Queenscourt Limited v The Commissioners for HMRC

Neutral Citation: [2026] UKUT 00195 (TCC) Case Number: UT/2024/000112 UPPER TRIBUNAL (Tax and Chancery Chamber) The Royal Courts of JusticeRolls Building, London VALUE ADDED TAX – whether dip pots supplied as part of a takeaway meal deal are a separate zero rated supply or whether they are part of a composite standard rated supply alongside the hot food in the relevant meal deal – jurisdiction of the First-tier Tribunal to consider arguments based on legitimate expectation in the context of an appeal under s 83(1)(t) Value Added Tax Act 1994 – whether HMRC precluded from recovering tax wrongly credited under s 80 Value Added Tax Act 1994 on the basis of legitimate expectation – appeal allowed Heard on: 11 and 12 March 2026 Judgment date: 19 May 2026 Before THE HONOURABLE MRS JUSTICE JOANNA SMITH DBE JUDGE MARK BALDWIN Between QUEENSCOURT LIMITED Appellant and THE COMMISSIONERS FOR HIS MAJESTY’S REVENUE AND CUSTOMS Respondents Representation: For the Appellant:Charlotte Brown of counsel, instructed by PricewaterhouseCoopers LLP For the Respondents: Dilpreet Dhanoa of counsel, instructed by the General Counsel and Solicitor to HM Revenue and Customs DECISION Introduction

1. This appeal is about two issues. The first (which we refer to as the “VAT liability issue”) is whether the supply of a dip pot as part of a KFC takeaway meal deal is a separate zero-rated supply for VAT purposes, or whether it is part of a standard rated supply of hot food comprised within the meal deal. The second (which we refer to as the “public law issue”) is whether the First-tier Tribunal (the “FTT”) has jurisdiction to consider arguments based on legitimate expectation in the context of an appeal under section 83(1)(t) Value Added Tax Act 1994 (“VATA”) against a recovery assessment made under section 80(4A) VATA, and, if it does, whether it was sufficiently unfair for HMRC to resile from their initial acceptance of a claim made by the Appellant (“Queenscourt”) and to seek to reclaim the money they had paid to Queenscourt following their initial acceptance of that claim.

2. By way of background, until early 2019 Queenscourt accounted for VAT on the basis that the dip pots it supplied as part of a KFC takeaway meal deal formed part of a single standard rated supply. However, it subsequently decided that this was the wrong approach and that a meal deal should be treated as a multiple supply so that, where appropriate, component parts (such as coleslaw, cookies, yoghurts and milkshakes, as well as dip pots) could be zero rated if that is how they would be treated if they were sold on their own.

3. On 29 March 2019 Queenscourt submitted an error correction notice (“ECN1”) to reclaim the VAT it considered it had wrongly accounted for on meal deals between October 2015 and September 2018. This included £75,502 relating to dip pots included in takeaway meal deals. After some debate as to quantification, HMRC agreed to repay the VAT.

4. Queenscourt submitted a further error correction notice (“ECN2”) on 22 April 2020 to reclaim VAT which it had accounted for on items included in takeaway meal deals during the VAT periods 12/18 to 09/19. The claim included £30,936.64 related to dip pots. This claim was reviewed by a different HMRC officer, who accepted that cookies and yoghurts could be treated as separate supplies, as they could be consumed on their own, but considered that dip pots formed part of a single supply alongside the hot food that formed part of the relevant meal deal as, in his view, the dip pots were ancillary to the supply of the hot food and were simply a means of better enjoying that food. The officer also considered HMRC’s decision to repay the VAT relating to the dip pots supplied as part of meal deals following ECN1 to have been incorrect. HMRC therefore issued an assessment under section 80(4A) VATA to recover the amounts relating to dip pots which had been repaid following HMRC’s initial acceptance of ECN1.

5. Queenscourt appealed against the decision to refuse the repayment claimed in ECN2 and against the recovery assessments relating to ECN1. Its primary ground of appeal was that the supply of the dip pots as part of a takeaway meal deal was a separate zero-rated supply.

6. Queenscourt also argued that, if it was wrong on the VAT liability issue, HMRC were prevented from recovering the tax repaid following their acceptance of ECN1, based on either legitimate expectation or estoppel by convention. Queenscourt did not argue that HMRC were bound, based on legitimate expectation or estoppel by convention, to accept ECN2, even though the VAT that Queenscourt sought to reclaim related to periods before (and, of course, ECN2 was submitted before) the second HMRC officer reconsidered the VAT treatment of the dip pots.

7. In its decision (“the Decision”) released on 3 June 2024 with neutral citation [2024] UKFTT 460 (TC), the FTT decided that: (1) the supply of dip pots as part of a takeaway meal deal was part of a single standard rated supply of the hot food and dips; (2) whilst the Tribunal did have jurisdiction to consider arguments based on legitimate expectation in the context of an appeal under section 83(1)(t) VATA against a recovery assessment made under section 80(4A) VATA, HMRC’s decision to resile from their initial acceptance of the claim made in ECN1 was not so “outrageously unfair” that it should not be allowed to stand; and (3) HMRC were not estopped from making the recovery assessments as there had been no detrimental reliance on the original position taken by HMRC in connection with any subsequent mutual dealings.

8. Queenscourt obtained permission to appeal on the following grounds: (1) Ground 1(a): the FTT erred in law in concluding at paragraph [88] of the Decision that there is “no reason in principle” why two or more elements of a single transaction cannot constitute a single supply whilst, at the same time, other elements of the same transaction may constitute a separate supply. (2) Ground 1(b): the FTT erred in law in concluding at paragraphs [75] and [76] of the Decision that the element of choice bore less significance because the dip pots could not be obtained from a third party. (3) Ground 2: when deciding whether it was conspicuously unfair for HMRC to depart from its decision to repay the tax for the purposes of deciding Queenscourt’s legitimate expectation claim, the FTT erred in law by confining its analysis to weighing up the financial detriment suffered by Queenscourt and failing to consider whether permitting HMRC to resile from its previous decision to repay the VAT was not conducive to good administration. (4) Ground 3: the FTT erred in law at paragraph [230] of the Decision in concluding, in relation to the estoppel ground of appeal, that the detriment suffered by Queenscourt did not result from its reliance on the expressly shared common assumption in its subsequent mutual dealings with HMRC.

9. Queenscourt abandoned Ground 3 before the substantive hearing of its appeal against the Decision.

10. HMRC contend that the FTT does not have jurisdiction to determine Queenscourt’s legitimate expectation claim in an appeal under section 83(1)(t) VATA. As required by rule 24(1B) of the Tribunal Procedure (Upper Tribunal) Rules 2008, HMRC gave notice, in its Response to the Notice of Appeal, of its intention to ask the Upper Tribunal (“Upper Tribunal”) to uphold the FTT’s decision on Queenscourt’s legitimate expectation claim for this reason.

11. As a result, there are two live issues before us: (1) the VAT liability issue (Grounds 1(a) and 1(b)); and (2) the public law issue (Ground 2 and HMRC’s assertion of the FTT’s lack of jurisdiction). The facts

12. The facts relevant to the appeals were not in dispute and were set out at paragraphs [33]-[51] of the Decision. We set out the key facts relevant to the VAT liability issue below. The facts relevant to the public law issue are summarised at [2]-[4] above.

13. The menu at the KFC outlets operated by Queenscourt offers items to be purchased individually but also offers various meal deals which bundle together popular items at a discounted price. By way of example, the “boneless banquet” includes three boneless mini fillets, a small portion of popcorn chicken, fries and the customer’s choice of a side, drink and a dip pot. In December 2022, the cost of the boneless banquet was £7.99. If the items were all purchased separately, they would cost £12.43.

14. Customers who buy a meal deal are not obliged to receive a dip pot. They are free to say that this is something they do not want, even though it is included in the price. However, even if they decline the dip pot, they may receive one in any event as the packagers do not always take note of the fact that one is not needed. In any event, there is no reduction in the purchase price of the meal deal if the dip pot is declined.

15. Not all the meal deals include a dip pot. For example, the “mighty bucket for one” does not do so. Some meal deals also include other cold items such as coleslaw (which is one of the choice of sides), cookies or yoghurts.

16. All items comprised in a meal deal are available for purchase separately. The dip pots cost 40p per pot at the time of the FTT hearing. The dip pots are larger than the sachets of ordinary sauces such as ketchup which are provided free of charge.

17. Dip pots are, therefore, available to be purchased on their own, as part of a meal deal or as part of an order comprising several individual items from the menu which do not form part of a meal deal. The FTT’s Decision on the VAT Liability Issue

18. The FTT considered the principles to be applied in determining the VAT liability issue at paragraphs [53]-[89] of the Decision and then at [90]-[116] it applied the principles, as it found them, to determine the VAT liability of the supply of dip pots as part of a meal deal.

19. Having noted that the sale of a portion of takeaway chicken by one of Queenscourt’s KFC outlets (being hot food) is standard rated for VAT purposes, whilst the sale of a dip pot (being cold food) is zero rated, it identified the issue it needed to resolve as being whether this remains the case when the two items are purchased together (along with other items) as part of a takeaway meal deal.

20. At [55] it noted that the case law from the European Court of Justice (“ECJ”) makes it clear that, whilst the starting point is that every supply should normally be regarded as distinct and independent, there are circumstances where the supply of more than one item in a single transaction may be treated as a single supply which takes its VAT treatment from whichever part of the supply predominates. That case law has identified two situations where a supply which contains one or more elements as part of a single transaction may be treated as a single supply.

21. The first situation was considered by the ECJ in Card Protection Plan Limited v Customs and Excise Commissioners (case C-349/96) (“CPP”) and arises where one element constitutes the principal service to which all other elements are ancillary.

22. The second situation was considered by the ECJ in Levob Verzekeringen BV v Staatssecretaris van Financien (Case C-41/04) (“Levob”) and arises where two or more elements or acts supplied by the taxable person to the customer are so closely linked that they form, objectively, a single indivisible economic supply, which it would be artificial to split.

23. The FTT recorded that it was common ground that, if there is a single supply in this case, it must be on the basis that the supply of the dip pots as part of the meal deal is ancillary to the supply of other items contained within the meal deal, so it is the principles set out in CPP rather than Levob which are relevant.

24. The FTT discussed the decision of the Upper Tribunal in HMRC v The Honourable Society of Middle Temple, [2013] UKUT 0250 (TCC) (“Middle Temple”), in which the Upper Tribunal reviewed both CPP and Levob, as well as a number of subsequent decisions of the ECJ, and, based on all of those decisions, set out at [60] a number of key principles to be applied in determining whether a particular transaction should be regarded as a single composite supply or as several independent supplies.

25. The FTT discussed two of those principles in detail. The first was the relevance of the “typical consumer”, which the FTT considered at [64]-[67]. The FTT concluded that it must be relevant to consider whether the typical consumer of a meal deal would view the supply of a dip pot as ancillary to the supply of other elements within the meal deal such as the chicken. That question was “highly fact sensitive” and all relevant circumstances must be taken into consideration.

