Bond House Systems Ltd v Customs And Excise [2003] UKVAT V18100 (08 May 2003)

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

United Kingdom VAT & Duties Tribunals Decisions

You are here: BAILII >> Databases >> United Kingdom VAT & Duties Tribunals Decisions >> Bond House Systems Ltd v Customs And Excise [2003] UKVAT V18100 (08 May 2003)
URL: http://www.bailii.org/uk/cases/UKVAT/2003/V18100.html
Cite as: [2003] UKVAT V18100

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

18100

VALUE ADDED TAX — input tax — "carousel" fraud — whether established — appellant unwitting participant — whether transactions without economic substance — whether payments made as VAT properly so regarded — First Directive, preamble, art 2 — Sixth Directive, arts 2, 4, 5, 17, 18, 28a — VATA 1994, ss 4, 24, 25, 26 — legal certainty — proportionality — Human Rights Convention, art 14, First Protocol, art 1

MANCHESTER TRIBUNAL CENTRE

BOND HOUSE SYSTEMS LIMITED

Appellant

- and -

THE COMMISSIONERS OF CUSTOMS AND EXCISE

Respondents

Tribunal: Colin Bishopp (Chairman)

C B H Gill

Sitting in public in Manchester on 6, 7, 8, 9, 13, 14, 15 and 16 January and 4 March 2003

Paul Lasok QC and Michael Patchett-Joyce, instructed by Baker & McKenzie, for the appellant

Jonathan Peacock QC and Francis Fitzpatrick, instructed by their solicitor's office, for the respondents

© CROWN COPYRIGHT 2003

DECISION

Introduction

Bond House Systems Limited ("Bond House") was, at the times material to this decision, a dealer in computer components, and principally processor chips ("CPUs"). It was incorporated in 1999, but its business had previously been carried on by a partnership, formed in about 1992, of two of the present directors. By far the larger part by value of its business was the purchase from other traders registered for VAT, like itself, in the United Kingdom, of CPUs which it then sold to traders registered in other European Union countries. In normal circumstances a trader such as Bond House pays VAT on its purchases and acquires, in principle, a right to a VAT credit, while its sales are zero-rated. The preponderance of Bond House's exports to other Member States over its other sales was such that for several years it had been a repayment trader. It rendered its VAT returns monthly and, for each month until May 2002, received a substantial repayment from the respondent Commissioners.

In that month, it made a number of purchases of CPUs, and 157 sales to an aggregate value (net of tax where applicable) of £95,421,407. Of its sales, 51 (or 32.5% by number) were to customers in other Member States but, at a total value of £94,622,611, they represented 99.1% by value of Bond House's turnover in that month. We need to concern ourselves only with those sales.

In each of those 51 cases the transactions were as we have described: the purchase was from another UK trader, and the sale was to a trader located in another Member State, in almost every case in the Republic of Ireland. In each case, Bond House took delivery of the CPUs it had bought, at what it considered to be a fair market value, paid the agreed price plus the VAT the supplier added, and then arranged delivery of the chips to its customer, which in turn paid the agreed price, also at what Bond House considered to be a fair market value, though in every case somewhat higher than the amount Bond House had paid. Bond House's supplier in the United Kingdom was invariably registered for VAT, provided Bond House with a VAT invoice complying in every particular with the requirements of regulation 14 of the Value Added Tax Regulations 1995 (SI 1995/2518) and accounted to the Commissioners for the VAT it had charged and received. The customer, too, was invariably registered for VAT in its own Member State and Bond House satisfied the UK requirements which must be met if supplies to registered traders in other Member States are to be zero-rated, that is its invoices quoted the overseas customer's VAT number and the goods left the UK. We heard no direct evidence that the overseas customers accounted for acquisition tax, but we have no reason to suppose that they did not. The Commissioners accept that the transactions were as we have outlined them, that the goods and the payments changed hands, and that the goods were as they were described in the invoices.

Anyone acquainted with the workings of VAT would assume that Bond House had an unassailable right to credit for the input tax it had incurred in making its purchases. Bond House certainly made that assumption when it rendered its May 2002 return on Friday 7 June. The amount of tax for which it sought credit and repayment by that return (since, as usual, the value of its zero-rated supplies in the month substantially exceeded that of its standard-rated supplies) was £16,287,204.94. As in previous months, and as it had agreed with the Commissioners, it submitted documentation supporting the amount claimed with the return. Although the Commissioners had carried out occasional inspections of Bond House's records by means of a site visit in the past, its claims had always been met. On this occasion the Commissioners refused to pay.

They first decided to make more detailed enquiries than usual into the transactions which underlay the claim. Those enquiries began with a visit to Bond House's premises on Tuesday 11 June, when a start was made on examining all of the documentation it had relating to its May 2002 trading; the examination was not limited to the purchases, but extended to sales records and beyond those to such matters as telephone call logs. The purpose of the investigation was not only to determine the genuineness or otherwise of Bond House's own transactions but also to identify, if possible, the chain of supply. It was made clear that the Commissioners were intent on investigating the entire sequence. Bond House's directors realised immediately that it could not look forward to early payment of the amount claimed in its 05/02 return and on the following day, understandably concerned about its position and in particular its overdraft arrangements with its bankers (which were dependent on prompt payment of its input tax credit claims) and its ability to continue its business lawfully, Bond House decided to cease trading.

The meeting of 11 June was followed by intense correspondence between the solicitors whom Bond House had by then instructed and the Commissioners, in which the Commissioners made it perfectly clear that a decision about whether the claim would be accepted, in whole or in part, would not be made until they considered they had completed their enquiries, and that, even at this early stage, they had significant misgivings about the claim. Part of Bond House's complaint is that they had already decided not to meet it and the purpose of the investigation was to find evidence to fit the conclusion they had reached. Despite their doubts, but recognising the financial difficulties occasioned to Bond House by their refusal to meet the input tax claim, the Commissioners did relent to the extent of £6m, which they paid to Bond House in July 2002 without prejudice to their objections in principle to the claim. They refused to pay more, and indicated that action to recover the amount which they had paid was a distinct possibility.

Bond House then applied to the Administrative Court for judicial review of the Commissioners' refusal to meet its claim. That application led to the making by the Commissioners to Bond House of a further payment of £2 million, though again without prejudice to their ultimate conclusion, and to their undertaking to give a fully reasoned decision for their refusal to meet its claim in full. The reasoned decision was contained in a letter dated 19 September 2002, written by the officer who led the enquiries we have mentioned and who had determined that the claim should not be met, Andrew Lumb. Mr Lumb's letter informed Bond House that the Commissioners now accepted that, of the total sum claimed, £2,732,200.54 was allowable but the remainder – £13,555,004.40 – was not. This sum represents the aggregate of the input tax incurred by Bond House in making 27 of its May 2002 purchases. We shall return to the reasons for the refusal in due course. Mr Lumb's letter was followed, in October 2002, by an assessment for the difference between the aggregate amount which the Commissioners had paid, of £8 million, and the amount which Mr Lumb considered allowable – that is, for £5,267,799.46.

The appeal to this tribunal was launched in August 2002 (and therefore before Mr Lumb's letter had been written and the assessment had been made) but it has been amended so as to enable Bond House to challenge both the assessment and the Commissioners' refusal to meet the balance of its May 2002 repayment claim. If it is successful in its appeal against the refusal to pay the balance of its claim, it also seeks a direction that the Commissioners pay repayment supplement.

The Commissioners contend that the transactions which gave rise to the disputed sum were in each case part of a series, or circle – a "carousel" – of transactions whose purpose was not the buying or selling of computer chips in the course of a business activity properly so regarded, but the perpetration of a fraud. They do not contend that Bond House was an active participant in the fraud, or even aware of it; indeed they expressly disavow any such contention and do not assert recklessness, although they do not go so far as to say Bond House is above criticism – they maintain that it was naïve or imprudent, or that its directors suspended disbelief. Instead, they argue that the fraudulent purpose of the series, as they see it, has the consequence that the entire sequence of transactions, including those to which Bond House was a party, were devoid of economic substance. If, as they maintain, the purpose of the sequence of transactions was not to supply goods, in the ordinary sense of that phrase, but to act as a device for the gaining of a pecuniary benefit by fraud, Bond House's lack of involvement in, and ignorance of, that fraud did not confer economic substance on those transactions in the sequence in which it participated. Since, they argue, illegal transactions are outside the scope of VAT what Bond House paid to its suppliers under the guise of VAT was not VAT at all, and it is not entitled to input tax credit.

Bond House's retort is an essentially simple one: that it has made taxable purchases in the course of its business of making taxable, even if zero-rated, onward supplies, that is it has the requisite documentary evidence and that it is accordingly entitled, as of right, to the input tax credit which it has claimed. There is no case for the Commissioners to look beyond Bond House's own transactions once it is accepted (as is the case here) that Bond House was not party to any fraudulent activity – indeed, they are not permitted to do so.

If the matter were limited to that issue, we would have needed to hear little or no evidence, and would instead have dealt only with the parties' arguments on the law. However, Bond House disputes not only the Commissioners' interpretation of the applicable law, but also their conclusion that there is evidence or, at least, sufficient evidence to justify their stance of fraudulent activity in the chain of supply. It raised, in addition, two further arguments about the manner in which the Commissioners had approached the matter. The first related to their treatment of Bond House, by comparison with other traders identified to have been participants in the sequence of transactions, in that they had deprived Bond House of its input tax credit, maintaining that what was claimed was not VAT, while treating transactions entered into by those other companies normally (that is, by accepting the output tax offered and giving credit for the input tax claimed), and while prosecuting others on charges which necessarily implied that they had sought to evade payment of VAT. Their doing so, it was argued, manifested an inability on their part to distinguish properly between what was, and what was not, VAT. Secondly (though these are perhaps three discrete arguments), it contended that the Commissioners' approach offends its human rights and the principles of proportionality and of legal certainty.

It has consequently been necessary for us to consider a large number of documents, to hear oral evidence from several witnesses and to read the statements of others. We were asked by counsel to read further material, some of which is evidential while the remainder consists of legal authority, and we have done so.

Although we heard Bond House's opening submissions on the law, followed by evidence and then further legal argument, after which the Commissioners responded in a similar fashion and the appellant then replied, it seems to us simpler, for the purposes of this decision, if we deal with the evidence and our findings of fact before we proceed to the parties' various arguments. First, however, we need to set out the Commissioners' description of a "carousel" fraud, which is the foundation of their case.

The "carousel" fraud

We take what follows from the respondents' statement of case, from the evidence of, in particular, Mr Lumb and Peter Birchfield, another Customs officer whose statement we had, although he did not give oral evidence, and from what was said on the respondents' behalf by their leading counsel, Jonathan Peacock QC. Bond House – represented by Paul Lasok QC and Michael Patchett-Joyce – did not demur from the description of a carousel fraud as it was advanced by Mr Peacock; they limited themselves to disputing that there was sufficient evidence of such a fraud behind any of the transactions with which we are concerned.

In its simplest form a carousel fraud works in this way. A VAT-registered trader ("A") in one European Union Member State sells taxable goods to a VAT-registered trader ("B") in another Member State. A's sale to B is zero-rated in A's Member State. B should declare the purchase and pay acquisition tax in its own Member State, and, upon the premise that it is intending to use those goods in order to make an onward taxable supply, then claim credit for the same amount as input tax. Usually, if it is a participant in a carousel fraud, it does neither. B then sells the goods to another VAT-registered trader ("C") in its own Member State, charging and receiving VAT on the consideration. However, it fails to account to the tax authorities for that VAT and effectively disappears; it becomes what the respondents refer to as a "missing trader". Nevertheless, at the time of making its sale to C, while it is still registered for VAT and before the Commissioners are aware that it is or might become a missing trader and have been able to intervene (for example, by de-registering B) it provides a VAT invoice to C, which claims the VAT it has paid to B as input tax. C (to whom the Commissioners refer as a "broker") then sells the goods to a registered trader in another Member State: the hallmark of the simplest fraud is that this purchaser is A, and it is this circularity which gives rise to the term "carousel fraud". C has an input tax claim but, because its sale to A is zero-rated in C's own Member State, is not required to account for any output tax. The result, if the fraud is successful, is that B has received, but not accounted for, the VAT which the tax authorities must pay to C. In this situation B is certainly fraudulent and A probably so; C may well be entirely ignorant of what is happening, and of the use which is being made of its participation.

B's objective, as the Commissioners perceive the matter, is not to buy and sell goods in the ordinary course of business, but to put itself in a position to receive the VAT – or, as the Commissioners contend, the purported VAT – on its sale, for which it has no intention of accounting; the sole purpose of the transaction, from B's point of view, is to profit by the receipt of the VAT, or purported VAT, from C. The goods are no more than a token, necessary to lend verisimilitude to the transactions. Whether its activities amount to a fraud on the taxing authorities or on C is, in effect, the issue we must determine.

B is indifferent as to whether C is allowed input tax credit, but plainly B must be able to deliver a valid VAT invoice if C is to be persuaded to buy the goods and its indifference may be coloured if it intends to repeat the exercise by selling more goods to C. A (at least, if it is a participant in the fraud) likewise has no genuine business motive in buying back that which it has just sold. Though it is possible that A, too, is an unwitting participant if the goods are sold to it by C for a price lower than that at which it has sold them to B (or lower than a price for which, allowing for market movements, it could sell them in the future) the Commissioners' theory is that A and B are acting in concert and that a re-purchase by A is desirable, partly for reasons of control and partly because it is simpler and cheaper to use a finite quantity of goods several times rather than buy a fresh supply in the open market for each sequence.

Part of the Commissioners' case is that B, knowing it will gain from its failure to account for the output tax it has incurred, can afford to sell the goods to C for less than it paid to A, and will do so if this is necessary in order to secure a sale. In some cases, we were told, the goods are fictitious and the "transactions" are merely paper entries, but it is not suggested there was any fiction, or sham, about any of the transactions in which Bond House was involved.

It is said in Mr Birchfield's statement that the loss to the UK exchequer of carousel fraud in the year 2000-2001 was estimated at between £1.7 and £2.6 billion. The appellant challenged that estimate, on the basis that the source of the information on which it was based is not stated and its accuracy must therefore be doubtful. That may or may not be so but it seems to us to be an immaterial objection since the fact is that that is the Commissioners' perception and the reason why they attempt to detect and prevent such frauds. It was apparent from the documentation we saw that there is a similar perception in other EU Member States, and frauds perpetrated in intra-community trade (of which carousel fraud is an example) were the subject of the Court of Auditors' Special Report No 9/98. We are in no doubt that carousel fraud, whatever the scale of the loss may truly be, is a significant problem.

