Skip to Main Content
Alpha

Help us to improve this service by completing our feedback survey (opens in new tab).

HM Revenue and Customs v Citibank NA & Anor

[2017] EWCA Civ 1416

Case No: A3/2016/2015 &

A3/2016/2017
Neutral Citation Number: [2017] EWCA Civ 1416
IN THE COURT OF APPEAL (CIVIL DIVISION)

ON APPEAL FROM THE UPPER TRIBUNAL (TAX AND CHANCERY CHAMBER)

MR JUSTICE NUGEE AND JUDGE TIMOTHY HERRINGTON

[2016] UKUT 0123 (TCC) (10 TH MARCH 2016)

ON APPEAL FROM THE FIRST-TIER TRIBUNAL (TAX CHAMBER)

JUDGE BARBARA MOSEDALE [2014] UKFTT 1063 (TC) (28 TH NOVEMBER 2014)(Citibank)

AND JUDGE JOHN WALTERS QC [2014] UKFTT 921 (TC) (24 TH SEPTEMBER 2014)(E Buyer)

Royal Courts of Justice

Strand, London, WC2A 2LL

Date: 29 September 2017

Before:

SIR TERENCE ETHERTON, MASTER OF THE ROLLS

SIR GEOFFREY VOS, CHANCELLOR OF THE HIGH COURT

and

LADY JUSTICE HALLETT D.B.E

Between:

THE COMMISSIONERS FOR HER MAJESTY’S REVENUE AND CUSTOMS

Claimants/

Respondents

- and -

CITIBANK NA

and

E BUYER UK LIMITED

Defendants/Appellants

Mr Jonathan Kinnear QC, Mr James Puzey and Mr Howard Watkinson (instructed by the Solicitor to HMRC) for the Appellants

Mr Michael Conlon QC and Mr David Scorey QC (instructed by Hogan Lovells International) for Citibank NA

Mr John Wardell QC and Mr David Scorey QC (instructed by Banks Kelly) for E Buyer

Hearing date: 18th July 2017

Judgment

Sir Geoffrey Vos, Chancellor of the High Court:

Introduction

1.

This appeal raises a short point that may have important consequences for the way in which the Commissioners for Her Majesty’s Revenue and Customs (“HMRC”) can challenge a taxpayer’s right to deduct input VAT on the basis that the taxpayer knew, or should have known, that its transactions were connected with the fraudulent evasion of VAT as part of a Missing Trader Intra-Community fraud (an “MTIC fraud”).

2.

The parties do not, however, entirely agree on the short point that the court has to consider. HMRC contend that the question is whether the Upper Tribunal (“UT”) was wrong to conclude that an allegation that a taxpayer knew that its transactions were part of an orchestrated scheme to defraud HMRC required HMRC to plead and particularise, and therefore prove, an allegation of dishonesty. The taxpayers submit in response that the UT understood full well that, as a matter of law, an allegation that a taxpayer knew that its transactions were part of an orchestrated scheme to defraud HMRC might not involve an allegation of dishonesty or fraud (which was common ground). As a result, the taxpayers contend that the question for this court is only whether in the circumstances of the pleadings in these cases, the UT was right to say that what HMRC had actually pleaded as to knowledge amounted to dishonesty and had to be particularised as such. The taxpayers say that they are entitled, as a matter of good procedural case management, to have the serious allegations that HMRC has made fully and properly particularised.

3.

To untangle this basic dispute, it will be necessary to look in rather more detail than might otherwise have been necessary at the decisions of the two First-tier Tribunals (the “FTT”) in these cases, and at the decision of the UT.

4.

The UT upheld the decision of the FTT (Tribunal Judge Barbara Mosedale) in the case against Citibank NA (“Citibank”), but reversed the decision of the FTT (Tribunal Judge John Walters QC) in the case against E Buyer UK Limited (“E Buyer”). HMRC are appealing the UT’s decision in both cases. In E Buyer’s case, HMRC also appeals the UT’s decision to overrule Judge Walters’ decision to refuse E Buyer’s application for standard disclosure by list of the kind envisaged by the Civil Procedure Rules (“CPR”) in High Court litigation in respect of the knowledge allegation.

5.

In both cases, however, HMRC’s main ground of appeal is as follows:-

“The [UT] erred in law in concluding that the allegation that [Citibank and E Buyer] knew both that there was an overall scheme, involving an orchestrated and contrived series of transactions, to defraud [HMRC], and that its transactions were part of this scheme, required a plea of dishonesty to be made as part of [HMRC’s] Kittel case, because the Kittel test contains no element of dishonesty”.

6.

The Kittel test arises from the decision of the Court of Justice of the European Union (“CJEU”) in the case of Axel Kittel v. Belgium; Belgium v. Recolta (C-439/04 and C-440/04), [2008] STC 1537 (the “Kittel” case). It is not in dispute. It provides that a trader will not be able to reclaim input VAT if it knew or should have known that the transaction in which it was involved was connected with a scheme for the fraudulent evasion of VAT. Demonstrating actual knowledge has been referred to as the first limb of the Kittel test, whilst demonstrating that the taxpayer should have known has been referred to as the second limb of the Kittel test. HMRC wish to rely on both limbs in each of these cases and are not prepared to give up their reliance on the first limb of the Kittel test.

7.

HMRC say, in essence, that they are not necessarily alleging fraud or dishonesty merely by contending that the trader knew that its transaction was connected with a scheme for the fraudulent evasion of VAT. The taxpayers submit that, whilst it would be possible to allege that the first limb of the Kittel test was satisfied without alleging fraud, in these cases HMRC are doing so because, despite the fact that HMRC have disavowed reliance on the taxpayers being either dishonest or fraudulent, HMRC’s pleadings can only sensibly be construed as making such allegations.

8.

I will start with a brief resume of the essential factual background in each case.

Factual background

9.

E Buyer is a large online retailer of electronics, with an annual turnover of approximately £200 million. The appeal concerns 289 wholesale export transactions between June 2010 and September 2011 in connection with 37 defaulting traders and three so-called “contra-traders”. HMRC refused E Buyer’s claim to deduct input tax by various decisions notified in letters dated between 31st January 2013 and 29th April 2014 in an assessed sum of £6,740,113.23. E Buyer lodged Notices of Appeal in respect of each decision with the FTT. HMRC filed their Statement of Case on 12th February 2014 alleging that E Buyer knew or should have known that the transactions were connected with VAT fraud. On 21st March 2014, HMRC applied for a direction that Rule 27(2) of the Tribunal Procedure (First Tier Tribunal) (Tax Chamber) Rules 2009 (the “2009 Rules”) be dispensed with, and sought a direction that the exhibits to the witness statements should stand as the documents on which the parties intended to rely. Rule 27(2) provides that “[s]ubject to any direction to the contrary, within 42 days after the date [HMRC] sent the statement of case … each party must send or deliver to [the FTT] a list of documents … which the party providing the list intends to rely upon …”. E Buyer objected to this application on 2nd June 2014 and made its own application for HMRC’s Statement of Case to be struck out, alternatively for a direction that HMRC should provide E Buyer further and better particulars in response to its 31-page request, and also for a direction that both parties should provide disclosure by list of all documents relevant to the appeal (which was, in effect, standard disclosure of the kind envisaged by the CPR).

10.

After a hearing on 15th and 16th September 2014, Judge Walters QC directed on 23rd September 2014 that HMRC’s application should be allowed and that E Buyer’s applications should be dismissed.

11.

The relevant Citibank transactions were the purchase and sale of EU Emissions Allowances (“EUAs”) or carbon credits in September 2009. EUAs were at the time a new type of financial derivative treated for VAT purposes as a supply of services. On 27th August 2013, HMRC assessed Citibank for VAT that it had reclaimed in the total sum of £10,081,126. A review of the relevant decisions dated 17th December 2013 upheld the assessment, however a subsequent substantive review dated 27th February 2014 reduced the assessment to the current sum claimed of £9,893,821. Citibank filed its Notice of Appeal in the FTT on 16th January 2014 on the basis that the assessment failed to identify the alleged fraud or to make good the assertion that Citibank knew or ought to have known of the fraud and therefore failed to displace Citibank’s prima facie right to recover the input tax. HMRC’s Statement of Case dated 22nd April 2014 alleged that the bank knew or ought to have known that the transactions were connected with VAT fraud. On 23rd May 2014 Citibank issued an application for directions seeking answers to its request for further and better particulars of the Statement of Case. The matter was heard by Judge Mosedale on 3rd November 2014, who decided on 24th November 2014 that HMRC’s Statement of Case alleged dishonesty against Citibank, and that this should have been expressly pleaded. She therefore directed that HMRC should amend their Statement of Claim, which HMRC did on a protective basis on 23rd December 2014.

12.

The appeals were heard together in the UT on 7th and 8th December 2015 before Mr Justice Nugee and Judge Timothy Herrington. The UT gave its decision on 10th March 2016. I shall return to the details of that decision in due course.

The allegations in the Statements of Case

13.

The Statements of Case in both cases are long and complex. But the critical aspects of the decisions considered by both the FTTs and the UT are confined to a small number of paragraphs.

14.

The following allegations were made by HMRC against E Buyer in its consolidated Statement of Case:-

“38. [HMRC] contend that [E Buyer’s] transactions formed part of an overall scheme to defraud the Revenue, that the scheme involved an orchestrated and contrived series of transactions, and that there were features of those transactions which demonstrate that [E Buyer] knew or ought to have known that this was the case. It is not necessary to prove the identities of those responsible for orchestrating the overall scheme in order to resolve the appeals brought by [E Buyer] …

80. The circumstances of [E Buyer’s] trading and those it was trading with could or should have left no doubt that its transactions were connected to fraud. [E Buyer] was well aware of the risks presented by trading in the wholesale market for these goods with companies that were small, recently established and unknown in the industry and yet it continued to do so even when informed of tax losses in its transaction claims. The inevitable conclusion from this is that it must have known or should have known of its connection to deals and traders that were fraudulent. The information known to [E Buyer] about its suppliers and customers and how it reacted to that information make good that point.

104. It is evident from the above that [E Buyer’s] checks upon its trading partners were often instigated after trading had commenced, that risk factors were overlooked, that checks required by [E Buyer’s] own procedures were left uncompleted and that even where companies were not approved trading went ahead.

119. Prior to 2008 [E Buyer] as a company had no history or experience in the wholesale market for electrical goods nor had it experience in the dispatch or export of such goods. It created the B2B team to address this market. However, so far as [HMRC] have been able to establish its trading in this wholesale market was almost exclusively connected to fraud. This cannot be coincidental.”

15.

HMRC’s original Statement of Case against Citibank, which was the document considered by Judge Mosedale before she ordered it to be amended, included the following allegations:-

“49. [HMRC] contend that [Citibank’s] transactions formed part of an overall scheme to defraud [HMRC], that the scheme involved an orchestrated and contrived series of transactions, and that there were features of those transactions which demonstrate that [Citibank] knew or ought to have known that this was the case. It is not necessary to prove the identities of those responsible for orchestrating the overall scheme in order to resolve the appeals brought by [Citibank] …

96. The trade by [Citibank] in EUAs in July 2009 was connected with the fraudulent evasion of VAT by 3 defaulting traders. [Citibank] was well aware of the risks of trading in this market by the 8th July 2009 and recognised that its principal supplier at that time, SVS was a suspect trader. However, [Citibank] continued to buy millions of pounds worth of EUAs regardless of its own expressed concerns, both general and particular. The only reasonable explanation for its behavior is that it knew or should have known of its connection with VAT fraud”.

The FTT’s decision in the Citibank case

16.

Judge Mosedale was dealing with Citibank’s 24-page request for further and better particulars of HMRC’s 57-page Statement of Case. Citibank contended that there were possible inferences of fraud in the Statement of Case, but that fraud was not clearly pleaded, and that the Statement of Case lacked the necessary detail to let Citibank know what was alleged against it.

17.