26. The second was the relevance of choice, which the FTT considered at [68]-[76]. The FTT held (at [75]) that what is important is not whether the customer wanted to receive a particular element of the supply at all, but whether they could realistically receive it from a third party. As the dip pots were only available from KFC outlets, the FTT considered (at [76]) that, whilst the element of choice as to how to obtain the dip pots from a KFC outlet was a relevant factor in determining whether the supply of the dip pots was ancillary to other items received as part of a meal deal, it was not as important as it would have been if the dip pots could realistically be refused by a consumer and obtained elsewhere from a third party.

27. The FTT considered Queenscourt’s submission that it was not open to HMRC to argue that dip pots forming part of a meal deal are part of a single standard rated supply of hot food in circumstances where they accept that other items which form part of a meal deal, such as coleslaw, cookies and yoghurts, are separate elements of a multiple supply. The FTT described this submission as being that a meal deal must either be a single supply in its entirety or, if it is not, then each item comprised in the meal deal must be a separate supply. HMRC argued that there is no inconsistency between accepting that the meal deal is not a single supply and holding that two or more of the elements within that meal deal are treated as a single supply on the basis that one is ancillary to the other.

28. The FTT observed at [80] that neither party was able to refer it to any authority which shed light on this question, and so it decided that it would approach the question from first principles. At [83] it recorded its conclusion that the ECJ’s comments in CPP at [29] and [30] indicate that it is perfectly possible for some elements of a single transaction to be principal/ancillary supplies (and therefore to constitute a single supply) whilst there may be other elements of the same transaction which (based on the starting point that all elements of a supply are separate) do not form part of that single supply as they are distinct and independent.

29. At [84] the FTT went on to hold that, given the starting point that each element of the supply is separate, it must follow that it is necessary to look separately at each element of the supply in order to determine, applying the principles developed by the ECJ, whether that element stands alone or whether it should be treated as part of a single supply along with one or more (but not necessarily all) other elements of the transaction.

30. At [85] the FTT acknowledged that the cases it had been referred to envisage that either all elements of the transaction will constitute distinct principal supplies or will alternatively be treated as one single supply, but at [86] it observed that there was no suggestion in any of those cases that the Court was being asked to consider the possibility that two or more elements of a transaction might together form a single supply whilst other elements are treated as separate supplies. In that context, the FTT thought it was worth noting that the ECJ in CPP cautioned at [27] that “having regard to the diversity of commercial operations, it is not possible to give exhaustive guidance on how to approach the problem correctly in all cases.” The FTT thought that there was some support in Field FisherWaterhouse LLP v HMRC (Case C-392/11) (“FieldFisher”) at [24]-[25] for the proposition that some elements of a single transaction may constitute a single supply whilst other elements may be separate.

31. This reasoning led the FTT to its conclusion (at [88]) that there is no reason in principle why two or more elements of a single transaction cannot constitute a single supply whilst, at the same time, other elements of the same transaction constitute a separate supply. So (at [89]) it held that the fact that HMRC accepted that some elements of a meal deal were a separate supply did not mean that the dip pots must also form a separate and distinct supply on their own. Instead, the FTT held that it was necessary to consider whether, in the context of a meal deal, the dip pots were ancillary to the supply of one or more of the items of hot food contained in the meal deal.

32. Starting at [90] the FTT turned to the characterisation of the supply of dip pots. Save for the question of the weight to be attached to the relevance of choice (as to which see [26] above), Queenscourt makes no criticism of the FTT’s analysis of whether, if the two items were looked at in isolation, the supply of dip pots would be ancillary to the hot food with which they are supplied. Its fundamental criticism is that this is not an exercise the FTT should have embarked upon. The FTT’s Decision on the public law issue

33. The FTT began its discussion of the jurisdiction question by noting the uncertainty that prevails in this area, despite the issue having been considered in numerous previous decisions.

34. At [131] the FTT set out three general principles which emerge from the authorities, including that whether the FTT has jurisdiction to consider public law arguments such as legitimate expectation is a question of interpreting the relevant statutory provisions.

35. At [138] the FTT noted that a “significant development” in the approach to the task of statutory interpretation in this area is to be found in the decision of the Court of Appeal in David Beadle v HMRC, [2020] EWCA Civ 562 (“Beadle”). Pre-Beadle courts and tribunals had generally (but not always) started from the perspective that, if Parliament had intended the FTT to be able to consider arguments based on public law, the relevant provisions should be expected to make this clear

36. In Beadle the Court of Appeal held that, in the context of an enforcement action by a public body, the person against whom that action is taken may rely on public law arguments to defend themselves unless such a right is excluded by the relevant legislation either by express words or by necessary implication. From this the FTT concluded that the starting point in the statutory interpretation exercise is not whether the relevant statutory provisions confer an ability on the court or tribunal to consider public law arguments, but is instead whether the statutory scheme excludes such rights.

37. The FTT considered that the decision in Beadle led the Upper Tribunal in KSM Henryk Zeman Sp Z.o.o. v HMRC, [2021] UKUT 182 (TCC) (“Zeman”), to conclude that its starting point in deciding whether the FTT could consider a legitimate expectation argument when deciding an appeal under section 83(1)(p) VATA was whether that ability was excluded (expressly or by implication).

38. The FTT held that the principle in Beadle only applies where the proceedings in question concern “enforcement action”. Following Zeman, the FTT accepted that the raising of an assessment by HMRC constitutes enforcement action for this purpose and that, in appealing against the assessment, a taxpayer is seeking to defend an enforcement action, so that the principles set out in Beadle are engaged.

39. The FTT noted that the statutory provisions relevant to this appeal were materially identical to the statutory provisions considered in Zeman and that HMRC had not identified any specific flaw in the Upper Tribunal’s reasoning in Zeman, other than taking the principles set out in Beadle as its starting point. The FTT considered that, although it was not bound by Zeman, it should follow the reasoning in Zeman in determining whether there is anything in the relevant legislation which excludes the FTT’s jurisdiction to consider an argument based on legitimate expectation. It concluded (at [175]) that there is nothing in the statutory scheme of section 80(4A) and section 83(1)(t) VATA which either expressly or implicitly excludes the FTT’s jurisdiction to consider public law arguments and, in particular, arguments based on legitimate expectation in relation to HMRC’s decision to make the recovery assessments.

40. Although this was the FTT’s conclusion, it recorded (at [176]) that it reached that conclusion with some hesitation. The FTT’s own view, in line with the authorities prior to Beadle, was that, for the reasons explained in those cases, it would be surprising if Parliament intended to confer on the FTT an ability routinely to consider arguments based on public law grounds in the context of appeals under section 83(1) VATA.

41. Turning to the question of whether Queenscourt had a legitimate expectation in this case, the FTT held that it did not. It accepted that Queenscourt had put all its cards face upwards on the table and that the ruling or statement relied upon was clear, unambiguous and devoid of relevant qualification.

42. On the question whether it would be fair to allow HMRC to depart from that clear and unambiguous ruling, the FTT began by noting that the correction of a mistake by HMRC is not, on its own, sufficient to reach the necessary threshold of unfairness. What the FTT needed to do, once it had decided that a taxpayer had a particular legitimate expectation, was to weigh the requirements of fairness against any overriding interest relied upon for the change of policy. In that context it accepted HMRC’s submission that it needed to consider the extent and nature of any detriment which may have been caused to Queenscourt as a result of its reliance on the ruling made by HMRC and weigh this in the balance in determining whether there was a sufficient element of unfairness such that it would be wrong for HMRC to be permitted to depart from that ruling.

43. Although the FTT accepted that detrimental reliance was not a requirement, it noted that Queenscourt did not suggest that there was anything which would reach the required level of unfairness other than the fact that Queenscourt had to repay the VAT which it thought it had successfully reclaimed.

44. The FTT examined the evidence relating to the extent of the detriment suffered by Queenscourt itself (as opposed to the wider corporate group of which it was a member). It described this evidence as “sketchy”.

45. The FTT concluded that, whilst Queenscourt may have suffered some detriment to its business as a result of having to repay the VAT which had been reclaimed, it had not shown that there had been, or was likely to be, any serious detriment to it which resulted solely or primarily from HMRC making the recovery assessments.

46. At [215] the FTT said that this conclusion needed to be balanced against HMRC’s duty to collect the right amount of tax. The FTT considered that, if HMRC’s error were not corrected, Queenscourt would be put in a better position than other taxpayers in respect of whom HMRC had applied the law correctly, and noted that the Court of Appeal in Samarkand Film Partnership No 3 v HMRC, [2017] EWCA Civ 77,considered that there was a strong public interest in ensuring that no individual taxpayer is unfairly advantaged at the expense of other taxpayers. Further, at [216] the FTT expressed the view that Queenscourt must also bear some responsibility for HMRC’s mistake. Queenscourt’s Submissions

47. On Ground 1(a) Queenscourt submits that the FTT made a fundamental error of principle when it held (at [88]) that there was no reason in principle why two or more elements of a single transaction cannot constitute a single supply whilst, at the same time, other elements of the same transaction may constitute a separate supply. Queenscourt says this for the following reasons: (1) All the authorities treat supplies containing multiple elements as either a single or a multiple supply. (2) Where a transaction is a multiple supply, each component element must be taxed at its own rate. The only exception (not relevant here) is where legislation specifically provides that a specific component in a single composite supply may be taxed at a different rate; Talacre Beach Caravan Sales Ltd v HMRC (Case C-251/05) (“Talacre”). (3) HMRC has already accepted that the meal deal is a multiple supply. Each component part must therefore be taxed at its own rate. (4) The FTT misinterpreted the references to the plural “elements” in paragraph [30] of the ECJ judgment in CPP as indicating that within a single supply there may be some elements which are separate and distinct and some which are principal/ancillary supplies. There is no suggestion in CPP that a single supply can contain some distinct and independent components and some principal/ancillary components. Nor does CPP suggest that some of the distinct and independent components of a multiple supply can be conflated into a single supply within that multiple supply. (5) The FTT was wrong to read the ECJ’s refusal to give exhaustive guidance owing to the diversity of commercial operations (at paragraph [27] of CPP) as supporting its conclusion. This comment was directed at the factors to be considered when determining whether a supply is a single or multiple supply and has nothing to say about whether elements within a multiple supply can be combined. (6) The FTT was also wrong to read the CJEU’s comments in Field Fisher at [24] as suggesting that some elements of a single transaction may constitute a single supply whilst other elements may be separate.

48. On Ground 1(b) Queenscourt says that the FTT’s (wrong) conclusion at [76] was based on the understanding that the relevant case law dealt with fact patterns where the item in question was available from a third party. However, the underlying factual matrix is irrelevant to the legal principle. In addition, even though the FTT at [76] accepted that choice was a relevant consideration, at [111] it (inconsistently) held that the ability to buy a dip pot from the same KFC outlet in a different way was “irrelevant”.

49. On the FTT’s jurisdiction to entertain public law arguments, Queenscourt’s position is that the FTT did not err in law and was right to conclude, for the reasons it gave, that it had jurisdiction to consider the legitimate expectation claim.

50. In short, Ms Brown submits that the correct legal position is that the FTT may have jurisdiction to consider appeal grounds based on public law arguments depending upon the statutory provisions under consideration. As the legislation is permissive not mandatory, it affords HMRC some discretion and therefore the FTT does have such a jurisdiction in relation to this appeal.

