R T Rate Limited v The Commissioners for HMRC

In joined cases C-181/04 to C-183/04 Elmeka NE v Ipourgos Ikonomikon, the CJEU gave guidance to national courts on the EU doctrine of legitimate expectation. In that case, the taxpayer company (Elmeka) was established in Greece and provided services relating to the transport of petroleum products to a recipient (Oceanic) which was established outside the Community. With effect from 1 January 1993, harmonised Community VAT law required Elmeka to subject its services to VAT because, those services were supplied in Greece, even though they were supplied to a person established outside the Community. However, in June 1994, the Piraeus tax authorities wrongly confirmed to Elmeka that the supplies were exempt from VAT. Elmeka relied on this confirmation, by not accounting for VAT on its services and by not passing on any VAT cost to the price it charged Oceanic. Subsequently the Greek tax authorities discovered the error and sought to charge Elmeka VAT, interest and penalties for tax years 1994, 1995 and 1996.

Litigation ensued in Greece which resulted in the Greek courts referring the following question to the CJEU:

The CJEU gave the following explanation of the applicable principles at [31] and [32] of its judgment:

There was some doubt in Elmeka whether the Piraeus tax authority approached for a ruling on the issue was competent under applicable Greek law to give the ruling. The CJEU accordingly stated in paragraph 35 of its judgment that it was for the Greek national court to decide whether Elmeka could reasonably have believed that the Piraeus tax authorities were competent to give the ruling.

The decision of the CJEU in Case C-144/14 Cabinet MedicalVeterinar Dr. Tomoiagă Andrei concerned the VAT treatment of the services of veterinary practitioners. Before Romania joined the EU, such services were exempt from VAT in Romania. However, when Romania joined the EU, with effect from 1 January 2007, such services were taxable under harmonised Community VAT legislation. The Romanian legislature amended the law to remove references to veterinary care from the list of transactions exempt from VAT, but the taxpayer argued that he had a residual legitimate expectation that he could benefit from the predecessor exempt treatment. The CJEU formulated the concept of legitimate expectation in the following terms, in paragraphs 43 and 44 of its judgment:

From those authorities, we derive the following propositions:

RT Rate argues, by reference to the judgment of the Supreme Court in R (on the application of Davies and another) v HMRC and R (on the application of Gaines-Cooper) v HMRC [2011] UKSC 47, that the test set out in paragraph 33(3) involves a question of law. We disagree. Davies and Gaines-Cooper involved an application of the English law concept of legitimate expectation in the context of judicial review proceedings brought by two taxpayers. One of the issues arising in that case was the question whether HMRC had engendered a particular legitimate expectation by the publication of a booklet, known as IR20, which in turn involved an examination of what the relevant paragraphs of IR20 meant. We quite accept that the Supreme Court treated the meaning of the IR20 booklet as a question of law. However, in this case, the meaning of the Italian Tables and accompanying HMRC guidance is not in dispute. The test in paragraph 34(3) does not, therefore, necessitate any determination of the meaning of the Italian Tables. Rather the question is what, if any, expectation a “reasonably prudent economic operator” would have obtained from a reading of the Italian Tables and surrounding circumstances. In our judgment, that involves the application of an objective factual test.

In deciding whether to entertain the 2016 Claim, HMRC were not determining whether supplies were taxable or exempt under provisions of the Directive (the situations considered in Elmeka and Andrei). Rather, they were deciding whether a claim was made in time or not. The determination of the procedure for enforcing EU law rights, and the time limit for making claims, is generally within the competence of individual member states, as distinct from a matter that is harmonised at the level of the Community. However, in their submissions, HMRC did not seek to argue that this provided a reason why the EU law principle of legitimate expectation was not in issue. We therefore proceed on the basis that it was common ground that, if the Italian Tables gave rise to a legitimate expectation in the EU law sense, HMRC were obliged to respect that expectation.

In our judgment, the short answer to RT Rate’s Ground 2 is that the FTT, a specialist tribunal, has expressed its evaluative conclusion that a reasonably prudent economic operator would not have derived the expectation from the Italian Tables for which RT Rate argued. The FTT’s conclusions in this regard were available to it and, since an appeal to the UT lies only on a point of law, we should not interfere with those conclusions.

RT Rate seeks to escape from this conclusion by arguing that the FTT’s evaluative findings were vitiated by either a failure to take into account relevant factors, or by the reaching of a conclusion that was so outside the reasonable range of conclusions as to involve an error of law of the kind set out in Edwards v Bairstow 36 TC 207.