The judge reviewed the special rules of pleading that operate in the Tribunal. Rule 2(2) of the 2009 Rules provided that the objective was to deal with cases fairly and justly. The judge said that litigation by ambush was not fair and just. Rule 25(2)(b) of the 2009 Rules provided that the Statement of Case must “set out the respondent’s position in relation to the case”. Further she concluded that, while the strict rules of pleading in the courts which might prevent HMRC relying on an allegation which was only contained in a witness statement, did not apply in the FTT, those strict rules offered a guide to what a Statement of Case ought to contain. She asked by reference to Andrew Smith J’s dictum in Gamatronic (UK) Ltd [2013] EWHC 3287 (QB) whether HMRC’s Statement of Case made its allegation plain. She considered that question in the light of the 70 requests for more information.

18.

Under the heading “Allegations of Dishonesty?”, Judge Mosedale said that she agreed with Citibank, though HMRC did not really disagree, that a pleading of fraud must be made plain in the Statement of Case, citing Lord Millett in Three Rivers DC v. Bank of England (No 3) [2003] 2 AC 1 at paragraphs 184-6, to the effect that fraud or dishonesty had, if alleged, to be distinctly pleaded, particularised and proved as a matter of both pleading and substance. Proof of primary facts that tilt the balance and justify an inference of dishonesty must be both pleaded and proved.

19.

The judge then moved on to ask “[w]as dishonesty/fraud alleged?”. She recorded HMRC’s submission that dishonesty was not alleged, because the first limb of the Kittel test did not involve making an allegation of dishonesty when input tax was denied on the basis that the taxpayer knew its transaction was connected to fraud. The judge then made a finding that HMRC had described an MTIC fraud in paragraphs 14-21 of their Statement of Case. She recorded HMRC’s position that they did not have to prove that the taxpayer understood the nature of that fraud in order to succeed.

20.

The judge then said at paragraph 23 that “HMRC’s position disintegrates … when I consider what facts HMRC seek to prove that the appellant [Citibank] knew that its transactions were connected to fraud”. She said that “Most significantly, as is almost inevitable in a case where MTIC fraud is alleged, the [Statement of Case] here alleges that [Citibank] knew that its transactions were contrived”. She cited paragraph 49 of the Statement of Case supra, and concluded as follows:-

“24. In my view, if an appellant is shown to know in advance that its purchase and sale were orchestrated by a third party in order to perpetrate a fraud on HMRC (or indeed on anyone), its decision to proceed with the transactions knowing this would be dishonest.

25. In other words, as part of its case that the appellant knew that its transactions were connected to fraud, HMRC seek to prove that the appellant acted in a dishonest fashion: they seek to prove that it went ahead with transactions which it knew were orchestrated for the purpose of fraud”.

21.

The judge then went on to consider the authorities. She cited Briggs J’s dictum in Megtian Ltd v. HMRC [2010] EWHC 18 (Ch) at paragraph 41 (“Megtian”) as good authority for the proposition that “[a] person who knows that a transaction in which he participates is connected with fraudulent tax evasion is a participant in that fraud” and “has a dishonest state of mind”. She rejected HMRC’s submission that this position was inconsistent with the subsequent decision of the Court of Appeal in Mobilx Ltd v. Revenue and Customs Commissioners [2010] EWCA Civ 517 (“Mobilx”). She placed some reliance on a suggestion made by Moses LJ (whom she mistakenly referred to as Millett LJ) that the behaviour there had to be put to a witness in cross-examination. Thus, although Moses LJ twice drew a distinction in his judgment between a fraudulent trader and one with knowledge of another’s fraud, she interpreted his judgment on balance as being consistent with her view that:-

“31. The context of what [Moses LJ] said was in a reference back to what the CJEU said at [53] and [56] of its decision [in Kittel]. In these two paragraphs, the CJEU drew a distinction between ‘where tax is evaded by the taxable person himself’ ([53]) and a person who knew or should have known its transaction was connected to fraud ([56]). The former transaction described at [53] was not an economic activity, so there would be no right to recover input tax; whereas the transaction described at [56] was economic activity but one in which the right to deduct input tax was lost. In these two paragraphs, the CJEU was not drawing a distinction between honesty and dishonesty; it was drawing a distinction based on whether the transactions were an economic activity or not. [Moses LJ], no more than the CJEU, stated that a person in the second category was honest if, although not actually fraudulently evading payment of VAT by its transaction, that person nevertheless entered into a transaction knowing it facilitated fraud by someone else. On the contrary, it stands to reason that such a person is not acting honestly. And that explains why later at [84] [Moses] LJ said such allegations had to be specifically put to the witness.

32. So I do not consider that there is anything in Mobilx which casts any doubt on what was said by Briggs J in the passage above, and which, for the reasons I have given at 25, must be right in law.”

22.

She then went on to hold that the Kittel test, which refers to transactions “connected with fraud” means that a transaction must “facilitate” the fraud, and that “a person entering into such a transaction necessarily has a dishonest state of mind”. This culminated in a direction for HMRC to amend its case to plead dishonesty expressly. The following paragraph of her decision attracted the most attention in oral argument before us:-

“40. HMRC ought to amend its statement of case to make it clear whether or not it is alleging a dishonest state of mind against the appellant. And, if it does not amend its [Statement of Case] to allege a dishonest frame of mind, at the hearing it must not ask the Tribunal to find that the appellant knew its transactions were contrived, nor the appellant knew its transactions facilitated fraud by others, nor, indeed, that the appellant knew its transactions were connected to fraud”.

23.

Judge Mosedale then reviewed the nine heads of objection made against the Statement of Case. She held, in effect, that HMRC could not plead the elements in the first limb of Kittel without alleging dishonesty:-

“129. … I do consider that the [Statement of Case] was seriously flawed and some of [Citibank’s] complaints about it were justified. The [Statement of Case] was seriously flawed as either HMRC intended to allege behaviour amounting to dishonesty but failed to plead it with clarity or HMRC did not intend to allege behaviour amounting to dishonesty but nevertheless insinuated it. In a few instances, it failed to plead the primary facts relied on. In summary, the defects were: (1) The [Statement of Case] failed to make clear whether or not HMRC is alleging a dishonest state of mind against [Citibank]. If HMRC do not (successfully) apply to amend the SOC to allege a dishonest frame of mind, they must not make allegations at the hearing that [Citibank] knew its transactions were contrived, or that [Citibank] knew its transactions facilitated fraud by others, or indeed that [Citibank] knew its transactions were connected to fraud”.

The FTT’s decision in the E Buyer case

24.

The basis of E Buyer’s case was that HMRC’s Statement of Case did not provide the necessary clarity to allow it to prepare its defence. Judge Walters held that the Tribunal had no power to strike out the Statement of Case under Rule 8 of the 2009 Rules. He considered directing HMRC to amend their Statement of Case, but came to the view that it adequately set out HMRC’s position as it stood. Unlike Judge Mosedale, Judge Walters accepted HMRC’s submission that no allegation of fraud was made in the Statement of Case, and held E Buyer’s applications to be misconceived. He held as follows:-

“35. I accept that there is no allegation made by HMRC of involvement as a co-conspirator in any fraud, but instead there is an allegation of knowledge of the connection with fraud. The references in the [Statement of Case] to an overall scheme to defraud the revenue involving an orchestrated and contrived series of transactions are, I accept, details as to relevant circumstances surrounding [E Buyer’s] transactions, which HMRC may relevantly allege in order to establish, if they can, the context in which [E Buyer’s] transactions took place. Establishing this context will be relevant as it will assist the tribunal in determining what [E Buyer] knew or ought to have known (see: the passage from the judgment of Christopher Clarke J in Red 12 Trading Limited v R&C Commissioners [2009] EWHC 26523 (Ch) at [109] to [111] cited with approval by Moses LJ in [Mobilx]”.

25.

This passage from Christopher Clarke J’s judgment in Red 12 supra reads as follows:-

“109. Examining individual transactions on their merits does not, however, require them to be regarded in isolation without regard to their attendant circumstances and context. Nor does it require the tribunal to ignore compelling similarities between one transaction and another or preclude the drawing of inferences, where appropriate, from a pattern of transactions of which the individual transaction in question forms part, as to its true nature e.g. that it is part of a fraudulent scheme. The character of an individual transaction may be discerned from material other than the bare facts of the transaction itself, including circumstantial and “similar fact” evidence. That is not to alter its character by reference to earlier or later transactions but to discern it.

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

111. Further in determining what it was that the taxpayer knew or ought to have known the tribunal is entitled to look at the totality of the deals effected by the taxpayer (and their characteristics), and at what the taxpayer did or omitted to do, and what it could have done, together with the surrounding circumstances in respect of all of them.”

26.

I digress to record that we were told in the course of oral argument that this passage from Red 12 supra is regarded as a road map for FTT hearings in this kind of case.

27.

Judge Walters then said that the proper course would be for E Buyer to apply for a properly focussed direction requiring HMRC to provide documents, information, or submissions pursuant to rule 5(2)(d) of the 2009 Rules. Turning to the applications concerning disclosure, Judge Walters said:-

“40. As to [E Buyer’s] application at this stage for ‘standard disclosure’ going beyond what is provided for by rule 27 of the [2009] Rules, I reject that also. Litigation in this tribunal is intended to conform to a different model from litigation in the High Court and the [2009] Rules establish the framework within which litigation in this tribunal is to be carried on. Rule 27 provides for the normal disclosure in a standard or complex case and I consider it would not be appropriate for me, at this stage in this litigation, to require wider disclosure than that required by rule 27. It is, in my view, no answer to complain in this forum about the inadequacy of the terms in which rule 27 is framed.

41. I also consider that HMRC’s application to dispense with the requirement on each party to send or deliver a list of documents ought to be allowed in the interests of convenience and saving costs. I will direct, as HMRC requested, that the exhibits to the parties’ witness statements shall stand as the documents which they respectively intend to rely upon or produce in these proceedings. In my view, no substantial injustice will arise from this direction and it is in harmony with the view I take that [E Buyer] must ascertain the full particulars of the case it has to meet from a close examination of HMRC’s witness evidence”.

The UT’s decision

28.

Having referred to the two FTT decisions, the UT recorded at paragraph 14 of its decision that it could be seen that “two experienced FTT judges who have each heard many MTIC appeals have come to opposite conclusions on the question as to whether allegations of the nature pleaded [and set out above] have to be pleaded with the level of particularity to be expected when allegations of dishonesty are made in civil fraud proceedings”. The UT said that, in these circumstances, it was appropriate for it to give “general guidance as to the proper approach to be taken by HMRC in [Statements of Case] which relate to MTIC appeals that allege knowledge on the appellant’s part that its transactions in respect of which input tax is denied are connected to fraud in the manner alleged in the E Buyer and Citibank [Statements of Case]”. It said it was conscious the UT should be slow to interfere in the FTT’s case management decisions (referring to Norris J in Goldman Sachs v. Revenue and Customs [2009] UKUT 90 (TCC) at [23] and [24]). The UT then set out the relevant parts of the 2009 Rules, which it is not necessary to repeat here.

29.

After dealing with the facts and the FTT decisions, the UT set out the English law principles relating to fraud and dishonesty at paragraphs 45-49, including the well-known dicta of Millett LJ in Armitage v. Nurse [1998] Ch 241 at page 256 and of Lord Millett in Three Rivers DC supra at paragraph 186. The UT cited Judge Wallace’s dictum in Blue Sphere Global Ltd v. HMRC [2008] UKVAT 20694 at paragraph 30 to the effect that the principle that any allegation of fraud must be clearly pleaded with particulars, applies as much to tax appeals as to any other litigation.

30.

The UT then turned to the CJEU cases at paragraphs 50-63, starting with the principle in Halifax plc v. Customs and Excise Commissioners (Case C-255/01) [2006] STC 919 (“Halifax”), which it described as the “starting point”. It cited paragraph 59 of the CJEU’s judgment in Halifax where it said that it was true that “those criteria are not satisfied where tax is evaded, for example by means of untruthful tax returns or the issue of improper invoices”. That “finding” meant that the “Halifax principle” was that “where the tax is evaded by the taxable person himself the requirements of there being a supply of goods by the taxable person acting as such and the carrying on of economic activity will not be met and there will be no right to deduct input tax in relation to the transactions concerned”.