51. On Ground 2, Queenscourt asserts that the FTT erred in law when it rejected the legitimate expectation claim because it confined its analysis of whether there had been sufficient unfairness to weighing up the financial detriment suffered by Queenscourt and failed to consider the question whether permitting HMRC to resile from its previous decision to repay the VAT was conducive to good administration or not, which is a relevant factor. HMRC’s Submissions

52. On Ground 1(a), HMRC accept that the correct starting point is that every supply should normally be regarded as distinct and independent. The single composite supply doctrine is an exception to this approach.

53. In paragraph [29] of its judgment in CPP, the ECJ referred to “several distinct principal supplies”. HMRC submit that this language and approach (which were reflected by the Upper Tribunal in Middle Temple at [60]) are “fatal” to what HMRC describe as Queenscourt’s “all or nothing” approach. HMRC say that the ECJ here expressly contemplates the possibility of there being more than one principal supply in a complex transaction, and so it envisaged the principal/ancillary analysis being applied at the level of individual elements in the overall transaction. CPP does not, HMRC submit, lay down a taxonomy of multi-element transactions where there is either one supply or as many supplies as there are elements.

54. HMRC accept that the language used by the ECJ in CPP, and in other cases, naturally reads as operating at the level of the transaction as a whole, but HMRC say there is nothing in the language used which states that the analysis cannot be applied to parts of the whole. HMRC contend that the opinion of the Advocate General in Frenetikexito–Unipessoal Lda v Autoridade Tributária e Aduaneira (Case C-581/19) (“Frenetikexito”) supports this position.

55. HMRC also rely upon paragraphs [41] and [46] of the decision of the Court of Appeal in HMRC v Gray & Farrar International LLP, [2023] EWCA Civ 121 (“Gray & Farrar”), and paragraphs [17]-[20] of Field Fisher.

56. HMRC contend, as the FTT held, that the observations of the ECJ in CPP at [27] (referred to above) leave open the question of multi-element transactions being analysed in ways that are different to the approach adopted in CPP.

57. If it is accepted that the analysis is not a binary one, then the question of how many supplies are to be found in a multi-element transaction becomes fact sensitive and the factors used to analyse whether a transaction constitutes a single supply or a number of supplies can be deployed below the level of the transaction as a whole to consider whether some (but not all) of the elements in the transaction constitute a single supply.

58. On the element of choice, HMRC submit that the question is not whether a customer wants to receive a particular element of a supply at all, but whether they can realistically receive it from a third party.

59. Choice is relevant only in so far as it illuminates whether a particular element is an independent object of consumption. The fact that a consumer may choose which dip pot to take, or choose whether to add a dip pot, does not of itself establish that a dip pot supplied in a meal deal is a separate supply. Many ancillary elements in composite supplies may be selectable without ceasing to be ancillary.

60. Equally, the fact that an item can be purchased separately in other circumstances does not compel the conclusion that it is a separate supply when included within a composite product. A component can be a freestanding supply in one context but ancillary in another; the proper characterisation is context-sensitive.

61. On the question of the FTT’s jurisdiction to entertain legitimate expectation arguments, HMRC agree that the question whether (and to what extent) public law arguments may be entertained depends on the construction of the relevant statutory appeal provisions, although HMRC submit that the question is one of statutory conferral, not the absence of ‘exclusionary’ language. The correct question is therefore whether section 83(1)(t) (read with section 84 and the wider scheme) confers jurisdiction to determine a freestanding public law case.

62. Public law challenges to HMRC’s conduct and decision-making are ordinarily justiciable by judicial review, subject to well-understood safeguards (a permission filter, time limits, and judicial review remedial discretion). It would cut across that allocation if freestanding public law challenges could routinely be litigated alongside ordinary VAT assessment appeals in the FTT without those safeguards. The structure of the tribunal system (including the Upper Tribunal’s specific public law jurisdiction in defined circumstances) points away from an implied general public law jurisdiction in the FTT under section 83(1)(t).

63. Thus, submit HMRC, neither Beadle nor Zeman assist here: (1) Beadle is concerned with a different problem (collateral challenge in a ‘gap’ context) and does not supply the starting point where the decision under challenge is itself subject to a statutory appeal; (2) Zeman should not be followed on this issue – its jurisdictional analysis wrongly treated Beadle’s ‘exclusionary’ reasoning as the starting point in a statutory appeal regime, contrary to the orthodox requirement of statutory conferral.

64. On Ground 2, HMRC submit that the Tribunal must weigh: (1) the fairness to Queenscourt in holding HMRC to the representation (including the nature of any reliance and consequences for the taxpayer), against (2) the public interest in the correct application of the law, consistent treatment of taxpayers, and the orderly administration of the tax system.

65. “Good administration” is not a free-standing requirement; it is the underpinning rationale for the doctrine and is addressed through the balancing exercise. It embraces (amongst other things) certainty, predictability, consistency, transparency, and the avoidance of arbitrary shifts of position. But it also includes the public interest in ensuring that the correct tax is collected and that mistakes – particularly mistakes conferring a windfall – can be corrected by proper statutory means.

66. The FTT’s balancing exercise expressly engaged with the core components of good administration in tax: the duty to collect the correct tax and to apply the law equally, so that one taxpayer is not unfairly advantaged. That is the institutional ‘good administration’ point in this context. Further Written Submissions

67. During the hearing we raised with the parties a passage in the judgment of Nolan LJ in Bophuthatswana National Commercial Corp Ltd v CCE, [1993] STC 702 (“Bophuthatswana”). BNCC was established and funded by the government of Bophuthatswana to carry on such activities on a non-profit-making basis as might have been carried on by a diplomatic mission if the government had been recognised internationally. The Commissioners assessed BNCC to value added tax on the basis that it had made a single standard-rated supply of diplomatic services. BNCC appealed to a value added tax tribunal contending that it had supplied several separate services to the government, some or all of which were zero-rated. At p708 Nolan LJ said: “ … [A]lthough there may be only a single commercial relationship between BNCC and the government of Bophuthatswana, the individual supplies of goods and services in the course of that relationship appear to vary widely both in nature and in taxability or potential taxability. It cannot be right in my judgment to cast over them a blanket label 'services of the sort ordinarily provided by a diplomatic mission' and to conclude that, since this label does not appear in the relieving provisions, the whole of the services must be charged at the standard rate. It is essential, to my mind, to analyse the individual supplies of goods and services by reference to the specific taxing and relieving provisions of the 1983 Act, as a preliminary to deciding whether any of them are no more than ancillary or incidental to another or others, and to determining whether and if so how the money paid by the Bophuthatswanan government should appropriately and fairly be apportioned between them.” (emphasis added)

68. The words we have underlined in that passage might be read as suggesting that Nolan LJ thought that the various elements of the services supplied by BNCC could be analysed in a way which resulted in BNCC making a number of supplies some (or all) of which comprised more than one element. As the parties were not on notice of this point, we invited them to make further submissions.

69. Ms Brown submits that the parties in Bophuthatswana did not suggest that the supply was anything but a single or a multiple supply and the Court of Appeal was not being asked to rule outside of this binary choice. It was simply setting out a pre-CPP analysis of whether a single transaction involving multiple supplies could have one tax treatment under a blanket label or each supply should be treated separately and taxed at its own rate. In any event, the alternative approach was not adopted by the subsequent binding authorities, starting with CPP, so any other reading of the Court of Appeal judgment would not be relevant.

70. Ms Brown contends that the plural “others” is used in this passage because there may be more than one ancillary element to a principal supply and/or more than one element making up the principal supply, depending on the number of elements of the transaction. For example, in De Montfort (see below) it was argued that both the drink and the crisps were ancillary to the sandwich. In Bophuthatswana, more than one of the 8 elements might have fallen within a principal supply of “diplomatic services” but the remaining elements would all have to be ancillary to the diplomatic services for there to be a single supply of diplomatic services.

71. Ms Dhanoa submits that Bophuthatswana pre-dates CPP and Levob. She says that it is cited only as an authority militating against overly unitary “single transaction” labelling and in favour of an objective assessment of all the elements and circumstances of the transaction (including, where relevant, ancillary/incidental and apportionment issues), tested against the economic reality. HMRC rely on Bophuthatswana only to support its position that Queenscourt’s proposed ‘monolithic’ binary rule is not compelled as a matter of domestic VAT analysis.

72. We also asked the parties to reflect on the VAT Tribunal decision in De Montfort University Students’ Union v HMRC, [2003] Lexis Citation 984 (“De Montfort”), as that case also concerned a “meal deal”, which comprised a zero-rated sandwich, a standard rated drink and a standard rated packet of crisps supplied together for a discounted aggregate price. The taxpayer claimed it was a single, zero-rated supply with the sandwich being the principal element and the drink and crisps being ancillary, applying the analysis in CPP. HMRC argued that it was a multiple supply and that each component part should be taxed separately. The Tribunal found that it was a multiple supply as the crisps and drink were not a means of better enjoying the sandwich.

73. Ms Brown submits that the VAT Tribunal considered paragraphs [26]-[32] of CPP before going on to state at [7] that, “ultimately, it is a question of fact whether there are two or more supplies, each with its own tax treatment, or a single supply whose treatment is governed by that of the principal element, all others being regarded as ancillary” (her emphasis). This confirms the binary approach Queenscourt submits is correct. The Tribunal went on to state at [10] that, “[e]ven if-which we doubt-it could properly be said (in the sense intended by the Court of Justice) that a drink is a means of “better enjoying” a sandwich, perhaps by making it easier to swallow, the same cannot possibly be said of a packet of crisps”. Just as both the drink and crisps needed to be ancillary to the sandwich in De Montfort for it to be a single supply, the cookie, yoghurt and dip pot elements of the takeaway meal deal in this case would need to be ancillary to the hot chicken for there to be a single standard rated supply.

74. Ms Dhanoa submits that De Montfort does not itself grapple with the precise question raised here (whether a wider transaction can contain both a composite supply and separate supplies). It considered whether the taxpayer was supplying: (i) a sandwich with ancillary items; or, (ii) three supplies at a discounted aggregate price. It cannot be relied upon as direct authority either way on the ‘two or three supplies’ issue. Its value is instead as a reminder of: (i) the CPP fact-sensitivity point; and, (ii) the need to resist ‘artificial’ characterisations driven by rating outcomes, themes which run through the CPP jurisprudence and which the FTT deployed. For completeness, HMRC note that there is nothing conceptually impermissible about analysing a transaction at an appropriate level of economic reality: even a "packet of crisps” involves elements (such as packaging) which are integral/ancillary to the crisps and do not constitute separate consumer aims.

75. In their further written submissions HMRC also referred us to the CJEU decision in Dyrektor Krajowej Informacji Skarbowej v P. in W. (Case C-282/22) (“P in W”). The taxpayer in that case operated electric vehicle recharging stations. The supply provided during each recharging session could, in principle, include, depending on the needs of the user concerned, access to recharging devices, including integration of the charger with the vehicle operating system, the supply of electricity to the batteries of the vehicle, and the necessary technical support. The taxpayer also planned to create a special platform, a website or an IT application which would enable a user to reserve a particular connector and to view his or her transaction and payment history. For all those supplies, P in W billed a single price. The local court held that the 3 elements of the transaction at a recharging station were so closely linked that they constituted a single complex transaction. By contrast, the special platform would constitute an ancillary supply forming part of that single complex transaction.