The argument that the FTT failed to take into account relevant considerations centred on the fact that the Italian Tables referred to the “typical sale price” of vehicles but the FTT failed to address the significance of that wording. We reject that argument. The FTT quoted an extract from the Italian Tables at [14] and in doing so demonstrated that it was aware that the column containing figures applicable to “Volume” cars was headed “Typical sale price”. (In fact, the corresponding columns applicable to “Prestige” and “Other” cars were headed “Sale Price”, without the use of the word “typical” though it was not suggested to us that there was any significance in this difference of wording). The FTT also referred in detail, at [90], to RT Rate’s arguments that the reasonable expectation was that the figures in the Italian Tables would be correct as typical figures. The FTT did not fail to take this consideration into account. Having properly taken it into account, it was a matter for the FTT to decide how much weight to give that consideration.

RT Rate also argues that the FTT’s evaluative conclusion fell outside the reasonable range of conclusions available to it. Effectively, this is an argument that the only reasonable conclusion available to the FTT was that a reasonably prudent economic operator would have concluded from the Italian Tables that HMRC were (i) representing that those figures were materially correct as “typical/average” figures and (ii) that therefore any claim made in accordance with the Italian Tables would be closed (if it was closed) on a materially correct basis.

RT Rate submits that HMRC must have been giving some representation as to the figures’ accuracy because those figures were not presented as arbitrary. In RT Rate’s submission, the reasonable basis on which it relied upon the Italian Tables would have been fundamentally undermined if it became clear that HMRC had arbitrarily applied a 50% discount to the figures set out in the tables. One problem with that submission is that it assumes what it seeks to prove. In effect it is said that, because RT Rate and other taxpayers would expect the figures in the Italian Tables to be “accurate”, it would be objectionable for HMRC to apply an arbitrary discount when finalising and publishing the figures.

However, even putting that objection to one side, RT Rate’s argument is incapable of establishing that the only reasonable conclusion was that HMRC were providing some confirmation that the figures in the Italian Tables were accurate. The FTT’s conclusion to the contrary set out [96] and [97] was entirely reasonable and, indeed, we respectfully share it. The Italian Tables set out a relaxation to what would otherwise have been a requirement for a trader making a claim for repayment in respect of output tax paid on sales of demonstrator vehicles to establish with precision the precise margin that had been subjected to VAT. The relaxation consisted of HMRC indicating that, provided a claim was otherwise plausible and satisfactory, they would accept claims based on the figures in the Italian Tables. Traders were free to avail themselves of that relaxation or not. If they thought that they had, or could obtain, evidence of their actual margins that would lead to a higher VAT repayment they were free to obtain that evidence and seek to persuade HMRC accordingly. If they could not obtain that evidence, or were unwilling to go to the effort of obtaining it, or if they thought that the figures contained in the Italian Tables were sufficient, they were free to base claims on the Italian Tables. It may well be that, in publishing the Italian Tables, HMRC was giving some assurance that claims based on them would at least be considered. However, the Italian Tables contained no express confirmation of the accuracy of the figures and in the light of the purpose and function of the Italian Tables that we have just highlighted, we do not consider that any such representation should be implied. Still less do we consider that the only reasonable conclusion open to the FTT was that the Italian Tables contain an implicit representation as to the accuracy of those figures.

RT Rate also criticises the FTT’s conclusion, in the final sentence of [96] to the effect that, even if there was a legitimate expectation that figures in the Italian Tables were accurate, there was no expectation that HMRC would allow closed claims made on the basis of the Italian Table to be re-opened at any point in the future. That, argues RT Rate, is to confuse the nature of the expectation with the nature of the remedy that should be ordered where the expectation is not honoured. We reject that submission. RT Rate’s case before the FTT was that its legitimate expectation consisted of two aspects: first that the figures in the Italian Tables were materially correct as the typical/average figures that they purported to be and second that claims it made would be closed on such a materially correct basis. It follows from RT Rate’s formulation of its case that a reasonably prudent economic operator would have expected that claims settled on the basis of materially “incorrect” figures would not be “closed” and, therefore, must necessarily have remained “open”. In our judgment, in the final sentence of [96], the FTT was considering whether such an expectation could be “justified”. Its conclusion was that any such expectation could not be justified because it would result in apparently settled claims remaining open indefinitely. That evaluative conclusion was open to the FTT.

For those reasons, in our judgment we are not entitled to interfere with the FTT’s evaluative conclusions as to the absence of a legitimate expectation in the EU law sense and RT Rate’s appeal on Ground 2 accordingly fails.