31.

The UT then contrasted Halifax with Optigen Ltd, Fulcrum Ltd and Bond House Systems Ltd v. Commissioners of Customs & Excise (Cases C-354/03, C-355/03 and C-484/03) [2006] ECR I-483, which concerned traders implicated in a so-called “carousel fraud” who neither knew, nor could be expected to know, of the fraud. These two cases, according to the UT, marked the outer limits of the effect of fraud on the right to deduct or reclaim VAT. I shall return in due course to say something specifically about this analysis.

32.

The UT summarised its view of the Kittel case as follows:-

“57. Before coming to the later authorities, we will give our own views of what can be taken from this decision. First, the Court of Justice was concerned to lay down a clear line between those who do not, and those who do, lose their right to deduct input tax. That line is drawn between those who come within the Optigen principle—variously expressed as applying where a taxpayer “had no knowledge and no means of knowledge” of the possible fraudulent nature of another transaction in the chain of supply ([44]); or where a prior or later transaction is vitiated by fraud, or the seller himself is committing fraud, without the taxpayer “knowing or having any means of knowing” ([45]-[46]); or the taxable person “did not and could not know” that the transaction concerned was connected with a fraud committed by the seller ([52], [60]) – and those who by contrast come within the Kittel principle, repeatedly identified as those who “knew or should have known” that they were taking part in, or participating in, a transaction connected with a fraudulent evasion of VAT ([56], [58], [61]). We will call this class of taxable persons “the Kittel class”.

58. Second, as these citations show, the line was expressed by the Court of Justice in terms of knowledge. As Mr Kinnear [for HMRC] said, the word “dishonesty” does not feature in the judgment. This is perhaps unsurprising given the questions that were asked, but we certainly accept that the Court of Justice did not express its conclusions in terms of dishonesty. And it would have been quite odd if it had, given that the Court of Justice included within the Kittel class those who “should have known” of the connection: someone who should have known that his transaction was connected to a fraud, but did not in fact, would not in any usual sense of the term be regarded as dishonest. Having included those who should have known, as well as those who knew, within this class, there is no particular reason why the Court should have said anything about the honesty of the Kittel class as a whole.

59. Third, on the other hand, the Court of Justice held that a person in the Kittel class should be “regarded as a participant in the fraud” ([56]) who “aids the perpetrators of the fraud and becomes their accomplice” ([57]). This is the basis on which the Kittel class is denied the right to deduct—they are equated for VAT purposes with those who evade tax themselves (see the opening words of [56]: “In the same way…”).

60. Fourth, there are two limbs of the Kittel principle, namely that the Kittel class includes both those who have actual knowledge of the connection to fraud and those who should have known. The Court of Justice does not distinguish between them in any of its remarks, nor, as we have said, does it say anything in terms about the honesty or otherwise of the Kittel class, but in the case of those who come within the first limb, that of actual knowledge, it may perhaps be doubted whether the Court would have regarded their knowing participation in a transaction connected with fraud such that they must be regarded as participants in fraud, or as knowing accomplices to and aiding fraud, as consistent with honesty.

61. In summary therefore, Kittel laid down at the European level a test of “knew or ought to have known”; this is plainly not an honesty test, and the word “dishonesty” does not feature in the judgment; but this does not answer what we regard as the real question in these appeals which is whether, in a case where HMRC choose to allege actual knowledge of a fraud under what we have called the first limb of Kittel, such an allegation is tantamount to, or attracts the same consequences in domestic law as, an allegation of dishonesty”.

33.

The UT then dealt at paragraphs 64-88 with the English authorities that had considered the Kittel principles. In Revenue and Customs Commissioners v. Livewire Telecom, Revenue & Customs Commissioners v Olympia Technology Ltd [2009] EWHC 15 (Ch) (“Livewire”), a joined appeal from the VAT and Duties Tribunal, Lewison J expressly rejected the submission that Kittel laid down a test of dishonesty. The UT stressed the importance of reading statements of principle against the background of the facts of each case, and observed that Livewire was a case falling under the second limb:-

“66. That is the context in which Lewison J, in an extended passage at [84]-[85], considered and rejected a submission on behalf of both taxpayers that the Court of Justice was laying down a test of dishonesty. It is not necessary to cite it all: it is sufficient to refer by way of example to the following extracts from [85]:

‘… it seems to me that the ECJ was at pains to stress that the test was not one of dishonesty … A requirement to take all reasonable precautions … is incompatible with a simple test of dishonesty … In addition, it seems to me that the proposition that whether a person knew or should have known is to be tested by objective facts or factors … is also inconsistent with a simple test of dishonesty”.

34.

The UT accepted Lewison J’s statement of the law as “plainly right”, and held that the allegation that a taxpayer “should have known” that his transaction was connected with a fraud but did not in fact know, was not an allegation of dishonesty.

35.

The UT then dealt with Megtian supra saying:-

“68. Briggs J in the result rejected that contention [i.e. that HMRC’s pleadings against the taxpayer were insufficiently clear] being satisfied that from start to finish HMRC advanced a case that Megtian actually knew that its transactions were connected with fraud, as well as the alternative case that it ought to have known (see at [40]). That conclusion of course turned on its own facts, but in the course of discussing this ground of appeal Briggs J said:

‘41. It is important to bear in mind, although the phrase “knew or ought to have known” slips easily off the tongue, that when applied for the purpose of identifying the state of mind of a person who has participated in a transaction which is in fact connected with a fraud, it encompasses two very different states of mind. A person who knows that a transaction in which he participates is connected with a fraudulent tax evasion on a participant in that fraud. That person has a dishonest state of mind. By contrast, a person who merely ought to have known of the relevant connection is not dishonest, but has a state of mind broadly equivalent to negligence.

42. The distinction between dishonesty and negligence is of fundamental importance, even in cases such as the present where proof of either of them will suffice for the opposing party’s purpose. For that reason, an allegation of dishonesty in civil litigation must be clearly and specifically pleaded, and, if the person against whom dishonesty is alleged gives oral evidence, it must be specifically put in cross-examination. These principles apply to all civil litigation, including tax appeals: see for example Revenue and Customs Commissioners v Noel Dempster [2008] EWHC 63 (Ch)’”.

36.

The UT noted that, although there was an apparent tension between Megtian, and Livewire, that tension could be resolved. Briggs J was not concerned with the technicalities of common law dishonesty and negligence, nor was he equating the first limb of Kittel with the common law concept of dishonesty:-

“72. Rather, as we read this part of his judgment, [Briggs J] was making a broader point, which is that the rolled-up plea of ‘knew or ought to have known’ conceals two very different states of mind, there being a significant difference between accusing someone of knowing that his transaction is connected with fraudulent tax evasion, and merely alleging that they ought to have known. The former, which he characterised as involving a dishonest state of mind, requires to be clearly and specifically pleaded and specifically put in cross-examination. This is all preparatory to his examination of the question whether on the facts before him HMRC’s case before the Tribunal was clearly based on the first limb of Kittel as well as the second. As already mentioned, he concluded that it was, but it is of some interest how he expresses his conclusion, namely:

‘49. Those references demonstrated to me that, from start to finish, HMRC advanced a case that Megtian knew that the transactions upon which it based its input tax claims were connected with tax fraud, and a case in the alternative that, if it did not know of the connection, it ought so to have known. Indeed, my reading of those materials is that HMRC’s primary case was that, from start to finish, Megtian was a knowing participant in a contrived, pre-ordained series of transactions designed to achieve the evasion of tax rather than, as Mr Andreou maintained in his evidence (and upon which he was disbelieved) that those transactions were separate arm’s length commercial deals negotiated with individual and independent traders in a competitive fast-moving market.

50. Mr Patchett-Joyce frankly and very properly acknowledged (so as to avoid a time consuming trawl through the transcript of the lengthy cross-examination of Mr Andreou) that a case that Megtian knew that the relevant transactions were connected with tax fraud, in the sense which I have just described, was properly put by way of cross-examination.’

It can be seen that what he concludes is that HMRC did advance a case that Megtian ‘knew that the transactions… were connected with tax fraud’, and that the case that Megtian knew this was properly put in cross-examination. In other words, Briggs J was not saying that in a case brought under the first limb of Kittel, what HMRC needed to allege in terms, and put to a witness in terms, was that the taxpayer had acted ‘dishonestly’. What he was saying was that where HMRC advanced a case under the first limb, they needed to make it clear, with sufficient particularity, and to put in cross-examination, the allegation that the taxpayer actually knew that the relevant transactions were connected with tax fraud; and that this was because to allege such a thing was in effect to allege a dishonest state of mind and attracted the usual consequences of making such an allegation”.

37.

In paragraphs 73 and 74, the UT acknowledged the common ground between the parties that not all cases where knowledge within the first limb of Kittel was alleged amounted to an allegation of dishonesty, and said they did not read Briggs J as seeking to lay down any universal rule to the contrary. The UT did think, however, that in many first limb cases, it would be quite plain that the taxpayer would have been dishonest by any standards.

38.

The UT also rejected HMRC’s submission that Megtian could not stand in the light of Moses LJ’s judgment in Mobilx. The UT examined the facts of the three cases joined in Mobilx in detail, and observed that the first of them, Calltel Telecom Ltd, was a clear case of actual knowledge; it was “not surprising in those circumstances that the judgment of Moses LJ contain[ed] little or no discussion of the first limb” (paragraph 76). The other two appeals, Blue Sphere Global Ltd and Mobilx, were cases decided under the second limb, and “none [of the challenges raised in the appeals] raised any question as to the precise ambit of the first limb; far less [did] they involve any discussion as to whether conduct within the first limb does or does not involve ‘dishonesty’, a word which does not feature in the judgment” (paragraph 79).

39.

The UT considered Moses LJ’s comment in paragraph 41 that Kittel “enlarged the category of participants to those who themselves had no intention of committing fraud”, but rejected the submission that this contradicted the approach it took on the basis of Megtian:-

“81. … This does not seem to us to be what Moses LJ was saying. As we read Moses LJ’s judgment, all he meant was that the Court of Justice had previously (in Halifax) dealt with the case where the taxable person evaded or intended to evade tax himself, for example by failing to account for tax that was due from him; and then (in Kittel) extended the same treatment to others who did not themselves intend to evade their own tax, but who were to be treated as participants in the fraud. There is here no warrant for reading Moses LJ as saying anything about whether persons treated as participants in this way would or would not be dishonest, a question that was not raised before him and was not relevant to the issue he was discussing. Whether such a person would be dishonest would depend on the facts. A person within the Kittel class might have no intention of himself defaulting on any tax due from him and yet be centrally involved in the scheme to defraud HMRC, and undeniably honest; or he might in fact be unaware of the scheme to defraud HMRC and only within the Kittel class because he ought to have been aware, in which case as we have already said we think it plain that he would not be dishonest. In short, we do not think that what Moses LJ says in [41] can be used to resolve this question; nor for that matter do we think it undermines what Briggs J said in Megtian”.

40.

Following this review, the UT said this at paragraphs 85 and 86:-

“85. … We agree, as we have already said, that the test laid down in Kittel is one of knowledge, not of dishonesty; it is certainly not to be regarded as synonymous with dishonesty as understood by the common law. But again we do not regard this as the real question, which is whether in a case where HMRC choose to allege actual knowledge of a fraud under the first limb of Kittel, such an allegation is tantamount to, or attracts the same consequences in domestic law as, an allegation of dishonesty.

86. We have now considered the main authorities cited to us on the Kittel principle, including all the relevant decisions of the Court of Justice or Court of Appeal. Our conclusions from the review of these authorities are as follows:

(1) The Court of Justice in Kittel did not lay down a test of dishonesty but of knowledge, or to be more specific whether the taxpayer knew of should have known that he was taking part in a transaction connected with fraudulent evasion of VAT.