Since the simplest carousel fraud, consisting of a succession of sales of precisely the same goods from a registered trader in another Member State (an "EU trader") to the missing trader, from the missing trader to the broker, and from the broker back to the originating EU trader, is easily detected and easily proven, and those perpetrating frauds of this kind expose themselves to prosecution (we have mentioned already that some of those involved in the series of transactions which preceded Bond House's purchases were currently being prosecuted), attempts are almost always made to conceal the true nature of the arrangements. Additional steps – that is, further transactions – may be inserted between B and C. The parties to the additional transactions (to whom the Commissioners refer as "buffers") may or may not be ignorant of what is happening. The additional transactions are designed to make it harder for the taxing authorities to trace the sequence. Other variations, we were told, are that B may not be a VAT–registered trader at all, but may have used (the Commissioners say "hijacked") a genuine trader's VAT number. More sophisticated arrangements still require the participation of more than one EU trader acting in concert and of more than one broker; usually, we were told, there will be more than one missing trader (each in a different chain) and more than one buffer company in each chain, though the latter was not always the case. The goods might progress, not round a circle, but round a figure-of-eight or in another more complicated pattern; the word "spirograph" was suggested during the course of argument. Plainly combinations of the variations are possible, although the initial sale from A to B – that is, a sale from one Member State to another – must remain since this step, and the consequent ability of the missing trader to avoid paying output tax, is necessary if the fraud is to succeed at all.

The "motive power", as the Commissioner described it, of those carousel frauds in which goods actually changed hands and in which some of the parties were innocent of knowing participation was the notional "loss" suffered by B, in selling the goods to the first buffer company, or direct to the broker, C, for less than the price it paid to A. That "loss" was always less than the amount of tax for which B did not account (since otherwise there was no purpose to the sequence) but was sufficient to enable each of the buffer companies and the broker to buy and sell at a profit (although we were to hear evidence of sales by buffer companies at cost). The longer the series of transactions before the goods were sold back to the originating Member State the more participants there were who needed to make some profit, and the greater B's notional loss had to be. There was therefore, we deduce, a balance to be struck between concealing the fraud by extending the chain, and keeping it short so as to maximise the profit.

If the Commissioners are right, it follows that carousel frauds all have the same essential characteristic of circularity – that is, the chips pass through the hands of several traders but always, ultimately, find their way back to their starting point. Plainly they will not do so, except very occasionally, by coincidence and the carousel must be organised by one of the participants – the "ringmaster", as Mr Peacock described that participant. In the simpler case only one participant might be fraudulent, but in the more sophisticated situation, it seemed to us, more than one would need to be a knowing participant, even though only one would be the ringmaster. The ringmaster would, in particular, so organise the chain that an unwitting broker would be offered a supply of goods and a ready buyer would be easily found, or the broker would instead receive an offer of purchase and quickly identify the necessary supply, in each case with a sufficiently attractive profit margin.

In their efforts to detect and prevent such frauds, the Commissioners have developed three concepts, which they term "specific circularity", "general circularity" and the "ring fence." These concepts were developed in detail in some of the statements produced to us, and in an annex to the respondents' statement of case. However many parties may be involved, the ringmaster's need to control the carousel requires that the totality of the transactions remain within what the Commissioners term a "ring fence." Within the ring fence, there may be instances of the same goods reaching the same trader more than once: the Commissioners refer to that as "specific circularity." Where there is an identifiable series of linked transactions, but it cannot be shown that the identical goods reached the same trader more than once, they say there is "general circularity." That arises particularly where consignments are split and combined.

A simple example of this process, the Commissioners say, is where the missing trader divides a consignment, selling one part to each of two (or more) buffers. They may sell their parts to the same broker who, unwittingly, combines the two or more parts and sells what is now the reconstituted original consignment back to the EU trader. More commonly, the Commissioners' theory goes, a split consignment passes through two different brokers, or a buffer or broker combines goods acquired from two different missing traders or buffers respectively. Clearly many permutations are possible, and the number of possibilities multiplies as the number of traders involved increases.

Mr Birchfield states that the Commissioners' verification process, when they are presented with large repayment claims, requires them to consider whether there is a carousel, whether the claimant is, or appears to be, a broker, and whether one of five other factors is present – that the whole sequence is fraudulent in that the goods are fictitious or the apparent participants are in reality the same person; some or all of the steps are a sham; some or all of the transactions are devoid of economic substance in the sense developed in Halifax plc and others v Commissioners of Customs and Excise [2001] V & DR 73; there has been an "abuse of rights"; or the claimant has, in reality imported the goods itself (the missing trader being no more than an agent, or "front", for the EU trader) and its claim should be offset by the acquisition tax for which it was required to account. It was accepted by Mr Peacock that of these, only the third could apply in this case (although he put the argument in a rather different way from that in which it had been pursued in Halifax). It was on this ground that the Commissioners had relied in the Administrative Court.

The Burden of Proof

Before we turn to consider the evidence and to set out our findings of fact, we think it appropriate that we examine the incidence of the burden of proof which, the appellant contends, does not rest entirely on its shoulders.

Although articles 17 and 18 of the Sixth Directive confer a right to deduct input tax (or, in the case of a repayment trader such as Bond House, to claim payment of the surplus of input tax over output tax) they do not do so entirely without qualification. The primary qualification, at least in the context of this case, is that it is for the trader to establish his entitlement. In Rompelman v Minister van Financiën (Case 268/83) [1985] ECR 655 at 665, paragraph 24, the Court of Justice said "that it is for the person applying to deduct VAT to show that the conditions for deduction are met and in particular that he is a taxable person."

Here, there is no dispute that Bond House is, or was, a taxable person or that it had complied with the documentary requirements; the issue is whether the condition is met that the relevant transactions were economic activities and correspondingly within the scope of the tax. Ordinarily, too, the burden of proof before this tribunal lies on an appellant: see, for example, Tynewydd Labour Working Men's Club and Institute Ltd v Customs and Excise Commissioners [1979] STC 570 at 580, per Forbes J:

" … any taxpayer who appeals to the tribunal takes upon himself the burden of proving the assertion he makes, namely that the assessment is wrong … ".

See also Grunwick Processing Laboratories Ltd v Customs and Excise Commissioners [1986] STC 441.

Dr Lasok, however, made the point that his clients, having done all that was normally considered necessary in order to receive an input tax credit, could not reasonably be expected to discharge the burden of displacing, on the balance of probabilities, the factual premise of the Commissioners' contention that the transaction lacked economic substance (if, contrary to his principal arguments, the lack of economic substance, even if it should be established, was the relevant criterion). The facts and other matters on which the Commissioners relied – essentially, the nature of the transactions to which Bond House was not a party – were within their knowledge, or at least available to them, but were unknown to Bond House, and inaccessible to it: it could look no further than its own suppliers and customers.

Even without recourse to article 6 of the European Convention on Human Rights it seems to us that there is some merit in Dr Lasok's argument. With a few exceptions, such as s 60(7) of the VAT Act and s 16(6) of the Finance Act 1994, the rule that the burden of proof before this tribunal rests upon an appellant does not depend upon statute, but on the general (and common-sense) rule that he who seeks to advance a case bears the burden of establishing it. That is, in essence, what was said by Forbes J in Tynewydd Working Men's Club. Here, it is the respondents who seek to persuade the tribunal that the transactions in which Bond House was engaged were not what they seem, but in reality something else; and in our view there is logic in the proposition that the burden of doing so must rest on them.

Although he perhaps did not quite concede the point, Mr Peacock, while relying on Rompelman as the demonstration of the fundamental rule, rather side-stepped the issue by maintaining that the respondents had discharged any burden which might lie upon them of proving the facts; thereafter, he said, it was for the appellant to show that, despite those facts, it was entitled to the input tax credit it claimed.

Since the respondents acknowledge that the relevant transactions were genuine in that the goods and the consideration changed hands, that Bond House has not knowingly been party to any fraudulent or otherwise dishonest activity and that it has obtained all the necessary documentation, we are of the view that it is for them to satisfy us, on the balance of probabilities, of the facts of each of the 27 individual transactions for which input tax credit has been withheld (or is sought to be recovered), and to demonstrate, to the same standard, that each was undertaken in the furtherance of fraudulent – or, at least, non-economic – activity. We have therefore approached our analysis of the evidence and the finding of facts on that footing. What follows our conclusion on those issues is, in our view, entirely a matter of law to which burden of proof is immaterial.

The evidence and our findings of fact

We propose in this section of our decision to deal with the evidence not by relating the detail of what was said by the witnesses, or is set out in their statements, one by one, thereafter describing our conclusions, but instead to deal with the evidence, oral and written, on the various matters of dispute, setting out our analysis of and conclusions on that evidence before moving on to the next issue. Even though few of the matters in dispute can be regarded as self-contained, since each has a bearing in one way or another on one or more other issues, we have concluded that understanding is improved by the division of the matters we must consider, so far as is possible, into discrete elements. At the end of this section we will draw the threads together.

We heard oral evidence from Michael Cheetham, Bond House's technical director (and one of the partners in the original business); from Mr Lumb, the author of the letter of 19 September 2002 and who was responsible for the decision that the input tax claim should not be met and for the assessment which together are the subject-matter of the appeal; from Angela Norris, another Customs officer who had been closely involved in the case; from Richard Marshall, who is a security officer of Intel, the manufacturers of all of the chips sold in the transactions with which we are concerned; and from Philip Costello, an officer of the Irish Revenue department. We had one or more statements from each of those witnesses, apart from Mr Marshall. Some of the statements had been before the Administrative Court in connection with the application for judicial review. The statements of several other witnesses, not called to give oral evidence, were also available to us. Dr Cheetham's evidence included a visit to Bond House's offices at Castleford. We had extensive documentary evidence to which counsel – and in particular Mr Patchett-Joyce, who dealt on its behalf with factual matters – were to take us in some detail.

After the hearing was concluded, we were asked by the parties to admit further argument about documents which had been identified during the hearing but which had not been produced to us and we duly heard that further argument. In consequence it was agreed we would allow in evidence the transcripts of various telephone conversations between Dr Cheetham and a number of Customs officers (with the transcript of a single conversation between Bond House's sales director and a Customs officer), those conversations having been tape recorded by Bond House, together with extracts from the notebooks of several Customs officers, some of whom were also parties to the telephone conversations. We were provided with the latter in unedited form, and also in the redacted form in which they were provided to Bond House and its advisers; sensitive material had been deleted from the latter. We should say at this point that we detected from comparison of the two versions no eliminated material which had any bearing on the matters which we must decide.

Lastly, we had extracts from replies given by Customs officers to enquiries made by their counterparts in other EU Member States and which related to Bond House. Those replies were given in accordance with Council Regulation 218/92/EEC, on administrative co-operation, and, because the consent to their disclosure of the other Member States had not been obtained we had only the replies, and not the questions. The parties put in an agreed summary, though in terms which left it open to us to come to our own conclusions.

The market in computer chips

The trend of the market in personal computers and components, particularly CPUs, was a matter of some debate. Among the documentation produced to us was some published material relating to the subject. We were not taken to it in any detail and it was not formally proved or, in most cases, even clearly identified, but we were invited to consider it and have done so. The gist of it was put to Dr Cheetham as he gave evidence.

The documentation contained some statistical evidence that, in 2001 and continuing into 2002, the worldwide demand for personal computers declined – in 2001 by some 5% by comparison with the level in 2000. Dr Cheetham, while accepting that there was an overall reduction in demand for new computers, told us that the diminution in sales of new computers was, at least in part, balanced by an increase in sales of processor chips for the upgrade market; that possibility was not excluded by the documentation, though we did not form the impression that the increase in the latter wholly balanced the decline in the former.

Dr Cheetham also said that most of Bond House's sales were to the Irish Republic, whose economy was very buoyant. We are willing to accept that proposition and to accept too that Bond House viewed buoyancy of the Irish economy as the factor fuelling demand but we are sure it was a false perception since (as will become apparent later in this decision) many, if not all, of the chips sold by Bond House to Irish companies were immediately resold to UK traders, and were plainly not satisfying any demand in Ireland. In addition, while we understood Dr Cheetham's evidence that Bond House preferred to buy its supplies within the UK, to minimise transport and security costs (even though these were generally the seller's responsibility), we did not understand why it preferred to sell, not within the UK, but by export to Ireland – simultaneously increasing both the cost of carriage and its exposure to risk which, sometimes, it bore itself.

Dr Cheetham explained (this was borne out by the market information we had) that CPUs for the personal computer market are almost all made by one of two manufacturers – Intel, which has about 85% of the world market, and AMD. Although Bond House dealt in AMD chips all of the sales and purchases with which we are concerned were of Intel Pentium 4 chips, mainly those with a clock speed of 1.7 GHz, though a few were rated at 2.0 GHz. Large quantities of such chips, he said, are sold by the manufacturer direct to companies which install them in new computers – such a company is known as an original equipment manufacturer ("OEM"). Sometimes an OEM will order too many chips for its needs either by mistake or by design; conversely, an OEM may find it has ordered too few chips for its manufacturing needs and cannot make up the shortage quickly by ordering from the chip manufacturer. There is thus a "grey" market in which OEMs with too many chips sell, and those with too few buy.

Since large OEMs, buying many chips at a time, can do so at discounted prices, lower than those prevailing in the grey market, they commonly order more than they need for their own manufacturing volumes, with a view to selling the surplus into the grey market at a profit. There are, in addition, small OEMS, who cannot buy in the large volumes in which the manufacturers are prepared to sell, and traders whose business is upgrading existing computers and who need CPUs in order to do so. They, too, buy (and sometimes sell) in the same market. There is a comparatively modest retail market, aimed primarily at amateurs, but the chips with which we are concerned were packaged for bulk purchase and not for the retail market, for which chips are packed singly.