76. At [27]-[30] the CJEU summarised the familiar principles to be derived from the ECJ/CJEU caselaw in this area and then (at [32]) commented that “it does not appear that that classification [adopted by the referring national court] disregards any of the criteria set out in paragraphs 27 to 30 of the present judgment”. HMRC submit that the CJEU treated the transaction as, in principle, a single complex supply with dependent ancillary elements. HMRC say that this analysis is consistent with CPP and provides a modern worked example of the principal/ancillary structure in a complex transaction.

77. Although we raised Bophuthatswana and De Montfort with the parties, on reflection we agree that they are of limited value when it comes to deciding Ground 1(a). Bophuthatswana predates CPP and Nolan LJ’s comments need to be read in the light of the issue he was concerned with: whether the various supplies of goods and services should be given a single “blanket” description. He was not focused on the issue that confronts us. Similarly, in De Montfort the tribunal was addressing whether there was a single supply or whether every item in the meal deal should be taxed separately. As Ms Brown commented, there is some incidental support for her position in the way the tribunal framed the issue, but there was no detailed analysis of the question before us and it would be wrong to read too much into this. Discussion The VAT Liability Issue Ground 1(a) Must each supply in a multi-element transaction be analysed separately unless the transaction is a single supply?

78. As the FTT observed, whilst there are many decisions, at both domestic court/tribunal and ECJ/CJEU level, analysing the circumstances in which a transaction comprising several elements should be treated as a single supply for VAT purposes, none of those decisions directly addresses the issue we are concerned with, which is whether, in the context of a multi-element transaction which is not a single supply for VAT purposes, it is possible to isolate two or more (but not all) of the elements of that transaction and conclude that, viewed in isolation, they constitute a single supply, so that the number of supplies for VAT purposes is greater than one but less than the number of elements in the transaction. HMRC say that the answer to that question is “yes”, and the FTT (at [88]) agreed with them, whereas Queenscourt says that the answer is “no”. CPP

79. We start our discussion of this issue with the decision of the ECJ in CPP. CPP offered holders of credit cards a plan intended to protect them against financial loss and inconvenience resulting from the loss or theft of their cards or of certain other items such as car keys, passports and insurance documents. The plan incorporated several (at least 14) elements, including insurance elements. The dispute between CPP and the HM Customs and Excise (now HMRC) turned on whether the plan was a single supply or several supplies and the extent to which any supply or supplies amounted to the provision of insurance.

80. At [26] the ECJ framed the question (so far as relevant to us) as “whether a transaction which comprises several elements is to be regarded as a single supply or as two or more distinct supplies to be assessed separately”.

81. At [29] the ECJ set out two guiding principles. It said: “In this respect, taking into account, first, that it follows from Article 2(1) of the Sixth Directive that every supply of a service must normally be regarded as distinct and independent and, second, that a supply which comprises a single service from an economic point of view should not be artificially split, so as not to distort the functioning of the, VAT system, the essential features of the transaction must be ascertained in order to determine whether the taxable person is supplying the customer, being a typical consumer, with several distinct principal services or with a single service.”

82. There is, as the Advocate General pointed out in Frenetikexito, a tension between these principles.

83. At [30] the ECJ stated that: “There is a single supply in particular in cases where one or more elements are to be regarded as constituting the principal service, whilst one or more elements are to be regarded, by contrast, as ancillary services which share the tax treatment of the principal service. A service must be regarded as ancillary to a principal service if it does not constitute for customers an aim in itself, but a means of better enjoying the principal service supplied Joined Cases C-308/96 and C-94/97 Commissioners of Customs and Excise v Madgett and Baldwin [1998) ECR 1-6229, paragraph 24.”

84. We agree with Ms Brown that the ECJ is here addressing itself to a situation where there is one principal service, although it may comprise more than one element, to which all the other services are ancillary. The opening words of [30] and the closing words of [29] make this clear and it is the only way in which one tax treatment (“the tax treatment of the principal service”) for the whole transaction would emerge. There is nothing here to support the idea that the principal/ancillary service analysis can produce more than one principal service or to suggest that that analysis has any place outside the context of deciding whether a multi-element transaction comprises one supply/service or not. What did the ECJ mean by “several distinct principal services” in [29]?

85. We consider that it is important to understand this expression in the context of the overall ECJ judgment, which is focused on deciding “what the appropriate criteria are for deciding, for VAT purposes, whether a transaction which comprises several elements is to be regarded as a single supply or as two or more distinct supplies to be assessed separately” (see [26]). The questions the UK court asked the ECJ were about how to decide whether there was “a single composite supply or two or more supplies”. Beyond that, the ECJ was never asked to think about how to work out how many supplies there were. The ECJ answers the question at [32], when it says: “The answer to the first two questions must therefore be that it is for the national court to determine, in the light of the above criteria, whether transactions such as those performed by CPP are to be regarded for VAT purposes as comprising two independent supplies, namely an exempt insurance supply and a taxable card registration service, or whether one of those two supplies is the principal supply to which the other is ancillary, so that it receives the same tax treatment as the principal supply.” The intervening paragraphs ([27]-[31]) discuss those criteria, but we consider that it is important to keep this context in mind when looking at the language used here. The ECJ is not exploring how many supplies there are. It has taken the 14+ elements and distilled them into two supplies (card registration and insurance), so the answer is only ever going to be one (if one supply is the principal supply and the other is ancillary) or two (if this is not the case).

86. The ECJ has set up a clear dichotomy here between one supply, on the one hand, and more than one supply, on the other, but it does not always articulate this dichotomy in quite the same way, as is clear from a close reading of [26], [29] and [32].

87. The ECJ contrasts a single supply/service with distinct supplies/distinct principal services/independent supplies. In all cases a single supply/service is being contrasted with the same thing (more than one supply/service). Given that the ECJ never suggested that anything turned on these slight linguistic differences and remembering that the analysis the ECJ pointed to in [26] and [32] was of “transactions” and that the principal/ancillary concept did not make an appearance in the judgment before paragraph [30], we consider that it is reading too much into the words “distinct principal services” in [29] to read them as a considered invitation to carry out a principal/ancillary supply analysis below the level of the totality of the transaction. What did the ECJ mean when it said in [29] that “every supply of a service must normally be regarded as distinct and independent”?

88. In the context of a multi-element transaction, we need to ask whether the ECJ meant that every element (here meaning everything which would be a supply if provided on its own) in the transaction should be analysed separately or that every supply in a technical VAT sense (including treating as a single supply elements which, supplied together in isolation, would amount to a composite supply) to be found in a multi-element transaction should be treated separately? This is clearly important because, if the ECJ intended to convey the second meaning, that would support HMRC’s case.

89. At [81] the FTT said that “it is clear from the ECJ authorities that the starting point is that every element of a transaction is a separate supply (see for example Card Protection Plan at [29])”. We are not sure that every element of a transaction being a separate supply can be distilled from the words used in CPP. At [29] in CPP the ECJ simply says that every supply should be regarded as distinct. Its distillation of 14+ elements into two supplies at [32] and its comment (at [30]) that a principal service might comprise more than one element rather suggest that it might not have seen the concepts of element and supply as necessarily always being the same. The ECJ never explained how it reduced 14+ elements to two supplies and what (if anything) it thought turned on that or whether these were just convenient labels, particularly given that the real issue in CPP was whether what was supplied could be analysed as a single insurance supply.

90. Another example of an imprecise boundary is at [26], where the ECJ says that the question to be asked is “whether a transaction which comprises several elements is to be regarded as a single supply or as two or more distinct supplies to be assessed separately”. “Transaction” clearly refers to the overall commercial transaction, the “bundle of features and acts” being analysed (CPP at [28] referring to Case C-231/94 Faaborg-Gelting Linien v Finanzamt Flensburg (Case C-231/94) (“Faaborg-Gelting”)), and “supply” clearly refers to the VAT concept of supply. What is less clear is whether “element” refers to something which would be a supply if provided on its own or whether it has a less precise meaning.

91. Although there may be some imprecision in the words the ECJ used, the most natural reading of “every supply of a service must normally be regarded as distinct and independent” in the light of the very limited exception (where a transaction is analysed as amounting to a single supply) is that every element (every item which would be treated as a supply if provided separately) in a multi-element transaction is analysed separately unless the transaction as a whole is a single composite transaction. Subsequent articulations of this principle

92. Subsequent cases bear out this conclusion. In Purple ParkingLtd v HMRC, (Case C-117/11) (“PurpleParking”), the Court said (at [26]): “According to settled case-law, it follows from Article 2 of the Sixth Directive that every supply must normally be regarded as distinct and independent. However, a transaction which comprises a single supply from an economic point of view should not be artificially split, so as not to distort the functioning of the VAT system …”.

93. The Court considered the circumstances in which “several formally distinct services, which could be supplied separately and thus give rise, separately, to taxation or exemption, must be considered to be a single transaction when they are not independent”. It said: “27. Furthermore, in certain circumstances, several formally distinct services, which could be supplied separately and thus give rise, separately, to taxation or exemption, must be considered to be a single transaction when they are not independent (see Ministero dell'Economia e delle Finanze v Part Service Srl (Case C-425/06) [2010] BVC 443 ; [2008] ECR I-897, paragraph 51; RLRE Tellmer Property, paragraph 18; Don Bosco Onroerend Goed BV v Staatssecretaris van Financiën (Case C-461/08) [2010] BVC 1,084 ; [2009] ECR I-11079, paragraph 36; and Everything Everywhere, paragraph 23).

28. Such is the case particularly where one or more elements are to be regarded as constituting the principal supply, while other elements are to be regarded, by contrast, as one or more ancillary supplies which share the tax treatment of the principal supply. In particular, a supply must be regarded as ancillary to a principal supply if it does not constitute for customers an end in itself but a means of better enjoying the principal service supplied (see, inter alia, CPP, paragraph 30; C & E Commrs v Primback Ltd (Case C-34/99) [2001] BVC 315 ; [2001] ECR I-3833, paragraph 45; RLRE Tellmer Property, paragraph 18; Everything Everywhere, paragraphs 24 and 25; and Bog, paragraph 54).

29. Further, there is a single supply where two or more elements or acts supplied by the taxable person to the customer are so closely linked that they form, objectively, a single, indivisible economic supply, which it would be artificial to split (see LevobVerzekeringen, paragraphs 22 and 30; Aktiebolaget NN, paragraph 23; Part Service, paragraph 53; RLRE Tellmer Property, paragraph 19; Don Bosco Onroerend Goed, paragraph 37; and Bog, paragraph 53)”.

94. As we understand these passages, the Court is explaining (in more familiar language) that: (1) The rule of general application is that supplies are treated independently; (2) The exception is where the elements (“formally distinct services, which could be supplied separately”) in a multi-element transaction (a “bundle of features and acts” to use the expression in Faaborg-Gelting adopted in Levob at [19]) fall to be treated as a single supply.