(2) Ultimately the question in every Kittel case is whether HMRC has established that the test has been met. The test is to be applied in accordance with the guidance given by the Court of Appeal in Mobilx and Fonecomp.

(3) It is certainly possible for that test to be satisfied without the taxpayer being dishonest. This is likely to be the case for example where HMRC rely exclusively on the second limb of Kittel as (by the time the appeal reached the High Court) they did in Livewire: see Livewire at [84]-[85] per Lewison J.

(4) On the other hand where HMRC rely on the first limb of Kittel and allege that the taxpayer actually knew that he was taking part in a transaction connected with fraudulent evasion of VAT, this will very often amount to an accusation of conduct that would be regarded as dishonest by an English court: see Megtian at [41] per Briggs J.

(5) It does not necessarily follow that all cases of actual knowledge within the first limb of Kittel allege conduct that would be regarded as dishonest by an English court, and there may be cases where the taxpayer satisfies the first limb of Kittel but would nevertheless not be regarded as dishonest. As we have said above, this was in fact common ground between the parties.

(6) But whether the case advanced by HMRC in any particular case does or does not allege conduct that would be regarded as dishonest does not turn on whether they chose to allege that the taxpayer is himself evading tax under the Halifax principle or has knowledge that his transaction is connected with fraudulent evasion under the Kittel principle; it turns on what HMRC actually allege in each particular case.

87. We add a few comments on the last point. Some of Mr Kinnear’s submissions seemed to us to amount to saying that there was a distinction between (i) a case where HMRC alleged that the taxpayer was himself party to a fraud and dishonest under the Halifax principle and (ii) a case where HMRC chose to make no such allegation but to proceed under the first limb of the Kittel principle, in which case all that they had to allege and prove was knowledge, and it followed that they were not accusing the taxpayer of dishonesty at all. This seems to us to be a false dichotomy. As we understand the Court of Justice in Kittel, the distinction between the Halifax principle and the Kittel principle is not that one applies to dishonest conspirators in a fraud and the other to those who “merely” know of a connection; but that the Halifax principle applies to those who are evading their own tax (and operates by denying their transactions the status of a “supply of goods” and “economic activity” within the meaning of the VAT Directive), whereas the Kittel principle applies to those who know (or should have known) that their transactions are connected to such evasion. This latter class, the Kittel class, are not evading their own tax but are treated as participants in the fraudulent evasion with the result that their own transactions, even though otherwise meeting the objective criteria for supply of goods, are also denied that status. The Kittel class, as we have said, may include both those who are dishonest and those who are not; whether in any particular case HMRC are alleging dishonesty against a taxpayer therefore does not depend on whether HMRC have brought the case under the Halifax principle or the Kittel principle, but on what HMRC allege the facts to be in the particular case”.

41.

The UT then went on to apply the principles it had set out to Citibank in paragraphs 89-98. It found that HMRC’s Statement of Case made a rolled-up plea of both actual and constructive knowledge (paragraph 91), which it found in the circumstances to have presented alternative allegations (despite the opposite default principle in Armitage v. Nurse supra), including an allegation of actual knowledge (paragraph 92). The UT then asked whether the allegation against Citibank amounted to an allegation of dishonest conduct. It said this:-

“93. In our view it does. We do not see how a trader in the position of [Citibank] could honestly proceed with a transaction if it knew that the transaction in question was part of an orchestrated and contrived scheme to defraud the Revenue. No argument was we think addressed to us to the effect that such conduct could be consistent with honesty.

94. When we come to Judge Mosedale’s decision in the FTT, we therefore agree with her at [23] that the Statement of Case alleges that the appellant knew that its transactions were contrived; at [24] that if an appellant is shown to know in advance that its purchase and sale were orchestrated by a third party in order to perpetrate a fraud on HMRC, its decision to proceed would be dishonest; and at [25] that HMRC were seeking to prove that the appellant acted in a dishonest fashion, by seeking to prove that it went ahead with transactions which it knew were orchestrated for the purpose of fraud. We see no error of law in these passages.

95. The FTT was therefore in our view fully justified in the view that the allegation in fact made in this case was an allegation of dishonesty. We would accept that in general it is not necessary to use the word “dishonesty” in a pleading if it is clear what is being alleged: see the passage cited from Millett LJ in Armitage v Nurse at [45] above. But in the present case, Judge Mosedale records (at [37]) that HMRC’s position at the hearing was that they did not allege dishonesty against the appellant. It was not suggested that she had misunderstood HMRC’s position. In these circumstances we consider that she was justified in taking the view that HMRC’s position was unclear, being on the one hand willing to allege that the appellant knew that its transactions formed part of a contrived scheme designed to defraud HMRC, but on the other unwilling to accept that they were alleging dishonesty.

96. In these circumstances we see nothing wrong in her direction that HMRC should clarify their stance by either expressly alleging dishonesty, or expressly disclaiming it. In our view the order she made at [40] was therefore justified, and we dismiss the appeal against this order”.

42.

The UT applied these principles to E Buyer in paragraphs 99-109. The Statement of Case against E Buyer was in identical terms to that against Citibank (paragraph 100), again using a rolled-up plea which the UT held in the circumstances made an allegation of actual knowledge and constructive knowledge in the alternative. For the same reasons, it held that this amounted to an allegation of dishonest conduct against E Buyer:-

“101. On this basis HMRC here allege that E Buyer actually knew both that there was an overall scheme, involving an overall scheme, involving an orchestrated and contrived series of transactions, to defraud the Revenue, and that its transactions were part of this scheme. For reasons given in relation to the Citibank appeal we consider that this does amount to an allegation of dishonest conduct against E Buyer.

102. Given this, we take the view that the general principles applicable to the pleading of a case of fraud or dishonesty (which, as we say, were not disputed before us) do apply, with the result that E Buyer is entitled to have the allegation that it knew that its transactions were part of an orchestrated and contrived scheme to defraud the Revenue not only clearly alleged against it but properly particularised: see [45]-[48] above. Judge Walters refused the application for particulars because he took the view that there was no allegation made by HMRC of involvement as a co-conspirator in any fraud, but merely an allegation of knowledge of the connection with fraud, the references to an overall scheme to defraud the Revenue involving an orchestrated and contrived series of actions being a reference to the relevant circumstances surrounding E Buyer’s transactions, which HMRC might relevantly allege to establish the context (at [35] of the FTT’s Decision) … For reasons we have given, we think this understates the effect of HMRC’s allegation; whether or not it is an allegation of being a co-conspirator in a fraud, it is an allegation of dishonesty, and the references to the overall scheme to defraud the Revenue are not just a reference to the relevant circumstances surrounding E Buyer’s transactions, but are part of what E Buyer is alleged to have known, namely that its transactions formed part of such a scheme.

103. In these circumstances we think that E Buyer was entitled to proper particulars of what it is said to have known and the facts on which such a case is based…”.

43.

This did not mean, in the UT’s judgment, that E Buyer had to be given particulars of every aspect of the alleged VAT fraud. Exhaustive and hugely detailed requests for information would not necessarily be justified or proportionate, and should not be treated as a game to be played nor as a rigid entitlement. The proper course of action, it held, was to remit the question of particulars back to the FTT for reconsideration (paragraph 105).

44.

Turning to the question of disclosure on E Buyer’s appeal from the decision of Judge Walters, the UT said:-

108. We accept that rule 27 demonstrates that disclosure in the FTT does not simply mirror that in ordinary civil litigation. But we do not think it follows that this means that E Buyer’s disclosure request should have been rejected. Rule 27 applies to a wide variety of cases, and as we said above (at [19]) the Rules provide a wide variety of powers to give effect to the overriding objective. This includes the power in rule to order disclosure of other documents. It does not seem to us to be an answer to a request for wider disclosure that, in effect, a litigant before the tribunal is stuck with rule 27 and cannot complain about its inadequacy. If a reasoned case is made for wider disclosure being proportionate to the importance of the case, the complexity of the issues, the anticipated costs and the resources of the parties, the tribunal should consider whether such disclosure is appropriate without being trammelled by the limited terms of rule 27. Notwithstanding that it was a case management decision, we do not think Judge Walters’ decision did address the specific points made by E Buyer.

109. It therefore falls to us to consider this application afresh. We consider there is force in Mr Wardell’s submissions on the facts of this case. The burden on proof is on HMRC. They hold all the cards, having carried out extensive investigations and gathered information, as they are able to do, from various sources. E Buyer is necessarily unable to have any insight into what material HMRC have uncovered other than that which HMRC chooses to rely on, and is not in a position to carry out a similar investigation itself. In the circumstances it does seem to us to be appropriate for HMRC to disclose not only the material that they wish to rely on, but also any other material uncovered in the course of their investigations which might undermine that case. So far as proportionality is concerned, there is little doubt that this case, alleging dishonesty against E Buyer and involving some £7m of tax, is both of very great importance to the taxpayer and highly complex. The case is also no doubt important to HMRC: we do not underestimate the significance of the Kittel principle to HMRC’s continuous attempts to combat the ingenuity of those who wish to abuse the VAT system for fraudulent ends. But it does seem to us that fairness to E Buyer requires that it be given the disclosure that it seeks”.

45.

The UT therefore dismissed HMRC’s appeal in relation to Citibank, set aside Judge Walters’ decision in relation to E Buyer and remade that part of the decision which involved an error of law in the disclosure direction by directing disclosure in the terms of E Buyer’s application. The UT remitted E Buyer’s application for further and better particulars to the FTT for consideration.

The arguments advanced by HMRC

46.

HMRC’s arguments can be shortly stated. The UT was wrong to conclude that an allegation, made under the first limb of Kittel, that a taxpayer knew that its transactions were part of an orchestrated scheme to defraud HMRC required a pleading of dishonesty. Actual knowledge should not be equated with dishonesty, which was what the UT did, despite acknowledging that it should not be done. The UT’s decision creates another hurdle for HMRC to overcome in these cases by requiring them to plead and prove dishonesty against the taxpayer, when the true test is knowledge.

47.

The UT wrongly concluded that the facts alleged in support of HMRC’s reliance on the first limb of the Kittel test determine whether or not HMRC are alleging dishonesty as a state of mind against the taxpayer.

48.

It was an error of law to require HMRC to plead dishonesty, which is not alleged and is not required under the first limb of the Kittel test. The UT should have held that the relevant test for the cause of action, rather than the specific facts alleged to make out the cause of action, dictated the requirements of the pleadings. Mr Jonathan Kinnear QC, counsel for HMRC, put the point succinctly in closing that, as HMRC was not required to prove actual dishonesty to make out its cause of action, it should not be required to plead it. Dishonesty is a complex nuanced concept in English law that HMRC should not be required to plead.

49.