It was in this grey market that Bond House traded. It was, Dr Cheetham maintained, much like any other commodity market in which dealers constantly bought and sold. Bond House always did so on its own account, and never as an agent for others. It employed a team of dealers, supervised by its sales director. Because computer chips are constantly undergoing improvement – Dr Cheetham explained that Intel, for example, produced a faster (and therefore more desirable) chip at regular intervals – chips tended to lose value (or, more accurately perhaps, their prices moved downwards) over the fairly short term. The result was that Bond House and similar traders were almost always dealing in a bear market; there were exceptions when there was a surge in demand or an unexpected drop in supply (Dr Cheetham gave the example of an earthquake in the Far East which had disrupted supplies), but it was evident to us that such occasions were unusual.

Bond House avoided holding stocks of a potentially depreciating asset (save for modest quantities destined for small manufacturers, upgrades and the retail market) but instead, with very limited exceptions, always secured a sale before concluding a purchase. Dr Cheetham told us that a pair of transactions – that is, a purchase by Bond House and a corresponding sale – might be initiated by a purchaser whose requirements Bond House would endeavour to meet, or by a seller, which would offer chips which Bond House would endeavour to sell on. In either case Bond House would conclude both deals at the same time. It never sold "short" and it only very unusually bought for stock, although it so happened that it had held stocks of chips, worth about £1.8 million, in May 2002. Dr Cheetham was unable to recall why that should have been so. We observe in passing that of the 27 purchases with which we are concerned, only five had a pre-tax value of less than £1.8 million.

Bond House's business methods

Dr Cheetham described Bond House's history, as we have very briefly set it out at the beginning of this decision. We record at the outset that we have no doubt that Bond House is, or was, a serious, professionally managed concern which carried on its business honestly. We were particularly impressed with the extensive and, without doubt, expensive security measures it had installed at its premises, designed to protect the stocks it had on hand at any time from theft which, Dr Cheetham told us, was an ever-present threat (indeed, by unhappy coincidence, a large theft of computer chips took place during the course of the hearing, though Bond House was not its victim).

It was explained to us at the site visit that Bond House had been one of a group of three associated companies which shared the accommodation within the building although, save for the reception facility and a few other common parts, each had its own segregated area. Because Bond House had ceased trading it had let off most of its area of the building to an unrelated transport company. A second company within the group had been sold to its managing director, partly because he wished to protect his livelihood and partly in order that Bond House could raise money to meet the legal fees it had incurred in dealing with the present dispute. The third company continued to trade, in the computer components business (selling mainly to small manufacturers and into the retail trade) and we were able to see its operation, including the goods stocked within the warehouse. We also saw those areas of the building which had been used by Bond House, and were able to form an impression of manner in which it carried on its business, on which Dr Cheetham was also to give some detailed evidence.

He told us that the dealers spent much of their working day on telephone calls, to potential suppliers and customers, keeping abreast of market trends, particularly in relation to price, and changes in supply and demand. Some of their information was obtained from the internet; we understood that chips were commonly offered for sale on dealers' websites. Most deals, however, were negotiated and agreed by telephone, and later confirmed in writing, often by fax but sometimes by e-mail. Bond House's practice varied depending upon its course of dealing with the supplier or customer in question. With those it knew well, it would take a fairly relaxed approach, and accept subsequent written confirmation; with dealers with whom it had not established a relationship of trust it would insist upon written confirmation before it considered an agreement was concluded. In either case, its rule was that no single team member could undertake a deal without its being reviewed by at least one other team member – and in many cases by the sales director. This rule was imposed for two reasons: to ensure that deals were made at a realistic market value, and to safeguard against fraud. We were to hear a good deal more evidence about the efforts Bond House made to safeguard itself against entering into fraudulent transactions, to which we will return.

Bond House also had a "rule of thumb" that its dealers were expected to aim for a profit margin of £2 per CPU, though they would sometimes accept as little as £1 if they could do no better. If the margin were less, the profit would be overtaken by the expenses of the deal; the acceptable margin had to take account of the incidence of carriage, insurance and storage costs. There was no hard and fast rule in the trade about these costs, which were borne sometimes by the seller and sometimes by the buyer: this was a matter negotiated individually for each transaction. Bond House aimed to pass on the costs of carriage and insurance to its purchaser, paying the charges itself but including the amount incurred on its invoice. Sometimes, in order to secure a sale, it bore them itself

Dr Cheetham told us that the appellant also had strict rules about payment. It would in no circumstances release goods to a purchaser before the price had been paid (with very rare exceptions for credit sales) and it would not itself pay for goods until they had been inspected and a representative of Bond House had taken possession of them. This evidence was borne out by Bond House's printed terms of trade, which, as was demonstrated, were also reproduced on its website. Dr Cheetham's evidence was that Bond House's reputation and standing in the trade allowed it to impose this apparently somewhat one-sided arrangement, though as the evidence developed and we examined the documentation it seemed to us that, in a practical sense, Bond House's approach did not differ greatly from that adopted by other traders, most of whom imposed (or claimed to impose) similar terms of trade. They all seemed to include retention of title clauses. We also learnt from the transcript of one of the telephone conversations with which we were provided that the system was not entirely foolproof and that, by judicious prevarication, it was possible for a seller to receive Bond House's money and use it to pay its own supplier for the goods before Bond House's driver removed them from the place of delivery.

Because of the high value of the goods, and the relatively compact size of a consignment (a single box of chips of the kind in which Bond House traded, capable of being carried in one hand, could have a value of as much as £500,000) there was an ever-present risk of theft and consignments of chips were commonly transferred from seller to buyer within a secure warehouse – that is, a facility owned by an independent warehouse keeper from whom the seller rented storage space. Delivery was effected by the seller instructing the warehouse keeper to "release" the goods to the buyer; this was done when the seller was satisfied that payment for the goods had been received. In some cases both seller and buyer might be customers of the same warehouse keeper and the goods would be "released" from one customer to the other without the goods themselves moving at all, though it appeared that those of Bond House's purchases with which we are concerned were invariably, or almost invariably, removed from the warehouse (perhaps after overnight storage if delivery had been taken late in the day) for immediate transfer to an aircraft for removal to Ireland.

It appeared to us that Bond House was unusual, in its trade, by being based in the north of England – in Castleford, West Yorkshire. Most of the other UK-based companies with which it dealt were located in the London area, and we understood that all of the relevant imports and exports (which terms we use as a convenient, even if pedantically inaccurate, description of acquisitions of goods from, and their removal to another Member State for acquisition by a registered trader there respectively) arrived or left via Heathrow airport. All of the secure warehousing facilities in the United Kingdom of which we heard are close to Heathrow (except for Bond House's own secure facilities, which are at Castleford), and it appeared that corresponding facilities in Ireland are, similarly, close to Dublin airport. When goods sold by Bond House arrived in Ireland, the consignment was, we understood, usually removed immediately to a secure warehouse where it awaited release to the Irish purchaser, which was effected once Bond House had received payment.

Security risks

We have mentioned Dr Cheetham's evidence that consignments were inspected before Bond House would accept delivery of them. At the material time, Intel chips intended for bulk users were usually packed in boxes of 288 or 336. The explanation which we were later to hear from Mr Marshall was that those were convenient numbers for packing purposes at that time, allowing for the size and shape of the chips. As new chips, of a different size or, possibly, shape were introduced, it would be convenient to pack different numbers in a box. Intel were willing, Mr Marshall said, to produce boxes containing different quantities of chips, but most larger OEMs (for which such boxes were mainly destined) were content to accept boxes containing those quantities. They were willing to order, and accept, chips in quantities which approximated fairly closely to, but did not exactly correspond with, their needs. Dr Cheetham told us that this practice was reflected generally in the trade; an order for, say, 20,000 chips would be regarded as satisfied by the supply of a close approximation made up of multiples of 288, 336 or a combination of the two.

The boxes, of which we saw some examples, are made of stout brown cardboard. Inside, there are light plastic frames in which the individual chips are placed and secured for transport; unlike those intended for retail sale they are not packed in separate boxes and they have no cooling fans attached. Each cardboard box is sealed with adhesive tape and with other seals designed to ensure that if a box is opened, that fact is obvious, and to make counterfeiting – apparently another ever-present problem – more difficult. Each box also bears a label on which are printed various numbers and bar codes. Mr Marshall explained that the numbers identify which of Intel's manufacturing plants produced the chips, in which production run they were made, their specifications and details of that kind. The numbers are reflected in the bar code and each box should bear a unique set of numbers and, correspondingly, a unique bar code.

Dr Cheetham told us that Bond House had detected some duplications. Mr Marshall did not rule out that possibility entirely – it might happen, he said, if there were a breakdown in the equipment with which Intel printed the labels – but he also described the safeguards which Intel had introduced to prevent such an occurrence, and to rectify it without delay (and before the goods left Intel's premises) should it occur. We are quite sure that Mr Marshall was going out of his way to be scrupulously fair, and that in truth the possibility that Intel would encode two boxes identically, and that they would be released to a purchaser before correction, is so remote that it can be regarded as fanciful. On the other hand, we accept that a counterfeiter might well copy a genuine Intel label, and that duplications might come onto the market by this means. Indeed, we heard evidence from Dr Cheetham about a box of counterfeited chips which had been detected by reason of its bearing a duplicated number.

Dr Cheetham was at pains to explain that Bond House was very conscious of the security risks of its business, not only of theft, but of counterfeiting; and in other respects to which we will come shortly. Before it accepted delivery of a consignment Bond House would usually (though we understood not invariably) use a bar code scanner to check the bar code printed on the box label, and have the boxes x-rayed.

Each of Bond House's van drivers – to whom the task of checking consignments before they were accepted for delivery was largely delegated – was equipped with a bar code scanner. The scanners were capable of reading the bar codes and of checking that they contained the information they should contain, in a way which made sense, and of verifying that the code did not correspond to that of a stolen box; details of known stolen boxes were stored in the scanners' memories for the purpose of the verification. Dr Cheetham told us that Bond House was, in early 2002, in the process of developing its scanners further, so that they could communicate by radio with Bond House's offices at Castleford in order to carry out a more sophisticated check of the bar codes, to verify in particular whether the label was a duplication, or whether it identified a box which had already been within Bond House's ownership. We shall return to the greater significance of scanning later in this decision.

The purpose of subjecting boxes to x-ray examination was to ensure that they contained what they purported to contain, that is a certain number of processor chips of a particular type. The examination was done in this way because a box which had been opened, for visual inspection, and whose seals had consequently been broken, so Dr Cheetham told us, was worth less than one with its seals intact. He said that boxes had as many as five seals, and that one with, say, three seals intact was worth more than one with only one seal intact, though it did not become clear to us why boxes needed more than one seal, nor why anyone might break some but not all of the seals. Bond House did not have its own x-ray facilities but instead relied upon x-rays undertaken by the secure warehouses it used for overnight storage of chips which it had bought and which were awaiting delivery to a customer, and another at whose premises several of its purchases were released to it, since the sellers were customers of that warehouse. Some of the x-ray certificates produced by that warehouse, Forward Logistics (Heathrow) Ltd, were to be of importance.

The second type of risk of which Bond House was very conscious, Dr Cheetham explained, was that it could find itself dealing with other traders who were dishonest or in some other way not trustworthy, and, he said, it went to considerable lengths to minimise that risk. We have already mentioned its rule that that no single trading team member could conclude a deal. We were shown, at the site visit, the files which Bond House kept of the results of searches and enquiries it had made about the traders with whom it did, or had considered doing, some business. There were also copies of some of the search results in the bundles. These included searches at the Companies Registry and credit enquiries, made of the usual trade reference agencies, bank account details and copy VAT registration certificates.

In addition Bond House (in common, it emerged, with many of the other traders with which it dealt) checked with the Commissioners that the particulars given to it by a trader corresponded with the Commissioners' own records – particularly that the VAT registration number given was genuine and belonged to that trader, and that other details, such as address and telephone number, matched. All of the checks were repeated by Bond House at fairly frequent intervals.

We mention at this point that we accept that Bond House undertook those enquiries and that, if it received an adverse report from the Commissioners, it refused to deal with the trader concerned. Although we think it probable that, by and large, Bond House was also cautious about the results of other enquiries it made, and did not take chances in order to secure particularly attractive contracts, we were to hear more detailed evidence about its enquiries relating to two of its suppliers, with which we will deal when we come to consider the 27 transactions.

Bond House's relations with the respondents

Dr Cheetham's evidence was that he (as the director primarily responsible for such matters) maintained particularly close contacts with the respondents, initially at Bond House's local office at Leeds, but latterly at Redhill when responsibility for matters of this kind – essentially assurance of traders employed in high-risk transactions – was moved there; this change occurred as recently as the beginning of May 2002. Those contacts, he said, were not limited to enquiries about other traders. He provided the officers whom he met and to whom he spoke, on many occasions, by telephone, with large quantities of market intelligence which, he believed, was of considerable help to the Commissioners in their efforts to tackle fraud in the industry. In return, the officers were more forthcoming with him than they otherwise might have been, warning him in some cases, while in others giving a trader a clean bill of health. In one case, he said, they went so far as to give a trader a glowing reference.

We did not hear oral evidence from any of the officers at Leeds or Redhill who were present at the discussions to which Dr Cheetham referred, but we had the transcripts of the telephone conversations and the officers' notebooks, admitted in evidence following the supplementary hearing, and the unchallenged statement of Roderick Stone, an officer employed in the Redhill office, with which we will deal first.

Mr Stone is evidently a senior Customs and Excise officer, although we have not been told his precise rank. He is in charge of the office at Redhill whose task is the detection of carousel frauds in the computer chip and mobile telephone markets. In late April 2002 it was decided that the work of verifying traders' VAT registrations, in the manner Dr Cheetham had described, should be transferred from local offices to Mr Stone's team at Redhill, and it was for this reason Bond House had been asked to deal with Redhill rather than Leeds. Mr Stone's statement also describes the Commissioners' policy of asking traders such as Bond House for their help in preventing fraud by providing information about other traders and their transactions although, as he mentions, there were no powers of compulsion. It is also evident from his statement that the information-gathering was for the Commissioners', rather than the trader's, benefit.

Mr Stone and Dr Cheetham had a meeting at Redhill on 28 May 2002, at Dr Cheetham's request. Unfortunately, their respective recollections of that meeting are at some variance, though we have also the (apparently, although it is undated) contemporaneous note of David Sweeting, a VAT consultant who accompanied Dr Cheetham. Dr Cheetham made an extensive note of the meeting shortly afterwards; Mr Stone, it seems, did not, and his statement about it was not made until nearly three months later. Nevertheless, and despite our not having heard his oral evidence, we prefer Mr Stone's recollection of the meeting. We are fortified in that view by the contents of Mr Sweeting's account, which is much closer to Mr Stone's recollection than to Dr Cheetham's.

