Ammanford Recycling Limited v The Commissioners for HMRC

The main point Ammanford argues under this ground is that the FTT erred in agreeing that HMRC could put to witnesses that they actually knew that the transactions were connected with fraud even where its case was that it did not have to identify such a person and did not seek to. That, Mr Watkinson submits, was a clear error of law: HMRC had to first plead the individual had actual knowledge before it could be put to the individual in cross-examination that they had actual knowledge.

In support, Ammanford relies on a number of authorities on the additional pleading and particularisation obligations in cases involving pleading dishonesty (which it submits applies commensurately to the serious accusation that a person had actual knowledge that the transactions were connected to the fraudulent evasion of VAT). It also relies on Ronald Hull for the proposition that once HMRC’s case became that a particular officer of the company had actual knowledge the relevant transactions were connected to fraud which should then be attributed to the corporate appellant, that allegation then had to be pleaded together with primary facts in support of it.

HMRC maintain there is no difficulty with them pleading a case that someone had actual knowledge yet still putting to a particular appellant’s witness that the witness had actual knowledge. In summary, they argue that the pleading that someone had actual knowledge gave sufficient notice that an allegation of actual knowledge would be made to any of the appellant’s witnesses who were speaking to the relevant transactions. HMRC have pleaded primary facts upon which it is alleged that the company, through one or more of its officers, employees and/or agents must have known of connection to fraud. The case that it has to meet will be clear to the company. If it calls witnesses who give evidence of involvement in the relevant business and transactions, those witnesses will be fully aware of allegations made and it would come as no surprise to them if they were to be asked whether they had requisite knowledge. A finding that a specified individual had actual knowledge was, Mr Foulkes argued, a route to establishing that someone had actual knowledge (which was all HMRC need to establish). It did not entail HMRC changing its case. What was important was that the witness was given the opportunity to answer the allegation in cross-examination.

There is much agreement between the parties about some of the basic relevant legal principles concerning the content and function of pleadings, and how those principles apply in a Kittel case (including those the FTT set out - see [12] – [13] above). Although, as established in Citbank /E-buyer CA it is established that an allegation of actual knowledge does not necessarily entail dishonesty, HMRC accept, correctly in our view, Ammanford’s position that the allegation of actual knowledge of connection with fraud is a serious one. There is no dispute also that the burden of proving this allegation is on HMRC. It is also agreed (although it should be noted this was not Ammanford’s position before the FTT – see [15] above) that, as a general principle, HMRC does not have to identify any named individual in order to make out their case on corporate actual knowledge.

Furthermore, it is accepted that, in addressing the various information requests, the FTT was correct to take stock, not only of the pleadings (i.e. HMRC’s Statement of Case) but the evidence HMRC had served in assessing particularity (in line with McPhilemy) (although Ammanford’s position is that on the particular facts of this case HMRC’s evidence does not add anything to the analysis).

In addressing this ground, it is important, we consider, to keep clear the two distinct issues concerning pleading which arose before the FTT and before us. The first is what HMRC need to plead, at a minimum, in order to show that Ammanford, a corporate body, had the requisite actual knowledge. This was the issue raised by the information request the FTT determined in HMRC’s favour.

The second issue is what consequence flows from HMRC putting its case in that way in terms of the further conduct of the case. This is the issue Ammanford principally raises in this appeal; the specific issue being whether HMRC have to amend their pleading if they want to put to a witness of the appellant in cross-examination that that witness had actual knowledge. (Mr Watkinson emphasised that the appellant was not saying HMRC were closed off from asking such questions of the witness, but they needed to seek and obtain permission to amend their pleadings before they could do it.)

