Terms and expressions and description of MTIC

Law

10.  Despite four weeks of hearing and voluminous paperwork, the question which this Tribunal is called upon to decide is short.  The European Court of Justice (“ECJ”) ruled in Axel Kittel v Etat Belge (C-439/04) and Etat Belge v Recolta Recyling SPRL (C-440/04) in July 2006 that (paragraph 61):

“where it is ascertained, having regard to objective factors, that the supply is to a taxable person who knew or should have known that, by his purchase, he was participating in a transaction connected with fraudulent evasion of VAT, it is for the national court to refuse that taxable person entitlement to the right to deduct.”

Is Kittel part of English law?

11.  Mr Beal’s point is that the test enunciated by the ECJ in Kittel has not been expressly enacted into English legislation.  The Kittel test goes further than the common law doctrine of fraud:  in making a taxpayer liable for what they “ought to have known” it is introducing liability for negligence and carelessness whereas for fraud a person could only be liable for what they actually knew (or recklessly chose to ignore).

12.  Mr Beal accepts that if Pars can be shown to have actually known of a connection to fraud, then that is fraud and Pars would as a matter of English law of fraud lose its entitlement to input tax recovery.  But if Pars is shown merely to have been careless then the law of fraud does not apply, and, in Mr Beal’s view, Kittel does not apply either as it has not been enacted into English law.

Connection to fraud

Credibility of witnesses

79.  We consider the credibility of Pars’ witnesses in the second section of this decision notice where we consider what was Pars’ knowledge and/or means of knowledge.  This first section of the decision notice deals with the evidence of the deal chains and the alleged frauds by defaulting traders:  this evidence was entirely adduced by HMRC (with the exception of Pars’ own invoices, purchase orders and CPU spreadsheet) and so here we deal with the credibility of HMRC’s witnesses.

80.  A number of witnesses gave evidence for HMRC to the Tribunal.  The main witness was Mr Churchouse who was the officer assigned to Pars shortly after the deals the subject of this appeal took place.  He was therefore the officer who took the decision to deny Pars the input tax the subject of this appeal.

81.  The Appellant challenged his credibility, in particular expressing the view that Mr Churchouse was biased against Pars.  It is indeed the case that Mr Churchouse expressed his view in both his witness statement and at the hearing that Pars by its directors and officers was knowingly involved in a fraud.

82.  The Tribunal is not interested in the opinions of witnesses of fact save to the extent that we consider they might have coloured the evidence of fact which they give to the Tribunal.  Mr Churchouse’s evidence was largely documentary:  he produced the deal packs and the missing trader packs and also the logs of the meetings between HMRC officers and Pars officers before and during the deals in this appeal.  (The Tribunal would have preferred to hear from the HMRC officers themselves).  We do not think that the view he had formed, which he openly admitted, that Pars were fraudulent, coloured this evidence. 

83.  We agree with Mr Beal that the opinions he drew from the evidence tended to be unfavourable to Pars, but we do not base our conclusions about Pars on his opinions.  We also agree that in a number of instances Mr Churchouse made mistakes (most significantly in believing that Pars did not insure air freighted goods) but the Tribunal found that he was prepared to accept that he had made mistakes where this was shown to be the case.  We also found the mistakes he made (such as believing sales at less then market price indicated fraud and that the airway bills indicated a lack of insurance) arose from ignorance of matters such as how a grey market would operate and how freight insurance was organised, rather than any intention to be misleading.

84.  In all, Mr Churchouse was subjected to fairly sustained criticism by Mr Beal but, in the round, we found Mr Churchouse to be a good witness of fact (although we ignore his opinions).  Much of Mr Beal’s criticisms of Mr Churchouse seemed to be that his decision to deny Pars its input tax was based on incorrect information and not reached (Mr Beal alleged) in a fair and balanced way.  However, even if true (and we make no comment on this) this is not relevant to this Tribunal:  we are not reviewing Mr Churchouse’s decision.  We have original jurisdiction and will decide whether or not on the evidence in front of this Tribunal Pars are entitled to recover the input tax at stake.

85.  Mr Beal also suggested to Mr Churchouse (which was denied) that HMRC had a general policy to deny wholesale exporters in CPUs their input tax:  again the Tribunal is not here to judicially review HMRC’s actions (it does not have the jurisdiction) but to decide whether as a matter of law Pars is entitled to recover its input tax.  So we make no comment on this allegation.

86.  A number of other officers gave evidence in relation to the alleged defaulting traders at the top of the alleged deal chains.  We mention them by name when considering the defaulters:  Mr Beal did not challenge their credibility and we accepted their evidence as set out in our conclusions on the defaulters.

87.  There were two witnesses for HMRC (Ms Carter and Mr Stone) for whom Mr Beal had no questions and so they were not called.  We have considered the undisputed evidence in their witness statements and mention it where we have found it relevant. 

Connection

88.  It was HMRC’s case that Pars’ 59 purchases (and sales) of CPUs which are the subject of this appeal were connected to fraud.  They sought to show connection by showing that the CPUs purchased by Pars were purchases at the end of chain that traced back to a fraudulent defaulting trader.  So the “connection” element was a trail of invoices and/or purchase orders relating to the same CPUs traded from seller to buyer down a line to Pars and then to Pars’ customer.

89.  We find that if HMRC can show that the CPUs purchased by Pars were previously bought and sold up a chain of transactions then this is sufficient “connection” for the purpose of the test as set out by the ECJ in Kittel. 

90.  Pars did not accept that its purchases were so connected and so we set out below our findings on the evidence and to which sales we have found Pars’ purchases connected.  For ease of reference for us as much as the parties reading this we have set out a schedule showing the connections we have found proved on the evidence and this is attached as an appendix to this decision notice.

91.  There is little dispute between HMRC and Pars on the question of from whom Pars bought the CPUs.  The evidence put by both sides appears to coincide in most cases:  HMRC produced their deal packs showing a chain of invoices and purchase going back (in most cases) to a company they allege is an importer.  Pars’ evidence is its contemporaneous CPU spreadsheet in which they recorded all their wholesale CPU trades showing their seller and their customer, quantities bought and sold and often the identification numbers on the wholesale boxes of CPUs.  Where there was a difference between HMRC’s evidence and Pars, we preferred to rely on Pars’ spreadsheet as it was contemporaneous and we have no reason to suppose anything other than entirely accurate. 

92.  Indeed in a number of cases we found that this spreadsheet gave useful detailed evidence of Pars’ purchases and sales.  As an example, the spreadsheet makes it clear that the source of Pars’ sale of 4,500 retail CPUs to All Trading was 3,500 from deal 46 and 1,000 from deal 47.  This was not apparent from HMRC’s deal packs presented in evidence by them and could only have been inferred as a logical solution if it were not for the spreadsheet showing that that indeed was what had happened.

93.  Pars’ case is that Pars has no knowledge of its suppliers’ supply chain and can only challenge the evidence which HMRC have produced.  We have accepted HMRC’s evidence where the invoice and/or purchase order chain relates to the same description of goods in sufficient quantities at the “right” price and on the same date as more likely than not (in the absence of any other evidence) to be the chain from which Pars’ seller obtained the goods sold to Pars.

94.  Mr Beal questioned whether we could make that inference in the absence of release notes actually identifying that it was the same goods being passed down the chain of invoices.  However this Tribunal deals in probabilities and not certainties:  we find where the invoice/purchase order chain is on the right date at the right price and the same specification goods in the right quantity then more likely than not it is the correct chain of goods (unless there was evidence to the contrary.) 

95.  Invoice/purchase order:  In most cases the connection in the chain is supported by both an invoice and purchase order.  Sometimes there is only an invoice and in a very few cases only a purchase order.  Mr Beal did not suggest that the evidence of connection was insufficient if only an invoice or purchase order were produced in evidence and we do not find it to be so.

96.  As mentioned in paragraphs 107-109 below in a very few cases we have found a connection proved where there is no invoice or purchase order.  Where this is the case we explain our reasoning in the paragraphs which deals with the particular alleged chain.

97.  The quantity:  In many of the chains the quantity purchased by Pars (and sometimes by its immediate predecessor or predecessors in the chain) is less than the quantity traded by the companies higher in the chain.  It was not suggested to us that this meant that the transactions were not connected:  it is a feature of normal trading that a vendor might split a purchase between a number of buyers.  It is HMRC’s case of course that this is not normal trading but MTIC fraud and we find the splitting of consignments is also likely to be a feature of MTIC fraud.  This is because the fraudulent importer buys (or purports to buy) very large quantities (to maximise the VAT take).  These are likely then to be split up and despatched or exported by the brokers in smaller quantities because the broker can only buy in quantities for which it has the capital to fund the VAT for which he will expect to be out of pocket until HMRC refunds (whether or not the broker knows or has the means of knowing of the fraud).

98.  And we have found that split deals have been a particular feature of the alleged chains in this appeal.  A glance at the spreadsheet shows that not only are quantities of CPU split into smaller quantities (eg for chains 4,12/13, 23/26, 31/33)  but also loads have been split and recombined (eg 17/20, 42, 44, 54,65).  Where a load has been split we have considered the evidence as a whole to decide whether or not the connection is proved as we explain on a deal by deal basis below.

99.  the price:  It is the Appellant’s case that HMRC is put to strict proof of the chains.  If HMRC are right and these chains are MTIC fraud chains then in our view it is inherently unlikely that any sale will show a loss.  This is because if it is MTIC fraud the deals will be pre-arranged and no one would deliberately enter into a purchase combined with a sale where it was certain to realise a loss.  Where the alleged chain shows a loss, this is likely to indicate that HMRC have not identified the correct chain of sales and purchases or, if they have, that the chain is not fraudulent. 

100.         We refer to “usual” or “normal” price increments and by this we mean that there is a small price increase at each step in the transaction.  As explained in paragraphs 447-451  the profits made by the buffers are very small.

101.         the date:  In most cases the deals were back to back in that they took place on the same day.  This means that all the purchase orders and invoices in the entire chain and evidence of despatch or export by Pars would bear the same date with the exception of Pars’ invoice which was often dated the day or so later. Pars explained that their invoice might be dated later than the date of the customer’s purchase order and the date of shipment because paperwork in the accounts department would get behind and/or they did not issue the invoice until paid which could be delayed until the goods were received and checked by their buyer abroad.  Nevertheless, in many cases it is apparent the deals were back to back because Pars’ shipped on the same day as the date of all the documents apart from their own invoice. Where the evidence is that the chain of transactions did not take place on the same day we specify this below and explain whether or not because of this we consider that on the balance of probabilities it is the right chain.

102.         The description.  We found that not all sales invoices and purchase orders describe the goods in the same way and some of the invoices and purchase orders carried very little detail.  Many of the transactions related to SL7Z9 processors and we adopt this as shorthand for the full description of the CPUs, which would, we find,  not have been used on all the invoices in the alleged chain in any event.  Where we have found that the alleged chain does not appear to relate to the same description of goods we specify this with our reasons for either accepting or rejecting HMRC’s alleged chain as proving connection.

103.         Time’s supplier:  In addition to the above evidence, we have found that even without any documentation such as invoices or purchase orders, it is more likely than not that up to 7 May any purchase made by Time would have been from KEP and that any purchase after that date would have been from Okeda.

104.         Our grounds for coming to this conclusion are that with exceptions mentioned below, the documentary evidence which we have accepted for the chains evidenced in this appeal (not all of which we have found to be connected to Pars) in which Time makes a purchase is that Time invariably purchases from KEP with its last documented purchase being on 27 April.  On 8 May Time purchases from Okeda and thereafter all its documented purchases in this appeal (which evidence we accept) are from Okeda.  The exceptions mentioned above are chains 36, 45 (M4), 47 and 50 for which there is no documentary evidence from whom Time made its purchase.

105.         We have also taken into account Mr Armstrong’s evidence (which we accept) which was that KEP was visited on 9 May following enquiries made by HMRC the previous day.  This visit led to its immediate deregistration for suspected involvement in MTIC fraud.  This has led us to conclude that more likely than not, bearing in mind the evidence of active involvement in fraud of KEP (or the person purporting to be KEP) and Time, it was HMRC’s interest in KEP and its imminent deregistration on 8 May that led Time to switch suppliers.  We have therefore concluded that Time’s purchases in deal chains 36, 45 (M4), 47 and 50 were from KEP as the last of these took place on 5 May.

106.         We have also taken into account (but without giving it almost any weight) evidence that Mr Jhaj, Time’s director gave to HMRC that at this period up to including early May Time dealt only with KEP.  We agree with Mr Beal that the evidence was only hearsay and Mr Jhaj is very likely to be an unreliable witness.  Nevertheless for the reasons given in the above few paragraphs we think the evidence to be truthful in this regard. 

107.         Apart from the evidence in relation to Time and KEP, we also find a connection proved without documentary evidence in relation to chain 9/10 and we explain our reasoning under that heading below.

108.         In Deal 29A as explained below there is no purchase order or invoice for the connection we find proved between Connect and Apollo:  we explain the evidence – a supplier declaration - in that section which led us to find the connection proved.

109.         In Deal 42 alleged chain M1 there is no purchase order or invoice for the connection we find to be proved between Ability and Apollo.  We find the connection proved on the basis of a third party payment instruction as explained below where we deal with this chain.

110.         Deal numbering:  We have chosen to identify the 59 purchases made by Pars on which HMRC have refused to refund the VAT by Pars’ purchase order number to try to avoid the confusion by the different numbering systems adopted by the Appellant and HMRC in the hearing, and the HMRC’s habit of treating as more than one deal a single purchase by Pars which was then split. 

111.         Nevertheless, we have referred to the alleged deal chains using HMRC’s reference numbers as well to help make it clear to the parties to which evidence we are referring when deciding whether or not a purchase made by Pars was connected to the chain.

Deal 1 Pars’ Purchase Order number 32737

112.         The documents in this chain are all dated 3 April and the goods are shipped on the same date.  They all relate to SL7Z9 CPUs. Twelve boxes were traded down to Connect, who bought the twelve but only sold 10 of them to Futuristic.  Bearing in mind the identity of the product and dates and that the chain involves normal price increments, we find it is more likely than not that this is the correct chain.  Pars sold and exported the goods to Mashall.

113.         In conclusion, the Tribunal finds that this purchase of CPUs by Pars was connected by the chain of sales and purchases of CPUs which traced back to FX Drona as alleged by HMRC as chain A1 and as set out in the Appendix. 

Deal 2/5/6/8  Pars Purchase Order number 32763

114.         This was a single purchase by Pars from Leisure.  Pars then split the consignment of 16 boxes of CPUs and despatched/exported them to GIG in Switzerland (3 boxes), Alvegar in Spain (2 boxes), Emerald in Spain (1 box) and Inco Pars in the Netherlands (10 boxes).

115.         All the transactions and the export/despatch of the goods took place on 5 April and related to SL7Z9 CPUs.  The chain of documents shows usual price increments.  The chain involved 23 boxes down to Leisure who bought the 23 boxes but sold only 16 to Pars.  Bearing in mind the identity of date and produce, and the price increments, and that splits are not unusual in these chains, we find it to be more likely than not to be the correct chain of supply of the goods bought by Pars.

116.         The Tribunal finds that this purchase of CPUs by Pars was connected by a chain of sales and purchases of CPUs which traced back to FX Drona as set out in the Appendix and as alleged by HMRC as deal chains A2,5,6, & 8.

Deal 3 Pars Purchase Order number 32739

117.         All the documents were dated 3 April with the normal exception of Pars’ invoice.  The goods were shipped on 3 April.  All the documents relate to 800 retail SL7Z9 CPUs.  The chain shows the normal price increments.

118.         The Tribunal finds that this purchase of CPUs by Pars was connected by a chain which traced back to Samson as set out in the Appendix and as alleged by HMRC as chain A3.  Pars sold and despatched the goods to All Trading. 

Deal 4 Pars Purchase Order number 32773

119.         Pars bought 5 boxes of SL7Z9s from Futuristic but Futuristic bought 2 lots of 5 boxes of Sl7Z9 CPUs on the same day at the same price, and we cannot therefore know which of the 2 lots were sold on to Pars (or if it was a mixture of the 2 lots).  This does not matter, as we find that the supplier to Futuristic of both lots was Connect, who purchased 10 boxes from Innovate, who purchased 10 boxes from Ability.  Pars sold and exported the goods to Mashall.

120.         The chain shows the usual price increments but the invoices and purchase orders are not all back to back.  Most of the documents bear the same date of 6 April including Pars’ own purchase order and Mashall’s purchase order.  However, Futuristic’s invoice to Pars is dated 12 April and we find that it only purchased the goods on 12 April from Connect.  Mr Hamakarim explained that having placed the order with Futuristic, Futuristic let them down and were only able to supply the requested goods 6 days later.  Pars communicated this to Mashall who were happy to wait.  The invoice trail shows that the hiccup occurred with Connect who bought 10 boxes on 6 April but did not sell until 12 April.  The deal went through for the same quantity and at same price but 6 days later than the purchase orders from Mashall and Pars. The goods were exported on 12 April.

121.         On the basis of this evidence, in particular that the deal was intended to take place on 6 April but was for an unexplained reason held up by Futuristic until 12 April, we find that Pars’ purchase from Futuristic was connected by the alleged chain A4 which traced back to Ability and was as set out in the Appendix. 

Deal 7&11 Pars Purchase Order number 32776

122.         We find from the invoices and Pars’ spreadsheet that Pars purchased 12 boxes of SL7Z9s from Leisure on 6 April.  HMRC gave evidence of two chains (A7a and A7b) ending in Leisure either of which they considered might have been the source of the 12 boxes.  14 boxes of SL7Z9s originate with KEP in a deal chain taking place on 6 April. 15 boxes of SL7Z9s originate in a chain from FX Drona on 3 April but are sold to Pro Force on 6 April.  Both chains are at similar prices. 

123.         We agree that on the face of it either chain could be the source of the CPUs sold by Leisure to Pars.  Leisure’s purchase order numbers do not tie in with its invoice numbers (there is no reason why they should) so they are no help in resolving the issue.  However, Leisure’s release note to the freight forwarder and in favour of Pars is exhibited giving the box types and box numbers.  Pro Force’s release note to the same freight forwarder and in favour of Leisure is also exhibited.  Pro Force was the line 3 buffer in the FX Drona chain. This has been annotated in hand to show 12 boxes of “60234A” and 3 boxes of “60233A”.  The note is more likely than not to have been contemporaneous as the box type could only be verified when the boxes were physically present. These type numbers are not the same as on the Leisure release note, which shows 5 types covering 12 boxes and none of them being identical to these numbers.

124.         On the basis of this the Tribunal finds it more likely than not that the 12 boxes purchased by Pars from Leisure originated with the KEP chain, and that the boxes the subject of the FX Drona chain were not the boxes sold to Pars. 

125.         We therefore find on the balance of probabilities that this purchase of CPUs by Pars was connected by a chain which traced back to KEP as set out in the Appendix.  Pars purchased the 12 boxes and then split them up and despatched 7 boxes to Inco Parts in the Netherlands.  One box was sold to Made on Trade and 4 boxes sold to Datec.  Datec, unlike the other companies, was based in the UK, and this sale did not involve a despatch of goods outside the UK.  On the contrary Pars charged and accounted for VAT on this sale. We deal with this in paragraphs 493-507 below.  All the transactions (including the despatch) took place on 6 April and related to  SL7Z9  CPUs. 

Deal 9 & 10 Pars Purchase order number 32776

126.         Although shown by HMRC as two chains, A9 and A10, it was in fact a single purchase by Pars which Pars then split into two.  The documents in this alleged chain are all dated 5 April except, (as normal) Pars’ invoices which are dated 7 April but  additionally both CMRs and the purchase order from Dutch Team are also dated 6 April. 

127.         Pars bought 5,000 retail SL7Z9s and sold and despatched 3,000 of them to Inco Parts and 2,000 of them to Dutch Team.  HMRC led evidence (their chain A9) that the chain could be traced back beyond Ultimate to KEP.  However we do not find that on the balance of probability that the documents they produced in evidence in fact relate to the same chain.  Firstly, as pointed out by Mr Beal, the chain from KEP through Time, Resolutions and down to a purchase order from Ultimate dated 5 April relates to SL7Z8 CPUs.  This might, of course, be an error in the paperwork particularly as the dates are the same and the price increment on what Ultimate pays (on a Z8) to what it makes (on a Z9) is 5p per unit whereas in fact other evidence in this case we find shows Z8s are normally priced significantly higher than Z9s.  Nevertheless, from evidence in other deal chains in this appeal, we find that Ultimate’s invoice number matches a reference number which it puts on its purchase orders.  In this alleged chain Ultimate’s invoice number is 1371 which does not match the reference on its purchase order of 1682.  We therefore conclude the difference in specification is not an error but that it shows that this is a separate chain.

128.         Nevertheless we do find that the chain on the balance of probabilities traces back to KEP.  This is because the evidence in this appeal is that of the 13 chains in which we have found Ultimate proven to be a participant it purchased from Resolutions in all of them (except that there is no documentary evidence for this chain 9/10).  It is more likely than not therefore that in chain 9/10 it also purchased from Resolutions (but not on the above invoice which we have found does not relate to this chain).  In the same 13 chains, Resolutions always purchased from Time, except of course that there is no documentary evidence for this particular chain.  Again we find therefore that on the balance of probabilities it purchased from Time in chain 9/10.  Lastly, as discussed in paragraphs 103-106 we find that Time always purchased from KEP up to 7 May.

129.         In conclusion, based on all this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from Leisure was connected by a chain of sales and purchases of CPUs back to KEP and as set out in the Appendix. 

Deal 12 & 13.  Pars Purchase Order number 32799

130.          This was a single purchase by Pars but HMRC show it as two: A12 and A13.  Pars ordered on purchase order number 32799 and were invoiced by Leisure for 4 boxes of CPUs.  However, Pars’ evidence, which we accept, is that 1 box was not delivered and they received a credit note for this.  We accept this evidence as it is consistent with Pars’ contemporaneous CPU spreadsheet which shows purchase order number 32799 as linked with only 3 boxes being onward sold. It is also consistent with the release note given by Leisure to Forward Logistics (the UK freight forwarder) which shows only 3 boxes being released to Pars.

131.         Of the 3 boxes delivered, 2 were sold and despatched by Pars to Emerald in Spain on 10 April.  The other box was combined with a separate chain of 4 boxes (see deal 15 below) and sold and despatched to Skinlight also in Spain and on 10 April.