95. We see the same language in Field Fisher at [14]-[15]: “14. It should be recalled, as a preliminary point, that for VAT purposes every supply must normally be regarded as distinct and independent, …

15. Where, however, a transaction comprises several elements, the question arises whether it is to be regarded as consisting of a single supply or of several distinct and independent supplies which must be assessed separately from the point of view of VAT. According to the Courtʼs case-law, in certain circumstances several formally distinct services, which could be supplied separately and thus give rise, in turn, to taxation or exemption, must be considered to be a single transaction when they are not independent …”

96. In both Purple Parking and Field Fisher “elements” are equated with “formally distinct services, which could be supplied separately” (our emphasis). That may suggest a limit on the independent analysis of elements and answer the point Ms Dhanoa raised in her written submissions based on the facts in De Montfort. The “meal deal” in that case involved packets of crisps and she asked whether Queenscourt’s analysis would mean that the transaction would have to be broken down as far as possible with the result that the crisps and the packets they were sold in would be treated as separate supplies. This reference to items which “could be supplied separately” suggests that, if there are elements which realistically could not be supplied separately, they would continue to be analysed as a single element. Accordingly, we do not consider there to be any reason to think that Queenscourt’s analysis is flawed because it is at risk of producing bizarre outcomes.

97. In Frenetikexito (discussed below) Advocate General Kokott ‘s analysis is that something which would be a supply if provided on its own, continues to be treated independently, even if the way it is supplied links it to something else, unless one of the CPP/Levob “exceptions to the principle that a supply is independent” applies.

98. Middle Temple concerned one of the Inns of Court, which let premises principally to barristers as chambers, and to some others for business use. It opted to tax its land and as a result the leases it granted were subject to VAT. For historical reasons, the Middle Temple owned a network of underground pipes through which cold water was supplied to chambers in the Inn. Cold water was supplied to that network by a water utility company. The supply of water to the network was metered, and the Middle Temple was charged accordingly. The cold water supplied by Middle Temple to each set of chambers or other premises in the Inn was not metered and was charged by reference to the area occupied by the premises concerned. Tenants in the Inn had no choice but to obtain their cold water from the Middle Temple. HMRC decided that the grant of a lease of the land, together with the supply of cold water, was a single supply, which was chargeable to VAT at the standard rate.

99. At [60] in Middle Temple the Upper Tribunal set out some principles it had derived from CJEU cases for the purposes of determining whether the transaction comprises one supply or two or more: “The key principles for determining whether a particular transaction should be regarded as a single composite supply or as several independent supplies may be summarised as follows: (1) Every supply must normally be regarded as distinct and independent, although a supply which comprises a single transaction from an economic point of view should not be artificially split. (2) The essential features or characteristic elements of the transaction must be examined in order to determine whether, from the point of view of a typical consumer, the supplies constitute several distinct principal supplies or a single economic supply. (3) There is no absolute rule and all the circumstances must be considered in every transaction. (4) Formally distinct services, which could be supplied separately, must be considered to be a single transaction if they are not independent. (5) There is a single supply where two or more elements are so closely linked that they form a single, indivisible economic supply which it would be artificial to split. (6) In order for different elements to form a single economic supply which it would be artificial to split, they must, from the point of view of a typical consumer, be equally inseparable and indispensable. (7) The fact that, in other circumstances, the different elements can be or are supplied separately by a third party is irrelevant. (8) There is also a single supply where one or more elements are to be regarded as constituting the principal services, while one or more elements are to be regarded as ancillary services which share the tax treatment of the principal element. (9) A service must be regarded as ancillary if it does not constitute for the customer an aim in itself, but is a means of better enjoying the principal service supplied. (10) The ability of the customer to choose whether or not to be supplied with an element is an important factor in determining whether there is a single supply or several independent supplies, although it is not decisive, and there must be a genuine freedom to choose which reflects the economic reality of the arrangements between the parties. (11) Separate invoicing and pricing, if it reflects the interests of the parties, support the view that the elements are independent supplies, without being decisive. (12) A single supply consisting of several elements is not automatically similar to the supply of those elements separately and so different tax treatment does not necessarily offend the principle of fiscal neutrality.” Of course, the language used by the Upper Tribunal will be conditioned by the question it is answering, but we consider it to be clear from this passage that the Upper Tribunal was of the view that the principles derived from the CJEU cases are to be applied at transaction level to decide whether the transaction should be regarded as a single composite supply.

100. We acknowledge that the reference to “principal services” in [60(8)] is slightly odd when the other references are in the singular (perhaps it follows from the earlier reference to “elements”), but principle (8) appears only to operate where there is a principal service/element to which other elements are ancillary. For the transaction to be regarded as a single composite supply, this must absorb all the elements in the transaction. There is no suggestion that this principle operates below the level of the whole transaction.

101. In Middle Temple the Upper Tribunal considered the relevance of the concept of fiscal neutrality and concluded (at [39]): “It is clear from the paragraphs set out above [paragraphs [38]-[40] in Purple Parking] that the principle of fiscal neutrality is concerned with ensuring that supplies of similar goods and services, which are thus in competition with each other, are treated the same way for VAT purposes. The CJEU clearly states, at para 39 of Purple Parking, that ‘the treatment of several services as a single supply for the purposes of VAT necessarily leads to tax treatment different from that that those services would have received if they had been supplied separately … Accordingly, a complex supply of services consisting of several elements is not automatically similar to the supply of those elements separately’. This passage shows that the principle of fiscal neutrality is not a factor to be taken into account in determining whether a transaction consisting of more than one element should be regarded as a single supply or as several independent supplies. Further, para 40 indicates that the fact that the UK zero rates the supply of cold water is not relevant to the question whether the different elements provided by the Middle Temple are a single supply of services or several supplies”.

102. There was no evidence before the FTT of dip pots being supplied as part of a meal deal just with hot chicken products, and so we do not know what the VAT analysis of such a supply would be, although it is not difficult to imagine what the parties’ analyses would be. However, even if such a supply would have constituted a single standard rated supply, these passages from Purple Parking and Middle Temple make it clear that the inclusion of those elements in a larger more complex supply could change the VAT analysis and that there should be nothing problematic or surprising in such an outcome.

103. The observation that a supply of one element on its own can be very different from a supply of the same element in combination with other elements is an important one. If the VAT analysis of a particular element can change depending on whether it is supplied on its own or in combination with other elements, it must follow that it would be artificial and potentially distortive to analyse an element otherwise than on the actual basis on which it was supplied. Put simply, if there is a transaction with 6 elements, the VAT liability should be determined on exactly that basis (a single 6-element transaction), not on a division of that transaction into (say) 3 separate transactions each of 2 elements. HMRC never explained on what principled basis the elements in a multi-element transaction could be divided between such hypothetical transactions. Such a level of unpredictability would not, in our opinion, be consistent with legal certainty or the efficient operation of the VAT system. Is there anything in Frenetikexito, Gray & Farrar or Field Fisher that impacts on this conclusion?

104. HMRC submits that support for its position is to be found in the Opinion of Advocate General Kokott in Frenetikexito.

105. At [16] she observed that for VAT purposes every supply must normally be regarded as distinct and independent. She explained the reason for this (at [17]) as being that VAT law has different rules governing the place of performance, exemptions and the tax rate. If supplies, which would be assessed differently if looked at individually, were subject to a uniform analysis for VAT purposes “merely because there are certain geographical, temporal or substantive links between them, that would circumvent this differentiated system”. So (at [18]) she said that each individual supply must be assessed separately for VAT purposes even where there are certain links between multiple supplies because they pursue a single economic aim. This is not an absolute rule, and at [20]-[21] he explained the exceptions to the principle that every supply is independent: “20. However, the principle that every supply is independent is not absolute. Transactions should not be artificially split, so as not to distort the functioning of the VAT system. The VAT treatment of bundles of supplies is thus caught in tension between the principle that supplies are independent, on the one hand, and the prohibition on the artificial splitting of single transactions, on the other.

21. In this connection, the Court has developed two exceptions to the principle that a supply is independent: single complex supplies (see under a.) and dependent ancillary supplies (see under b.). In addition, the VAT Directive also contains the exception for closely related activities (see under c.)”.

106. She went on to discuss these exceptions in more detail (a, b and c in [21] being references to the subsequent sections of his opinion where he does this).

107. In [21] she refers to “dependent ancillary supplies (see under b)” without referring to a single principal supply; she simply says that “dependent ancillary supplies” are an exception to the principle that each supply is independent. This can be contrasted with her earlier reference to “single complex supplies (see under a)”. As she later makes clear (at [22]) the single complex supply (Levob) exception can only produce one supply.

108. Starting at [34] she expands on the “dependent ancillary supply” exception. Taking the relevant passages in isolation and at face value, some of the language she uses could be applied both at the level of the multi-element transaction and below that level (asking whether one element is ancillary to a particular principal service). So, in paragraph [34] she says: “A further derogation to the principle that every supply is independent is required if a supply constitutes a merely dependent ancillary supply to a principal supply. A supply must be regarded as ancillary to a principal supply if it does not constitute for customers an end in itself but a means of better enjoying the principal service supplied. The ancillary supply has only secondary importance compared with the principal supply, which is why it ‘shares the tax treatment of the principal supply’. This means that the ancillary supply is to be treated for VAT purposes in exactly the same way as the principal supply”. In the first two sentences she uses the indefinite article and her switch to the definite article may be no more than a reference to “the” principal/ancillary service in question. Nothing in the language used here expressly shuts out the idea of the dependent/ancillary supply analysis taking place below the level of the overall transaction.

109. However, in [35] she says that “dividing a bundle of supplies into a principal and an ancillary supply does not give rise to any artificial splitting. The principal supply and the ancillary supply are clearly divisible from one another.” At this point she is clearly looking at an analysis being carried out at the level of the transaction (the bundle of supplies), albeit a relatively simple transaction (with one principal and one ancillary supply).

110. Her examples at [36] reinforce this. Importantly, at [37] she comments that fiscal neutrality does not require “transactions to be split” where negligible ancillary supplies are concerned and concludes that “For the sake of practicability, a single transaction should therefore be taken to exist.” Her analysis is focused at the level of the transaction and we consider that her use of the indefinite article in [34] needs to be understood in that light.

111. There is nothing in her opinion that suggests that she was endorsing the idea of the dependent/ancillary supply analysis taking place below the level of the overall transaction.

112. Field Fisher concerned a firm of solicitors which occupied office buildings as tenants. The landlords of the buildings had not opted to waive exemption so that the rent was exempt from VAT. The terms of the lease required the taxpayer to pay a service charge and the landlords had treated the services as exempt, in accordance with HMRC’s interpretation of the law. The service charge included such things as the taxpayer’s share of the cost of insuring the building and charges for the supply of water, heating, repairs, lift maintenance, cleaning and security. The taxpayer challenged that interpretation and sought to reclaim, as input tax, the VAT included in the payments made. HMRC refused the claim on the principal ground that the lease and the services constituted a single supply since they formed one indivisible economic supply which it would be artificial to split.

113. At [13] the CJEU described the question put to it as “whether the VAT Directive must be interpreted as meaning that, in the circumstances of the main proceedings, the leasing of immovable property and the supplies of services linked to that leasing must be regarded as constituting a single supply, entirely exempt from VAT, or several independent supplies, assessed separately as regards whether they are subject to VAT”. The Court also identified a particular issue raised as being the fact that the lease provided that the tenant must receive the services supplied by the landlord, even though it could in principle be supplied with at least some of the services by a third party.

