there is one element which we consider must be wrong, and that is the element the appellant relies upon. We understood that the appellant particularly relied on the Upper Tribunal’s assumption that the purpose of contra-trading was to distance the broker from the default such ‘that he could not know of his connection to it’.

46. But it is illogical to presume, as the Upper Tribunal appeared to do, that the fraudster introduced contra-trading in order to hide the fraud from the ‘third trader’ (in other words, the broker). It is illogical to think that because contra-trading would not have had that effect: contra-trading does not hide the fraud from the broker as the broker won’t be investigating its supply chain: no innocent broker would expect its supplier to tell it from where it obtained its supply. Contra-trading only hides the fraud from a person who investigates the supply chain. The only person who would normally do that is HMRC.

47. We do agree with the Upper Tribunal that it is logical to presume that the purpose of contra-trading in an MTIC fraud was to hide from HMRC the connection to fraud between the broker’s transaction and the default. If HMRC investigated the broker’s chain of transactions, they would find that all participants in it correctly accounted for VAT and that there was no default. It would be a logical inference by the fraudster that contra-trading would make HMRC more likely to repay the VAT to the broker. Contra-trading must have evolved to hide the connection to fraud from HMRC when undertaking extended verifications.

48. We note that the evidence (see §432) was that in 2003 CCA did pay for an independent firm of accountants to undertake a verification of one of its supply chains because HMRC was withholding a VAT repayment from it. However, the evidence was also clear that, although CCA paid for the line check, its suppliers only consented to it on the basis that CCA was not informed of the identity of any persons above its supplier. The supply chain information was only shared with HMRC. So, while it might be true to say that in 2003, CCA as well as HMRC, discovered that that there was no fraud that particular chain, in general CCA could not know its supplier’s suppliers.

49. In any event, it was not suggested by either HMRC or the appellant that the appellant knew its supplier’s suppliers: apart from that one, limited, instance in 2003, CCA did not and could not investigate its supply chain any further back than its supplier or any further forward than its customer. The fact that the fraud (if proved) involved contra-trading, therefore, says nothing about the appellant’s state of knowledge. The appellant’s state of knowledge depends on the same sorts of things that it would depend upon if the fraudsters had used a chain of buffers or some other means of distancing the default from the broker: those sorts of indicators would be matters such as whether the appellant was offered deals that were too good to be true; whether the appellant truly attempted to avoid connection with fraud, and so on, and we consider these factors below.

Fraudsters preferred ignorant brokers?

50. As we have said, there was no allegation CCA was a party to the fraud. While, subject to the points already made, the appellant accepted that it had acted as a broker in MTIC chains, its position was that it had not known that and could not have known that. Part of its case on this was that it was logical to suppose that a fraudster would do what it could to keep its brokers ignorant of knowledge of the fraud. HMRC’s case was that, on the contrary, the fraudster was likely to prefer a knowing broker.

51. The appellant’s case appeared to be based on the assumption that a broker who knew that the transactions were connected to fraud would refuse to engage in the transaction as it would understand that, if HMRC were to discover the connection to fraud, they were likely to refuse to repay the VAT, leaving the broker out of pocket.

52. However, the fault with the appellant’s submission is that it is undertaking risk analysis with the benefit of hindsight. In May 2006, HMRC refused to repay its entire input tax claim to the appellant and has continued to refuse to repay it; HMRC later refused to repay CCA’s input tax claim for the next two months and the appellant was forced to cease trading. That step was unprecedented; withholding some VAT in 2003 and sometimes being slow to repay (§§349, 431), HMRC had not withheld its entire VAT claim from CCA. It seems to us likely that CCA, or anyone in the position of a broker, would not have anticipated such action by HMRC because people are likely to have assessed risk by reference to the past.

53. Another fault with the appellant’s submission is that it assumes that only brokers with knowledge of fraud would be concerned with risk. But even a broker without any knowledge of fraud in its own supply chain, should have known of fraud in the mobile phone market, and CCA certainly did (see §§344-350). A broker should have understood that trading in that market carried some risk of non-VAT repayment. So, an innocent, as well as knowing broker, might be wary of the risk of non-repayment.

54. Another fault with the appellant’s analysis is that the facts of this particular case (described below at §§329-333) show that a knowing party to the fraud (Future) did choose to take the risk of non-repayment to CCA by HMRC, in transactions where in practice Future’s own cash was at stake, as described below. Future’s contemporaneous assessment of risk was, in the face of its knowledge that the transactions were connected to fraud, that it was a risk worth taking. If CCA had the knowledge of connection to fraud, it may have made the same risk assessment as Future: there is certainly no justification for assuming, as Mr Pickup asks us to do, that CCA would have seen it as too risky.

55. The appellant’s case that the fraudster would prefer an innocent broker was central to its case that the fraudster took active steps to preserve CCA’s innocence; in particular it was the appellant’s case that the fraudster positioned traders on either side of CCA who would manipulate CCA’s director (Mr Trees) into believing he was undertaking negotiations in the genuine grey market for mobile phones which (all parties accepted) existed at the time.

56. The appellant relied on Mr Birchfield’s opinion (see §§161-163 below for discussion of his evidence) that if one person in the supply chain was going to be kept innocent of the fraud, then it would be the broker. And, in theory, we accept Mr Birchfield’s opinion makes some sense. HMRC’s response was that it made no sense on the facts of this case because in 117 transactions in the period in issue, CCA was in the role of buffer and not broker: the necessary implication of the appellant’s case was that the buffers were not innocent of knowledge of the fraud. Mr Pickup’s response to that was that in all these 117 transactions, the broker was actually the contra-trader; so if the contra-traders were parties to the fraud, it was possible for CCA, as the broker-turned-buffer, to still have fraudsters on either side of it intent on pretending to CCA that the transactions were taking place on the open market.

57. Again, at first glance, that appears to make sense. But it overlooks the fact that the fraudster’s alleged purpose in keeping the broker innocent would be to keep the broker happy to enter into transactions time and time again in which it paid more in cash for the goods than it received. It would make no sense to involve such an innocent broker in buffer deals: there would be no profit to be made from the broker because, on the contrary, buffers made a small cash profit on these deals (as they were within the UK - see §26) whereas every transaction with an innocent dupe must increase the risk of the dupe becoming suspicious. So we consider that HMRC are right to say that, logically, it would make no sense for a fraudster who goes to the lengths of setting up mock negotiations in order to give verisimilitude to the transactions with the broker, to then use that broker as a buffer in other transactions in which it would again have to go to trouble to set up the appearance of a real negotiations. There would be no point in taking such a risk as, unlike brokers, buffers did not introduce their own cash into the chain.

58. On the other hand, if the independent broker knew of the connection to fraud, there was no reason why they would not be given the opportunity to participate as a buffer in other chains.

59. So we agree with the appellant that, while in theory, if the fraudster was going to set up mock negotiations to keep one person in the chain innocent of the knowledge that the transactions were orchestrated, that one person would be the broker, we agree with HMRC that that analysis does not hold good where the broker acted as a buffer in other chains. And CCA was broker in 39 chains and buffer in 117 chains in the periods in issue.

60. In conclusion, for the reasons given, we reject the appellant’s case that logic suggests a fraudster would prefer an innocent broker because a knowing broker would refuse to be involved due to perception of risk; we also consider that it tells against CCA’s case that the fraudster nevertheless did try to keep it innocent of the fraud, because we think that logically if that was the case, the fraudster would not have used CCA as a buffer. Nevertheless, we do not think it is appropriate to decide the case on assumptions about what an unidentified fraudster would or would not prefer in the absence of consideration of the specific facts of the case. And we consider those at §§126-469.

The fraudsters preferred a knowing broker?

61. In contrast to the appellant’s suggestion that the fraudster would prefer an innocent broker, HMRC suggested the fraudster would prefer a knowing broker: an innocent broker might try to sell the goods to someone other than the fraudster’s designated EU purchaser. An innocent broker might inspect the goods. They might ask awkward questions. A knowing broker, on the other hand, could be reassured about risk: a knowing broker could be given an explanation of the steps taken to distance it from the default.

62. While we accept there is logic in what HMRC says, as we have said our view is that it is not appropriate to decide the case on the basis of assumptions about what the unidentified fraudster would have preferred, in the absence of consideration of the specific facts in this case. And we consider those at §§126-469.

63. Having described the fraud, and the parties’ competing submissions on it, we now consider the law and what the parties had to say about that.

The law

the kittel doctrine

64. The Court of Justice of the European Union (‘CJEU’) decided in the case of Kittel (C-439/04) in 2006 that:

‘where it is ascertained, having regard to objective factors, that the taxable person knew or should have known, that by his purchase he was participating in a transaction connected with fraudulent evasion of VAT….’

input VAT recovery could be denied.

65. It is clear that this doctrine applies in the UK even though it was not enacted by statute: the Court of Appeal in Mobilx Ltd [2010] EWCA Civ 517 ruled that:

[47] …..Applying the principle in Kittel, the objective criteria are not met where a taxable person knew or should have known that by his purchase he was participating in a transaction connected with fraudulent evasion of VAT. That principle merely requires consideration of whether the objective criteria relevant to those provisions of the VAT Act 1994 are met. It does not require the introduction of any further domestic legislation.

66. The appellant accepted that principle of law but stressed that it was very much an exception to the normal rule that permitted input tax to be deducted as of right. The appellant stated that the right to deduct could only be lost where the taxpayer was proved to have essentially been an accomplice of the VAT fraudster. The appellant relied on Mobilx for this proposition:

[43]…a taxable person who knows or should have known that the transaction which he is undertaking is connected with fraudulent evasion of VAT is to be regarded as a participant, and equally, fails to meet the objective criteria which determines the scope of the right to deduct.

Mr Pickup said the Tribunal must bear this in mind and the Tribunal must not be too quick to disenfranchise taxpayers from right to deduct VAT. We are not entirely sure what he meant by this: the law is clear: if it is proved that the appellant knew or ought to have known that its transactions were connected to fraud, then it loses the right to recover input tax in respect of those transactions. And if that is not proved, the appellant is entitled to its input tax.

Knowledge of what?

67. The trader must be shown to know, or have the means of knowing, its transactions were connected to fraud. In Fonecomp [2015] EWCA civ 39 the Court of Appeal explained:

‘[the trader] …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.

68. In other words, the trader must know (or have the means of knowing) of the connection to fraud but he does not need to know when it is to take place, the identity of the fraudster(s) nor exactly how the fraud was, or is to be, committed.

Means of knowing

69. If HMRC do not prove actual knowledge of connection to fraud, input tax can be denied if they can prove the appellant had the means of knowledge of connection to fraud. What this means was explained by the Court of Appeal in Mobilx, where the judges said it applies to:

[59]…‘those who should have known from the circumstances which surround their transactions that they were connected to fraudulent evasion. 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 fraudulent evasion of VAT then he should have known of that fact.

But the Court also said:

[56] It must be remembered that the approach of the court in Kittel was to enlarge the category of participants. A trader who should have known that he was running the risk that by his purchase he might be taking part in a transaction connected with fraudulent evasion of VAT cannot be regarded as a participant in that fraud. The highest it could be put is that he was running the risk that he might be a participant. That is not the approach of the Court in Kittel. In those circumstances, I am of the view that it must be established that the trader knew or should have known that by his purchase he was taking part in such a transaction.

[60]…Kittel does not extend to circumstances in which a taxable person should have known by his purchase that it was more likely than not that his transaction was connected with fraudulent evasion. But a trader may be regarded as a participant where he should have known that the only reasonable explanation for the circumstances in which his purchase took place was that it was a transaction connected with such fraudulent evasion.

Later, in the case of Davis and Dann [2016] EWCA Civ 142, the Court of Appeal stated:

[4] It was common ground that in order to show that the [taxpayer] ought to have known of the connection between their purchases and fraudulent evasion HMRC had to reach a high hurdle under EU law of showing that they ought to have known that the only reasonable explanation for the transactions was that they were connected to VAT fraud.

70. There was no dispute before us over these principles: it is not enough to deny a person the right to input tax recovery if it can be shown that they had grounds for suspecting that their transaction was connected to fraud. HMRC must go further and show that they had the means of knowing this. A person has the means of knowing it if the only reasonable explanation for the circumstances in which the transaction took place was that it was connected to fraud. HMRC are unlikely to be able to prove this if the circumstances of the transaction were consistent with those of similar transactions on a genuine market: see Upper Tribunal decision in Davies and Dann [2013] UKUT 374 at [56].

How is knowledge and means of knowledge ascertained?

The Court of Appeal in Mobilx at [83] approved what was said in Red 12 [2009] EWHC 2563 about how knowledge/means of knowledge should be ascertained:

To look only at the purchase in respect of which input tax was sought to be deducted would be wholly artificial. A sale of 1,000 mobile telephones may be entirely regular, or entirely regular so far as the taxpayer is (or ought to be) aware. If so, the fact that there is fraud somewhere else in the chain cannot disentitle the taxpayer to a return of input tax. The same transaction may be viewed differently if it is the fourth in line of a chain of transactions all of which have identical percentage mark ups, made by a trader who has practically no capital as part of a huge and unexplained turnover with no left over stock, and mirrored by over 40 other similar chains in all of which the taxpayer has participated and in each of which there has been a defaulting trader. A tribunal could legitimately think it unlikely that the fact that all 46 of the transactions in issue can be traced to tax losses to HMRC is a result of innocent coincidence. Similarly, three suspicious involvements may pale into insignificance if the trader has been obviously honest in thousands.

… 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.

71. The Court of Appeal in Mobilx specifically approved, although it should not really have needed saying, that it is appropriate for the Tribunal to draw inferences from circumstantial evidence; Moses LJ concluded by saying, by reference to the circumstantial evidence available in that appeal:

Such circumstantial evidence… will often indicate that a trader has chosen to ignore the obvious explanation as to why he was presented with the opportunity to reap a large and predictable reward over a short space of time.

thus indicating that it may be legitimate to draw inferences from the trader’s response to the circumstances of the transactions.

Burden of proof

72. There was no issue between the parties on the burden of proof: it is well established that it lies on HMRC: see [81] of Mobilx.

Must HMRC prove that contra-traders were a party to the fraud?

73. There was, however, a disagreement between the parties over whether HMRC had to prove that the alleged contra-traders were a party to the fraud in order to deny the broker its input tax. The appellant relied on what the Chancellor said in Blue Sphere Global Ltd [2009] EWHC 1150 for this proposition:

[55] In my view it is an inescapable consequence of contra-trading that for HMRC to refuse a reclaim by [a taxpayer] it must be in a position to prove that [the contra-trader] was party to a conspiracy also involving [the defaulter]. Although the fact that [the contra-trader] is party to both the clean chain with [the taxpayer] and dirty chain with [the defaulter] constitutes a sufficient connection it is not enough to show that [the taxpayer] ought to have known of the fraudulent evasion of VAT involved in the subsequent dirty chain. At the time he entered into the clean chain there was no such dirty chain of which he could have known, nor was the occurrence of such dirty chain inevitable in the sense of being pre-planned.

74. This passage was presumably influenced what was said by Dr Avery Jones in Olympia Technology Ltd [2008] UKVAT 20570 at [3] which was approved at [4] in Blue Sphere Global where the VAT chairman had said:

…the only way for [HMRC] to refuse repayment of [the broker’s] input tax is to show that [the broker] knew or ought to have known of [the defaulter’s] fraud in a completely different chain, and of [the contra-trader’s] involvement in the fraud.

75. We agree with HMRC that the Chancellor in Blue Sphere was not saying that the broker’s reclaim was only connected with the default if the contra-trader was knowingly a party to the fraud; as Mr Kerr pointed out, the Chancellor had already implied that connection to fraud did not depend on the contra-trader being proved to be a part of the conspiracy:

‘[40] …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 transactions in which there is on common party whether or not the commodity sold is the same. If there is connection in that sense it matters not which transaction or chain came first. ….

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…..

Moreover, in the passage at [55] relied upon, the Chancellor specifically said that it was the act of the contra-trader in being a party to both clean and dirty chain which constituted sufficient connection.

76. On the contrary, the Chancellor, in the passage relied upon, seemed to be saying something about the nature of the broker’s knowledge, rather than anything to do with connection. The difficulty with that is that it is clear from the later Court of Appeal decision in Fonecomp [2015] EWCA Civ 39 that the broker only has to be shown to know of a general connection to fraud: the broker does not have to be shown to know of the details of the fraud. The broker will lose its right to a VAT refund if it knows its transactions are connected to fraud, even if it does not know it was an MTIC fraud involving contra-trading. At [48] Arden LJ in Fonecomp said that

‘what is required is simply participation with knowledge in a transaction “connected with fraudulent evasion of VAT”’

The Judge went on to cite with approval Briggs J in Megtian Ltd [2010] STC 840 where he said:

[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.

And Arden LJ went on to say:

[51] However, in my judgment, the holding of Moses LJ [in Mobilx] 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.

77. Indeed, any proposition that the claimant had to have actual knowledge of the details of the fraud would appear quite inconsistent with what was said in Mobilx, where persons said to know of connection to fraud included:

[59]… those who should have known from the circumstances which surround their transactions that they were connected to fraudulent evasion. 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 if it turns out that the transaction was connected with fraudulent evasion of VAT then he should have known of that fact. He may properly be regarded as a participant for the reasons explained in Kittel.

78. Paragraph [55] of Blue Sphere might therefore appear inconsistent with the later and binding decision of the Court of Appeal in Fonecomp, but as Mr Pickup pointed out, the Court of Appeal in Fonecomp approved the decision in Blue Sphere. So [55] of Blue Sphere, he says, represents good law. But if it is good law, what is it saying? Its logic appears to be that if the contra-trader was not a party to the fraud, its ‘clean’ chain transactions can’t have been orchestrated by the fraudster and so the broker couldn’t know about the fraud as there would be nothing to know: its transactions were taking place on the open market. Indeed, in Blue Sphere it was a finding of fact that the alleged contra-trader (Infinity) was innocent of any knowledge of the fraud.

79. If [55] of Blue Sphere is good law, it seems it is saying that if the clean chain is not orchestrated by the fraudster, there is no connection to fraud, or the broker in the clean chain could not know of the fraud as its transactions were not orchestrated. Either way, the premise of the statement in [55] was that the contra-trader in the dirty chain was not a party to the fraud and/or the clean chain was not orchestrated. Whether [55] is good law therefore only matters if we find that the alleged clean chains were not orchestrated and/or we find that the alleged contras were not parties to the fraud.

80. And that brings us to another legal issue which arose which was whether, as a matter of law, it was open to HMRC to allege fraud against Infinity, one of the three alleged contra-traders. But, if we were required to decide whether [55] was good law, we would agree with HMRC that it is not good law, despite the apparent approval of Blue Sphere as a whole in Fonecomp, as it is inconsistent with the principle set out by the Court of Appeal in Mobilx and Fonecomp that a trader who enters into a transaction which is connected with fraud (even if the connection is via an offset by a contra-trader) and who knows that or ought to know that, is not entitled to recover its input VAT. Therefore, we do not think we have to resolve whether it is open to HMRC to allege fraud against Infinity.

81. Nevertheless, because there is clearly doubt on whether [55] of Blue Sphere is good law, we will decide as a matter of fact whether the clean chain transactions were orchestrated and whether the contra-traders, including Infinity, were knowing parties to the fraud. Because if HMRC can prove both these matters, it is irrelevant whether [55] of Blue Sphere is good law or not. We reach our conclusions on the facts on orchestration at §§135-209 and on Infinity’s knowing participation in the fraud at §§224-233 below.

Could HMRC allege fraud against Infinity?

82. And, as we have said, that brings us to the legal question whether HMRC could allege fraud against Infinity. To explain that legal question, we need to outline some undisputed facts.

83. And those facts are that there were two ‘Infinity’ companies, both owned and controlled by a Mr Simon Thakor. The question of whether either company was fraudulent was really a question whether Mr Thakor had acted fraudulently.

84. The first of the two companies was Infinity which was one of the three alleged contra-traders in this appeal. Distribution was the other company. It successfully appealed against HMRC’s decision to refuse to repay it £12 million input tax on the basis it knew or ought to have known its deals were connected to fraud: its success came about because HMRC were barred from the appeal due to non-compliance with directions from the Tribunal. The substantive issues were never tried.

85. As well as that litigation, HMRC assessed Distribution for a very significant sum on the basis that it had reclaimed input tax on its sales of Samsung Serenes and P990s, and also zero rated sales which HMRC now maintained were ineligible for such treatment. Those decisions are under appeal with active litigation in this Tribunal. And while HMRC had originally served evidence in that appeal which (HMRC claimed) tended to show fraud, HMRC had not pleaded fraud, and had withdrawn that evidence on the basis that they did not allege fraud against Distribution.

86. HMRC’s position in this appeal was that they had chosen not to allege fraud in Distribution’s appeal as they could win that appeal simply be proving the phones did not exist and the sales did not meet the formal criteria for zero-rating; there was nothing abusive, they said, in alleging fraud against a company owned by Mr Thakor in some proceedings but not others. The appellant did not agree.

87. Putting aside the fact that Infinity and Distribution are not in law the same entity, and proceeding on the basis that they should be treated as single entity because of their mutual owner, we do not agree that it is abusive as a matter of law for HMRC to allege fraud against Infinity. Mr Pickup did not cite any authority or really explain why it would be abusive for HMRC to allege fraud against Infinity in these proceedings.