132.         All the transactions in alleged chain A12 related to SL7Z9 CPUs.  The transactions are not all back to back. The sale by FX Drona to Innovate and by Innovate to Connect all take place on 3 April.  The sale by Connect to Pro Force and by Pro Force to Leisure all take place on 6 April.  Pars’ purchase order is dated 10 April but the invoice to it is dated 7 April.  It despatches the goods on the same date.

133.         We considered whether this was all part of a single chain or whether different chains had been conflated.  We conclude that it is more likely than not that the chain from FX Drona to Leisure are a single chain despite the date-change as Connect’s purchase order of 3 April carries the reference number 2006040601 and its invoice of 6 April carries the same number.  Evidence from other chains in this appeal is that Connect used the same number on a purchase order and invoice in a single deal chain.  In all other respects the deals appear connected as they all relate to 15 boxes of SL7Z9s and there are the same small price increments as shown in other deal chains in this appeal.

134.         Mr Cunningham for HMRC suggested that Pars’ documents were in some way indicative of Pars being involved in the fraud because there are two copies of purchase order number 32799, one dated 10 April and one dated 8 May.  However, we accept Mr Hamakarim’s evidence that there was no discrepancy in that the one dated 8 May is merely, as it states on its face, a copy and the 8 May date is merely the date on which it was printed.  Both copies of the purchase order otherwise have the same details and both in particular show the “date required” as 10 April.

135.         The oddity is that Leisure’s invoice is dated 7 April and Pars’ purchase order is dated 10 April.  Nevertheless we conclude that this is the same chain of goods.  This is because Pars itself identifies this Leisure invoice as showing their purchase of the goods Pars then sold to Emerald.

136.         This leaves the date discrepancy that Leisure’s own purchase order is dated 6 April and its invoice the day later.  We find there is nothing inherently improbable in buying goods on one day and selling them the next, in either a commercial or MTIC context. 

137.         There is however one further discrepancy is that Leisure’s release note states that its suppliers are “Blue and pro” whereas in the alleged chain only Pro is the supplier to Leisure.  Otherwise there is correlation of quantity and specification and the increment in price is similar to that seen elsewhere in this appeal. 

138.         We note that we have found that Leisure also made a purchase of the same type of CPU on the same day from Blue Star:  see Deal 7/11 above.  We find this is no doubt the explanation of why the release note refers to both “Blue” and “Pro.”  As Leisure bought 14 boxes in Deal 7/11 and only sold 12 to Pars in that chain, it is possible that 2 boxes from that chain were sold to Pars in this chain.  But it follows that at least one box originated in this chain tracing back to FX Drona.

139.         On the balance of probability we conclude that it is more likely than not the Pars’ purchase from Leisure on PO number 32799 is connected to this purchase by Leisure and thence to Pro Force and up the chain to FX Drona because it was the source of at least one of the boxes.  We find that all three boxes were connected to this chain back to FX Drona or to chain 7/11 back to KEP.  We do not need to decide which as it makes no difference to our conclusion as we find both chains to be connected to fraud as explained below.

Deal 15 Pars Purchase Order number 32793

140.         No copy of Pars’ purchase order was produced but from their CPU spreadsheet it was numbered 32793 and dated 7 April.  All the other documents in the alleged chain down to Pars were dated 7 April as well.  The product quantity and description was consistent down the alleged chain. The price increments were also consistent with those shown on the other deals in this appeal.  Pars despatched the goods to Skinlight in Spain on 10 April. There was no purchase order from Skinlight produced to the Tribunal.  No explanation of the 3 day delay from purchase to despatch was given and it may have been caused because Pars was waiting to acquire a fifth box which it did on 10 April.  In any event it is clear from Pars’ own evidence in its CPU spreadsheet that its sale to Skinlight was connected to its purchase of the 4 boxes from 21^st in this Deal 15 chain. 

141.         The Tribunal finds this alleged chain (A15) proved.  This purchase of SL7z9 CPUs by Pars was connected by a chain which traced back to KEP as set out in the Appendix.  Pars purchased the 4 boxes of SL7Z9s which it then sold and despatched on a single invoice with 1 box purchased on PO number 32799 (as mentioned above) to Skinlight in Spain. 

Deal 14 Pars Purchase Order Number 32798.

142.         The Tribunal finds that this purchase of CPUs by Pars was connected by a chain which traced back (in respect of 400 CPUs)  to KEP as set out in the Appendix and as alleged by HMRC as chain A14.  Pars sold and despatched the 500 CPUs to AAB in the Netherlands.

143.         Mr Beal for the Appellant suggested that the evidence produced by HMRC actually showed two separate chains and that the chain from Pars could only be traced back to its vendor, Aim.  He identified two discrepancies in the chain as alleged by HMRC.  Firstly, the paperwork from KEP to 21^st is dated 11 April but Aim’s and Pars’ paperwork is dated 10 April.  Secondly, the quantities do not match:  the dealers in the chain from KEP to 21^st supply 400 retail CPUs but Aim and Pars supply 500 retail CPUs.

144.         We bear in mind that otherwise the chains appear to relate:  they are for the same specification CPU “INTP4/3.2/800/2MB/775 Retail #640” and the price increments follow the same pattern as seen on all the other deals in this appeal.

145.         The date discrepancy appears a red herring.  Aim’s purchase order is dated 10 April but 21^st’s sales invoice to Aim is dated 11 April but they are clearly in relation to the same deal as 21^st’s invoice cites Aim’s purchase order number.  It appears to be the case that the purchase orders from 21^st, DDM and Time are all dated one day later than they should be because Pars actually despatched the goods on 10 April.  Bearing in mind that our overall conclusion (see paragraph 480) is that the evidence shows all the chains to which we have found Pars’ purchases the subject of this appeal to be connected were fraudulently orchestrated, we find that the mismatch in the paperwork is simply an error in the paperwork.

146.         The other discrepancy is in the quantity.  Aim purchased 400 CPUs from 21^st on 10 April but sold 500 to Pars on the same day.  Bearing in mind that other chains in this appeal have involved quantities of goods been split and/or joined, it is not improbable that Aim purchased the extra 100 CPUs from another company.  We find it to be more likely than not that the 500 sold to Pars included the 400 purchased by Aim from 21^st in the alleged A14 chain and therefore that Pars’ purchase (of 400 out of 500) was connected to that chain.

147.         We might have come to a different conclusion had Aim’s purchase order been dated 11 April as that might have indicated its purchase of 400 CPUs was not related to its sale of 500 CPUs the day before.  But, as noted, the evidence is that Aim purchased the 400 CPUs on the same day it sold them to Pars.

Deal 16 Pars Purchase Order number 32800

148.         With one peculiarity relating to Pars’ purchase order the documents in this chain are all dated 10 April.  The documents all relate to 15 boxes of SL7Z9 CPUs.  The price increments are the normal increments seen in the other chains in this appeal.

149.         The Tribunal finds that this purchase of CPUs by Pars was connected by the sale and purchase of a chain of 15 boxes of wholesale CPUs which traced back to KEP as set out in the Appendix and as alleged by HMRC as chain A16.  Pars sold 10 boxes to Datec on 11 April, a company registered for VAT in the UK.  The other five boxes were sold and despatched to Alvegar in Spain, one on a single invoice and the other 4 invoiced together on 12 April.  The units were shipped by Pars via UPS and delivered 11 April. 

150.         We note that this conclusion (as with a number of others we reach) is not entirely consistent with HMRC’s alleged chain.  In this instance they see Deal 13 as being the source of the single box of CPUs sold to Alvegar.  We do not agree.  We rely on Pars’ contemporaneous CPU spreadsheet as we consider it is more likely to be correct and is consistent with the invoices. 

151.         The oddity in this deal is that Pars’ original purchase order dated 10 April relates to 11 boxes of SL7Z9s.  A new purchase order raised on 12 April but for goods still required on 10 April is identical save that it now requires 15 boxes at the same unit price as before but obviously for a higher total price. 

152.         We find (from Pars’ evidence including their 4 invoices and CPU spreadsheet) that the deal as originally understood by Pars was that they would buy 11 boxes and sell 7 to Datec and 4 to Alvegar.  By 11 April or earlier the deal changed and was that they would buy 15 boxes and sell 10 to Datec and 5 to Alvegar.  The sale to Datec was split into two invoices:  invoice 82873 & 82874.  The first was for 7 boxes at £80.25 per unit and the second was for 3 boxes at £80 per unit.  They were both dated 11 April.  Pars’ records also show that its sales to Alvegar was split into two:  the invoice on 11 April for one box is at £122.50. And the second invoice on 12 April is at £122.40 per unit. 

153.         Mr Doherty for Pars accepted that the deal changed and that this was the explanation for the replacement purchase order.  His explanation for this was that the he presumed the buyer wanted more and vendor had more to sell.  However, it is clear that the chain from KEP always comprised 15 boxes so the increase in Pars’ purchase was much more likely to be driven by its vendor than its buyers, apart from the inherent unlikelihood of both buyers deciding simultaneously to increase their order. 

154.         We find nothing turns on this change in the deal.  It could be equally consistent with a deal on the open market or an organised MTIC chain.  But we do find the connection to KEP proved as stated above.

Deal 18 Pars Purchase Order number 32823

155.         The documents in this chain are all dated 13 April (with the normal exception of Pars’ invoice which is dated 18 April).  The documents all relate to 4 boxes of SL7Z9 CPUs.  The price increments are the normal increments seen in the other chains in this appeal.

156.         The Tribunal finds that this purchase of CPUs by Pars was connected by this chain of sales and purchases of CPUs as set out in the appendix to KEP and as alleged by HMRC as chain A18.  Pars sold and despatched the CPUs to IncoParts in the Netherlands on 13 April.

Deal 19 Pars Purchase Order number 32824

157.         The documents in this chain are all dated 13 April (with the normal exception of Pars’ invoice which is dated 18 April).  The documents all relate to 1000 SL7Z9 retail CPUs.  The price increments are the normal increments seen in the other chains in this appeal.

158.         The Tribunal finds that this purchase of CPUs by Pars was connected by this  chain of sales and purchases of CPUs which traced back to KEP as set out in the Appendix and as alleged by HMRC as chain A19.  Pars sold and arranged for the despatch of the CPUS to All Trading in the Netherlands on 13 April.

Deal 17 Pars Purchase Order number 32814 and Deal 20 Pars Purchase Order number 32821

159.         Pars’ documentation shows Deal 17 and Deal 20 to be separate transactions.  However, we have to look at them together because the alleged chains (A17a and A17b/20) leading to  Pars’ purchases involve the same players and (if we accept that they are the right chains) it is not possible to say which of the two chains was the source of the goods in each of the two Pars’ deals.

160.         One chain (referred to by HMRC as A17a) related to 2,400 retail SL7Z9s sold in a chain from KEP as outlined in the appendix.  The other alleged chain (referred to by HMRC as A17b and A20) was identical save that it related to 1,970 retail SL7Z9s.  The documents in both chains are all dated 12 April.  The price increments are the normal increments seen in the other chains in this appeal.  The oddity in the chains down to Blue Star is that it shows that the same party is selling the same goods to the same buyer at two slightly different prices on the same day.

161.         Leisure buys both lots of the CPUs on 12 April (as has its predecessors up the chain).  It sells one lot of 2,000 CPUs to Pars on 12 April and another lot of 2,370 CPUs to Pars the day after.  Pars sold and despatched 2,000 CPUs to Inco Pars in  the Netherlands on 12 April.  It sold and despatched 2,370 CPUs to Dutch Team in the Netherlands on 13 April.

162.         Taking into account that in total Leisure bought 4,370 CPUs and sold the same amount to Pars, and taking into account the identity in quantity and description of the goods back up the two chains, and the price increments, the Tribunal finds it more likely than not that these two chains as set out in the Appendix were the source of the 4,370 CPUs purchased by Pars.  Therefore Pars’ purchase of all these CPUs was connected by two chains of sales and purchases of CPUs to KEP.

Deal 21 Pars Purchase Order Number 32842

163.         The documents in this chain are all dated 20 April (including Pars’ invoice).  The participants in the alleged chain (A21) above Pars all deal in 27 boxes of SL7Z9s. The price increments are the normal increments seen in the other chains in this appeal.  It is clear that Culmain’s purchase of 27 boxes is linked to its sale to Pars of 23 boxes as (as can be seen from other transactions in this appeal) Culmain’s invoice number is used as its purchase order reference number where the purchase and sale are linked.  In this deal the invoice is number 1108A/1109 and its purchase order reference number is 1109.  Pars buys 23 boxes of SL7Z9s and sells and despatches the same to Mashall on 20 April.

164.         Based on this evidence, the Tribunal finds it more likely than not that this purchase of CPUs by Pars from Culmain was connected by this chain of sales and purchases of CPUs which traced back to Ability as set out in the Appendix. 

Deal 22 Pars’ Purchase Order Number 32849

165.         The documents in this alleged chain (A22) are all dated 21 April (including Pars’ invoice).  The participants in the alleged chain above Pars all buy 33 boxes of SL7Z9s (with the exception of Ability at the top who sells 33 boxes but there is no evidence of how many it purchased). The price increments are the normal increments seen in the other chains in this appeal.  Pars buys 17 boxes of SL7Z9s and sells and despatches the same to Mashall on 21 April. It is clear that Culmain’s purchase of 33 boxes is linked to its sale of Pars of 17 boxes as (as can be seen from other transactions in this appeal) Culmain’s invoice number is used as its purchase order reference number where the purchase and sale are linked.  In this deal the invoice is number 1113 and its purchase order reference number is 1113. 

166.         Based on this evidence, the Tribunal finds it more likely than not that this purchase of CPUs by Pars from Culmain was connected by this chain of sales and purchases of CPUs which traced back to Ability as set out in the Appendix. 

Deal 23 & 26 Pars’ Purchase Order Number 32837

167.         The documents in this single alleged chain (A23/A26) are all dated 19 April with the usual exception of Pars’ invoices which are both dated 21 April (although the pro forma is dated 19 April) and also with the exception of Inco Parts’ purchase order which is dated 18 April.  The participants in the alleged chain above Pars all buy 7 boxes of SL7Z9s (with the exception of Ability at the top who sells 7 boxes but there is no evidence of how many it purchased). The price increments are the normal increments seen in the other chains in this appeal.  Pars buys 4 boxes of SL7Z9s and sells and despatches 3 of them to Mashall (Netherlands) on 19 April and the remaining box to Emerald (Spain) also on 19 April. It is clear that Culmain’s purchase of 7 boxes is linked to its sale to Pars of 4 boxes as (as can be seen from other transactions in this appeal) Culmain’s invoice number is used as its purchase order reference number where the purchase and sale are linked.  In this deal the invoice is number 1106 and its purchase order reference number is 1106. 

168.         Based on this evidence, the Tribunal finds it more likely than not that this purchase of CPUs by Pars from Culmain was connected by this chain of sales and purchases of CPUs which traced back to Ability as set out in the Appendix. 

Deal 24 Pars Purchase Order Number 32839

169.         The documents in this alleged chain (A24) are all dated either 18 or 19 April (with the exception of Pars’ invoice which is as usual dated after the shipment). From KEP to 21^st’s purchase order they are dated 18 April.  21^st’s invoice down to Inco Parts’ purchase order are dated 19 April, the day after.  We find on the balance of probabilities 21^st’s purchase order on 18 April is related to its invoice dated the day after.  The participants in the alleged chain all buy 5 boxes of SL7Z9s (with the exception of KEP at the top who sells 5 boxes but there is no evidence of how many it purchased). The price increments are the normal increments seen in the other chains in this appeal.  Pars sells and despatches the 5 boxes to Inco Parts in the Netherlands on 19 April. 

170.         Based on this evidence, the Tribunal finds it more likely than not that this purchase of CPUs by Pars from 21st was connected by this chain of sales and purchases of CPUs which traced back to KEP as set out in the Appendix. 

Deal 25 Pars’ Purchase Order Number 32840

171.         The documents in this alleged chain (A25) are all dated 20 April (with the usual exception of Pars’ invoices which is dated 21 April).  The participants in the alleged chain including and above Pars all buy 800 retail boxes of SL7Z9s (with the exception of Ability at the top who sells 800 retail CPUs but there is no evidence of how many it purchased). The price increments are the normal increments seen in the other chains in this appeal.  Pars buys 800 SL7Z9s but only sells and despatches 750 of them to All Trading in the Netherlands.  The invoice/reference number used in Culmain’s documentation indicates that its purchase in this chain is linked with its sale.

172.         Mr Hamakarim’s evidence was that the discrepancy in the amounts bought and sold by Pars arose because Pars kept back 50 of the retail CPUs for its own use and we accept this evidence.

173.         Based on all this evidence, the Tribunal finds it more likely than not that this purchase of CPUs by Pars from Culmain was connected by this chain of sales and purchases of CPUs which traced back to Ability as set out in the Appendix. 

Deal 27 Pars’ Purchase Order Number 32843

174.         The documents in this alleged chain (A27) are all dated 20 April with the usual exception of Pars’ invoice which is dated 25 April, and also with the exception of Inco Parts’ purchase order which is dated 21 April. The participants in the alleged chain buy 17 boxes of SL7Z9s.  The price increments are the normal increments seen in the other chains in this appeal.  Pars despatches the 17 boxes to Inco Parts in the Netherlands on 21 April. Pars’ own spreadsheet connects its sale on 21^st to its purchase on 20^th.

175.         The alleged purchase by Midwest from Megatek is in fact an acquisition as Megatek is based in France and therefore Midwest’s purchase order (there is no invoice) which shows the purchase is free of VAT.  The purchase order covers 18 separate items being purchased in varying quantities.  The total value of the order is nearly £11million.  6 of these items (some 100 boxes in total) relate to wholesale SL7Z9s and in a commercial world it makes no sense for them to be itemised separately.  It is perhaps possible that they are all SL7Z9s with separate specifications but since the specifications are not given this also does not make sense in a commercial world. The unit prices also vary. One of the items is for “5355 SL7Z9” which ties in with Midwest’s sale of 5355 SL7Z9s to Data Solutions in the chain put forward by HMRC.  In any event the quantities of SL7Z9s on this purchase order would easily have allowed Midwest to make its sale to Data Solutions. 

176.         Based on this evidence, the Tribunal finds it more likely than not that this purchase of CPUs by Pars from TGT was connected by this chain of sales and purchases of CPUs which traced back to Midwest and Megatek as set out in the Appendix. 

Deal 28 Pars Purchase Order Number 32859

177.         With two exceptions, the documents in this alleged chain (A28) are all dated 24 April.  One exception is that as usual Pars’ invoice (but not of course its purchase order) is dated 25 April.  The other exception is that Innovate’s invoice is dated 24 June.  Its purchase order is, like all the other documents in the chain dated 24 April and we conclude it is more likely than not that it is a typographical error and that the transactions did take place on 24 April.  The documents all related to 8 boxes of SL7Z9s.  The price increments are the normal increments seen in the other chains in this appeal.  Pars sells and despatches 8 boxes to ASAP in the Netherlands on 24 April. 

178.         Based on this evidence, the Tribunal finds it more likely than not that this purchase of CPUs by Pars from Culmain was connected by this chain of sales and purchases of CPUs which traced back to Ability as set out in the Appendix. 

Deals 29A and 29B  Pars’ Purchase Order Numbers  32893 and 32892

179.         Deal 29 is two alleged chains (A29a and A29b) as Pars was denied input tax on 2 invoices.  What happened (as evidenced by the invoices and Pars’ CPU spreadsheet) is that Pars bought one lot of 20 boxes and one lot of 4 boxes from Culmain and sold them on a single invoice to Mashall.  They were exported to Mashall in Switzerland on 27 April.

180.         All the documents in the first alleged chain (A29a) relate to SL7Z9s and the documents from Connect to Culmain are dated 26 April and relate to 6 boxes.  From Connect to Pars they also show the price increments that are of the type shown in all other chains in this appeal.  HMRC allege that Connect bought from Apollo and Apollo from Ability.  However, the documents for this part of the chain only relate to 5 boxes and are dated 27 April, the day after the later parts of the chain.  There is a mismatch in price between Connect’s purchase order addressed to Apollo and Apollo’s invoice addressed to Connect and further Connect’s supplier declaration to Apollo dated 26 April in relation to the 6 boxes refers to invoice “ASL-014” whereas Apollo’s invoice for the 5 boxes in the chain alleged by HMRC is numbered “ASL-015”.

181.         There is also the discrepancy in that Culmain buys 6 boxes but only sells 4 to Pars.  Taking into account the identical date and normal price increments and fact that Culmain links this purchase of 6 boxes to its sale of 4 boxes as its invoice is numbered “1125” and its purchase order is referenced 1125, we find that the link from Pars to this chain with Culmain back to Connect is established.

182.         We find that the documents produced by HMRC to evidence a connection beyond Connect to Apollo and Ability do not in fact relate to the chain of 6 boxes sold on 26 April but a separate supply of 5 boxes which took place a day later.  However, we find it more likely than not that Connect purchased those 6 boxes from Apollo because it issued a supplier declaration to Apollo on 26 April in relation to a purchase of 6 boxes of CPUs at the right price (judging by the routine price increments) and quotes what we find to be an Apollo invoice number (albeit no copy of invoice ASL-014 has been produced to us.) 

183.         All the documents in the second alleged chain (A29b) relate to SL7Z9s and are dated 27 April (including Pars’ invoice).  The alleged chain comprises 19 boxes save for Culmain’s sale to Pars which is of 20 boxes.  However, as chain A29a shows, Culmain had 2 spare boxes as it had purchased 6 boxes the day before and only sold 4 to Pars.  The alleged chain shows the sort of price increments which are normal in the chains in this appeal. 

184.         Therefore, we find it more likely than not that the purchase of 19 of 20 of the boxes of CPUs by Pars from Culmain was connected to the chain of 19 boxes (A29a) as set out in the Appendix going back to Ability. The other box we find more likely than not was connected to chain A29a as shown in the appendix and tracing back to Apollo.

Deal 30 Pars’ Purchase Order Number 32889

185.         With the exceptions mentioned below the documents in this alleged chain (A30) are all dated 26 April.  The exceptions are that 21^st’s invoice, but not its purchase order, is dated 27 April.  Pars’ invoice is also dated 27 April and the goods were shipped on the same date to Switzerland.  There is no copy of Pars’ purchase order but its CPU spreadsheet indicates that Pars considered the deal took place on 27 April so more likely than not this would have been the date of the purchase order.  The other exception is the purchase order from Gig. 

