Nicholas and Charlotte Sandham T/A Premier Metals Leeds v The Commissioners for HMRC

If a person enters into transactions that are connected with fraudulent evasion of VAT either knowing, or having the means of knowing, that the transactions were so connected, the person is not entitled to credit for associated input tax. That is the familiar effect of the decision of the CJEU in Kittel v Belgium (Case 439/04) [2008] STC 1537. The short question raised by this appeal is how the principle in Kittel applies where the transactions in question were effected by a partnership through an agent and, while the partners did not themselves know that the transactions were connected with fraudulent evasion of VAT, the agent did.

In the decision under appeal released on 2 April 2019 (the “Decision”) the First-tier Tribunal (Judge Rupert Jones) made detailed findings of fact. None of those findings are under appeal and we therefore set out the relevant factual background briefly, with references to numbers in square brackets being to paragraphs of the Decision unless we state otherwise.

Nicholas and Charlotte Sandham carried on business together through a partnership (the “Partnership”) that traded under the name of Premier Metals. The Partnership was registered for VAT with effect from 7 March 2007. Initially, it bought and sold scrap metal. However, in February 2013 the Partnership began to trade in more valuable “primary metals” ([63]).

In its VAT period 02/13 and its final VAT period which ended on 31 March 2013 (which HMRC refer to as VAT period “99/99”), the Partnership entered into 56 transactions involving the purchase and immediate resale of lots of primary metals. Those transactions were all entered into by Mr Jonathan France who acted as agent for the Partnership and was authorised to purchase and sell primary metals on the Partnership’s behalf ([114]). There was no written document setting out the scope of his authority to act as agent ([113]). He was not authorised or instructed by the Partnership to commit fraud ([119]). Mr France knew that the transactions he was effecting on behalf of the Partnership were connected with fraudulent evasion of VAT ([115]).

On 30 March 2015, HMRC decided to disallow the Partnership’s claims for input tax in respect of the 56 relevant transactions. They subsequently made assessments to recover input tax for which the Partnership had already claimed credit through its VAT returns. The total amount of those assessments was £1,930,951 ([8]).

The Partnership appealed to the FTT against those decisions. Since the Partnership accepted ([14]) that all 56 purchases of primary metals could be traced back to a fraudulent VAT loss, the live issues before the FTT concerned the question whether the Partnership knew, or should have known, that the acquisitions were connected with fraudulent evasion of VAT. The FTT directed itself, at [19] to [21] that, in order to determine the Partnership’s appeals, it needed to determine the following two issues:

Whether Mr France’s knowledge that the transactions were connected to the fraudulent evasion of VAT should be attributed to the Partnership so that, when applying the Kittel test, the Partnership should be taken as knowing that the transactions were so connected (“Issue 1”).

Whether, if Mr France’s knowledge was not attributed to the Partnership, Mr and Mrs Sandham themselves either knew, or should have known, that the transactions were connected with fraudulent evasion of VAT (“Issue 2”).

The FTT accepted HMRC’s arguments on Issue 1. That meant that the Partnership’s appeal was dismissed. However, the FTT went on to consider Issue 2 concluding that:

Neither Mr Sandham nor Mrs Sandham themselves knew that the relevant transactions were connected to fraudulent evasion of VAT ([442]).

While they did fail to take reasonable care by asking basic questions about the way in which Mr France was carrying on the primary metals business ([565]), neither Mr Sandham nor Mrs Sandham had the means of knowing that the transactions were connected with fraudulent evasion of VAT ([576]).

With the permission of the FTT, the Appellants appeal to this tribunal on the ground that the FTT erred in law in concluding that Mr France’s knowledge that the 56 transactions were connected to fraudulent evasion of VAT was sufficient for the principle in Kittel to apply. Rather, in the Appellants’ submission, the FTT’s conclusions on Issue 2, which neither party seeks to disturb, mean that the Kittel principle was not engaged.