88. While it was true that HMRC did not allege fraud against Distribution in the other proceedings, those proceedings would not determine whether or not Distribution or Mr Thakor was fraudulent as it was not a live issue in that appeal. So by alleging fraud in these proceedings, HMRC would not be undermining a finding by another Tribunal in proceedings in which they were or are a party; moreover, even if we found Infinity to be fraudulent, it would not be open to HMRC to rely on that finding in proceedings to which Infinity, Distribution or Mr Thakor was a party, as those entities were not parties to, nor Mr Thakor a witness in, these proceedings.

89. We did not think it abusive for HMRC to maintain in these proceedings that Infinity was a knowing party to the fraud.

90. We note, almost in passing, that Mr Pickup did point out that it was a finding of fact in Blue Sphere that Infinity was not a knowing contra-trader (see §78 above). However, that finding only binds HMRC in the litigation with Blue Sphere. It is not abusive for HMRC to allege otherwise in different appeals with a different broker. We do not know the evidence which was produced in Blue Sphere: we can only base our findings of fact on the evidence before this tribunal. We will consider that evidence and draw our own conclusion on whether Infinity was a knowing party to the fraud: see §§224-233.

Interlocutory matters

91. Having dealt with the substantive law which is relevant to this appeal, we turn to record the interlocutory decisions which were reached. When the hearing commenced on 30 April 2019, the Tribunal was asked to resolve three applications by the appellant:

(1) To strike from HMRC’s statement of case various statements which the appellant considered amounted to an allegation of dishonesty or fraud on the part of the appellant;

(2) To strike from HMRC’s statement of case the reference to evidence given at the previous hearing;

(3) To adduce the evidence of a Mr Smith as an expert on grey market trading.

The Tribunal decided these applications and announced its decision orally at the time, but we record the decisions now in writing.

The application to adduce Mr Smith’s evidence

92. In the original hearing, both parties had each adduced an expert witness on grey market trading in mobile phones. At the time directions for the re-hearing were given in July 2018, the parties were uncertain whether expert evidence would be called: the directions envisaged that if there was any expert evidence, the parties would serve it at the same time as their factual evidence. HMRC’s evidence was due on 17 September 2018 and the appellant’s was due on 19 November 2018. The directions envisaged that if it was not going to be possible to serve the expert evidence by then, the parties would no later than 30 September 2018 agree a new timetable for it. No such timetable was agreed, it seems because in August 2018 the parties agreed that salient opinions of the previous experts’ evidence were not really contentious (covering matters such as the size of the grey market) and could be reduced to an agreed statement to be provided to the Tribunal. And such a statement was produced and is before this Tribunal (see §§435-437).

93. Then in March 2019, some two months before the hearing, the appellant applied to rely on expert witness evidence from a Mr Smith. HMRC opposed the application. The judge directed that the application be heard at the start of the hearing and it was so heard and determined. For reasons explained below, the Judge sat alone on the application.

94. The appellant’s position was that it had not made the application so late that HMRC had been unable to respond with their own expert evidence; moreover, the evidence concerned a matter not covered by the agreed statement of expert opinion referred to above. Mr Smith was put forward as an expert in trading practices in the grey market in the mobile phone industry at the time of the transactions at issue in this appeal. Having worked in the mobile phone industry for over 25 years, at the relevant time, Mr Smith had been the director of JDI which traded in mobile phones. Mr Pickup explained that the importance of his evidence was that HMRC relied on a number of features of CCA’s trading as being indicators of fraud, whereas Mr Smith’s evidence was that those features were normal in legitimate grey market trading and not indicators of fraud at all. Therefore, it would be the appellant’s case that CCA could not have known from those factors that its transactions were connected with fraud (if they were).

95. HMRC objected to Mr Smith’s opinion evidence being adduced on the basis it was not relevant because they considered Mr Smith’s experience was in deals that were engineered for the purpose of MTIC fraud and not in the genuine grey market at all. HMRC said this because JDI had been denied input tax by HMRC on some 9 deals which occurred in July 2006 on the basis they were connected to fraud. While JDI’s appeal against this decision had succeeded, it succeeded on the grounds that the tribunal had found that the company acting by its directors had not known, nor could have known, that its transactions were so connected: [226] of JDI Trading Ltd [2012] UKFTT 642 (TC). The Tribunal recorded at [210]:

… we have no hesitation in finding that there was a contrived scheme for the fraudulent evasion of VAT …. It is also clear and undisputed that the transactions with which this appeal is concerned were connected to that fraud.

HMRC’s position was that the JDI decision showed that JDI’s only trades in the entirety of 2006 had been those at issue in the appeal, bar one which was in an earlier period and on its face very similar to those at issue in the appeal. So HMRC’s conclusion was that JDI had not traded in the genuine grey market so Mr Smith was not an expert in grey market trading.

96. HMRC also considered that this Tribunal was not bound by the findings of fact in JDI and HMRC did not accept the tribunal’s finding that Mr Smith had not known, nor should have known, that JDI’s transactions were connected to fraud: they said, if Mr Smith was a witness in this hearing, they would want to adduce evidence that he had known, or should have known, of the fraud.

97. The appellant accepted that defaults were found in some of JDI’s supply chains but it was its position that JDI traded in a legitimate grey market notwithstanding the defaults. The appellant also relied on Mr Smith’s prior experience working for authorised distributors, which had included transactions with traders on the grey market. The appellant considered that this Tribunal was bound by the findings of fact in JDI, as findings of fact between the same parties could not be re-opened, and so it was not open to HMRC to question Mr Smith about his alleged knowledge of the fraud.

98. The day before the hearing commenced it had become apparent to the Tribunal panel that the Tribunal Member, Ms Gill Hunter, would probably be unable to continue to sit as a member of the Tribunal hearing this appeal if Mr Smith was a witness because of the appearance of bias: Ms Hunter had been a member of the panel that had sat on JDI Trading Ltd. It followed that the Tribunal had recognised, as the parties also agreed, that it was not appropriate for Ms Hunter to be a part of the panel while it was deciding this application. Mrs Hunter had therefore left the hearing room for the entirety of the application and its determination and, as we have said, the Judge heard and decided the matter on her own. The following ruling was by the judge alone.

99. My full ruling was recorded in the transcript and I will only summarise it here. I said that in so far as Mr Smith, from his trading experience prior to his time at JDI, could give evidence about white market trading, and white to grey market trading, this evidence was unnecessary as the parties had already agreed a statement on this and it was irrelevant in any event as no one was suggesting that was the type of trading at issue in this appeal.

100. The real significance of Mr Smith’s evidence was on his evidence on sales within the grey market: grey to grey trading. I thought HMRC’s challenge should properly be characterised as a challenge to Mr Smith’s expertise rather than to the relevance of his evidence, but how it was characterised did not matter. The question was whether Mr Smith actually had knowledge of trading between grey market traders in a genuine grey market.

101. A fair reading of the previous Tribunal’s decision was that virtually all of JDI’s trading was considered in that hearing, and all of it that was considered was found to be connected to MTIC fraud of the type alleged in this appeal. That meant that that Tribunal had found that JDI’s transactions had been brought about by fraudsters for the purpose of the fraud and by necessary implication were not trades in a genuine grey market.

102. There was a dispute between the parties whether this tribunal could go behind the findings of another tribunal or whether it was bound by such findings. I considered, based on Calyon v Michailidis [2009] UKPC 34 that the Tribunal’s finding in JDI was not binding on this Tribunal as the proceedings were between different parties: the respondent was the same but the appellants were different. That conclusion was unaffected by the fact that the main witness from JDI would also (if the application was allowed) be a witness in this case.

103. But my view was that either way, I would not admit the evidence. If the Tribunal was bound by the findings in JDI, then it appeared JDI’s known trading was not on a genuine grey market: if the appellant sought to show that JDI had other trading which was on a genuine grey market then this point was clearly contentious and there was no time for the relevant evidence to be served, considered or tested. So, if the Tribunal was bound by the JDI findings, the appellant had failed to satisfy me that Mr Smith was an expert in the matter on which the appellant tendered him as an expert witness: put another way, I was not satisfied that he could give relevant evidence.

104. But if this Tribunal could go behind the findings in JDI Trading Ltd, then the appellant could put the case that JDI’s trading was on a genuine grey market and HMRC could put the case that it wasn’t, and moreover, that Mr Smith knew or should have known of this. That would lead to an inevitable adjournment as it would amount almost to a re-hearing of JDI, which was not desirable in any event, and there was certainly no time for the relevant evidence to be served, considered or tested in time for this hearing. If this hearing was not adjourned, and neither party applied for it to be adjourned, there was therefore no time for the appellant to satisfy me on the very contentious point of whether Mr Smith was a person who was expert in trading in the genuine mobile phone grey market. The evidence should be excluded.

105. In any event, I considered that the appellant had made its application too late. It was clear that the appellant had known by July 2018 that it might be calling Mr Smith: they had given HMRC no warning of this despite the necessary implication of agreeing a statement of expert evidence that they were not intending to call expert evidence. The first HMRC had known that the appellant was intending to call expert evidence was two months before the hearing commenced: that had been far too late for HMRC to deal with it as they would need to adduce evidence from the JDI appeal in order to seek to undermine Mr Smith’s evidence, and there was clearly not enough time for that. It was not even enough time for HMRC to identify a competing expert on the same subject. I considered that the appellant ought to have been aware that two months’ warning was insufficient and, as the appellant had known of its intentions since the previous summer, the prejudice to HMRC was unnecessary and avoidable and in the circumstances, it was correct to rectify the prejudice by refusing to admit the evidence.

106. As I refused the application to admit Mr Smith’s evidence, the parties were agreed, as I was, that there was no reason for Ms Hunter to be recused from the hearing of the appeal, and she re-joined the panel for rest of hearing and this decision.

Application to strike out fraud allegations

107. The Court of Appeal in the joined cases of E-buyer and Citibank [2017] EWCA Civ 1416 ruled at [97] that it was not necessary for HMRC to plead, particularise or prove dishonesty or fraud where the allegation was actual knowledge by the trader that its transactions were connected to fraud. The appellant’s position was that, subsequent to this, it had asked for CPR-style disclosure from HMRC and HMRC had refused on the grounds that they made no allegations of fraud or dishonesty against CCA or Mr Trees. The appellant had chosen not to pursue the disclosure any further on the basis of this concession.

108. However, shortly before the hearing, it took the position that HMRC was putting a case to the tribunal that was inconsistent with their assurance that they made no allegations of fraud or dishonesty against the appellant. The appellant’s position was that allegations that were only consistent with fraud or dishonesty should be removed from HMRC’s statement of case, could not be made in submissions at the hearing, should not be put to any of the appellant’s witnesses, and that the Tribunal could not reach findings of fact on them.

109. The appellant identified the particular allegations to which it referred: there were about 10 of them. We do not need to set them all out in full; they can be summarised as either allegations that the appellant, acting by its director, was a participant in an overall scheme of fraud or allegations that the appellant was not a free agent. We understood that the implication of not being a free agent was that the appellant knew it was being directed in its trading so that it had to know that it was a participant in overall scheme of fraud. We set out examples of the allegations complained about:

It is highly improbable that such a high proportion of the appellant’s deals would have led back to these defaulting entities if it were free to select its suppliers.

It is implausible that the appellant was duped by all 15 counterparties in respect of the 156 deals

We note in passing that one of the allegations complained about (concerning IP addresses) was later withdrawn during the hearing for a different reason.

110. The appellant’s point was that Kittel deemed a person who had knowledge or means of knowledge of connection to fraud to be a participant in the fraud, but HMRC were going further and alleging that CCA was actually a knowing participant in the fraud. That, said the appellant, amounted to an allegation of fraud or dishonesty on the part of CCA which was inconsistent with HMRC’s statement that they made no such allegation.

111. HMRC’s point was, they said, that they did not allege that CCA was conspiring with the fraudsters, but they did say that CCA (by its director) was being manipulated by his suppliers and customers and instructed on what to do in its dealing such that it was obvious to Mr Trees and CCA that its deals had nothing to do with commercial activity. HMRC relied on what the Court of Appeal said in Citibank and E-buyer that it was possible to allege knowledge of connection to fraud without making an allegation of dishonesty.

112. Our ruling was given orally and recorded in the transcript; we only summarise it here. Firstly, we agreed that the effect of the Court of Appeal’s ruling was that it was possible to enter into a transaction knowing that it was connected to fraud without being dishonest - [78], [86] [107] and [120]. It followed that a pleading of knowledge of connection to fraud was not a pleading of dishonesty.

113. However, the appellant was wrong to say that that meant HMRC could not plead factors which supported the allegation of knowledge if they also supported an allegation of dishonesty or indeed necessarily were pleadings of dishonest behaviour. For this, we relied on statements by the judges of the Court of Appeal as follows:

[85]…. It might be, of course, that if some or all of the allegations made in the Statement of Case were proved, that might (in theory, though not, of course, in practice) have allowed a tribunal to go on to make a finding that the taxpayer had been dishonest. But if HMRC does not seek such a finding, and if such a finding is not needed to support the conclusion that the taxpayer cannot recover its input tax, there is neither any need nor any utility in asking the FTT to undertake that exercise.

Sir Geoffrey Vos

[109] In summary, in my view, if HMRC do not wish and do not need to plead dishonesty, the concept of dishonesty should not be raising its head. As the Chancellor has observed, an analysis of whether the allegations amounted to dishonesty was unnecessary and inappropriate in litigation of this kind. Traders should not have to face a plea of dishonesty, HMRC should not be obliged to take on the burden of proving dishonesty, and judges should not have to address the added unnecessary complication of dishonesty simply on the basis HMRC seeks to prove actual knowledge of a fraud, in accordance with the Kittel test, and relies on facts and or inferences from facts that could support a finding of dishonesty. (our emphasis)

Lady Justice Hallett

[120] … Unless dishonesty is expressly alleged, the only question is whether the pleaded allegations are relevant to the issue of actual or constructive knowledge for the purposes of the Kittel test: …. For that reason, it is entirely irrelevant whether dishonesty is implicitly alleged in HMRC's statements of case. (our emphasis)

Sir Terence Etherton

114. In conclusion, the allegations complained about were allegations which supported HMRC’s case that the appellant had actual or constructive knowledge of the (alleged) connection to fraud; even if the allegations were consistent with a state of dishonesty on the part of the appellant and/or its director, the allegations were not allegations of dishonesty. The Tribunal would not strike out the allegations and would reach a conclusion on the allegations, but, as HMRC accepted, the Tribunal would reach no conclusion about whether the appellant’s and/or its director’s behaviour was dishonest or fraudulent: nor would The Tribunal undertake the exercise of deciding whether the allegations were only consistent with a dishonest state of mind; we would only make a finding on whether or not HMRC had proved actual or constructive knowledge of connection to fraud. The Tribunal noted that it would not be open to HMRC’s counsel to suggest to the appellant’s witnesses that they were dishonest.

Application to exclude evidence from previous hearing

115. The third interlocutory matter was the appellant’s application to exclude from the hearing the evidence from the previous tribunal hearing. HMRC had put the transcripts from the 2012 hearing into the hearing bundle and the appellant did not object to that; we were informed that the appellant did not object to that because it accepted that it would be legitimate to refer to them if a witness gave evidence in this hearing which was thought to be inconsistent with the evidence that witness had given in the previous hearing.

116. What the appellant objected to was HMRC’s stated intention to refer to passages from the transcript during the hearing; they objected to the fact that HMRC’s statement of case quoted evidence which Mr Trees had given in 2012. The appellant’s case was that this was an entirely new hearing: none of the findings from the previous hearing bound this tribunal in any way and indeed all the findings had been set aside. Moreover, it was unwise to rely on a transcript of oral evidence as the Tribunal did not have the benefit of seeing the witnesses’ demeanours.

117. HMRC’s position was that the transcript recorded the sworn testimony of persons who were witnesses in this tribunal and on the exact same subject as this tribunal: it was relevant to see what the witnesses had had previously to say about the facts in issue.

118. Again, our ruling was given orally at the start of the hearing and recorded in the transcript. This is only a summary of it.

119. Our ruling was that submissions from the previous hearing were only expressions of opinion by counsel and were not relevant in this hearing, although of course there was nothing to prevent counsel for either party renewing submissions before us that they had made to the tribunal in 2012. But the transcript of their 2012 submissions were not relevant and we did not think either party suggested that it was.

120. We also said that evidence given by a person in the 2012 hearing who was not a witness in this hearing could not to be relied on as to the truth of what was said in this hearing as that person was not giving evidence to us. Nevertheless, we considered that it was possible that such evidence could be relevant, in that a witness in this hearing could be asked about what had been said in 2012 by another person no longer a witness.

121. But we thought the greater part of the transcript of the oral evidence was plainly relevant because it was earlier statements made by witnesses in this hearing about the facts at issue in this hearing. As relevant evidence, it should be admitted unless there was a compelling reason not to do so.

122. We did not think that there was such a compelling reason. Firstly, we did not understand that the order of the Upper Tribunal (upheld by the Court of Appeal) that this case be remitted for a complete re-hearing was a bar on this Tribunal looking at the oral evidence given at the first hearing. We said this because the reason that the Upper Tribunal remitted the case, and the reason why the Court of Appeal upheld that decision, was because they considered that the 2012 FTT judge had reached his conclusion via a line of reasoning which the appeal courts considered to be unsafe. There was no suggestion that the hearing itself had been unfair and therefore no reason to suggest that evidence received in that hearing should be excluded as elicited under unfair circumstances.

123. Moreover, while we accepted that a written transcript of oral evidence may not be the best way of considering the evidence given as the judicial panel could not assess demeanour, that went to the weight put on the evidence rather than its admission.

124. We also did not think it unfair to admit the evidence as this hearing would be the opportunity for any witness confronted with what they had said in 2012 to explain what they had meant; no witness could be taken by surprise as they had attended the 2012 hearing and had given the evidence recorded in the transcript.

125. The transcript was admitted in evidence: that part of it which comprised the record of submissions was not admitted but we did not ask for it to be removed from the bundles. We would ignore it. We did point out that if either party considered that there was inconsistency between what a witness said in this hearing and what they had said in 2012, then the witness needed to be given the opportunity to explain it.

The facts

126. The factual evidence falls into two halves, which reflects the binary nature of the Kittel test.

(1) Alleged connection to fraud of the deals in respect of which input tax is claimed;

(2) Alleged knowledge or means of knowledge of any such proved connection.

The two halves are not entirely discrete as a part of HMRC’s case is that the nature of the connection to fraud was such that the appellant did or should have known of it, but we will so far as possible deal with the evidence in two parts. In any event, we will start the consideration of the facts with the appellant’s background and an outline of the disputed transactions before moving on to consider these two main areas of dispute.

127. An appendix to this decision records the full names of the companies which we below refer to only by a part of their name.

The appellant’s background

128. We start with an explanation of CCA and the background of its director, Mr Trees. While HMRC challenged Mr Trees’ evidence extensively, his evidence on his background was not in dispute.

129. Mr Ashley Trees had worked in the computer industry for some years; then in 1996 he set up a company with a Mr Carl Bennett: Appleco Ltd (t/a AC Computer Warehouse). Initially, Mr Trees and Mr Bennett continued with part-time employment, but as Appleco became successful they were able to work for it full time. By 1999, Appleco had expanded to employing about 30 staff. And having started in Mr Trees’ back bedroom, the business had gradually expanded into larger and larger business premises.

130. The initial impetus to form the company was an opportunity via Carl to remove old computer equipment from Prudential Insurance for a low price: they were able to do the equipment up and sell it and make a substantial profit. This gave the business a good start. Later they had a large contract with HSBC to remove old computers from offices and call centres: again Appleco was able to recondition and sell them for significant profit.

131. The company also manufactured computers from purchased components. When pentium processers reached the market, new computers were built with them, leaving Appleco with a profitable opportunity to build new computers using newly obsolete processers: this was profitable as such computers were in demand by businesses which had expensive software incompatible with pentium processors. For instance, ICL commissioned 1,000 new computers from Appleco with the old style processors for this reason.

132. By 2000, Appleco’s business model had become diverse. While its original business model was to buy used computers and then clean, repair and sometimes upgrade them and re-sell them (mostly retail) at a profit, by 2000 it also sold new computers, sold new computers which it had built from grey-market sourced components, carried out computer servicing, and rented out computers to businesses. It was effectively trading in the ‘grey’ market in that it was not an authorised distributor for any computer manufacturer.

133. In 1999, Appleco started to get offers to supply mobile phone handsets. Mr Trees suggested that company’s name became known to phone traders as it advertised in computer trade magazines which were also used by mobile phone companies. By 2002, Appleco had started to trade in mobile phones. In 2002, its turnover was £6.2 million and by the following year £22 million.

134. The appellant was incorporated in November 2001. It was incorporated to keep a joint venture between Mr Trees and someone else separate from Appleco. The joint venture was not successful and CCA became dormant shortly after its incorporation. However, from March 2003, Mr Trees chose to put the mobile phone trading previously carried out by Appleco through CCA, and CCA ceased to be dormant. The explanation for this decision was it would make checking the mobile phone transactions easier if they were split off from Appleco’s other business transactions. Mr Trees was, and remained, CCA’s sole director. All CCA’s trading at issue in this appeal was undertaken by Mr Trees.

The transactions at issue in this appeal

April 2006 broker deals

135. It was agreed that the appellant had undertaken 14 transactions in April 2006 in which it had purchased mobile phone handsets from a UK trader and sold them to an EU trader. The appellant’s immediate supplier was always Infinity, who acted as acquirer. HMRC referred to CCA’s deals as broker deals because they considered that the appellant was in the position of the broker in an MTIC chain of transactions in these 14 deals. In particular, they considered that CCA was a broker in a ‘clean’ MTIC chain and that Infinity was in the position of contra-trader.