186.         Bearing in mind that we find that the chain overall appears to relate to the same goods as it shows 4 boxes of SL7Z9s being sold for increments in price of the type shown in all other proved chains in this appeal,  the change in date in the later part of the chain does not appear significant.  We find 21^st bought the goods on 26^th but did not sell them until the following day. 

187.         The date on Gig’s purchase order was not explained to us.  It shows a date of 22 March (with a delivery date of 27 March) which is a month before the other documents in this chain.  Nevertheless, we still conclude that this purchase order from Gig relates to the chain as set out in the Appendix.  This is because Pars’ CPU spreadsheet links its purchase of 4 boxes of SL7Z9s on 27 April from 21^st to its sale of the same to Gig.  Pars informed us that all its CPU deals were recorded on this spreadsheet and there is no other sale of 4 boxes of SL7Z9s in the period 23 March to 27 April to Gig.  So we find this purchase order does relate to this deal.

188.         Based on this evidence, the Tribunal finds it more likely than not that this purchase of CPUs by Pars from 21^st was connected by this chain of CPUs which traced back to Ability and as set out in the Appendix. 

Deal 31& 33 Pars Purchase Order Number 32863

189.         With the exceptions mentioned below the documents in this single alleged chain (A31/33) are all dated 24 April.  The exceptions are that Culmain’s invoice to Pars and Pars’ and Inco Parts’ purchase orders date to the following day.  Pars invoices are dated 27 April.  The goods were shipped on 25 April. Although Culmain’s invoice is dated 25 April its purchase order is dated 24 April.  The alleged chain on 24 April comprised 15 boxes of SL7Z9s.  Culmain then sold 5 boxes to Pars the next day and Pars sold and despatched 1 box to Alvegar and 4 boxes to Inco Parts.

190.         It is clear that Culmain’s purchase of 15 boxes is linked to its sale to Pars of 5 boxes as (as can be seen from other transactions in this appeal) Culmain’s invoice number is used as its purchase order reference number where the purchase and sale are linked.  In this deal the invoice is number 1118A and its purchase order reference number is 1118.  We also note that this chain shows the price increments of the type shown in all other proved chains in this appeal.

191.         Based on this evidence, the Tribunal finds it more likely than not that this purchase of CPUs by Pars from Culmain was connected by this chain of CPUs which traced back to Ability and as set out in the Appendix. 

Deal 32 & 34 Pars Purchase Order Number 32881

192.         With the normal exception of Pars’ invoices (which are dated the day after) the documents in this single alleged chain (A32/34) are all dated 26 April.  The goods were shipped on 26 April. The alleged chain comprised 8 boxes of SL7Z9s.  Pars bought 8 boxes from 21^st and then sold and despatched 4 boxes to Skinlight and 4 to Inco Parts.  The documents show the price increments of the type shown in all other proved chains in this appeal.

193.         Based on this evidence, the Tribunal finds it more likely than not that this purchase of CPUs by Pars from Culmain was connected by this chain of sales and purchases of CPUs which traced back to KEP and as set out in the Appendix. 

Deal 35 Pars Purchase Order Number 32868

194.         With the normal exception of Pars’ invoice (which is dated 27 April) the documents in this alleged chain (A35) are all dated 25 April.  The goods were shipped on 25 April. The alleged chain comprised 10 boxes of SL7Z9s.  Pars sold and despatched the 10 boxes to ASAP.  The documents show the price increments of the type shown in all other proved chains in this appeal.

195.         Based on this evidence, the Tribunal finds it more likely than not that this purchase of CPUs by Pars from Culmain was connected by this chain of sales and purchases of CPUs which traced back to Ability and as set out in the Appendix. 

Deal 36 Pars’ Purchase Order Number 32879

196.         With the following exceptions, the documents in this alleged chain (A36) are all dated 25 April.  The exceptions are that Leisure’s invoice to Pars, Pars’ and All Trading’s purchase orders are dated the day after. The goods were also shipped on 26 April. The alleged chain down to Leisure comprised 11 boxes of SL7Z9s. Pars bought 9 boxes from Leisure and sold and despatched them to All Trading.  The documents show the price increments of the type shown in all other proved chains in this appeal.

197.         Leisure’s sale to Pars on 26 April is we find linked to its purchase from Blue Star on 25 April in HMRC’s alleged chain.  The reasons are as follows.  Mr Churchouse’s evidence is that inspection reports in this deal pack were found in Leisure’s possession.  They are dated 25 April and show the box numbers that match those in Pars’ CPU spreadsheet.  Leisure’s release note to Forward Logistics also identifies one of the boxes by one of the numbers also used in Pars’ spreadsheet and shows the supplier to be Blue Star.

198.         As mentioned above there is no documentary evidence linking Time to KEP but we have accepted the chain to be as alleged by HMRC for the reasons given in paragraphs 103-106 above.

199.         Based on this evidence, the Tribunal finds it more likely than not that this purchase of CPUs by Pars from Leisure was connected by this chain of sales and purchases of CPUs which traced back to KEP and as set out in the Appendix. 

Deal 37 Pars’ Purchase Order Number 32904

200.         The documents in this alleged chain (A37) are all dated 28 April.  The chain comprises 21 boxes of SL7Z9s and shows the price increments of the type shown in all other proved chains in this appeal.  Pars sold and exported the 11 boxes to Mashall on 28 April.

201.         Based on this evidence, the Tribunal finds it more likely than not that this purchase of CPUs by Pars from Futuristic was connected by this chain of sales and purchases of CPUs which traced back to Ability and as set out in the Appendix. 

Deal 38 Pars’ Purchase Order Number 32890

202.         The documents in this alleged chain (A38) are all dated 27 April with the exception of Pars’ invoice which is dated 30 April.  The despatch to ASAP took place on 27 April.  The chain comprised 10 boxes of SL7Z9s and shows the price increments of the type shown in all other proved chains in this appeal. 

203.         Based on this evidence, the Tribunal finds it more likely than not that this purchase of CPUs by Pars from Futuristic was connected by this chain of sales and purchases of CPUs which traced back to Ability and as set out in the Appendix. 

Deal 39 Pars’ Purchase Order number 32888

204.         The documents in this alleged chain (A39) are all dated 27 April with the exception of Pars’ invoice which is dated 30 April.  The despatch to Inco Parts took place on 27 April.  The chain comprised 1000 retail SL7Z9 CPUs and shows the price increments of the type shown in all other proved chains in this appeal. 

205.         Based on this evidence, the Tribunal finds it more likely than not that this purchase of CPUs by Pars from Culmain was connected by this chain of sales and purchases of CPUs which traced back to Ability and as set out in the Appendix. 

Deals 40 & 41 Pars’ purchase order number 32905

206.         The documents in this single alleged chain (A40/41) are all dated 28 April with the exception of Pars’ invoices which are dated 30 April and the purchase order from Inco Parts which is dated 27 April.  The despatches to Inco Parts and Bridema took place on 28 April.  The chain comprised 5100 retail SL7Z9s and shows the price increments of the type shown in all other proved chains in this appeal.  Pars sold 1,500 of the units to Bridema and 3,500 of the units to Inco Parts.  No evidence was given on what happened to the missing 100 units other than a general comment made by Mr Hamakarim that on a few occasions Pars took for own use a few of the CPUs in the deals and we find that this was one of the occasions.

207.         Based on this evidence, the Tribunal finds it more likely than not that this purchase of CPUs by Pars from Culmain was connected by this chain of sales and purchases of CPUs which traced back to Ability and as set out in the Appendix. 

Deal 42 Pars’ Purchase order number 32918

208.         The documents in these alleged chains (M1a and M1) are all dated 2 May with the only exception being Pars’ invoices which are dated 5 May. The alleged chains featured Ability at the start selling 23 boxes (chain M1A) of SL7Z9s to Innovate and 12 boxes (chain M1) of SL7Z9s to Apollo. Both legs of the chain are sold through 2 more alleged buffers before both being sold to Culmain.  Culmain then sells 25 boxes to Pars, who splits this and sells 22 boxes to Mashall and 3 boxes to Inco Parts.  The goods were shipped on 2 May.  The price increments in both legs of the chain are of the sort evidence in all the other proved chains in this appeal.

209.          The evidence that Ability was at the start of the 23 box leg (M1A) is the invoice it issued to Innovate and Innovate’s purchase order to Ability.  No such evidence was produced to the Tribunal for the start of the 12 box leg (M1).  However, a third party payment instruction from Ability to Apollo was produced in evidence.  This carries the reference number “ASL-0021” and the date of 2 May.  Apollo’s invoice to its buyer (Connect) in this alleged chain carries the same reference and date.  As Mr Beal pointed out, Apollo’s reference number used by Ability on the third party payment instructions must have been taken off Apollo’s purchase order, but (as he recognised) Apollo uses the same reference number for its invoices as for its connected purchase orders, this connects Ability’s payment instruction to Apollo’s purchase order issued to Ability and to Apollo’s invoice issued to Connect.

210.         Based on this evidence, the Tribunal finds it more likely than not that this purchase of 25 boxes CPUs by Pars from Culmain was connected by these two legs of the  chain of sales and purchases of CPUs back to Ability and as set out in the Appendix. 

Deal 43 Pars Purchase Order Number 32930

211.         The documents in this alleged chain (M2) are all dated 2 May with the exception of Leisure’s invoice (dated 3 May), Pars’ purchase order (dated 3 May) and Pars’ invoice (dated 5 May).  The alleged chain features 20 boxes of SL7Z9s being sold along a chain from KEP to Leisure.  Leisure sells 2 boxes only to Pars, so bearing in mind that Leisure’s purchase order does not bear the same date or number of CPUs as its invoice to Pars, the question arises whether it is the same deal chain.  However, Leisure’s release note to Forward Logistics in respect of the 2 boxes sold to Pars (and which gives the same box numbers as shown on Pars’ spreadsheet) identifies their supplier as Blue Star, as in the alleged chain.  The price increments in the chain are also of the sort evidenced in all the other proved chains in this appeal.

212.          Based on this evidence, the Tribunal finds it more likely than not that this purchase of CPUs by Pars from Leisure was connected by this chain of sales and purchases of CPUs back to KEP and as set out in the Appendix. 

Deal 44 Pars’ Purchase Order Number 32917

213.         The documents in the two legs of this alleged chain are all dated 2 May with the only exception being Pars’ invoices which are dated 5 May. The alleged chains featured Ability at the start selling 3,500 retail SL7Z9 CPUs (M3a) to Apollo and the same to Innovate (M3).  Both legs of the chain are sold through 2 more alleged buffers before both being sold to Culmain.  Culmain then sells the 7,000 CPUs on a single invoice to Pars, who then splits the consignment again and sells half to Inco Parts and half to ASAP.  The goods were shipped on 2 May.  The price increments in both legs of the chain are of the sort evidenced in all the other proved chains in this appeal.

214.          Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from Culmain was connected by these two legs of a chain of sales and purchases of CPUs back to Ability and as set out in the Appendix. 

Deal 45 Pars’ Purchase orders numbers 32934

215.         Although described as a single deal because Pars combined two purchases and made a single sale to Inco Parts, we see this as two deals as there were two purchases by Pars and therefore two invoices on which HMRC has refused to repay input tax and we deal with them separately.

216.         Chain M4 which HMRC alleges leads to Pars’ purchase order number 32934 comprised documents all dated 3 May with the exception of Pars’ invoice which was dated 5 May.  The goods were shipped on 3 May to Inco Parts. The chain comprised 1,000 retail SL7Z9s and shows the price increments of the type shown in all other proved chains in this appeal.

217.         As mentioned above there is no documentary evidence linking Time to KEP but we have accepted the chain to be as alleged by HMRC for the reasons given in paragraphs 103-106 above.

218.          Based on this evidence, the Tribunal finds it more likely than not that this purchase of CPUs by Pars from 21^st was connected by this chain of CPUs which traced back to KEP and as set out in the Appendix. 

Deal 45 Pars’ Purchase orders numbers 32933

219.         Chain M4A which HMRC alleges leads to Pars’ purchase order number 32933 comprised documents which were dated 27 April with the exception of Leisure’s invoice to Pars and Pars’ purchase order (both dated 3 May), Inco Parts’ purchase order (dated 3 May) and Pars’ invoice dated 5 May.  All the documents dated 27 April relate to 4,000 retail SL7Z9s CPUs whereas Leisure only sells 500 of such CPUs to Pars.  The question arises therefore whether the purchase made by Leisure of 4,000 CPUs on 27 April is connected to its sale of 500 CPUs to Pars on 3 May.

220.         The Tribunal was given hearsay evidence in the form of a fax dated 31 October 2006 from Leisure to HMRC explaining that the 4,000 CPUs was purchased on their purchase order no P406026 dated 27 April.  They were sold to First Talk but 500 of them were then returned and sold to Pars on invoice I506004.  HMRC also exhibited the invoice from Leisure to First Talk.  The purchase order and invoice numbers cited match those in the chain exhibited by HMRC and alleged to be the chain for this purchase by Pars.  HMRC also exhibited the credit note for 500 CPUs from Leisure to First Talk dated 28 April, which corroborates the fax from Leisure.

221.         Based on this evidence, the Tribunal finds it more likely than not that this purchase of CPUs by Pars from Leisure was connected by this chain of sales and purchases of CPUs back to KEP and as set out in the Appendix. 

Deal 46 Pars’ Purchase Order Number 32947

222.         The documents in this alleged chain (M5) are all dated 4 May with the only exception being Pars’ invoices which are dated 5 May (sale to Inco Parts) and 10 May (sale to All Trading). The goods were shipped by Pars to Inco Parts and All Trading on 4 May.  The price increments in the chain are of the sort evidenced in all the other proved chains in this appeal.  The entire chain relates to 5,000 retail SL7Z9s except that Pars sells only 1,500 to Inco Parts and then combines the remaining 3,500 with the 1,000 purchased on order 32950 (Deal 47) and sells these 4,500 to All Trading.

223.         Mr Beal challenged this alleged chain on the basis that there was neither a purchase order from nor invoice to Pro Force.  In the event, this challenge turned out to be unfounded as HMRC located the missing invoice in the deal pack for deal 47.  HMRC also exhibited a release note dated 4 May from Connect to Forward Logistics instructing them to release to Pro Force 5,000 SL7Z9s.

224.         Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from Culmain was connected by the chain of CPUs back to Ability and as set out in the Appendix. 

Deal 47 Pars’ Purchase Order number 32950

225.         The documents in this alleged chain (M5a) are all dated 4 May with the only exception being Pars’ invoice which is dated 10 May, and is the same invoice to All Trading as referred to above in Deal 46. The goods were shipped by Pars to All Trading on 4 May.  The price increments in the chain are of the sort evidenced in all the other proved chains in this appeal.  The entire chain relates to 1,000 retail SL7Z9s except that, as explained above Pars combined 1,000 CPUs with the remaining 3,500 from deal 46 and sold these 4,500 CPUs to All Trading.

226.         There is no copy of Pars’ purchase order but there is no suggestion it did not exist.  We find it was numbered 32950 relying on 21^st’s invoice to Pars and on Pars’ own spreadsheet.

227.         As mentioned above there is no documentary evidence linking Time to KEP but we have accepted the chain to be as alleged by HMRC for the reasons given in paragraphs 103-106 above.

228.         Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from 21^st was connected by a chain of sales and purchases of CPUs back to KEP and as set out in the Appendix. 

Deal 48 Pars’ Purchase Order Number 32922

229.         The documents in this alleged chain (M6) are all dated 2 May with the only exception being Pars’ invoice to BM Euronet which is dated 5 May. The goods were shipped by Pars to BM Euronet on 4 May.  The price increments in the chain are of the sort evidenced in all the other proved chains in this appeal.  The first five companies in the chain sell 4 boxes of SL7Z9s.  The last of these companies, Delltronics, buys 4 boxes but sells only 2 to Pars.  Pars then sells these boxes to BM Euronet. 

230.         Mr Beal challenged whether the 2 boxes purchased by Pars from Delltronics were connected to the 4 boxes purchased by Delltronics in this alleged chain.  There is an invoice from Point of Logistics, the freight forwarder dated 8 May making a small charge for a release of goods:  2 boxes of CPUs to Pars and 2 boxes to a purchaser identified as Aston Tech.  There are allocation instructions to Point of Logistics from Delltronics dated 3 May referring to the two boxes to be released to Pars (and Pars shipped the following day).  As there is no other relevant sale by Delltronics to Pars at this time it appears both the allocation instructions and resulting invoice relate to this sale, and is therefore evidence that Delltronics split a purchase of 4 boxes into two.  We also bear in mind the evidence from other proven chains in this appeal that it was quite common for consignments to be split.  We also bear in mind that the goods’ specification and dates were the same.

231.         Mr Hamakarim gave evidence that the deal was negotiated on 2 May but because Pars had to wait for a Redhill check to come through the deal was not actually completed until 3 May.  In fact Pars’ purchase order is dated 2 May although payment is not made until 3 May.  This does explain the slight delay in the chain.

232.         Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from Delltronics was connected by a chain of sales and purchases of CPUs back to Bullfinch and as set out in the Appendix. 

Deal 49 Pars’ Purchase Order Number 93180

233.         The documents in this alleged chain (M7) are all dated 8 May. The goods were shipped by Pars to Formosa on 8 May.  The price increments in the chain are of the sort evidenced in all the other proved chains in this appeal.  All documents in the chain relate to 12 boxes of SL7Z9s. 

234.         Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from 21^st was connected by the chain of CPUs back to Okeda and as set out in the Appendix. 

Deal 50 Pars’ Purchase Order Number 32955

235.         The documents in this alleged chain (M8) are all dated 5 May apart from Pars’ invoice which is dated 10 May.  Pars shipped the goods to Inco Parts on 5 May.  The price increments in the chain are of the sort evidenced in all the other proved chains in this appeal.  All documents in the chain down to Leisure relate to 18 boxes of SL7Z9s. 

236.          There is an invoice and purchase order for every step in this transaction up to Time apart from the absence of a purchase order from Ultimate. Mr Beal also claimed that there was no invoice from Resolutions, meaning that the chain could not be traced back beyond Ultimate.  He is mistaken.  Resolutions’ invoice is dated 5 May, numbered R263 and relates to 5,670 CPUs (18 boxes) at a unit price of £76.65:  R8/1896.  He appears to have been misled by the inclusion of an unrelated invoice at R8/1890.  This invoice relates to a separate chain of 2205 CPUs tracing back to Resolutions from Blue Star.  This chain of 2205 CPUs is clearly the wrong chain because, in addition to the evidence of the right chain discussed in the previous paragraph, it is dated 9/5 and the price would involve Blue Star selling at a loss.

237.         The alleged chain shows Leisure buying 18 boxes but selling only 7 boxes of SL7Z9s to Pars.  However, there is also an inspection report and release instruction from Leisure to the freight forwards (Forward Logistics). It was Mr Churchouse’s evidence (which we accept) that the inspection reports in the deal packs came from Leisure’s records.  In any event this evidence is verified because Leisure’s release instruction identifies 7 boxes to be sold to Pars by reference to the item number on the inspection report. The inspection report has 17 items and items numbered 11-17 are the ones identified on the release instruction for release to Pars.  The box numbers for these (shown on the inspection report) tally with the ones shown in Pars’ CPU spreadsheet. 

238.         As mentioned above there is no documentary evidence linking Time to KEP but we have accepted the chain to be as alleged by HMRC for the reasons given in paragraph 103-106 above.

239.         Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from Leisure was connected by a chain of sales and purchases of 18 boxes CPUs back to KEP and as set out in the Appendix. 

Deal 51 Pars’ Purchase Order Number 32963

240.         The documents in this alleged chain (M9) are all dated 8 May except for Pars’ invoice which is dated 10 May.  The goods were shipped by Pars to Inco Parts on 8 May.  The price increments in the chain are of the sort evidenced in all the other proved chains in this appeal.  All documents in the chain relate to 1900 retail SL7Z9 630s.

241.         Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from TGT was connected by a chain of sales and purchases of CPUs back to Bullfinch and as set out in the Appendix. 

Deal 52 Pars’ Purchase Order Number 93213

242.         The documents in this alleged chain (M10) are all dated 8 May with the exception of Pars’ invoice which is dated 10 May. The goods were shipped by Pars to ASAP on 8 May.  The price increments in the chain are of the sort evidenced in all the other proved chains in this appeal.  All documents in the chain relate to 1,000 retail SL7Z9s. 

243.         Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from 21^st was connected by a chain of sales and purchases of CPUs back to Okeda and as set out in the Appendix. 

Deal 53 Pars’ Purchase Order Number 33000

244.         The documents in this alleged chain are all dated 11 May except for Formosa’s purchase order which is dated 8 May for shipment on 11 May. The goods were shipped by Pars to Formosa on 11 May.  All documents in the chain relate to 20 boxes of SL7Z9s. 

245.         The price increments in the alleged chain are of the sort evidenced in all the other proved chains in this appeal with the exception that Okeda sells at £81.50 and Time sells at £79.05.  This is a loss per unit of £2.45 for Time whereas in every other proved chain in this appeal where Okeda sells to Time, Time makes a profit of 5p per unit. 

246.         However, as discussed in paragraph 103-106 we have also found that on the evidence in this appeal, Time purchased exclusively from KEP before 8 May and exclusively from Okeda after this date.  Therefore, although we do not find the Okeda to Time invoice in the deal chain M11 as alleged by HMRC is the right invoice, we do find it more likely than not that Time did purchase the CPUs in the chain from Okeda.

247.         The Tribunal therefore finds it proved that this purchase of CPUs by Pars from 21^st was connected by a chain of sales and purchases of CPUs back to Okeda and as set out in the Appendix. 

Deal 54 Pars’ Purchase Order Number 32976

248.         The documents in this alleged chain (M12 and M12a) are all dated 9 May except for Pars’ invoice which is dated 15 May.  The alleged chains show Ability selling to Innovate (on 2 separate invoices) 1,000 retail CPUs (chain M12) and 2,500 retail CPUs (chain M12a).  The chains each go through 2 more buffers and are then both sold to Culmain.  Culmain sells the 3,500 CPUs on a single invoice to Pars.  The goods are sold and shipped by Pars to Inco Parts on 9 May.  The price increments in the chain are of the sort evidenced in all the other proved chains in this appeal.  All documents in the chain appear to relate to retail SL7Z9s.

249.         Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from Culmain was connected by two chains of sales and purchases of CPUs back to Ability and as set out in the Appendix. 