Regarding the Upper Tribunal’s jurisdiction on appeal, Mr Watkinson rightly identifies that the FTT’s decision was a matter of case management discretion. (As the FTT set out at FTT [50], the application was an exercise of the FTT’s power under Rule 5(3)(d) of the Tribunal Procedure (First-tier Tribunal)(Tax Chamber) Rules 2009 to require a party to provide information, which power was to be exercised in accordance with the overriding objective of dealing with a case fairly and justly.) Mr Watkinson also rightly accepts the well-established proposition that an appellate tribunal will not lightly interfere with such discretion except on limited grounds including that the FTT has made an error of principle. His case however is that the FTT’s agreement with HMRC’s stance represents just such an error of principle. HMRC did not, it appears to us, take issue with the alleged error being capable of being one of principle but argue there is no error. HMRC point out that as well as having the ability to put to the witness that they had actual knowledge, the situation could arise where they would be obliged to do that in meeting their pleaded case that someone acting on behalf of the company had actual knowledge.

As reflected by the detailed submissions the parties made to us, the point is a significant one. Before addressing that there is a preliminary question which arises over whether the FTT actually made a decision on the matter which is said to disclose an error of law, in other words, whether the FTT actually ruled on whether HMRC could put questions to the witnesses for the first time in cross-examination that they had actual knowledge but without having given advance notice of that.

That it did not so rule would not necessarily be a surprising outcome given the arguable premature nature of the point. As noted above, the appellant has not served its evidence yet. It is not known which witnesses will give evidence. It is also not known what questions HMRC will ask; it seems to us all HMRC do at this stage is flag that they want to keep the opportunity open to ask such questions of the appellant’s potential witnesses. The context in which the issue has arisen is also relevant. As clarified at the hearing before us, the genesis of the particular issue regarding what questions might be put later in cross-examination was not raised in the appellant’s information application but was something which arose through the comments HMRC’s counsel made in response to Ammanford’s oral arguments in the course of that hearing. Questions of what could or could not be put to witnesses at a subsequent hearing would normally be a matter of case management discretion by the FTT hearing the substantive appeal.

Before the FTT, the ability to run a “someone knew” case was contested, whereas in this appeal the appellant now accepts HMRC can do that as a matter of principle but argues the consequence of that is that HMRC cannot cross-examine the appellant’s witnesses on the basis the witness had actual knowledge without first amending HMRC’s pleadings to that effect.

Ammanford refers to various paragraphs (FTT [62][63] and [68]) to show that the FTT did make an error in relation to the issue of what cross-examination questions could be put. The first two of these paragraphs do not, however, in our view establish that such a ruling was made.

In contrast, FTT [68] does in terms reject Mr Watkinson’s submission that HMRC must identify the individuals who had knowledge of connection to fraud in order to put that to witnesses. There, the FTT said:

We acknowledge there is some room for debate around whether the FTT was making a specific ruling on the point, when this extract is set in the wider context of the decision (where the main dispute was around whether a “someone knew” type case was possible, and given the original scope of the hearing was to deal with Ammanford’s information request application). However, noting that HMRC did not take any point on the scope of the FTT decision under appeal, we consider the better view is that the FTT did make a ruling to the effect that HMRC did not have to identify individuals who had knowledge of connection to fraud in order to be able to put to a specific witness appearing on behalf of Ammanford that that witness knew of such connection. (Although this looks to be inconsistent with the FTT’s unqualified agreement with Ronald Hull (in particular the extract at [52] below), we note that the main point the FTT took away from Ronald Hull was that HMRC did not need to identify a specific individual with actual knowledge to make out its actual knowledge case Kittel. That was the key point relevant to the FTT’s agreement with Ronald Hull as it explained the FTT’s rejection of Ammanford’s information request.) We are satisfied that Ground 1 of Ammanford’s appeal is therefore properly directed towards an issue which the FTT decided. We therefore proceed to consider the ground’s merits.

In doing so, it is also convenient to note at this point the following aspects of the FTT’s analysis.

When addressing Mr Watkinson’s submission, the FTT went on to set out an extract from Citibank FTT 2014, that included the following:

The FTT explained the Citibank FTT passages (to the extent they thought it were capable of supporting Ammanford’s point) needed to be treated with caution for a variety of reasons:

It is also relevant to note the FTT agreed with the proposition advanced in Ronald Hull that if HMRC’s case became that a specified individual had actual knowledge then that had to be pleaded (Ronald Hull [39]). The FTT here did not therefore rule out the possibility that HMRC’s case could in due course become one which involved identification.