136. In this period, the appellant also undertook 70 of what HMRC describe as ‘buffer’ deals. HMRC referred to them as buffer deals because they considered that the appellant was in the position of a buffer in an MTIC chain of transactions. Most traced back to acquisitions by C&B, Kep and AS Genstar. CCA’s customer was invariably Future. HMRC considered that these buffer transactions were all in MTIC ‘dirty’ chains: in other words, the supply chain commenced with a VAT default by the acquirer.

May 2006 deals

137. It was agreed that the appellant undertook 12 broker transactions in mobile phones in May 2006. The appellant’s immediate supplier was always Future. HMRC considered these to be clean MITC chains and that Future was a contra-trader.

138. CCA also undertook 6 buffer transactions. They traced back to acquisitions by AS Genstar, Okeda and Open Line. CCA’s customer in these transactions was Future or Infinity. HMRC considered these to be dirty MTIC chains and all to trace back to a VAT default by the acquirer.

June 2006 deals

139. It was agreed that the appellant had undertaken 13 transactions in June 2006 in which it had purchased mobile phone handsets/GPS/camcorders/laptops from a UK trader and sold them to an EU trader. The appellant’s immediate supplier was Future, Infinity or Soul. HMRC considered these to be clean MITC chains in which the appellant’s immediate supplier was the contra-trader.

140. CCA had also undertaken 41 buffer deals in this period; we find on the balance of probabilities that all of them traced back to Wade. In all the deals, CCA sold the goods to Future. HMRC considered these to be dirty MTIC chains in which Wade defaulted.

Samsung Serenes and P990s

141. One of the broker deals in each of the three periods in each involved the purchase and sale of Samsung Serene and P990s mobile phones. HMRC denied the input tax in relation to these three purchases by a separate decision and on a separate basis: their case was that the phones did not exist and so the transactions as described on the invoices had not taken place.

142. The appellant did not appeal this decision; and, therefore, those deals are not part of this appeal. HMRC’s position was that the appellant had accepted that those phones did not exist; Mr Pickup said he was neutral on the question of whether or not the phones existed; Mr Trees’ position was somewhat contradictory. He insisted that he believed the phones existed (although CCA never inspected stock and had therefore never seen them) and it had merely been the negligence of his solicitors which meant that the appellant had not made a timely appeal of the denial decision. At the same time, he said that he would offset the value of these phones, supplied by Future to CCA, against sums which CCA owed Future on other supplies (discussed below at §329) on the basis that the phones did not exist. (Although not relevant to the hearing, we noted CCA might not be able to do this because the facts showed that CCA had supplied more Samsung Serenes and P990s to Future in buffer deals than had been supplied by Future to CCA in broker deals).

143. The evidence on the existence of these phones before the tribunal was effectively hearsay although Mr Pickup did not ask for it to be excluded nor did he challenge it. As we have said, his position on behalf of CCA was that it was neutral as to their existence. The evidence was contained in the witness statement of Mr Bishop, an employee of Samsung. Mr Bishop was not a witness in this appeal, but his witness statements were exhibited to the evidence of Mr Allyn Cunningham; Mr Cunningham was not a witness either but his evidence was adopted by Mr D’Rozario, who was HMRC’s main witness.

144. Moreover, it seemed to us that it might be an abuse of legal process for the appellant to assert in proceedings against HMRC that the phones did exist in circumstances where it had not pursued its appeal against the denial made by HMRC on the basis that they did not exist. In other words, there was in effect already a legal finding between the appellant and HMRC that the phones did not exist. We recognise that Mr Trees explained the failure to appeal on the basis that, firstly, his then solicitors had let down the company by failing to make a timely appeal despite being instructed to do so, and that he had not make an out of time appeal as he (quite wrongly) believed that that was not possible and nor had CCA pursued an action against his solicitors on the basis they were insolvent. We were not convinced by this: for reasons given elsewhere we did not consider Mr Trees a reliable witness. And his explanation was not really credible: if he had wanted to take the matter further, he could have asked for advice on the prospects of lodging a late appeal. In any event, we did not really have to decide whether it would be abusive to assert the phones existed, because CCA (acting via Mr Pickup) did not do so.

145. In so far as Mr Trees gave evidence of his belief that the phones existed, we put no weight on this: Mr Trees was not a reliable witness, he accepted elsewhere that CCA did not inspect any of the phones and so had never seen them, and the basis of his belief of their existence seemed to be that its customers had not rejected the supply nor asked for their money back. However, as we set out below our findings that the transactions were entirely artificially orchestrated for the purpose of the fraud and that the appellant’s customers must have been knowingly involved in that fraud, it is of no surprise that the transactions were not rejected by the appellant’s customers who would not have been interested in the goods the subject of the transactions, as they would have know the transactions were orchestrated for fraud.

146. In conclusion, we have unchallenged hearsay evidence that the Samsung Serenes and P990 phones did not exist at the time they were purportedly traded by the appellant. We have no reliable evidence that they actually existed at that time and indeed the nature of the fraud strongly indicates that valuable phones would not have been used. We find that the Samsung Serenes and P990s in which CCA purportedly traded did not exist.

Connection to fraud

Fraudulent tax defaults

147. We move on to consider the evidence of connection of CCA’s 39 broker and 117 buffer transactions to fraud. HMRC served the evidence of a great many witnesses. In the event, the appellant did not require many of them to be called as it did not dispute all of the evidence. We will summarise their evidence where relevant without necessarily indicating which officer gave the evidence; where evidence was in dispute we will set out the dispute and our findings on it.

148. Unchallenged evidence was given by Mr Parker (Future), Mr Armstrong (Kep), Mr D’Rozario (unchallenged only in so far as his evidence on AS Genstar), Mr Sharrock (Wade, RS Sales and C&B), Ms Parsons (Wade and RS Sales), Mr Goulding (C&B), Mr Cameron-Watson (UK Communications), Ms Kandola (Okeda), Ms Coelho (ET Global), Mr Foy (PM Wholesale), Mr Siddle (Eutex), Ms Carroll (Booming) and Mr Mercer.

149. From the mass of evidence we had, in very brief summary, as it was not in dispute, in their VAT accounting periods which covered the period with which this appeal is concerned, the alleged contras, Future, Infinity and Soul undertook transactions as acquirers and as brokers. In all transactions in which they acted as brokers (in other words, the alleged dirty chains) we find that all the supply chains commenced with an acquisition from the EU by a company which defaulted on its VAT liability in respect of that acquisition. (A few chains could not be traced back to an acquisition but because of the similarity between those chains and the others we find it was more likely than not that they traced back to an acquisition by a company which defaulted on its liability). These dirty chains included CCA’s 117 buffer deals.

150. The defaulters were RS Sales, Wade, Okeda, AS Genstar, ET Global, Booming, Universal Trade and Sound & Secure, Eutex, C&B, and Kep. There was a great deal of overlap in the sense in that many of these defaulters were at the start of the supply chain for all three contra traders, particularly Wade which was (more likely than not) the defaulter in all Soul’s chains, in half of Infinity’s chains and some of Future’s chains.

151. The defaults were huge: for instance, AS Genstar was assessed for £46 million. We are satisfied all the defaults were fraudulent; while the evidence was not identical for each company, we find the companies issued VAT invoices showing a VAT liability, and did not declare it to HMRC, and certainly did not pay it. While a failure to pay may be the result of insolvency, a failure to declare a liability, particularly a very large liability of which trader was clearly aware as it issued invoices, is most likely to be due to dishonesty. When combined with other factors mentioned by HMRC’s witnesses, (such as some VAT numbers being hijacked, some traders issuing invoices without being VAT registered, and some trading in phones that did not exist etc) it is beyond doubt that all these defaults were fraudulent.

152. Apart from anything else, the coincidence of all alleged contra-traders’ many broker chains involving millions of pounds of trade, plus all of CCA’s buffer chains, tracing back to a default is too great to be ignored: the defaults were deliberate.

CCA’s transactions’ connection to fraud

153. As we understood it, the appellant’s original position was that it denied that its transactions were part of an overall scheme to defraud HMRC although it did not challenge the evidence of fraudulent default. In the hearing, Mr Pickup conceded that the banking evidence might be thought to show a contrived overall scheme of fraud but said this was not conceded. However, later in closing he said that it was accepted there was an overall scheme to defraud HMRC.

154. Nevertheless, we will, briefly in comparison to the mountain of evidence, consider the fraud because, to some extent at least, the appellant put HMRC to proof on the matter. In any event, the nature of the fraud (if proved) has some relevance to the question of the appellant’s alleged knowledge and means of knowledge.

155. We did not understand the appellant to question CCA’s buffer transactions’ connection to fraud. There was a clear connection as the supply chain commenced with an acquisition on which there was a fraudulent VAT default. CCA’s VAT reclaim on these deals has not been denied and this connection is not directly relevant to the appeal, apart from the light it may cast on CCA’s broker deals.

156. CCA did question, at least to an extent, HMRC’s case that its broker deals were connected to fraud. As we have explained, all the broker deals were alleged to be in clean MTIC chains; the contras through which the VAT in the clean chains was allegedly offset against the VAT in the dirty chains were Future, Infinity and Soul. The appellant came to accept that in so far as the contra was Future or Soul, they were proved to have acted fraudulently and so it seems that the appellant came to accept the connection to fraud of its broker transactions where Future or Soul was its supplier. As we have already discussed (see §§80, 82-90), the appellant did not accept that Infinity could be shown to have acted fraudulently and therefore did not accept that its broker deals with Infinity (ie all those in April 2006 and some of those in June 2006) were shown to be connected to fraud.

157. We have in any event reached our own conclusion on HMRC’s evidence of connection to fraud. Nevertheless, as it was not really in dispute save as to Infinity, we do not set the evidence out in anything like the detail contained in the (largely unchallenged) witness statements. We set it out in some detail because to some extent it informs the question of knowledge and means of knowledge which we then go on to address.

The challenged witnesses on connection to fraud

Kenneth George Rhodes

158. Mr Rhodes was the compliance officer for Soul, allocated to the trader in July 2006. He looked at Soul’s trading from February 2006. He was cross examined but his evidence was not significantly challenged; indeed later Mr Pickup said the evidence of Soul’s involvement in the fraud was overwhelming. We accept Mr Rhodes’ evidence.

Candida Styles-Coles

159. Ms Judith Clifford had given the evidence with respect to Future in the 2012 hearing; by 2019 she was no longer an HMRC officer and Ms Styles-Coles adopted her witness statements. Ms Styles-Coles was cross examined but her evidence was not seriously challenged and, as we have said, Mr Pickup later said that the evidence of Future’s involvement in the fraud was overwhelming. We accepted Ms Styles-Coles evidence.

Jayne Louise Holden

160. Mr Simon Devine had been the compliance officer for Infinity and gave a witness statement in this appeal; he had left HMRC before the last hearing. His evidence was adopted and expanded upon by Ms Holden, an HMRC officer, who was cross examined in this hearing. Again we did not find her evidence to be seriously challenged and we accept it. We deal with the question of Infinity and whether HMRC can prove that it was fraudulent below at §§224-233.

Peter Birchfield

161. Mr Birchfield was an HMRC officer and he gave evidence about what he had found on the servers of the FCIB. He made two new statements for this hearing which largely replaced those he had given in the first hearing; the distinction was that since the first hearing, HMRC had obtained access to FCIB’s Paris server whereas for the first hearing they had only had access to the Dutch server. The Paris server contained additional information, in particular the timings of transactions and the narrative put with a transaction by the transferor.

162. Mr Birchfield was a witness of fact and not an expert witness; we accept his factual evidence as he appeared to be a reliable witness. We consider that he plainly strove to be fair to the appellant and not to make inferences of fact that the evidence did not justify. He accepted, for instance, that the money flows by themselves did not indicate that CCA knew of the connection to fraud. Moreover, the factual evidence about which he made his statement were the FCIB statements which were available for both parties to analyse and challenge if they saw fit. And the appellant did not really challenge his factual evidence; what was challenged was the inferences properly to be drawn from it, which we deal with below.

163. The appellant also asked Mr Birchfield’s opinion on certain matters, and in particular relied on his opinion that, if there was to be an innocent dupe in an MTIC chain, it would be in the position of the broker. We do not accept Mr Birchfield’s opinions as evidence because he was not an expert witness, nevertheless we have treated them as submissions to the extent that they were relied on by the appellant. We have already dealt with his opinion (treating it as a submission) that if there was to be an innocent dupe in the chain, it would be the broker: see §§55-57.

The expert witnesses

164. The Tribunal had in evidence a number of expert reports from two expert witnesses, Mr Letherby for HMRC and Mr Moore for the appellant. Their evidence concerned the IP addresses from which various entities in the money chains had given their instructions to FCIB, information available to this Tribunal from the Paris FCIB server and therefore not available to the tribunal in 2012. Having heard Mr Letherby’s evidence under cross examination by Mr Gurney, HMRC withdrew the allegations they made based on the IP addresses and accepted Mr Moore’s evidence. Mr Moore was therefore not called.

165. As the allegations were withdrawn, we have not considered them. The expert evidence does not really need recording. Suffice it to say that it appeared to us that HMRC came to accept, as Mr Letherby did, and as we do, Mr Moore’s evidence that the particular IP addresses concerned belonged to O2 and that where those IP addresses had been used it appeared therefore that the instruction had been given to FCIB via a mobile phone; and moreover the most likely explanation for the same or similar IP addresses being used was that O2 had routed all calls to FCIB via the same IP address (‘gateway’), thus making it appear, incorrectly, as if the instructions had all emanated from the same modem. In conclusion, very briefly, there was nothing to be read into the use of the same IP addresses by more than one of the entities. We do not refer to this expert evidence again.

Were the transactions contrived?

166. For the following reasons we find that all of CCA’s trades in the period April - June 2006, including all the deals the subject of this appeal and CCA’s buffer deals, were contrived and did not take place on the open market.

Coincidence

167. We have briefly set out above the details of the defaulters and the evidence that their massive defaults were fraudulent. It was not challenged. The evidence shows, and we find, that every deal by CCA in that 3-month period connected to a fraudulent VAT default. We agree with HMRC that this was no coincidence. We find it indicates that all CCA’s deals, broker and buffer, were contrived as part of an overall scheme to defraud HMRC and the tax-paying public.

Repeated patterns in the supply chains

168. We find other evidence which shows that the deals were contrived and did not take place on the open market. In particular, the supply chains show a great deal of repetition and patterns inconsistent with a genuine market:

(1) The chains largely had the same entities in the same order buying and selling to each other;

(2) The contra-traders were involved in a great many chains, involving brokers other than just CCA yet largely those chains involved the same entities as the chains with CCA;

(3) The many transaction chains, involving a great many companies, only involved a few defaulting acquirers; why did these companies all just happen to buy from the same defaulters if it was not an organised fraud?

(4) The goods in all the chains and not just those involving CCA were all sourced from the same 4 EU suppliers (Alpha, Esat, Vertex and Rachel Tel) and the goods ended up sold back to the same EU entities to which CCA sold the goods;

(5) CCA’s customers often sold the goods to the same entity. (For instance, CCA sold to Mohomedbhay, Universal, and Pielkenrood in April but all three of those entities sold the goods to Wizard; in other deals CCA sold to these 3 persons, and they on-supplied all the goods to Tolus; moreover, Wizard and Tolus both sold the goods to Bartonole.

Transactions inconsistent with commerce?

Back to back transactions

169. All the transactions were back to back with each stage making an incremental profit; in other words, each trader was able to sell the goods the moment they were purchased. By itself this does not indicate fraud, but it is a factor which is consistent with fraud as it is unlikely that in the real world all buyers and sellers would all be able to identify simultaneous purchases and sale and still make a profit.

Failure by traders to identify most profitable deals

170. The identified supply chains normally involved 5 to 6 entities: (1) EU supplier to (2) contra-trader to (3) CCA to (4) CCA’s EU customer, to (5) their customer, and in some cases to CCA’s customer’s customer’s customer. Each entity added on a mark-up. No one appeared to be adding in any value to these back to back wholesale transactions. Yet there were internet based trading platforms such as International Phone Traders website that would have enabled traders to identify the cheapest price.

171. Therefore , these long chains appeared inconsistent with commerce; it meant that, despite the product being mass-produced and uniform, the traders were consistently failing to identify that they could buy it cheaper elsewhere. Why was CCA’s customer’s customer not buying the phones direct from CCA’s supplier? Market forces would drive traders to identify cheaper deals and cut out middlemen. The fact this did not happen indicates contrived deals.

172. Not only were the chains long, but they involved the product entering and leaving the UK on a very short turnaround. Again, the point is much the same as in the previous paragraph: in a genuine market, the last identified customer ought to have been looking to purchase the goods from someone such as the contra’s EU supplier. That would have saved the costs of transportation to the UK and back again as well as allowing the purchase at a much lower price. But, instead, the same EU customers were content repeatedly to purchase from CCA in the UK despite an EU supplier being able to sell them cheaper to CCA’s supplier.

173. The uncommerciality of the behaviour goes further than this. In CCA’s broker deals, the three contras sold to CCA at a relatively small mark-up while CCA sold to its EU customers at a much larger mark up. It might appear, therefore, that the contras were commercially naïve in repeatedly failing to identify that they could have sold to EU entities at a much higher price than CCA was apparently prepared to pay. But naivety is not the explanation for their behaviour because the contras did sell to CCA’s EU customers in other deals and were clearly aware of their existence and their willingness to pay a much higher price. As an example, Future sold goods in May and June 2006 to Allimpex, Pielkenrood, Mohomedbhay and Universal Handels; it also sold goods to CCA in the same period which CCA sold to the Allimpex, Pielkenrood, Mohomedbhay and Handels.

174. This persistent uncommerciality in many deal chains is a strong indicator of their artificiality and is with all the other evidence clearly indicative that all the deal chains were orchestrated for the purpose of MTIC fraud.

Phones sold without retail demand

175. The quantity of particular phones in CCA’s transactions was in some instances so great it was in excess of the total demand for phones in Europe and the Middle East. For example, the appellant sold 14,000 Samsung SGHi 310 phones in May and June 2006: but only 10 were sold retail in the entire year and none were sold retail before September 2006.

176. Moreover, all of the phones were of European specification and had 2 pin chargers and could not have been used in the UK without an adaptor. It makes no sense to continually import such phones into the UK., particularly when it appears there was no retail demand for them as they were then, as pointed out above, almost immediately re-exported back onto mainland Europe.

Limited number of freight forwarders

177. In all of CCA’s broker deals in issue, save for 3 where the evidence is not available, we find that the goods were stored at Freightline Logistics in Luxembourg before being shipped to the UK. This would be something of a coincidence if the deals were commercial because CCA bought from three different suppliers (Infinity, Future and Soul).

178. Again, CCA always despatched the goods to the EU to the location requested by their customer; whichever entity was the customer, they always requested the goods to be despatched to Boston Freight in Belgium. This was the case wherever the customer was located and none of them was located in Belgium.

179. The same happened in other chains in which CCA did not appear; for instance, all of Soul’s despatches for several months in 2006 were to Boston Freight.

180. This would make no sense in a commercial situation, but if it was a single orchestrated MTIC fraud, it makes perfect sense: the goods were being circulated and so the fraudster would have wanted them delivered to the same complicit freight forwarder so that they could be sent back to the UK, so they could be continually carouselled.

Uniform profit margin

181. In many of the chains, the other entities all made the same profit. For instance, in all but a few of CCA’s broker deals, the supplier charged CCA a price that enabled the supplier to make a profit of precisely £2. In those few exceptions, which only concerned Future, Future made a profit of precisely £3. There were consistent profit margins in all of CCA’s buffer chains (in most cases, £1).

182. It is extremely unlikely that there would be such uniformity of profit margin, and round sum profits, in independently negotiated deals on an open market. It makes sense in the MTIC world where the deals were orchestrated as it can be supposed the fraudsters dictated set profits for the participants.

Deal documentation uncommercial?

183. The contractual terms as reflected in the deal documentation was very sparse and we find generally failed to deal with matters such as the passing of title, risk and terms of payment. Even where terms were specified, they appeared not to reflect what the parties had actually agreed. An example of this was Future’s contracts which stated that title was reserved until payment was made; yet CCA sold and shipped the goods before it paid Future for them.

184. The contractual terms also failed to specify the subject of the deals in the sort of detail that would be expected. They generally failed to specify whether the phones were new or used, whether they were locked or unlocked, whether a sim was included or not, whether they were of UK or European specification, whether they had a 2 or 3 pin charger, whether they had a warranty, whether they came with accessories, their colour, the language of the manuals and handset.

185. We had no expert evidence on this but the general knowledge of the tribunal about mobile phones suggests this is to some extent very odd: while the appellant said the phones could be configured for a number of languages and the manual would be printed in a number of languages, commons sense says that the value of the phones must have been affected by whether they were new or used and whether they were locked or unlocked, and that an adaptor would be an additional expense if the phone had the wrong charger. The failure to specify such basic information in the contractual documentation does suggest the traders had no real interest in what they were buying.