The meeting is of some importance for the light it casts upon Bond House's relations with the Commissioners. There is no dispute that the principal topics of conversation were Bond House's concern about its position in what Bond House itself perceived to be a high-risk industry, and its willingness to assist the Commissioners. It appears that Mr Stone was initially under the impression that Bond House was not a cooperative trader (it had, of course, only recently begun communicating with Redhill and we deduce that Mr Stone had had no occasion until then to make enquiries about it), but this misapprehension was evidently quickly dispelled and in his statement Mr Stone confirms Dr Cheetham's evidence that Bond House was regularly carrying out checks on other traders' VAT registrations, and that it was indeed considered by Customs to be a cooperative trader. However, Mr Stone's statement does not support the conclusion that Bond House did much more than verify other traders' VAT registrations, and one does not gain the impression that the Commissioners had, or had had, any sort of special relationship with Bond House. There is likewise no hint in Mr Sweeting's note that this was the case.

At the meeting Dr Cheetham referred to Bond House as a "legitimate pawn caught up in a world of mire" (as he put it in his note) and it is apparent from all three of the accounts of the meeting available to us that Dr Cheetham was anxious to ensure that Bond House should avoid any possibility of being penalised, and volunteered that it would do everything required of it to assist the Commissioners, although there is some difference between them about precisely what he offered, and what Mr Stone requested. The only one of those differences which we need to mention is Dr Cheetham's assertion that he discussed Bond House's scanning system and the development of it to which we have already referred, while Mr Stone recalls that scanning was mentioned briefly, but that Dr Cheetham did not elaborate. Mr Sweeting's note indicates that scanning was mentioned, but dismissed by Mr Stone as being of little utility, and it does not support Dr Cheetham's evidence that a new system was discussed, or even under development. Moreover, as we read Dr Cheetham's note, it too is inconsistent with his evidence that such a system was under development; rather, it appears to have been no more than an idea for the future. That conclusion is borne out by the analysis undertaken by Mr Peacock, in the course of his cross-examination of Dr Cheetham, of a report prepared for Bond House's bankers by Kidsons Impey, the accountants, on Bond House's security systems. That report, likewise, does not correspond with Dr Cheetham's recollection that a "third generation" scanning system, as he described it, was actually in development.

The most important point about the meeting, in our view, is that Dr Cheetham came away from it with the impression that Bond House was considered by the Commissioners to be doing everything it reasonably could to ensure that it entered only into genuine transactions, that it was regarded by them as a legitimate trader and that, as long as it continued to conduct its business as it was doing, it could rely on the Commissioners to police the industry, would not itself be the target of investigation and had nothing to fear. In particular, he believed that it was not the Commissioners' intention to withhold input tax recovery from legitimate traders. Mr Stone's recollection, by contrast, is that he indicated that a trader innocent of knowing involvement in fraudulent activity was unlikely to be prosecuted, but that he gave no similar assurance regarding any other investigative or regulatory measures which the Commissioners might take. On this issue too Mr Sweeting's note supports Mr Stone's evidence.

We observe, before leaving this meeting, that Mr Stone's statement refers to an exchange about the end use of the chips in which Bond House was trading, and his expression of doubt that there was any true end use. The exchange is mentioned in Mr Sweeting's account, but not in Dr Cheetham's note. Dr Cheetham was to tell us in his evidence that he did not know the identity of the end user of the chips, and concerned himself only with Bond House's transactions, looking no further in either direction; indeed, he could not do so since Bond House's supplier would not disclose the identity of its own supplier, and its customer the identity of its own customer, for fear of being by-passed.

Mr Stone's evidence deals only briefly with the checks which Bond House and similar traders made with the Commissioners' officers, but there was more information on that topic within the transcripts of the telephone conversations and the officers' notebooks. These reveal that Mr Stone had indeed initially under-estimated the extent of Bond House's communications with Customs officers and the scale of the information which Dr Cheetham and, to a lesser extent, others of Bond House's staff were passing on – that is, information which they had obtained, either from their dealings with other traders or from discussions within the trade. It is also clear from these documents that Bond House did indeed systematically carry out the checks on other traders' VAT registrations which we have mentioned.

However, the traffic is largely one way. Apart from the formal confirmations of VAT registrations (and, occasionally, an indication that the details do not tally) there is, with only one exception we could find, no evidence from these documents that the Customs officers made any overt further comment, either approving or adverse, about any other trader. On the contrary, it is recorded that Dr Cheetham was told that Customs officers could not express such opinions and the comment is made in one telephone conversation that he should know better than to ask. The one exception to which we have referred was the comment – we assume from the context unguarded – that another trader was "under suspicion".

That comment was made in response to one of Dr Cheetham's persistent, as we perceive them, suggestions that the officer to whom he was speaking had already made a comment – generally approving – of a trader, in the evident hope that the officer concerned would feel it necessary to confirm or deny what was being put to him. The transcripts suggest that the officers made only non-committal replies, though we recognise that much may have been conveyed by tone of voice, or indeed by silence, and there is one recorded reply which (given the avowed policy of making no such comment) can only be regarded as indiscreet, though not directly relevant to whether or not Bond House should do business with the object of the comment. Nevertheless, there is no recorded exchange which comes even close to a clean bill of health, still less to a "glowing reference", as Dr Cheetham contended had been given. The best that can be said is that the officer concerned did not take the opportunity to deny Dr Cheetham's prompt. With the limited exception of the two comments we have identified, the impression we have is that Customs were glad to have the information which Dr Cheetham was able to provide but gave very little in return.

In summary, it is apparent from all of the documentation, including that which we admitted following the supplementary hearing and from his own evidence, that Dr Cheetham, in particular, was in frequent contact with the respondents' officers at Leeds and, latterly, Redhill. It is apparent too that he was imparting information to those officers. Though we are unable to gauge the value of the information, it is also clear that the Commissioners were receptive to his telephone calls, and encouraged him to continue providing information. We accept therefore the proposition that Bond House was well-regarded by the Commissioners and was itself in no way the subject of suspicion. However, what is also apparent is that the officers gave little in return.

We have come to the conclusion, in relation to his meeting with Mr Stone and to Bond House's relations with the Commissioners as a whole, that Dr Cheetham's understandable concerns and his desire for reassurance have led him to believe that much greater reassurance was given than was in fact the case. We think it most unlikely that any Customs officer would give positive advice to any trader that he could safely do business with another; it is much more probable that an officer would go no further than to say that the Commissioners knew of no reason to the contrary. Even this goes rather further than the telephone conversation transcripts suggest, save for the two comments we have identified, although we are prepared to accept that hints may have been given by tone of voice or by silence. We are satisfied however that Mr Stone, a very experienced officer well aware of the nature of the Commissioners' continuing investigations, would have given no indication to Dr Cheetham that Bond House would not be the subject of investigation. We have concluded that this, too, was wishful thinking on Dr Cheetham's part and that Mr Stone is right in saying that co-operation by a trader would not absolve it from investigation or, indeed, any action that might appear appropriate.

The relevant transactions

The Commissioners have, as we have mentioned, disallowed the input tax attributable to 27 of Bond House's purchases in May 2002. We do not intend to deal with each of the 27 relevant transactions individually; to do so would add to the already considerable length of this decision without significant benefit, although we have ourselves nevertheless examined the details of each transaction. Rather, we intend to identify those features of the transactions as a group which we have found material, distinguishing where distinction is to be drawn, before outlining our conclusions about the transactions.

We start from the position that the Commissioners accept that all of Bond House's transactions were genuine in that the goods and the payments for them changed hands on every occasion. There was also no evidence before us to suggest that goods and payment did not change hands in respect of each transaction in the chain before and after Bond House had bought and sold.

There was evidence that in some cases the missing traders, or users of hijacked VAT registration numbers, were handing over the entire proceeds of sale of goods to the EU trader from whom they had acquired them, and not merely the price disclosed by the EU trader's invoice; and in others innuendo, rather than clear evidence, that they may have paid for the goods only after they had received payment from the first buffer trader. There was evidence that some at least of the buffers were using Bond House's money to fund their own purchases. The timing of the various payments and of the releases of the goods from the seller's warehouse keeper to the purchaser was a matter on which we heard a good deal of evidence, with which we will need to deal in some detail. However, there is nothing to suggest that the money did not ultimately change hands, nor that the transactions were shams in the sense that they did not happen at all, at any stage, or were of different goods, or of different quantities, from those represented to have been sold.

Dr Cheetham had prepared an analysis of all Bond House's purchases in May 2002 – not only of chips but of other items, such as stationery. It was apparent from his analysis that some of the chip purchases for which the input tax credit claimed had been allowed differed in no evident respect from some of those for which it had been disallowed; in particular, it had bought from the same suppliers and had sold to the same customers. Dr Cheetham, in his evidence, and Mr Patchett-Joyce, in his observations to us, were to make a good deal of this apparent inconsistency. It was, they suggested, indicative of the Commissioners' having based their conclusions on suspicion or guesswork, rather than on proper evidence; and it demonstrated, as they had complained from the outset, that the Commissioners had made up their minds and were trying to make the evidence fit their preconception. If their reasons for thinking that those transactions for which the input tax credit had been allowed were not offensive, on what basis, they asked, could it be said that those for which the credit had been disallowed were in some way not legitimate?

Mr Lumb's explanation, which we were to hear rather later, was that there was a distinction to be drawn between the two categories: in those cases where the credit had been disallowed, there was (as he regarded it) clear evidence of a carousel fraud, in that the necessary characteristics were demonstrable, whereas in the others, one or more elements were lacking, or at least he could not say with confidence that they were present, and he recognised therefore that the evidence was not sufficiently strong for him to feel justified in refusing the credit. Shortly put, he had given Bond House the benefit of the doubt.

Of the 27 purchases, nine were from SDP Ltd, 16 from V J Trading Ltd, and one each from companies called Larcom and TCHYP. In each case, the appellant's sale was by way of export to an Irish trader. It was not disputed by the appellant that all of the goods it bought from SDP were sold to Fancygrove Ltd, in Dublin; the other goods it acquired in these purchases were sold to one of three other Irish traders, but in no case to Fancygrove.

Among the documents were separate bundles of papers, each bundle headed by a summary or "deal sheet", providing details of the transactions before and after the appellant's purchases and sales, so far as the Commissioners have been able to trace them. We had in addition a narrative, forming annex D, reproduced in tabular format as annex E, to the statement of case, in which the Commissioners set out their perception of each series in which both Bond House and Fancygrove were participants; the nine transactions in which Bond House sold goods to Fancygrove form part of an overall total of 24 sets of transactions which, the Commissioners contend, demonstrate the carousel. We did not have a similar analysis of the remaining 18 of Bond House's transactions, but we have examined the bundles of papers relating to them. We have considered annexes D and E, those bundles and all the remaining documentation produced to us against the background of the oral and the other written evidence available, and we have borne in mind the Commissioners' descriptions of the characteristics of a carousel fraud, and our conclusions about the burden of proof.

In each case, if it as the Commissioners maintain, the essential constituents of a carousel fraud are to be found: the sale by an EU trader to a trader registered in the UK; a sale by that trader to another UK trader on which VAT is charged but not accounted for; and a sale by a UK trader to an EU trader. There is also at least one buffer trader in each series of transactions. We mention at the outset that we are satisfied from the evidence that in each of the cases the Commissioners have established, on the balance of probabilities, that the parties to the series of transactions, the quantities sold and the prices paid were as they are set out on the "deal sheets". There are, however, some dissimilarities between the transactions.

One striking feature of the appellant's trade is the number and value of the purchases it made from two suppliers – VJ Trading Ltd and SDP Ltd. We learnt from Dr Cheetham that VJ Trading had a single director, Victoria Powell, whom, and whose husband, Dr Cheetham had known for several years from earlier business dealings with a company by which they were then both employed. He told us that Mrs Powell wanted to set up her own business dealing in computer components but her husband – who, by all accounts, was in a financial position to do so – was for some reason unwilling to provide the necessary funds and Dr Cheetham did so instead, personally lending Mrs Powell as much as £500,000, at only 4% p.a. interest. He told us that the loan was secured by Mrs Powell's personal guarantee, backed by some investments which she had but which, we deduce, she could not realise at short notice, and that he had lent the money for only a fairly brief period while he considered what form his investment should take. In the event he decided not to invest and the loan had been repaid in full before May 2002. We are bound to say we found it puzzling that Dr Cheetham should lend money to someone else in order that she could set up a business in direct competition with his own.

More puzzling still was that Dr Cheetham did not seem to have questioned Mrs Powell's ability, within months after her business was established, to sell to Bond House nearly 400,000 chips worth (excluding VAT) almost £43 million in the space of a single month; indeed Bond House bought over £140 million worth of chips from V J Trading in March, April and May 2002. It does not seem to have occurred to Dr Cheetham or his fellow directors to wonder how Mrs Powell (then aged only 21) could so easily identify suppliers of such large quantities of chips at (they would have to assume) prices lower than those at which she was selling those chips to Bond House, and lower than those at which Bond House – by then established for about 10 years – could obtain them.

Dr Cheetham conceded in his evidence that V J Trading had not passed all of the checks which Bond House made; he was evidently prepared to overlook that failure because of his acquaintance with Mrs Powell, and because the check Bond House undertook with the Commissioners' office at Redhill, to verify V J Trading's VAT registration details, was satisfactory.

There was no connection (other than a trading relationship) between Bond House and SDP. This company had previously traded in the oil distribution business (SDP is the acronym of Storage Distribution Petroleum) – yet even in that business it had only a short history, having apparently been incorporated in March 2000. Its only director is shown, by the documents produced to us, to have decided in February 2002 to start dealing in computer chips; the same documents mention that the petroleum distribution business had by then been sold. A letter from the director (undated but evidently sent in February 2002) indicates that he intended to deal in 2000 chips per day. By Bond House's standards, this was a very modest quantity. However, in May 2002 Bond House alone bought over 300,000 chips, at a pre-tax value of £33.5 million, from SDP and, in the period of six weeks before it ceased to trade Bond House did business with SDP to a value of over £63 million. Again, neither Dr Cheetham nor any of his fellow-directors appears to have asked himself how a small company such as SDP could, in such a short space of time, have expanded its business to such a level, and how it too could identify a source of supply not available to Bond House.