There can be no doubt that, at least in certain circumstances, if an agent has knowledge of particular matters, the agent’s principal is also to be treated as having that knowledge. We doubt that authority is needed for that general proposition, but to the extent it is, it can be found in the words of Lord Halsbury in Blackburn, Low & Co v Vigors12 App Cas 531 at p 537:

Recent case law has emphasised the importance of context when deciding whether an agent’s actions or knowledge are to be attributed to a principal. In part, that development in the law has been driven by situations involving claims of fraud or breach of duty by a principal against an agent. It was obviously unattractive for an agent to be able to plead, as a defence to such a claim, that his or her knowledge of their own wrongdoing should be attributed to the principal so as to defeat the claim. The emphasis on context, therefore enabled the law to articulate a distinction between situations where an agent’s knowledge was to be attributed to a principal and those where it was not. Lord Toulson and Lord Hodge, with whom Lord Neuberger agreed, expressed the point as follows at [181] of their judgment in Bilta (UK) Ltd (in liquidation) and others v Nazir and others[2016] AC 1:

The next logical question is how context should be ascertained for the purpose of deciding whether an agent’s actions or knowledge are to be attributed to the principal. In Meridian Global Funds Management Asia Limited v Securities Commissioner 2 AC 500 gave some guidance on this issue. That case involved the question of whether the knowledge of an individual fell to be attributed to the company that employed him and, accordingly, some of the Privy Council’s analysis drew on principles of company law. For example, at 506A of the reported judgment, Lord Hoffman observed that:

Lord Hoffman then went on to explain that what he referred to as “primary rules of attribution” are typically to be found in a company’s constitution. A provision of articles of association to the effect that “the decisions of the board in managing the company’s business shall be decisions of the company” is an example of such a “primary rule”. A further example is the principle of company law to the effect that a unanimous decision of all shareholders in a solvent company to do anything within the company’s power is to be treated as the decision of the company.

Lord Hoffman went on to say (as 506F to 507F):

These passages are, of course, focusing on the rules of attribution as applied to companies. However, in our judgment they also give guidance as to how the question of context is to be approached more generally. Lord Hoffman makes it clear that the task is to construe the legal provision at issue in order to determine whether the knowledge or actions of an agent are to be attributed to a company. If the legal provision in question suggests that an agent’s knowledge or actions are not to be attributed, then it is necessary to fashion a “special rule of attribution” that decides, for the purposes of that legal provision, whose knowledge and actions are to be attributed to the company.

Therefore, the context to be identified when deciding whether the acts or knowledge of an agent are to be attributed to a principal is a specifically legal context. This is confirmed in the following passage from the judgment of Lord Toulson and Lord Hodge in Bilta:

In the circumstances of this appeal, the “context” in which the question of attribution arises is the rule set out in Kittel. We therefore need to decide whether, in that legal context, the knowledge of Mr France fell to be attributed to Mr and Mrs Sandham, the partners, for the purposes of deciding whether they knew that the transactions effected were connected to fraudulent evasion of VAT.

In this section we will consider the question we posed at [16] by reference to various aspects of the relevant context to which the parties referred us. As a preface to that analysis, we note that Mr Bott QC accepted in his oral submissions that, if Mr France’s knowledge fell to be attributed to the Appellants, he was not seeking to apply the “Hampshire Land principle” to displace that attribution. As we have already noted, following Bilta, it is unlikely that the decision in Hampshire Land sets out a free-standing exception to the law on attribution and instead should be regarded as setting out relevant aspects of the legal context. Nevertheless, the Appellants’ concession is significant as they accept, rightly in our judgment, that the fact that Mr France clearly breached duties owed to the Appellants when acting on their behalf does not of itself prevent his knowledge being attributed to them.

The Appellants rely strongly on the finding of the FTT, at [431], that the case had distinct and unusual features. Those findings, they argue, demonstrate that they were comprehensively deceived by Mr France, a serial fraudster with a track record of deceit, including of his own trustee in bankruptcy. Before Mr France’s involvement in their business, the Appellants had an “impeccable” reputation and trading history but, for his own ends, Mr France entered into dishonest and artificial transactions on their behalf. Mr Bott QC argued in his closing submissions that the truly exceptional circumstances of this case are a relevant aspect of context to which we should have regard in deciding whether it is fair and just for the Appellants to be fixed with Mr France’s knowledge. He drew an analogy with developments in the law of tort which, he submitted, adopted a “nuanced and case specific” approach to the question of whether a person should be vicariously liable for the acts of their employees and agents.

The FTT’s findings, both in [431] of the Decision and elsewhere clearly evoke sympathy for the Appellants. However, we do not accept that a consideration of context involves the kind of general analysis of overall “fairness” which the Appellants urge us to perform. As we have noted, the context we are required to consider is a specifically legal context arising from the legal principle laid down in Kittel.