We also consider that the point raised in HMRC’s Response, referred to in paragraph 27(1) above, supplies a further reason why RT Rate could not have held a legitimate expectation that EU law would protect. As we have noted in paragraph 33(3) above, any legitimate expectation has to be engendered by “precise assurances”. We acknowledge the point, made by the FTT at [94], that it can be meaningful to speak of a figure as being “materially” correct where that figure is wrong, but differs by a sufficiently slender margin from the true figure. However, in this case, there are no “true” figures against which the figures in the Italian Tables can be compared. There is no “true” price for the profit that could be made on resale of a “Prestige”, “Volume” or “Other” car for a variety of reasons. The three categories of car identified are very broad labels: at the margins it will not be clear which cars fall into which categories. Moreover, one car dealership may be more or less profitable than another. Profit margins may vary depending on where the dealership is located, or the particular month in the year in which a particular car is sold.

Therefore, the figures in the Italian Tables cannot easily be compared with “true” figures in order to gauge whether they are “materially” accurate even as typical or average figures. Instead, those figures are best understood as the outcome of a modelling exercise designed to produce a theoretical construct, namely a supposedly “typical” margin achievable on different types of car in different years. Doubtless, there are different ways of approaching that modelling exercise and some ways would be more robust than others. However, in the absence of any “true” figures against which those in the Italian Tables can be compared, we do not consider that there is sufficient precision in the statement that those figures are expected to be “materially” accurate to warrant the protection of the EU law doctrine of legitimate expectation.

For the reasons that we have given, RT Rate’s Ground 2 fails. That makes Ground 3 (which considers the question of remedy) academic and we will not, therefore, deal with that Ground.

We have carefully considered whether it would be right for us to deal with Ground 1 so as to give guidance to the FTT on the scope of its jurisdiction to deal with further claims of this nature. We have concluded, however, that this would not be the appropriate course to follow. We are concerned that any guidance we give to the FTT in a situation where, in our judgment, there was quite clearly no legitimate expectation of the kind that EU law would protect, would be answering a purely academic and theoretical question. We do not think that an issue of this kind is best addressed in a case where, on the facts, it does not actually arise. In short, we consider that the scope of the FTT’s jurisdiction is best addressed by a binding statement from this Tribunal only in a case where such a statement is necessary. We will not, therefore, address Ground 1.

There is less authority on the scope of the EU law principle of equal treatment than there is on the principle of legitimate expectation. For an articulation of that principle, we were referred to the judgment of the CJEU in Marks & Spencer Plc v Commissioners of Customs (Case C3-09/06) [2008] STC 1408 (“Marks & Spencer 2006”).

That judgment, so far as relevant to the principle of equal treatment, concerned legislation in s80 of VATA which entitled HMRC to refuse to repay output tax overpaid by a trader where that repayment would lead to unjust enrichment. In essence the legislation was intended to ensure that traders who had passed the cost of output VAT on to their own customers could not claim what the UK authorities regarded as a windfall consequent on reclaiming that output VAT from HMRC. Significantly, there could be no “unjust enrichment” within the scope of the statutory provision as drafted for “repayment traders”, namely taxpayers who were routinely net recipients of VAT refunds from HMRC (generally because they made significant zero-rated supplies and therefore enjoyed significant input tax recovery without any material corresponding obligation).

The taxpayer company was a “payment trader” that routinely had a net VAT liability due to HMRC. It successfully established that it had overpaid output tax. Eventually HMRC agreed to make a repayment. However, they would only repay 10% of the amount of output tax at issue, contending that the other 90% would have been passed on to the taxpayer’s customers with the result that the taxpayer would be unjustly enriched if paid the full amount of its claim. The taxpayer objected, arguing that it was being treated less favourably than a repayment trader which could have recovered the full 100% of output tax overpaid without any reduction for perceived “unjust enrichment”. The UK national courts referred the question of the applicability of the doctrine of equal treatment to the CJEU.

At [51] of its judgment, the CJEU summarised the doctrine of equal treatment in the following terms:

At [52] and [53], the CJEU concluded that domestic legislation did breach the principle of equal treatment because “payment traders” and “repayment traders” were in similar situations as regards their entitlement to repayments of VAT. Both categories of trader had overpaid VAT and that fact was not altered by the fact that repayment traders were net creditors of HMRC and payment traders tended to be net debtors to HMRC. Therefore, there was no objective justification for payment traders being subject to a restriction which did not apply to repayment traders.