HMRC relied particularly on four dicta as follows:-

i)

Moses LJ’s judgment in Mobilx supra at paragraph 41:-

Kittel did represent a development of the law because it enlarged the category of participants to those who themselves had no intention of committing fraud but who, by virtue of the fact that they knew or should have known that the transaction was connected with fraud, were to be treated as participants. Once such traders were treated as participants their transactions did not meet the objective criteria determining the scope of the right to deduct”.

ii)

Arden LJ’s judgment in Fonecomp Ltd v. Revenue & Customs Commissioners [2015] EWCA Civ 39, [2015] STC 2254 (“Fonecomp”), where she said this at paragraph 45:-

“The CJEU determined in Kittel that knowledge was the essential characteristic of a participant in a fraud who lost his right of deduction. The type of knowledge required is not necessarily derived from any national law test, still less common law fraud but has been devised by the CJEU: see generally Case 296/95 R (o/a EMU Tabac SARL) v Comrs of Customs & Excise [1998] ECR 1-1605 at [30]”.

iii)

Paragraph 100 of the Upper Tribunal’s decision in Universal Enterprises (EU) Ltd v. Revenue & Customs Commissioners [2015] UKUT 311 (TCC), where it decided that HMRC did not need to plead fraud as part of their case that the taxpayer knew that its transactions were connected with fraud:-

“In our view HMRC pleaded the matter properly, and it is apparent that Universal fully understood the case that it had to meet, recognising that the burden of proof lay with HMRC. In particular, what was pleaded required HMRC to satisfy the Kittel test. It may be that evidence to suggest that Universal was itself a participant in the fraud or that it was engaged in a conspiracy would ensure that the Kittel test was not met. The Kittel principle is, however, a principle of the EU VAT system and we do not consider that it requires HMRC to plead either fraud or conspiracy as part of their case. Assuming that HMRC is able to produce evidence sufficient to support its case on the application of the Kittel test to the civil standard, Universal would need to respond essentially by showing that there was in the circumstances a reasonable basis for its transactions so that it would be impossible or unsafe to conclude that the Kittel test was satisfied. Plainly, HMRC having satisfied the FTT that the contra-traders were fraudulent and that Universal’s transactions were connected with the fraud, Universal failed to displace the further conclusion that it had actual knowledge of the fraud.”

iv)

Lewison J’s judgment in Livewire, where he held that the CJEU in Kittel was not laying down a test of dishonesty, cited in part (but not in full) by the UT, above:-

“84. Both Mr Scorey (for Livewire) and Mr Beal (for Olympia) submitted that the ECJ was laying down a test of dishonesty. They buttressed this submission by reference to the principle of equivalence, and submitted that the nearest equivalent domestic concept was that of dishonest assistance in a breach of trust, for which dishonesty is a necessary condition. This was not a question that arose in Just Fabulous, because on the assumed facts the taxable persons were dishonest. In addition they submitted that the right to deduct was an essential part of the VAT scheme and that any restriction on that right was a derogation that must be strictly applied.

85. I do not accept these submissions. First, the very nature of the fraud under consideration is an intra-community fraud (i.e. it is a fraud in which the crossing of national frontiers is inextricably involved). It is very unlikely that the ECJ was laying down a principle that might be differently applied in different Member States to the same chain of transactions. Second, it seems to me that the ECJ was at pains to stress that the test was not one of dishonesty. This comes out most clearly from the ECJ’s answer to the question posed in Teleos, where good faith is not enough on its own (§ 68). The supplier must, in addition to being of good faith, have taken every reasonable measure to ensure that his supply was not participating in VAT evasion. But it is also evident from the court’s statement that a Member State may lawfully impose a requirement on suppliers to take all reasonable precautions in order to preserve their right to deduct (Teleos § 65). A requirement to take all reasonable precautions (or to act with all due diligence and care (Netto A-G § 45)) is incompatible with a simple test of dishonesty. The policy is that an honest and careful trader should not be liable for the frauds of others (Teleos A-G § 77 fn). In addition, it seems to me that the proposition that whether a person knew or should have known is to be tested by objective facts or factors (Kittel § 59) is also inconsistent with a simple test of dishonesty. Moreover, if dishonesty were the only test, it is difficult to see why the court in Kittel would have thought that their interpretation would make it more difficult to carry out fraudulent transactions (Kittel § 58). Third, the objective of preventing VAT evasion is positively encouraged by the Sixth Directive, and consequently measures taken in that respect are not a derogation. As Burton J put it in Just Fabulous (§ 45):

“The principle of legal certainty must be trumped by the ‘objective recognised and encouraged by the Sixth Directive’.””

50.

On the disclosure point, Mr James Puzey, junior counsel for HMRC, submitted that Judge Walters made no error in law, as disclosure is based on a different model in the FTT as compared to the High Court. The UT could only interfere with his decision if he was clearly wrong, which he was not. The UT’s error flowed logically from its error on the dishonesty and pleading point.

Arguments advanced by E Buyer and Citibank

51.

For E Buyer, Mr Wardell submitted that the first ground of HMRC’s appeal was utterly misconceived, and that the UT’s decision made plain that it was importing no gloss into the Kittel test or importing dishonesty as a requirement of the first limb generally. The UT rightly said that dishonesty was not part of the Kittel test, and that the first limb does not automatically attract dishonesty. However, where HMRC plead a case that is only consistent with dishonesty, the case needs the procedural safeguards of pleading particulars to ensure justice.

52.

It was HMRC’s choice, he submitted, whether or not to plead dishonesty on the part of a taxpayer under the first limb of Kittel. In either case, HMRC must give proper particulars to inform the taxpayer of the case against it; Kittel could not absolve HMRC from the usual rules of pleading fraud. It was never E Buyer’s case that the first limb of the test requires a plea of dishonesty, he said, only that HMRC’s Statement of Case against E Buyer contained an allegation which had to be particularised. He accepted that the UT put the case too high when it accepted Briggs J’s dictum in Megtian, but that it saved itself from error by moderating this position later in its judgment. The question whether a case of knowledge implies dishonesty is one for the facts of the individual case.

53.

Turning to HMRC’s Statement of Case against E Buyer, Mr Wardell pointed to the paragraphs cited above as well as numerous others, including paragraphs 105, 106, 108-109, and 111-114. He was rightly willing to concede, however, that the key paragraph 38 did not actually go beyond what is required to be proved under the first limb of Kittel, namely that E Buyer’s transactions (a) formed part of an overall scheme to defraud HMRC, (b) that the scheme involved an orchestrated and contrived series of transactions, and (c) that there were features of those transactions which demonstrate that E Buyer knew or ought to have known that this was the case. It was, however, an allegation “tantamount”, “akin to”, or “effectively” an allegation of dishonesty, because anyone who gets involved in such a scheme to evade tax is dishonest. An allegation “tantamount” to an allegation of dishonesty should, by parity of reasoning, attract the same safeguards, and no such plea should be allowed to be made by insinuation rather than express pleading.

54.

It was difficult, he said, to envisage a case where actual knowledge under the first limb of Kittel was not tantamount to dishonesty. He suggested a general distinction could be drawn between so-called “up-chain” and “carousel” MTIC frauds, and that the former might not entail any allegation of dishonesty. The question was not what must be proved but what must be pleaded. The question is one of English procedural law in the context of a particular type of Kittel case, not the content of the Kittel test itself. Mr Wardell did, however, somewhat retreat from this position by accepting that the UT had redefined the Kittel test’s relationship to Halifax where it criticised the “false dichotomy” between Halifax requiring dishonesty and Kittel requiring actual or constructive knowledge. The first limb of the Kittel test, he said, should be recognised to embrace some cases of dishonesty in appropriate cases.

55.

For Citibank, Mr Michael Conlon QC adopted Mr Wardell’s submissions on the dishonesty point. Unlike E Buyer, Citibank did ask for a direction that HMRC plead dishonesty in the FTT. Mr Conlon rightly accepted that Judge Mosedale may have gone too far at points in her judgment including paragraphs 40 and 129(1), but he said that the appeal really concerned a question of case management about the need for further and better particulars to be provided. Judge Mosedale was entitled to take the point that HMRC’s pleading was defective as originally drafted, and to direct that it be amended to provide further particularity, because Citibank had not been made aware of the case against it. Likewise, there was no error in the UT’s decision to uphold her direction; it could not find that she was plainly wrong and was in turn correct to uphold her case management decision. The UT’s decision was not infected by any error of law relating to the Kittel test and should be upheld.

56.

For E Buyer on the second ground of appeal, Mr David Scorey QC said that Kittel must be looked at through the prism of domestic law, as EU law does not dictate how knowledge or fraud are to be proved or what constitutes a proper plea of knowledge. Given that fraud was alleged against E Buyer, HMRC should have given standard disclosure. Judge Walters did not grapple with the arguments advanced in the FTT, but simply held that the “model” of disclosure was different under the 2009 Rules and the CPR. The UT was correct in holding at paragraph 108 that Judge Walters failed to address the specific points made by E Buyer and deciding to consider the application afresh.

57.

Before dealing with these arguments, it is important, I think, to understand the juridical basis of the applicable principles.

The Halifax principle

58.

As I have said, the UT stated the Halifax principle at paragraph 50 in these terms: “where the tax is evaded by the taxable person himself the requirements of there being a supply of goods by the taxable person acting as such and the carrying on of economic activity will not be met and there will be no right to deduct input tax in relation to the transactions concerned”.

59.

Despite the fact that no argument was addressed to this point, and that it may not ultimately matter to the outcome of these cases, the UT’s description of the Halifax principle does not seem to me to be quite accurate. In Halifax, the CJEU actually said the following:-

“58. It follows that transactions of the kind at issue in the main proceedings are supplies of goods or services and an economic activity within the meaning of arts 2(1), 4(1) and (2), 5(1) and 6(1) of the Sixth Directive, provided that they satisfy the objective criteria on which these concepts are based.

59. It is true that those criteria are not satisfied where tax is evaded, for example by means of untruthful tax returns or the issue of improper invoices. The fact nevertheless remains that the question whether a given transaction is carried out for the sole purpose of obtaining a tax advantage is entirely irrelevant in determining whether it constitutes a supply of goods or services and an economic activity.

60. It follows that the answer to Question 1(a) must be that transactions of the kind at issue in the main proceedings constitute supplies of goods or services and an economic activity within the meaning of Article 2(1), Article 4(1) and (2), Article 5(1) and Article 6(1) of the Sixth Directive, provided that they satisfy the objective criteria on which those concepts are based, even if they are carried out with the sole aim of obtaining a tax advantage, without any other economic objective. …

74. In view of the foregoing considerations, it would appear that, in the sphere of VAT, an abusive practice can be found to exist only if, first, the transactions concerned, notwithstanding formal application of the conditions laid down by the relevant provisions of the Sixth Directive and the national legislation transposing it, result in the accrual of a tax advantage the grant of which would be contrary to the purpose of those provisions.

75. Second, it must also be apparent from a number of objective factors that the essential aim of the transactions concerned is to obtain a tax advantage. As the Advocate General observed in point 89 of his Opinion, the prohibition of abuse is not relevant where the economic activity carried out may have some explanation other than the mere attainment of tax advantages. …

84. [In relation to Question 2, which was whether a taxable person had the right to deduct input VAT where the transactions on which that right was based constituted an “abusive practice”] However, as the Court has already had occasion to observe, it is only in the absence of fraud or abuse, and subject to adjustments which may be made in accordance with the conditions laid down in Article 20 of the Sixth Directive, that the right to deduct, once it has arisen, is retained. …

85. Accordingly, the answer to be given to the second question must be that the Sixth Directive must be interpreted as precluding any right of a taxable person to deduct input VAT where the transactions from which that right derives constitute an abusive practice.

86. For it to be found that an abusive practice exists, it is necessary, first, that the transactions concerned, notwithstanding formal application of the conditions laid down by the relevant provisions of the Sixth Directive and of national legislation transposing it, result in the accrual of a tax advantage the grant of which would be contrary to the purpose of those provisions. Second, it must also be apparent from a number of objective factors that the essential aim of the transactions concerned is to obtain a tax advantage”.

60.

Thus, if one looks beyond the examples stated of tax evasion by means of untruthful tax returns or improper invoices (paragraph 59 of Halifax), it seems that the Halifax principle, is not based on the premise that fraudulent transactions are not “economic activity” or a “supply of goods by the taxable person acting as such”. Instead, even a transaction carried out with the sole aim of obtaining a tax advantage will constitute a supply of goods and an economic activity provided it satisfies the objective criteria on which the concepts of a supply of goods and economic activity are based. The Halifax principle (or at least the more central Halifax principle) is, rather, that arising from the CJEU’s answer to the second question it was asked, namely that the right of a taxable person to deduct input VAT is precluded where the transactions from which that right derives constitute an abusive practice.

61.