Mr Bott QC sought support from the decision of the Supreme Court in Singularis Holdings Ltd (in liquidation) v Daiwa Capital Markets Limited[2019] UKSC 50. In that case, the question was whether, for the purposes of an action by a company against its bank and broker alleging breach of a duty of care in making transfers of funds, the company should be attributed with the knowledge and actions of its sole shareholder and director who instructed the bank to make the transfers. The Supreme Court held that the knowledge and actions of the agent should not, in this context, be attributed to company for the reasons set out at paragraph 35 of the judgment of Baroness Hale, with whom all members of the court agreed:

We do not consider that analysis of context to be based on general considerations of “fairness”. On the contrary, it is based on an analysis of the nature of the legal duty in connection with which the question of attribution arose.

For a similar reason, we derive little assistance from the common law on vicarious liability. The Appellants observe that, where a claimant seeks to make an employer vicariously liable in tort for the actions of an employee, a court is required to inquire as to the nature of the employee’s job and then to ask whether there was sufficient connection between that job and the employee’s wrongful conduct to make it right, as a matter of social justice, for the employer to be held liable (see Mohamud v WM Morrison Supermarkets Plc[2016] UKSC 11). However, that is a different test applied for a different purpose. The concept of vicarious liability involves a consideration of whether one person should be liable for the tortious acts of another. Here the question is whether, applying the Kittel principle, the Appellants had knowledge of connection of fraud so as to disqualify them from entitlement to input tax credits in connections with transactions which they had effected.

Mr Bott QC pointed out that, in the extract from Meridian which we have quoted at [13] above, Lord Hoffman stated that primary rules of attribution “together with the general principles of agency, vicarious liability and so forth are usually sufficient to enable one to determine [a company’s] rights and obligations”. However, in that passage, Lord Hoffman was not suggesting that the law governing the attribution of an agent’s acts to a principal was similar to the law on vicarious liability. He was just noting that the totality of a company’s rights and obligations could be determined by applying primary rules of attribution, the law of agency and principles of vicarious liability.

We therefore consider that the “distinct and unusual” features of this case shed little light on how the question of attribution should be approached.

The Appellants emphasise that they gave Mr France authority to enter into ordinary trading in primary metals but gave him no authority to engage in the contrived transactions, far removed from ordinary trading, forming part of an orchestrated VAT fraud. That, they argue, is part of the relevant context which indicates that they should not be attributed with Mr France’s knowledge of that fraud.

The FTT accepted at [119] that Mr France was not instructed, or authorised, to commit fraud finding as follows:

However, though Mr France was acting outside the scope of the authority he had been given, the FTT concluded that he still bound the Partnership when entering into the 56 transactions on its behalf concluding, at [114]:

Moreover, the Appellants acknowledged that Mr France had bound them into those 56 transactions by claiming credit for input tax associated with them. We therefore agree with HMRC that there is an inherent contradiction in the Appellants’ position. On one hand, they assert that Mr France bound them into the transactions for the purposes of substantiating their claim for input tax credit. However, when it comes to applying the Kittel principle, they seek to distance themselves from Mr France’s knowledge of fraud. We would respectfully endorse what the Upper Tribunal (Morgan J and Judge Sinfield) said of a similar argument in Mobile Sourcing Limited v HMRC[2016] UKUT 274 (TCC) as follows:

We also note that in the Mobile Sourcing case one of the assumed facts was that MSL relied upon the assurances of the officers and employees of Wigig that (a) transactions were carried out conscientiously and properly with neither the knowledge nor the means of knowledge of the alleged connection to missing traders and (b) the terms of their agreement were being observed. In other words, in acting as it did Wigig acted contrary to an express instruction that all instructions were to be proper. That was no bar to attribution of Wigig’s knowledge to MSL; what mattered was that Wigig had been acting on MSL’s behalf in entering the transactions under which MSL claimed the input tax.

We do not, therefore, consider that the fact that Mr France acted contrary to his express instructions prevents his knowledge of fraud from being attributed to the Appellants.

We analyse under this heading the submissions that Mr Bott QC made to the FTT summarised, fairly and accurately he accepted, at [151] to [174] of the Decision, which the Appellants renew on this appeal.