RT Rate approached the issue of equal treatment before the FTT on the basis that it required the FTT to make factual findings as to the similarity or otherwise between RT Rate’s circumstances and the circumstances of KAP. If satisfied that RT Rate’s circumstances were similar, or sufficiently similar, it was argued that RT Rate should be afforded the benefit of the same settlement that KAP obtained pursuant to the KAP Agreement. As we have noted, the FTT doubted whether the principle of equal treatment applied in relation to an isolated decision which treats a single taxpayer more favourably than another in the absence of any legislative or policy basis for the unequal treatment. However, the FTT considered that it did not need to address that issue because its factual conclusion, set out at [115], was that RT Rate had not discharged its burden of proving that its circumstances were sufficiently similar.

RT Rate has framed its Ground 4 as challenges to the FTT’s factual conclusions on the question of similarity. In essence, it is argued that, instead of resorting to the burden of proof, the FTT should have tried harder to find relevant facts based on (i) the evidence that it did have (consisting primarily of the KAP Agreement), (ii) inferences that could be drawn from the terms of the KAP Agreement and (iii) adverse inferences that should have been drawn from HMRC’s failure to adduce evidence on KAP’s circumstances which were particularly acute given that HMRC was a public authority with a duty to identify “the good, the bad and the ugly” aspects of its case (drawing on paragraph 106(3) of the judgment of the Singh LJ in R (oao Citizens UK) v Secretary of State for the Home Department [2018] EWCA Civ 1812. RT Rate also argues that the points on which the FTT relied, at [115], in support of its conclusion that the burden of proof was not discharged were matters raised of the FTT’s own motion and did not form the basis of HMRC’s submissions.

For reasons set out below, however, we do not accept the premise underpinning RT Rate’s challenge. The FTT was not obliged to make any factual findings as to the “similarity” between RT Rate’s case and that of KAP. That is because we accept HMRC’s argument, set out in HMRC’s Response to RT Rate’s appeal, that even if the circumstances of RT Rate and KAP could be said to be “similar”, the principle of equal treatment has no application in the circumstances of this case.

To put our reasoning into context, we note that RT Rate and KAP are by no means the only taxpayers who have relied on the Italian Tables when making claims. Rather, the Italian Tables were of relevance to a class of taxpayers comprising a potentially significant number of motor traders. It is important to note the consequences that would flow if a single taxpayer within a class of taxpayers affected by an HMRC practice or decision were able to advance a case on equal treatment of the kind that RT Rate sought to bring:

Of course, the fact that these consequences could flow, and the fact that we regard them as deleterious, does not determine the scope of the EU principle of equal treatment. The law sometimes produces consequences that might be thought unwelcome. However, those consequences demonstrate the central flaw in RT Rate’s approach to the principle of equal treatment.

The flaw is this. The Italian Tables were evidently targeted at a class of taxpayers: motor traders who considered that they may have overpaid output tax on sales of demonstrator vehicles. Therefore, whenever HMRC make a decision about the claim of a taxpayer in that class, there is nothing particularly surprising about the proposition that that taxpayer might, in some respects at least, be in a similar position to other taxpayers in the class. Some “similarity” follows inevitably from the fact that the taxpayers are members of the same class.

Where HMRC make a decision on the tax affairs of a taxpayer who is a member of a class, the comparison which the doctrine of equal treatment requires has to be applied in a way that is workable having due regard to the constituents of the class as a whole. A comparison between the situations of individual members of the class is neither workable, nor mandated by the doctrine. Suppose that, to build on an example considered by the FTT at [114], Taxpayer A, who is a member of the same class as Taxpayers B and C, receives a favourable HMRC decision. Taxpayer B receives an unfavourable decision, considers challenging it, but decides not to and Taxpayer C receives a similar unfavourable decision to Taxpayer B. There is simply no coherent answer to the question whether Taxpayer C has been the subject of “unequal” treatment given that its treatment is identical to that of Taxpayer B, albeit different from that of Taxpayer A.

It follows that, in the present case, where RT Rate is a member of a class of taxpayers capable of being affected by HMRC decisions, any failure to afford RT Rate equal treatment must similarly be evaluated by asking whether HMRC are differentiating between the situations of different classes of taxpayer, in much the same way as the question was approached in Marks & Spencer 2006. In the circumstances of this case, RT Rate could not have succeeded with a claim based on the EU doctrine of equal treatment simply by pointing to an allegedly different treatment of a single other taxpayer within the class.