Deal 55 & 58 Pars’ Purchase Order Number 32997

250.         The documents in this single alleged chain (M13/M16) are all dated 11 May except for Pars’ invoice which is dated 15 May.  Pars and the companies in the chain above Pars all deal in 3,420 SL7Z9s.  The alleged chain shows the price increments of a sort evidenced in all other proved chains in this appeal.  Pars sells 2,000 of the CPUs purchased from Culmain to ASAP.  The evidence for this is Pars’ own spreadsheet.  As can be seen from Pars’ spreadsheet, the remaining 1,420 are combined with its purchase from Futuristic on purchase order 32996 (see Deal 55 (M13a) below) of 1,000 units and sold to Inco Parts.  The goods were shipped by Pars to ASAP and to Inco Parts on 11 May.

251.         There is no invoice from Ability and Innovate’s purchase order does not name Ability.  Nevertheless it gives the supplier’s address and that address belonged to Ability as can be seen from Ability’s invoices which evidence other chains in this appeal.

252.         Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from Culmain was connected by a chain of sales and purchases of CPUs back to Ability and as set out in the Appendix. 

Deal 55 Pars’ Purchase Order Number 32996

253.         The documents in this alleged chain (M13a) are all dated 11 May except for Pars’ invoice which is dated 15 May, as mentioned above, and relate to 1,000 retail SL7Z9s.  The goods were shipped by Pars to Inco Parts on 11 May, as mentioned above, combined with 1,420 retail SL7Z9s bought on purchase order 32997(deal 55/58).  The price increments in the chain are of the sort evidenced in all the other proved chains in this appeal. 

254.         Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from Futuristic was connected by a chain of sales and purchases of CPUs back to Ability and as set out in the Appendix. 

Deal 55 Pars’ Purchase Order Number 32998

255.         Pars’ CPU spreadsheet as well as its purchase order and Aim’s invoice shows that it purchased 800 Z8 retail CPUs from Aim on 11 May.  It sold them to Inco Parts on the same invoice with which it sold Inco Pars the 2,420 retail Z9s mentioned in the above 2 deals. 

256.         HMRC advanced two possible chains as the source of these 800 Z8s. One chain, M13b(1) goes through 21^st back to KEP.  Although this was not noticed in the hearing, it comprises exactly the same documents as we have found proved to be connected with Deal 14.  They are in any event for the wrong quantity and one month too early for Deal 55, and would have involved Aim taking a loss per unit of £5.50 all of which factors which would in any event have made it highly improbable that they related to Deal 55.  We therefore do not find it proved that they relate to Deal 55.

257.         The other chain alleged by HMRC to be the source of the 800 Z8s was chain M13b(2). They produced documents evidencing a chain from SES to Booming to Miaotech on 25 May of 645 Z8s proceeding in orderly 10p price increments.  Mr Churchouse gave evidence that HMRC have a purchase order and invoice between Miaotech and Aim, but they were not produced to us.  Mr Churchouse’s evidence is that the missing documents are dated 11 May.  He also said that a bank transfer (also missing from the bundle) shows that the payment took place on 26 May but that as it had the numbers from the purchase order and invoice issued on 11 May showing that despite the date discrepancies the transactions are related. 

258.         Nevertheless, although we have found Mr Churchouse to be an honest and reliable witness, we will not accept his evidence on what he recollects the documents to say for two reasons.  Firstly, the Appellant has not had sight of these documents nor did HMRC apply to admit them so we will not allow HMRC to get the evidence in, so to speak, by the back door.  Secondly, Mr Churchouse’s recollection of them may be flawed and without the documents we cannot judge this.

259.         Based on this evidence, the Tribunal only finds it proved that this purchase of CPUs by Pars was from Aim and that there is no evidence of from whom Aim purchased the CPUs. 

Deal 56 Pars’ Purchase Order Number 32977

260.         The documents in this alleged chain (M14) are all dated 10 May with the exception of All Trading’s and Pars’ purchase orders which are dated the day before, and Pars’ invoice which is dated 15 May.  All documents (except All Trading’s purchase order) appear to relate to 730 retail SL7Z9 630s.  Pars shipped the goods to All Trading on 10 May.  The price increments in the chain are of the sort evidenced in all the other proved chains in this appeal. 

261.         Apart from the date on the two purchase orders being 9 May the other slight discrepancy with these documents is that All Trading’s purchase order was for 1,000 CPUs.  Nevertheless, Pars spreadsheet evidences that the CPUs purchased from 21^st on this purchase order were sold to All Trading.

262.         Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from 21^st was connected by a chain of sales and purchases of CPUs back to Okeda and as set out in the Appendix. 

Deal 57 Pars’ Purchase Order Number 33007

263.         The documents in this single alleged chain (M15(1)/M15(2)) are all dated 12 May except for Pars’ invoice which is dated 15 May.  The goods were shipped by Pars to All Trading on 12 May.  The price increments in the chain are of the sort evidenced in all the other proved chains in this appeal. 

264.         The product description is not consistent.  The chain appears to relate to 1,000 Z9s and 500 Z8s.  However on the first three invoices the description for the Z9s refers to them as 3.2 Ghz whereas the documents issued by 21^st, Pars and All Trading all refer to 3.0 Ghz Z9s.

265.         We find that this was an error and that the entire chain dealt with the same CPUs.  We find this because of the improbability (in the context of MTIC or of a commercial transaction) of 21^st having bought on the same day from the same supplier one lot of 1,000 3.2 Ghz Z9s and 500 Z8s and a second lot comprising 1,000 3.0 Ghz Z9s and 500 Z8s.  It is considerably more likely to be an error in the paperwork and we did not understand Mr Beal to suggest otherwise.

266.         Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from 21^st was connected by a chain of sales and purchases of CPUs back to Okeda and as set out in the Appendix. 

Deal 59 Pars’ Purchase Order Number 33051

267.         The documents in this alleged chain (M17) are all dated 17 May.  The goods were shipped by Pars to Mashall on 17 May.  The price increments in the chain are of the sort evidenced in all the other proved chains in this appeal.  All documents in the chain appear to relate to SL7Z9s.

268.         There is no invoice from Sweetlime at the start of the chain but HMRC adduced in evidence a purchase order from Akorn addressed to Sweetlime.

269.         Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from Culmain was connected by a chain of sales and purchases of CPUs back to Sweetlime and as set out in the Appendix. 

Deal 60 Pars’ Purchase Order Number 33044

270.         The documents in this alleged chain (M18) are all dated 16 May with two exceptions.  The goods were shipped by Pars to Fonelink on 17 May, actually leaving the country at just past midnight on 18 May.  The price increments in the chain are of the sort evidenced in all the other proved chains in this appeal.  All documents in the chain appear relate to 8,190 (26 boxes) SL7Z9s.

271.         The date discrepancies are that Bevtex’s invoice to Delltronics, and Pars’ invoice to Fonelink are dated, respectively, 17 and 18 May.  However, Bevtex’s invoice clearly relates to Delltronic’s purchase order the previous day as it carries its reference number.  And as noted it is usual for Pars’ invoice to carry a later date to the date of the deals.

272.         There are two invoices to Time from Okeda (04763 and 04746) both dated 16 May and another which appears to be a cancellation also numbered 04763 but dated 18 May.  Both invoices relate to 8,190 Z9s but the 04746 one (the one which does not appear to have been cancelled) carries a more detailed product description.  We find the paperwork in relation to Okeda muddled but that it is more likely than not that it sold the 8190 Z9s in issue to Time on one or other of these invoices.

273.         Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from Delltronics was connected by a chain of sales and purchases of CPUs back to Okeda and as set out in the Appendix. 

Deal 61 Pars’ Purchase Order Number 33030

274.         The documents in this alleged chain (M19) are all dated 16 May with the exception of Pars’ invoice which is dated 18 May.  The goods were shipped by Pars to Inco Parts on 16 May.  The price increments in the chain are of the sort evidenced in all the other proved chains in this appeal.  All documents in the chain appear relate to retail Z8 640s.

275.         However, Leisure buys only 1,500 units from Blue Star (and the chain above Leisure relates only to 1,500 units).  Leisure sells 2,000 units to Pars.  No evidence was adduced as to where the “missing” 500 came from.

276.         Nevertheless, based on this evidence, the Tribunal finds it proved that 1,500 of the units in this purchase of 2,000 CPUs by Pars from Leisure was connected by a chain of sales and purchases of 1,500 CPUs back to Okeda and as set out in the Appendix.  The other 500 is found to be connected by sale and purchase to Leisure (and indeed we did not understand Pars to dispute this as this is clear from their own records).

Deal 62 Pars’ Purchase Order Number 33031

277.         The documents in this alleged chain (M20) are all dated 16 May apart from Pars’ invoice which is dated 18 May. The price increments in the chain are of the sort evidenced in all the other proved chains in this appeal. 

278.         All documents in the chain down to Leisure relate to 25 boxes of SL7Z9s.  The alleged chain shows Leisure buying 25 boxes but selling only 9 boxes of SL7Z9s to Pars.  However, there is also an inspection report and release instruction from Leisure to the freight forwards (Forward Logistics). It was Mr Churchouse’s evidence (which we accept) that the inspection reports in the deal packs came from Leisure’s records.  In any event this evidence is verified because Leisure’s release instruction (dated 16 May) identifies 9 boxes to be sold to Pars by reference to the item numbers on the inspection reports.  It also identifies their supplier as being Blue Star. Curiously the inspection reports show 35 boxes inspected but they are clearly the correct reports as Leisure’s release note identifies them by a unique file reference number on each report. 8 boxes identified by item number on the release note bear the same box number as in Pars’ CPU spreadsheet:  the remaining box BH099Q08 is a box on one of the inspection report but not one which (according to the release note) Forward Logistics was instructed to release to Pars.  It appears Forward Logistics made a mistake.  We find that all this evidence shows that the 9 boxes sold to Pars were part of a much larger consignment sold to Leisure by Blue Star and more likely than not it was the one for which we have the evidence.

279.         Pars sold and despatched 2 of the boxes to Alvegar in Spain.  The remaining 7 boxes were sold to Datec in the UK. 

280.         Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from Leisure was connected by a chain of sales and purchases of CPUs back to Okeda and as set out in the Appendix. 

Deal 63 Pars’ Purchase Order Number 33017

281.         The documents in this alleged chain (M21) are all dated 15 May except for Pars’ invoice which is dated 18 May.  The price increments in the chain are of the sort evidenced in all the other proved chains in this appeal.  All documents in the chain appear relate to 2000 retail SL7Z9s.  The goods were shipped by Pars to All Trading on 15 May.

282.         Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from Culmain was connected by a chain of sales and purchases of CPUs back to Sweetlime and as set out in the Appendix. 

Deal 64 Pars’ Purchase Order Number 33065

283.         The documents in this alleged chain (M20) are all dated 19 May. The price increments in the chain are of the sort evidenced in all the other proved chains in this appeal.  All documents in the chain appear to relate to 16 boxes of SL7Z9s.  The goods were shipped by Pars to Mashall on 19 May.  (Mr Beal notices that the deal pack contains a purchase order from Culmain which is clearly not the correct one for this chain: R11/2489.  R11/2490 is a copy of a different purchase order form Culmain which we do find evidences a link in the chain as for the right price, quantity and date.)

284.         The documents trace back to an invoice to Connect from Akorn and a purchase order to Akorn from Connect.  HMRC considered that they had evidence which showed that Akorn purchased from Smartview, even though there was no invoice or purchase order which appeared to be connected to this chain of 16 boxes of SL7Z9s.  The only evidence they produced of this was the internal memo of an HMRC officer called Helen to Mr Churchouse stating that this chain traces back to Smartview.  However, no evidence was produced which supported this view and the witness statement of Mr Laing who was the defaulting officer for Smartview only gave evidence that the company was known to import phones from France.

285.         Mr Laing also gave evidence that Smartview was assessed for the missing VAT at the top of this chain:  but that is not evidence that Smartview did sell the CPUs to Akorn, merely that in HMRC’s opinion Smartview sold them to Akorn.

286.         Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from Culmain was connected by a chain of sales and purchases of CPUs back to Akorn and as set out in the Appendix.  We do not find it proved that Akorn purchased from Smartview.

Deal 65 Pars’ Purchase Order Number 33056

287.         So far as Pars’ documents show this was a purchase of 1,785 retail Z9s and 1,900 retail Z8s.  The chains alleged by HMRC (M23b, M23a(1) and M23a(2)) show that although all the CPUs originated with Sweetlime, Sweetlime divided them up and sold them all to Akorn but on three separate invoices, one for 1,785 Z9s, one for 1,000 Z8s and the other for the balance of 900 Z8s.  Akorn sold to Connect, still on 3 separate invoices.  Connect then sold the three parcels to three different “buffers”, Futuristic, Emmen and Maximise.  These three buffers then sold the parcels to Culmain.  Culmain sold them on a single invoice to Pars and Pars sold them on a single invoice to All Trading

288.         All the documents down to and including Culmain’s purchase orders are dated 17 May. The price increments in the chains are of the sort evidenced in all the other proved chains in this appeal.

289.         Culmain’s invoice to Pars and Pars’ and All Trading’s purchase orders are dated 18 May.  Pars’ invoice is dated later, as normal, on 24 May.  The goods were shipped to All Trading on 18 May.

290.         Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from Culmain was connected by chains of sales and purchases of CPUs back to Sweetlime and as set out in the Appendix. 

Deal 66 Pars’ Purchase Order Number 33063

291.         The documents in this alleged chain (M24) are all dated 19 May except for Pars’ invoice which is dated 24 May.  The price increments in the chain are of the sort evidenced in all the other proved chains in this appeal.  All documents in the chain appear relate to SL7Z9s.  The goods were shipped by Pars to Inco Parts on 19 May.

292.         Based on this evidence, the Tribunal finds it proved that this purchase of CPUs by Pars from Leisure was connected by a chain of sales and purchases of CPUs back to Okeda and as set out in the Appendix. 

293.         We also note that in all the deal chains as set out above we find it proved that Pars sold the goods to the companies as we have set out in our conclusions above in relation to each deal chain.

Connection to fraudulent evasion of VAT?

294.         As stated above, HMRC must prove that Pars’ transactions were connected to fraudulent evasion of VAT.  We have looked at the 59 purchases by Pars in the immediately preceding paragraphs of this decision notice and concluded that these purchases were connected to the chains as set out above.  We find connection to fraudulent evasion of VAT is made out if HMRC in addition prove that at the start of the proved chain there was evasion of VAT.

295.         As we have said before, it is convenient to split the question of fraudulent evasion into two:  was tax due to the exchequer left unpaid, and if so, was it unpaid due to fraudulent intent?

296.         On the question of whether tax was left unpaid, Mr Beal was of the opinion that to show this HMRC would have to prove that the taxpayer who was alleged to have defaulted on payment of VAT at the top of each chain was an importer.  We do not agree with him.

297.         As explained by Mr Justice Clark in Red 12 Trading Limited [2009] EWHC (CH) at paragraphs 81-84 in classic MTIC the defaulter will be the importer.  This is because in order to make money the fraudsters rely on the VAT free status of acquisitions into the UK.  However, it is not necessary for the defaulter to actually acquire goods  (ie to “import” them into the UK from the European Union):  the fraud is as effective if the defaulter buys from someone else who is the acquirer and neither of them account for VAT.

298.         It is true that if the defaulter was not the acquirer but purchased the goods in the UK then it would normally have a VAT credit and the VAT lost to HMRC would (normally) be much less than the unaccounted for output tax.  But if the true “acquirer” was not VAT registered, or did not chose to treat the sale as subject to VAT, then the defaulter, even thought not an importer,  would nevertheless acquire the goods free of VAT and without any legal right to a VAT credit (having no proper VAT invoice).  If it then sold the goods, it would be in the same position as the true acquirer: able to disappear with 7/47ths of the sale price.  So, as Mr Justice Clark says, it is not necessary to show the defaulter actually brought the goods into the UK:

“In order to justify denial of the right to deduct input tax there must be knowing participation in a transaction connected with fraudulent evasion of the tax.  If that is established, the right is lost.  It would be inconsistent with that principle, and an unmerited boon to fraudsters, to require that the authorities prove that the defaulter was the original importer.”

299.         It follows from this that where there is proof that the defaulter was the importer in the chain this is clear proof of tax loss.  But if the alleged defaulter has entered into transactions and gone missing without accounting for that VAT in circumstances where it was likely either it or its supplier (or its supplier etc) imported the goods then this is also evidence of tax loss: it does not have to be proved that the defaulter actually was the importer. 

300.         But it seems to us that although the defaulter does not have to be the importer, we nevertheless have to be satisfied that there was an acquisition.  It is the VAT free acquisition of goods in the UK which makes the fraud lucrative.  But the other side of this coin is the fact a person defaults (in circumstances where it was clear they always intended to default) itself may well be evidence that there was an acquisition (by the defaulter or its supplier) because otherwise the default is unlikely to be sufficiently lucrative.

301.         As an example of this, take Deal 1.  Innovate, the line 1 buffer made a profit of 10p per unit and sold 3780 units.  Its VAT liability (after credit for its purchase) would have been £66.15.  Common sense suggests that it would not intentionally be a defaulter for such a small sum.  FX Drona, the alleged acquirer, failed to account for VAT on its deal in circumstances where it made no claim for a VAT credit so it probably had none:  the VAT lost to the revenue on its sale to Innovate was £53,978.

302.         In conclusion we find that defaulting on tax liability in circumstances where no claim is made for a tax credit is likely to mean (subject to other evidence) that the defaulter was the acquirer or purchased the goods without paying VAT from an acquirer, and the tax loss is the output tax on which the defaulter failed to account for VAT.

303.         To show that the loss was fraudulent, HMRC needs to show criminal intent by the alleged defaulters.  Failure to pay tax that is due is clearly not necessarily fraudulent:  businesses with the best of intents can fail.  It is fraudulent where the taxpayer intended to incur the VAT liability without meeting it.

The alleged defaulting traders

In summary, of the 59 invoices on which Pars were denied input tax we have found them to be connected by a chain of invoices & purchases orders and in some cases other evidence as described to the following alleged defaulters:

KEP  (19 invoices)

Okeda  (9 invoices)

Ability (18 invoices)

FX Drona (3 invoices)

Samson (1 invoice)

Midwest (1 invoice)

Bullfinch (2 invoices)

Sweetlime (3 invoices)

In five cases we have found that Pars’ purchase was not connected in whole, or (in two cases) in part, to an alleged defaulter.  These are:

·       Deal 64 connected back to Akorn;

·       Deal 29A connected back to Apollo;

·       Deal  55 (M13b) connected back to Aim and 100 units from Deal 14 only trace back as far as Aim; and

·       500 units in Deal 61 only connect back to Leisure.

304.         So in relation to the chains apart from these five (to which we revert later) we consider whether the persons at the start of the chain were involved in fraudulent tax losses.  We agree with HMRC that certain conduct is indicative of fraud.  This includes giving third party payment instructions to a purchaser or charging VAT and then failing to make a return in respect of it.  We explain our views below.

Charging but not returning VAT

305.         Mr Beal put the point that failing to pay VAT that is due is not necessarily indicative of fraud:  it may merely indicate that the person has run out of money.  He also considers failing to appeal or pay an assessment is not necessarily indicative of fraud: as with non-payment of VAT it may only indicate that the person accepts liability but does not have the money to pay it.

306.         It is fraud where the person incurred the liability without intending to pay it.  In our view, a failure to make VAT returns showing the liability to HMRC is indicative of fraud.  An honest but insolvent person ought to have made a return of his liability even if now he cannot pay it.  Where the failure to make a return is coupled with clear evidence that earlier the person issued invoices charging VAT to third parties, showing that he knew he was liable to account for VAT on that transaction, it seems to us clear evidence (if there are no contra indications) of fraud.

Third party payments

307.         We find that some of the chains involved the payment of third parties. This feature was limited to the Pattern A chains (as explained below) and the evidence did not apply to all pattern A chains.  We find that third party payment instructions were given by either or both the defaulter and line 1 buffer in some 27 of these 32 chains.

308.         The third party payment instructions would instruct the buyer to pay the gross sum due on the invoice largely to someone other than the vendor.

309.         One beneficiary of the third party payment was invariably OPM, although its share was always a small percentage.  The main beneficiary was always one of four companies:

·       TMI

·       Adax

·       Bruins

·       ASAP

310.         The third and last beneficiary of the payment instructions would be the vendor itself.  Where the third party payment instruction was issued by the defaulter, the sum instructed to be paid to the defaulter would be a small amount, significantly lower than the VAT due on the transaction.  It would be approximately 0.5% of the total invoiced sum or about 2% of the VAT due.  Where the evidence included the third party instructions given by the line 1 buffers as well as those given by the defaulter in the same chain, it can be seen that the line 1 buffer duplicated the instructions from the defaulter to pay the third parties.  In other words, they discharged their own liability to their supplier (the defaulter) by instructing their buyers to make the same payments to the third parties which the defaulter had directed the line 1 buffer to make.

311.         The invoiced sum by the line 1 buffer would of course be slightly higher because the line 1 buffers in this appeal always made a small profit.  Therefore, the amount due to the line 1 buffer would be higher than that due to the defaulter and this can be seen from the payment instructions where the payer is instructed to pay a larger sum to the line 1 buffer, than the line 1 buffer was instructed to pay to the defaulter.  This small sum would have been enough for the line 1 buffer to pay the sum owed to the defaulter, and to pay the VAT due on its transaction and have a small sum left over.

312.          The reason why it would have been enough for the line 1 buffer to discharge its own VAT liability is that the profit margin on its sale was invariably small, often no more than 5p per unit.  As it was not an importer it would be entitled (subject to the ruling of the ECJ in Kittel) to offset the input tax on its purchase from the defaulter against its output tax on its sale to the line 2 buffer.  That the line 1 buffer issues third party instructions is not therefore per se evidence of fraud by the line 1 buffer:  it is not out of funds to pay its small VAT liability (putting to one side the issue that  Kittel might mean its VAT liability was much larger).

313.         However, we find that where we have evidence of third party payments by both defaulter and line 1 buffer in the same chain, the latter duplicates the former as explained above.  We therefore infer on the balance of probabilities that where there is only documentary evidence of third party payment instructions from the line 1 buffer that there would also have been third party payment instructions from the defaulter in that chain to the line 1 buffer.  This is because the evidence is that the line 1 buffers issued such instructions where they had to discharge their own liabilities under such instructions from the defaulter.