The particular issue raised under Ammanford’s ground before us is in essence whether by cross-examining a witness on the basis the witness had actual knowledge one has to regard HMRC’s case as being that the witness had actual knowledge (with the consequence that such case had to be pleaded).

As to the relevant general legal principles regarding pleadings, and as indicated above, there is agreement between the parties on 1) the general purpose of pleadings (to enable the other party to prepare their case), 2) that allegations of dishonesty and other serious allegations engage particular pleadings obligations, and 3) that allegations of actual knowledge amount to serious allegations.

On that latter point, Mr Watkinson’s skeleton helpfully referred to various authorities for the proposition that serious allegations short of dishonesty still need to be clearly pleaded: principally Lakatamia Shipping Co Ltd v Su & Ors [2021] EWHC 1907 (Comm), per Bryan, J., at [39] – [42]; and Mann J in The Deposit Guarantee Fund for Individuals v Bank Frick & Co AG & Anor [2022] EWHC 2221 (Ch), at [39] – [40]. In the latter case, Mann J said at [40]:

We also did not understand it to be in dispute that allegations of actual knowledge that transactions were connected to fraud were serious allegations and that as such they would require some form of notice. The level of particularity may not be directly equivalent to fraud or dishonesty allegations but it should be commensurate with the seriousness of the allegation. It also appears to us to be accepted by HMRC that where (as in the s69D appeal) serious allegations are made against individuals the primary facts from which the allegation is inferred must be set out. HMRC’s position however, in summary, is that, the specific allegation is included within the general; adequate notice to Ammanford is given in the form of the pleading that someone acting on behalf of the company had actual knowledge. HMRC also emphasise that what is key is that the witness would have chance to answer the allegation in cross-examination.

For the reasons below, and in agreement with Mr Watkinson, we consider that the FTT’s ruling, which had the effect that HMRC could put to Ammanford’s witnesses that they, as an individual, had actual knowledge of connection to fraud, but without HMRC having first identified the person as an individual in respect of whom HMRC were alleging actual knowledge) represented an error of legal principle.

We start by noting that there is now no disagreement between the parties that HMRC are able to run a Kittel case without specifying particular individuals who have actual knowledge. That is consistent with the FTT’s analysis in Tower Bridge, Citibank FTT and Ronald Hull. However, we consider that if HMRC wish to put, in cross-examination, to a witness appearing on behalf of the appellant, that they have actual knowledge the transactions were connected to fraud (with a view, it must be assumed, to inviting the FTT hearing the substantive appeal to make such finding) then HMRC must first have identified those witnesses as ones in respect of whom actual knowledge of fraud was alleged.

First, it is relevant to consider the rationale underpinning why it is sufficient for HMRC to plead that someone, rather than a specific person had actual knowledge. As explained in Ronald Hull and in Tower Bridge (see [15] and [16]) the reasons stem from the difficulties of pinning knowledge in relation to a specific person. It is relevant we think that the examples given in the FTT cases relate to larger organisations. The difficulties are not across the board. That indicates that, in situations where HMRC make a serious allegation against an identified individual, so as to be able to put questions to them in cross-examination, HMRC must be taken to have satisfied themselves that there is some evidential basis for that. In those circumstances the rationale for not requiring identification falls away. In other words, the fact that HMRC can plead their case on a “someone had actual knowledge” basis is effectively a concession insofar as it lowers the bar for what HMRC have to show. But there is no reason to allow reliance on that lower bar if HMRC’s case is that they are able to pin the knowledge on an identified person.

If HMRC intend to invite a finding that a particular person has actual knowledge that, properly understood, is not, as HMRC argue, a route to saying someone had knowledge and therefore the company had knowledge. Instead, it is a rather more direct route to saying the company had actual knowledge. The identification of a specific person effectively by-passes the need to rely on a case only that someone had actual knowledge.