It is conspicuous that, at the same time, Bond House's own turnover was increasing dramatically. One factor which attracted the Commissioners' attention was the rise in its input tax claims from an average of just over £2 million per month in 2000 to an average of £11 million in the first five months of 2002; the claim in May 2002 amounted in all to over £16 million. These figures reflected a corresponding increase in turnover from £258 million in 2001 to a projected £1.5 billion in 2002 (or £6 million per working day, and a near six-fold increase). Dr Cheetham explained that Bond House had been able to negotiate an increase in its overdraft limit from £2 million to £12 million at the end of 2001, and this enabled it to finance much larger deals than had previously been the case.

While we understand that point, it did not become apparent to us how Bond House was able to secure significantly larger deals against the background of a static, even if not declining, market; nor did we hear evidence that, before it secured an increase in its overdraft limit, Bond House was forced to turn down deals for want of the ability to finance them. Indeed, what Dr Cheetham told us on this point was difficult to reconcile with his evidence about the ability of VJ Trading and SDP to finance large deals despite their having little or no visible assets, at least if their balance sheets are to be relied upon. The explanation given by Dr Cheetham was that they were able to use the proceeds of each sale to finance the corresponding purchase. We did not understand how he reconciled that practice with Bond House's policy of refusing to pay for goods until the seller was able to deliver them with good title although, as we have mentioned already, that system does not seem to have been incapable of circumvention. It seems a little unlikely that VJ Trading and SDP, with very short trading histories, were able to buy such large quantities on credit. In addition, if they were able to finance deals in this way, it is by no means clear why Bond House was unable to do likewise, before its overdraft limit was increased.

Although other EU traders feature among Bond House's customers in May 2002, most were based in Ireland and the customer to which most of its sales, both in number and value, were made is Fancygrove Limited ("Fancygrove"), a company located in Dublin. The detailed evidence and argument we heard focused on the appellant's sales to that company.

Such evidence as was available to us about Fancygrove itself was to the effect that it was a substantial, established trader. Mr Peacock carefully avoided making any direct allegation that Fancygrove had indulged in fraudulent activity but he did not demur from the suggestion that, if the Commissioners' case was correct, Fancygrove must have been the ringmaster controlling the carousel, at least so far as the chips it bought itself are concerned. The Commissioners had asked their Irish counterparts to make some enquiries about Fancygrove, and those enquiries were undertaken by Mr Costello, whose evidence we heard on the last day of the hearing. He, too, was circumspect in what he said about Fancygrove. We were left with the impression that we had not been told of all of the Irish authorities' enquiries into the company. We bear in mind that Fancygrove is not a party to this appeal and has not been represented at the hearing, and that we have no evidence from any of its officers. We consider we must treat allegations, including implicit allegations, against that company with care.

The "missing trader" which appeared in those of Bond House's chains of transactions which also featured Fancygrove was Rangepace Ltd, a company registered for VAT in the United Kingdom. It appears to have had a sole director, one Luc Soudan, a Belgian national. It is not suggested by the Commissioners that Bond House had ever had any dealings with Rangepace, nor that any of its directors knew of it or of Mr Soudan, and Dr Cheetham told us that, until the current dispute began, he had never heard of either. We accept what he said.

It was demonstrated by unchallenged evidence included in the bundles that Rangepace was causing concern to the Commissioners by early 2002 to the extent that, using their powers under reg 25(1)(c) of the VAT Regulations 1995, they had required it to make at least one return to a non-standard date, and to render it within a week of that date. It appears that the return was made in accordance with that requirement. It showed that a very large amount of output tax was due, though almost wholly cancelled by the input tax for which credit was claimed. There remained a net sum due to the Commissioners of about £23,000, which was paid. We did not discover whether the accuracy of the return had ever been verified. The payment evidently allayed the Commissioners' concerns only to a very limited degree.

In July 2002 Rangepace was compulsorily deregistered. We did not discover the reasons for its deregistration, but were left to assume that it was due to the Commissioners' belief that Rangepace was involved in illicit activities. It was recorded in the Commissioners' documents of which we had copies that at deregistration it owed tax of about £17 million – that is, the Commissioners' estimate of the excess of its output tax over its input tax for which it had not accounted. That record, of course, begs the question whether it truly owed VAT at all.

Mr Patchett-Joyce was able to discover during an overnight adjournment, by using the internet, that one Luc Soudan was listed in the Belgian telephone directory at an address which appeared in Rangepace's documents filed at the Companies Registry, and that other addresses listed there also matched telephone directory entries. Mr Patchett-Joyce suggested that the Commissioners' suspicions about Rangepace might have been misplaced, or exaggerated. That is, of course, possible, just as it is possible that an entirely innocent Mr Soudan's identity has been "borrowed".

As in the case of Fancygrove, we bear in mind that we have heard nothing from Rangepace or from Mr Soudan. Nevertheless, and despite that caveat and what Mr Patchett-Joyce was able to discover, the evidence overwhelmingly supported the conclusion that Rangepace (whoever controlled it) was deregistered owing the Commissioners – if it truly owed anything – a large sum of money, and that it was a knowing participant in a fraud.

In every one of the nine cases, the Commissioners' analysis shows not only that the goods Bond House sold to Fancygrove were acquired by it from SDP, but also that SDP had in turn acquired them from Rangepace, and that Rangepace had acquired them from Fancygrove. That, by itself, suggests a very simple carousel, but the Commissioners' analysis shows that Fancygrove did not invariably sell the goods it had obtained from Bond House immediately back to Rangepace, but on several occasions to other UK traders who sold on, through one or more others, until the goods reached a broker which sold them back to Fancygrove. Fancygrove might then sell them to Rangepace or to another UK trader; thus there were several "loops" of transactions centred on Fancygrove. It is conspicuous that while SDP made a very small profit – generally as little as 25p per chip and on one occasion nothing – on its sales to Bond House, the profit margin achieved by Bond House on its sales to Fancygrove was in every case, and regardless of the quantity of chips or the purchase price it had paid, precisely £4 per chip.

We were able to examine Fancygrove's sales and purchase ledgers for May 2002, copies of which were produced by Mr Costello. Their authenticity and accuracy were not challenged. If they are correct it is in our view beyond doubt that Fancygrove sold a very large number of chips to Rangepace during that month. Although the evidence thereafter is rather less clear, we have no doubt that Rangepace sold a large number of chips to SDP which, in turn (and as is common ground), sold a large number to Bond House. The Commissioners contend that these were, in each case, the same chips. And that Bond House then sold the chips back to Fancygrove which then sold them, in most cases, to Rangepace again. If that hypothesis is right, it is difficult to understand how Rangepace alone could have ensured that it would happen.

The similar range of documents relating to the appellant's purchases from V J Trading reveal a more complicated pattern. Bond House's sales were made to one of three different Irish companies – none of them Fancygrove. V J Trading's purchases were all made from one of two companies, apparently themselves buffers, which had in turn bought their supplies from several missing traders, although it appears that each of those buffers had bought all of its supplies from a single missing trader. It is then shown that the Irish trader exported the goods back to the United Kingdom, selling them in each case directly to a missing trader. The profit margins achieved are again conspicuous – V J Trading consistently made 50p per chip, while Bond House, irrespective of quantity, purchase price or identity of purchaser, made precisely £3 per chip on every deal.

The chips obtained by Bond House from Larcom were sold to one of the Irish companies to which Bond House commonly sold chips bought from V J Trading, but the traders which preceded V J Trading in those chains did not feature in any of the transactions leading to Larcom. It was believed by Mr Lumb and Miss Norris that the chips had passed through the hands of a missing trader before reaching Larcom, but within the documents relating to this chain was an e-mail passing between two Customs officers indicating that the trader concerned had re-appeared. Whether or not it had accounted for the tax was not revealed. We observe that Larcom made only 50p per chip profit, while Bond House managed to achieve £3.50.

Bond House sold the chips it acquired from TCHYP to another of the Irish companies to which it had also sold many of the chips bought from V J Trading. In this case the Commissioners were able to demonstrate that TCHYP had bought the chips from a buffer which had in turn bought from a missing trader. By contrast with the other transactions, TCHYP achieved a profit of £4 per chip, and Bond House only £3.50.

Specific circularity

A considerable amount of time and energy was devoted, both at the hearing and in the antecedent correspondence, to the question of specific circularity – that is, whether the Commissioners could establish that any of the boxes of chips which had been traded in the 27 series of transactions had been owned by Bond House more than once. It will be apparent from our description of those series that none is demonstrably quite as simple as a sale by Bond House to an Irish trader immediately followed by a sale of the same chips to a UK missing trader leading in turn to a sale, through one or more buffers, of the same goods to Bond House, although it seems to us that those involving Fancygrove and SDP come close. Others appear to be a good deal more complicated. Nevertheless, Mr Lumb believed that Bond House must have owned the same goods more than once on several occasions and, in order to demonstrate that his conclusion was correct, he married the available documentary evidence relating to the various sales and purchases made by companies known, or at least believed, to have participated in transactions before and after Bond House to his extremely detailed analysis of the scanning records which had been obtained by the respondents from Forward Logistics.

These records were the subject of some detailed, but in part self-contradictory, evidence given, in the form of four statements, by the sole director of Forward Logistics, Angela Deane. Resolving the contradictions as best we can, we conclude that the bar codes on boxes of chips were scanned as the boxes arrived at Forward Logistics' warehouse, in order that Forward Logistics had a record of the goods on its premises, and primarily as evidence it could produce in the event of theft. It was also clear, however, that scans were not undertaken on every occasion, but only when time permitted.

The scan results were transferred to a computer, and they could be married to the files which Forward Logistics maintained for each of its customers' transactions. We deduce that no check was undertaken by Forward Logistics to ascertain whether the goods had been within its possession before; that was no part of its purpose in undertaking the scans. From those records which were available Mr Lumb was, however, able to show that large numbers of boxes had indeed been delivered to Forward Logistics on several occasions – up to seven times in some cases – and he was able in some instances to link those deliveries to Bond House's purchases.

Some of that evidence of specific circularity was compelling. Discarding, as we do, Dr Cheetham's suggestion that duplicated box numbers were a material factor, we accept the accuracy of Mr Lumb's analysis of the scanning records and his conclusion from that analysis that the same boxes were the subject of transactions in which they were temporarily stored and released by Forward Logistics on more than one occasion. The sales and purchase ledger records obtained by Mr Costello from Fancygrove demonstrate, beyond any reasonable doubt, that there are several instances of Fancygrove buying goods from Bond House and then immediately selling the same goods to Rangepace. Although box numbers are not listed in those records, Fancygrove have themselves identified goods purchased with goods sold, and we are satisfied they have done so by reference to the identity of the chips, and not merely their quantity; the nature of the records is inconsistent with the latter conclusion.

The frequency of the occasions on which a trader is shown to have bought a given number of chips and to have sold the same number of chips on the same day or the next is, by itself, striking. Dr Cheetham's evidence that Bond House, at least, did not buy chips unless it had already negotiated a sale (and therefore sold the identical chips, rather than merely the same number) must, we think, have been the practice of others within the chain. It is an inevitable conclusion from the evidence we heard about SDP and V J Trading that they could not have financed deals in any other way. It is equally clear, as we have already mentioned, that Fancygrove's sales to Rangepace led to sales from Rangepace to SDP and thence to Bond House. Bond House frequently supplied Fancygrove; even though the goods appear to have gone round at least one other circuit before returning to Fancygrove, it seems to us highly likely that they were eventually sold again to Rangepace. Those factors seem to us to lead inexorably to the conclusion that Bond House must, on the balance of probabilities, have owned the same chips on more than one occasion. Mr Lumb identified three specific occasions on which he believed it had occurred.

However, there was some evidence to the contrary. Mr Patchett-Joyce was able to show, by his cross-examination of Mr Lumb, that some of his conclusions on specific circularity were not necessarily correct. Indeed, even in the pre-hearing correspondence Mr Lumb conceded that his conclusion in respect of one of the three occasions was wrong. He made some further, albeit modest, concessions as he gave his evidence.

First, it was clear from the analyses which Mr Lumb and Miss Norris had undertaken that there had been a substantial measure of splitting and combining of consignments of chips. Mr Patchett-Joyce made much of their approach to the reconstruction of the deals, and of the assumptions which they necessarily made in order to show how the parts of a consignment which had been split could be traced to a later, combined consignment, or by some other means back to their originating trader. There was, we accept, some merit in his attack; although the officers' approach was arithmetically sound, other assumptions, also arithmetically sound, could be made which did not lead to the identical conclusion.

Second, Dr Cheetham told us that when the appellant's driver arrived at Forward Logistics' warehouse to collect chips which it had bought, and found that the seller had more chips stored in the warehouse than were necessary to fulfil the order, the driver would select the best boxes from the entire quantity, even if those destined for Bond House had already been set to one side. Although, at first, we found that a little surprising as a matter of ordinary commercial practice, it is confirmed by Mrs Deane's unchallenged evidence and by some of the evidence we heard about the X-ray examinations, which showed that there was at least one occasion during May 2002 on which such selection was possible, and we accept that it might happen. Nevertheless, in relation to that one occasion, the only inroad Mr Patchett-Joyce was able to make was to demonstrate that of 77 boxes, 33 might have been exchanged for better boxes. We did not hear evidence that such selection happened at all or, if it did, to what extent; nor was the possibility eliminated that some of any replacement boxes which were taken might themselves previously have been owned by Bond House.

Another of the officers' analyses was of the timings of purchase orders, of releases of goods held in secure warehouses from seller to purchaser, and of payments for the goods. This information was extracted from the date and time "stamps" applied automatically to faxes and e-mails confirming orders, from the release authorities held by Forward Logistics, also in the form of faxes and e-mails, and similar documents authorising the purchasers' bankers to release payments, or evidencing receipt. In several cases the analyses showed that, contrary to what appeared to be the industry practice, releases took place before the corresponding payment was made and it was possible to deduce that Bond House's money was being passed up the chain to pay the purchase price of the goods at earlier stages – in other words, Bond House's money was being used to finance the whole, or at least a substantial part, of the chain.

