The Commissioners for HM Revenue and Customs v The Rank Group PLC: [2012] UKUT 347 (TCC) [2012] UKUT 347 (TCC)
[1]The essence of betting is the placing of a bet or stake on the outcome of an 5 event in which the punter has no substantial involvement. The essence of gaming is the involvement of the punter in a game of chance for winnings or a prize. But there is no “brightline” boundary between the two. The substantive law regulating betting and gaming is complex and has significantly influenced the development of the machines (including slot machines) which punters and 10 players use to entertain themselves in betting and gaming.[2]Betting and gaming are taxed in different ways. The provisions of the VAT Act and its subsidiary statutory instruments in charging tax and granting exemptions from tax in relation to gaming machines reflect the complexity of 15 the substantive law. But all of those taxing provisions are subject to the overriding principle of “fiscal neutrality”. This principle precludes treating services which are similar, and thus in competition with each other, differently for VAT purposes. 20 3. Gaming machines operating under the regulations contained in Part III of the Gaming Act 1968 (“Part III machines”) are subject to the VAT regime. The Rank Group plc (“Rank”) has paid VAT of the order of £30 million on the revenue from its Part III machines for the period in question (from 1 October 2002 until 5 December 2005). There are many other operators who have done 25 likewise. Betting machines called “fixed-odds betting terminals” (“FOBTs”) are exempt from VAT (although they are subject to a charge to betting duty).[4]The question which lies at the heart of this appeal is whether charging VAT on Part III machines is a breach of the principle of fiscal neutrality, given that 30 FOBTs are exempt from VAT: and that involves deciding whether the services provided by Part III machines and by FOBTs are relevantly “similar”.[5]Before descending into the detail of the argument it is well to note a limitation on the principle of fiscal neutrality. The principle does not require all betting 35 and gaming to be subject to a single tax regime. The Sixth Directive 77/388/EEC (relating to the common system of value added tax) contained, in Article 13, a provision relating to exemptions. Article 13.B.(f) provides:- “….. Member States shall exempt the following under conditions 40 which they shall lay down for the purpose of ensuring the correct and straightforward application of the exemptions and of preventing any possible evasion, avoidance or abuse…. (f) betting, lotteries and other forms of gambling, subject to conditions and limitations laid down by each Member State…..” 45 3 In HMRC v The Rank Group plc (joined cases C-259/10 and C-260/10) (“the Reference”) the Court of Justice of the EU (“the ECJ”) explained how the principle of fiscal neutrality related to this exemption in these terms :- 5 “[53] It must first be observed that if Article 13.B.(f) of the Sixth Directive and the discretion which that provision grants to the Member States … are not to be deprived of all useful effect, the principle of fiscal neutrality cannot be interpreted as meaning that betting, lotteries and other games of chance must be considered similar services within 10 the meaning of the principle. A Member State may thus limit the VAT exemption to certain forms of games of chance….. [54] It follows ….that the principle is not breached where a Member State imposes VAT on services supplied by means of slot machines while exempting horse-race betting, fixed-odds bets, lotteries and 15 draws from VAT… [55] However, in order not to deprive the principle of fiscal neutrality of meaning and so as not to distort the common system of VAT, a difference of treatment for VAT purposes cannot be based on differences in the details of the structure, the arrangements or rules of 20 the games concerned which all fall within a single category of game, such as slot machines…”[6]I must now set out the context of the argument. What follows is such description as is necessary to explain the points for determination on this 25 appeal, and it is not intended to amount to findings of fact binding any Tribunal which has to consider these questions hereafter.[7]Part III machines had to be operated on premises licensed for gaming. The maximum stake and the maximum prize were both regulated (and varied 30 depending upon the type of premises licensed). Those that could be operated on any licensed premises were governed by the provisions of section 34. If the section 34 machine was located in a licensed betting office it could have a maximum stake of £.30p and a maximum prize of £25: if the section 34 machine was located on other licensed premises it could have a maximum 35 stake of £.30p and a maximum cash prize of £5 or a non-cash prize worth £8. The machines offered a game of chance or afforded amusement with prizes in return for the payment of a single stake. Typical formats were blackjack, poker or reel-based games (such as “fruit machines”). The range of outcomes could be very large since the punter was simply taking a chance that one of them 40 would occur. The punter played a game, and if he was experienced in playing then that experience might influence the outcome of the game, because he was required to make decisions in the course of the game.[8]FOBTs were a later development than the original gaming machines, but were 45 something like them. They operated in licensed betting offices. The legislation did not prescribe the maximum permitted stake or the maximum payout: but there was a voluntary code in place under which both stakes and payouts were 4 much higher than those prescribed for Part III machines (being £15 per bet (or £100 aggregate for multiple bets) and £500 respectively). FOBTs offered the punter a range of fixed-odds bets on an event, that event being produced by a random number generator located elsewhere and outside the control of the 5 operator. The punter could place multiple bets: but he made his decisions before the process producing the event was initiated. Typical formats were automated roulette or virtual horse- or greyhound-racing. (These formats were typical because the events had to be limited in number, since it had to be possible to bet on every outcome). 