Evidence of male fides of various entities

186. There was other evidence that some entities in the chain were not commercial entities or indeed were knowingly involved in fraud. The evidence is not of the best: it is mostly hearsay. For instance, the Austrian Tax authority said that Handels did not actually send on the goods it purchased to its customer but simply returned the goods to the UK. More hearsay evidence from a tax authority was in respect of Wizard. Its declared business was in real estate; it never submitted VAT returns and was deregistered in mid 2007 for failing to abide by its VAT obligations. Its director and shareholder was a UK national. Bartonole was a short-lived entity which was incorporated in January 2005 and de-registered for VAT in August 2006 because the director told the national VAT authorities that someone was using the VAT number for fraudulent purposes. Again, the director was British.

187. The most frequently used freight forwarder in the UK was A1. Its director was a Mr Lee Sellers. Mr Sellers was convicted of VAT fraud offences in Britain. An employee of Boston Freight, Mr Paul Smith, was convicted in Britain of VAT fraud offences.

188. This evidence of male fides of various entities in the chains supports what is overwhelmingly obvious from the banking (discussed below) and transaction chain evidence: all the transactions were part of an MTIC fraud; they were all engineered and did not take place on the open market. Companies were VAT registered to undertake the orchestrated transactions but were not genuinely trading. While the above evidence of male fides in so far as it is hearsay is weak, it is accepted as it supports what the other evidence makes clear.

189. We were, however, very wary of accepting the hearsay evidence in respect of Pielkenrood. As we have said, CCA sold only to 4 EU entities in its 39 broker deals. One of those four entities was Pielkenrood. The director of that company was Mr Simon Pielkenrood who lived in Taiwan. He had a son, Mr Jacob Alexander Pielkenrood, who appeared to go by the name Sander, and to whom we shall refer by that name. Sander lived in Russia and had some involvement in the sale of mobile phones by Pielkenrood, although the extent of the involvement was very much in dispute.

190. Pielkenrood failed to produce any CMRs and so the Dutch authorities investigated the company. In April 2009, they interviewed Sander who informed them that individuals based in the UK and Dubai had told him from whom to purchase and to whom to sell. He said there were no negotiations as the deals were pre-arranged. He also told them he had had to open an FCIB account and he would be informed that money was coming in, and he had to be at his computer and ensure he immediately transferred the money onwards.

191. Mr Pickup told us that this evidence was excluded from the 2012 hearing; he did not ask for it to be excluded from this hearing but did submit that no weight should be put on it. It was hearsay of the weakest kind. HMRC were repeating information third hand; ‘Sander’ appeared to be admitting to fraud yet it was not clear whether he had legal advice nor whether he was making admissions in return for non-prosecution. He had not signed the statement, his evidence was not tested in Dutch courts and he was certainly not a witness in this hearing.

192. We agree with Mr Pickup about the inherent weakness in the evidence and would be inclined to place no weight on it for the reasons he identified; we pause only because of the very plausibility of Sander’s evidence. The documentary evidence we have seen shows that this was a highly sophisticated fraud involving orchestrated chains of transactions; even the appellant appears to accept that other participants in the chain (apart from the broker) must have been told from whom to buy, to whom to sell, and at what price. The Paris server evidence (which we discuss in detail below and which was not available when this evidence was first produced) shows that the other traders must have been instructed to be ready to transfer on the sum of money received as soon as it came into their account. This hearsay evidence is entirely consistent with all the other evidence in the Tribunal and for that reason appears reliable.

193. However, the most crucial part of the evidence is Pielkenrood’s position as CCA’s customer: if this hearsay evidence is accepted as reliable, it directly contradicts CCA’s case that it believed it was negotiating its deals on the open market and that the fraudsters took pains to keep CCA in ignorance of the fraud by keeping up a pretence of genuine negotiations: Sander said there were no negotiations.

194. This is a critical question in the appeal. Our decision on the reliability of Sander’s evidence, is that although it appears highly plausible, it is very weak hearsay evidence and we would be very reluctant to base our determination on it. We put it out of our minds when considering all the other evidence.

The banking evidence

195. As we have said, the appellant made no fundamental challenge to the reliability of the overall banking evidence given by Mr Birchfield. It did dispute some of the inferences to be properly drawn from it.

196. The banking evidence came from the servers of the same bank, FCIB. As we have already said, for the 2012 hearing, Mr Birchfield had only seen the account information from the Dutch server of the FCIB; for this hearing he had seen the information contained on the Paris server. The distinction was that the Paris server information was more detailed: it contained the times as well as the dates of transactions; it also recorded any narrative that the transferor had attached to a payment.

197. Mr Birchfield created charts showing the movement of money which came into and left CCA’s account in the period covered by this appeal. He found that the money moved in loops, from one trader to another, some 61 in all, and all of which had FCIB accounts.

198. There were 45 payments into/out of CCA’s account in that period: in 30 out of 45 chains, the money started and ended with the same entity (Bartonole or Peoria). In the remaining 15 chains, either the funds went to other companies which appeared to be similarly owned or controlled (eg started with Bartonole but ended with Peoria, Praxi and Lassi) or the funds went into Future’s Barclays account and could not be traced further, or to a company (Destonia) which featured in other MTIC fraud chains.

199. We find that the banking evidence gave a very clear indication that a sophisticated, complicated and very lucrative fraud on the British public was at the root of all the transactions at issue in this appeal. That is the only logical explanation for money circulating in loops. The loops are only consistent with a fraud which involved money being moved between entities to give the appearance of commercial transactions; it was entirely inconsistent with genuine commercial transactions where the ultimate seller would not be the ultimate purchaser.

200. Moreover, Mr Birchfield was only able to identify these loops because all the entities in the chain banked with FCIB. It would be extremely unlikely that a chain of commercial, independent traders would all choose to bank with the same bank.

201. The appellant suggested this was not unlikely because UK banks (possibly at the instigation of HMRC) closed accounts for mobile phone traders while FCIB, an off-shore bank, offered modern banking facilities. We do not accept that explanation: it was clear that both CCA and Future were able to hold and use other bank accounts as they did so: Future had a Barclays account and CCA had a Bank of Ireland account. Moreover, the chains involved non-UK companies who would not be affected by the closure of UK bank accounts and so that could not be the explanation as to why they held FCIB accounts. We consider it more likely than not that the reason all the entities banked with FCIB is that it suited the fraudsters that they did so as it enabled the money to move swiftly and that is what they wanted.

202. We say that it was clear that the fraudsters wanted the money to move swiftly because, firstly, that is what it did, and secondly, it is a logical inference. Some of the loops, involving many entities, took less than 2 hours. It is also clear the same funds were used to circulate many times within the same period of 24 hours.

203. Mr Birchfield’s evidence which we accept and which was not in any event challenged, was that FCIB only executed transfer instructions in 3 minute batches; it was therefore impossible for any person to transfer money they received in less than 3 minutes. We find that in many instances the recipient of funds must have given instructions to the bank to transfer the funds onwards in three minutes or less of its receipt because there was only 3 minutes between the receipt and the transfer. Such swift receipt and transfer of funds enabled large loops to be completed very quickly, as we have said.

204. Moreover, when Mr Birchfield looked beyond the loops involving CCA and just followed the money, the evidence showed that the same sum of money might circulate in a loop 10 times within 24 hours: it looked like the money was circulating in loops as many times as possible in as short a time as possible.

205. The money movements on their face had little to do with commercial transactions as the entities were in most cases passing on very quickly exactly the amount they received. It was like a game of pass the parcel. It did not look commercial: an entity being paid for a commercial transaction might choose to keep the money: it would expect to retain its profit margin in any event and wouldn’t expect its supplier to know that it had been paid so its supplier wouldn’t be looking for payment the moment its customer paid.

206. The banking evidence did not look consistent with commerce. It was consistent with fraudsters trying to give credence to the orchestrated chain of transactions by setting up chains of payments to show HMRC that money had actually changed hands, while, at the same time being careful not to have their money out of their control for any length of time. It looked like the fraudsters were maximising the use of the money by ensuring as many loops took place within 24 hours as possible.

207. Only a highly sophisticated and organised MTIC fraud makes sense of the banking evidence and the appellant did not suggest a realistic alternative explanation. We agree with HMRC that this demonstrated very clearly that at the root of the circulation of funds was a fraud as it was entirely inconsistent with a commercial transaction where individuals were free to pay their suppliers at a time of their choosing.

Conclusion on connection to fraud

208. In conclusion there is overwhelming evidence that CCA’s supply chains, both buffer and broker, were contrived. There is no reason to contrive transactions other than fraud and in any event as we have already found all of them did trace back to a fraudulent default, albeit in the case of CCA’s broker transactions, the connection to the fraudulent default was via a contra-trade. The overwhelming likelihood is that all of these transactions were contrived for the purpose of MTIC fraud and so we find.

209. We find for the reasons given above that all of CCA’s transactions in the period in question, broker and buffer, were contrived by fraudsters; that its broker transactions were with alleged contra-traders, Future, Soul and Infinity, all of whose broker transactions could be traced back to fraudulent default; we also find CCA’s buffer transactions all traced back directly to a fraudulent default.

The connection via contra-trading

210. We discussed above at §§73-80 that HMRC’s view was that that was enough to prove connection to fraud; the appellant did not agree; it considered that HMRC had to prove that the contra-traders not only actually set off the VAT in CCA’s broker chains against VAT in a dirty chain, but that they did so as a knowing party to the fraud.

211. We have already expressed our view (see §80) that HMRC is right that the law does not require this, but, in case this appeal goes further, and because we have the facts before us, it seems sensible to determine whether or not the alleged contra -traders were knowing parties to the fraud. And, as we have already said, knowing participation in the fraud was in the end conceded by CCA with respect to Future and Soul. In so far as Infinity was concerned, the appellant’s challenge, rather than being to the evidence produced by HMRC, appeared to be one of law. That challenge was that HMRC could not allege fraud as to do so would be abusive, but we dismissed that at §§82-89. Nevertheless, as we had the evidence to consider, we set out briefly why we think it has been proved as a matter of fact that the three alleged contras were knowing parties to the fraud.

Future

212. Future’s roots were in a company called Ravjani Corporation which came into existence in 2001. It was run by Dilawar Ravjani although the company was owned, at least legally if not beneficially, by his father, Haider Ravjani. By the end in 2006, its turnover was £2.8 billion.

213. Future was on quarterly returns but its quarters did not align with the three VAT periods at issue in this appeal: one VAT quarter ended at the end of April, and its next at the end of July 2006. In those two quarters, Future undertook 6,791 sales; 560 of them were buffer deals, 3,399 were broker deals and the remaining 2,832 were acquisition deals. Most but not all the broker deal chains have been traced back; all of the deals which have been traced back, trace back to one of eight defaulters with UK tax losses of £125 million. As we have said, we find on the balance of probabilities all its broker deals traced back to defaults.

214. In approximately the two years to July 2006, it undertook deals worth £3.2 billion but its net VAT liability was £133.89; in the quarterly period to July 2006 it deals were worth £46 million but its net VAT liability was £12.36.

215. It could not be chance that all its deals appear to have been orchestrated; all its broker deals traced back to fraud; and that its output tax netted so exactly against its input tax. It was not chance: its deals were clearly orchestrated for the purposes of fraud, as set out above.

216. The appellant suggested that a contra-trader would not necessarily realise that it was being manipulated to enter into transactions such that its input tax netted so exactly against its output tax. The appellant suggested it would make good business sense to avoid incurring a substantial VAT debt from HMRC as there was always a delay in repayment. We find the proposition that a trader could be manipulated unknowingly into transactions that so neatly netted off difficult to believe: if there were no other indications of orchestration, the mere fact that it was offered such transactions would be something of an indication. This is particularly the case when the evidence shows that Future undertook almost exclusively acquisitions in the first month of each quarter, a mixture in the second month, and almost exclusively broker deals in the last month of each quarter.

217. But in reality the deals were orchestrated which means that they did not occur on the open market and could not have been negotiated: while the appellant put forward a case that it was in the interests of the fraudster to keep the broker innocent such that the fraudster would set up pretend negotiations to keep the broker innocent, obviously if the fraudster did that, it could not keep the parties on either side of the broker ignorant. So, if CCA was ignorant, that could only be because Future, its trading partner in so many of its broker and buffer deals, was not. And if the fraudster did not attempt to keep CCA ignorant, it would have no motive to keep the contra-trader ignorant.

218. Putting all that aside, Future dealt in phones that could not have existed in the quantities in which it traded them. Moreover, Dilawar Ravjani and its manager, Rajesh Gathani, were convicted of fraud in connection with Future’s trading. The appellant did not even suggest that these convictions were wrong: indeed, Mr Gathani had pleaded guilty.

219. Further, Future acted as a company which was aware that its deals were contrived and did not take place on the open market would act: it did not undertake due diligence and despite the vast quantity of phones in which it traded, it did not carry insurance. The company did not generally cooperate with HMRC.

220. The appellant conceded Future was a fraudulent entity and we find it was. It was clearly a knowing participant in the fraud; its role was that of contra-trader.

Soul

221. Soul was the appellant’s supplier in all its June deals apart from those mentioned above where the supplier was Infinity or Future. Soul was also on quarterly returns and in its 3-month period ending 31 July 2006 it conducted 100 broker deals worth £68.3million and acquisition deals worth £67 million. As we have said, we find all of its broker deals traced back to a default.

222. It traded in phones that could not have existed in the quantities in which they existed; it had close connections to Future (Soul was Future’s tenant) whose director and manager were convicted of fraud in respect of the part they played; Soul’s director went so far as to admit in his director’s disqualification order to recklessness or gross negligence to fraud in its trading although he did not admit actual fraud. Soul was incorporated in only 2001 but by 2006 had a massive turnover; at the same time there was no evidence that it ever carried out due diligence. Despite the massive value of the goods in which it deals, it held no insurance. Soul did not cooperate in HMRC’s investigation.

223. We find, as the appellant accepted, that Soul was a party to the fraud. We find that all its transactions, both in clean and dirty chains, were orchestrated; it managed its VAT returns such that its output tax was nearly perfectly offset by its input tax. Doing so was vital to the fraud in order to provide distance from the default to the broker. It cannot have been accidental. Despite the vast sums in which it was dealing, it did not appear to conduct any due diligence or hold any insurance so that it appears it was not interested in protecting its position as if trading on a genuine market. We find it must have known that it was not.

The Infinity companies

224. As we have said, there were two companies owned and controlled by a Mr Simon Thakor. Infinity and Distribution. We have already reached the conclusion that it is not abusive for HMRC to maintain in these proceedings that Infinity was a knowing party to the fraud. Reliance on the concept of abuse of process seemed to be the extent of the appellant’s case that Infinity was not a party to the fraud. Mr Pickup did not really challenge the mass of evidence put forward by HMRC which they relied upon as indicating Infinity was a party to the fraud, other than to suggest that in theory a contra-trader could be manipulated into the netting off. We rejected that theory at §§216-217 above.

225. It seemed to us that the evidence with respect to Infinity was very similar to the evidence with respect to Future and Soul, which the appellant had conceded was overwhelmingly evidence of fraud. But as the point was not conceded with respect to Infinity, we make up our own mind.

226. Infinity accounted for VAT on a quarterly basis; one of its quarters covered the period 03.06 to 06.06. In that quarter, it undertook 304 broker deals all of which we find, as set out above, traced back to defaulting traders. We find this was true for its previous quarter as well.

227. We find its deals in the period in question in this appeal were clearly orchestrated for the purposes of fraud, for the reasons given above. As we have said, it must have known this because the deals clearly were not negotiated but dictated to the participants for the reasons given above at §§167-209.

228. And while Infinity’s entire turnover in period 06/06 was £343million, it had net output tax liability of only £471. This pattern occurred in its previous quarters: a massive turnover and almost negligible output tax liability. This negligible VAT liability was critical to the fraud as its purpose was to distance the broker from the fraud by having an ostensibly VAT compliant acquirer in the broker’s chain. It was critical to the fraudster that the contra-trader carried out the transactions to permit this netting off so it logically follows that the contra-trader must have been a party to the fraud. For reasons already given, we reject the appellant’s case that a contra-trader could have been manipulated into this position without it knowing it was a party to the fraud.

229. Another pattern to its trading was that, like Future, the vast majority of its acquisition deals were completed in the first month of the quarter; it then undertook sufficient broker deals in the next two months of the quarter to almost exactly set off its VAT liability to HMRC. This cannot have been accidental; for the reasons given above we do not consider it could or would have been manipulated into doing this.

230. There are other indicia that Infinity knew that it was not trading on a genuine market some of which we summarise here: it always made the same mark-up, some of the goods were sold at uncommercial prices, it sold goods to companies, which then sold them at profit to companies with which Distribution was in contact (in other words, it failed to maximise its profits), it undertook no due diligence other than to check the VAT registration of its counterparties, it sold goods which could not have existed at that time in the quantities traded.

231. We also find that Mr Thakor had substantial funding the source of which he was reluctant to identify to HMRC; we find Infinity did not produce documents to support its VAT claims made to HMRC and did not appeal the refusal by HMRC to repay tax it claimed.

232. We find the evidence that Infinity was a knowing party to the fraud as clear as it was in the cases of Future and Soul.

Connection to fraud

233. That concludes our fact finding on the first part of the Kittel test: for the reasons given above we find that all of the transactions at issue in this appeal were connected to fraud. Similarly, we find that all of CCA’s transactions in the period in question, both broker and buffer, were connected to fraud. We find all three contra-traders (Future, Soul and Infinity) were knowing parties to that fraud.

Factors which indicate knowledge and/or MOK

234. The next question is whether the appellant knew or should have known of this connection to fraud at the time the transactions were entered into. We will deal with knowledge first, and then with means of knowledge.

235. The allegation of knowledge and means of knowledge was the main dispute between the parties and the evidence hotly disputed. We will start our summary of our findings of fact by considering the witnesses.

The witnesses on knowledge and means of knowledge

Vincent D’Rozario

236. Officer D’Rozario was the compliance officer for Appleco and then for CCA from its inception. However, due to pressure of work, a different officer, Mr Cunnningham undertook the extended verification on the deals the subject of this appeal and issued the denial letters.

237. The appellant criticised Mr Cunningham’s decision letters: they said he applied the wrong test and made factual errors. All of this might be true, but it is quite irrelevant. The question is whether CCA was entitled to its input tax and not whether Mr Cunningham put the correct reasons for denying the input tax in the decision letter. Mr Pickup also suggested that neither Mr Cunningham nor Mr D’Rozario had seen the appellant’s full due diligence files when its input tax was denied and had they done so, it would have changed their decision. Even if that was true, which Mr D’Rozario did not accept, it was not relevant. The tribunal was not here to review the adequacy of the reasons for which HMRC denied CCA’s input tax but to decide whether HMRC were legally entitled to deny repayment on the facts as now before this Tribunal.

238. Mr Cunningham was not a witness in this hearing; he had been a witness, we were told, in the previous hearing but had since retired. Mr D’Rozario had adopted Mr Cunningham’s witness statements as his own, as well as providing new witness statements.

239. Overall, we found Mr D’Rozario to be careful and precise in his answers to the Tribunal; he was willing to accept that some points put to him in cross examination were valid. We considered him a good and reliable witness.

240. The appellant placed significant emphasis on the relationship between Mr D’Rozario, Mr Trees, and CCA’s bookkeeper, Mrs Ryan. We will deal with that below. Suffice it to say here that there was some differences between them as to the nature of the relationship and we prefer Mr D’Rozario’s evidence on this as we found him to be a good witness, and the same was not true of Mr Trees, as explained below.

Wesley Gordon

241. Mr Gordon gave evidence on behalf of CCA in the form of a witness statement. HMRC did not challenge his evidence, so he was not called to give oral evidence. We accept his evidence as it was unchallenged. We summarise his evidence as follows.

242. Mr Wesley was taken on as an employee by Appleco in 2004. He also became company secretary to CCA. His role in Appleco was sales and marketing. He had no involvement in the day to day business of CCA. From early 2006 he was involved in updating and reviewing CCA’s due diligence files on its trading partners. He was instructed by Mr Trees on how to do this, Mr Trees having previously done all the due diligence himself.

243. Mr Wesley would tell Mr Trees if he was unhappy with a company he was verifying, for instance if they provided insufficient material. The files he worked on were removed by the officers executing a search warrant on 1 June 2006.

244. He worked in Appleco’s open plan office, as did Mr Trees, and heard Mr Trees speaking to staff throughout the day and negotiating with people on the phone.

245. At no point did he think Mr Trees was involved in anything dubious

Pat Ryan

246. Mrs Ryan gave evidence on behalf of CCA in the form of a witness statement. HMRC did not challenge her evidence, so she was not called to give oral evidence. We accept her evidence as it was unchallenged. We summarise her evidence as follows:

247. We find Mrs Ryan was an experienced book-keeper who became Appleco’s bookkeeper in 1998; when CCA commenced trading, she also took on the role of CCA’s bookkeeper. She worked in a small office at Appleco’s premises, which was very close to Mr Trees’ desk which was in the open plan office space. She kept the books and prepared and submitted both companies’ VAT returns. Mrs Ryan retired mid-2006.

248. We find, from her evidence, that Mr Trees was hard-wording and always busy; he could be heard throughout the day talking to people either in person or on the phone.

249. She never thought that there was anything dishonest going on; she believes Mr Trees to be honest and that he would not involve her in anything dubious.

250. She said, and it was not in dispute, that Mr Trees was entirely responsible for CCA’s trading; he put all the deals together and draw up all the paperwork (such as invoices). Her role was to collate the paperwork for submission to HMRC; it was also her role to allocate the payments received to debts due, which she did on the basis of allocating incoming payments to oldest outstanding invoice. She annotated the invoices to show which payment in was allocated to which liability: she accepted her method of allocation might well not match up with Mr Trees’ allocation.