114. Although (inevitably) the Court’s comments are conditioned by the question it was seeking to answer, we see the Court articulating (as set out above, at [14]) the general rule, that every supply should be regarded as distinct and independent, subject (in [15]) to an exception in certain circumstances, where formally distinct services must be considered to be a single transaction. At [16] and [17] it explains that those two circumstances are where one supply is ancillary to “the principal supply” (the CPP situation) and the other where two or more elements or acts are so closely linked that they form, objectively, a single, indivisible economic supply, which it would be artificial to split (the Levob situation). The important point for us is that an exception to the general rule is only in point where, in the context of a multi-element transaction, the otherwise distinct supplies “must be considered to be a single transaction”.

115. At [26] the Court addressed the relevance of the fact that a third party could in principle supply certain services. The Court considered that the existence of such a possibility is not decisive, as the possibility that elements of a single supply may, in other circumstances, be supplied separately is inherent in the concept of a single composite transaction.

116. We do not consider Field Fisher to support HMRC’s position. Indeed, paragraphs [15]-[17] are positively unhelpful to HMRC as they explicitly restrict departures from the norm (every element in a multi-element transaction to be treated independently) to cases where the otherwise distinct supplies “must be considered to be a single transaction”.

117. At [87] of the Decision, the FTT held that there was some support in Field Fisher at [24]-[25] for the suggestion that some elements of a single transaction may constitute a single supply whilst other elements may be separate. We tend to disagree.

118. In Field Fisher at [24] the Court held that supplies of services, which are linked to the leasing of immovable property and are supplied in accordance with the provisions of a lease, might constitute ancillary supplies or be indivisible from that leasing. At [25] it made the point that the mere fact that a supply is included in a lease cannot be decisive. If a lease were to provide for the inclusion of supplies which by their nature could not objectively be regarded as indivisible from or ancillary to the principal supply of the leasing of immovable property, but were independent of it, such supplies having only an artificial link to the principal supply (by being included in the lease), those supplies would not form part of a single supply of the leasing of immovable property, exempt from VAT.

119. We agree with Ms Brown that the point the CJEU is making here is that artificially including services in a lease is not enough for them to be treated as part of the leasing of real estate. The CJEU was not looking at a transaction which included elements which were genuinely (i.e. not just because of the parties creating an artificial link) part of the transaction but not ancillary to or indivisible from the principal supply.

120. Nor have we found anything in Gray & Farrar to support HMRC’s position. In that case (as recorded at [14]) there was no suggestion that what was supplied was not a single composite service. Ms Dhanoa directs us to two paragraphs ([41] and [46]) in the judgment of Simler LJ. Neither of these paragraphs assists her.

121. At paragraph [41] Simler LJ (as she then was) discusses the predominant element test identified in Město Žamberk Finanční ředitelství v Hradci Králové, (Case C-18/12) (“Mesto”), as a way of characterising a single supply. She notes that in Mesto the CJEU referred to Levob, where the predominant element test was used, in the context of supplies of standard software that was customised to meet the customer’s business needs, to identify whether, if this was a single supply, it was a supply of goods or a supply of services, and if a supply of services, where those services were supplied. We already know that Levob also provides a second (single economic transaction) test for a single supply, but nothing in the points Simler LJ is discussing here or her quotations from paragraphs [29] and [30] in the Court’s judgment in Mesto, which discuss how the predominant element is to be determined, are of any assistance to us in answering the question we are confronted with.

122. At [46] Simler LJ noted that the CJEU “has itself recognised the difficulty of prescribing a definitive test in the context of identifying the number of supplies for VAT purposes (see CPP at [27] to [29]) because the diversity of commercial operations made it impossible to give exhaustive guidance as to how to approach the problem correctly in all cases”. She considered that the same “may also be said of the closely related characterisation question”. Again, nothing here addresses the question we are answering. We note that Simler LJ readily considered that at [27] and [29] in CPP the ECJ was recognising that the task of identifying the number of supplies in a complex transaction was difficult. She did not suggest that the ECJ considered that supplies should not be treated independently in any circumstance outside an analysis of the overall transaction as a single supply, although clearly that issue was not one which she was concerned with. Did the ECJ in CPP leave open the possibility of individual elements in a multiple supply not being analysed separately?

123. HMRC says that in CPP the ECJ at [27] clearly appreciated that the answer to the question posed at [26] would not be reached in the same way in all cases. HMRC says that this supports its submission that, where there are multiple supplies made in the course of a wider commercial relationship or transaction, the tribunal’s task is to identify their true nature and VAT consequences by reference to the statutory scheme, rather than treating the existence of multiple elements as compelling a single characterisation at the level of the entire transaction.

124. We agree with HMRC that the analysis of multi-element transactions is acutely fact sensitive.

125. We also agree with HMRC that the need to address a wide range of different commercial operations means that there will be evolution in how the question is answered. We can see that in the way the ECJ developed the principal/ancillary test in CPP itself and then later (in Levob) developed the single complex transaction test. Most recently, in P in W the CJEU did not question the local court’s analysis that 3 of the elements made up a single complex transaction to which the fourth element (the reservation platform) “would constitute an ancillary supply forming part of that single complex transaction”. We agree with Ms Dhanoa’s submission that the CJEU appears here to be endorsing a combined application of the CPP and Levob analyses, which is a further demonstration of the flexibility to be adopted in this area.

126. However, we do not agree that this passage in CPP suggests that the question to be answered might change. In [27] the ECJ says that it is not possible “to give exhaustive guidance on how to approach the problem correctly in all cases”. We consider that the reference to “the problem” is not to the earlier commentary in [27], which explains why “the question of the extent of a transaction” is important and comprises the first part of the ECJ’s commentary on the problem posed in [26], but to the problem posed by the national court set out in [26].

127. That this is the correct reading of paragraph [27] of CPP is made clear by the ECJ in Purple Parking (at [30]): “In order to determine whether the taxable person is making to the customer, envisaged as being a typical consumer, several distinct principal supplies or a single supply, the essential features of the transaction must be ascertained and regard must be had to all the circumstances in which that transaction takes place …”

128. A similar point was made by the Upper Tribunal in Middle Temple (at [46]): “Although it might appear a statement of the obvious, the CJEU had never previously acknowledged that there is no absolute rule for determining whether two or more elements or acts are a single composite supply subject to the same VAT liability or are several independent supplies which must be treated separately for VAT purposes. There are some indicators that may point in one direction or another but they are not individually conclusive. These indicators must be evaluated once all the circumstances of the case have been taken into consideration. The CJEU’s statement in para 19 of Field Fisher Waterhouse clearly shows that the assessment of whether transactions are a single composite supply or several independent supplies is highly fact sensitive.”

129. This distinction between the question to be asked (whether the supplies constitute several distinct principal supplies or a single economic supply), which stays the same, and how that question is answered in a particular case (where there is no absolute rule) can also be seen in Middle Temple at [60] (set out in full above), where the Upper Tribunal sets out the “key principles for determining whether a particular transaction should be regarded as a single composite supply or as several independent supplies” it deduced from the CJEU cases. The list of principles includes a number of points not discussed in CPP, so again we see evolution in how the test is answered, but the question to be answered stays the same. Unsurprisingly, given the passage just quoted, the principles include a direction that: “(2) The essential features or characteristic elements of the transaction must be examined in order to determine whether, from the point of view of a typical consumer, the supplies constitute several distinct principal supplies or a single economic supply. (3) There is no absolute rule and all the circumstances must be considered in every transaction.”

130. Put simply, the point the ECJ is making at [27] in CPP, as we understand it, is no more than that there needs to be flexibility in the approach taken to answering the question whether a multi-element transaction is to be analysed as comprising a single supply as opposed to two or more distinct supplies. However, it is not suggesting that there are circumstances (other than where a multi-element transaction is properly analysed as a single supply) where the elements in a multi-element transaction should not be analysed separately. Conclusion on Ground 1(a)

131. For the reasons discussed above, our conclusion is that every supply (including every element in a multi-element transaction which would be a supply if provided on its own) must be regarded as distinct and independent unless it falls within one of the few exceptional cases where a derogation from that fundamental principle is permitted. Leaving aside legislative overrides (such as was seen in Talacre), these exceptions require all the elements in a single transaction to constitute a single economic supply following the principles in CPP or Levob (including a possible application of both principles to a single case, as in P in W).

132. Our conclusion on Ground 1(a), therefore, is that the FTT made an error of law when it concluded (at [88]) that there is no reason in principle why two or more elements of a single transaction cannot constitute a single supply whilst, at the same time, other elements of the same transaction may constitute a separate supply. Ground 1(b)

133. Ground 1(b) is a criticism of one aspect of the way the FTT concluded that dip pots were ancillary to the hot food supplied in a meal deal. Given our conclusion on Ground 1(a), that the dip pots and the other elements of a meal deal must be analysed independently, our decision on this Ground is not material to the outcome of this appeal, but we consider it briefly.

134. On Ground 1(b) Queenscourt says that the FTT’s conclusion (at [76]), that the element of choice here was not as important as it is in circumstances where elements of a transaction can realistically be refused by a consumer and obtained elsewhere from a third party, was based on a conclusion drawn from the factual matrix in the relevant caselaw (identified at [74]). However, Ms Brown submits, the underlying factual matrix of the case law authorities is irrelevant to the legal principle, which is summarised at [60(10)] of the Upper Tribunal decision in Middle Temple, as set out above in paragraph

99.

135. Against that background, we do not consider that the FTT made an error of law in its conclusion on the importance of choice. Here, as the FTT found (at [35]), if a customer did not want a dip pot, they might well receive one even if they asked not to, and the price would not be reduced even if no dip pot was delivered. If they wanted a dip pot, there was no point choosing not to take the dip pot included in the meal deal as the price would not be reduced and they would then need to pay more to get a dip pot from that outlet or elsewhere and dip pots could only be obtained from KFC outlets. There was a theoretical ability to choose not to take a dip pot as part of a meal deal, but making such a choice would not have any real impact on the relationship between the parties and so the element of choice was not important and the FTT was right to point that out.

136. Ms Brown also says that, at [111], the FTT dismissed choice as being irrelevant. The FTT said: “In any event, we do not consider the ability to choose to purchase a dip from the same KFC outlet but in a different way to be relevant to the question as to whether, from the point of view of the typical consumer of a meal deal, the supply of a dip pot as part of the meal deal is an end in itself or as a means of better enjoying the other items. As noted by the ECJ in Purple Parking at [31], “the fact that, in other circumstances, the elements in issue can be or are supplied separately is of no importance, given that that possibility is inherent in the concept of a single composite transaction”

137. We do not read this passage as the FTT saying that choice was completely irrelevant. It had already acknowledged the potential significance of choice. All it said was that a customer’s ability to buy a dip pot in a different way was irrelevant to the question of whether (from the perspective of the typical customer) the supply of a dip pot as part of the meal deal is an end in itself or a means of better enjoying other items. The possibility that elements of a composite transaction can be acquired separately is inherent in the concept of a single composite transaction. That point is obviously correct and is supported by authority (Purple Parking at [31]).