314.         We find that the issue of third party payment instructions by the defaulter is evidence of fraud. Where we have found that the alleged defaulter defaulted, because they acquired the goods without VAT they would have had no input vat to offset the output tax and their VAT liability on each transaction would have been 17.5% of the net sale price.  By causing to be paid away over 99% of the sales monies they were depriving themselves of the ability to pay the VAT on that transaction (which would have been 7/47ths of the sum paid).  We also note that Mr Beal (not surprisingly) conceded that issuing third party payment instructions was evidence of fraud.

315.         We find that issuing third party payment instructions such that the issuer was deprived of funds to pay its VAT liability indicates an intention from the first not to pay the VAT.  Such an intention is fraudulent.

Findings made by other Tribunals

316.         HMRC pointed out that various of the alleged defaulters in this appeal have been found to have acted fraudulently as acquirers in MTIC frauds in other tribunal decisions, and asked us to the draw the conclusion that these defaulters acted as fraudulent acquirers in the chains in this appeal too.  Mr Beal objected to such conclusions being drawn.

317.         Mr Beal also wanted us to draw conclusions from findings of fact made by other tribunals.  For instance, he wanted us to accept the conclusion of the Tribunal in Red 12 that there was no evidence of importation by Samson, and the finding in Euro Stock that UK Comms were a defaulter (and not as alleged in this case a line 1 buffer).

318.         A tribunal comes to its conclusions on the evidence in front of it.  Although a Tribunal has great latitude in what it can admit as evidence, evidence comprising  the findings of another tribunal, rather than the evidence which caused them to reach those findings, is unlikely to be of any great probative weight in another tribunal hearing.  We have not given weight to these findings of fact made in other tribunals.

400.         Mr Beal also considers that there could be some sort of half way house where the defaulter organises the sale to the first 1 or 2 buffers but no more, the subsequent sales all taking place on the grey market.  But we find orchestration of a chain of 3 or 4 or 5 buffers makes no sense other than in the context of the absence of a genuine market.  The effect of long chains of buffers is to make it hard to trace the sale by the broker to a default earlier in the chain.  This does not protect the defaulter:  he needs no protection as he is planning to go missing. Rather it protects the position of the broker by making it hard for HMRC to link its purchase and sale with a default earlier in the chain:  and the defaulter would only be concerned about the position of the broker if he intended to route goods to the same broker again.  And he would only be concerned about the route of the deal chain if there was no genuine market on which to sell the goods.

401.         We find it is highly improbable that a fraudster would organise 3 to 4 buffers sales but no more, leaving further sales to be on the open market.  On the contrary we consider the evidence of 3 to 4 sales being orchestrated is evidence that this is organised MTIC fraud at every step and there was no genuine market for the goods.

402.         Another reason Mr Beal said that the frauds were not orchestrated was because we should “follow the money”: He said it was wrong to compare Pars’ profits to the buffers’ (much smaller) profits but instead note that the VAT is stolen by the defaulters and paid out to third parties such as Bruins and not to Pars.  We consider Pars’ profit as explained below in paragraph 719.

403.         The third reason he said we should not infer the frauds were orchestrated is that it depends on hindsight of matters which were not known to his client in 2006.  This is not right:  it is muddling two entirely separate questions.  It is entirely proper for the Tribunal to use evidence which may not have been known to Pars in 2006 to decide with hindsight from our vantage point of 2010 whether the deals were orchestrated in 2006.  What we cannot and do not do is assume that Pars knew something in 2006 just because it is known to us in 2010.

404.         The fourth reason he said it was unlikely the deals were orchestrated was because (he said) that would make Pars a “complicit friend” to the fraudsters, yet it would involve the fraudster exposing Pars to the risk of VAT not being repaid.  We don’t agree with this argument:  if Pars were a complicit friend as postulated hypothetically by Mr Beal (and we make no decision on this as complicity in the fraud was not alleged, merely knowledge of the fraud) it may be the fraudsters assumed that the fraud would not be challenged and Pars’ refund would be paid.  We certainly would not be dissuaded from finding orchestration because it means inferring that the person committing the criminal offence of fraud was content to leave his buffers and brokers, whether they were his conspirators, friends or innocent dupes, at financial risk.  Such behaviour is only too likely from a dishonest person.

405.         The fifth reason Mr Beal said that we could not find orchestration was that the scheme only operated for the two months April to May 2006 yet Pars’ CPU trading before and since that date has not been challenged by HMRC.  It is Pars’ case, he says, that the earlier deals were indistinguishable from the later.  Therefore, says Mr Beal, if the earlier deals were not connected to orchestrated fraud, why were those in April and May 2006?

406.         We are not prepared to infer from the fact that HMRC have not challenged the earlier deals that it necessarily follows that there was nothing to challenge:  HMRC may have made a decision based on policy or for some other reason.  We do not know nor is it the case even if we did know the reason we would necessarily agree with it.  We will consider whether the deals at issue in front of this Tribunal were orchestrated on the basis of the evidence which is in front of this Tribunal.

407.         Mr Beal’s sixth point is that we cannot (he says) find orchestration because HMRC has not alleged that the buffers were knowing parties to the fraud.  This has, he says, deprived his clients the chance of calling their suppliers as witnesses. We find, however that it was part of HMRC’s Statement of Case that the deals were orchestrated as they say in terms: “Appellant knew perfectly well that its suppliers and customers would not fail in their obligations because the transactions had all been pre-arranged”.  This allegation was elaborated upon in Mr Churchouse’s witness statement.  As we have said in paragraph 58, it seems to us that it was Pars’ choice not to call its suppliers:  it knew HMRC’s case on this.

408.         Mr Beal’s seventh point is that if HMRC are right and the deals were orchestrated, this means that not only were all the deals involving Pars orchestrated but also all the other broker deals resulting from the original acquisition of goods by the defaulter.  This he thinks would be an absolutely massive fraud and inherently unlikely.

409.         We are unable to agree.  Sadly for the exchequer of this country, it is clear that the frauds were absolutely massive.  The assessments referred to above against the defaulters exceed £300million.  We presume for that level of return, the fraudsters were prepared to engineer a scheme of some complexity. 

410.         Further it is clear to us from the evidence, that the fraud involves far more than the chains connected to Pars.  In any event, the logic of MTIC dictates that this is the case.  Defaulters will want to acquire as much as possible to maximise their VAT take:  but brokers (innocent or otherwise) will be limited in how much they can purchase because their capital will be tied up in VAT reclaims.  It is therefore the logic of MTIC that for each defaulter there will be a large number of brokers.  So we do not agree with Mr Beal on this.

411.         So we move on to consider the evidence in this case of whether the frauds were orchestrated in the sense that the fraudsters organised every buy and sell in the chain.  To do this, firstly we consider the evidence of whether Pars’ trading was on the grey market.

Grey market

412.         Mr Kazeroonian gave evidence in his witness statement on the operation of the grey market in CPUs.  This was not challenged by HMRC.  Mr Stone gave evidence in the form of a witness statement which covered briefly the operation of the grey market in mobile phones.  This was also not challenged by Mr Beal other than to question its relevance in an appeal concerning CPUs.  Indeed, we have relied on Mr Kazeroonian’s explanation of the grey market in CPUs.  Although, as we explain below, we did not find him to be a reliable witness on some issues, his explanation of the grey market in which Pars has experience (as described below) was logical and consistent with Mr Stone’s acceptance that a genuine (if small) grey market exists for mobile phones.  Mr Beal says that on the grey market Mr Kazeroonian’s evidence was unchallenged, extensive and commercially logical and we agree.

413.         Our summary of Mr Kazeroonian’s detailed explanation of the grey market in CPUs is as follows. As explained in paragraph 3, OEMs are original equipment manufacturers which means they buy computer components and assembles computer equipment. Tier 1 OEMs (ie large OEMs) get good deals from manufacturers (to whom they are tied) but have to buy in large volume and pre-order. They are not authorised distributors and their contract requires them to use the CPUs in their OEM business and not resell them. Manufacturers also sell direct to authorised distributors but at higher prices.  Authorised distributers will sell to smaller OEMs who don’t have direct contracts with the manufacturer. Tier 1 OEMs are often left with unused CPUs because to get the price discount they have been forced to buy more than they need for their OEM business.

414.         When the OEMs are left with surplus stock, in breach of their agreement with the manufacturer, they will seek to offload this stock at a profit.  The price at which they offer the stock is likely to be lower than the price that authorised distributors (“ADs”) can offer because the manufacturer sells at a lower price to OEMs than ADs.

415.         It is called a grey market because the market is not illegal but in selling the stock on the open market the OEMs are breaching their restrictive contract with the manufacturer.  It is important to the OEMs that they sell discreetly because the manufacturer might be forced to take action against them if they openly sold the stock.  If a manufacturer was seen to allow an OEM to undercut ADs, the manufacturer would find it hard to maintain the higher price at which it sells the CPUs to its ADs.

416.         Tier 2 OEMs might well chose to buy CPUs on the grey market rather than from an authorised distributor because either or both the stock would be immediately available and the price would be less than an AD would charge.

417.         Mr Kazeroonian gave evidence of grey market trading in which Pars had been involved as described in paragraph 4 of this decision notice. His description of Pars’ involvement in 1995 in the grey market was that it was seller led and he had to work hard to find buyers.

418.         Mr Kazeroonian also explained that as an OEM Pars had surplus stock which they sold into the grey market.  To avoid selling into the ADs’ distribution area they sold abroad but even so had a difficult relationship with the manufacturer and when they ceased this practice in February 2005 they were given a better price by the manufacturer.

419.         More recently, some time after the deals in this appeal Pars has bought CPUs on what Mr Kazeroonian describes as the grey market although the vendors appear to be ADs.  Pars has onward sold the goods to a company in Russia.  He says that Pars will be offered deals by ADs at a good price but will often turn them down as Pars would have to agree to buy more stock than it could shift.  He says:  “many of these opportunities are not straightforward and will take a degree of negotiation before a deal is struck”.  We also find that the evidence is that the turnover from these deals is considerably lower than the deals at issue in this appeal.

420.         Mr Kazeroonian’s description of the grey market makes sense.  Companies are forced to buy in large quantities to get price discounts.  When they can’t use the entire stock purchased, they sell it for as much as they can get, even if this means selling it discreetly in breach of their contract with their supplier.  Companies buying it will aim to make a profit and may be able to do this by buying the stock speculatively, hoping to find buyers, and/or by selling it on in smaller parcels.

421.         Based on the evidence of Mr Kazeroonian, the features of the grey market which we identify are:

·       The chain will trace back to an OEM (or possibly an AD);

·       The price may well be lower than the price an AD would charge;

·       The sale may be high volume and low margin (as the trades in the appeal);

·       The purchaser is likely to break bulk and/or buy speculatively:  it will have to buy in large quantities but likely to only be able to sell onwards in smaller quantities. 

·       It is not a commodities market;

·       It is counter-intuitive that a chain on the grey market would be long.  The longer the chain the less money to be made. 

·       There will be risk.  The risk may be because the person buying from the OEM will need to buy in bulk and then locate buyers.  There is no reason why the deals would be back to back:  indeed Mr Kazeroonian’s evidence that the grey market was a source of “just in time” products indicates that the vendors were buying speculatively and holding stock.  This is consistent with his description of how OEMs would dump stock and inconsistent with all deals being back to back.  This means that traders in a grey market are at risk of not being able to sell all the stock they buy and/or fluctuations in price.

422.         It was originally part of HMRC’s case that Pars must have known that the deals at issue in this appeal were part of an MTIC fraud because the price was below the price an AD would charge.  However, the low price would be a feature of grey market trading and by itself would not indicate that there was anything unusual about the transaction.  At the hearing HMRC did not pursue the point and we do not find that buying and selling below the authorised distributor’s price indicated or should have been taken to indicate that the transactions were fraudulent.

423.         The chains found proved in this appeal exhibit very few of the features of a true grey market.  No chain traces back to an OEM (or even authorised distributor or manufacturer).  Most in fact we have found proved to originate with a defaulter without any genuine business.  The chains appear to involve no risk:  the vast majority are back to back or with only a day or two at most between purchase and sale.  They rarely involve breaking bulk: most deals in the chain are for the same quantity.  In many cases where there are splits they appear to be artificial, such as in deal 65 where the load is split only to be recombined.  In the majority of cases where Pars (or the buffer above it) buys a split load, the two chains originated with the same defaulter.  The profits are (for the buffers) very small and regular.  And there is always a profit.  We do not find that Pars’ purchases and sales were trades on the grey market.  We go on to consider whether they were orchestrated.

Retention of title clauses

424.         It was part of HMRC’s case that many of the companies in the chain including Pars itself imposed retention of title clauses.  In other words it was a term of trading that title remained with the vendor until payment was made for the goods (an example is the invoices from DDM and Qiass in deal 19).  Yet it was also Pars’ case that the goods would not be paid for until delivered to the purchaser’s customer.  HMRC’s point was that this showed the chains were not on commercial terms because the clauses had no meaning:  a company could not retain (or pass) title to goods which it itself did not have because it had not paid for the goods. 

425.         Mr Beal disagreed.  Title could be passed despite a retention of title clause if the purchaser was able to pass possession of the goods to its purchaser.  In his view it was not commercial nonsense.

426.         Our conclusion is that HMRC have not made out their case on this point.  We agree that with back to back deals, if payment is also back to back, a retention of title  clause is meaningless.  But for HMRC to use the existence of these clauses as evidence that the deals were contrived and not on the open market, HMRC would have to show us that all payments by the buffers were also back to back and that it is not normal for commercial contracts under which large quantities of CPUs are sold to have meaningless retention of title clauses.  HMRC did not lead this sort of evidence.  On the contrary, it appears quite possible to us that a retention of title clause is a standard clause inserted into many commercial sales contracts without much thought being given to its enforceability.

Back to back deals

427.         The deals were almost invariably back to back in that they took place on the same day.  As explained above, Pars’ invoice was often not dated the same day as other documents in the chain but the day on which the deal took place was reflected in the day that Pars despatched or exported the goods:  in almost all chains that was the same day as the documentation issued by the other parties in the chains.

428.         Of the 66 chains in all but 15 the documentation at every stage of the chain was dated the same (with the exception of Pars’ invoice as outlined above).  Of the 15 where this was not so, in all but 3 of the chains the date variation was very minor (and outlined in our analysis of the deal chains), such as a purchase order dated a day before or afterwards.  It was nevertheless clear that the deal took place in a single day, or at most two days. 

429.         The back to back chains indicate a lack of commercial rationale.  It shows no one appeared to be taking any risk of being left with stock that they could not sell. 

Knowledge or means of knowledge

508.         As mentioned at the start of this decision notice, the Tribunal is here to determine whether Pars, having regard to objective factors, knew or should have known that by making the purchases which are at issue in this appeal it was participating in transactions which were connected with the fraudulent evasion of VAT.  We have concluded that all its transactions at issue in this appeal were connected with the fraudulent evasion of VAT.  Did Pars or should Pars have known this?

509.         We agree with the Appellant and its counsel that to answer this question this Tribunal is not concerned with hindsight.  The question is what Pars knew or ought to have known at the time of the disputed deals in 2006.

510.         HMRC advanced its case as one of knowledge:  their secondary case was that Pars ought to have known of the connection to fraud. We dealt with the issue of whether Pars were given sufficient notice of HMRC’s case that they had actual knowledge of the connection to fraud at the start of this decision notice and concluded that they were:  see paragraphs 37-59.

511.         We start by considering (as it was put in issue) the credibility of Pars’ three witnesses, Mr Kazeroonian (director and principle shareholder), Mr Hamakarim (general manager) and Mr Doherty (financial controller).

Control of the company

512.         We find that the Pars’ purchases at issue in this appeal were put in place on a day-by day basis by Mr Hamakarim.  This is clear not only from the evidence of all three witnesses for Pars, but also from the many Pars’ documents in the appeal bearing Mr Hamakarim’s name.  Mr Kazeroonian said he trusted his employees:  Mr Hamakarim to negotiate the deals and Mr Doherty to deal with the finances. He gave them the guidelines in how to carry out the CPU trading but left it to them on a day to day basis.  He spent most of his time dealing with Pars’ core business as an OEM.  We find that he would expect his employees to report back to him. This is not surprising: he owned and controlled the company.  Bearing in mind that he was (as owner) the one who benefited from the CPU trading, we find that he was well aware of the nature of Pars’ wholesale CPU trading and effectively took the decisions even though he did not oversee each individual deal. 

Mr Kazeroonian

513.         Mr Kazeroonian set up Pars from scratch.  Prior to setting up his own business he worked in other companies gaining experience in building and repairing computers.  Pars was a successful business with about 25 employees.  Nevertheless, we find Mr Kazeroonian’s profits were fairly modest until the company began to trade in wholesale CPUs.  In 2003 his salary was approximately £60,000.  In 2004 (wholesale trading in CPUs having commenced the previous year) it increases to £221,000.  In 2005 it increases to £350,000 and in 2006 to £1.28million.

514.         As a witness under cross examination we found he often avoided giving straight answers but used cross examination as a platform to put across his view (somewhat repetitively) that it was HMRC’s responsibility to prevent MTIC fraud and that Pars had done what it could to avoid being involved in fraud. When asked about business practices in the context of the CPU trading, he would often give a reply that related to Pars’ core business as an OEM even though he was clearly aware that the case was about the CPU trading.  He clearly disliked being asked questions which he appeared to think would show him in a bad light:  when asked about his profits (set out in the preceding paragraph) he first doubted the figures that were put to him (despite these not ever having been in dispute and being apparent from the Company’s accounts) and then emphasised he was liable to income tax on the sum.

515.         We found him to be an evasive and defensive witness.

516.         We note that his evidence contained some self-serving contradictions.  His witness statement dated 16 October 2008 explained in some detail how HMRC’s action in denying input tax had forced Pars to make six staff redundant and how this had been a “considerable upset” for Mr Kazeroonian.  Yet under cross examination in 2009, somewhat heated as he was being accused of dishonesty, he replied “…if I was a dishonest person…the first thing they would do is to make everybody redundant…I did not make anybody redundant due to this.  We worked even harder to secure our people, our employees’ jobs.  Yes, there were people who left from our business.  We didn’t make them redundant…we just did not replace people that they left….”  His counsel gave him the opportunity to clarify the apparent inconsistency with his witness statement in re-examination.  His explanation was that Pars “must” have made people redundant but they were not from the core part of the business.  This explanation left us unclear whether or not Pars had made anyone redundant but quite clear that Mr Kazeroonian had given untrue evidence on this issue with the purpose of showing himself in a good light.  We did not think the contradiction could be explained by a lapse of time as Mr Kazeroonian would certainly remember in 2008 or 2009 whether or not Pars had made people redundant in 2006 particularly as he said the issue was so important to him. 

517.         As with Mr Hamakarim and Mr Doherty, Mr Kazeroonian denied that he knew much about MTIC fraud in 2006.  For the reasons explained below when we consider Pars’ knowledge of MTIC in 2006, we do not accept this.  He also repeated Mr Doherty’s statement that he thought it was like credit card fraud.  We don’t accept he really thought this. 

518.         We concluded that Mr Kazeroonian was not a reliable witness and anything he said should be treated with caution if not capable of verification.  We did accept Mr Kazeroonian’s evidence about the grey market in general:  it made commercial sense and was not challenged by HMRC.

Mr Hamakarim

519.         Mr Hamakarim was responsible for the day to day running of Pars’ CPU wholesale business.

520.         Inspection reports

521.         His original witness statement was dated 17 October 2008. In this he explains he was responsible for checking the goods to be purchased and that therefore he always requested an inspection report from the freight forwarders before finalising a deal.  He was also aware that some of the reports showed that trays had been damaged but explains that he considered that the damage was only to the packaging and not the CPUs themselves.

522.          He produced a second witness statement during the course of the hearing to deal with the question of inspection reports in more detail as HMRC had indicated in their opening that they would be cross examining on why so few inspection reports had been disclosed by Pars.  In this he explained his practice of having the inspection report faxed to him by the freight forwarders so he could read it and use it as the source of the box numbers so Pars could check that it was not buying the same goods twice.  He then gave evidence that his practice was to throw away the inspection reports as they were no longer required and this explained why Pars were able to produce so few.

523.         He said that the reports were no longer required because the faxed copy of the inspection report on which he made the decision whether or not to buy the CPUs was really a duplicate copy and unnecessary as the original hard copy would arrive in the post shortly afterwards together with the freight forwarder’s bill for carrying out the inspection. 

524.         In fact we find (as Pars accept) that the bill would not arrive for a month or two, would often arrive without the report attached or arrive with the wrong report attached.  In fact, so few inspection reports were held by Pars that during the hearing they instructed their solicitors to go to the freight forwarders to retrieve copies of them.  By the end of the hearing from having being able to produce only a few reports, Pars was able to produce to the tribunal reports for about 80% of the deals. 

525.         It was put to Mr Hamakarim that his evidence was untruthful and that in reality he never received a faxed copy of an inspection report.  He denied this. He said his story was backed up because Pars had to get the box numbers of the wholesale CPUs to enter into their CPU spreadsheet to ensure that they did not trade in the same goods twice:  if the box numbers did not come from the inspection reports from where did they come? 

526.         It was pointed out to him that some of his instructions to freight forwarders, which included instructions to inspect the goods, included the box numbers.  This would indicate that Pars had the box numbers before they instructed the freight forwarders to inspect the goods. 

527.         Mr Hamakarim denied this:  he said that the instructions to freight forwarders were merely a template document on the computer and sometimes he forgot to delete the instructions to inspect even though the inspection had already taken place.  However, we find that the document was not a template in that the contents and typeface vary from deal to deal.  And some of these instructions (as they include instructions to inspect) are inconsistent with his statement that the inspection reports were the source of the box numbers. 

528.         He also made the point that the box numbers in the spreadsheet (Excel) were merely “copy & paste” into the instructions (Word) to the freight forwarders thus showing (he said) that the former predated the latter.  We agree (and HMRC did not dispute) that on the evidence of the typeface and in particular because a formatting error was repeated that one was a “copy & paste” of the other.  But that does not tell us which way around this happened:  we find it as easy to “copy & paste” from Word to Excel as from Excel to Word.  In any event, even if the spreadsheet was completed before the instructions, that does not tell the Tribunal that the inspection report must have been the source of the box numbers.

529.         We accept that he may on some occasions have made a mistake and left in instructions to carry out an inspection which had already been carried out, but even if this happened it does not show the tribunal that the inspection reports had to have been the source of the box numbers.  Mr Hamakarim said that Pars did not know the box numbers before it received the inspection report as it relied on merely instructing the freight forwarder to inspect the boxes as allocated to Pars by the vendor. 