We do not accept the submission summarised in those paragraphs that to attribute Mr France’s knowledge to the Partnership or to the Appellants would involve a mistaken application of the law on corporate attribution. As we have explained above, the principle that the knowledge of an agent can, depending on the applicable legal context, be attributed to a principal does not apply only where companies are concerned. It is a rule of more general application. Indeed, at 506F of Meridian, Lord Hoffman said:

Indeed, as HMRC pointed out, there are authorities dealing with the attribution of the knowledge of an agent to a partnership which demonstrate that this is not a principle of purely corporate attribution. For example, in Re Drabble Brothers[1930] 2 Ch 211, two brothers carrying on business in partnership delegated to their agent, Tiley, the task of deciding which of the partnership’s creditors to pay, and when. The brothers became insolvent and, in full knowledge of their insolvency, Tiley arranged for a particular creditor to be paid so as to confer a fraudulent preference on that creditor. The Court of Appeal rejected the argument that there was no fraudulent preference because, while Tiley had the necessary knowledge, the partners did not, with Lord Hanworth MR saying at pp.235 to 236:

At the hearing before us, the Appellants made a related submission namely that, because the Appellants entered into the transactions as part of a business carried on as a family partnership, the question of knowledge should appropriately be tested by reference to their own state of mind rather than the state of mind of an agent acting on their behalf. We reject that submission since, if correct, it would mean that persons acting in partnership could never be treated as possessing the knowledge of their agents. That would leave partnerships free to delegate all aspects of their business to potentially dishonest actors without being answerable for the consequences so long as they ensured that they were kept uninformed of what was going on. We do not need to decide how likely it is that partnerships would act in this way. It is sufficient to note that, for reasons we expand upon in the next section, such an interpretation would be contrary to the purpose and effect of the Kittel principle.

Thus far, our analysis has focused on our reasons for rejecting the Appellants’ submissions that considerations of context mean that Mr France’s knowledge should not be attributed to the Appellants. We will conclude our analysis by explaining why, in the context of the rule in Kittel, his knowledge should be so attributed.

The Upper Tribunal, in its decision in Mobile Sourcing, to which we have already referred, has already considered this issue. That analysis was performed in the context of transactions effected by a company but, as we have explained in the section above, we consider that identical considerations apply where an agent enters into transactions on behalf of persons carrying on business in partnership (or, indeed, as sole traders). We would say quite simply that we respectfully agree with the analysis in paragraphs [47] to [55] of Mobile Sourcing.

In the interests of brevity, we will not quote the entirety of the reasoning in Mobile Sourcing. Nor will we attempt to summarise it since any summary would run the risk of omitting important aspects of that reasoning. Rather, we will simply highlight what we see as three key aspects of the context provided by the Kittel test as highlighted in Mobile Sourcing:

The Kittel test involves a consideration of an issue lying between HMRC and Appellants, namely whether the Appellants should obtain credit for input tax said to have been incurred in connection with particular transactions. Since that issue lies between HMRC and the Appellants, there is no obvious reason why the Appellants should not be attributed with the knowledge of the agent who entered into those very transactions.

That conclusion is only reinforced once it is appreciated that the Appellants’ claim for input tax depends on the assertion that Mr France entered into the 56 transactions on their behalf. Therefore, if the Appellants were not attributed with Mr France’s knowledge, they would simultaneously be relying on Mr France’s acts as their agent to substantiate their claim for input tax credit yet denying that his knowledge should be attributed to them when considering whether the right to input tax credit could be restricted under the principle in Kittel.

The Kittel principle is designed to protect member states from VAT fraud. The rationale for the principle is that a person who enters into transactions either knowing, or having means of knowledge, that they are connected with fraudulent evasion of VAT is not entitled to credit for input tax associated with those transactions on the basis that the taxpayer becomes, in effect, the fraudsters’ accomplice. We do not see any reason why that should be treated as excluding the normal rule to the effect that a principal is fixed with the knowledge of an agent. Indeed, if the knowledge of the agent who entered into the very transactions giving rise to input tax was not attributed to the principal, we accept HMRC’s submission that it might be possible for taxpayers to avoid the consequence of the Kittel rule by entering into transactions through agents while ensuring that they remain ignorant of the full circumstances of those transactions. We recognise that there is no suggestion that the Appellants themselves engaged in such conduct, but we see no reason why the Kittel principle, intended as it is to guard against fraud, should even give taxpayers the option of doing so.

Finally, we do not doubt the Appellants’ assertion, advanced in the skeleton argument of Mr Bott QC that, as a matter of criminal law, partnerships are not generally capable of being guilty of offences except where a statutory provision provides otherwise. However, we see no reason why that should inform a consideration of the context provided, in the civil arena, by the Kittel principle.

For the reasons we have given, the FTT was correct to conclude that when applying the Kittel principle the Appellants were to be attributed with Mr France’s knowledge that the 56 transactions were connected with fraudulent evasion of VAT. The appeal is dismissed.