Secondly, even if it is considered that HMRC’s case is not to be regarded as having changed to being that a specific person had actual knowledge, there would still, we consider, be a requirement to plead that the particular individual had actual knowledge. This would be on the basis that alleging a named individual had actual knowledge the transactions were connected to fraud for the purposes of making out a case on Kittel self-evidently amounts to a primary fact which is relevant to the overall allegation on corporate actual knowledge, (whether that is put in terms of the company having actual knowledge or an allegation that someone acting on behalf of the company had actual knowledge).

Thus on either of the above two bases, when it is argued that an identified individual had actual knowledge, that becomes HMRC’s case or a primary fact in relation to at least part of it (in that HMRC would not be precluded from also arguing that some other unidentified individuals also had actual knowledge).

This is consistent with the principle stated by Judge Mosedale in Ronald Hull (at [43]) that:

HMRC argues that Ronald Hull should be distinguished because it was given before the Court of Appeal in Citbank / Ebuyer which, as noted above, confirmed dishonesty was not necessarily entailed in an actual knowledge allegation. We do not think that is a valid basis of distinction. The principle Judge Mosedale stated would apply equally to the serious allegation that a person had actual knowledge the relevant transactions were connected to fraud.

HMRC argue that in any case Judge Mosedale’s view at [43] was wrong in principle. The evidence may change, and the witness’ oral evidence might be different from the statement. It cannot be right, Mr Foulkes argued, that HMRC would, in such circumstances have to apply to amend its pleadings, or else be barred from running a case that the individual had actual knowledge.

We disagree. There would be nothing surprising about imposing the discipline of HMRC having to request permission to amend the pleadings and conversely good reason to conform to that discipline. Pleadings are there so the opposing party knows what case it has to answer. They shape the evidence which is advanced and the scope of disclosure. Having to request permission to amend pleadings will mean any prejudicial impact on the evidence or disclosure that might otherwise have taken place can be aired and resolved fairly and justly.

As to Mr Foulkes’ proposition that the case-law suggests it is enough that the witness can answer questions in cross-examination, that is not reflected by the insistence in several authorities, that Mr Watkinson drew our attention to, that relevant allegations are to both be pleaded and cross-examined. (Although these cases concerned allegations of dishonesty, there is no reason to suppose the same would not apply commensurately to allegations which fell short of dishonesty but which were serious nonetheless.)

So in HMRC v Dempster (t/a Boulevard) [2008] EWHC 63 (Ch) Briggs J as he then was noted at [26] the “cardinal principle …that if serious allegations, in particular allegations of dishonesty are to made against a party who is called as a witness they must be both fairly and squarely pleaded, and fairly and squarely put to the witness in cross-examination” (emphasis added). Similarly he later noted in Abbey Forwarding Ltd (in liquidation) v Hone and others [2010] EWHC 2029 (Ch) at [47] that before a finding of dishonesty can be made it must not only be pleaded but also put in cross-examination.

While it is correct that the facts of the cases those authorities in turn referred to turned on deficient cross-examination that does not undermine what Briggs J said regarding a dual requirement of pleading and cross-examination. In those underlying authorities there was as a result equally no specific need to consider the proposition Mr Foulkes advances that cross-examination, but without advance pleading, would be sufficient to enable a court or tribunal to find the serious allegation was made out. His proposition would also go against the rationale for advance pleading more generally namely of enabling the other party to know the case they had to meet.

As to Mr Foulkes’s argument that the “someone had actual knowledge” allegation provides sufficient notice to Ammanford, knowing that a particular individual is alleged to have actual knowledge might well, in our view, engage different and more detailed and targeted responsive evidence or disclosure than a general allegation that someone knew. Simply providing an opportunity to deal with it for the first time in cross-examination would not address that concern. The issue is not only about fairness to the witness of knowing the allegation in advance, but also about the party knowing what case it has to meet. The point is not that the taxpayer would be surprised, given the overall allegation, that a particular individual or individuals were alleged to have actual knowledge, but that if that is a finding that HMRC will be inviting the tribunal to make it is only fair that the taxpayer can provide its evidence and disclosure knowing that specific allegation.