10[9]At the time when FOBTs were developed some of those who operated them were at pains to distinguish them from Part III machines. Thus Mr Alan Ross (the managing director of Ladbrokes Limited) wrote on 6 August 2003 to the Department of Culture Media and Sport emphasising 15 “Fixed odd (sic) betting terminals are quite different from any gaming machines….. It is neither appropriate nor possible to transfer the betting activity currently carried out on FOBTs to gaming machines. The activity involved by the customer in using FOBTs is betting 20 whereas the activity involved in playing gaming machines is gaining.”[10]As an operator of Part III machines Rank took the opposite view. Rank wished to argue that Part III machines (in particular section 34 machines) were relevantly “similar” to FOBTs, and so in competition with them; and that 25 accordingly the differential VAT treatment of Part III machines was a breach of the principle of fiscal neutrality.[11]The same issue had arisen in the context of comparing(a) gaming machines operating under section 14 of the Gaming Act 1968 with(b) gaming machines 30 regulated under section 21 of the Gaming Act 1968 or section 16 of the Lotteries and Amusements Act 1976 (“section 16/21 machines”). The former were subject to VAT but the latter were exempt from VAT. Rank had commenced proceedings before the Value Added Tax Tribunal (which became the First Tier Tribunal (Tax Tribunal)) (“the Tribunal”) seeking repayment of 35 the VAT paid in respect of the revenues earned on its gaming machines. Into those proceedings it introduced the argument that not only were section 16/21 machines relevant comparators for the purpose of applying the principle of fiscal neutrality, but so also were FOBTs. In opposition HMRC advanced a number of arguments of which only one is now relevant to this appeal. HMRC 40 argued that FOBTs and Part III machines were not relevantly “similar” so that the principle of fiscal neutrality did not require similar treatment for VAT purposes. The arguments relating to the section 16/21 machines and those relating to FOBTs were separated out, the former being called “Slots 1” and the latter “Slots 2” 45[12]The Tribunal granted the Slots 1 claim and on appeal I upheld that decision: HMRC v Rank Group [2009] EWHC (Ch) 1244. In the Slots 1 judgment 5 (which also considered similar issues in the context of mechanised cash bingo) I said of the principle of fiscal neutrality (at paragraph [21]):- “The first element of the principle is that the goods or services supplied 5 must be apparently identical or similar. In the instant case it was common ground that games of chance conducted under section 14 (subject to VAT) and those conducted under section 21 (exempt) were all but identical. This concession was rightly made because the doctrine requires the comparison to be made between the relevant 10 goods and services at a high level of abstraction and on the basis of broadly defined categories. This is illustrated by the opinion of the Advocate General in Linneweber [2005] ECR 1-1131 where (at paragraph 58) the view is expressed that: 15 “ In assessing similarity of gaming machines the national court must focus on whether the use of gaming machines operated in public casinos is comparable from the average consumer’s point of view to the use of gaming machines operated elsewhere, those machines therefore being in 20 competition with each other, factors which must be taken into account in this regard being in particular potential scale of winnings and the gambling risk”. That was an elaboration of the view earlier expressed … that where the 25 similarity of goods or services is being assessed it must be considered whether they have similar characteristics and meet the same needs from the point of view of consumers”.[13]When the Tribunal came to consider the Slots 2 arguments it took that 30 paragraph as giving appropriate guidance as to the approach to be adopted: see paragraphs [23] and [26] of the Tribunal’s decision of 11 December 2009. (It did, however, take a different view of the significance of paragraph 58 of the Advocate General’s opinion: see paragraph [31]). After considering Linneweber the Tribunal held that the maximum stake, the maximum 35 winnings and the game rules were not legally relevant to the issue of “similarity”, saying (in paragraph [32]) :- “Mr Vadja relied on a series of differences between FOBTs and Part III machines which in our judgment are therefore not relevant on the 40 basis of the decision in Linneweber. In particular he relied on the different stake and prize limits and the fact that complex betting patterns were available on FOBTs which the rules prevented on Part III machines. Mr Justice Norris decided that the comparison is to be made at a high level of abstraction……….” 45[14]The Tribunal went on to consider a number of other differences between Part III machines and FOBTs (payout ratios, multiple bets, the ability of a player to 6 make decisions in the course of a game, the difference between betting and gaming). But the Tribunal rejected all of the suggested distinctions as insufficient to prevent the generality of players considering the machines as similar at a high level of abstraction, saying (in paragraph [38]): 5 “On the evidence before us to the generality of players they were all just gambling machines”. The Tribunal further held (at paragraph [50]) that “from the players’ 10 perspective there would be very few differences between a FOBT and a gaming machine” and that “the customer was primarily interested in the opportunity to gamble”. It accordingly considered that Part III machines and FOBTs were relevantly “similar” for the purpose of applying the principle of fiscal neutrality. 