Guidance as to the nature of an “abusive practice” is available not only from paragraphs 74 and 75 of the CJEU’s judgment in Halifax (set out above), but also from paragraphs 84-91 of Advocate General Poiares Maduro’s opinion in Halifax, and from Lord Sumption’s judgment in Commissioners for HMRC v. Pendragon plc [2015] UKSC 37 (“Pendragon”) at paragraphs 4-13. As Lord Sumption explains at paragraphs 5-6, “abuse of law” is a principle derived from civilian jurisprudence, perhaps better described by the French term “détournement de droit”, which has become part of the law of the European Union and thence part of the law of this country. Since Halifax, the CJEU has generally determined whether a scheme is abusive by asking whether the relevant transactions have any independent commercial purpose beyond tax evasion.

62.

Paragraph 53 of the CJEU’s decision in Kittel, where it said that “the objective criteria which form the basis of the concepts of ‘supply of goods effected by a taxable person acting as such’ and ‘economic activity’ are not met where tax is evaded by the taxable person himself (see Halifax … para 59)” may justify the UT’s summary at paragraph 53 to the effect that “if the taxpayer is himself seeking to evade tax, his fraudulent transaction does not meet the objective criteria to qualify as a supply of goods and he has no right to deduct tax (Halifax)”. The UT’s summary could also be supported by the later formulation of the CJEU in paragraphs 37 and 38 of Bonik EOOD v. Direktor na Direktsia ‘Obzhalvane i upravlenie na ispalnenieto’ - Varna pri Tsentralno upravlenie na Natsionalnata agentsia za prihodite (case C- 285/11) [2013] STC 773 to the effect that “national courts [should] refuse the right of deduction, if it is shown, in the light of objective factors, that that right is being relied on for fraudulent or abusive ends”, and “[t]hat is the position where a tax fraud is committed by the taxable person himself. In such a case, the objective criteria which form the basis of the concepts of ‘supply of goods or services effected by a taxable person acting as such’ and ‘economic activity’ are not met”.

63.

These dicta do not, however, tell the whole story, because, as I have said, the underlying EU law principle to be extracted from Halifax is that the right of a taxable person to deduct input VAT is precluded where the transactions from which that right derives constitute an abusive practice. As it seems to me, the two limbs of Kittel are derived from, and provide a workable approach to, the operation of that underlying principle.

64.

The underlying ‘abuse of law’ approach also, I think, supports the argument that there can be knowledge, without dishonesty, under the first limb of Kittel. The taxpayer’s abuse of its prima facie legal rights would obviously not involve dishonesty in every case: cf. the way Lord Sumption approached the question of “concurrent purposes” in paragraphs 12ff of Pendragon.

The first ground of appeal

65.

With that introduction, I turn to deal with HMRC’s first ground of appeal in each of the two cases, namely that the UT was wrong to conclude that the allegation that Citibank and E Buyer knew that there was an overall scheme to defraud HMRC and that its transactions were part of that scheme required a plea of dishonesty.

66.

It is first useful to remind ourselves of Lord Carnwath’s comments in paragraphs 50 and 51 in Pendragon as to how the Court of Appeal should approach an appeal of this kind:-

“50. The difficult concept of ‘abuse of law’ as developed by the European court, though not strictly one of statutory construction, is a general principle of central importance to the operation of the VAT scheme … it was clearly one which was particularly well suited to detailed consideration by the Upper Tribunal, with a view to giving guidance for future cases. Having found errors of approach in the consideration by the First Tier Tribunal, it was appropriate for them to exercise their power to remake the decision, making such factual and legal judgments as were necessary for the purpose …

51. Against this background, it was unhelpful, in my view for the Court of Appeal to identify the main issue as to whether the Upper Tribunal went beyond its proper appellate role. The appeal to the Court of Appeal (under section 13) was from the decision of the Upper Tribunal, not from the First Tier, and their function was to determine whether the Upper Tribunal had erred in law. That was best approached by looking primarily at the merits of the Upper Tribunal’s reasoning in its own terms, rather than by reference to their evaluation of the First Tier’s decision. True it is that the Upper Tribunal’s jurisdiction to intervene had to begin from a finding of an error of “law”. But that was not the main issue in the appeal, which was one of more general principle. … it may not be productive for the higher courts to spend time inquiring whether a difference between the two tribunals was one of law or fact, or a mixture of the two … The Upper Tribunal reached a carefully reasoned conclusion on law and fact. The task of the Court of Appeal was to determine whether that conclusion disclosed any error of law”.

67.

In this case, it will be remembered, we are dealing with a case in which the UT reversed one decision of the FTT (E Buyer) and upheld another (Citibank). The order of the formal questions that must be answered, therefore, are as follows:-

i)

Was the UT right to hold that there was no error of law in Judge Mosedale’s decision in the Citibank case?

ii)

Was the UT right to hold that there were errors of law in Judge Walters’ decision in the E Buyer case?

iii)

If the UT was wrong as a matter of law in either case, then the question is whether this court should remake the decision that was in error itself or send the case back to the UT or the FTT (bearing in mind that the UT actually remitted part of the E Buyer case to the FTT for the decision to be remade on the grounds that detailed arguments as to the pleadings had not been addressed to them) to remake that decision?

68.

As I have already said, the UT did not expressly conclude that the allegation of actual knowledge of the dishonest scheme and that the appellants’ transactions were part of that scheme did require a plea of dishonesty. Instead, as Mr Wardell submitted, in several places, the UT said the reverse to the effect that the allegation of actual knowledge of these matters could be made without alleging dishonesty (see particularly paragraphs 58, 60, and 86(5)). It would have been different, he submitted, if the Grounds of Appeal had criticised the UT for concluding that the allegations of actual knowledge of the dishonest scheme and that its transactions were part of that scheme amounted to a plea of dishonesty. Had that criticism been made, it would have been, in effect, a criticism of the UT’s interpretation of the pleadings, which is what Mr Wardell submitted he had been complaining about all along. All he wanted and all he had ever asked for on E Buyer’s behalf, he argued, was full and proper particulars of the allegation that E Buyer had the knowledge alleged.

69.

The problem, in my judgment, with the UT’s decision is that, whilst it does indeed say that the pleaded knowledge will not always require a pleading of dishonesty, in other parts of its decision it says something quite different. The UT clearly implies in paragraphs 61, 68, 72 and 102 that it thought that the allegation of actual knowledge of these matters automatically involved an allegation of fraud or dishonesty against the taxpayer.

70.

In paragraph 61, the UT said in the same breath that Kittel was “plainly not an honesty test” and that the “real question in this case” was whether a given allegation of knowledge is “tantamount” to an allegation of dishonesty. In paragraph 68 the UT effectively endorsed Briggs J’s dictum in Megtian to the effect that “a person who knows that a transaction in which he participates is connected with fraudulent tax evasion is a participant in that fraud [and] has a dishonest state of mind”. Likewise, in paragraph 72 (again by reference to Megtian) the UT drew a stark distinction between knowledge and constructive knowledge and said that to allege knowledge in the terms HMRC did in this case was “in effect” to allege a dishonest state of mind. In paragraphs 101 and 102, the UT held that an allegation of knowledge of connection to a fraudulent scheme was an allegation of dishonesty.

Citibank

71.

Against this background, the first question is whether the UT was right to hold that Judge Mosedale had made no error of law. At paragraph 94, the UT said that it agreed with Judge Mosedale at paragraphs 23-25 when she said that (a) if an appellant is shown to know in advance that its purchase and sale were orchestrated by a third party in order to perpetrate a fraud on HMRC, its decision to proceed with the transactions knowing this would be dishonest, and (b) HMRC were seeking to prove that Citibank acted in a dishonest fashion by saying that it went ahead with transactions which it knew were orchestrated for the purpose of fraud.

72.

At paragraph 96, the UT upheld Judge Mosedale’s decision in paragraph 40 that (1) HMRC had to amend its statement of case to make it clear whether or not it was alleging a dishonest state of mind against Citibank, and (2) if it did not do so, it would not be permitted to seek findings that Citibank knew that its transactions (a) were contrived, (b) facilitated fraud by others, or (c) were connected to fraud. The UT dismissed the appeal from Judge Mosedale on that basis.

73.

By the end of the argument, it was accepted on all sides that paragraph 40 of Judge Mosedale’s decision was wrong, at least in part, in that she had put the principle too high, and had elided the distinction between knowledge and dishonesty.

74.

I agree that paragraph 40 of Judge Mosedale’s decision was wrong. There was a discussion in argument about whether she may have been right to say that that HMRC could not allege that Citibank knew that its own transactions were contrived without alleging dishonesty. I am not sure that this is any more correct than Judge Mosedale’s other propositions in paragraph 40, but that may not be central to our decision.

75.

More importantly, it was on the basis of her decision in paragraph 40 that Judge Mosedale required HMRC to amend their Statement of Case to plead allegations of dishonesty if it wished to allege any of the three aspects of Citibank’s knowledge (which they have now done). It follows that, in my judgment, the UT was also wrong to uphold the order she made based on her conclusion in paragraph 40.

76.

Judge Mosedale was also wrong, I think, in paragraphs 23-25 of her decision to find that the allegations at paragraph 49 of HMRC’s original Statement of Case against Citibank necessarily involved an allegation of dishonesty. Those allegations were classic first limb Kittel allegations of actual knowledge. The allegations are not pleadings of fact, but inferences from facts. The facts alleged are (1) that Citibank’s transactions formed part of an overall scheme to defraud HMRC, (2) that the scheme involved an orchestrated and contrived series of transactions. The inference that HMRC contend for is that certain features of those transactions demonstrate that Citibank knew or ought to have known that this was the case. It follows that, in my judgment, that inference was not an allegation of dishonesty, but simply of knowledge.

77.

Judge Mosedale’s conclusions in her paragraphs 24 and 25 were, in effect, very similar to each other. They were that HMRC was seeking to prove that Citibank was dishonest when it went ahead with its purchase and sale knowing they were orchestrated by a third party in order to perpetrate a fraud on HMRC. But that is not what HMRC were trying to prove. HMRC were trying to prove what they pleaded at paragraph 96 of their Statement of Case to the effect that “the only reasonable explanation for Citibank’s behaviour was that it knew or should have known of [the connection between its transactions and] VAT fraud”.

78.

Finally in this regard, I should say something about what Judge Mosedale said at paragraphs 31 and 32 of her decision. She concluded there that there was nothing in Mobilx supra which cast any doubt on what Briggs J had said to the effect that “[a] person who knows that a transaction in which he participates is connected with fraudulent tax evasion is a participant in that fraud” and “has a dishonest state of mind”, so that what he (Briggs J) had said must be right in law. For the reasons I have already given, I do not think that such a bald proposition is right in law, because, as is acknowledged by all parties to this case, a person who knows that a transaction in which he participates is connected with fraudulent tax evasion may or may not have a dishonest state of mind.

79.

There was, in my judgment, on the basis of these allegations, no foundation for either the FTT or the UT to have required HMRC to plead dishonesty against Citibank in order to be allowed to allege that Citibank knew that its transactions (a) were contrived, (b) facilitated fraud by others, or (c) were connected to fraud. The UT was therefore wrong not to have identified these clear errors of law in Judge Mosedale’s decision.

80.

Some exiguous argument was addressed to other aspects of HMRC’s Statement of Case against Citibank. But in my judgment, it is not appropriate for this court to undertake a de novo review of the pleadings. Judge Mosedale and the UT decided the case on the basis of the paragraphs in the Statement of Case to which I have referred. This court should confine itself to identifying the errors of law in the tribunals below, and not seek to decide the facts again unless it is remaking the decision – a matter to which I shall turn in due course.

E Buyer

81.

Substantially the same allegations were made in HMRC’s Statement of Case against E Buyer as are made in their original Statement of Case against Citibank. It was the allegations in paragraphs 38 and 79 of HMRC’s then Statement of Case (numbered 38 and 80 cited above) which Judge Walters set out at paragraphs 9 and 10 of his decision and on which he essentially relied in paragraphs 34-39 of his reasoning. He cited other paragraphs, but those were the essential allegations of knowledge that E Buyer’s transactions were connected with fraud.

82.