In coming to this view, it should be emphasised that the relevant allegation of actual knowledge here is an issue on which HMRC bears the burden. None of this should be taken to suggest that an advance pleading of dishonesty or other serious allegations would be required in order to be able to put such allegations to the witness of the party who bore the burden for instance to displace an assessment (as explained by Henderson J in Ingenious Games LLP & Ors v HMRC [2015] UKUT 105 (TCC) at [62] to [65]). (The written or oral evidence underlying those serious allegations will of course have to be before the court.)

We address the consequences of the FTT’s error after we deal with the remaining aspects of this ground of appeal and Ground 2.

As regards the remaining points Ammanford argues under Ground 1, these properly concern matters the FTT decided. In relation to the s69D appeal it is argued the FTT erred in agreeing that HMRC had set out primary facts of the allegation because it had set out case that the directors “were responsible for the actions of the company during the periods under appeal”. Apart from a “one man band” set up, simply being responsible for the actions of a company as a director would not be a primary fact supporting actual knowledge. Thus, Mr Watkinson argues, the FTT ought to have directed HMRC to provide the primary facts for the allegation (according to Ronald Hull which FTT purported to follow). Furthermore, he also argues the FTT erroneously concluded that mere directorship was sufficient to found the s69D personal liability which would render the “attributability” requirement under that provision redundant.

In agreement with Mr Foulkes, neither of these points however amount to an error of law by the FTT. The FTT’s reference to the directors being responsible cannot be viewed in isolation but must take account that it had “carefully examined the Statement of Case and the evidence which HMRC [had] served” noting the particular set up of Ammanford i.e. that it had “only two directors, one book-keeper and a small number of yard staff”. Its statement regarding responsibility was thus grounded in an analysis of the particular circumstances of the appellant company as pleaded in the Statement of Case and the served evidence, not simply on the footing of the directors’ mere status as directors. There was accordingly no error of principle on the FTT’s part. As to the FTT’s view regarding the sufficiency of HMRC’s statement of the primary facts, the FTT was clearly satisfied that the Statement of Case and evidence had done enough to make the primary facts clear. That was plainly a matter of judgment and nothing in Ammanford’s grounds demonstrated why it was not open to the FTT to reach that conclusion.

Under this ground Ammanford submits the FTT was wrong to refuse appellant’s application that HMRC explain 1) what type of fraud they alleged the asserted scheme to defraud had been set up to achieve – was it an “MTIC fraud” or an “acquisition fraud” 2) the appellant’s alleged role in the fraud.

At FTT [90] the FTT explained:

As the FTT had set out earlier at FTT [64] the Court of Appeal in Fonecomp Ltd v HMRC [2015] EWCA Civ 39 (at [51]) had held that for the purposes Kittel knowledge, the trader just needed to know or have the means of knowing “that the fraud has occurred, or will occur, at some point in some transaction to which his transaction is connected”. The trader did not need to know the specific details of the fraud.

Ammanford does not disagree with that. Its point is that the FTT wrongly conflated proof of knowledge by Ammanford of the details of the scheme (which was not per Fonecomp required) and the primary facts of the scheme to defraud on which HMRC sought to found the inference that Ammanford knew of connection with fraudulent evasion of VAT. It says that the type of fraud was important to know because different types of fraud permitted different inferences to be drawn. (As explained by the FTT in CCA Distribution Ltd (In Administration) [2020] UKFTT 222(TC), in an “acquisition fraud” the fraudster imports goods free from VAT (the relevant terminology for EU imports being “acquire”), then sells the goods to a genuine buyer within the UK and then absconds without paying HMRC the VAT due on the sale. MTIC fraud also involved a default by the UK entity acquiring goods from the EU. But the way the fraud worked was different. It relied on a trader further down the chain having, under VAT law, to pay VAT to UK suppliers but not the EU customer they sold onto and then being able to recover the VAT they paid their UK supplier from HMRC. In contrast to acquisition fraud, which needed a genuine buyer for the goods, MTIC fraud did not take place in a genuine market but relied on engineering purchases and sales of goods to create the situation where tax repayments from HMRC were due.)

Ammanford also says that it needed the information detailing the fraud to answer the Fairford direction (which as mentioned above ([5]) typically asks the appellant to state in writing whether they accepted the transaction chains “were part of an orchestrated overall scheme to defraud the revenue”).