Rather graphically, it was possible to show by examination of the timings that in at least one case Fancygrove had placed an order with Bond House for a number of chips a few minutes before it instructed Forward Logistics to release the same quantity of chips to Rangepace, which – since the intervals are too short to be consistent with any other conclusion – had already arranged a sale of the same goods to SDP; and SDP, in turn, had already agreed to supply them to Bond House, which had bought them in order that it could sell them back to Fancygrove to satisfy the order Fancygrove had just placed (but without Bond House knowing, as we accept, that Fancygrove was their source). If that analysis is correct, it too is compelling evidence that there was indeed a carousel, and that Fancygrove was organising it by ensuring that all of the necessary steps were in place before setting the sequence running.

Mr Patchett-Joyce argued that the timings shown on fax and e-mail "stamps" were unreliable, since they showed no more than the time to which the sending machine was set, which might or might not be accurate. There was no certainty that the user of the machine had set the time at all, that the time, if once set, had been corrected for summer time or, in the case of Rangepace, no certainty that the time was set to British rather than continental time. Mr Peacock retorted that the timings shown on the documents were the best available evidence of the times at which the various events took place. True though that may be, it does not make the time "stamps" any more reliable and does not dispose of Mr Patchett-Joyce's objection which, in our view, has some merit. We have concluded that the timing evidence should be treated with some caution, but it cannot be disregarded if it is consistent with other available information, and it can, where appropriate, be treated as corroborative evidence.

The conclusion we have reached on this issue is that there were, almost certainly, several instances of specific circularity directly affecting Bond House. However, with the exception of the two remaining of the three occasions initially identified by Mr Lumb, the Commissioners have not in our view discharged the burden of establishing, even on the balance of probabilities, that there was specific circularity in any clearly identifiable individual case. In our view, the multiplicity of circuits and the numerous identifiable instances of splitting and combination of consignments, when added to the possibility that boxes were exchanged, must lead to the conclusion that there is no real certainty that precisely the same chips were passed from one link in any particular chain to the next, still less that they can be traced throughout a chain. There are, we accept, strong grounds for suspicion and we are persuaded that the likelihood is that in some, if not most, of the cases that is what happened; the difficulty lies in determining in which cases it did, and in which it did not.

General circularity and the "ring fence"

If the respondents' case rested on specific circularity alone, we might be persuaded that the two remaining identified occasions on which Bond House owned the same chips twice could be explained by movements in market price, or by some similar factor, which would bring their overall conclusion into question.

We do not, however, suffer from the same doubts when we consider general circularity, as the Commissioners describe it. We are quite satisfied that in each of those 25 transactions in which Bond House bought from SDP or from V J Trading and in respect of which the Commissioners have disallowed its input tax claim, such a pattern has been established and, for what it is worth (since in our view it really amounts to the same thing put differently), we also find that there was a ring fence. The evidence relating to the purchases from SDP was clearer and better developed in its presentation to us. A larger measure of inference was necessary when we came to consider the purchases from V J Trading; we were told that Mr Lumb and Miss Norris had not had sufficient time to analyse that evidence as closely. Nevertheless we have no real doubt that the same pattern of general circularity is established in both cases.

In reaching that conclusion we have been impressed by a number of factors. We found it remarkable that, in a single month, and including those transactions to which the respondents have taken no objection, Bond House sold over 800,000 CPUs which, according to the market information it adduced, represented over 10% of the worldwide demand for chips. Such high volumes might, perhaps, be capable of explanation if Bond House truly was in the position of a dealer in a market, such as a stockbroker or a dealer in a futures market who buys and sells on behalf of clients who are either investing, or disposing of investments, or speculating on market movements; or was dealing on its own account by speculating on market movements. Although Dr Lasok drew our attention to the Value Added Tax (Terminal Markets) Order 1973 (SI 1973/173) and its treatment of transactions undertaken on futures markets as being within the scope of VAT, we are not persuaded by his argument that Bond House is in an analogous position. It did not trade as an agent for others, and it did not speculate on its own account. Rather, it would (with rare exceptions) buy only when it had a purchaser, and sell only when it had secured a supply. It did not, as one might expect of a dealer trading in a permanent bear market, sell short. Nor did it buy forward, in anticipation of a rise in the market. We consider the closer analogy is with a wholesaler in a chain of supply leading from a manufacturer to a retailer.

We also find it inconsistent with legitimate trading activity that in every single case in which Bond House bought chips from SDP, it sold them to Fancygrove. Conversely, in no case in which it bought chips from V J Trading did it sell them to Fancygrove. When this factor is coupled with the evidence that, in every instance of a purchase of chips from SDP Bond House sold at a profit of £4 per chip, and in every instance of a purchase of chips from VJ Trading it sold at a profit of £3 per chip – and, moreover, in the latter case irrespective of the identify of the purchaser – it is an irresistible conclusion that the deals were orchestrated. It defies belief that it was pure coincidence that, whenever SDP was selling, Fancygrove was found to be demanding – or, put the other way round, that whenever Fancygrove wanted chips, SDP just happened to have the right number available, and was willing to sell for £4 less than Fancygrove was prepared to pay. Unless Phoenix, FTR and Eagle Mill (to which companies all of the supplies Bond House obtained from V J Trading were sold) were in reality the same organisation (and since we have no evidence on the matter we make no such finding) those points have less force when we turn to Bond House's purchases from V J Trading. Nevertheless if the transactions were entirely genuine we could expect to see, at the least, a few acquisitions of chips from V J Trading sold on to Fancygrove as well as some sales of chips sourced from SDP to customers other than Fancygrove, and would expect to find some differences in the profit margin Bond House was able to achieve.

Another (to our minds) remarkable feature is the rapid growth in the turnover of SDP, VJ Trading and (though to a lesser extent and in a rather different context) of Bond House itself. SDP was shown by the evidence to have had ambitions, and as recently as February 2002, to trade in 2000 chips per day. In May 2002 Bond House alone bought over 300,000 chips from SDP, or rather more than 14,000 for each of the 21 working days in the month. Its purchases from V J Trading were even greater: over 475,000 chips, or nearly 23,000 per day. It is a matter for no little surprise that two small companies, very recently established and, as the evidence showed, with negligible assets, should be able to obtain such large volumes of goods for onward sale to Bond House – by its own account, a well-established and well-known trader in chips. The explanation cannot be that SDP and VJ Trading were willing to pay higher prices than Bond House – on the contrary, the evidence shows that they were each making profits, albeit small, on the transactions. The irresistible inference is that they were (whether or not knowingly) being used in the perpetration of the frauds, interposed as they were between Rangepace and other missing traders and Bond House. It is less remarkable that Bond House was able to increase its own volume of trade for the very reason that it was already established in the market, and because it had recently increased its overdraft limit. Nevertheless, as we have already commented, there was no evidence before us that it was unable to enter into transactions for lack of funds before the limit was increased, nor does that factor explain why Bond House increased its trade by buying, not from other large dealers or from OEMs with surplus stock, as one might have expected from Dr Cheetham's evidence about the source of the chips in which Bond House was dealing, but from much smaller newcomers such as SDP and V J Trading.

The transactions in which Bond House bought from Larcom and TCHYP are a little different. Mr Patchett-Joyce did not direct any of his questions of Mr Lumb and Miss Norris to these transactions, nor were we addressed on them individually. Nevertheless, and particularly in the light of our other conclusions, we considered it was appropriate that we should examine the evidence available to us about them, and we have done so.

Although Bond House made other purchases from these traders in and before May 2002, we had no detailed evidence about those transactions and were unable to detect any pattern such as is evident from an analysis of the appellant's purchases from SDP and V J Trading. The notable features of the TCHYP purchase are that Bond House sold the goods to an Irish trader which featured in its sales of chips bought from V J Trading, and the Commissioners have identified a missing trader in the chain of supply to TCHYP. The fact that TCHYP made a significant profit for itself might suggest that the deals were genuine, but it is equally consistent with the possibility that TCHYP was in a similar position to Bond House, and needed to be tempted by a significant, rather than nominal, profit if it was to be persuaded to enter into the transaction at all. On the balance of probabilities we are satisfied that this chain does have the characteristics of a carousel fraud.

On the other hand, although the appellant's acquisition from Larcom has some of those characteristics – in that Larcom made a very small profit and the chips were exported by Bond House to one of the Irish traders which featured in other chains – we cannot disregard the evidence, within the e-mail we have mentioned, which undermines Mr Lumb's and Miss Norris's belief that there was a missing trader in the chain of supply. That evidence was unclear, but we consider it appropriate (particularly taking into account our conclusions about the burden of proof) that we resolve any doubt there might be in the appellant's favour. We therefore find as a fact that circularity, as it is defined by the Commissioners themselves, is not made out in this case. Consistency alone demands that the appellant's input tax claim in respect of this purchase, amounting to £96,616.80, should be allowed.

We turn last, in this section of our decision, to consider the argument that Bond House was naïve, imprudent or suspended disbelief. This contention is relevant to the appellant's arguments about its legitimate expectations, proportionality and its human rights.

Having heard Dr Cheetham for ourselves we are entirely satisfied that the Commissioners were right not to allege any dishonesty or recklessness – "Nelsonian blindness", as Mr Peacock put it – but we do consider there is copious evidence of imprudence and of the appellant's directors having failed to ask themselves obvious questions. We have already described the extraordinarily rapid growth in Bond House's trade with SDP and V J Trading and its evident failure to consider how those companies could source such large quantities of chips and finance their purchases. There was evidence that Bond House had dealt with, in particular, V J Trading despite its not meeting all Bond House's claimed criteria. It seems to us that Bond House relied far too much on the perception which Dr Cheetham had gained – although, as we have indicated, largely through wishful thinking – that the very limited checks on VAT registration details undertaken by the Commissioners and communicated to Bond House were sufficient assurance that trading with those (and other) companies was safe. The appellant had no real answer to any of these points.

Summary

We are satisfied that—

By the respondents' own criteria the evidence of carousel fraud in the chain of transactions leading to Bond House's purchase from Larcom is insufficient and its input tax claim in respect of that purchase (£96,616.80) should be allowed;

Each of the remaining 26 transactions identified by the respondents formed part of a circular series whose objective was fraudulent;

Bond House was ignorant of that fact; but

Bond House was imprudent in its dealings, and its directors failed to ask themselves obvious questions.

Did the transactions lack economic substance?

The appellant's argument

Dr Lasok's starting point was the absolute right, as he described it, to deduct the input tax which Bond House had incurred in its various purchases of chips. That right is derived initially from article 17 of the Sixth VAT Directive (77/388/EEC), which is in unqualified terms. Article 17.2 provides that

"In so far as the goods and services are used for the purposes of his taxable transactions, the taxable person shall be entitled to deduct from the tax which he is liable to pay:

(a) value added tax due or paid within the territory of the country in respect of goods or services supplied or to be supplied to him by another taxable person …."

Article 18 sets out the rules which must be followed by the taxable person seeking to effect the deduction. The principles of article 17 and the rules laid down by article 18 have been correctly transposed into the domestic legislation, appearing as ss 25(2), (3) and 26 of the Value Added Tax Act 1994, and reg 29 of the Value Added Tax Regulations 1995 respectively. These too are in unqualified terms: s 25(2), provides that the taxable person "is entitled" to deduct the input tax and s 25(3) that "the amount of the excess shall be paid to the taxable person by the Commissioners". Although some conditions are imposed, none is relevant here. The appellant's entitlement is therefore absolute, quite independent of any exercise of judgment, still less is it dependent on the Commissioners' opinion.

If (as we have recorded) Bond House had complied with the formalities imposed by article 18 and by reg 29; had acquired the chips from a taxable person; had incurred a liability to pay, and had paid, VAT on the consideration for those chips; and had disposed of them by making a taxable (even if zero-rated) supply in the course of its business, its right to deduct the input tax was established and could not be defeated, as the Commissioners maintained, because of what another trader, not connected with Bond House in relation to either its purchase or its sale, or in any other way, might or might not have done. The Commissioners' argument to the contrary, Dr Lasok said, could not withstand examination. There was no basis upon which a genuine trader could have his legitimate claim for input tax credit refused because another trader, of which Bond House knew nothing, coming before Bond House in a chain of transactions, had failed to account to the Commissioners for the output tax due from it.

There is nothing in the Sixth Directive or in the domestic legislation which supports the Commissioners' case that the activities of a fraudster, such as those they alleged to be involved in the relevant transactions, transforms otherwise legitimate business transactions into activities which are not "economic." Not only is there no legislative provision which supports that proposition, there is no decided case which could be interpreted in that way. On the contrary, the notion that the Commissioners are entitled to take a global view of a series of transactions is incapable of reconciliation with the thrust of the legislation, and with the decisions of the courts. The First VAT Directive (67/227/EEC) provides, at article 2, that

"The principle of the common system of value added tax involves the application to goods and services of a general tax on consumption exactly proportional to the price of the goods and services, whatever the number of transactions which take place in the production and distribution process before the stage at which tax is charged.

"On each transaction, value added tax, calculated on the price of the goods or services at the rate applicable, to such goods or services, shall be chargeable after deduction of the amount of value added tax borne directly by the various cost components."

In BLP Group Plc v Customs and Excise Commissioners [1995] STC 424 at 429 the Advocate General described the effect of article 2 in the following way:

"29. According to art 2(1) of the First Directive, the common system of VAT is based on the principle of the application to goods and services of a general tax on consumption exactly proportional to the price of the goods and services, whatever the number of transactions which take place in the production and distribution process before the stage at which tax is charged. In order for the 'number of transactions which take place in the production and distribution process before the stage at which tax is charged' not to influence the amount of VAT ultimately due to the revenue authorities, art 2(2) of the First Directive introduces the mechanism of deduction of input tax.

"30. A consideration of those provisions together shows that the Community legislature, proceeding from an ideal image of 'chains of transactions' — to adopt the neat phrase used at the hearing by the representative of the United Kingdom —intended to attach to each transaction only so much VAT liability as corresponds to the added value accruing in that transaction, so that there is to be deducted from the total amount the tax which has been occasioned by the preceding 'link in the chain' … ".