530.         We find that this evidence about the freight forwarder instructions does not resolve the issue of whether Pars received the inspection reports either way so we consider the destruction of the inspection reports themselves. 

531.         We find his story of throwing away inspection reports quite improbable.  On the premise that Pars’ was, as its witnesses said, unaware that the deals were orchestrated, Mr Hamakarim must have been concerned about the condition of the CPUs and ensuring that his buyer was happy with the condition of the goods, or, if the goods were damaged tracking down when and where they were damaged and whose fault it was.  Therefore, Pars must have wanted a copy of the inspection report to hand in case the buyer in Europe had any queries about the condition of the goods as they left Pars’ control.  Apart from this it seems extremely unlikely a businessman would choose to destroy any report for which the business had paid and particularly unlikely he would destroy a report that they might need to refer to in the next day or so and was evidence they had had the goods inspected, a check which HMRC had recommended (in Notice 726) that they undertake to protect themselves from MTIC fraud.  Pars maintained files of due diligence on suppliers:  it seems extremely improbable that they would have destroyed inspection reports relating to the goods.

532.         We note Mr Hamakarim’s explanation that it was a duplicate in the sense another would be arriving.  But we do not believe he would have destroyed the faxed copy without having another copy to hand.  He says he did not know it took 6-8 weeks for the second copy to arrive:  but we cannot believe he, if genuinely unaware of the fraud, would destroy a faxed copy without ensuring that he had another copy to hand.

533.         This is especially the case where the inspection reports themselves indicated quite a lot of damage at least to the packaging.  Although Mr Hamakarim’s evidence is that on two occasions they had some boxes delivered to Pars and found the CPUs to be undamaged, nevertheless any reasonable person (who did not know the buyer was not interested in the condition of the goods) would have kept what evidence they had of the state in which the goods were despatched. 

534.         We also note that Mr Hamakarim, although he said he read them to decide whether or not to buy the goods on the basis of the condition the report disclosed, he also  said he did not notice indicators that the inspection was not carried out on behalf of Pars.  In particular, we find approximately half of the reports were dated earlier than the date of the deal and/or related to more boxes than Pars was buying.  It was obvious to the Tribunal where the reports related to more boxes than Pars was buying as (a) there would be lines of box numbers deleted or typex’d out and (b) the report would show a greater number of boxes being inspected.  Further, because some of the reports contained some general comments such as stating the number of boxes with knife marks, as these did not relate to specific boxes it was not apparent whether the comments applied to the boxes Pars was buying or other boxes inspected.

535.         We think that a reasonable person who was paying for an inspection to take place on very valuable goods and  who did not know that the deals were orchestrated, would have noticed and been concerned by the fact many of the inspection reports carried indicators that the inspection was not carried out on Pars’ behalf.  This is yet one more factor which suggests to us that Mr Hamakarim did not read the reports before committing Pars to the deals because he had not noticed this.

536.         We also note that prior to the hearing there was no hint from Mr Hamakarim that throwing away inspection reports was part of his practice despite his lengthy witness statement in which he said “we were in regular contact with the staff at the forwarding company regarding the physical condition and status of the goods to make sure that the goods conformed to the suppliers’ description”.

537.         In conclusion, we do not accept that Mr Hamakarim threw away the inspection reports. We believe this was an untrue explanation he gave to explain why Pars had so few copies of inspection report when the real explanation was that Pars had never received them (apart from a few copies received with the freight forwarder’s invoice some weeks after the deal). 

538.         We note that even if he was telling the truth (and we do not find that he was) this would mean that he knew he did not need to keep the inspection report even though the goods were in transit and would arrive in the next day or so and be inspected by Pars’ buyer.  This would indicate that he did not expect the buyer to query the condition of the goods:  this would indicate to us that he knew the buyer was not interested in the condition of the goods because it was a deal orchestrated for the purpose of fraud.

Knowledge of MTIC fraud

539.          We also did not accept what Mr Hamakarim said about his knowledge of MTIC fraud.  His position was that he had a general awareness of MTIC fraud in CPUs but “certainly did not know the extent of the wider problem of MTIC fraud at the time I negotiated the deals that are the subject of this appeal”.  He denied understanding that a letter written by HMRC told him categorically that some of his deals had been traced back to a defaulter.  He also repeated what Mr Doherty (who gave evidence first) had said that he saw MTIC fraud as being similar to credit card fraud, implying that he had not understood it to be a particular risk to Pars.  He said Pars knew MTIC fraud existed but did not consider themselves at risk because they knew their suppliers and did their due diligence on them.

540.         This evidence was quite inconsistent with the evidence of what happened in 2006.  Pars undertook due diligence:  this included meeting and photographing their new business associates and asking them to complete suppliers’ questionnaires specifically aimed at MTIC fraud risk.  They kept due diligence files and claimed to comply with Notice 726.  HMRC visited once a month where from the officers’ notes (which Pars did not dispute) it appears MTIC fraud was clearly the concern.

541.         We find that at the time of the deals in question Pars was well aware of the risk of MTIC fraud in CPU trading and knew it was of far more immediate and serious concern to them than credit card fraud.  As Mr Hamakarim suggested otherwise in his oral evidence we find him to be less than a reliable witness.

542.         We deal below with Mr Hamakarim’s evidence on how the deals were negotiated.  We also did not believe this evidence for the reasons as explained below.

543.         In conclusion, we did not find Mr Hamakarim to be a reliable witness. We do not accept his evidence that he destroyed the inspection reports nor that he did not understand how relevant to Pars the risk of MTIC fraud was. It was put to him that his evidence on how the deals were negotiated was preposterous and that in respect of the inspection reports he was making it up.  He denied this but we do not believe him. 

Mr Doherty

544.         Mr Doherty was Pars’ financial controller.  He joined the company in November 1999.  He dealt with all financial matters including VAT and was the Pars’ officer who met with HMRC officers when they visited. 

545.         We found him to be a defensive witness under cross-examination:  although he gave relatively short and to the point answers to questions from his own counsel, under cross examination he gave extremely long-winded answers that would only sometimes eventually answer the question.

546.         Mr Doherty’s repeated view was that all Pars could reasonably in law be required to do was check on their own suppliers.  It was not for Pars to effectively police the chain and look beyond what it could obviously see.  And in his view it was for HMRC to tell them if there was anything wrong with their due diligence.

547.         Mr Doherty was asked to explain why he did not question how Pars was able to make such large profits for so little outlay.  Mr Doherty’s response was that they took the risk of not being able to sell on at a reasonable profit yet Mr Doherty was clearly aware that Pars did not agree to buy if they did not have a buyer lined up. He ultimately agreed that Pars’ expenses were limited to freight, insurance and inspection costs and the financial expense of being out of pocket on the VAT for a few months until HMRC met the VAT repayment.

548.         We found in his evidence he chose to play down what he had understood of the risks of MTIC fraud in 2006.  He said that in 2006 he understood that HMRC’s primary reason in visiting Pars was to check that they held the proper proof of export.  He said that they understood that MTIC fraud was mostly in mobile phones and that to protect themselves they only needed to ensure that they did not deal in the same goods twice (which was why he devised the CPU spreadsheet already mentioned). He considered (he said) the risk to Pars of MTIC fraud to be on a par with the risk from credit card fraud. He denied that he had understood from the letter of 15 November 2004, that HMRC had found fraud higher in chains involving Pars although he later agreed he had understood the risk of fraud higher in chains from reading Notice 726.

549.         Later in the cross examination, the emphasis in his evidence changed.  He agreed that from March 2003 Pars knew MTIC fraud affected high value CPUs but said Pars was protected because HMRC visited them once a month and never told them their due diligence was insufficient.  We find that Mr Doherty was well aware that MTIC fraud was such a concern to HMRC that they were visiting Pars once a month and reviewing their due diligence.  Mr Doherty himself at the time was clearly concerned as he repeatedly asked for HMRC to effectively sign off on their due diligence (which HMRC refused to do).

550.         Overall our impression was of a witness who wished to deny knowing anything very much about MTIC fraud but as the cross examination developed and more things were presented to him, gradually came to agree that he was well aware of the very real risk to Pars of MTIC fraud in 2006. 

551.         He said he was aware of Mr Hamakarim’s practice of throwing away inspection reports but he only said this after Mr Hamakarim’s second witness statement was put in evidence. Further, Mr Doherty said what we find must be true, which was that he was aware that the freight forwarder invoices did not come to Pars for 6 to 8 weeks after the deal and often did not contain the inspection report.  If the parties were genuine and if, as Mr Doherty says, he knew Mr Hamakarim did not keep the faxed copies of the report because it merely duplicated the hard copy that would come with the bills, it seems extremely improbable to us that Mr Doherty would not have made Mr Hamakarim aware of the position and caused him to revise his practice the result of which was that Pars had very few inspection reports. For these reasons and because we have found that Mr Hamarkarim did not throw away the inspection reports, we therefore do not accept that Mr Doherty is being truthful when he says he was aware of Mr Hamakarim’s practice of throwing away the reports.

552.         And as we have already said, if we did accept (which we don’t) Mr Doherty’s and Mr Hamakarim’s evidence, this would mean that they both knew that the inspection reports were unnecessary to be kept because they knew that their buyer would not query the condition of the goods.

553.         We did not find Mr Doherty to be a reliable witness. 

554.         We consider whether Pars (by its director and its two chief officers) knew or ought to have known that all its wholesale CPU deals in the two months in issue were (as we have found) connected to fraudulent loss of tax.  We look at this under three main headings:

·       What was Pars’ general knowledge of the risks of MTIC fraud at the time of the deals in question?

·       What did they do to safeguard Pars from these risks?

·       What did or should they have known from the nature of the deals themselves?

General knowledge of MTIC

555.         As mentioned above we found that Mr Doherty, Mr Kazeroonian and Mr Hamakarim in evidence played down the extent to which they were aware at the time of the deals in issue in this appeal of MTIC fraud.  We found this partly because they admitted to knowing more as matters were presented to them in their cross examination but also because of the evidence of what they actually did know at the time.

Notice 726

556.         Notice 726 is actually concerned with legislation that could give traders joint and several liability should their transactions be connected with fraud:  it does not deal with the “knowledge or means of knowledge” test from European case law.  Nevertheless, it explains the risks of MTIC fraud which it describes as “widespread” in the industry in which Pars operated and recommended checks to prevent involvement.

557.         It states on more than one occasion that a trader must verify its supply chain but the recommended checks in the notice only verify the trader’s supplier.  The notice itself recognises that a trader can not be expected to discover the identity of its supplier’s supplier.

558.         Mr Beal points out that Notice 726 is not legally binding.  We entirely agree.  However, this does not seem relevant to the issue of knowledge or means of knowledge.  HMRC’s case is that Notice 726 put Pars on notice of both the risks in what they were doing and of checks they could undertake to mitigate those risks.  If a trader did not carry out those checks, the question is why not, and what would have been revealed if the checks had been carried out?

559.         Mr Beal also says Notice 726 puts a disproportionate and unreasonable requirement on traders and may be in breach of EU law.  Yet this could only be relevant to the extent Pars did not carry out checks which HMRC recommended.  Evidence from Pars officers was that they did the checks that they considered reasonable and not those that they did not.  We deal with this below but suffice it to say we do not consider that credit checks and trade references (which HMRC recommended and Pars did not undertake) to be a disproportionate and unreasonable requirement on traders:  both seem to be of the type that a business might well undertake to assess the bona fides and financial status of a trading partner in any event even where MTIC fraud is not a known risk.

560.         HMRC had issued Pars with Notice 726 (and its reissue) before the deals in this appeal and both Mr Doherty and Mr Hamakarim agree that they were familiar with Notice 726 from 2005 if not earlier.  They understood from it that the fraud was widespread in the computer industry.  The notice says as much.

561.         It was Mr Doherty’s evidence that if he asked HMRC to confirm his due diligence was satisfactory the stock response was to refer him to Notice 726.  He says he was regularly referred to the Notice by HMRC officers.

562.         We know that Pars understood  HMRC recommended checks such as due diligence and inspection reports in notice 726 because Pars carried out due diligence specifically to counter MTIC fraud and claimed to have the goods inspected and were doing this long before the deals in this appeal.

563.         We find Pars’ by its officers was well aware of the contents of Notice 726 at the time of the deals in question and undertook some of the checks that were recommended in it.

HMRC monthly visits

564.         HMRC and Mr Doherty agree that HMRC officers made over 40 visits:  they would visit once a month and sometimes more frequently.  The visits commenced in March 2003. 

565.         HMRC did not call the officers who had carried out these visits but the notes of their visits were put in evidence.  The Appellant did not directly challenge the evidence of the notebooks:  indeed to some extent they relied on it to put the point that they were never told their due diligence was inadequate. 

566.         Nevertheless, at the hearing Mr Doherty (as we have said) played down what he knew in 2006. He said he thought the visits were primarily about checking Pars’ proof of export and not about MTIC fraud.  He also said that HMRC (Mr Evans) had informed him that carrying out Redhill checks would protect them from MTIC fraud.

567.         We consider the hearsay evidence of the notebooks themselves.  These make it clear that the visits were primarily concerned with MTIC fraud.  Mr Evans on 5 March 2003 records:  “I have stressed to Jonathan [Mr  Doherty] the importance of clearing all new suppliers and customers with Redhill and he has agreed to do this. He is now well aware of what to look for with MTIC fraud.”

568.         Later (but before 2006) there was a visit by Officers Saunders and Christopher which involved a detailed examination by HMRC of Mr Doherty as to the checks he made.  The officers did not say that they thought the checks inadequate.

569.         An Aide Memoire was completed by Officers Saunders and Marsh in conjunction with Mr Doherty on 16 June 2005.  This was a very detailed document going through due diligence on suppliers.  Export evidence was only a very minor element.

570.         We consider Mr Doherty’s witness statement.  This indicates that HMRC were looking at evidence relating to MTIC fraud as Mr Doherty records that HMRC looked at invoices, purchase orders, inspection sheets, due diligence checks as well as the export documentation. 

571.         The CPU spreadsheet Mr Doherty devised and handed to HMRC each month contains no information on export (it does not distinguish between its non-UK and UK buyers).  This also shows as we have already found that Mr Doherty knew that HMRC’s main interest was MTIC and not whether Pars held sufficient evidence of export.

572.         Mr Hamakarim gave evidence that Mr Doherty, Mr Kazeroonian and himself would meet after each monthly visit by HMRC to discuss the position.  He agreed that the three of them would discuss MTIC fraud and how to protect Pars from it.

573.         From all of this it is clear to us that Mr Doherty and the officers of Pars were well aware of the risk in their business of CPU wholesaling of MTIC fraud.  They were aware that HMRC’s keen interest in their business was because the risk of MTIC fraud.

Conclusion

574.         We find that in April and May 2006 (and before) Pars was well aware of the risks that purchases it made of wholesale CPUs could be connected to fraud. 

575.         We consider that Pars themselves emphasise (and as is verified in the notebooks) that they would ask HMRC to confirm that their due diligence was acceptable and were frustrated that HMRC’s stance was to neither confirm nor deny this (although they did on one occasion make a recommendation for Pars to carry out credit checks).

576.         We consider that Pars’ own evidence is that they carried out due diligence and considered it to be exemplary and that it is clear from their evidence that they understood at the time that they were doing this to protect Pars from MTIC fraud.  It is clear that they were well aware that Redhill checks alone were not sufficient.

577.         Our conclusion is that Mr Doherty and the other two officers of Pars were well aware of the risks of MTIC fraud in wholesale CPU trading at the time of the deals at issue in this appeal, and that they understood from Notice 726 that the risk was of a fraud somewhere in the supply chain and not just with their immediate supplier.  We accept that Notice 726 did not in terms suggest any actual checks other than checks on the immediate supplier.

Supplier’s supplier

578.         Pars put the point that in a commercial world they could not be expected to discover who was their supplier’s supplier or otherwise police the supply chain.  A supplier would not reveal to its customer from whom it purchased.  They would not even be able to carry out a retrospective extended verification of the supply chain as HMRC had:  they certainly could not do a contemporaneous one.

579.         We accept that this is self-evidently true and that (unless Pars were actively involved in organising the fraud which is not alleged) they would not be able to discover who was in the supply chain above their own supplier and the prices at which the goods were bought and sold.

580.         We find that Pars did understand that HMRC could verify a deal chain:  this was stated in the Saunders letters and elsewhere; Notice 726 told traders to take a view on the deal chain; in evidence Mr Doherty agreed that he cooperated with HMRC and handed over documentation because he knew HMRC might attempt to verify the deal chain.  Pars’ point is that HMRC could do this but not Pars.

581.         We agree.  But we also conclude that Pars were well aware that the fraud could occur higher in the deal chain than just with their supplier and from this it follows that they did know verifying their supplier alone was insufficient.  They knew could not know their supplier’s supplier but knew they did need to take a view on the integrity of the supply chain as a whole from what information that was available to them.

582.         Mr Doherty puts the point in cross examination that he thought Pars’ duty was merely to verify its supplier and HMRC to verify the rest of the chain.  Mr Doherty said they would make a value judgement on the supplier but implied they would not do this on the supply chain as they did not see this as something on which they could make a judgment.  We do not accept his evidence on this:  we did not find him a reliable witness and it must have been obvious to anyone that if they thought there was fraud higher in the chain they should not trade. We consider that Pars was well aware of this.

Repayments without prejudice

583.         HMRC had some evidence that Pars were repaid their input tax on a “without prejudice” basis pending extended verification.  This is recorded in an officer’s notebook in relation to a claim for early 2004 “repayment authorised on a without prejudice basis 10/3/06 and trader notified in writing”  but the letter was not produced to the Tribunal.  Nevertheless Mr Doherty accepted that he saw that some payments were on a without prejudice basis but said that he did not understand to what this related.  As the letters were not produced to the Tribunal we could not take a view on what Pars did or ought to have understood from them and we do not find that the “without prejudice” repayments alerted Pars to anything than we have already alluded to.

Other HMRC contact

584.         Pars also received a letter dated 15 November 2004 from Officer Saunders in which he said that some of Pars’ deals earlier in 2004 had been traced back to defaulters.  The letter also dealt with the need for Pars to provide export evidence in relation to some current deals, and informed them that their current claim for repayment would be withheld until the export evidence was provided and verification of the new deals had been undertaken.

585.         The export evidence was provided and the VAT reclaim met by HMRC.  HMRC did not refer to the matter again and in particular did not tell Pars which of their wholesale CPU trades traced back to a fraud.

586.         Mr Doherty and Mr Hamakarim’s evidence is that they only understood the letter to relate to export evidence and did not understand that HMRC had traced some of their deals back to defaulters. 

587.         We do not accept that Mr Doherty and Mr Hamakarim did not understand the plain words of the letter.  What we find they understood, and what a reasonable person would have understood is that some of their chains were connected with MTIC fraud but that nevertheless HMRC would not take the matter any further (as long as the export evidence for the current deals was provided).

588.         It should have put Pars on notice that MTIC fraud was a very real issue for Pars itself.  We do not accept that Pars’ officers only had a very general understanding of MTIC fraud in early/mid 2006 and that they genuinely thought that it was of no more significance to Pars than credit card fraud.  They knew Pars had been actually involved in at least one fraudulent chain earlier in 2004, even though they also knew HMRC was taking the matter no further.

589.         In January 2005 Mr Doherty had a call from Officer Walkerdine saying that HMRC had traced another one of their purchases back to a defaulter.  Mr Doherty’s undisputed evidence was that Miss Walkerdine later retracted this and said the fraud was not in one of Pars’ chains:  nevertheless it shows Pars was aware of the risk to it should fraud be discovered higher up in the supply chain culminating in a purchase by them.  It did not tell them that one of their transactions had been found to be connected to MTIC.

R S Parmer

590.         A Mr Parmer was employed by Pars on 17/2/3 and left them on 22/11/5.  We were informed that neither side could have called him as a witness as he has since died.  Mr Churchouse’s evidence was that Mr Parmer had a history of being involved with MTIC trading and we were given details of this.  From this it was not clear to us how much Mr Palmer actually knew and in particular it was not clear to us from the evidence whether he was knowingly involved in fraud.  It was certainly not clear to us how much Pars’ officers would have been told by Mr Parmer.  We disregarded this evidence. 

Conclusion

591.         We have found that Pars did understand MTIC fraud was a very real risk to it and its officers did understand that the risk to Pars was not just that their supplier might be fraudulent but that their supplier’s supplier (or a supplier higher in the chain) would commit fraud.  But we find they could in practice only verify their supplier.

Reassurance from HMRC?

592.         It is Pars’ case that HMRC’s actions led its officers to believe that it was not at risk of MTIC fraud.  Pars case’ is that HMRC visited them once a month without telling them that their due diligence was inadequate which led them to believe that HMRC approved of what they were doing.  Indeed, Pars’ officers say that until HMRC withheld the repayment for April and May 2006, they had not had any intimation that anything was wrong and in particular that any of their purchases could be connected to fraud.

593.         If Pars actually knew of the fraud then the continued repayments by HMRC could only have led Pars to believe that, despite the fraud in the chains, HMRC would continue to repay the input tax.  If Pars did not know of the fraud, we have to consider whether HMRC’s actions would reasonably have led them to be off their guard.

594.         As Mr Beal for the Appellant put it, Pars was under regular and intensive scrutiny by HMRC.  The Appellant points out that HMRC knew the sort of due diligence that was being carried out by Pars, including the fact that it did not undertake credit checks, but nevertheless continued to make repayments of VAT to them.

595.         Pars’ officers also complain that when they asked HMRC if Pars’ due diligence was satisfactory the response was always non-committal:  Mr Doherty would be recommended to read Notice 726 or told that it was for the company to satisfy itself.  The exception was the HMRC officer who was informed that Pars did not carry out credit checks recommended that they do so.  Pars’ continued failure to do so was not checked upon by HMRC and the repayments of VAT continued.

596.         Mr Beal’s point is that Pars was therefore entitled to think that what it was doing was right and that it should not have its input tax withheld.  We do not agree. HMRC did not in fact reassure them that their due diligence was satisfactory.  They refused to do so and Pars knew this.

597.         Pars knew that HMRC were taking a very active interest in them by continuing to visit once a month, occasionally withholding tax for a short period and never confirming that their practices were sufficient in HMRC’s view.  They had had the Saunders letter mentioned above saying that fraud had been detected in a supply chain to a Pars’ purchase.

598.         They could not have reasonably supposed that HMRC by doing these things that HMRC were giving them an assurance that they were not involved in fraud.  The contrary has to be true:  that HMRC were concerned and Pars was required to be vigilant.