251. Her recollection was that she had a good personal relationship with Mr D’Rozario. They spoke on the phone many times, and Mr D’Rozario visited the office on occasions.

Mr Andrew Tidey

252. Mr Tidey was an insurance broker. He sold CCA the policy we discuss at §§395-407. His evidence was unchallenged by HMRC, on the basis, as agreed with the appellant, that his evidence was taken to include not just his witness statement but his oral evidence given at the previous hearing. We accept his evidence and discuss it below.

Mr Martin O’Neill

253. Mr O’Neill, who was the solicitor now acting for CCA, made a written witness statement on behalf of the appellant which was not challenged by HMRC and so he was not called as a witness. We accepted his evidence.

254. This evidence explained that most of CCA’s due diligence files were seized in the raid in June 2006 and copies provided to the appellant in 2009. They were evidence in this hearing. The remainder of CCA’s due diligence files were given by CCA to its original solicitors which later went into liquidation; due to that liquidation, those papers were not available to CCA in the 2012 hearing but were found before this hearing. The appellant selected extracts to be made available to us in this hearing and we refer to all the due diligence below.

Mr Ashley Trees

255. We deal with Mr Trees’ evidence as we go through our findings of fact. For the reasons we explain in detail below, we did not find his evidence about what he knew, and why he acted as he did, at the relevant time, to be credible. It was inconsistent and implausible. We also found that he played down certain matters in an attempt to present them as something other than they were.

256. For instance, in his witness statement he said CCA only occasionally got credit from his suppliers. When it was pointed out to him that CCA did not pay its suppliers until it was in funds from its customers, he said this happened ‘some of the time’. He was challenged and went on to accept it happened in every deal. Mr Trees’ unconvincing explanation was that he thought ‘credit’ meant agreed credit for a set number of days; he also said it was not real credit as the goods were held at all times by the freight forwarder; when challenged, he accepted that the freight forwarder held the goods to CCA’s order. Another instance of this was that his witness statement said his suppliers would take back goods if CCA’s customers reneged on the deal; in evidence, he appeared to retract that to an extent by saying his customers would ‘listen’ if CCA wanted to renege. It seemed to us that Mr Trees was playing down his original evidence on just how benign CCA’s trading environment actually was.

257. We also found that he avoided giving straight answers. For instance, when it was put to him that CCA’s was a seamless and risk-free business, his answer was that CCA didn’t need staff because it was selling mobile phones and not assembling computers. When asked if was important to CCA if it could return purchased goods to its suppliers if its customer let it down, he talked around the question without giving an answer. These were just examples: it seemed to us that he avoided answering questions to which he had no good answer.

258. In the main, we deal with why we found his evidence unreliable as we go through our findings of fact so we can deal with them in context. But we mention two further instances of unreliable evidence here because their only relevance is to the reliability of Mr Trees, rather than to our other findings of fact.

259. The first was in respect of Mr Trees’ evidence with respect to Pielkenrood and in particular that he only dealt with Simon Pielkenrood, the director of the company. He said he never dealt with Sander.

260. There was some documentary evidence which suggested otherwise: there was a fax from Sander to ‘Dear Ashley’; CCA’s due diligence showed Sander as the contact; Pielkenrood’s trade application form showed Sander as the contact; CCA’s database had Sander down as a contact.

261. HMRC suggested Mr Trees was trying to distance himself from Sander’s evidence (discussed above at §§189-194). Mr Trees denied this and pointed out that he had always accepted he dealt directly with Raj Gathani at Future even though Mr Gathani had pleaded guilty to fraud. We don’t see the situations as similar; it was always clear that Mr Trees dealt directly with Mr Gathani and in any event there was no suggestion that Mr Gathani’s evidence implicated CCA (indeed he pleaded guilty so presumably gave no evidence). Sander’s evidence, however, directly contradicted a key part of Mr Trees’ evidence.

262. We agree with HMRC that the documentary evidence does show that it was more likely than not that it was Sander and not Simon who was CCA’s main contact at Pielkenrood. We consider this yet one more reason why we do not consider Mr Trees’ evidence as reliable.

263. The second instance was that Mr Trees also gave evidence we could not rely on in respect of Soul. There were documents from Soul, stamped with what purported to be CCA’s stamp, which indicated delivery of goods to CCA’s premises. Mr Trees was adamant that goods were not delivered; he was also quite adamant that the stamp was not CCA’s and that CCA did not have a stamp. This was odd because in the previous hearing he had accepted it was CCA’s stamp and then later in this hearing, when confronted with a different stamped document in a different context, did accept that CCA had a stamp.

264. We consider Mr Trees’ evidence on this unreliable. We consider it more likely than not that CCA did have a stamp and it was used on the Soul documents in question. We think Mr Trees gave unreliable evidence as he had no reasonable explanation for why CCA would have stamped a document to say goods were delivered to CCA when they had not been.

Did the appellant know its transactions were connected to fraud?

265. It was assumed, rightly, by all parties that the appellant would be taken to know anything known by Mr Trees. Mr Trees was CCA’s alter ego. So the question is what Mr Trees knew at the time of the deals in question, and we use ‘the appellant’ and ‘Mr Trees’ interchangeably.

266. HMRC’s case is that there were very many factors which would have indicated to the appellant that its transactions were not taking place on the open market and that, therefore, particularly knowing that there was fraud in the mobile phone market, the appellant must have known, and did know, its transactions were connected with fraud. We consider these factors. We then go on to consider the counter-indicators put forward by the appellant.

All deals connected to fraud

267. HMRC’s case was that all of CCA’s deals in the period were connected to fraud and the Tribunal was entitled to conclude that it was unlikely to be the result of innocent coincidence. While the appellant accepted to the extent set out above (and we have found proved) that all of the deals were connected to fraud, the appellant’s case was that this did not indicate knowledge of connection: on the contrary, the appellant’s case is that it was clearly manipulated by fraudsters in every one of its transactions in period in question.

268. We agree with both parties that it was clearly no coincidence that every one of CCA’s deals was connected to fraud: all its deals were, on the evidence set out above, orchestrated by fraudsters for the purpose of fraud. We consider the evidence below when answering the question whether CCA knew this, or whether it was manipulated without its knowledge.

Too good to be true?

269. One such factor was that CCA had, we find, what appeared to be an extraordinarily successful business, with profits increasing exponentially over time. CCA turned over £9.7m in 03/04, £65m next year and £402 million in the following year to April 2006. Its gross profits in the last 7 months of trading were £1.35 million and its costs were not high (just freight forwarding fees and insurance in the main).

270. It was put to Mr Trees that this was phenomenal success; he did not accept this. He pointed out that his background in Appleco meant that, on occasion, particularly in the earlier years, he had had some very profitable deals. It was suggested to him that the situations were not comparable as Appleco traded with a large staff whereas CCA’s profits were entirely down to trades negotiated by him alone. His answer was that he was a workaholic.

271. As HMRC pointed out, after a few years of trading, CCA had a turnover approaching half a billion pounds. At the same time, Appleco, with a large staff with infrastructure, stock and business premises had a turnover some 250 times smaller. It was pointed out to Mr Trees that in 3 months, CCA sold 330,000 mobile phones which amounted to about 10% of the grey market at the time. Mr Trees’ answer was that he was not aware of that and would have been more concerned with his position in the European market.

272. It was pointed out to him that he made no losses on any deals; he said he had made a loss in 2003 (which was accepted) and that he was not in the market to make a loss.

273. We found Mr Trees’ evidence on his knowledge that the deals were too good to be true unconvincing. It was obvious that CCA’s success, if this was genuine trading, was staggering. Mr Trees gave no convincing explanation for why at the time he thought this phenomenal success arose from genuine trading. It was ridiculous to suggest, as he did, that one person’s hard work could be a rational explanation for such success: many people are workaholics without realising any profit at all.

Easy deals

274. All the broker and buffer deals of CCA were ‘back to back’ which meant that the appellant simultaneously bought and sold goods; moreover, it appeared and we find that the deal documentation was all raised in a single day. In other words, what the appellant had to sell was always exactly what his customer wished to purchase. It never had to split up a purchase between a number of customers, nor did it have to purchase from more than one supplier in order to meet a customer’s requirement.

275. Mr Trees’ answer was that in order to ensure that he met a customer’s demand, sometimes he had to buy less than was on offer. We don’t find this answer reliable as we find other evidence from Mr Trees unreliable. In any event, it was put to him that it was all too easy and we think that it was.

No need to advertise

276. Another factor is that CCA did not have any difficulty in locating persons from whom to buy stock and to whom to sell stock. Mr Trees accepted that CCA received many uninvited offers from persons wanting to buy and sell to CCA. He accepted that CCA did not need to advertise; while it had a webpage with its contact details, it did not need to advertise any stock for sale. It seems willing suppliers and customers simply materialised.

277. This was in stark contrast to Appleco which advertised very widely in order to trade.

Very few costs

278. As we have said, CCA’s business had very few costs. Its turnover increased to an enormous amount in a short space of time without requiring much in the way of effort or costs. Mr Trees’ response, when this was put to him, was to play it down. He did not choose to recognise what was plain to us.

279. The only staff required by CCA to turnover millions of pounds was Mr Trees, who put together all the deals, with a little part-time help from Appleco’s bookkeeper (Mrs Ryan) on the accounts and another Appleco staff member (Mr Gordon) who did some due diligence paperwork. Mr Trees suggested an ‘enormous’ amount of work was put in on CCA’s behalf, but all it seemed to amount to was the willingness of Appleco staff to pass on messages to Mr Trees, for paperwork to be faxed, and VAT numbers verified.

280. We agree with HMRC that the ease with which CCA made huge profits should have been obvious to Mr Trees; Appleco employed about 25-30 staff in the labour intensive business of reconditioning and building computers and in 18 months 2004/5 had sales of less than £1million and gross profits of less than £130,000. Mr Trees denied that it was obvious that the comparison with Appleco must have made it clear that CCA’s profits were the result of fraud: Mr Trees denied he knew that but we cannot accept his denial when all the evidence is taken into account.

Very few risks

281. Mr Trees’ evidence was that his suppliers would take stock back if his customer reneged on a deal. This had in fact happened he said with one deal with Infinity some time earlier in their trading relationship. It was put to him that this was all too good to be true. He denied this: he said that it was normal to be able to go back on a contract if payment had not been made.

282. We do not agree that it is normal, but it is clear that Mr Trees knew that he was not at risk if a customer let him down. It was put to him that the deals were too easy and stress-free as he described them: his only, and unconvincing, answer was that CCA had made a loss on a deal in 2003.

283. We consider it was just one more factor which meant that Mr Trees knew at the time that the deals were too good to be true and that the only possible explanation for CCA’s trading was that fraud was at the root of it.

Benign trading environment

Customers and suppliers willing to accept uncommercial terms

284. Linked with the fact that the deals were too good to be true, there were reasons for finding that the trading environment described by Mr Trees was uncommercially benign and he must have known this. Mr Trees accepted that CCA’s suppliers sold what appeared to be very high value goods to CCA without requiring immediate payment; moreover, even though some had terms which retained title, in practice CCA’s suppliers always released control of the goods to CCA long before CCA paid.

285. Mr Trees’ evidence was that the terms of credit was that CCA would pay its supplier when it was paid by its customer. He could not give a rational explanation for why his suppliers were so generous. We also note that CCA was not even required to stick to these terms: Mr Trees’ evidence elsewhere was that CCA, when it received funds, used the money to pay its most pressing supplier which would not necessarily be the supplier of the goods for which it had just received the payment. It all goes to show that CCA’s suppliers extended credit to CCA with no set date for repayment and no requirement for interest. This was obviously not commercial behaviour.

286. Mr Trees also accepted that CCA’s customers were in most cases (35 out of 39) willing to pay for the goods before CCA shipped them. It was put to Mr Trees that its customers were not acting in a commercial fashion: his reply was that they could have checked that the goods existed before paying for them had they wanted to. That was no answer : we find he was unable to give a rational explanation for why a third party customer would pay for goods before receiving them and without checking on their existence and quality. The fact that they did so was a clear indicator to Mr Trees at the time of fraud being at the root of the transactions.

287. We note that in most cases the delay between payment to CCA and shipment by CCA was normally only a few days, in some cases it was a matter of weeks and in one case a matter of 42 days. Mr Trees said he could not remember why this delay occurred and after 12 years we are not surprised; nevertheless, the impression we had is that CCA’s customers were not clamouring for their goods to be shipped even once they had paid. Mr Trees knew that at the time.

288. So, CCA never had to pay for the goods before CCA was paid by its customers. CCA’s customers normally paid CCA before CCA shipped the goods, and sometimes a long time before they were shipped. Mr Trees told the Tribunal that he was not concerned that his suppliers extended credit to him on such generous terms with no interest liability and no set date for payment. In so far as his customer’s willingness to pay in advance, he said that this was the way the business operated: the customers would not get the goods until they paid.

289. But this was clearly not true for CCA which was able to take control of the goods without payment. He was given the chance to explain this but gave no real explanation other than to suggest that it was different when the supply was UK to UK rather than UK to EU because (he said) it was easier for the supplier to get the goods back if the supply was UK to UK. This clearly made no sense whatsoever as a UK supplier giving control of the goods to a UK company had no way of preventing them being exported, and indeed that is exactly what happened. CCA did export them.

290. Mr Trees appeared to be a rational person and was unable to give to us a rational explanation for why at the time he thought that CCA was able to operate with such very accommodating suppliers and customers. We conclude his evidence was not reliable and that he knew at the time that the reason for this benign trading environment was that it was connected to fraud.

Customers and suppliers materialised without any effort

291. We have already mentioned that CCA’s customers and suppliers materialised without any effort on CCA’s part: it was put to him that he must have asked himself why he was not being cut out of market : why didn’t his suppliers supply direct to his customers and cut out CCA? His answer was that he had not thought about it at the time and it wasn’t something a businessman would think about.

292. It was put to him that he knew Future, Soul and Infinity were very large businesses; he accepted he knew that; he was asked why he had not questioned why they sold to CCA rather than directly to customers in the EU? Again his answer was that he had not thought about it, although he did suggest CCA had contacts that they might not have had.

293. We think, particularly taking into account Mr Trees’ knowledge of how easy CCA’s trading environment was, and how customers materialised with no effort on CCA’s part, and knowledge that there was fraud in the market, he would have considered CCA’s role very odd if he was concerned to avoid deals connected to fraud. It was obviously very strange that CCA was in a position to make very large profits for doing very little, when there was no obvious commercial reason why CCA’s suppliers couldn’t have cut out CCA and realised CCA’s profits for themselves. We do not consider Mr Trees’ answer that he did not think about it reliable. We think he knew why CCA was in the supply chain: he knew CCA was a broker in transactions orchestrated for the purposes of fraud.

CCA knew trading irrational?

294. Mr Trees accepted that he knew that the phones CCA traded in were recently imported into UK and he knew CCA was re-exporting them back to the continent; he also accepted he knew the phones would need adapters if they were sold retail in the UK as (mostly) they had 2 pin plugs.

295. His answer was that traders would buy what was available; in effect he was saying it was rational for a trader to participate in a deal in which it could make a profit, and up to a point we agree. But this was his opportunity to explain why he thought the market CCA was trading in existed; it was his opportunity to explain its commercial rationale. He had been able to explain the rationale for deals made by Appleco (see §§130-132): but he gave no real explanation for the trading CCA took part in. We think this indicates that at the time he knew the deals were connected to fraud.

296. Similarly, Mr Trees gave no convincing explanation for why it so happened that if CCA bought from Future, Soul or Infinity, it always sold the goods to an EU customer, but if it bought from its 8 other suppliers, it always sold the goods to a UK registered company. Again this is just another indicator of knowledge because such patterns in trading appear commercially irrational and Mr Trees, who must have known of it, had no explanation for it.

CCA knew customers not interested in goods?

297. We found Mr Trees reluctant to admit that he knew that his customers did not inspect the goods that they were purchasing and had to pay for before they were transported to them. He did accept, when questioned, that not only was there no evidence that CCA’s customers ever inspected the goods, but that (as they were held to CCA’s order) he would have had to give permission to enable his customers to inspect them. His answer seemed to be that there was some level of trust between the parties and the freight forwarder and that his customers did not need to inspect the goods.

298. We did not find his evidence on this convincing: he was describing an uncommercially benign trading environment and he must have known that at the time. His customers gave little specificity as to what they wanted and did not inspect the goods before or after they purchased them, and never rejected them after they received them. He knew about this lack of interest in the goods and was not concerned. Again, we think this indicates he had actual knowledge of the connection to fraud.

CCA used FCIB account

299. Moving on from the uncommercial trading environment, we consider CCA’s use of an FCIB bank account. We found Mr Trees’ evidence about why CCA had an FCIB account unreliable. His answers went round in circles; with an explanation given at one time inconsistent with another explanation he gave earlier or later.

300. His evidence was that FCIB offered a very flexible, online account and CCA chose to use it as it was quick and convenient. CCA also had a Bank of Ireland account which in the 2012 hearing Mr Trees said was opened because it offered a good rate of interest. Nevertheless, when asked in the same hearing why CCA kept a large balance in its FCIB account, Mr Trees said it was because the interest from the Bank of Ireland account was low. It was also his evidence that he chose not to use the Bank of Ireland account for trading.

301. Yet when asked in this hearing why he paid its suppliers within minutes of receiving money from his customer he said it was in part because FCIB, every 4 to 6 weeks, locked accounts down for 1 to 2 weeks such that no money could be moved. This evidence had never been given before and was inconsistent with his suggestion that FCIB was a convenient business account.

302. It seemed to us that Mr Trees gave whatever explanation he thought of at the time in answer to any particular question but was unable to give answers that were, when viewed in the round, consistent. So FCIB was a convenient bank for a business to use when the question was why he happened to have an FCIB bank just the same as all other traders in the MTIC supply chain; but when the question was why he moved the money so swiftly, it was because FCIB might out of the blue lock his account down. There were many other examples of answers which were inconsistent.

303. We do not accept any of the answers he gave as to why CCA held an FCIB account: his evidence was inconsistent and unreliable. Taking into account that it was clearly convenient, even essential, to the fraudsters that the money circulated within the same bank so that it could move as fast as possible, we think the reason CCA held an FCIB account was because it was told to do so. It was put to Mr Trees that he was told to open an FCIB account but he denied this. We don’t accept the denial as reliable: we think more likely than not that CCA held an FCIB account because he had been told that in order to participate in the transactions CCA needed an FCIB account. And that means Mr Trees knew that the transactions were connected to fraud.

Margins must have been determined and not negotiated

304. We move on to consider the margins. We are not dealing with matters in the same order as HMRC as this matter was the one with which Mr Kerr commenced his cross examination. Mr Trees stood by his 2012 evidence that every deal was negotiated individually and appeared to him to be just like any other commercial deal in a highly competitive market. He said he was a keen negotiator and implied that he was out to get the best deal.

305. Nevertheless, a short time later, when asked to explain why it so happened that CCA got a margin of exactly £1 per phone in 115 out of its 117 buffer deals in the period in question, his answer was that he wasn’t trying to maximise profit but realise a pound per phone. That evidence was not only inconsistent with what he had first said but also with the broker deals, where the profits made per phone were much higher. His explanation for that was that his costs in broker deals were higher (transportation, insurance and being out of the money pending the VAT reclaim). But we find his profit was much more than £1 per phone even once these costs were considered.

306. We concluded that he was trying to explain what could not really be explained, other than by knowledge of the fraud, which was the consistency in margin on the buffer deals, and the consistently much higher profits on broker deals.

307. The question of the alleged negotiations was very important because it was CCA’s case that the fraudsters wanted CCA to remain ignorant of the fraud and so would have undertaken negotiations with Mr Trees to make it appear like genuine trading. Yet we found his evidence on negotiations quite unconvincing: it was internally inconsistent for the reasons given above and inconsistent with the documentary evidence which showed that buffer deals always had consistent, round figure margins of nearly always £1, and broker deals had much higher margins. All this suggested Mr Trees did not undertake individual negotiations for each deal.

308. We take account of Mrs Ryan’s and Mr Gordon’s evidence referred to above. In so far as they heard Mr Trees on the phone during the day, negotiating deals, this tells us nothing. We have no evidence at all about what was actually said and to whom and even how many calls were in respect of Appleco’s or CCA’s business. Therefore, in so far as their evidence is relied upon as showing real negotiations took place on behalf of CCA, it is very weak evidence compared to the evidence which we have seen which indicates that negotiations in any real sense cannot have taken place.

309. In conclusion, we find Mr Trees’ evidence that there was a rigorous negotiation unreliable. Mr Trees denied he was rung up, like Sander Pielkenrood, and told from whom to buy, to whom to sell, what to buy and sell and at what price. But Mr Trees’ denial is not reliable: we think the evidence shows that that is what happened. As we have said at §194 above we have excluded reliance on Sander’s evidence: but our finding is consistent with that evidence which indicates, that, had we put weight on it, it was reliable evidence.

Banking evidence

310. As we have said, the appellant accepted that the banking evidence showed that there was an overall scheme to defraud HMRC. The appellant did not accept that the banking evidence indicated knowledge or means of knowledge of that fraud by CCA: on the contrary, it was its position that the banking evidence showed the special position of CCA in the fraud. That special position was (says the appellant) that of a dupe whose innocence was carefully preserved by the fraudster.