138. Ground 1(b) is not made out. The Public Law Issue HMRC’s assertion of lack of jurisdiction

139. As the FTT observed (at [128]), despite the point having been considered in numerous previous decisions (including by the Upper Tribunal and the Court of Appeal), there remains uncertainty as to the circumstances in which the FTT can consider arguments based on public law principles (including legitimate expectation) in the context of an appeal by a taxpayer.

140. Not long before the hearing of this appeal, the Upper Tribunal released its decision in MWL International Ltd & Anor v HMRC, [2026] UKUT 00062 (TCC) (“MWL”). One of the issues dealt with in MWL was whether the FTT had jurisdiction to decide whether the appellant in that case had a legitimate expectation that HMRC would not retrospectively change a particular agreement. MWL concerned national insurance contributions (“NICs”) and HMRC decided that a liability to NICs arose under mandatory provisions which did not incorporate any exercise of discretion. The relevant appeal rights were held to be narrowly framed, and not to signal a supervisory jurisdiction for the FTT. On that basis the Upper Tribunal held that the FTT had been right to conclude that the provisions in question did not confer jurisdiction on the FTT to decide the legitimate expectation argument.

141. We also drew the parties’ attention to MWL and to the recent decision of the FTT in WM Morrison Supermarkets Limited v HMRC [2025] UKFTT 1542 (TC) (“Morrisons”), which contains a discussion (at [288]-[392]) of some (but by no means all) of the case law on the FTT’s jurisdiction in relation to legitimate expectation in the context of tax appeals.

142. Although MWL dealt with very different statutory provisions to those which concern us, there are four paragraphs in that decision which we consider to be an important and useful summary of the current state of the authorities and the approach a tribunal should take when confronted with a suggestion that it should entertain a legitimate expectation (or other public law) argument: “74. In our opinion, there is a tension between the authorities, but it should not be overstated. The tension is explicable by a number of factors. First, and most importantly, many of the decisions relate to different statutory provisions, both in terms of the relevant substantive legislation and also in terms of the related appeal rights. As we emphasise below, the question is always one of statutory construction. Second, it is not always clear whether some statements in the authorities are focussed on whether the FTT might be expected to have a judicial review jurisdiction, or should have one, rather than on whether jurisdiction in fact arises under the particular provisions. Third, comments on the issue which are obiter carry less weight.

75. Nevertheless, the remaining uncertainty, which is being grappled with by the FTT as taxpayers increasingly seek to raise legitimate expectation arguments in tax appeals, would clearly benefit from clarification, either by legislation or by a decision of a superior court.

76. In our opinion, the approach which best reconciles the authorities is to focus on a conventional purposive construction of the particular statutory provisions in determining the extent of the FTT’s jurisdiction in an appeal. In any such exercise, the context and purpose of the provisions will clearly be material. The relevant provisions will include the substantive legislation which is the subject of the appeal, any provisions which relate to the decision making process in question (does the appeal relate to what Simler LJ in Beadle describes as “proceedings that are dependent on the validity of an underlying administrative act”?), and the related appeal rights, including in particular the extent to which they confer a supervisory jurisdiction on the FTT. It will also be relevant to determine whether the FTT proceedings are “enforcement action”.

77. We endorse and emphasise what the Upper Tribunal said to this effect in R & J Birkett v HMRC [2017] UKUT 89 (TCC) at [30]: “(4) In each case therefore when assessing whether a particular public law point is one that the FTT can consider, it is necessary to consider the specific jurisdiction that the FTT is exercising, and whether the particular point that is sought to be raised is one that falls to the FTT to consider in either exercising that jurisdiction, or deciding whether it has jurisdiction. (5) Since the FTT's jurisdiction is statutory, this is ultimately a question of statutory construction.””

143. The recovery assessments in this case were raised under section 80(4A) VATA, which provides as follows: “(4A) Where— (a) an amount has been credited under sub-section (1) or (1a) above to any person at any time on or after 26 May 2005, and (b) the amount so credited exceeded the amount which the Commissioners were liable at that time to credit to that person, the Commissioners may, to the best of their judgment, assess the excess credited to that person and notify it to him”.

144. This provision is cast in very different terms to section 8 of the Social Security Contributions (Transfer of Functions etc) Act 1999, the provision in point in MWL, which (the Upper Tribunal found) “does not require or confer a discretion on an officer. Rather, it envisages a decision that a person is liable to pay NICs.” The language of being able to (“may”) raise an assessment to the best of the Commissioners’ judgment is very different from a bald statement that it is for an HMRC officer to decide whether a person is or was liable to pay NICs of any particular class and, if so, the amount that he or she is, or was, liable to pay. Section 80(4A) clearly confers an element of discretion on HMRC. HMRC’s exercise of that discretion will be subject to supervision by the Administrative Court, but can the FTT exercise a parallel supervisory jurisdiction?

145. Section 83 VATA provides for the bringing of appeals and, so far as relevant for us, provides as follows: “(1) Subject to sections 83G and 84 an appeal shall lie to the tribunal with respect to any of the following matters – … (t) a claim for the crediting or repayment of an amount under section 80, an assessment under subsection (4A) of that section or the amount of such an assessment;”

146. As we pointed out to the parties, some of the heads of appeal in section 83(1) refer to “a decision of the Commissioners” whereas others do not, and the opening words of the predecessor provision (section 40(1) of the Value Added Tax Act 1983) were different to those in section 83(1). They read as follows: “(1) An appeal shall lie to a value added tax tribunal constituted in accordance with Schedule 8 to this Act against the decision of the Commissioners with respect to any of the following matters …”. A reference to “a decision of the Commissioners” might be thought more clearly to bring the process by which HMRC took a particular course of action before the tribunal (rather than just asking whether whatever step they took satisfied the statutory requirements) and so to signpost a supervisory jurisdiction. Although we could see that the language of the statute has changed, we were unable to identify for ourselves when that change took place or whether at the time the change was made anything was thought to turn on it. We asked Ms Dhanoa and Ms Brown whether they thought anything turned on these linguistic differences. Without giving any particular reasons in support of their view, Ms Dhanoa thought these linguistic differences might be material, whereas Ms Brown thought they were not.

147. The first limb of section 83(1)(t) (an appeal in respect of a claim for repayment) was considered by Jacob J in CCE v National Westminster Bank plc, [2003] STC 1072 (“NatWest”). The taxpayer complained of unfair treatment (not being treated in the same way as other comparable car leasing companies) when the commissioners decided not to repay overpaid VAT based unjust enrichment. Following an earlier decision of Moses J also dealing with an appeal in respect of a claim for repayment of overpaid VAT, Marks and Spencer plc v CCE, [1999] STC 205 (“M&S”), Jacob J held that the tribunal did not have jurisdiction to determine the unfair treatment point.

148. The second limb of section 83(1)(t), with which we are concerned, is very similar to the provision considered in Zeman. The assessment under appeal in that case was raised under section 73(1) VATA (which allows HMRC to make best judgment assessments in a wide range of circumstances) and the appeal was brought under section 83(1)(p) VATA (“an appeal … with respect to … an assessment”). In Zeman the Upper Tribunal held that the FTT had jurisdiction to consider a legitimate expectation argument on such an appeal, expressing its conclusion (at [84]) in these terms: “Coming back then to where we started our analysis, the critical question in this case (see Beadle at [44]) is whether the relevant statutory scheme expressly or by implication excludes the ability to raise a public law defence of legitimate expectation (again, see Beadle at [44]). For all the reasons given above, we do not consider that s 83(1)(p) does exclude that ability. On the contrary, on the facts of this case and given the broad subject-matter of s 83(1)(p), we see strong reasons for thinking that it would be artificial and unworkable to exclude a defence based on the public law principle of legitimate expectation from the tribunal’s appellate jurisdiction. We therefore consider that the FTT did have jurisdiction to determine that question in this case.”

149. Although the Upper Tribunal’s decision on jurisdiction has not been universally approved, it was held to be obiter but correct (albeit with reservations) by the FTT in this case, and as binding and correct by the FTT in Treasures of Brazil Limited v HMRC, [2024] UKFTT 00929 (TC), and United Carpets (Franchisor) Ltd v HMRC, [2025] UKFTT 895 (TC).

150. In Morrisons the FTT considered that the conclusions of the Upper Tribunal in Zeman on the scope of the tribunal’s jurisdiction were not binding on it, but decided that it should follow them, partly because it considered that they were conclusions of a superior tribunal which, whilst wrong in its opinion, were not plainly wrong and partly because of the three FTT decisions which considered those observations to be correct and the need for the FTT to speak consistently with one voice on an issue as fundamental as jurisdiction.

151. Put briefly, as the outcome of this appeal does not turn on the public law issue, we consider that the FTT was wrong to conclude that Zeman was correctly decided and on that basis to conclude that it had jurisdiction to entertain a legitimate expectation argument. In our judgment: (1) Whilst we respectfully agree with the Upper Tribunal in MWL that this question should not be approached from any preconceived position (as to whether a public law jurisdiction needs to be expressly conferred or excluded), the overall statutory context is (as the Upper Tribunal in MWL made clear) relevant. A reading that there is no jurisdiction for the FTT to consider legitimate expectation arguments would be consistent with the tribunal structure established by the Tribunals Courts and Enforcement Act 2007 (“TCEA”) and the rule in O’Reilly v Mackman, [1983] 2 AC 237,  that public law issues should generally be raised in judicial review proceedings in the High Court. (2) Such a reading would also be consistent with the run of authorities before Beadle. With the exception of Oxfam v Revenue and Customs Commissioners, [2010] STC 686, (which was not followed in HMRC v Noor, [2013] UKUT 71 (TCC)), the authorities before Beadle (including CCE v JH Corbitt (Numismatists) Ltd, [1980] STC 231, a House of Lords decision where essentially the same appeal gateway was in point as that considered in Zeman) lean heavily against the tribunal having a supervisory jurisdiction (to consider public law issues), even where the appeal is “in respect of” a decision which the tax authority “may” make. (3) In Beadle itself the Court of Appeal concluded that the tribunal had jurisdiction to consider an appeal against a penalty assessment, basing its decision on the accepted exception to the rule in O’Reilly v Mackman for “enforcement proceedings” except where (as was held to be the case there) the statutory scheme in question excludes the ability to raise a public law defence. However, nowhere in her judgment in Beadle did Simler LJ equate an appeal against an ordinary tax assessment with enforcement proceedings, and assessment and enforcement are quite different matters.  (4) We consider that the Upper Tribunal in Zeman was wrong to equate a taxpayer bringing an appeal against a tax assessment with, in substance, a defendant in enforcement proceedings, and then go on to hold that the taxpayer should be entitled to challenge HMRC’s decision to raise an assessment on public law grounds in proceedings before the FTT “unless that entitlement is excluded by the relevant statutory language”. (5) It is going too far to read the requirement for “best judgment” in an assessment under section 73 VATA as carrying with it an obligation on HMRC to consider issues which might impact on a decision to assess, such that the appeal right in section 83(1)(p) gives the FTT jurisdiction to consider public law issues. As the Upper Tribunal itself observed in Zeman at [80], the focus of the “best judgment” test is on whether HMRC have made a genuine, reasonable attempt to work out what is due from a taxpayer. In the light of the context in which this calculation is being made, “best judgment” speaks to how HMRC are to calculate the amount of VAT due in the unsatisfactory position they find themselves in. There is no suggestion that the requirement for “best judgment” goes any wider than that. In MWL the Upper Tribunal (at [80]) described a best judgment assessment as giving a “discretion to estimate”. In one of the “best judgment” cases referred to by the Upper Tribunal in Zeman, CCE v Pegasus Birds Ltd, [2004] EWCA Civ 1015, the Court of Appeal  held that an assessment would only fail the “best judgment” test if there had been no honest and genuine attempt to make a reasoned assessment of the VAT payable. Even in such a case, the Court of Appeal was not persuaded that the correct response was to set aside the whole assessment. .