599.         Apart from this at the time Pars knew matters which it must have known were not known to HMRC.  They knew that they paid for inspection reports which they did not receive (or did not receive until long after the event).  They knew they were being able to take part in very profitable deals for no commercial risk and little effort on their part (see our findings below).  Knowing that HMRC did not have the full picture of what they were doing they could not reasonably have taken comfort from HMRC’s continued repayments that all was well.

Checks undertaken by Pars

Box numbers and its spreadsheet

600.         Pars took care to ensure that they did not trade in the same box of wholesale CPUs.  Mr Doherty designed a spreadsheet to record the box numbers on wholesale CPUs.  This is the same spreadsheet that we have referred to above in the first part of this decision notice:  we found it to be good evidence of from and to whom Pars bought and sold CPUs. It was in operation from July 2003 to July 2006.  It was designed so that the same box number could not be entered twice. It appears to have been successful:  there was no evidence that Pars had traded the same box of wholesale CPUs twice. 

601.         Pars’ counsel suggested that the spreadsheet by itself was sufficient due diligence.  At the time, he says, the only MTIC that was understood was carousel fraud.  Ensuring that they did not trade in the same goods twice was sufficient to ensure that they did not participate in fraud.

602.         We do not accept this.  We have found above that they had sufficient general awareness of MTIC from HMRC visits and Notice 726 to know that it was more than ensuring that they did not trade in the same goods twice. Further they clearly did undertake other checks, such as due diligence on their suppliers, so Pars clearly knew it was about more than checking that they did not trade in the same goods twice. 

603.         In any event, box numbers could not be recorded for their wholesale trades in retail CPUs.  Approximately half of Pars’ trades were in retail CPUs.  Mr Hamakarim’s evidence was that up to 9 March 2006 Pars instructed its freight forwarder to mark the retail boxes with a UV pen so that it did not buy them a second time.  It was not explained to us why this practice ceased nor did we have any evidence about whether it was successful.

604.         In conclusion, we find Pars did not trade in the same wholesale CPUs more than once but had no system in place to ensure that it did not trade in the same retail CPUs.  As Pars was aware, in either case this was insufficient to ensure that there was no MTIC fraud in the chain.

Due Diligence - outline

605.         HMRC’s case is that Pars’ due diligence was undertaken not to ensure that they did not become the broker in a deal chain connected with fraud but as “window dressing” to go through the motions of appearing to check because Pars knew that its chains were orchestrated for the purposes of fraud. Pars denies this.  Its officers say that firstly at the time that the due diligence was successful in keeping the company out of deals connected with fraud and that, secondly, if it wasn’t satisfactory HMRC never told them this.

606.         Of course Pars’ due diligence was unsuccessful if its purpose was (as Pars says it was) to keep the company free of deals connected with MTIC fraud.  We have found that every single one of its deals in this appeal was connected with MTIC fraud.  Pars’ due diligence was totally ineffective.

607.         The question for this Tribunal is whether Pars knew or ought to have known that its transactions on which it seeks to recover tax were connected to fraud.  So we will consider what due diligence was undertaken and what this tells us about what Pars knew or ought to have known.

608.         Mr Doherty said that Pars maintained a file on each customer and supplier.  He agrees that Pars did not undertake all the checks which HMRC recommended in their Notice 726 but they made their decisions on the integrity of a potential supplier or customer when they had (in their view) garnered enough information.

609.         Pars’ due diligence  would cover

·       Checking that the company was validly incorporated by obtaining a copy of its certification of  incorporation;

·       Checking that the company was validly VAT registered by obtaining a copy of its VAT registration certification;

·       Checking with HMRC at Redhill that the company was still validly VAT registered;

·       Often (but not invariably) obtaining a letter of introduction on the company’s headed notepaper;

·       Obtaining the company’s bank details;

·       (normally) obtaining a letter of introduction on the supplier’s headed notepaper and details of the supplier’s bank;

·       (sometimes) proof of the directors’ ID

·       Mr Hamakarim would visit the company’s premises and meet the directors.  Photographs of the premises would be taken.

·       Suppliers would be asked to complete a declaration saying (amongst other things) that they carried out due diligence on their suppliers.

Incorporation and VRN checks

610.         Pars checked their suppliers’ certificates of incorporation.  Pars also verified their suppliers’ VAT registration with HMRC at Redhill.  In the hearing HMRC made some criticism that Pars did not repeat the checks regularly.  However, Mr Mendes’ evidence was that HMRC’s advice to traders was to verify new suppliers and that HMRC would then inform a trader (with a “veto” letter) if a supplier whom a business had verified later became deregistered.  This was consistent with Mr Doherty’s evidence that he understood HMRC would send out veto letters if someone whom they had checked earlier was later deregistered.

611.         We find that there was nothing untoward in Pars relying on HMRC to send out a veto letter.  We also find that all of their suppliers were validly VAT registered at the time of their deals with Pars.

612.         However, checking that the supplier was incorporated and validly VAT registered would tell Pars that the VAT was being properly charged but would not protect against MTIC fraud.  Indeed, the fraud would be unsuccessful if the companies were not validly VAT registered.  It is a basic check for any business wishing to recover VAT which they incur.

613.         We find Pars knew this.  It was clear from their background knowledge of MTIC and from notice 723 and from the fact they did more than check registration, that they knew that this was not enough.  We also find nothing in Pars’ Redhill checks that indicate to us that they knew or ought to have known of MTIC fraud in the chains.  They carried out Redhill checks on their suppliers and occasionally repeated them but relied on receiving veto letters in the meantime.  This seems to us reasonable behaviour.

Bank account and premises

614.         Information that a company has a bank account and premises and directors would tell Pars that the company was not a complete sham:  it did exist.  It does not tell them whether the company is likely to be involved in fraud. 

Supplier’s declaration

615.         A supplier declaration form also tells Pars virtually nothing as it depends on honesty and the honesty of their supplier is what they are seeking to test. 

Directors’ identities

616.         Pars did not normally check the ID of the supplier’s directors nor the director’s proof of address. Mr Hamakarim’s explanation was that his meetings with the directors had satisfied him as to their integrity and so further investigation was unnecessary.  We do not agree that this was a reasonable view to take:  where you are seeking to test a person’s honesty, asking them whether they are honest is not the way to do it.

617.         Nevertheless, it was not suggested by HMRC that, had Pars’ checked the directors’ identities and proof of address, that anything untoward would have been discovered.

Trade References

618.         Pars did not obtain trade references.  Mr Hamakarim’s evidence was Pars’ experience with its OEM business was that trade references would give false positives: companies needing a reference would only ask for trade references from business connections who they knew would give a good reference.  For this reason Pars did not ask for trade references from its customers and suppliers in its core business as an OEM.

619.         Under cross examination, he says it was his recollection that in some cases with the CPU wholesaling he tried to obtain a trade reference from the freight forwarder but the freight forwarder refused to give any.  He did not find this suspicious.

620.         We find that trade references were a check recommended by HMRC plus one that might reveal the extent (or otherwise) of the proposed supplier’s business.  Pars’ failure to carry them out indicates a lack of interest in their suppliers of very high value consignments in a trade they knew to be rife with fraud.

621.         However, it can not be relevant to means of knowledge as HMRC have not (and presumably could not) show us what asking their new suppliers for a trade reference would have revealed to Pars’ officers.

Credit checks

622.         Credit checks were recommended by HMRC in Notice 726.  Mr Saunders (Pars’ visiting HMRC officer) in his visit on 16 June 2005 specifically recommended them when Mr Doherty informed him that they did not carry them out.  Pars continued not to carry out credit checks.

623.         Pars’ officers explanation for this was that they did not take out credit references in their core OEM business because credit references could give false positives: they therefore saw no reason for undertaking them in relation to wholesale CPU business in which they were extending no credit.  It was also their case that visiting the suppliers was far more revealing than carrying out credit checks, and that credit checks were unnecessary where they had a long trading history.

624.         We agree with Mr Beal’s point that a credit check is far from conclusive:  a company of good financial standing could commit fraud as much as one with no assets.  But a negative result would indicate there were questions to be answered.  We also agree with Mr Beal that it might be reasonable to suppose that long established trading partners were less of a risk than new ones and that less due diligence might be needed for such trading partners.  He also makes the point that newly established companies may not have filed accounts and so a credit check would reveal very little.

625.         We cannot agree that credit checks were unnecessary.  We find Pars was aware of the distinction between their core business and their CPU business as far as risk of MTIC was concerned. The credit checks were to find out if there was anything odd about the party to the deal which would suggest that all was not as it should be and in particular if the deal was orchestrated for the purpose of fraud.  A positive credit report indicating that a company had financial standing would be of some reassurance – although, as Pars says, perhaps not good enough on which to base a decision to lend credit.  On the other hand, a negative credit report would suggest to a reasonable person that further checks should be carried out.  Knowing that credit reports might give a false positive is no reason for not carrying them out when the object is to look for negative indicators.

626.         Further, we do not accept that a visit in person was a reasonable substitute check:  financial standing cannot be objectively verified by simply asking the subject.  Mr Hamakarim says that it is their choice to do business based on trust and indeed there is no other rational way to conduct business:  but it is no answer to the question of whether they did enough to ensure that they were not connected to fraud to say that they trusted the persons with whom they traded.  They have to act reasonably.  A reasonable person who wished to ensure that they are not caught up in fraudulent deals would not take everything on trust.  They would carry out checks, particularly checks they could carry out without the other party being aware of them, such as credit checks.

627.         We also do not agree that Pars had traded with many of their suppliers for very long.  The longest relationship was with Culmain and it only dated to October 2004 which was just over 18 months before the deals in question.  Aim and 21^st Pars had traded with for nearly a year.  Leisure it had traded with for 5 months; it had no trading history with Delltronics, Futuristic and TGT before the deals in issue.  Pars conducted no credit checks irrespective of how long they traded with anyone.

628.         Pars knew MTIC fraud was a real and present danger in its business and that credit checks were recommended by HMRC: we think they ought to have carried them out. 

Due diligence on Futuristic

629.         Pars carried out their normal due diligence checks as set out above.  The company was validly incorporated and VAT registered.  Futuristic completed a suppliers questionnaire confirming that it vetted its suppliers.  Mr Hamakarim visited the premises and met the staff and director (a Mr Ali).  He took photographs and checked Mr Ali’s driving licence.  He obtained a copy of a utilities bill from the premises.  He was given a formal letter of introduction from the company.

630.         This satisfied Mr Hamakarim (or so he says) of the bona fides of the company.  We do not think anyone could have been reasonably satisfied that these checks provided much security.  They showed no more than the company was genuine, with offices and a VAT registration.

631.         Perusal of the supplier’s questionnaire would reveal that the company had only 2 employees.  Its certification of incorporation would reveal it was incorporated on 3 December 2004 and VAT registered on 10 February 2005.  A credit check would have revealed it had filed no accounts (it was only just incorporated) and (not surprisingly) had no credit rating and no net worth. 

632.         At the very least this would in our view have suggested to a reasonable trader with the knowledge of MTIC that Pars had that it needed to assure itself that the reason why a newly incorporated company with 2 employees, no assets and no credit rating was able to get CPUs at a better price than Pars was not fraud.  Pars’ decision was to commence trading.

Due Diligence on Leisure

633.         Pars carried out their normal due diligence checks as set out above.  The company was validly incorporated and VAT registered.  Leisure completed a supplier’s questionnaire confirming that it vetted its suppliers.  Mr Hamakarim visited the premises and met the staff and director (a Mr M Mohammed).  He took photographs of the premises. He was given a formal letter of introduction from the company and their banking details.

634.         This satisfied Mr Hamakarim (or so he says) of the bona fides of the company.  We do not think anyone could have been reasonably satisfied that these checks provided much security.  They showed no more than that the company was genuine, with offices, five employees and a VAT registration.

635.         We find that a credit check undertaken in 2006 would have revealed that Leisure had a high risk credit rating, although with a historical credit score of “normal”.  Had Pars undertaken the credit check it would have suggested to it that there were questions which needed an answer:  why would a company with a risky credit score be able to get better deals than Pars?

21^st Trading

636.         Pars had traded with 21^st for nearly a year before the deals at issue in this appeal.  It carried out its normal due diligence checks as set out above.  21^st was validly incorporated and VAT registered.  21^st completed a suppliers questionnaire confirming that it vetted its suppliers.  Mr Hamakarim visited the premises and met the staff and director (a Mr Al Hasani).  He took photographs of the staff at the premises.

637.         This satisfied Mr Hamakarim of the bona fides of the company.  We do not think anyone could have been reasonably satisfied that these checks provided much security against participation in MTIC fraud. They showed the company was genuine, with offices and a VAT registration.

638.         No credit check was carried out.  We find it would have revealed that at the time it had a “caution” credit risk having previously been “high risk”.  The company records a profit and no fixed assets.  As a small company it did not have to file full accounts.

639.         Had Pars undertaken the credit check it would have suggested to them that there were questions which needed an answer:  why would a company with a poor credit score and no fixed assets be able to source goods more cheaply than Pars?

Aim International

640.         Pars had traded with Aim for nearly a year before the deals at issue in this appeal.  It carried out its normal due diligence checks as set out above.  Aim was validly incorporated and VAT registered.  Mr Hamakarim visited the premises and met the director and company secretary.  He took photographs of the staff at the premises. He verified the director’s ID and checked a utilities bill for the company at the premises.

641.         Aim completed a suppliers questionnaire confirming that it vetted its suppliers but not until after the deals in question in this appeal.

642.         This satisfied Mr Hamakarim of the bona fides of the company.  We do not think anyone could have been reasonably satisfied that these checks provided much security against connection to MTIC fraud.  They showed the company was genuine, with offices and a VAT registration.

643.         No credit check was carried out.  We find it would have revealed that at the time Aim had a stable credit risk but no fixed assets.

Due Diligence on Culmain

644.         Pars had traded with Culmain for nearly two years before the deals at issue in this appeal. It carried out its normal due diligence checks as set out above.  Culmain was validly incorporated and VAT registered.  Mr Hamakarim visited the premises and met the employees, whom he photographed. He was given a letter of introduction on headed notepaper and the company’s bank details.

645.         This satisfied Mr Hamakarim of the bona fides of the company.  We do not think anyone could have been reasonably satisfied that these checks provided much security against connection to MTIC fraud.  They showed the company was genuine, with offices and a VAT registration.

646.         No credit check was carried out.  We find it would have revealed that at the time Culmain had a stable credit risk (from “caution” before) but fixed assets of only £2,000 and total assets of £200,000.  It had no debtors, creditors, work in progress or fixed assets.  These were not the accounts of a normal trading company and we think would have meant a reasonable person would have investigated further.

Com 2K Ltd t/a Deltronics

647.         Pars had no trading history with Deltronics before the deals at issue in this appeal.  Its first deal with Deltronics was on 2 May and its last on 6 July. 

648.          It carried out its normal due diligence checks as set out above.  Deltronics was validly incorporated and VAT registered.  Mr Hamakarim visited the premises and met the two directors.  He took photographs of the staff at the premises. He verfified the director’s ID and checked a utilities bill for the company at the premises.

649.         Deltronics completed a supplier’s questionnaire confirming that it vetted its suppliers.  Deltronics also provided a letter of introduction on headed notepaper and details of their bank account.  A directors’ statement was provided.

650.         This satisfied Mr Hamakarim of the bona fides of the company.  We do not think anyone could have been reasonably satisfied that these checks provided much security against connection to MTIC fraud.  They merely showed the company was genuine, with offices and a VAT registration.

651.         No credit check was carried out.  We find it would have revealed that at the time Deltronics had a stable credit risk, fixed assets of £31,000 but no debtors or creditors.

TGT

652.         Pars’ first deal with TGT was on 20 April 2006.  Pars carried out its normal due diligence checks as set out above.  TGT was validly incorporated and VAT registered.  Mr Hamakarim visited the premises and took photographs of bare desks at the premises.

653.         TGT completed a supplier’s questionnaire confirming that it vetted its suppliers.  The company provided a letter of introduction on headed notepaper and details of their bank account.  The director provided a directors’ statement.

654.         This satisfied Mr Hamakarim of the bona fides of the company.  We do not think anyone could have been reasonably satisfied that these checks provided much security against connection to MTIC fraud.  They showed the company existed, with offices, VAT registration and a bank account.  It had two employees.

655.         No credit check was carried out.  We find it would have revealed that at the time TGT had a stable credit risk up from caution but no fixed assets, debtors, creditors, stock or work in progress.

Due Diligence on customers

656.         It seems very little due diligence was carried out on Pars’ customers.  Pars’ explanation is that they had a long trading history with many of them.  HMRC did not suggest that more extensive due diligence would have revealed anything untoward.

Due diligence on freight forwarders

657.         Pars did not carry out due diligence on their freight forwarders.  Mr Kazeroonian said that due diligence was unnecessary because they had met with them and they had an excellent reputation in the market place.

658.         We have no evidence to suggest that due diligence would have revealed anything untoward:  but we have commented elsewhere that it was clear that the inspections of the goods for which Pars paid the freight forwarders were not being carried out:  Pars was given inspection reports that in many cases had been carried out before the instructions were given and in relation to much larger consignments.  Whoever instructed these inspections, it was not Pars.  This we find should have suggested to them that there was something seriously wrong.

Inspection of the goods

659.         Notice 726 and HMRC recommended that traders check the goods that they purchase.  This is relevant to MTIC as goods used for the fraud might not exist, or might show signs of frequent handling as they circulate without reaching the hands of final user.

660.         Pars did not suggest that inspection was unnecessary:  Mr Hamakarim said he was responsible for the physical checking and it to be a crucial part of the due diligence.  It was Pars’ case that it paid the freight forwarders to carry out inspections and that the results of such inspections were satisfactory.

661.         We have already dealt with the evidence of Pars’ inspections in some detail in relation to Mr Hamakarim’s lack of credibility as a witness.  We find as a fact that although Pars paid its freight forwarders for goods to be inspected, it did not receive and read these reports before going ahead with the deals.  A  minority of them were ultimately received by Pars when sent the freight forwarder’s bill but this would be long after the transaction was completed and formed no part of the decision making process.

662.         Even if Pars had received and read the reports, they ought not to have been reassured by them.  Many of them contained indications that the inspections were not carried out on Pars’ behalf as the date was too early and/or the number of boxes inspected was too many.

663.         If Mr Hamakarim had read them, he would have noticed the packaging of many of the boxes of CPUs was in poor condition.  He explains (as we have reported above) that he believes that it would only be the packaging and not the CPUs themselves that would be damaged.  He says he knows this because Pars had two boxes delivered to it and used the CPUs themselves.  He says the reports merely tell him that the transporter did not handle the box carefully. 

664.         We do not agree that the damaged packaging was not a concern:  a reasonable person knowing the risks of MTIC fraud should have been concerned that so many of the boxes showed signs of wear and tear as this might indicate they had been circulating in MTIC fraud.  And there ought to have been a concern that the CPUs could have been damaged and a customer have to cause complain and reject the consignment.

Conclusions

665.         Pars put quite some effort into its due diligence (such as the visits to each supplier) but it was totally ineffective to protect it from connection to fraud.  All its deals were connected to fraud.

666.         Apart from basic checks such that the company was VAT registered and properly incorporated, it undertook no checks that gave any independent verification of the company.  It relied, as it says, on Mr Hamakarim’s impression having met the directors’ personally.  Bearing in mind that the very real risks of MTIC were  known to Pars, the failure to undertake any independent checks, such as to ask for trade references and carry out a credit check begs a very real question in our mind.  We find Pars did not do this because it had every intention of trading in any event. 

667.         We are not persuaded that the fact it did not do such checks in its core OEM business is the real reason why it did not do them in its CPU wholesaling business:  the position was quite different.  The latter but not the former was known to Pars to be at real risk of MTIC fraud and Pars knew that the checks were recommended in Notice 726.  If Pars’ officers seriously intended to avoid involvement in fraud they would have done everything reasonable recommended by HMRC.

668.         On the question of means of knowledge, carrying out the credit checks recommended by HMRC would have raised questions over many of the suppliers.  They were new companies with poor credit rating, or slightly older companies with no asset base.  Pars should have asked itself how these companies could get better deals than they could, and how could companies with low credit ratings or low assets manage to trade in such valuable consignments.

669.         Mr Beal’s view is that this checks would do no more than point to the risk of fraud.  It would not show that the transactions definitely were fraudulent. We agree.

670.         It is true that Pars itself as a new company back in 1995 had traded for a short period on the grey market and managed to do this because it was extended credit by its supplier.  Nevertheless 10 years later with MTIC fraud known to be a real risk in the market, we think Pars should have asked itself in respect of its deals how these companies were managing to offer Pars deals in such valuable consignments.  We are inclined to think that Pars would have concluded (had it asked itself this question) that a likely explanation was that the deals were connected to MTIC fraud.

What was the nature of the deals Pars entered into?

Importance to Pars

671.         HMRC’s unchallenged evidence was that between 1997-2002 Pars’ turnover fluctuates between £7-£14 million per annum.  At this time Pars did not trade in wholesale CPUs but conducted its core business as an OEM.  In 2003 it commenced dealing in wholesale CPUs and its turnover increased to £20million.  In 2004 its turnover increased to £35million.  In 2005, its turnover increased to £109million.  In 2006 its turnover increases at the start of the year but drops back as Pars conduct no broker deals after May 2006:  its turnover for the entire year is £91million.  In 2007 there is no wholesale CPU trading and Pars’ turnover has returned to approximately £14million.

672.         In money terms, the size of Pars’ wholesale CPU trading dwarfed its core business as an OEM:  in April 2007 it was denied input tax on 90% by value of its purchases and in May 2006 it was denied input tax on 79% by value of all its purchases in that month:  that denial of input tax is the issue in this appeal. 

673.         Nevertheless, this vast increase in turnover was not matched by increased resources.  The deals were negotiated by Mr Hamakarim and the finances dealt with by Mr Doherty in addition to their normal duties on the OEM side of the business.  No extra staff or premises were needed to carry out these trades. 

Commercial risk

674.         Pars’ case is that it was financially at risk in the transactions.

675.         We find that in 46 out of the 59 deals in issue Pars paid its supplier before it was paid by its buyer.  However, the risk to Pars in these deals was considerably lessened as in almost all cases where Pars was not paid in advance, it shipped the goods “on hold” so that the goods were shipped to a warehouse abroad and not released into the possession of the buyer until the buyer had paid.  On the face of it, if the buyer did not pay, Pars would be left with the goods which it would have to resell, possibly at a loss.  There were no instances where this actually happened.