311. There was no suggestion by HMRC that Mr Trees knew or could have known of the money movements other than the payments into and out of CCA’s FCIB account. And so it is those movements we consider.

312. The first thing to note is that the money received and paid by CCA did not clearly tie up with particular invoices. The explanation for this was not simply that CCA received and made payments covering more than one invoice. It was not in fact possible to be sure, at least retrospectively, which payments covered which invoices. While there were some notes on the transfers indicating which invoices were being paid, these did not normally match up with Mrs Ryan’s annotations on the invoices made some time later when she reconciled the payments made to CCA’s liabilities (see §250). We do not consider this confusion by itself any indication of knowledge.

313. We also note that CCA did not normally use the money it received from its customer to pay the supplier which had supplied the goods sold to that customer: as the appellant points out, that is only to be expected in genuine commercial trading. We agree but we also think it is not an indicator of knowledge or the lack of it as this lack of linkage is also consistent with the very sophisticated fraud revealed by the banking evidence, where the same sum of money was circulating many times in 24 hours apparently to support a number of different transaction chains.

314. What is relevant is that we find that Mr Trees moved the money which was received quickly on to another entity. Sometimes CCA moved it in no more than 3 minutes, sometimes CCA took a few minutes longer; it never took longer than half an hour save on 2 occasions when it kept the money for some hours (probably overnight). We accept that CCA did not always move the money on as fast as most of the other entities. We also accept that Mr Trees, unlike most of the other entities in the chain (except Future), did not normally transfer exactly the same amount; he normally kept back a sum which varied from about £25 to £1,500.

315. We were given no real explanation by Mr Trees of why such (relatively) small sums were retained. We agree with the appellant that it shows CCA was in a special position in the fraud. But what was that special position? The appellant’s case was that it was retaining small sums as it was a free agent and it could do so if it chose to do so. But it could equally have been that CCA was permitted to hold back these sums as knowing broker in an MTIC fraud. So the retention of these small sums does not help prove the case one way or another.

316. With respect to the differences in speed of transfer, again we accept the appellant’s case that this shows that it was in a special position. Its case is that its special position was as a free agent, choosing when to pay its supplier. But while we accept it was in a special position compared to most other entities, again CCA’s special position may simply have been as the knowing broker in a chain of transactions. In this respect, it is interesting to note that Future was also clearly in a special position as it also held onto the money, sometimes for about 12 hours; even the appellant accepted that Future was a conspirator in the fraud so Future’s special position did not indicate any lack of knowledge by Future.

317. So, was CCA’s special position because it was a free agent or because it knew it occupied the important position of broker in an MTIC fraud chain? And why did CCA always swiftly transfer on to another entity the vast majority of any sum it received, even if it did not always move the money quite as swiftly as most other entities in the chain?

318. Mr Trees’ explanation of this did not really make sense and was not consistent with evidence he gave at the previous hearing.

319. Mr Trees’ evidence in this hearing was that he always had his laptop open showing CCA’s FCIB account and this was not in dispute. He said his suppliers were pressing for payment, so, as soon as he saw a payment had arrived from a customer, he moved on the money (or very nearly all of it) to the most pressing supplier. He said he was not told to whom to transfer the money; he made up his own mind at the time by reference to whom he knew was pressing CCA the most for payment.

320. This evidence was not consistent with the picture he described elsewhere of suppliers who were very laid back about payment: his suppliers were, we find, happy to release the goods to him without payment or any formal credit terms. They were, we find, quite happy to lose control of the goods pending payment. This picture is quite inconsistent with Mr Trees’ suggestion that they pressed him so hard for payment that he would pay them almost the moment money came into the account (even though his suppliers should not have known when he was paid).

321. Moreover, the suggestion CCA was a free agent and paid the money at the time and to the person of its choosing is quite inconsistent with the banking evidence. Mr Trees’ evidence was that he was a free agent in deciding who to pay; it was put to him that he moved the money so swiftly he did not have time to decide which supplier was the most pressing. His answer was that he kept the information in his head. Yet the loops of money were clearly pre-ordained by the fraudsters as the money circulated in a loop very fast. That was only really possible if every person in the loop was sitting at their computer terminal ready to pass the money on as soon as it arrived. It is clear that in many cases CCA’s suppliers were passing on the money from CCA shortly after it was received. Indeed, in approximately half of CCA’s money movements the money sent by CCA was transferred on by the recipient within 9 minutes of receipt and on occasions within 3 minutes of receipt. We find this means that the recipients must have been expecting it: yet how could they have been expecting it if CCA was truly a free agent?

322. In this hearing, Mr Trees also gave evidence he moved the money swiftly not only because of pressure from his suppliers but because of the risk of FCIB suspending his account. But we don’t accept this: it had never been mentioned before and if he was truly a free agent, he would have been free to use his Bank of Ireland bank account for trading: there was no suggestion that this account had ever been suspended without notice and for no apparent reason.

323. Lastly, we note that the evidence given by him in this hearing could not really be reconciled with the evidence he gave in the previous hearing when the Paris server evidence was not available, In the first hearing, he had said that he would check at end of day whether the promised money had come in and then pay it on to his supplier. There was no suggestion from his evidence in 2012 that he paid it on with the speed with which he actually did.

324. In summary, we do not consider what Mr Trees said about the reasons he moved the money as swiftly as he did, and to whom he transferred it, to be in any way reliable. On the contrary, we think the banking evidence indicates that CCA was told when it was receiving money, and how much it was to transfer to whom and (normally) to transfer it very fast. We think that the banking evidence indicates very strongly that Mr Trees was well aware that CCA’s transactions were connected to fraud. It was put to Mr Trees that he was told to be at his computer and to transfer the money as it came in: he denied this but in light of the evidence we do not accept his denial as reliable. It was put to him that the money movements meant he must have known of the connection to fraud: he denied this but we do not accept the denial for the reasons given above.

CCA’s closing liabilities

325. Before leaving the banking evidence, we deal with one more financial issue: this issue is whether and to the extent that CCA owed the three contra-traders money and what inferences should be properly drawn from this.

326. We explained above that a broker in an MTIC fraud chain in theory could be at either end of a spectrum or anywhere along it: it could be owned and controlled by the fraudster; at the other end of the spectrum it could be entirely independent of the fraudster and ignorant of the connection to fraud. The appellant’s case was that that was its position: it was free agent ignorant that its transactions were connected to fraud. It suggested that the fraudster would cause its buyers and sellers to negotiate the transactions with it in order to maintain the appearance of commercial reality so that CCA would not appreciate the connection with fraud and therefore the risk of non-repayment by HMRC.

327. HMRC suggested that by June 2006, CCA was only able to continue to participate in its transactions because it was trading on credit provided by its suppliers, Future, Infinity and Soul. In other words, its suppliers were not demanding payment for previous transactions but were happy to continue selling to CCA on credit.

328. The appellant did not accept that this had happened (bar that it accepted that its suppliers did not require to be paid immediately). We agree with the appellant that HMRC did not satisfy us that CCA ended up as a debtor of Infinity or Soul. While the transactions in the FCIB accounts showed substantial unpaid liabilities, especially to Soul, CCA had another bank account and statements from this had never been provided to HMRC. HMRC suggested the failure to provide them was deliberate; the appellant’s position was that HMRC asked for them many years too late and they no longer existed. We note that both Infinity and Soul are in liquidation but their liquidators have not claimed any debts owing to those companies from CCA. We find it unproved that CCA owed either of them substantial sums of money.

329. The position was different with Future. While HMRC claimed (but we consider were unable to prove) that the figure was much higher, even the appellant accepted that it owed Future £5.6 million in unpaid invoices. Future’s management receiver had originally claimed £9.2 million but, as it appeared he may have accepted some offsetting, we find that the figure owing in unpaid invoices was only £5.6 million.

330. We also take account of the appellant’s case was that very little of this sum was actually owed to Future because CCA now intended to offset the full amount of the invoices for the Samsung Serenes and P990s for which it had been denied input tax. This was on the basis that the phones did not exist: Future had, therefore, failed to supply them and, said Mr Trees, CCA ought to be repaid the full invoiced amounts. However, whether or not the appellant is right about its entitlement to offset its debt for these phones (which we doubt as CCA appears to have sold more such phones to Future in its buffer deals than it purchased from Future in its broker deals), the point remains that in June 2006, CCA was only able to continue trading because of the credit provided by Future. This indicates a number of things.

331. Firstly, it is inconsistent with the model of an innocent broker whose role was to pay more in cash for the goods than it received for their sale, thus enabling the fraudster to extract the profit form the fraud. In this case, there is no dispute that Future was a party to the fraud; we find it was a party to the fraud for reasons given above at §§212-220. Future, a fraudster, was clearly not going to get its money back, let alone realise its profit, unless CCA was repaid its input tax by HMRC and therefore able to repay its debt to Future.

332. This suggested a rather different, and closer, relationship between Future and CCA. It indicated a degree of trust in CCA by Future; we consider it indicated that CCA was not an intended victim of the fraud in which Future was a knowing participant.

333. It also indicated that Future, a knowing party to the fraud, was prepared to risk its own money in the fraud. The appellant suggested that the fraudster would go to a lot of trouble, such as setting up artificial negotiations with CCA, to keep CCA from realising its transactions were connected to fraud because, said the appellant, no one would knowingly put their own money at risk in transactions connected with fraud. But it is clear that Future did so. Future, a knowing party to the fraud, must have thought the risk of non-repayment to CCA from HMRC to be low enough to be worth funding CCA to continue trading. So, there is no reason to think that Future or any other of the fraudsters would have been concerned to keep CCA ignorant of the fraud.

CCA’s relationship with the contras

334. And that financial evidence brings us to a more general question about CCA’s relationship with the contra-traders and in particular with Future.

335. The appellant’s case is that Future appeared to Mr Trees to be a large and successful business and he was entirely taken in by it. He was horrified to learn, as he did shortly before the hearing in 2012, that CCA’s main contact at Future, Mr Raj Gathani, whom Mr Trees had always thought trustworthy, had pleaded guilty to fraud.

336. We note, almost in passing, that Mr Trees gave evidence about what he considered a very impressive brochure given to him by Future about ‘IGB’, the Innovative Global Business Group Ltd, which was by the end of 2006 the (new) holding company for Future. We found that, whether or not it was reasonable to rely on Future’s self-generated self-promoting brochure, Mr Trees had not relied on it at the time of the deals in question as it did not come into existence until October 2006.

337. We agree that Mr Trees trusted Future. We find that the appellant’s trust in Future lasted even when there were indicators of serious trouble: as the end of 2006 approached, Mr Trees knew that HMRC had refused to repay CCA some £10 million in VAT relating to all its trading in April - June 2006, much of it with Future, such that CCA had had to cease trading; he knew that officers had raided CCA’s premises at the start of June 2006; he knew that HMRC were saying that over £5million in Samsung Serene and P990 phones which Future had supplied to CCA did not exist.

338. Mr Trees was nevertheless content to continue to put a great deal of trust in Future. On 5 December 2006, CCA paid a company (Crystal Cars) which it knew was related to Future, and on whose behalf Raj Gathani acted, some £237,000 and then on 11 December a further £110,000. The payments were, said Mr Trees, to purchase two second-hand luxury cars, although he could produce no paperwork to prove this. CCA, it seems, did not take possession of the cars, but cancelled the deals a few days later and received a refund of all the money, with an additional £584, on 19 December 2006.

339. Mr Trees’ explanation was that he had customers lined up to buy the cars who had backed out when they had discovered the cars were imported: Crystal Cars was happy to cancel the deal and refund the money. He still trusted Future, he said, because Mr Gathani had shown him witness statements from the manufacturers which had satisfied him that the Samsung Serenes and P990s had existed. He could not explain the £584 discrepancy.

340. HMRC insinuated, but did not prove, that Future used the money from CCA to carry out a deal with Unique Distribution, a company which HMRC suggested (but did not try to prove) was a conspirator in MTIC fraud.

341. What we do find is that Mr Trees did have a trusting relationship with Raj Gathani and Future. Future clearly trusted CCA to the tune of over £5million (see §329); Mr Trees trusted Mr Gathani and a company in the same group as Future with over £300,000 at a time when he had real reason to be concerned about the connection to fraud in deals he had undertaken with Future.

342. Even though CCA’s main contact at Future, Mr Gathani, later admitted fraud, the trust between Mr Gathani and Mr Trees does not by itself prove that CCA knew of the connection to fraud. But when this factor is taken with all the other factors discussed above, it does add colour to the overall picture described at §§269-333 of Mr Trees (and through him CCA) as someone prepared to trade knowing of the connection to fraud.

The appellant’s actions only consistent with knowledge of fraud?

343. HMRC’s case is that decisions taken by and actions of CCA indicate that it knew its transactions were connected to fraud, because it would have acted differently if it had not known this.

General knowledge of fraud in mobile phone market?

344. HMRC’s case was in part based on saying that Mr Trees knew of MTIC fraud in the mobile phone market place and so, had he wished to ensure CCA’s deals were not connected to fraud, he would have acted differently. So, the first question to address is what Mr Trees actually knew of MTIC fraud at the time of the deals in question.

345. HMRC’s case was that Mr Trees had a good general knowledge of MTIC fraud and would have known enough to know that he was at risk of trading in MTIC fraud chains and that he knew enough to know that he had to be wary of deals offered to him by both suppliers and customers.

346. We consider that in the hearing Mr Trees played down his knowledge of MTIC fraud. While he accepted he had some knowledge, he indicated that he had no idea how serious a risk it was. He suggested carouselling was not explained to him although it was clear from his contemporaneous notes that Mr D’Rosario had explained this in 2004 and in any event Mr Trees understood that HMRC wanted IMEI numbers because of their desire to look for circulation of the same phones. On more than one occasion in the hearing he said that he had not understood that his customers presented a risk as well as his suppliers: yet when challenged he accepted he had known this.

347. We find that he had a good general understanding of MTIC fraud and he knew that the risk was of orchestrated deals and carouselling. We think this because it is clear from Mr D’Rozario’s log that Mr Trees was told this; moreover, he knew HMRC had concerns with CCA’s former customers such as Y International and Easy Trading; he knew HMRC was concerned that phones were being continually re-sold and that was why they wanted IMEI numbers recorded.

348. He also knew that his own supply chains could be suspect as in 2003 CCA received letters from HMRC stating they had found fraud in CCA’s supply chains and in particular tax losses were identified in chains in its 09/03 and 04/04 VAT return periods; periodically, its repayment claims were withheld pending verification. This happened for periods 09/03, 10/03, 11/03, 03/04, 04/04 and 05/04.

349. We accept that contra-trading was not explained to Mr Trees before CCA ceased to trade: we accept this as HMRC themselves did not appear to understand contra-trading until some time in 2006 and Mr D’Rosario himself not until after CCA had ceased to trade. We do not see that this lack of knowledge is significant as, as we have already said, contra-trading would not hide the fraud from a broker because the broker would not be able to investigate its own supply chain. An ignorant broker informed about contra-trading would not be in any better position to spot the fraud than an ignorant broker in the same circumstances but who was ignorant about contra-trading.

350. So, we find Mr Trees was well aware of the risk of MTIC fraud in mobile phone trading; he knew the risk was with the entire supply chain, including his customers; he was aware phones might be carouselled: what did he do about it?

Lack of interest in goods

351. CCA’s invoices and purchase orders did not contain much detail about the phones that CCA was purchasing and selling. For example, the invoices did not specify that the phone had to be SIM-free. They did not specify that the phones had to be new. Mr Trees’ answer was that neither criteria needed specifying because they were obviously terms of the contract. He suggested there was an ‘understanding’ in the industry all phones would be new and sim-free.

352. Whether a phone is tied to a particular network operator obviously affects its retail value: Mr Trees accepted the significance of the phone being SIM-free in his answer. And it is patently obvious that whether or not the phone is new affects its value. And we don’t agree, as Mr Trees’ implies, that the need for the phones to be new and Sim-free would be necessarily implied in the contract: on the contrary, the failure to specify it suggests that the parties weren’t really concerned.

353. HMRC pointed out that the contracts did not specify which kind of adaptor (ie UK or EU) was supplied: there was a clear cost to a retailer if a second adaptor had to be supplied if the phone came with the wrong one for the country in which it was sold. Mr Trees’ answer was that it wasn’t important even though he accepted the cost might be as much as 50p per phone and in the buffer transactions his margin was only £1.

354. His evidence was that he was not concerned by the lack of specification and read nothing into his customers’ lack of interest in the phones’ specification. This seemed inconsistent with his later evidence that he ensured that his supplier declaration required his suppliers to confirm the phones had original manufacturers’ specification. He said he did this because Appleco had been caught out by lack of specification on a deal in the past which meant it paid too much for computers which it thought had a higher spec than they actually did.

355. We did not find his explanation of why there was a lack of specification on the phones convincing. He clearly did understand the need to specify what was being bought and sold and yet at the time had not done so in commercial detail in CCA’s deals nor been concerned by the failure of most of his trading partners to do so. This suggests to us that he wasn’t really interested in the stock he was trading as he knew the transactions were orchestrated for fraud.

Deal documentation inadequate for commercial deals

356. CCA’s deal documentation, such as its invoices and purchase orders, did not contain terms: they did not specify time of payment nor contain any clauses about when title would pass. As the goods changed hands at a time other than the time of payment, we think the question of title would have been quite important, particularly for valuable goods.

357. Mr Trees’ explanation for this was not particularly convincing: he said terms couldn’t be put down in writing as they would be too long. This was clearly not correct: important terms could be set down quite shortly. It was pointed out to him that some of CCA’s counterparties did specify terms (eg Pielkenrood’s paperwork specified the type of phones it wanted and required despatch within 3 days of payment). He also said that it was unnecessary to specify anything in writing as it was all agreed orally: again this was not a convincing answer as the whole point of writing things down is to record what was verbally agreed.

358. We consider that he failed to give a sensible reason as to why basic contract terms for contracts involving very large sums of money and what appeared to be very valuable goods were not reduced to writing. At best his answer amounted to saying that everyone trusted each other so terms did not need to be put in writing.

359. The evidence, however, is that the deals were orchestrated and whether or not Mr Trees knew that, it is clear that every other participant in the supply chains must have known that. They were therefore not interested in the goods or the contractual terms but simply in playing their part in the charade: this is reflected in the lack of contractual terms and specification of the goods. The fact that CCA’s own contracts also lacked such specification and terms suggests that Mr Trees similarly was not interested and for the same reason.

Due diligence inadequate for commercial risks

Date of due diligence

360. Exactly how much due diligence CCA undertook was a matter of some dispute between the parties. From the start CCA kept a record of its trading partners’ VAT numbers and always checked their validity with HMRC Redhill: this tells us nothing as CCA would be unable to reclaim its input tax if either its supplier or customer was not validly VAT registered.

361. It had undertaken more detailed checks on many of its trading partners but HMRC’s position was that most of this due diligence had been undertaken after Mr D’Rozario informed Mr Trees in mid-May 2006 that he was coming to check it. CCA’s position was that it was in the process of revamping its due diligence in any event and much of its original due diligence had been superseded by its revised May/June 2006 due diligence and its inability to produce very much from an earlier period was for this reason.

362. CCA’s due diligence database was extensive, covering some 2,500 entities, but of the material in the hearing almost all of the entries dated to mid-2006. We find there was some due diligence conducted before that date: for example, Mr Trees had undertaken some visits and had a good credit rating report for Future from 2004. We find there was little evidence that much due diligence was conducted before that date. That is consistent with Mr Gordon’s evidence which was that he was only asked to start work on CCA’s due diligence some time in 2006. In conclusion, we find most of the due diligence was done in mid-2006 which was after CCA’s trading relationship with most of its trading partners had commenced.

Thoroughness of due diligence

363. CCA’s case was that it carried out thorough due diligence requiring its trading partners to produce their certificate of incorporation, VAT registration certificate, letter of introduction, trade application form, director’s ID eg passport; director’s proof of address eg utility bill; and banking details. It also obtained a credit rating report on its trading partners.

364. The database recorded all due diligence undertaken and included many companies it had never traded with. Due diligence for some of the companies it had not traded with had comments such as ‘do not deal with’ and gave a reason, such as the company refusing to cooperate with CCA’s due diligence enquiries. For Zena International Ltd the note on CCA’s database said not to deal with them as they had limited product knowledge. Mr Trees also said he reported that company to Mr D’Rozario and we accept that he did.

365. On the surface, at least it appears as if CCA, by mid-2006, took due diligence seriously. We consider it in greater detail.

Credit rating reports

366. CCA obtained Creditsafe reports for various counterparties. There was a common theme in that many showed that the subject of the report had a very low or bad credit rating and/or no accounts. For instance, the report for Universal Trade Supplies Ltd (incorporated less than a year earlier) showed they had no accounts and no credit rating; the report for Sound and Secure (established for over a decade) showed they had zero credit rating.

367. Mr Trees’ answer was that he was not concerned by any of this as most the companies were new as it was a new market and a good credit rating was not to be expected; even where the company was not new, he was not concerned as CCA did not have a good credit rating itself, and, moreover, CCA was not extending credit to its trading partners.

368. It was put to Mr Trees that the reports told him nothing: his answer was that they would show the company was real and existed, and the other companies with which the directors were connected. He would not ask a company for its accounts, he said, as it was rude.

369. But it seems to us fairly pointless to obtain credit rating reports if the credit rating was to be ignored. We found it unconvincing when Mr Trees said he did not ask to see his trading partners’ accounts because he thought it was rude.