152. It follows from this that we consider that the FTT erred in law when it held (at [173]) that it should follow the reasoning of the Upper Tribunal in Zeman in determining whether there is anything in the relevant legislation which excluded the jurisdiction of the Tribunal to consider an argument based on legitimate expectation. What the FTT should have done is to undertake a conventional purposive construction of the relevant provisions, as explained in MWL at [76].

153. We also do not consider that the FTT could properly arrive at its conclusion (at [175]) that there was nothing in the statutory scheme of section 80(4A) and section 83(1)(t) VATA which expressly or impliedly excluded the jurisdiction of the Tribunal to consider an argument based on legitimate expectation.

154. We know from M&S and NatWest that the first limb of section 83(1)(t) gives the tribunal jurisdiction to hear argument about whether the statutory conditions making HMRC liable to credit amounts are met, but not to entertain public law arguments (in those cases that HMRC were operating the unjust enrichment defence inconsistently as between similar traders). Despite the language of this limb of section 83(1)(t) being very wide (“an appeal … with respect to …a claim for the crediting or repayment of an amount under section 80”) and seemingly encompassing anything to do with a claim, the tribunal can only hold HMRC liable to credit an amount if the relevant statutory conditions are met. Bearing in mind that there is no permissive, discretionary language in the substantive provision, the tribunal’s appeal jurisdiction being confined to deciding whether the statutory conditions are met and not including public law issues is perhaps not surprising.

155. The second limb of section 83(1)(t) gives an appeal right where HMRC decide that an amount credited to a person is excessive and raises an assessment to recover that amount. Again, the language is on its face very wide (“an appeal … with respect to … an assessment under subsection (4A) of [section 80] or the amount of such an assessment”). In context we consider it unlikely that Parliament would have intended in so few words to create an appeal regime which does not give the tribunal jurisdiction to consider public law arguments when considering whether HMRC are liable to credit an amount under section 80, but does give the tribunal such a jurisdiction when considering an appeal against an assessment raised to recover a credit they gave under that provision but which (HMRC say) has turned out to be excessive.

156. The FTT did not take both limbs of section 83(1)(t) into account in its analysis; it focused on the relationship between the second limb of section 83(1)(t) and section 80(4A), treating those provisions as a hermetically sealed unit, and failed to consider the wider context, which we have just discussed.

157. Although Ms Brown submitted that, as the substantive legislation (section 80(4A)) is permissive not mandatory, the FTT has jurisdiction to review HMRC’s exercise of this statutory discretion, neither she nor Ms Dhanoa addressed us on the conventional purposive interpretation of the second limb of section 83(1)(t). This would not necessarily be an easy task to undertake. Given our conclusions on Ground 1(a) and Ground 2, to which we turn next, nothing turns on any conclusion we might reach on this issue, and it would be wasteful and inconsistent with the overriding objective to embark on that exercise at this stage.

158. Accordingly, although we agree with HMRC that in reaching its conclusions the FTT made the errors of law we have identified, we leave the question whether the FTT has jurisdiction to consider arguments based on legitimate expectation in the context of an appeal under section 83(1)(t) VATA against a recovery assessment made under section 80(4A) VATA to be considered in a case where the court has heard full submissions on the relevant statutory scheme and the issue would make a difference to the outcome. Ground 2

159. Queenscourt’s criticism of the FTT here (its only criticism of the way the FTT dealt with its legitimate expectation argument) is that the FTT erred in law by confining its analysis to weighing up the financial detriment suffered by Queenscourt and failing to consider whether permitting HMRC to resile from its previous decision to repay the VAT was not conducive to good administration.

160. The FTT’s consideration of the merits of Queenscourt’s legitimate expectation argument begins at [180]. So far as relevant to Ground 2, it correctly observed at [192] that there is no requirement that the person to whom a ruling is given has relied on it to their detriment. That said, the FTT examined (at [202]-[214]) the detriment alleged by Queenscourt and held that Queenscourt may have suffered some detriment.

161. The FTT’s consideration of the merits of Queenscourt’s case on legitimate expectation included identifying (at [192]) that, for Queenscourt to succeed, it needed to be the case that it would not be fair for HMRC to depart from the clear and unambiguous ruling. In that context, the FTT considered (at [194]-[197]) what unfairness means in the context of tax. In particular, it noted the observation by the Court of Appeal in HMRC v Hely Hutchinson, [2017] EWCA Civ 1075 (at [37]), that part of the context is that the law imposes on HMRC a duty to collect tax and that taxpayers must expect to pay the right amount of tax, and its reference at [40] to the comment in R v Inland Revenue Commissioners, ex parte MFK Underwriting Agencies Limited, [1990] 1 WLR 1545 at 1569, that ‘the taxpayer’s only legitimate expectation is, prima facie, that he will be taxed according to statute not concessions or a wrong view of the law’. At [48] the Court of Appeal had noted that a decision-maker ‘is not bound, and is not entitled, to follow a previous decision that he considered erroneous’, and the FTT drew on this when it held that the correction of a mistake by HMRC is not, on its own, sufficient to reach the necessary threshold of unfairness.

162. Both cases relied on by Queenscourt (In the matter of an application by Geraldine Finucane for Judicial Review (Northern Ireland), [2019] UKSC 7, (“Finucane”) and R (Nadarajah) v Secretary of State for the Home Department, [2005] EWCA Civ 1363 (“Nadarajah”)) endorse the fundamental point (with which we respectfully agree) that, at least as a starting point, subjects have a right to expect that public authorities will behave towards them in a straightforward and consistent way and do what they have promised. However, that is not an absolute.

163. In Nadarajah, when discussing this premise, Laws LJ said (at [68]) that “Where a public authority has issued a promise or adopted a practice which represents how it proposes to act in a given area, the law will require the promise or practice to be honoured unless there is good reason not to do so” (our emphasis).

164. In Finucane Lord Kerr said (again our emphasis) (at [72]): “I would disagree with any suggestion that it must be shown that the applicant suffered a detriment before maintaining a claim for frustration of legitimate expectation for a fundamental reason. A recurring theme of many of the judgments in this field is that the substantive legitimate expectation principle is underpinned by the requirements of good administration. It cannot conduce to good standards of administration to permit public authorities to resile at whim from undertakings which they give simply because the person or group to whom such promises were made are unable to demonstrate a tangible disadvantage. Since the matter does not arise, however, it is better that the point be addressed in a future case when it is truly in issue”.

165. This reference to resiling from undertakings “at a whim” makes it clear that departures from undertakings for a good reason may constitute a fair and reasonable course for an authority to take. Although Ms Brown pointed us to this passage, Lord Kerr is clearly not advocating as a rule of good administration that public authorities should always be held to their promises or practices.

166. Whilst it did not use the expression “good administration”, the FTT clearly identified the factors in play in the tax context at [194]-[197]. These include the fundamental duty of HMRC to collect the taxes Parliament has decided should be paid, the expectation taxpayers should hold that they will pay this tax (no more or less) and the need to be fair to all taxpayers. At least as a starting point, it is not fair to one taxpayer that, because of an error on HMRC’s part in relation to their dealings with someone else, that taxpayer should pay more tax than another taxpayer in a similar situation.

167. After considering the level of detriment Queenscourt suffered, the FTT returned to these factors at [215], when it said that it needed to weigh HMRC’s duty to collect the right amount of tax and the strong public interest in ensuring that no individual taxpayer is unfairly advantaged at the expense of other taxpayers against the detriment to Queenscourt that it had identified.

168. Whether we call these factors particular features of good administration in the tax context or potential good reasons for HMRC’s departure from previous rulings, “good administration” does not require public authorities to honour prior undertakings in all circumstances, and the FTT correctly identified the issues that needed to be taken into account when deciding whether HMRC’s decision was so outrageously unfair that it should not be allowed to stand. It reflected those issues in its conclusion.

169. Ground 2 is not made out. Disposition

170. For the reasons discussed above, Queenscourt succeeds on Ground 1(a), but not Ground 1(b) or Ground 2, and HMRC succeeds in its argument that the FTT made errors of law while concluding that it had jurisdiction to consider public law arguments in this appeal.

171. Section 12 of the Tribunals Courts and Enforcement Act 2007 provides as follows: “(1) Subsection (2) applies if the Upper Tribunal, in deciding an appeal under section 11, finds that the making of the decision concerned involved the making of an error on a point of law. (2) The Upper Tribunal— (a) may (but need not) set aside the decision of the First-tier Tribunal, and (b) if it does must either— (i) remit the case to the First-tier Tribunal with directions for its reconsideration, or (ii) re-make the decision.”

172. In Degorce v HMRC, [2017] EWCA Civ 1427, Henderson LJ (with whom Thirlwall and Longmore LJJ agreed) confirmed (at [95]) that “if the Upper Tribunal finds an error of law to have been made, it then has a broad discretion whether or not to set aside the decision of the FTT” and went on to add “I find it difficult to envisage circumstances in which the Upper Tribunal could properly leave the decision of the FTT to stand, once it is satisfied that the error of law might (not would) have made a difference to that decision”.

173. It is common ground that the overall supply of a meal deal is a multiple supply, not a single composite supply. Against that background, if the FTT had correctly concluded that each element of a multiple supply must be analysed separately for VAT purposes, it could only have concluded that dip pots supplied as part of a takeaway meal deal were separate zero-rated supplies. In this case the FTT’s error is a material one; if it had not made it, the FTT would inevitably have reached the opposite conclusion.

174. As we have just explained, finding for Queenscourt on Ground 1(a) inevitably leads to the conclusion that dip pots supplied as part of a takeaway meal deal are separate zero-rated supplies. For that reason, we agree with Ms Brown that nothing would be gained by remitting the case to the FTT and we should re-make the FTT’s decision to reflect the outcome it would inevitably have reached had it not made this error of law.

175. Queenscourt having succeeded on the VAT liability issue, the public law issue (which was run as an alternative were it to fail on the VAT liability issue) is irrelevant to the outcome of this appeal.

176. Queenscourt’s appeal under Ground 1(a) is allowed, and we set aside the FTT’s decision and remake it to allow Queenscourt’s appeal against HMRC’s decision to refuse the repayment claimed in ECN2 and against the recovery assessments relating to ECN1. THE HONOURABLE MRS JUSTICE JOANNA SMITH DBE JUDGE MARK BALDWIN Release date: 19 May 2026