676.         It was Pars’ case that they took a risk because they paid before they were paid but at the same time say that they undertook no financial due diligence (such as credit checks) as they extended no credit.  However, this makes little sense because (if what they say is true) they were at risk at being left with unpaid-for goods:  a financial check on their buyers would have made commercial sense.  Their answer was that they had a long trading history with many of their buyers and they trusted them.  This means that they did not consider Pars to be at risk.

677.          We find that Pars did not really consider themselves to be at risk.  And it appears that they were correct:  their buyers always paid and Pars was never left with unsold stock.  We find the transactions involved no real risk for Pars and it knew this.

Contractual terms

678.         Pars did not have written contractual terms for its CPU trades.  Its website does contain its standard terms and conditions but these were written for its core business.  We find that its lack of specific terms and conditions did not cause Pars any problems:  goods were never returned, unpaid for, nor complaints made about their condition.

679.         We find it odd that little thought was given to the terms and conditions for Pars’ CPU trading which in value (although not effort) dwarfed its core business and in particular whether it would have been appropriate to have different terms and conditions for a wholesale export business than a retail OEM business.

Insurance

680.         An allegation made by HMRC before but not pursued at the hearing was that the goods were not insured by Pars indicating that they knew these were not normal commercial deals.  We find however that the goods were insured.  Pars held static insurance up to £1million for goods held in freight forwarders’ warehouses and £350,000 worth of freight insurance for goods on the move.  This cover was not always sufficient for some of their deals but would be increased by Pars on a deal by deal basis.

Credibility of deals

681.         Pars were buying from newer and smaller companies than themselves who did not have an established business (apart from CPU wholesaling).  These companies were sourcing the goods more cheaply than Pars.  HMRC ask why didn’t this alert Pars to the fact that the deals were being orchestrated?

682.         Mr Doherty’s answer was that Pars believed that these companies just had better contacts that Pars. It is Pars’ case that it believed that these were genuine grey market trades.  It knew that it itself was not buying from an OEM or authorised dealer so its case must be that it assumed that its vendor was buying from someone who had an interest in keeping their dumping activities secret.  After all, it had had the experience in 1995 as a new company of being selected as a contact to dump excess stock on behalf of a much larger company.

683.         Pars did not know its vendors’ profit margins but it knew it own.  It made a profit on all its deals:  on some it was making a very large profit.  What did it do in return?  It knew that it did not take any real risk:  see paragraph 674-677 above.  It was never left with unsold stock.  It was able to put together deals involving a buy and sell of identical quantities of stock.  It was able to put a deal together so that they were back to back.  When Pars exported goods on hold its buyer always paid. 

684.         Even putting the deal together seemed to require very little effort:  Mr Hamakarim’s description was that the whole process would be spread out over 24-48 hours and merely involve offers being made by or to vendors and buyers.  We do not accept this evidence in any event (as explained below). But even if we did accept it, putting the deal together was hardly a great effort and no real risk.

685.         So Pars knew it was making easy and large profits.  Even if it was possible that its vendor would be a small, new company, why would this vendor give up the profit made by Pars?  Why didn’t Leisure, say, sell direct to Inco Parts?  Neither were OEMs with a need to keep their activities secret.  Leisure could have advertised its stock and Inco Parts its requirements.

686.         If Pars had no knowledge of the fraud, they must have asked themselves these questions or shut their eyes to it.  It was put to Mr Kazeroonian that it was too good to be true but he denied this.  We do not believe him. He has to have been aware, and we find he was aware,  that Pars’ CPU wholesale trading gave Pars a turnover and profit that dwarfed its core business yet cost Pars very little in effort, risk or expenses.

687.         It also makes no commercial sense that Pars would not seek to buy at the cheapest price.  In Deals 25, 44, 64 & 65 it bought from Culmain.  But Culmain bought from Futuristic (at a cheaper price than Culmain sold to Pars).  Pars’ case is that they did not know that:  but Futuristic was a supplier to them (see deals 1, 4, 37, 38, and 55) and it is their evidence that they would shop around.  Clearly they did not.

Split deals

688.         Mr Beal put the view that because deals were sometimes split this indicates genuine and not artificially contrived trading.  We find Pars split the consignment it purchased on about 18 (approximately one third) of the 59 invoices at issue in this appeal.  We have found the deals were orchestrated including the splits.  The question is whether Pars knew this or ought to have known it.

Length of deal chains

689.         It is Pars’ case that they did not know how long the deal chains were.  Its vendors would not give away their contacts.  This is a reasonable point to make.

690.         But we have to consider whether Pars should have done more to verify the length of the chain or whether it did not do so because it knew that the chain was orchestrated.

691.         Pars did not ask their freight forwarders for an indication of how many buffers there were in the chain nor did they enquire if there was an OEM or authorised distributor in the deal chain.  Mr Doherty replied that they did not do this because it did not occur to them and it was not a question which HMRC had ever suggested to them.  We accept that they did not do these checks and that HMRC had never suggested them.

692.         We find this lack of curiosity itself curious.  Pars is a business out to make profit:  it knew that it was not buying from the OEM who was dumping stock:  it should have been curious to find out what it could from its own business point of view.  In the context of knowing it was at risk of being involved in MTIC trade it should have been concerned to ensure that it was a grey market trade and not an orchestrated trade.

693.         Nevertheless, even if Pars had made enquiries of the freight forwarders they may have been rebuffed and found out nothing.

FCIB

694.         Many of Pars’ suppliers banked with FCIB.  We were told that this bank was closed down by the Dutch authorities in August 2006 for suspected involvement with MTIC fraud.  However, we find this was not common knowledge until (at the earliest) August 2006 so we do not agree with HMRC that Pars should have made inferences from the fact its suppliers chose to bank with FCIB.

Internal order forms

695.         An allegation made by HMRC was that Pars internal order forms were evidence that the deals were orchestrated.  This was because the forms were dated often much earlier than the date of the deal showing (said HMRC) that they were arranged in advance.  In some cases they were dated up to 3 weeks before the deal in question.

696.         Mr Doherty and Mr Hamakarim’s explanation was that the internal order forms were part of Pars’ internal accounting system that applied across the business including its core OEM business (for which it had been designed).  Each deal would have a unique reference number taken from the internal order form. Mr Hamakarim said he would ask staff members to print off a batch of forms for him which he would keep and use on the wholesale CPU deals.  In order to get staff to input the details of a wholesale CPU trade into the accounting system, Mr Hamakarim would make handwritten notes on the pre-printed internal order form.

697.         HMRC carried out an analysis of the internal order forms used by Pars on their wholesale CPU trades.  This showed that least 11 were dated 6 April 2006; 14 were dated 20/4/06; 21 were dated 28 April and 4 were dated 16 May.  From looking at the dates of the deals it seemed that the 11 dated 6 April were used up to 20 April, the 14 dated 20 April were used on deals dated up to 28 April and so on.  This is consistent with Mr Hamakarim’s explanation that he would use up a batch and then ask for another batch to be printed.

698.         HMRC were able to demonstrate from the numbers (they were consecutive) that the “batch” printed on 20 April were printed almost alternatively by two members of staff “Emery” and “Adidi”.  This showed, they said, that they were not truly printed in batches.  Pars’ explanation was that the forms could only be printed individually:  and indeed they all carry a unique reference numbers so we accept this.  Therefore, says Pars, a member of staff would have to print one after the other to get a batch.  If two members of staff worked on this together this would explain why the batch would show alternating names of the members of staff.  We accept this explanation as it makes sense of the evidence.

699.         We find there was nothing in Mr Hamakarim’s use of these forms that indicated knowledge of the orchestration of the deals.  In particular, there is nothing in them that indicated foreknowledge of the deals.  They were printed in batches of a sort and carried the date of printing.  As Mr Hamakarim did not date his handwritten notes on the deal it is apparent to the Tribunal that they must have been made contemporaneously with the deal:  how else would the accounts department know the date of the deal unless they knew it was the date on which they were handed the form?  How would Mr Hamakarim himself have known the date of the deal if HMRC were right and the forms were completed before the day of the deal in question? If HMRC were right, surely Mr Hamakarim would have had to have entered the intended date of the deal.  But he didn’t.  We accept Mr Hamakarim’s evidence that he completed them by hand on the day of the deal.

Invoices

700.         HMRC made a point that many invoices post-date the deal and post-date the payment and release.  Mr Doherty’s reply is that the invoice was done as soon as possible.  We do not find the date of the invoice to be significant:  it is not a legal requirement for them to be produced on the day of the transaction and the fact that they were not does not appear to indicate anything relevant to us on the question of knowledge or means of knowledge.

Negotiation of the deals

701.         How were Pars’ deals negotiated and did the way they were negotiated tell Pars, or should it have told Pars, that the deals were orchestrated and connected to fraud?

702.         Mr Hamakarim negotiated the deals and he was questioned about it under lengthy cross examination.  We found him reluctant and vague in his evidence.  He was asked to restrict his answers to the CPU wholesale trade but would talk about the business as a whole.  He appeared to want to step back from the evidence in his witness statement on negotiation and suggested that it applied to Pars’ business as a whole and had not been limited to CPU trading:  we find do not accept this because his witness statement appears to relate solely to the CPU trading and Mr Hamakarim was well aware it was just the CPU trades that were in issue. 

703.          He was asked if the deals were customer or supplier driven and to confirm whether Mr Kazeroonian’s statement they were customer driven was correct.  He was reluctant to give any clear answer:  we considered whether this was just because he did not wish to contradict the owner of the business which employed him but we found it to be the case he wished to avoid being tied down to a definite statement.

704.         He was reluctant under cross examination to confirm his witness statement that he would call 3 to 4 suppliers when he had a customer wanting to trade, and then indicated the calls would be spread over a number of hours which we find improbable if the business was a fast-moving as he claimed.

705.         He denied that deals were back to back because he said that they could not always match quantity:  a customer might want more than Pars could supply and the deal was concluded on the amount that could be supplied.

706.         He was cross examined on the lack of terms and conditions.  Terms were set out in the invoice but invoices would not be issued on the day of the deal.  His witness statement said Pars would provide details to its customers, but in cross examination he said they would know the terms from invoices issued on previous deals.  He then appears to adopt Mr Cunningham’s suggestion that the terms and conditions were read out over the phone.  We find this most improbable particularly when his evidence was that the only discussion about terms was terms of payment.

707.         Our conclusion is that Mr Hamakarim gave the Tribunal no very clear understanding of how the deals were negotiated:  we do find that terms and conditions were not discussed other than perhaps terms of payment.  We find his vagueness was because he did not want to be pinned down into saying something that could be disproved and this was because he knew that the deals were not negotiated in anything like a normal manner.  We consider that his vague evidence reflected the fact that he was being less than honest with the tribunal because he knew that these deals were not negotiated on the open market.

708.         When we add to these considerations the fact we have already concluded in the first part of this decision notice that Pars’ purchases and sales were all orchestrated by the fraudsters, Mr Hamakarim’s vague evidence is entirely explained.  There were no normal negotiation on these deals and Mr Hamakarim knew this.

Profit margin

709.         Mr Beal considers that HMRC should not make a case on Pars’ profits on these deals as there is no evidence in front of tribunal that profit margin obtained by PARS was other than could have been obtained in the market and no evidence that prices paid by Pars’ buyers weren’t the prevailing prices in the countries in which they were located.

710.         HMRC point out that Pars’ profit normally greatly exceeds the profit made by the buffers.  Pars’ response is that they did not know their supplier’s suppliers and they certainly did not know the profit that they made.

711.         Unlike the buffers, Pars’ profit margin is not consistent and varies between the deals.  It was nevertheless a substantial profit on every deal.  It was only £1,656 on Deal 13 with a sale of 315 CPUs amounting to a profit of £5.26 per CPU.  A more average deal was number 44 with a profit of £17,500 on a sale of 3,500 CPUs at £5 profit per CPU.  There were a few deals which broke the mould: deal 46 had a profit of £46.50 per CPU which led to overall profit of £69,750.  The second highest profit per CPU was £7.39 (deals 32 and 62).  The more average profit per CPU  seemed to be around £5 per CPU and was never lower than £3.55 (deal 41).

712.         For the two months at issue in the appeal Pars stood to make £907,000 in profit from its trade in wholesale CPUs.  Mr Doherty’s evidence in cross examination was that at the most Pars’ expenses to achieve this profit would have been about £40,000. This was an estimate of the freight and inspection charges of £32,000 and £8,000 for two months’ insurance premiums.  Pars’ expenses were limited to freight, insurance, and inspection and the loss of use of capital tied up in a VAT reclaim.  We have already commented that little staff time and no premises were needed to conduct the business. 

713.         The Tribunal finds this to be an extraordinarily high net profit for a business which was risk free.

714.         Pars’ profits in each deal were greatly in excess of the buffers.  HMRC’s point is that the fact that the orchestrator of the fraud allowed Pars to pocket the lion’s share of the profit in the deal chain indicated that Pars were knowingly involved in the fraud:  if not, why were they paid so much more than anyone else?

715.         It was put to Mr Kazeroonian that he did very well out of the wholesaling of CPUs.  In 2003 his salary was nearly £60,000.  In 2004 (wholesale trading in CPUs having commenced the previous year) it increases to £221,000.  In 2005 it increases to £350,000 and in 2006 to £1.28million.

716.         Mr Beal described this as the politics of greed:  a high salary does not necessarily indicate fraud and we agree.  Mr Kazeroonian implied that his salary increases were due to him ploughing back profits in previous years to build up the business but in later years deciding to reap the rewards.  This is not the case.  Pars profits increased greatly due to its CPU trading. We think Pars and Mr Kazeroonian in particular should have asked why was Pars suddenly able to make such large profits. Why was it able to do this at a time when MTIC fraud was (to the knowledge of Pars’ officers) rife in the industry?

717.         It was put to Mr Kazeroonian that it was too good to be true and that he was doing very little to earn the £1.28million that he was paid in 2006.  Mr Kazeroonian’s answer was that nothing was too good to be true and that the market was “difficult” and “volatile”.  He said they risked their capital and used their market knowledge and good judgement to trade.

718.         We see no evidence of this.  Pars made a profit, normally a very large profit in all its CPU wholesale trades.  Its costs (freight insurance and inspection) were low in comparison to the profits made.  Even we accepted Mr Hamakarim’s explanation of how a deal was negotiated, it did not take a great deal of input and only really involved one member of staff on-and-off over a few hours, with help from accounts to draw up the purchase order and invoice.  The deals were without risk as Pars did not buy without a sale being arranged, and in practice a buyer never reneged. There is no indication that the deals would have required market knowledge:  nothing in Mr Hamakarim’s description of the negotiation process indicated that finding a buyer or seller was difficult or time-consuming.  There is nothing in Mr Kazeroonian’s description of the grey market which could explain why buyers and sellers (other than the OEM actually doing the dumping) would keep their identity secret.

719.         Mr Kazeroonian would not accept that it was too good to be true.  We take account of the fact that we have not found him to be a credible witness. We find that Pars must have been aware of the orchestration of the fraud because the level of their profit was so high.  The fraudsters would not allow an innocent party to participate in such a large share of the “profit” from the fraud. We consider that he knew that the transactions were orchestrated for the purposes of fraud.

720.         It also follows that even if we had found Pars did not actually know of the fraudulent orchestration, then they ought to have known that the only explanation for their ability to make such high profits without risk was that the deals were orchestrated and fraudulent.

Weights

721.         There had been some suggestion by HMRC that some of the Appellant’s CMRs might be false and/or the goods as described did not exist because HMRC did not consider that the declared weights were correct.  HMRC did not in fact pursue this allegation at the hearing and the Appellant demonstrated that (because of the inclusion of a fan in a box with a retail CPU) the weights on shipments of trays of wholesale CPUs would be quite different to those on shipments of retail CPUs.  We entirely accept this evidence.

Findings in other Tribunals

722.         As we have already mentioned in paragraphs 316-319 above we do not give weight to findings of fact made in other tribunals.  Mr Beal asked us to bear in mind that in his view the evidence of knowledge/means of knowledge in this appeal was similar to that in Bond House where the finding was that the traders were not involved in the fraud even though their deals were connected to it.

723.         In addition to the reasons given above why a tribunal would be reluctant to give any weight to findings of fact of another tribunal that must particularly be the case where the finding of fact was not relevant to the findings of law made in the Tribunal.  Bond House pre-dated the ECJ’s ruling in Optigen and Kittel  and HMRC’s case was that the knowledge or  means of knowledge on the part of the trader was not relevant to liability.  That HMRC conceded in that case the trader was not knowingly involved cannot tell this Tribunal anything in this case.

Deals not entered into

724.         Mr Beal put forward as one reason the Tribunal should see that his client did not have actual knowledge of the fraud was that they checked out some 14 potential suppliers with Redhill but then did not go ahead with any trades.  If, he says, the Appellant was intending to fraudulently extract VAT, why did it not trade as many times as possible in order to maximise returns from fraud?

725.         We do not know why the Appellant checked out more traders than it traded with.  Checking out more traders than traded with could have a dishonest as well as an innocent explanation:  it could be that that the fraudsters set up deals in advance that for whatever reason were not proceeded with.  So we look at the evidence overall when deciding what the Appellant knew.  And we have concluded from this that they did know of the connection of fraud in all the deals at issue in this appeal.

No connection with fraudsters

726.         Mr Beal also puts the case that no evidence was led that Pars had any knowledge of or connection with the persons who organised the frauds, being, he suggests, the persons to whom the third party payments were made.  This does not persuade us of Pars’ innocence either.  The question for this Tribunal is whether Pars knew (or ought to have known) that its transactions were connected to fraud.  We have concluded that they did know this.  The question is not whether they actually knew the persons who organised the fraud.

727.         In any event, if Mr Beal is right and the beneficiaries of the third party payments were indeed the organisers of the fraud, then the evidence is that Pars did indeed know one of them:  ASAP was one of its trading partners.  However, we do not determine who orchestrated the fraud:  we do not need to do so to decide the case and we do not have the evidence to make such a determination.

Conclusion on knowledge and means of knowledge

728.         We found that Pars by its Director and officers was very well aware of the risks of MTIC in the market in which it traded.  It knew there was a real risk of a fraud somewhere in the supply chain and that checking its suppliers would not be enough to protect it.  We did not find Pars’ witnesses to be credible and we did not believe them when they denied this for the reasons explained above.

729.         HMRC did continue to repay input tax to Pars during the years immediately preceding the deals in question.  Because we find that Pars actually knew of the fraud this is not relevant.  Putting this aside, on the question of means of knowledge, we find that Pars could not reasonable have supposed from this that all was well because Pars knew that it did not undertake all the checks recommended by HMRC (such as credit checks and inspections of the goods), it knew that it had undertaken transactions connected to fraud in the past but HMRC had nevertheless continued to repay Pars, and it knew things about the deals that were not known to HMRC. 

730.         We consider the checks undertaken by Pars.  They failed to undertake the two independent tests of their supplier (trade references and credit checks) that were recommended and instead chose to rely solely on their judgment of their supplier’s honesty based on their meeting with the supplier.  We find this indicates that they did know of the connection to fraud as that is the most likely explanation of the reason for acting in this way. 

731.         Pars did not inspect the goods.  More than that it paid for inspection reports that it did not receive (if at all) until long after the deals were completed.  It is clear to us from this that Pars knew that its transactions were arranged for the purposes of fraud and that it intended to hide that fact by appearing to undertake a check recommended by Customs (the inspection) without actually doing so.  We find the “window dressing” alleged by HMRC is proved.

732.         Even independent of the evidence of the inspections, we find that Pars was entering into deals the value of which eclipsed its core business, in respect of which it (and its Director) made huge profits vastly in excess of its core business, without undertaking any real risk, with relatively very low expenses, and very little expenditure of effort.  We do not believe that the fraudsters would have allowed Pars to realise such large profits if Pars were not “in the know” about the fraud.  We find Pars did know of the fraud.

733.         We also find that Pars did not seem to consider that it should be concerned about the contractual terms on which the deals were carried out (relying they suggested on their OEM terms of business although Mr Hamakarim’s evidence on this was confused), that Pars did not appear to shop around for the best deal as explained in paragraph 687 above, and they evinced no curiosity about the chain, and that we found the evidence of all three of their witnesses to be unreliable.  This conduct is all consistent with our finding of actual knowledge.

734.         Lastly, but not least, we have found that the deals were orchestrated.  They could not have been negotiated in a normal commercial manner.  Mr Hamakarim, who told us he negotiated the deals, had to have known from this that they were orchestrated for the purposes of fraud.  And we find he did know this.

735.         We find that Mr Kazeroonian (and by his agency Pars) knew all that was known to Mr Hamakarim and Mr Doherty.  He controlled the company, caused the company to employ them and stood to reap the profits of the deals in issue.  It was also clear from the evidence that although he left day to day control to Mr Hamakarim and Mr Doherty nevertheless in practice he retained overall control of the CPU wholesaling. 

736.         In conclusion Pars (by its Director and officers) knew that the transactions at issue in this appeal were connected to evasion of VAT. 

737.         On the question of means of knowledge, if Pars had undertaken credit checks it would have discovered information that should have led it to question how those suppliers were trading in such large quantities of high value goods. Futuristic would have been shown to be a credit risk:  the other companies showed a more stable credit risk but that they were companies without much substance.  We agree with Mr Beal that by itself this is not enough to show that the only likely explanation of the transaction was that it was MTIC.  We think it should have indicated to Pars that it was a possible explanation.

738.         We also find that the ability to make such large, frequent and easy profits should have alerted Pars to the fact that the only explanation for them must be that the transactions were connected to fraud.  This is especially the case as they should have known that they were trading in very large figures in many cases with fairly new companies with little or no assets and should already have been very alert to the risk of MTIC fraud.

739.         The fact that the deals could not have been negotiated in a normal commercial manner should also have told Mr Hamakarim that the only explanation for this was that they were being orchestrated for the purposes of fraud.

740.         Therefore, in the alternative, we also find that Pars (by its Director and officers) ought to have known that the transactions at issue in this appeal were connected to fraud and the evasion of VAT.

741.         We dismiss the appeal in its entirety.

742.         We award HMRC their costs (to be assessed on the standard basis if not agreed) as this is a transitional case in respect of which a successful application was made for the old costs rules in the Value Added Tax Tribunal Rules 1986.

743.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: 15 December 2010

Appendix 1

Full name of companies mentioned in decision notice