The issue in BLP Group was whether the taxpayer was entitled to recover the VAT charged on various professional fees it had paid in connection with a share disposal which it had undertaken in order to reduce the level of debt it had incurred in carrying on its ordinary business activity of making taxable supplies. The taxpayer's case was that if the debts were incurred in the course of carrying on its business, then fees incurred in taking measures to reduce those debts were "for the purposes of [its] taxable transactions" and within article 17. The Commissioners argued that the fees were incurred in making an exempt supply of shares, and that one could not go behind that fact in order to re-characterise the supplies. The Advocate General's analysis of article 2 made it clear that each transaction in the chain must be examined individually, and by its own objective characteristics; the taxpayer's intentions were immaterial. This theme was taken up by the Court which (at p 437) said:

"24. … if BLP's interpretation were accepted, the authorities, when confronted with supplies which, as in the present case, are not objectively linked to taxable transactions, would have to carry out inquiries to determine the intention of the taxable person. Such an obligation would be contrary to the VAT system's objectives of ensuring legal certainty and facilitating application of the tax by having regard, save in exceptional cases, to the objective character of the transaction in question."

The need to consider transactions individually and objectively was amply demonstrated by the House of Lords' decision in Commissioners of Customs and Excise v Robert Gordon's College [1995] STC 1093 in which, at p 1100, Lord Hoffmann, giving the only substantive speech and after referring to BLP Group said " … that decision makes it clear that for the purposes of European VAT legislation, it is not permissible to take a global view of the series of transactions in the chain of supply."

The "objective character" of Bond House's transactions could not, indeed was not, in dispute. The Commissioners accepted that the goods bought and sold by Bond House existed, that they had been purchased and sold in the course of Bond House's business, that the purchases and sales were made from and to unconnected persons for profit, and that in each case "the transfer of the right to dispose of tangible property as owner" – the definition, by article 5.1 of the Sixth Directive, of a "supply of goods" – had taken place. There was no suggestion, nor could there be, that Bond House's transactions themselves were to be disregarded as supplies because they were artificial in the manner which the Commissioners had alleged in Halifax plc v Commissioners of Customs and Excise [2002] STC 402, that is, that they were transactions entered into for the sole purpose of avoiding or mitigating Bond House's liability for VAT. That others, elsewhere in the chain of supply, might have had improper motives was an immaterial consideration.

If the Commissioners considered that other traders were evading their tax liabilities their proper remedy was obvious: they should pursue those other traders. The care and management of VAT are their responsibility and a range of sanctions was open to them: see VAT Act ss 58, 60, 61 and Sch 11. It was not for traders such as Bond House to undertake the Commissioners' responsibilities for them, and it was not open to the Commissioners to argue that because they had not received VAT properly due from one trader, they could recoup the loss by depriving another trader – not even adjacent in the chain of transactions – of credit for the input tax it had incurred.

The Commissioners' argument that what Bond House had paid and which was described on its suppliers' invoices as VAT was in fact not VAT at all could not be sustained. Bond House had bought chips from another trader registered for VAT in the United Kingdom. Its purchase was indisputably a taxable transaction. The chips had been exported to a trader registered in another Member State, again indisputably a taxable transaction, even if zero-rated. Both the purchase and the sale were carried out in the course of Bond House's and its suppliers' businesses, and were therefore in each case supplies made by a taxable person acting as such. Their character could not be transformed by the activities of other traders, remote from the two transactions in which Bond House had been engaged. Such a conclusion was incapable of reconciliation with the decision of the High Court in Halifax plc [2002] STC 402 where it was determined that the reasons why each participant in a scheme had become a participant must be considered; it was not sufficient to consider only the motives of the "prime mover". That proposition was consistent only with the notion that Bond House's transactions, and its purpose in entering into them, must be viewed in isolation, without regard to what had gone before, in order to determine their character.

Even if there were an analogy to be drawn with Halifax plc, as the tribunal had viewed the matter (see [2001] V & DR 73), the impugning of some transactions did not have the result that the whole chain was regarded, for VAT purposes, as a nullity. There remained at least one supply: see Interleasing Ltd and others v Commissioners of Customs & Excise (2002, Decision No 17819, especially paragraph 64). In Halifax plc the tribunal decided that transactions which involved two companies interposed between the arm's length trader which made the supply and its ultimate recipient should be disregarded, but the result was not that there was found to be no supply at all – the supply remained, but was found to have been made by the arm's length trader directly to the ultimate recipient.

The High Court had questioned that outcome because the tribunal's approach ignored the subjective purpose of the interposed companies; and the tribunal had itself elsewhere concluded that "when determining whether a person is carrying on an economic activity his purpose cannot be ignored": see BUPA Hospitals Ltd and another v Commissioners of Customs & Excise (2002, Decision No 17588 at paragraph 73). Bond House's purpose was manifestly to do as it had done for the previous ten years – to buy and sell computer chips by way of trade. It was, Dr Lasok maintained, absurd to regard that long-standing trade as one devoid of economic substance. It was the essence of Bond House's business to buy and sell chips, from and to unconnected third parties, with a view to making a profit, and that was what it had done in the 27 instances which the Commissioners had selected for the disallowance of the input tax it had incurred. Their approach could not be reconciled with what the tribunal said in RBS Property Developments Ltd v Commissioners of Customs & Excise (2002, Decision No 17789):

"Unless therefore the matter can be described as wholly artificial in the sense of being a device for avoiding tax which had no reality then the Commissioners are not entitled to ignore what did in fact happen."

If the Commissioners were correct, he said, the result would be absurd. A trader could never know whether his supplier might decide to suppress the sale from his next VAT return. The supplier might change his mind, deciding to suppress the sale and then having second thoughts; and even then he might decide after all to conceal the sale. It could not possibly be the intention of the Sixth Directive that a payment made by a purchaser to his supplier should change its character on the whim of the supplier – and even more so when the supplier concerned was remote from the trader claiming the deduction, was unknown to him, and when his intentions could not possibly be guessed. This very absurdity demonstrated that the Commissioners could not be right.

The respondents' arguments

Mr Peacock's primary argument depended upon the interpretation of the Sixth Directive, and particularly articles 2, 4, 5, 6 and 28a, and upon our finding as a fact that the relevant transactions – that is, the chains of transactions, and not merely those to which Bond House was a party – were designed not to buy and sell CPUs as a normal trading activity but to operate as the vehicle for a fraud. It was immaterial that Bond House was ignorant of that purpose; all that mattered was the objectively determined character of the series.

We confine ourselves below to the provisions of the Sixth Directive since it was not disputed that the UK domestic legislation correctly implemented them. Article 2 reads:

"The following shall be subject to value added tax:

(a) the supply of goods or services effected for consideration within the territory by a taxable person acting as such;

(a) the importation of goods."

A "taxable person" is defined by article 4.1 as a person who carries out "any economic activity", the latter being defined by article 4.2 to "comprise all activities of producers, traders and persons supplying services including mining and agricultural activities and activities of the professions". Wide though that definition is, it seems to us that Mr Peacock is right in his argument that it is intended to encompass supplies ordinarily made by those engaged in legitimate business activities. However, whether or not the activity is carried on with a view to making a profit is not decisive since (in order to avoid distortions of competition) activities of the same kind carried on without the intention of making a profit – by public bodies, for example – are not excluded: see article 4.5.

Mr Peacock recognised that article 4.2 demanded a broad interpretation: see Van Tiem v Staatsecretaris van Financiën (Case C-186/89) [1990] ECR I-4363 at 4386 where the Court of Justice said "… it should … be underlined that article 4 of the Sixth Directive confers a very wide scope on value added tax, comprising all stages of production, distribution and the provision of services …"; and that exceptions were to be construed narrowly. It was entirely possible that a taxable person might make some supplies which were within, and some which were not within, the scope of VAT. The phrase "acting as such", used in articles 2.1 and 28a.1(a) exclude from the scope of value added tax those transactions in which a registered trader engages which are not carried out in the course of his business. Thus (to use Mr Peacock's own example) the sale by a taxable person of an article of furniture used exclusively in his home is not a supply made by him as a taxable person acting as such, but in a private, or personal, capacity. Such a sale could not constitute an economic activity within the meaning of the directive and is outside the scope of VAT. See, too, Polysar Investments Netherlands BV v Inspecteur der Invoerrechten en Accijnzen, Arnhem (Case C-60/90) [1993] STC 22 at 238, paragraph 11: "It follows from art 2 of the Sixth Directive, which defines the scope of VAT, that … only activities of an economic nature are subject to VAT." By parity of reasoning, any supply which is made, not in the ordinary course of business but in order to perpetrate a fraud, is outside the scope of VAT.

Article 5.1 provides that a "'Supply of goods' shall mean the transfer of the right to dispose of tangible property as owner." Mr Peacock acknowledged that title, and with it the right to dispose of the goods, had passed and that supplies of goods within this definition had been made.

Superficially, therefore, even if not in substance, Bond House's purchases fell within article 2.1. Similarly its disposals to registered traders in Ireland came within article 28a1:

"The following shall also be subject to value added tax:

(a) inter-Community acquisitions of goods for consideration within the territory of the country by a taxable person acting as such . . ."

We do not need to dwell on the detailed mechanics of inter-Community trade; it is sufficient to mention that Mr Peacock relied on the use in each of those provisions of the phrase "a taxable person acting as such", and on the use of the phrase "economic activity" in the definition of a taxable person.

It is well established, Mr Peacock continued, that a purposive interpretation is to be placed on EC Directives – a proposition which we take as axiomatic. The purposes of the Sixth Directive can be derived from its recitals and from those of the First Council VAT Directive (67/227/EEC) on which it built, and they are to remove factors which distort or hinder trade and competition. It was an obvious inference that both Directives were concerned and (he emphasised) concerned only with real economic activity – that is, in respect of goods, trade in the generally understood sense of that word. Correspondingly the Sixth Directive had no application to transactions which were not effected in the course of trade, as so understood. The purpose of the Directives was to promote free trade, and not fraud, which was quite alien to its scope.

Similarly, in Mol v Inspecteur der Invoerrechten en Accijnzen (Case 269/86) [1988] ECR 3627 at 3650, paragraph 17, the Court of Justice determined that unlawful supplies of narcotic drugs (that is, supplies other than those made for strict medical or scientific purposes), being illegal in all Member States, "are wholly alien to the provision of the Sixth Directive". Excluding such supplies from the scope of VAT could not distort competition as all such supplies were prohibited. The Court reached the same conclusion in Witzemann v Hauptzollamt München-Mitte (Case C-343/89) [1993] STC 108, in which the relevant event was the importation of counterfeit currency.

However, in Fischer v Finanzamt Donaueschingen (Case C-283/95) [1998] STC 708 the taxpayer carried on an activity in an unlawful manner by promoting games of chance without a permit; he was in competition with others who had permits enabling them to promote such games lawfully. The Court concluded that activities carried on unlawfully and in competition with similar activities carried on lawfully were within the scope of the Sixth Directive, and there could be no discrimination between them by imposing tax on the latter while exempting the former. In other words, if an activity which is capable of being carried on lawfully is carried on unlawfully, it bears the same tax consequences as would apply if it were carried on lawfully; but if it is impossible to carry on the activity lawfully, it is outside the scope of value added tax. Fraud, by definition, cannot be carried on lawfully.

From these propositions it followed, Mr Peacock argued, that transactions carried out for the purpose of perpetrating a fraud could not come within the scope of the Sixth Directive and correspondingly were not subject to VAT. Since the 27 transactions identified by the Commissioners (or, as we have found, 26) in which the appellant had engaged were (whatever the appellant may have believed) instrumental in the perpetration of a fraud they were not within the scope of the Sixth Directive. What the appellant thought it was paying to its supplier as VAT was, consequently, not VAT at all, and therefore could not give rise to an input tax credit. The appellant was not, in the case of these transactions, a taxable person acting as such, but a mere instrument used to perpetrate the fraud – the "pawn caught up in a world of mire" to adopt Dr Cheetham's own phrase.

Mr Peacock argued the same point in a rather different way by the analogy of a road journey which included a roundabout. The journey along the road leading into the roundabout, the passage around the roundabout to the appropriate exit road and the journey along the exit road to the intended destination all contributed to the traveller's moving from the start to the finish of the journey. If, however, he circled the roundabout several times, his doing so made no such contribution. Likewise, supplies by the manufacturer of chips, whether or not through other traders, to the entry point of a carousel, any supplies necessary to get the chips from that entry point to an appropriate exit point and any further supplies necessary to allow the chips to reach an end consumer constituted economic activities. But transactions which merely caused the chips to go round and round the carousel made no contribution to the chain of supply, properly so called, and as it was understood for the purposes of the Sixth Directive, and did not constitute an economic activity. Since that was the character of the appellant's relevant transactions, and they could therefore not rank as economic activities, what had purportedly been charged as VAT was not VAT and Bond House was not entitled to credit for it.

That Bond House's belief about the nature of the transactions was immaterial was demonstrated by Genius Holdings BV v Staatsecretaris van Financiën (Case 342/87) [1991] STC 239. The taxpayer sought to recover as input tax sums invoiced to it, as VAT, by sub-contractors it had engaged. Dutch law provided that the sub-contractors should not have claimed VAT on their charges, and the tax was therefore not legally due. The Court of Justice determined that the right to deduct extended only to tax actually due and not to tax which, however innocently, had been incorrectly included on an otherwise valid invoice and paid by the recipient of the invoice. The same principle applied if the taxpayer, innocently but mistakenly, thought he was engaging in an economic activity: see Wellcome Trust Ltd v Customs and Excise Commissioners (Case C-155/94) [1996] STC 945. That, said Mr Peacock, was the appellant's position here: however innocent, it could not claim to deduct that which was not truly VAT.

There is, Mr Peacock continued, a line of authority which demonstrates that, in determining whether a supply of goods (or services) has been made regard must be had not only to the legal analysis of the contract between the parties – which might even be misleading – but to the "commercial reality of the situation", the phrase used by Advocate General Jacobs in H J Glawe Spiel- und Unterhaltungsgeräte Aufstellungsgesellschaft mbH KG v Finanzamt Hamburg (Case C-38/93) [1994] STC 543 at 547, paragraph 18. The relevance of the commercial reality was shown not only by that case, where it was held that only the proportion of the stakes paid by players of gaming machines which could reach the owner of the machine, rather than the entire stakes, were taxable, but by Customs and Excise Commissioners v First National Bank of Chicago (Case C-172/96) [1998] STC 850, where the court concluded that a supply (of foreign currency) made free of commission was, as a matter of economic reality, made for consideration; and, in the domestic case-law, Customs and Excise Commissioners v Reed Personnel Services Ltd [1995] STC 588 in which, at p 595, Laws J said "… the concept of supply for the purposes of VAT is not identical with that of contractual obligation". Indeed, in Customs and Excise Commissioners v DFDS A/S (Case C-260/95) [1997] ECR I-1005 the Court of Justice went even further by concluding that economic reality must take precedence over legal form when there would otherwise be distortion of competition.