It follows, therefore, that there was no basis on which the UT could have held that Judge Walters had been wrong in law to dismiss E Buyers’ claim to be provided with further particulars of the alleged dishonesty. In my judgment, the UT was wrong at paragraph 101 of its decision to apply the reasoning from its decision in relation to Citibank to conclude that paragraph 38 of HMRC’s Statement of Case against E Buyer “amount[s] to an allegation of dishonest conduct against E Buyer”. Paragraph 102 of the UT’s decision was also wrong to hold that these allegations were allegations of dishonesty. In these circumstances, the UT ought to have held that there was no error of law in Judge Walters’ decision to refuse E Buyer’s request for further particulars of these allegations, which was made on the premise that HMRC were alleging that E Buyer was a co-conspirator in the fraud (see paragraph 35 of Judge Walters’s decision).

83.

Again, some argument was addressed to other aspects of HMRC’s Statement of Case against E Buyer. But, again, it is not appropriate, in my view, for this court to undertake a de novo review of the pleadings. The central attack on Judge Walters’ decision in the UT was that the FTT “should have accepted the main thrust of [E Buyer’s] application, namely that (a) the case advanced by [HMRC] was akin to a civil fraud case …”. That attack ought to have failed in the UT for the reasons I have given. This court should, as I have said, confine itself to identifying the errors of law in the UT.

The law as stated by the UT and the consequences

84.

All that I have said does not wholly invalidate the UT’s summary of the law at paragraph 86. It was largely accurate, although I would not have framed it as the UT did. The UT made repeated reference to the possibility that the taxpayer might be guilty of dishonesty and fraud, when all are agreed that any dishonesty and fraud by the taxpayer forms no part of the Kittel test under either the first or the second limb.

85.

The key point, in my judgment, is that, whilst HMRC can, of course, allege that a taxpayer has acted dishonestly and fraudulently in relation to the transactions to which it was a party, they do not need to do so in order to deny that taxpayer the right to reclaim input tax under the Kittel test. The exercise upon which Judge Mosedale was engaged was, therefore, inappropriate. It was simply irrelevant for the FTT to ask whether the allegations in the Statement of Case, if all proved, would necessarily lead to the conclusion that the taxpayer had been dishonest or fraudulent. It was even more inappropriate for Judge Mosedale to direct HMRC to plead dishonesty when it had expressly informed her that it did not wish to make any such allegation. It might be, of course, that if some or all of the allegations made in the Statement of Case were proved, that might (in theory, though not, of course, in practice) have allowed a tribunal to go on to make a finding that the taxpayer had been dishonest. But if HMRC does not seek such a finding, and if such a finding is not needed to support the conclusion that the taxpayer cannot recover its input tax, there is neither any need nor any utility in asking the FTT to undertake that exercise.

86.

The UT’s concern was that, because the allegations made in the Statement of Case, might be sufficient to support a plea of dishonesty, further particulars of that plea were required. But as I have sought to show by reference to what Judge Mosedale actually found in the Citibank case, the core pleading in that case did not automatically mean that it was alleged that Citibank behaved dishonestly. As the UT said, such conduct might or might not, depending on a number of factors, be conduct that would be regarded as dishonest.

87.

In these circumstances, in my judgment, the UT was wrong to imply that the FTT was justified in undertaking the task of seeking to ascertain from the Statement of Case whether or not the conduct alleged automatically amounted to dishonesty or fraud. Such a process was unnecessary and inappropriate and particularly so when HMRC disavowed an allegation of that character.

88.

None of what I have said should be construed as suggesting that properly informative particulars of the allegation of both actual and constructive knowledge need not be given. Of course, they must. This court cannot, however, trawl through the lengthy requests for particulars that have been made in either case to see whether they are appropriate or not. All we need to do is to say that Judge Mosedale was wrong to require HMRC to plead dishonesty against Citibank, and that the UT was wrong to uphold her decision. I would remit the Citibank case to the FTT for it to decide what particulars of knowledge, if any, still need to be provided, once HMRC’s amended pleading alleging dishonesty has been withdrawn.

89.

As regards the E Buyer case, as I have said, there was no error of law in Judge Walters’ decision. I would, therefore, allow the appeal from the UT’s decision, insofar as it remitted the case to the FTT to decide what further and better particulars ought to be provided by HMRC. I would reinstate Judge Walters’ order in that regard.

90.

Finally, if a summary of the applicable law is required along the lines of paragraphs 86 and 87 of the UT’s decision, I would simply summarise the principles as follows:-

i)

The test promulgated by the CJEU in Kittel was whether the taxpayer knew or should have known that he was taking part in a transaction connected with fraudulent evasion of VAT.

ii)

Ultimately the question in every Kittel case is whether HMRC has established that the test has been met. The test is to be applied in accordance with the guidance given by the Court of Appeal in Mobilx and Fonecomp.

iii)

It is not relevant for the FTT to determine whether the conduct alleged by HMRC might amount to dishonesty or fraud by the taxpayer, unless dishonesty or fraud is expressly alleged by HMRC against the taxpayer. If it is, then that dishonesty or fraud must be pleaded, particularised and proved in the same way as it would have to be in civil proceedings in the High Court.

iv)

In all Kittel cases, HMRC must give properly informative particulars of the allegations of both actual and constructive knowledge by the taxpayer.

91.

Moreover, I do not agree with the suggestion that the UT made in paragraph 87 of its decision as to there being a false dichotomy between (i) a case where HMRC alleged that the taxpayer was himself party to a fraud and dishonest under the Halifax principle and (ii) a case where HMRC chose to make no such allegation but to proceed under the first limb of the Kittel principle. As I have already pointed out, it is true that at paragraph 53 of Kittel, the CJEU echoed the first sentence of paragraph 59 of Halifax to the effect that the “objective criteria which form the basis of the concepts of ‘supply of goods effected by a taxable person acting as such’ and ‘economic activity’ are not met where tax is evaded by the taxable person himself”. But in Halifax, the denial of the right to deduct input tax was based upon the fact that the “transactions from which that right derives constitute an abusive practice”. Kittel decided that a taxable person who knew or should have known that, by his purchase, he was taking part in a transaction connected with fraudulent evasion of VAT should be regarded as a participant in that fraud, irrespective of whether or not he profited by the resale of the goods, because the taxable person aids the perpetrators and becomes their accomplice. The national court in that situation must refuse the right to deduct input tax “even where the transaction in question meets the objective criteria which form the basis of the concepts of ‘supply of goods effected by a taxable person acting as such’ and ‘economic activity’” (paragraph 59 of Kittel). In other words, the basis of the requirement to deny input tax must be the abuse of right or abusive practice adumbrated by the CJEU in answer to the second question in Halifax.

92.

There is therefore, perhaps, a dichotomy, but not a false one. Even if the transaction is to be regarded as an economic activity and, therefore, prima facie subject to the deduction of input tax, that tax will not be deductible if either the transactions from which the right derives constitute an abusive practice (Halifax), or if the taxable person knew or should have known that, by his purchase, he was taking part in a transaction connected with fraudulent evasion of VAT (Kittel). The CJEU may have explained the rationale for the second (Kittel) part of that statement of the position as being that the taxpayer is to be regarded as an accomplice of the perpetrators of the fraud. That explanation does not, however, undermine the facts that (a) the principle underlying the deduction of input tax is the question of whether there is an abusive practice, and (b) proof of dishonesty is not a necessary prerequisite for deduction of input tax to be disallowed.

Disclosure in the E Buyer case

93.

I can deal with the disclosure appeal more shortly. HMRC appeal on the basis that the UT had failed to identify any error of law made Judge Walters, and failed to recognise the wide case management discretion of the FTT. Judge Walters thought that he should apply rule 27 of the 2009 Rules and can hardly be faulted for that. The UT’s complaint was that Judge Walters did not address the specific points made by E Buyer as to why it should have wider disclosure. The problem, however, is that the UT thought that the reasons for giving wider disclosure were that (a) HMRC had the burden of proof and held all the cards from their extensive investigation, and (b) it was, therefore, appropriate for HMRC to disclose not only the material that they wish to rely on, but also any other material which might undermine that case, and (c) there was little doubt that this case, in which dishonesty was alleged against E Buyer and which involved some £7 million in tax, was complex and of great importance to the taxpayer.

94.

In my judgment, the third reason must have been the most important in the UT’s decision to displace rule 27 and order CPR-style disclosure. But the UT was wrong to think that dishonesty was alleged against E Buyer as I have already explained. In these circumstances, I cannot see how Judge Walters’ simple reasoning can be faulted. It is true that this is an important case, but the 2009 Rules were made for important as well as simple cases. The plain fact is that the procedure is different in the FTT. If fraud or dishonesty had been alleged, there is, as I have said, authority for the proposition that CPR-style disclosure would have been appropriate, but all that is alleged here is knowledge of a fraud, not direct dishonest participation in a fraud.

95.

In these circumstances, the UT’s decision on the disclosure question cannot, in my judgment, stand. The appeal on this point should be allowed and Judge Walters’ decision should be reinstated.

Conclusions

96.

For the reasons I have tried shortly to state, I take the clear view that this appeal must be allowed.

97.

It will by now be obvious that I agree with HMRC’s submission as to the fundamental issue that is required to be resolved by this appeal. HMRC said that the question was whether the UT was wrong to conclude that an allegation that a taxpayer knew that its transactions were part of an orchestrated scheme to defraud HMRC required HMRC to plead and particularise, and therefore to prove, an allegation of dishonesty. I agree that that was the question raised by this appeal. I also agree that the allegation, which is a classic Kittel first limb contention, does not require HMRC to plead, particularise and prove dishonesty or fraud.

98.

The main point in this case was not, as the taxpayers suggested a simple pleading question. The UT failed, I think, to identify the basic error that Judge Mosedale had made in the Citibank case, where she said, in effect, that making a first limb Kittel allegation required a plea of dishonesty. It does not; even if in some cases, the findings of knowledge made by the FTT could have led the FTT to uphold a plea of dishonesty had it been made. HMRC is entitled to stop short of alleging dishonesty and content itself with pleading, particularising and proving first limb Kittel knowledge. If, however, HMRC do expressly allege dishonesty, they will be required to comply with the normal rules of pleading and disclosure applicable to such cases. In future, it might be helpful in these cases for HMRC to say expressly in their Statements of Case whether or not they set out to prove the dishonesty of the appellant taxpayer.

99.

I would therefore allow these appeals, and remit the Citibank case to the FTT for reconsideration, in the light of the judgments of this court, of whether the further and better particulars of the original Statement of Case requested by Citibank should be provided. I would reinstate Judge Walters’ orders in the E Buyer case.

Lady Justice Hallett D.B.E:

100.

I am indebted to the Chancellor for the depth and clarity of his analysis. As a result, I can keep my observations relatively short. Given the focus in these appeals on the issue of dishonesty and a trader’s knowledge of a connection between its transaction and an MTIC fraud, I begin by considering the definitions of an MTIC fraud and of dishonesty and the requirements attached to a plea of dishonesty.

101.

At its most basic level, an MTIC fraud involves a trader importing goods VAT free, selling them on for a sum including VAT and then absconding without paying the VAT to the relevant authority. It is rare for such a fraud to consist of only one trade. Most, if not all, MTIC frauds (also known as carousel frauds) involve a chain of trades in which the same goods are repeatedly exported and imported and VAT charged each time but not necessarily accounted for. Many of the trades are contrived, in the sense they serve no purpose other than to distance the fraudsters from the fraud and/or to provide them with additional dishonest gain. The definition of an MTIC fraud may be simple to state, but MTIC frauds require considerable planning and sophistication and are usually perpetrated (or orchestrated) by organised gangs of professional criminals. Most MTIC frauds are therefore based on a ‘scheme designed to defraud HMRC involving an orchestrated and contrived series of transactions’.

102.

Reputable companies may be caught up in an MTIC fraud unknowingly. The fact that there is fraud somewhere in the chain of trades does not necessarily disentitle the trader to claim a return of the input VAT they have paid. For HMRC to deny a claim, it will have to prove that the trader knew or should have known that the transaction in which it was involved was connected with a scheme for the fraudulent evasion of VAT (the Kittel test).