Independent references

370. We find that, while by mid-2006 CCA was checking that some of its trading partners had appointed accountants, it did not ask for a reference from a professional or bank.

371. Its counter parties might give the names of trade references but it appears CCA almost never approached the referee for a reference. For instance, around mid-2006, CCA checked its accountants had Future on its books and asked for (but did not appear to receive) their trade references. In any event, Mr Trees accepted that most of CCA’s due diligence on Future was done in May 2006. His view seemed to be that they trusted each other.

Visits

372. Mr Trees did visit some of CCA’s counter parties, and we accept in many cases this type of due diligence was carried out before May 2006. In general, Mr Trees visited CCA’s suppliers but not its customers. He did not visit Allimpex, Pielkenrood, Mohomedbhay nor Universal (although he did meet the manager of Universal in the UK). Mr Trees said he thought the directors of Allimpex and Mohomedbhay attended CCA’s premises and we have no particular reason to doubt this.

373. Mr Trees made a number of visits to Future but he created no notes of any meetings, just descriptions of the premises. He suggested that he did not carry out many visits because all that would happen is that he would see people working in offices. We agree that the visits appeared to be ineffectual for due diligence.

Due diligence on freight forwarders

374. We find CCA undertook virtually no due diligence on the freight forwarders. While Mr Trees insists some was undertaken, the only evidence was copies of A1’s letters of introduction. He suggests CCA got rid of its due diligence. We did not find Mr Trees to be a reliable witness so we do not accept this: it seems to us more likely and in keeping with CCA’s general attitude to its due diligence, such as it was, that it looked at its suppliers rather than its customers or freight forwarders.

375. So far as his customers’ freight forwarder was concerned, he told the Tribunal he was not concerned that all his EU customers wanted the goods shipped to Boston Freight in Belgium. He said it was not his problem; when it was put to him it was an indicator that the phones were circulating, his reply was that he had not understood the risk of circulation at the time; when it was put to him that his evidence on IMEI numbers showed that at the time he did understand the risk of circulation of the phones, he accepted that, but then said he thought the coincidence was explained by Boston Freight being used as they were specialists, although he admitted he never undertook due diligence on them nor checked whether they were specialists.

376. This was yet one more example of unreliable evidence from Mr Trees; he changed what he said. Moreover, he was unable to give a good answer as to why he was not suspicious of the universal use of Boston Freight.

Negative indicators ignored

377. We find that the due diligence which was carried out appeared to have very little influence on CCA’s decision to trade. While there is evidence of CCA noting its files that it would not trade with a particular company, that was for a glaring matter such as a refusal to complete a supplier declaration form.

378. Mr Trees was not concerned if a company had changed direction: due diligence on Allimpex indicated that the company specialised in renting out machinery; CCA was happy to sell them mobile phones. Similarly, the report showed Pielkenrood’s business had been servo drives (a computing part); Mr Trees was not concerned with the change of direction. We have already commented he obtained credit rating reports but was unconcerned about a negative credit rating.

379. Infinity gave Future and A1 as its referees despite Infinity apparently being in competition with Future; Infinity then gave Soul as its trade reference; Mr Trees was not concerned by the coincidence but said it gave him confidence as Future was the largest mobile phone distributor in the UK.

380. Soul and Mohomedbhay gave the same trade reference (Eliyon), although one was CCA’s supplier and the other a customer. Pielkenrood gave CCA in July 2006 a reference from Infinity: as CCA knew it had sold goods to Pielkenrood that it had purchased from Infinity, this should have looked odd. Mr Trees’ reply was that the connection between those companies gave him confidence rather than concerned him.

381. The due diligence on Soul was obtained in late June 2006: it showed their address as being next door to Future. CCA did not check on who was Soul’s landlord but if it had, it ought to have discovered it was Future. Mr Trees accepted that the physical proximity between Future and Soul should have been a negative indicator to him but said that he never saw the due diligence, which was carried out by Mr Weston: he had visited Soul at different premises.

382. Bearing in mind that Mr Trees was well aware of the risk of fraud, his statements that the connections between the companies of which he was aware gave him confidence rather than concern are, we find, consistent with CCA’s general lack of interest in negative indicators. And the fact he did not see the due diligence on Soul, which was a major trading partner of CCA’s suggests, consistent with other evidence, it was a paper exercise undertaken by Mr Gordon on Mr Trees’ instructions in order to satisfy HMRC rather than inform Mr Trees’ decision making.

Purpose of due diligence

383. Mr Trees’ position seemed to be that the due diligence was undertaken primarily for VAT risks; so, for instance, he accepted that his due diligence on his suppliers was better than on his customers. It seems he did not see due diligence as protecting CCA from commercial risks such as customers who reneged on deals; that is perhaps not surprising bearing in mind, as we have already said, CCA had an uncommercially benign trading environment and Mr Trees knew that.

384. Moreover, although Mr Trees reverted to saying that he thought the risk of fraud was in the supply chain and not connected to his customers, we have found (see §§346-7 and 350) he was aware that the risk of fraud applied as much to customers as to suppliers. Indeed, in some cases he did carry out a little due diligence on customers.

385. So the patchiness of the due diligence, the fact that most of it was carried out after many of the deals in this appeal, its lack of thoroughness, and the fact negative indicators were routinely ignored, means we find it was more likely than not that its purpose was not to protect CCA from being involved in MTIC fraud but as window dressing to be shown to HMRC, to bolster CCA’s claim to refund of the input tax.

Knowledge his trading partners did not carry out due diligence?

386. Mr Trees clearly knew that a failure to carry out due diligence on a trading partner was a negative indicator as it was his evidence that he told Wesley Gordon that CCA would not trade with such companies. He insisted in evidence that all CCA’s trading partners had carried out due diligence on it.

387. Yet there was very little evidence that any of CCA’s trading partners had conducted much due diligence on CCA. Future asked for CCA’s certificate of incorporation and its VAT registration certificate. Representatives of Infinity and a few others (see §372) visited CCA’s premises, but it appears neither side kept notes of the meetings.

388. Apart from that, there is very little evidence of due diligence being conducted on CCA. We consider it more likely than not that virtually none would have been done: as we have said the deals were orchestrated for the purpose of fraud and it is clear everyone else in the chain must have known this. They would have no interest in genuine due diligence and (unlike CCA) no interest in window-dressing as they would not be in a VAT repayment position. We have already commented on the fact that the contra-traders (CCA’s suppliers) did not cooperate with HMRC and did not appear to carry out due diligence (bar as mentioned above). Mr Trees must have known its counterparties were not carrying out real due diligence on CCA and yet he was content to trade with them.

Conclusion on due diligence

389. We don’t accept Mr Trees’ evidence that much due diligence was carried out before May/June 2006; apart from visits, there is little evidence of this and we do not consider him a reliable witness. We also think that the due diligence that was actually carried out was carried out to satisfy HMRC: it was clear that CCA was happy to trade before May/June 2006 with the same companies without much due diligence so the purpose of the due diligence was clearly not to protect CCA against trading risks. We do not think Mr Trees really suggested that it was.

390. Mr Trees implied that CCA did not need due diligence as it traded with companies with which it had an established trading history and it trusted: but it appears CCA was also content to trade with new companies as well with little in the way of due diligence.

391. And when due diligence was undertaken, negative indicators were ignored. The due diligence was of little real value in any event as Mr Trees did not look at his trading partners’ accounts or obtain professional or banking references; the few trade references obtained appeared to raise questions rather than provide reassurance.

392. The appellant’s case was that thorough due diligence would not have discovered the fraud and we agree with that. Thorough due diligence might well have made CCA suspicious (eg it should have revealed the connection between Future and Soul) it would not by itself have shown a definite connection to fraud.

393. But the evidence on due diligence is relevant as it shows, we find, that Mr Trees understood CCA was trading without taking any real commercial risks from which it needed due diligence to protect it. This is yet more evidence that Mr Trees was not worried about the deals going wrong, or being left holding goods and liable to pay for them without having a customer. It also shows that Mr Trees was not looking at the evidence with a view to avoiding a connection to fraud: he was not concerned by the patterns in the due diligence, of odd connections between companies and of consistently poor credit ratings. The purpose of CCA’s due diligence was to satisfy HMRC.

394. The only explanation for all of this which makes sense is that at the time of the deals CCA knew that they were connected to fraud. It was put to Mr Trees that his due diligence was not consistent with wanting to check the legitimacy of his customers; he denied that but we do not accept his denial as reliable for the reasons given above. It was put to him that his due diligence was ineffective because he didn’t want to find out anything because he knew his deals were connected to fraud; he denied this as well but we do not accept his denial as reliable for the above reasons.

Inadequate insurance

395. CCA was buying and selling what appeared to be very valuable mobile phones. It was responsible for shipping them to its buyers. CCA’s position was that up to the end of 2005 it relied on the insurance held by the freight forwarder to cover them in transit; HMRC did not challenge this. From January 2006, the law changed and it had to carry its own insurance in order to be covered. Mr Trees arrange a marine freight insurance policy through an agent, Mr Tidey. The policy cost CCA £19,200.

396. HMRC’s case was the policy was ineffective to cover CCA and that it was just window dressing as CCA was not really concerned about the risk of loss in transit.

397. We consider the policy was inadequate for a number of reasons. Firstly, it listed the countries to which deliveries were covered. They did not include Belgium, when all of CCA’s deliveries were made to Belgium. Mr Trees accepted that this was true: his point was that adding Belgium to the policy would not have increased its price and so it was simply an oversight on his part.

398. We agree that whether or not the policy was intended as window dressing or genuine insurance, it was clearly an error that Belgium was not specified. Nevertheless, such carelessness indicated a lack of interest in the details on Mr Trees’ part.

399. The policy covered transit of goods worth £20 million per year. This was clearly inadequate: CCA sent goods worth £18 million in April 2006 alone. Mr Trees’ explanation was that he understood that that was just an estimate which could be revised upwards. Mr Tidey’s evidence was rather different: he considered the estimate had to be a genuine estimate. We find it was not. Even though Mr Trees denied this, we do not accept his evidence as reliable: the insurance was obviously inadequate.

400. We think it was obvious to Mr Trees that the insurance was inadequate; £20 million a year was clearly a massive underestimate. CCA’s previous year’s carryings were £79 million. There was no suggestion CCA had even attempted to revise the estimate when (by June 2006) they had exceeded it several times over. Mr Trees said he did not think it necessary to do so; he suggested he could pay an additional premium at the end of the year.

401. We find this was not consistent with Mr Tidey’s evidence nor with good sense if Mr Trees wished to have insurance that was effective.

402. We also note that the policy had a limit of £750,000 per ship yet the evidence shows that CCA’s shipments exceeded this by many multiples meaning in the event of loss CCA was, for this reason as well, uninsured. Mr Trees accepted he became aware of this but stated it was the fault of Aquarius the freight forwarder and against his instructions; he suggested that he refused to pay them because of this.

403. CCA’s insurance was inadequate in other ways; the marine cargo policy only covered goods in transit abroad, it did not cover goods in transit or in storage in the UK. Yet CCA owned the goods and was at risk while they were stored and they were stored for days or weeks as CCA did not ship them until the earliest when it was paid. Mr Trees’ explanation was that he relied on insurance held by the freight forwarder (Aquarius); he said he had fallen out with his previous freight forwarder (A1) for not providing a copy of the insurance. He did not agree with Mr Kerr that the freight forwarder made it clear that it was for owners to hold insurance.

404. The policy also only covered mobile phones; whereas some of the deals included DVD payers, GPS and laptops. Their value was £4 million. Mr Trees’ reply was that CCA self-insured them although he accepted that meant CCA was uninsured. He said he was happy to take the risk. We note that in the context of his willingness to take risk, it was his unchallenged evidence that he did not insure his home and, having spent a lot of money on securing Appleco’s premises, did not insure them either.

405. HMRC’s case, which Mr Trees denied, was that he bought the marine cargo policy in order to assist his repayment claims rather than because he wanted any insurance. We note that at this point it was Mr Trees’ evidence that he did not know that HMRC would check CCA held insurance when considering whether to release an input tax reclaim; it was pointed out to him this was inconsistent with what he had said earlier; he then accepted that he had known at the time that Mr D’Rozario was interested in CCA’s insurance position.

406. We consider that Mr Trees’ evidence was clearly unreliable in some areas (such as his knowledge of HMRC’s interest in insurance policies) and we had to be careful of all he said. We consider that in other spheres (such as Appleco and his home) he was prepared to take the risk of no insurance, but at least in respect of Appleco, it was a calculated risk as he spent significant sums on security. With CCA, we think he was also prepared to trade without insurance; it is more likely than not on the evidence we have that the reason CCA bought the inadequate marine insurance policy which it did was to assist with its VAT refund claims and not because it genuinely wanted insurance to protect itself from risk. This explains the policy’s inadequacy and the error over Belgium.

407. The question is why it did not want protection from the risk: with Appleco, Mr Trees took a calculated risk as he was able to diminish the risk of theft: CCA’s stock appeared to be many multiples more valuable and yet no steps were taken to diminish CCA’s risk of damage or theft of them, other than the purchase of this obviously inadequate insurance policy. It suggests to us that Mr Trees was not concerned with the risk, which strongly suggests he knew that there was no real risk because he knew the trading was orchestrated and his trading partners not really interested in the goods supposedly the object of it.

Lack of willingness to provide IMEI numbers

408. Mr D’Rozario kept a log of his contact with Mr Trees; he also produced the correspondence between the parties. It was clear from this that starting in 2003, Mr D’Rozario asked Mr Trees to keep records of the IMEI numbers of the phones which he was trading. It was also clear that Mr D’Rozario explained to Mr Trees why it would help HMRC identify fraud: it would help them identify phones which were being repeatedly sold (in other words, the ‘carousel’ element of MTIC fraud). Mr Trees accepted he knew that IMEI numbers would identify circularity.

409. We find that, from 2003, Mr Trees’ normal reply to Mr D’Rozario was that he would consider providing the numbers to HMRC. It was HMRC’s case that nevertheless he never provided any IMEI numbers to HMRC.

410. At the start of the hearing before us, it was the appellant’s case that Mr Trees had provided a list of IMEI numbers to HMRC in January 2006. The list was produced: it had not been before the previous hearing as it had only been found more recently. Mr D’Rozario’s logs did not record any receipt of IMEI numbers and Mr D’Rozario himself had no recollection of receiving any.

411. On closer investigation, as the hearing progressed, it became apparent that the list could not have been sent to HMRC in January 2006. There was a record that the appellant’s freight forwarders had told Mr Trees none could be provided for any period before April 2006; moreover, the list of phones, while it appeared to Mr D’Rozario to contain genuine IMEI numbers, did not clearly relate to any of CCA’s deals because of the mixture of model numbers.

412. We accept that at some point in early to mid-2006, CCA did obtain a list of some IMEI numbers; we do not accept that they were ever provided to HMRC. We prefer Mr D’Rozario’s evidence; it is consistent with the documents and he was a reliable witness. Mr Trees was not.

413. The significance of the IMEI numbers is the reason why CCA failed to provide them to HMRC while at the same time kept promising HMRC that it would consider doing so. It was put to Mr Trees that he was stringing HMRC along, and we think that he was.

414. In the hearing, he gave a number of reasons for not providing the numbers to HMRC. He said it was too expensive (50p per phone he suggested); he also said he believed HMRC would not tell him the results of the check if he provided the numbers to them.

415. He was also insistent in the hearing, repeating something he said first in his 2018 witness statement, that a computer could not, in a reasonable time, carry out a check as each IMEI number was 15 digits long and there were about 2 million IMEI numbers in HMRC’s Nemesis database. He suggested it would take weeks for any computer to work out if one 15 digit number matched any of the 2 million IMEI numbers held by the computer. We had no expert evidence on this, but we agree with HMRC that we could take judicial notice of the powers of computers in 2006 and this evidence was not credible. In any event, it made no sense: if computers were incapable of making this match, why would HMRC have had the Nemesis database? Moreover, it was clear that Mr Trees knew that the Nemesis database was used by HMRC. We thought this evidence about computer processing inadequacy was a poor attempt by Mr Trees to explain his reluctance to assist HMRC by providing IMEI numbers.

416. We do not accept that obtaining lists of IMEI numbers was too expensive: it was clear that at some point CCA did obtain some IMEI numbers. We do accept the appellant’s point that HMRC would not necessarily have informed CCA of the outcome of any IMEI number check but we comment that this would not have concerned Mr Trees if, as Mr Trees, said, CCA genuinely wanted to assist HMRC in its fight against MTIC fraud.

417. We also accept Mr Pickup’s point that there is no allegation or attempt to prove that CCA actually dealt in phones that were being carouselled, any more than HMRC attempted to prove (other than in respect of the Samsung Serenes and P990s) that CCA was dealing in phones that did not exist or did not meet the specification in the invoices. We agree that (other than in respect of the Samsung Serenes and P990s) it is not proved the phones did not exist or did not meet specification or were being carouselled; but the contrary is also not proved.

418. It seems to us that, as Mr Trees did not obtain the IMEI numbers and did not inspect the goods himself, he could not know that the phones existed, met specification and had not been carouselled. If he was ignorant of connection to fraud, he would have nothing to fear from providing the IMEI numbers to HMRC; but if he did know of the connection, or suspected it, sensibly he would have to be concerned about the risk to CCA if HMRC saw the IMEI numbers of the phones it traded.

419. We think that Mr Trees did not want to give HMRC IMEI numbers because he was aware that the check might reveal that CCA was dealing in phones that either it or someone else had dealt in before and that the phones were circulating between wholesalers rather than reaching a retailer. If Mr Trees had truly believed that CCA’s transactions were not connected with fraud, he would, we think, have been happy to provide what Mr D’Rozario wanted to satisfy him of CCA’s bona fides.

420. Having said that, we accept that Mr Trees’ failure to provide the IMEI numbers by itself might indicate nothing more than that he suspected his transactions were connected to fraud: suspicion is not enough to deny input tax. However, when this factor is considered with all the other factors which we have discussed, we think it is just one more element that shows Mr Trees knew CCA’s transactions were connected with fraud.

Failure to inspect goods

421. Mr Trees never inspected the goods which CCA bought and sold. It was his case that CCA occasionally paid its freight forwarders to undertake an outer box inspection, although we find no evidence that it ever paid for an inspection.

422. The evidence shows that by mid-May 2006 Mr Trees knew that Mr D’Rozario was questioning whether certain models of phones existed in the quantity in which they were bring traded: it was put to Mr Trees that this should have led him to inspect the goods carefully.

423. Mr Trees’ answer was that he did not need to inspect the goods because clearly his customer would not pay for them if they did not exist and meet specification. It was suggested to him that he didn’t inspect them because he knew his deals were connected to fraud: he denied this. For all the reasons we have already given about finding CCA knew of the connection to fraud, we did not find the denial convincing.

Reaction to raid

424. On 1 June 2006, Appleco’s (and therefore CCA’s) premises were raided by law enforcement officers wearing flak jackets and carrying a search warrant. Mr Trees’ evidence was that the search was cordial: at his request the officers removed their flak jackets, and Mr Trees was more than happy for them to search the offices and take whatever they wanted to take, which they did.

425. Mr Trees was not told the purpose of the raid. He did not seem concerned by it; he said he did not discuss it with his suppliers or customers; he did not increase his due diligence because of it. He carried on trading as before other than that a week or so later he did try to cancel a meeting arranged with Mr D’Rozario as most of CCA’s papers had been taken.

426. What we take from this is that it is quite clear that Mr Trees really did understand that MTIC fraud was a serious risk in mobile phone trading but that he was content to carry on trading in the manner in which he had always done despite this. He was clearly confident that either there was no fraud in his supply chain or that HMRC would not discover it. For the reasons given above, we think that it was the latter because we think he did know of the connection to fraud.

Factors which indicate no knowledge?

427. Having dealt with the factors which HMRC said indicated knowledge by CCA and Mr Trees, we move on to the appellant’s case was that there were many factors which indicated that Mr Trees did not know of any connection to fraud.

No third party payments

428. There was no suggestion by HMRC nor any evidence that any profits made by CCA were paid out to third parties. Such payments would potentially indicate actual knowledge of fraud, as they might indicate the fraudster controlled the broker (see §13). But the lack of any such payments was neutral: it supported neither party’s case. This is because, as explained above, the fraud works even if the broker is entirely independent of the fraudster. There is no obvious reason why an independent broker would make third party payments; but an independent broker may or may not know its transactions are connected to fraud.

CCA made a loss on some deals

429. It was accepted by HMRC that the appellant made a loss on 2 deals in 2003 with a company called YM International. It appears Mr Trees put the wrong figure on the documents and its customer would not let it correct this. The appellant’s case seemed to be that this indicated that its trading was on the open market and it was liable to make losses.

430. We do not agree: so far as the trading at issue in this appeal is concerned, it is clear that CCA was not trading on the open market and it never made losses. We make no findings as to what happened in 2003: it seems to us that YM International may have acted in this fashion whether or not the trade was genuine or orchestrated.

Prior trading history

431. The appellant’s case was that CCA used the same business model from 2003 onwards; HMRC had repaid its VAT reclaims, sometimes after an extended verification; HMRC did not tell CCA to improve its contractual terms and deal documentation. CCA was therefore entitled to think in 2006, it said, that there was nothing wrong with its manner of trading; it had no reason to be suspicious of its deals in April -June 2006.