That one must look at the economic reality was, Mr Peacock maintained, consistent with the European jurisprudence in other areas. The use, or abuse, of legal form in order to obtain a benefit was disapproved by the Court of Justice in such cases as Leclerc and others v Sàrl "Au blé vert" and others (Case 229/83) [1985] ECR 1 (a device to circumvent the French net book agreement) and, more recently, Emsland-Stärke GmbH v Hauptzollamt Hamburg (Case C-110/99) [2000] ECR I-11569, in which goods were exported from Germany to Switzerland, satisfying the formal requirements for the payment of export refund, and then immediately transported back to Germany. If one accepted the proposition that only the objectively determined character of the transactions was relevant, and that Bond House's belief or intention was immaterial, it must follow from the application of the principles to be drawn from those cases to the facts of this that the transactions were alien to the Sixth Directive and correspondingly outside the scope of VAT.

Discussion

In our view, neither BLP nor Robert Gordon's College supports the argument that Bond House's dealings must be considered in isolation, without regard to transactions which precede or follow them. The issue in those cases was whether one transaction was "coloured" by borrowing, or transferring, the character of another. The attempt to do so was, as we see the matter, contrary to the principle that the "commercial reality" of the transactions should be identified. In BLP, it could not realistically be said that the shares were issued in the course of the taxpayer's normal business activity: the argument that the share issue was designed to reduce its indebtedness could not transform an otherwise exempt supply by association. Conversely, in Robert Gordon's College, the Commissioners attempted to introduce extraneous, even hypothetical, matters into the analysis of the supply but were not permitted to do so where the economic reality of the supply could be determined and its tax status could be ascertained from that economic reality alone.

We agree, therefore, with Mr Peacock that it is permissible to examine the whole chain of supply in order to determine the character of the appellant's dealings: we accept his submission that that is the conclusion to be drawn from such cases as, in particular, DFDS and Emsland-Stärke. It is also a conclusion consistent with recent decisions of this tribunal, such as Interleasing, which we have already mentioned, and Blackqueen v Commissioners of Customs and Excise (2002, Decision No 17680), in which the very fact that a series of transactions was circular was critical.

Even so, it seems at first sight startling that a trader which has carried out transactions of the type normally carried out by it over several years should find that those transactions are impugned, not because of their inherent unlawfulness – the illegality here is not dealing in chips, but failing to pay VAT, or abusing the system – but because of the activities of others to which it has no direct connection. In DFDS, Emsland-Stärke, Interleasing and Blackqueen the various participants were all associated in one way or another, and the association was an important factor, to put it no higher. However, we have come to the conclusion that Mr Peacock is right that an objective approach to the series of transactions, without regard to the appellant's belief or intentions, is correct and that his analogy of a road journey including a roundabout is a useful means of illustrating the concept.

It seems to us clear that the Sixth Directive is concerned with transactions which lead from producer to consumer. Its entire thrust is that the tax on consumption should be borne entirely by the final consumer and that, by means of the right to deduct, those who deal with goods and services by producing or distributing them are in a fiscally neutral position. Sometimes producer and consumer are in direct contact, as might usually be the case in the provision of professional services, and in others remote, where goods mass-produced in one Member State are consumed singly by customers in another; and the chain of supply may be disrupted or frustrated, for example by accidental destruction of the goods. Nevertheless, even in those cases there is a distribution chain of the kind envisaged by the directive. So much is, we think, evident from the passage from the recitals to the First Directive (whose theme is continued by the Sixth Directive) ("production and distribution process") and from remarks such as that made by the Court of Justice in Van Tiem ("all stages of production, distribution"), both of which we have already set out more fully. Thus, since they find no place in the directives, transactions which have some other purpose are not within their contemplation.

Once it is found – as we have done in 26 of the 27 cases identified by the Commissioners – that Bond House's transactions did not contribute in any proper sense to the process of production or distribution, but merely played a part in the circulation of chips round a circle which had nothing to do with their distribution to a final consumer, just as repeatedly circumnavigating a roundabout does not contribute to the traveller's journey, it seems to us an inevitable conclusion that they were not transactions within the contemplation of the Sixth Directive; they were not "economic activities". It is certainly unfortunate for it that Bond House has been unwittingly used as the tool of others, but we are satisfied that Genius Holdings and Wellcome Trust cannot be meaningfully distinguished and that they support Mr Peacock's argument that its ignorance is not a material factor. Moreover, one has only to postulate the proposition that the tax liability of a transaction is dependent on the parties' belief to conclude that the proposition is manifestly wrong.

It is not, in our view, open to the appellant to argue that the analysis, by the present chairman, in Interleasing, of the tribunal's decision in Halifax plc assists it. The essence of that analysis is that where artificial transactions, such as those found by the tribunal to have been undertaken by The Halifax's subsidiaries, are interposed between the arm's length supplier and the ultimate recipient, the artificial transactions are "collapsed" so as to leave the "true" supply remaining. If that exercise were undertaken here, by eliminating or collapsing the fraudulent transactions in the circle, the logical outcome could only be that the appellant itself made an entirely circular supply, buying from and selling to itself if all the others were party to the fraud, but still within a circle even if some others were to be regarded as innocent: there might, for example, be a series from SDP to Bond House, back to SDP. It cannot, in our view, be said that the "economic reality" of Bond House's transactions, even on this hypothesis, contributed to a distribution chain – they only ever formed part of a fraudulent series.

For the same reason we do not accept Dr Lasok's argument that what the Advocate General said in Polysar ([1993] STC 222 at 234, paragraph 9) – "…in order to establish whether there is liability to tax, it is necessary to focus on the activities of each legal person separately, and not on the activities of the concern as a whole" – helps his case. Even if the observation is right (it was not necessary for the Court's decision, and is not easy to reconcile with some of the other cases) it is impossible to consider transactions such as those entered into by the appellant, for the purchase and sale of goods, as if they took place in a vacuum. Any such consideration must be undertaken against the background of the Sixth Directive and the production and distribution chain which it envisages. A purchase and a sale by themselves are, in conceptual terms, an abstraction; they must of necessity be set in their context if their nature and effect are to be determined. One is inevitably driven back to the same conclusion, that they were not economic activities.

Lastly, we are unimpressed with Dr Lasok's argument that the prior trader might change his mind about whether or not to account for the output tax. The need to establish both that and circularity, which the Commissioners accept, seems to us to make that argument irrelevant to the matters we are required to consider. Dr Lasok's argument might be appropriate in the case of linear transactions but not, we think, where there is circularity. Moreover, the whole theory of the carousel fraud presupposes that the missing trader has sold on at a notional loss; his doing so is quite incompatible with the possibility that he will after all account for the tax.

Subject to the subsidiary arguments to which we now turn, we conclude that, as the objectively determined characteristics of the 26 purchases show that they were entered into, not in the course of distributing goods to a final consumer, but with a view to obtaining an advantage by fraud, they cannot be regarded as economic activities within the meaning of the Sixth Directive, and are correspondingly outside the scope of VAT. We agree with the Commissioners that the sums claimed by Bond House as input tax, in respect of those 26 transactions, are not VAT.

Proportionality, human rights and legal certainty

We intend to deal with these topics together, as most of the argument we heard about them overlapped to a considerable degree.

Dr Lasok relied upon the fact that the Commissioners are prosecuting some of those involved in the series of transactions in which Bond House participated for offences involving VAT, and upon their treatment of others in the chains, such as SDP and V J Trading, in the normal way – that is, by allowing credit for input tax against the output tax for which they were accounting. Their treatment of Bond House was, he said, inconsistent with that approach, discriminatory, and offensive to article 14 of the Human Rights Convention; the appellant was, but others were not, being deprived of its rights under article 1 of the First Protocol and, even if that could be justified at all (and he was also to argue that it could not) it was certainly not justified if it was done in a discriminatory fashion. Likewise the fact that Bond House had been deprived of its input tax in some cases but not in other, apparently identical, cases was arbitrary and unjustified.

Mr Peacock's response was that, although the prosecutions had been launched in terms which suggested a VAT fraud, the indictments had not yet been framed and it could not be said that the Commissioners were arguing in one way in one tribunal and in a different, contradictory, way in another; the position in that regard was unclear. He acknowledged that other traders had been treated as Dr Lasok had contended, but said that the Commissioners were in the process of reconsidering the matter and would revisit those traders' VAT returns in due course; they were still in time to do so and would also want to consider this decision and its impact on the correct tax treatment of traders in such chains. It was, he said, reasonable that they should be given that opportunity, and wrong to criticise them at this stage. So far as there was apparent inconsistency in the Commissioners' treatment of the appellant's own transactions, he relied on Mr Lumb's evidence that the benefit of the doubt had been given where the Commissioners accepted they could not show circularity.

Article 1 of the First Protocol protects a person's right to peaceful enjoyment of his possessions. Those possessions, Dr Lasok argued, include the right (as article 18.4 of the Sixth Directive makes clear that it is) of a trader to an input tax credit. He acknowledged that the proviso to article 1 allowed the State to deprive a person of his possessions if that was necessary to secure the payment of taxes. However, what was at issue here was not the payment of tax but the trader's right to receive a repayment. Moreover, any measure which a State adopted must be proportionate to the objective sought to be achieved. On the latter point he relied heavily on Garage Molenheide BVBA and others v Belgium [1998] STC 126.

There, the Belgian tax authorities withheld input tax credit claims in a number of cases where they suspected that the claimants had been engaged in tax evasion. The Court of Justice determined that article 18.4 of the Sixth Directive did not preclude the taking of such measures, but that they should be taken in a manner which was proportionate, was the least detrimental to the objectives of the VAT directives, and did not systematically undermine the right to deduct input tax; the measures must go no further than was necessary to achieve their objective. Put another way, as in James v United Kingdom (1986) 8 ECHR 123 at 144, "there must be a reasonable relationship of proportionality between the means employed and the aim sought to be realised".

Here, the aim was to secure the payment of the output tax due from traders who had not accounted for it. It could not be proportionate to that aim to penalise legitimate, honest traders such as Bond House for no better reason than that they had bought and sold goods which, unknown to them, had previously been dealt in by a trader which decided not to account for the output tax due from it. Traders in Bond House's position could never know whether any goods they bought might, at some time in the past have been through the hands of such a trader. The appellant was entitled to the certainty that transactions undertaken by it in the ordinary course of its normal, lawful business would not be attacked for reasons of which it had no knowledge or for circumstances over which it had no control, and any measures which were taken which had that effect must be disproportionate.

That was particularly so in a case where, as here, the trader concerned had cooperated with the tax authorities, had done everything required of it and more, had taken every reasonable precaution to ensure that it entered only into legitimate transactions and had been led to believe by the Commissioners that, provided it continued to conduct its business in that way, had nothing to fear.

Mr Peacock's response was the simple one that the Human Rights Convention, and the similar principles developed in the jurisprudence of the Court of Justice, were designed to protect rights; and if a person did not have relevant rights, neither the Convention nor that jurisprudence could be prayed in aid. In any event, the appellant had been imprudent in its dealings, and had failed to take heed sufficiently of the adverse indications of its own enquiries; and the suggestion that the Commissioners had led it to believe that it was immune from investigation or other action was not borne out by the evidence.

In our view Mr Peacock's first argument, short and simple as it is, effectively disposes of Dr Lasok's contentions. It cannot have been the intention of the signatory states that the Human Rights Convention should be available to protect perceived, rather than genuine rights; and just as a trader's belief cannot affect the correct tax treatment of his transactions, so a person's mistaken belief that he has a right cannot create that right for him. We likewise accept that it is unrealistic to argue that because the Commissioners have treated one trader in a particular fashion then, even at a stage when they are considering their position and are not out of time to change their treatment, they are bound to treat all similar traders in the same (benevolent) way. We also take the view that it is, and cannot be, disproportionate, to decline to pay to a trader money which is not, as a matter of fact and law, due to him.

Accordingly we reject the contention that the respondents' actions in this case have offended the appellant's human rights, or the principles of proportionality or of legal certainty.

Conclusion

Our conclusions are these—

By the respondents' own criteria the evidence of carousel fraud in the chain of transactions leading to Bond House's purchase from Larcom is insufficient and its input tax claim in respect of that purchase (£96,616.80) should be allowed;

Each of the remaining 26 transactions identified by the respondents formed part of a series whose objective was fraudulent;

Notwithstanding Bond House's ignorance of that objective, and its innocence of any wrongdoing, its relevant transactions were devoid of economic substance;

The criteria by which such transactions are to be judged are wholly objective, thus Bond House's innocence of wrongdoing is immaterial;

Bond House could have no legitimate expectation that its input tax claims would be met;

The respondents' depriving Bond House of its input tax credit does not offended the principles of proportionality or of legal certainty, nor any of the appellant's human rights; thus

The respondents' refusal of Bond House's claim for input tax credit (save for the one exception we have mentioned) is justified.

With the limited exception we have identified, which results in a reduction of the assessment from £5,267,799.46 to £5,171,182.66, we determine this appeal in the respondents' favour.

Mr Peacock sought a direction in the respondents' favour for the costs of the appeal. Although we have directed the reduction of the assessment, we have done so to a very modest extent and for a reason which was not fully developed before us. In every other respect we have determined the appeal in the respondents' favour. The request was not opposed in principle. We consider it is appropriate that we make a direction in the respondents' favour and accordingly do so; if the amount cannot be agreed we direct that the costs shall be subject to detailed assessment, on the basis applied in the High Court, by a Master or District Judge, on the application of either party to be made within three months of the release of this decision.

COLIN BISHOPP

CHAIRMAN

RELEASE DATE: 8 May 2003

MAN/02/534

BAILII: Copyright Policy | Disclaimers | Privacy Policy | Feedback | Donate to BAILII
URL: http://www.bailii.org/uk/cases/UKVAT/2003/V18100.html

	[Home] [Databases] [World Law] [Multidatabase Search] [Help] [Feedback]
	United Kingdom VAT & Duties Tribunals Decisions
You are here: BAILII >> Databases >> United Kingdom VAT & Duties Tribunals Decisions >> Bond House Systems Ltd v Customs And Excise [2003] UKVAT V18100 (08 May 2003) URL: http://www.bailii.org/uk/cases/UKVAT/2003/V18100.html Cite as: [2003] UKVAT V18100