In any event, even if the FTT was required to turn its mind to the question whether RT Rate and KAP were in “similar” positions, we see no error of law in the FTT’s conclusions in that regard set out at [115]. The only evidence that RT Rate had put forward on the issue of “similarity” consisted of the KAP Agreement. That agreement largely documented the end result, namely the terms on which settlement was reached and the FTT was entitled to conclude that it was not, on its own, sufficient to discharge the burden of proving any necessary similarity between RT Rate’s situation and that of KAP. Parties to disputes reach settlements for a vast variety of reasons. Even acknowledging Mr Firth’s point that HMRC’s freedom to compromise disputes is constrained by its “Litigation and Settlement Strategy”, the FTT made a sustainable finding of fact in the last two sentences of [115] in concluding that the KAP Agreement itself contained insufficient evidence as to the specific circumstances of KAP that had resulted in the settlement being reached. RT Rate’s argument, that the FTT should have “striven to make a finding on the evidence before it”, is on closer inspection simply a disagreement with the finding that the FTT did make, namely that the evidence RT Rate put forward was insufficient and did not establish even a prima facie similarity between RT Rate’s circumstances and those of KAP.

We are only reinforced in this conclusion by our awareness, derived from email exchanges that RT Rate showed us during the hearing, that RT Rate was in touch with Mr Furneaux, a director of KAP, prior to the FTT hearing. That email exchange indicates that Mr Furneaux had some concerns about revealing details of the KAP Agreement, which contains a confidentiality clause. It also indicates that Mr Furneaux ultimately declined to provide RT Rate with much relevant information. However, RT Rate had the burden of proving its case. Even if Mr Furneaux was unco-operative, there was more that RT Rate could have done to obtain evidence of the averred similarity between its situation and that of KAP. It could, for example, have applied for Mr Furneaux to be made subject to a witness summons under the FTT’s rules of procedure. It could have applied for further disclosure of documents from HMRC.

RT Rate submits that HMRC’s duties as a public authority should have obviated the need for it to take such steps. It argues that HMRC should itself have drawn relevant material to the FTT’s attention, relying on the formulation of the duties of “candour and co-operation” of a public law authority set out in paragraphs 86 of the judgment of Singh LJ in Citizens UK to which we have already referred. We reject that submission. Citizens UK sets out the formulation of duties applicable in judicial review proceedings. Such proceedings are not concerned with the establishing of facts and indeed, as Singh LJ noted at [106(1)] and [106(2)] of Citizens UK, it is precisely because the ordinary rules of disclosure of documents do not apply in judicial review proceedings, that public authorities are subject to the duties of candour and co-operation. The proceedings before the FTT were not judicial review proceedings. They were, in part, proceedings concerned with the establishment of facts, and the ordinary burden on RT Rate, to prove its factual case, applied. That was particularly so given that HMRC was not, before the FTT, advancing any positive case that the situations of RT Rate and KAP were materially different but was instead relying on an argument that, even if similarity could be proved, the EU doctrine of equal treatment was still inapplicable.

We would perhaps place less weight than the FTT did, at [115], on the fact that the KAP Agreement does not mention the Italian Tables. Read as a whole, we consider that the KAP Agreement was demonstrably settling some kind of claim based on the Italian Tables. However, that does not alter our conclusion that, overall, the FTT was entitled to conclude that RT Rate had not discharged its burden of proof.

RT Rate also argues that the matters relied on at [115] in support of the FTT’s conclusion on the burden of proof were unprompted by submissions made by HMRC and so were identified for the first time in the Decision. We see nothing objectionable about that in the circumstances of this case. We accept that HMRC would have made no submissions as to how RT Rate’s case could be distinguished from that of KAP since, as we have noted, HMRC’s position was that, even if similarity could be proved, RT Rate’s case on equal treatment could not succeed. However, having decided that it would not decide the appeal on HMRC’s preferred basis, the FTT then needed to decide whether the evidence that RT Rate had put forward was enough. If that evidence had been more extensive, it might well not have been appropriate for the FTT to perform its own review of that evidence that was not tested by submissions from the parties. However, RT Rate’s evidence consisted entirely of the KAP Agreement. This ran to three pages (not including the cover page setting out the parties to it) of which one page (at least) can fairly be described as “boiler plate”. The FTT was entitled to decide for itself whether such scanty evidence could satisfy the burden of proof and, as we have explained, was entitled to its conclusion that the burden was not satisfied.

It follows that we also do not accept RT Rate’s argument that the FTT should have drawn an inference from HMRC’s failure to adduce evidence of their own as to the circumstance of KAP’s case. Having concluded, permissibly, that the KAP Agreement was simply insufficient to discharge RT Rate’s burden of proof, the FTT was entitled to conclude that no question of any adverse inference arose. Moreover, since the FTT had concluded, permissibly, that RT Rate’s evidence did not even disclose a prima facie case on “similarity”, it was entitled to treat that case as not proved, even though HMRC had adduced no evidence of their own.

For all of those reasons, we dismiss RT Rate’s appeal on Ground 4.

RT Rate’s appeal is dismissed on all grounds.