Ammanford also argues the FTT erred in concluding there was no requirement on the part of HMRC to provide information on Ammanford’s purported “role” in the asserted scheme to defraud. (The reference to Ammanford’s role arose in the context of paragraph 81 of HMRC’s Statement of Case where HMRC alleged the transactions “were undertaken as a part of a plan to defraud the Revenue. Given the role played by the Appellant, and the information known to it, it must have known of the connection with the fraudulent evasion of VAT.”)

The FTT dealt with that particular information request at [93] as follows:

Ammanford says none of these examples however identifies any “role” Ammanford was said to play in the asserted scheme.

Mr Foulkes’s response was that there was “no magic” in whether the fraud was acquisition or MTIC fraud, indeed CCA Distribution acknowledged any variation between two methods could be possible. What was important, in his submission, was not the label of the fraud but the actual nature and features of the chain of transactions which HMRC relied on to say the fraudulent scheme existed. It was that which enabled the appellant to address evidence as to what it would and should have been aware of within supply chain. We agree with that submission. We also agree that for the same reason there was no difficulty in Ammanford being able to respond to the Fairford direction. HMRC identify factors for why it is said the deals lacked commercial logic. Ammanford can take a view on whether it considers those indicate an overall orchestrated scheme to defraud. If they are not satisfied then it is of course open to them not to agree the point.

We also agree with Mr Foulkes’s submission that the FTT was entitled to conclude the information provided by the Statement of Case on Ammanford’s role was sufficient. The role HMRC were referring to, as the FTT correctly appreciated, was simply that Ammanford was involved in transactions which had no commercial logic. The reference to role was not speaking to a particular role in terms of MTIC jargon such as a “broker”.

As for Ammanford’s submission that the FTT wrongly conflated the fact that HMRC did not have to show Ammanford knew the particulars of the fraud with there being no obligation on HMRC to provide further information on the fraud for the purpose of allowing inferences to be drawn, Mr Foulkes accepts that these are separate issues but emphasises we should look at the totality of the FTT’s reasoning. He acknowledges, as he submits the FTT did, that some information regarding the alleged fraud needed to be provided but argues that the FTT was correct to conclude that HMRC had provided sufficient information and that no further information was required. He underscores the FTT’s references to “further information” in FTT [90] and FTT [93] and to “Again” at the beginning of FTT [93].

We consider however that the wording of FTT [90] clearly does link the question of what HMRC had to prove in terms of knowledge of the fraud with HMRC not having to provide further information. As both parties acknowledge, the one did not follow from the other. That HMRC did not need to show knowledge of the specifics of the fraud did not mean there could not be relevant inferences that might be drawn from information pertaining to the specific nature of the fraud. That error would apply just as much to further information as it would to the information already provided. Nevertheless, in agreement with Mr Foulkes’ case in the alternative, we consider the point does not assist Ammanford for the reasons already discussed. The FTT rightly came to the view that the information HMRC had already provided (the lack of commercial logic, and the various indicators of that) gave the appellant sufficient information on what HMRC’s case was regarding the inferences to be drawn from the alleged fraud. The FTT’s error was immaterial. It did not affect the ultimate analysis that HMRC’s references to factors which indicated a lack of commercial logic were sufficient information for the purpose of the appellant preparing its case.

This ground of appeal must therefore be rejected.

We set aside the FTT Decision and remake it, but only to the extent necessary to correct the error of law we have identified above at [46] in relation to Ground 1, to incorporate the reasoning at [47] to [59], and to remove the error under Ground 2 noted above at [75]. The remade decision accordingly rejects the proposition that HMRC may put to a witness appearing on behalf of Ammanford that the witness had actual knowledge of connection to fraud without having pleaded a case that identified such witness. The effect is that should HMRC wish to put to specific witnesses appearing on behalf of Ammanford that they had actual knowledge, HMRC will need to apply to amend its pleadings. In all other respects our remade decision adopts the FTT’s refusal of Ammanford’s information request and does so for the reasons the FTT gave.

The appellant’s appeal is allowed in part.