432. The appellant also points out that in 2003 it paid for independent accountants to undertake a line check on its supply chain because HMRC had refused to make a repayment to it: this followed the discovery of fraud in its supply chain in 09/03. The line checks found no fraud in the supply chain. The appellant relied on this, we understand, as proof of its bona fides and as reason for believing its deals were not connected to fraud.

433. We do not agree with the propositions in either of the above two paragraphs. Firstly, while HMRC did not seek to prove that CCA’s previous trading was connected to fraud, they did not concede that it was not. The appellant has not sought to prove that it was not connected to fraud. So, we leave open the question of whether CCA’s previous trading was connected to fraud.

434. Secondly, in so far as HMRC kept making repayments to the appellant right up until April 2006, all the appellant could reasonably read into that is that HMRC had not discovered any connection to fraud. In particular, if because of other factors, the appellant knew that its transactions were connected to fraud, HMRC’s decision to repay could only reassure CCA that HMRC had not discovered the connection.

Grey market in phones

435. Both parties accepted that at the time of these deals there was a real grey market in mobile phones in the UK. By ‘grey market’ the parties were referring to wholesale sales between persons who were not manufacturers nor authorised distributors of phones. The parties agreed certain facts about this grey market: it was a substantial and thriving market on which up to 12 million handsets might have been traded in that year. It could involve phones being imported into the UK, and on occasion re-exported from the UK; transactions could well be back to back.

436. The appellant’s case is that it thought it was trading on this genuine market and, although with the benefit of hindsight it can see that it was not, the fraudsters were careful to give the negotiations the appearance of genuine grey market trading.

437. We have already considered this. We do not accept Mr Trees’ evidence about the negotiations for the reasons given at §304-309. For that reason, and all our other findings set out above, we do not think, therefore, that CCA thought it was trading on the genuine grey market.

Worked in open plan/employees unaware of anything wrong

438. The appellant said that the unchallenged evidence of Mrs Ryan and Mr Gordon substantially undermined HMRC’s case. We understood this was said because their evidence was that they could hear Mr Trees on the phone all day. Moreover, said the appellant, Mrs Ryan was a respectable lady who considered Mr Trees to be a very respectable person who would never involve her in anything dubious.

439. Mrs Ryan’s and Mr Gordon’s opinion that Mr Trees would not involve CCA in anything dubious was opinion evidence; it was an opinion given by people who appeared to be loyal to their employer and, most significantly, without the benefit of the documentary evidence which we have seen which shows a highly sophisticated fraud in which all the deals were orchestrated and that CCA must have known that. Mrs Ryan and Mr Gordon were witnesses of fact; we are not obliged to accept their opinions and for this reason do not do so.

440. While we accept Mrs Ryan’s and Mr Gordon’s factual evidence, it does not take the appellant’s case anywhere. It is too general. We are prepared to accept Mr Trees spent a large part of each day on his phone and that what he said did not make his colleagues suspicious: but we do not know how much time he spent on the phone in respect of Appleco or CCA; we do not know what he said and what the other person on the phone was saying. This evidence is too vague to counter our findings at §§302-309 that there were no genuine negotiations on the CCA transactions.

Mr Trees put own money at risk

441. Mr Trees invested his own and his company’s money in CCA: the evidence was that Appleco had funded CCA with over £600,000 in interest free loans, funded partly by a mortgage over its business premises. Mr Trees has also re-mortgaged his own home in order to put CCA in funds. There was some suggestion that at least some of the investment had been repaid by the time of the denials in 2006, but we accept the appellant’s basic premise that Mr Trees had a substantial sum of his own money at risk in CCA at the time of the transactions at issue in this appeal.

442. The appellant’s case was that this was a strong indicator that Mr Trees knew nothing about the fraud: he would not have risked his own money in transactions which he knew were connected to fraud; he would not have risked losing everything if HMRC were to deny input tax recovery to CCA.

443. We do not consider this an indicator that Mr Trees lacked knowledge. What it indicates is that Mr Trees was confident that HMRC would repay the input tax. While a person might have had such confidence if they believed that there was no connection to fraud, we consider there were so many suspicious factors (such as the easy ability to realise enormous profits and the uncommercially benign trading environment) that even a person who did not know of connection to fraud would have been concerned. We think in Mr Trees’ case the confidence was engendered because, although he knew of the connection to fraud, he believed the risk to be low because either HMRC would not discover the connection to fraud or that they would repay CCA even if they did (as they had on occasions before).

444. We have already commented that Future, which undoubtedly knew of the connection to fraud, was content to risk its own money in the same manner as CCA: this is because by June 2006 CCA was effectively trading on money lent by Future. So the appellant is wrong to say no one would be prepared to take the risk. Future was, and so we find, was CCA. As we have said, risk should not be viewed with the benefit of hindsight.

Mr Trees’ relationship with Mr D’Rozario

445. In closing, the appellant said it relied strongly in its defence on the relationship between Mr Trees, Mrs Ryan and Mr D’Rozario. Its case was that Mr D’Rozario was a thorough officer who put CCA through rigorous scrutiny; Mr D’Rozario did not find any fraud at the time; the fraud was only discovered later when Mr Cunningham carried out extended verification. The appellant relied on this relationship to evidence a number of propositions:

(i) How could Mr Trees discover a fraud that Mr D’Rozario could not?

(ii) Mr Trees took a great deal of comfort, says the appellant, from Mr D’Rozario’s rigorous scrutiny that there was in fact no fraud.

(iii) Mr D’Rozario’s thoroughness would have meant Mr Trees would not have knowingly undertaken any transaction connected with fraud for fear of it being discovered.

(iv) Mr Trees’ clear desire (said the appellant) to assist Mr D’Rozario indicated that Mr Trees had no awareness of the fraud.

446. We will deal with these propositions in turn.

(i) Mr Trees no better placed to spot the fraud than HMRC

447. We do not accept this proposition. Mr Trees had access to a great deal of information that Mr D’Rozario did not. Mr Trees knew how the deals were put in place, he knew the funding arrangements, he knew about CCA’s benign trading environment. We have found he knew they were orchestrated transactions.

448. If Mr D’Rozario had known what Mr Trees knew, he would have known of the connection to fraud. Mr D’Rozario did not know this; moreover he was looking for fraud in CCA’s supply chain and not finding it, because he did not at that time understand contra-trading.

(ii) Mr Trees’ relied on Mr D’Rozario

449. Mr D’Rozario was cross-examined at length. Mr Pickup went through his logbook and the letters between him and Mr Trees. It was his case that the relationship was very cordial and Mr Trees’ relied on Mr D’Rozario’s opinions.

450. We do not read the correspondence in quite the same way. We agree that Mrs Ryan had a cordial relationship with Mr D’Rozario and we agree that on the whole Mr Trees had a professional and cordial relationship with Mr D’Rozario most of the time. Annual half-day compliance meetings took place which Mr D’Rozario and Mr Trees attended at CCA’s offices; letters were exchanged.

451. But we do not accept all of Mr Trees’ evidence on this. Mr D’Rozario did not accept Mr Trees’ evidence he dropped in unannounced at CCA’s offices, apart from once in December 2006). We rely on his logs which record no such visits but which are otherwise quite detailed.

452. Mr D’Rozario also did not agree that he told CCA how to conduct its business and in particular he did not agree that in 2003 he had told Mr Trees to use CCA for the mobile phone business which at that point Mr Trees put through Appleco (see §134). To some extent, we consider the difference in evidence on this was a difference in perception. Mr D’Rozario accepted he did make suggestions to Mr Trees (eg over using a supplier declaration form). Mr Trees might well have seen such suggestions as being told how to conduct his business.

453. But we do not accept that Mr Trees relied on Mr D’Rozario’s opinions. Having the benefit of the contemporary letters between the parties and Mr D’Rozario’s call log, we conclude that the relationship was not as close as Mr Trees liked to paint it. It was a professional relationship, which was normally polite and cordial, but at times Mr Trees was testy and confrontational. Mr Trees was also clearly not prepared to provide Mr D’Rozario with the IMEI numbers despite numerous requests.

454. In any event, while Mr Trees may well have taken comfort from Mr D’Rozario’s continued repayment of CCA’s input tax claims, that does not tell us whether or not Mr Trees knew of the connection to fraud. If he knew of the connection to fraud, which we find he did, the comfort it offered was that HMRC did not, and would continue to make repayments.

(iii) Mr Trees would not have risked discovery by Mr D’Rozario

455. In closing, Mr Pickup suggested that Mr D’Rozario was such a thorough VAT officer who, for instance, insisted on seeing Mr Trees’ parking ticket to verify that he had visited Infinity’s offices, that Mr Trees would never have risked entering into transactions which he knew were connected to fraud. It was only too likely Mr D’Rozario would spot the fraud.

456. We do not agree that Mr Trees considered Mr D’Rozario as likely to spot fraud. Mr D’Rozario may well have paid attention to the details but he did not spot the fraud: the connection to fraud only became clear to HMRC after extended verification involving the contras’ other deal chains. Moreover, we did not see any evidence that Mr Trees respected Mr D’Rozario’s abilities to spot a fraud: Mr Trees would not assist him by providing IMEI numbers and Mr Trees seriously doubted Mr D’Rozario’s statement in mid-May 2006 that there were issues with phones not existing in the quantities in which they were traded. Mr Trees says he still does not accept that the Samsung Serenes and P990s did not exist.

457. In reality, this is the same case on risk that the appellant has made and which we have already rejected: see §§51-54.

(iv) CCA’s cooperation with HMRC

458. We do accept that to an extent CCA cooperated with HMRC and Mr D’Rozario. Having said that, it was obviously in CCA’s interests to provide its deal documentation to HMRC as without it, HMRC would not pay it the claimed VAT refund. It is no surprise, and no evidence of anything other than self-interest, that CCA did so.

459. The appellant’s case is that its cooperation went further. When its suppliers did not cooperate with HMRC and provide their deal documentation, Mr Trees would use his influence with them to obtain the cooperation. Again, we read nothing into this: without its suppliers’ deal documentation, HMRC would not pay CCA’s VAT refund. It was in its own interests to obtain its suppliers’ cooperation.

460. Its case is that its cooperation went still further. On a few occasions, Mr Trees gave names of entities to HMRC that he said he considered suspicious (§364). We read nothing much into this: whether CCA was innocent of knowledge of the fraud or not, it lost nothing, but stood to gain HMRC’s confidence, by reporting to HMRC a few traders with which it no longer traded or had never traded.

461. It also said its cooperation with HMRC was disinterested when, at Mr D’Rozario’s request, it required its freight forwarder to specify the type of goods (ie mobile phones) carried on the CMRs. But we read nothing into this: having a correct CMR was essential for CCA’s VAT reclaim. A correct CMR was a requirement for zero rating. So, whether innocent of knowledge of the fraud or not, it was in CCA’s interests to ensure its CMRs met HMRC’s requirements.

462. It also suggested its readiness to carry out and improve its due diligence was an indication of its disinterested cooperation with HMRC. We do not agree; putting aside the question of the quality of its due diligence, the normal purpose of any due diligence is to protect the trader from bad debts and cancelled transactions. Nevertheless, it is our opinion that in this case that CCA undertook due diligence to satisfy HMRC: again, its willingness to undertake due diligence does not indicate innocence, as whether or not CCA knew of the connection to fraud, it would be likely to undertake due diligence to increase its chances of VAT repayment by HMRC. We dealt with its due diligence in more detail above.

463. Some mention was made of the fact that Mr Trees’ considered the search warrant shown to him by the officers carrying out the raid as invalid as it contained (he said) the wrong address. Nevertheless, he cooperated with the search and was friendly with the officers. Again we read nothing into this: we do not think it would have been in CCA’s interests, whether they knew of the fraud or not, to appear to be difficult and there was no suggestion that the papers removed (deal documentation and due diligence) contained anything CCA was not happy to show to HMRC.

464. So, while there was at least some cooperation by CCA with HMRC, in very large part it was in CCA’s interests to undertake that cooperation in order to obtain its VAT refunds. Its cooperation does not therefore indicate a lack of knowledge of connection to fraud.

465. And in one area it did not cooperate with HMRC and that was with respect to IMEI numbers. We have already dealt with this. The conclusion we draw is that CCA cooperated with HMRC to the extent that it needed to do so to attempt to ensure it got a VAT refund, and sometimes in some minor matters, it went further. But it did nothing that might have provided real assistance to HMRC in identifying the fraud and in particular it did not provide IEMI numbers to HMRC.

466. In conclusion, we reject the appellant’s case that the cooperation CCA offered to HMRC indicated that it had a clear desire to assist HMRC combat fraud and was not aware of the connection of its own deals to fraud.

Lack of cooperation by the contras

467. The appellant contrasted its own cooperation with HMRC with that of the contra-traders, Future, Soul and Infinity. We have already said that in very large part CCA’s cooperation with HMRC was self-interested. But we agree with the appellant that it exceeded any cooperation offered by Future, Soul or Infinity (see §§219, 222, 231).

468. Mr Pickup’s point is that if the contra-traders (as part of the fraud) wanted CCA to get its refund then they needed to cooperate with HMRC but they did not. We do not entirely agree: we think the contras would have had opposing concerns: on the one hand, they would not want HMRC to know too much about their own VAT affairs as HMRC might realise their role in concealing the connection between the brokers’ (including CCA’s) transactions in the clean chain and the defaults in the dirty chains; on the other hand, they would want to support the brokers’ input tax reclaim (as Mr Pickup said and for the same reasons we have given at §21 for why the person organising the fraud would wish the broker’s VAT claim to succeed). And we note, consistently with this dilemma, that while in general the contra-traders offered little or no cooperation with HMRC, the evidence was that they did provide the deal documentation to HMRC whenever CCA requested them to do so in order to assist with its VAT reclaim.

469. We can therefore see nothing in the contrast between CCA’s cooperation and the contra’s lack of cooperation to indicate a lack of knowledge of fraud by CCA. If anything, it is neutral.

Conclusion on knowledge

470. Mr Pickup suggested that HMRC’s case amounted to no more than saying that because CCA was broker in an MTIC fraud, it must have known that. We agree, as we have said, that that is not the case: we have to consider the facts of the particular case to decide if HMRC have proved on the balance of probability that Mr Trees knew at the time that CCA’s transactions were connected to fraud.

471. It should be clear from what we have said above that we do find that it is proved. There were a number of factors even by themselves showed on the balance of probabilities that Mr Trees did know of the fraud: all of them combined put the matter beyond any doubt.

472. As explained in more detail, a major plank of the appellant’s case was that the fraudster would want to keep a broker innocent of knowledge of the fraud and therefore would have carefully manipulated the broker, such as causing the entities on either side of it to enter into negotiations with it to preserve the illusion to the broker that it was a free agent undertaking transactions in the grey market.

473. We rejected this case: there was no convincing reason why the fraudster would do this. We note, for instance, that a knowing party to the fraud (Future) was prepared to take the risk of non-repayment to CCA, so it could not be assumed CCA would not have been prepared to take what was probably seen as quite a small risk at the time (if not with the benefit of hindsight). Moreover, we agree with HMRC that having an innocent broker would pose a real risk for a fraudster as an innocent broker might become suspicious and assist HMRC. In any event, while we do not want to decide this case on what the fraudster might be supposed to prefer, we note that it would be quite illogical for a fraudster who wanted to keep a broker innocent in its 39 broker transactions, to risk the broker getting more suspicious by involving it quite unnecessarily in 117 buffer transactions. If we had to decide the case on this basis, we would say on the balance of probability it is more likely that CCA knew of the connection to fraud.

474. But we do not decide it on that basis. Putting all of that out of our mind, we look at the facts. We have found that CCA did know that its deals were dictated to it (§§304-309) and that its banking was dictated to it (see §299-303 and §§310-324): knowing any one of these matters meant Mr Trees could not have been in any doubt that CCA was participating in transactions orchestrated for the purpose of VAT fraud. He knew CCA was not trading as a free agent in any grey market.

475. But even if we were to put those findings of fact aside, we would still conclude that Mr Trees knew that CCA’s transactions were connected to fraud. CCA was offered deals far too good to be true: it was offered the opportunity to make phenomenal profits over a short space of time for doing virtually nothing; the deals required no special skill nor utilisation of any carefully cultivated contacts, they involved few costs and no commercial risk. CCA knew that its customers had little interest in the product traded, it knew about the patterns in the trading that had no rational explanation, it knew that its trading environment was uncommercially benign (see §§269-298). The only explanation for all of this was clearly that the transactions were orchestrated for the purpose of fraud; that conclusion is all the more obvious when Mr Trees was well aware that there was VAT fraud taking place in mobile phone trading. From this we would conclude that it was very clear that Mr Trees knew that CCA’s transactions were connected to fraud.

476. But it is even more certain that he did: he acted as a person who did know of the connection; CCA’s deal documentation was inadequate reflecting that Mr Trees knew that it did not really matter; similarly CCA did not inspect the goods; CCA’s due diligence was inadequate and undertaken to satisfy HMRC; negative indicators were ignored; CCA held inadequate insurance against some risks and none against others despite the high value of the goods; CCA was not willing to cooperate with HMRC over IMEI numbers where such cooperation might have helped reveal circulation of the phones and thus the fraud. Mr Trees had a relationship of trust with Future which was clearly involved in the fraud and this relationship continued long after the suspicions of an innocent trader would have been raised.

477. None of factors put forward by appellant as indicating Mr Trees did not know actually countered or explained away any of these findings. CCA’s cooperation with HMRC was limited and self-interested; Mr Trees did not rely on Mr D’Rozario’s opinions; HMRC was definitely not better placed to spot the fraud than Mr Trees; while Mr Trees did choose to put his own money at risk, this was at least as consistent with confidence that HMRC would not discover any fraud, as with any innocence; CCA’s prior trading could only have given CCA confidence that HMRC would continue to repay it its input tax claims; its employees were not aware of anything untoward but there was no reason why they would know as it was Mr Trees who exclusively conducted the deals and in any event their evidence was so general it does not exonerate Mr Trees.

478. We are in no doubt that Mr Trees knew that all of CCA’s transactions in the period in question were connected to fraud. As we have found that they were so connected, CCA’s claim for input tax on the transactions was correctly denied by HMRC on the basis of Kittel. The appeal is DISMISSED.

means of knowledge

479. HMRC’s alternative case was that if CCA (acting via Mr Trees) did not know of the connection to fraud, then it ought to have known of it. We do not need to deal with this allegation in view of the fact that we have found that Mr Trees clearly did know of the connection to fraud, but for the sake of completeness we will deal with it briefly.

480. We are satisfied that, had Mr Trees not known of the fraud, he ought to have known of it. In summary, there are so many factors that ought to have made Mr Trees deeply suspicious, that when considered in combination, a person would realise that the only reasonable explanation for them was that the transactions were all orchestrated for the purpose of fraud.

481. As we said above, we have found that CCA did know that its deals were dictated to it and that its banking was dictated to it. But even if we were to put those findings of fact aside, we would still conclude from the rest of the matters that Mr Trees did know that CCA ought to have known its transactions were connected to fraud. Mr Trees knew that CCA was offered deals far too good to be true: it was offered the opportunity to make phenomenal profits over a short space of time for doing virtually nothing; the deals required no special skill nor utilisation of any carefully cultivated contacts, they involved few costs and no commercial risk. CCA knew that its customers had little interest in the product traded, it knew about the patterns in the trading that had no rational explanation, it knew that its trading environment was inexplicably benign. The only explanation for all of this was clearly that the transactions were orchestrated for the purpose of fraud; that conclusion is all the more obvious when Mr Trees was well aware that there was VAT fraud taking place in some mobile phone trading. From this we would conclude that it was very clear that Mr Trees ought to have known that CCA’s transactions were connected to fraud.

482. CCA’s appeal would have been dismissed on this ground if we had not already dismissed on basis Mr Trees did know of the connection to fraud.

Right to apply for permission to appeal

483. 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.

Barbara Mosedale

TRIBUNAL JUDGE

Release date: 14 May 2020

table of terms

A1	A1 Logistics and Freight Ltd
Allimpex	Allimpex Handels GmbH
AS Genstar	AS Genstar Ltd
Bartonole	Bartonole Trading Ltd
Booming	Booming Technologies Ltd
C&B	C&B Trading Ltd
ET Global	ET Global Solutions Ltd
Eutex	Eutex Ltd
FCIB	First Curacao International Bank
Future	Future Communications (UK) Ltd
Infinity	Infinity Holdings Ltd
Distribution	Infinity Distribution Ltd
JDI	JDI Trading Ltd
Kep	Kep 2004 Ltd
Mohomedbhay	Shabir Mohomedbhay or BHS Vertriebs GmbH
Okeda	Okeda Ltd
Pielkenrood	Pielkenrood Opto Electronics BV
RS Sales	Rafik Sodawala trading as RS Sales Agency Ltd
Soul	Soul Communications Ltd
Sound & Secure	Sound and Secure Ltd
UK Communications	UK Communications Ltd
Universal Trade	Universal Trade Supplies Ltd
Wade	Wade Tech Ltd
Wizard	Wizard Promotions

	[Home] [Databases] [World Law] [Multidatabase Search] [Help] [Feedback]
	First-tier Tribunal (Tax)
You are here: BAILII >> Databases >> First-tier Tribunal (Tax) >> CCA Distribution Ltd v Revenue & Customs (VAT - MTIC fraud) [2020] UKFTT 222 (TC) (14 May 2020) URL: http://www.bailii.org/uk/cases/UKFTT/TC/2020/TC07708.html Cite as: [2020] UKFTT 222 (TC)

First-tier Tribunal (Tax)