103.

It is common ground that HMRC does not need to allege or plead dishonesty in order to deny the trader its claims to repayment of input VAT. The Kittel test does not require proof of dishonesty. However, if it does allege dishonesty, HMRC is obliged to plead the facts, matters and circumstances relied upon to show that the trader is dishonest. This is to ensure the trader knows in advance the case it must meet and to ensure a court or tribunal does not make a finding of dishonesty, with all the serious consequences that such a finding entails, on an inappropriate basis. Where serious allegations of fraud are made, cogent evidence commensurate with the gravity of the allegations made must be adduced.

104.

Turning to the definition of dishonesty, this has proved controversial in both the criminal and civil law. The Court of Appeal Criminal Division in R v Ghosh [1982] 1 QB 1053 attempted to reconcile a line of authorities that had provided different definitions of dishonesty for different criminal offences. It concluded the authorities were irreconcilable and provided a two stage objective and subjective test for dishonesty namely:

(i) Was the conduct dishonest by the ordinary standards of reasonable and honest people?

(ii) Must the defendant have realised that what he/she was doing was, by those standards, dishonest?

However, the second, subjective part of the test has itself been the subject of significant criticism.

105.

Similarly, the issue of whether it is necessary to prove both a subjective and objective element to dishonesty in civil claims, and if so, in which civil claims, has been the subject of some debate; see for example, Royal Brunei Airlines v Tan [1995] 2 AC 378, Twinsectra Ltd v Yardley [2002] 2 AC 164 and Barlow Clowes International Ltd and Another v Eurotrust International Ltd and others [2005] 1 WLR 1476. A close analysis of those decisions and those that followed is not necessary for present purposes. Suffice it to say that not only must a claim of dishonesty be properly pleaded and appropriately cogent evidence advanced, proof of a claim of dishonesty requires, at the very least, proof of an objective element. The allegedly dishonest person in civil proceedings must have knowledge of the transaction such as to “render his participation contrary to normally acceptable standards of honest conduct” (per Lord Hoffman giving the opinion of the Privy Council at paragraph 15 of Barlow Clowes). Thus, even if HMRC would not be required to prove a subjective element, of the kind set out in Ghosh, to prove dishonesty against the respondents, (and the skeleton argument served on behalf of E Buyer suggests this remains a live issue), HMRC would still have to prove the additional objective element.

106.

It follows that an order requiring HMRC to plead dishonesty, on the basis it alleges actual knowledge of participation in a fraud, would have the effect of significantly and unnecessarily raising the bar, in terms of what it must prove to deny the respondents’ claims and the cogency of the evidence called.

107.

I recognise that proof of participation in an MTIC fraud with actual knowledge of the fraud may be powerful evidence of conduct contrary to normally acceptable standards of honest conduct, and dishonesty in that objective sense. It may also provide powerful evidence of dishonesty in the subjective sense, if that additional element is required (as E Buyer appears to maintain). The line between honest conduct and dishonest conduct may be a fine one, in such circumstances. Nonetheless, there is a line and entering into a transaction knowing that it is connected with fraud does not necessarily equate to dishonest conduct in either the objective or the subjective sense.

108.

I understood this to be accepted by the respondents, who expressly disavowed an intention to equate, in every case, an allegation of actual knowledge that a transaction was connected with fraud with an allegation of dishonesty. I was, therefore, puzzled by reliance on paragraph 41 of the judgment in Megtian (supra) in apparent support of the proposition that a “person who knows that a transaction in which he participates is connected with fraudulent tax evasion is a participant in the fraud” and has a dishonest state of mind. I do not believe that Briggs J (as he then was) intended to lay down any general proposition of law to that effect. If he did, I would respectfully disagree, because to do so would be to import the concept of dishonesty into every case in which actual knowledge is alleged, under the first limb of the Kittel test, and would not take account of the additional requirements attached to a plea and finding of dishonesty.

109.

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

110.

It is in that context, I have approached the judgments of the FTT and the UT and the SOCs against both Citibank and E Buyer. In both SOCs, HMRC alleges three things: (i) that there was an overall scheme involving an orchestrated and contrived series of transactions to defraud HMRC (ii) that the respondent trader’s transactions formed part of that overall scheme and (iii) that the respondent trader knew or ought to have known that this was the case. This amounts to an allegation of a classic MTIC fraud (perpetrated by someone) and actual or constructive knowledge on the part of the trader that its transactions formed part of the overall fraudulent scheme.

111.

If, contrary to my view already expressed, it is necessary to import the concept of dishonesty into such allegations, more would be required to justify the assertion that these allegations are ‘tantamount’ to allegations of dishonesty. It is not clear to me from the UT’s judgment why, having acknowledged that not every case of alleged actual knowledge of an MTIC fraud will involve an allegation of dishonesty, the UT concluded that in these two cases it did. In paragraphs 92, 93 and 101 of their judgment, the UT appear to have placed considerable reliance on HMRC’s use of the words “orchestrated and contrived” to describe the scheme in each case, but, as I have endeavoured to explain, most MTIC frauds are by their very nature “orchestrated and contrived”.

112.

I respectfully agree with the Chancellor, therefore, that although the UT expressly acknowledged the principle that not every case of alleged actual knowledge pursuant to the first limb of the Kittel test will involve an allegation of dishonesty, it has fallen into the same error as Judge Mosedale, and assumed that it does. To my mind, paragraph 93 of the judgment (in relation to Citibank) makes that clear:

“We do not see how a trader in the position of CGM could honestly proceed with a transaction if it knew that the transaction was part of an orchestrated and contrived scheme to defraud the Revenue”.

113.

The UT’s finding that Judge Walters QC had erred was based on precisely the same premise. This is clear from paragraph 101 of their judgment:

“… HMRC here allege that E Buyer actually knew both that there was an overall scheme, involving an orchestrated and contrived series of transactions, to defraud the Revenue and that its transactions were part of this scheme. For reasons given above to the Citibank appeal we consider that this does amount to an allegation of dishonest conduct against E Buyer.”

114.

It was on the basis that dishonesty is alleged, that the UT further concluded that E Buyer is entitled to the particulars demanded. I disagree with the UT in both respects. For my part, I am satisfied Judge Walters was right to reject E Buyer’s argument that the allegations against it amount to an allegation of dishonesty and that E Buyer requires further particulars of the alleged dishonesty. He was right to focus, as he did, on what was alleged (actual or constructive knowledge of the fraud) and on the question of whether any additional clarity or particularity of actual or constructive knowledge is required at this stage of the litigation.

115.

It follows that I agree with the Chancellor that Judge Mosedale was wrong to order as she did and the UT wrong to uphold her order, that Judge Walters QC made no error of law and the UT was wrong to find that he had.

116.

I too can deal with the Disclosure ground shortly. The UT overturned Judge Walters on this point because they believed this was a case in which normal disclosure pursuant to rule 27 did not apply. They based their decision in large part on their finding that HMRC’s allegations were tantamount to an allegation of dishonesty. Their decision was therefore flawed in at least one significant respect. Furthermore, I do not accept that Judge Walters refused the application for broader disclosure on a rigid application of rule 27, as claimed. It is clear from his judgment as a whole, delivered after a two day hotly contested hearing, that he was acutely conscious of the seriousness and complexities of the case. He acknowledged that rule 27 provides for normal disclosure in standard and complex cases and was satisfied there was no need to depart from the usual procedure, “at this stage in this litigation. This was a case management decision well open to him and one with which an appellate tribunal should be slow to interfere. I doubt that the UT would have done so but for their findings in respect of dishonesty.

117.

For those reasons and the reasons given by the Chancellor, I would allow the appeals on the one ground in the case of Citibank and on both grounds in the case of E Buyer. I am content that the Citibank case should be remitted to the FTT for re-consideration of the request for particulars if it is pursued (albeit I was far from convinced of the need for them) and I would reinstate the decision of Judge Walters QC.

Sir Terence Etherton, Master of the Rolls:

118.

These successive appeals to the FTT, the UT and then to the Court of Appeal have acquired a quite unwarranted complexity. The principal issue in each appeal is whether or not the allegations made by HMRC in its original statements of case for denying the taxpayers the right to reclaim input tax were sufficiently particularised. Particulars had to be sufficient to enable the taxpayers fully to understand the case against them and to enable the FTT to achieve the overriding objective of dealing with the cases fairly and justly (in accordance with the Tribunal Procedure (First-tier Tribunal) (Tax Chamber) Rules 2009 rule 2).

119.

The complexity has arisen through entanglement over the extent to which the first limb of the Kittel test necessarily or ordinarily will involve an allegation of dishonesty and, if so, the implications of that on the extent to which particulars must be given of that dishonesty. The issue of dishonesty in this context is, however, a digression from the critical issue, which is one of actual or constructive knowledge, namely that the trader will not be able to reclaim input tax if it knew or should have known that the transaction in which it was involved was connected with a scheme for the fraudulent evasion of VAT.

120.

It was common ground before the UT and is common ground before us that the Kittel test does not require dishonesty. Unless dishonesty is expressly alleged, the only question is whether the pleaded allegations are relevant to the issue of actual or constructive knowledge for the purposes of the Kittel test: if not, they are irrelevant and should be struck out or at least ignored, and, if they are, the only question is whether they are sufficiently particularised to enable the taxpayers fully to understand the case against them and to enable the FTT to achieve the overriding objective of dealing with the case fairly and justly. For that reason, it is entirely irrelevant whether dishonesty is implicitly alleged in HMRC’s statements of case.

121.

It follows that, in respect of the case against Citibank, Judge Mosedale was wrong to hold in paragraph [40] of her decision that, unless HMRC amended its statement of case to make it clear whether or not it was alleging a dishonest state of mind against Citibank, HMRC would not be entitled to advance a case that Citibank’s transactions were contrived or that Citibank knew its transactions were facilitated by fraud by others or that its transactions were connected to fraud. The UT was wrong to hold in paragraph [96] of its decision that there was nothing wrong with Judge Mosedale’s direction to that effect. It should have allowed the appeal against her direction.

122.

The Chancellor and Hallett LJ are of the view that Judge Mosedale was wrong to conclude that the paragraphs in HMRC’s original statement of case cited by her amounted to an implicit allegation of dishonesty. I do not find that point as straightforward as the Chancellor and Hallett LJ appear to consider, particularly if regard is had to the statement of case as a whole. I see no point, however, in conducting a detailed analysis of HMRC’s statement of case against Citibank on this issue since this is not the occasion for a detailed analysis of the ingredients of dishonesty in the civil law, there is a majority view in this court on what is HMRC’s pleaded case, and in any event, as I have said, implicit allegations of dishonesty are beside the point in applying the Kittel test.

123.

I agree, therefore, that the appeal in the Citibank case should be allowed, and that the case should be remitted to the FTT for reconsideration of whether further particulars relevant to knowledge should be provided by HMRC.

124.

I take the same approach on the appeal in the case against E Buyer. The Chancellor and Hallett LJ are of the view that Judge Walters QC was correct to conclude that the paragraphs in the statement of case cited by him did not contain an implicit allegation of dishonesty. Once again, I do not find that that point so straightforward, particularly if regard is had to the statement of case as a whole. For the reasons that I gave in relation to the Citibank case, however, I see no point in conducting a detailed analysis of HMRC’s statement of case against E Buyer on this issue.

125.

It follows that it cannot be said that Judge Walters acted outside the bounds of a valid exercise of judicial discretion in making the case management decision to refuse a special order for disclosure.

126.

Accordingly, I agree that the appeal in the E Buyer case should be allowed.

HM Revenue and Customs v Citibank NA & Anor

[2017] EWCA Civ 1416

Download options

Download this judgment as a PDF (775.0 KB)

The original format of the judgment as handed down by the court, for printing and downloading.

Download this judgment as XML

The judgment in machine-readable LegalDocML format for developers, data scientists and researchers.