15[15]HMRC appealed that decision arguing in its Notice of Appeal that the Tribunal had interpreted my approach of considering whether FOBTs and Part III machines had “similar characteristics and [met] the same needs of the consumers” at a “high level of abstraction” in such a way as to render any 20 differences between the machines, or types of gambling activities, entirely meaningless. An appeal lies only in respect of an error of law. The errors of law identified were:(a) That the Tribunal had misinterpreted Linneweber in holding that 25 the place where the activity was conducted, the stake and prize limits or the betting patterns available on machines and the difference in social environment were irrelevant as a matter of law;(b) That the Tribunal erred when undertaking a comparison at “a high level of abstraction” in treating as decisive that the generality of 30 players saw both Part III machines and FOBTs as “all just gambling machines” their interest being in “the opportunity to gamble”;(c) That no reasons were given for rejecting the identified distinctions as insufficiently important; 35(d) That even if at the margins some of the distinctions could be regarded as “arcane” it did not follow that there is no real distinction between the types of services offered, or that the distinction is not significant from a customer’s point of view. 40 16. The Court of Appeal granted permission to appeal my decision on Slots 1, and decided to make the Reference to the ECJ to assist in the determination of that appeal. I decided also to refer the issues arising on Slots 2 for a preliminary ruling as part of the Reference. The formulation of the questions was agreed and included the following:- 45 “In determining whether the principle of fiscal neutrality requires the same tax treatment of….FOBTs and Part III gaming machines, what 7 level of abstraction should be adopted by the national court in determining whether the products are similar? In particular, to what extent is it relevant to take into account the following matters: (a) similarities and differences in the permitted maximum stakes and 5 prizes between FOBTs and Part III gaming machines; (b) ………… (c) that the chances of winning the prize on the FOBTs were directly related to the published fixed odds, whereas the chances of winning on the Part III gaming machines could in some cases be varied by a 10 device that ensured a particular percentage return to the operator and player over time; (d) similarities and differences in those formats available on FOBTs and Part III gaming machines;(e) similarities and differences between the FOBTs and Part III 15 gaming machines in the interaction which could occur between the player and the machine;(f) whether or not the matters referred to above were either known to the generality of the players of the machines or regarded by them as relevant or important…………” 20[17]In giving its judgment on the reference the ECJ reformulated the question to avoid reference to any “level of abstraction”, saying ( at paragraph [52]):- “By this question the Upper Tribunal … seeks to know essentially, 25 whether or not, in order to assess whether, in the light of the principle of fiscal neutrality, two types of slot machine are similar and require the same treatment for VAT purposes, account must be taken of permitted minimum and maximum stakes and prizes, the chances of winning, the available formats and the possibility of interaction 30 between the player and the slot machine….”[18]The ECJ answered the question it so formulated in this way:- “[43] In order to determine whether two supplies of services are 35 similar …..account must be taken of the point of view of a typical consumer… avoiding artificial distinctions based on insignificant differences. [44] Two supplies of services are therefore similar where they have similar characteristics and meet the same needs from the point of view 40 of consumers, the test being whether their use is comparable, and where the differences between them do not have a significant influence on the decision of the average consumer to use one such service or the other….. [56] It is apparent from paragraphs 43 and 44 of the present judgment 45 that the determination whether games of chance which are taxed differently are similar, which it is for the national court to make in the light of the circumstances of the case, … must be made from the point 8 of view of the average consumer and take account of the relevant or significant evidence (sic) liable to have considerable influence on his decision to play one game or the other. [57] In that regard, differences relating to the minimum and maximum 5 stakes and prizes, the chances of winning, the formats available and the possibility of interaction between the player and slot machine are liable to have a considerable influence on the decision of the average consumer, as the attraction of games of chance lies chiefly in the possibility of winning. 10 [58] In the light of the foregoing considerations, the answer to the … question … is that, in order to assess whether, in the light of the principle of fiscal neutrality, two types of slot machine are similar and require the same treatment for VAT purposes it must be established whether the use of those types of machine is comparable from the point 15 of view of the average consumer and meets the same needs of that consumer, and the matters to be taken into account in that connection are, inter alia, the minimum and maximum permitted stakes and prizes and the chances of winning.” 20 19. In the light of that answer it can, in my judgment, now be seen that the decision of the Tribunal on Slots 2 contains two errors of law. First, differences in permitted stakes and prize limits and the available betting patterns are legally relevant to the determination of similarity because they are “liable to have a considerable influence on the decision of the average 25 consumer, as the attraction of games of chance lies chiefly in the possibility of winning”. In not recognising them as such and weighing them accordingly the Tribunal erred in law. Second, whereas the Tribunal (following my own approach) made the relevant comparison at “a high level of abstraction” the judgment of the ECJ avoids the using concept of “a level of abstraction”. It 30 instead requires the Tribunal to “take account of the relevant or significant [elements or circumstances] liable to have a considerable influence on [the consumer’s] decision to play one game or the other”. (The word “evidence” in the judgment can be seen, from a comparison of the English with French, Spanish and German texts to mean “elements” or “circumstances”). 35 “Comparison of facts at a high level of abstraction and on the basis of broad categories” seems to me may well be a different test from “taking account of significant elements liable to have a considerable influence on the consumer’s decision.” 40[20]Section 12 of the Tribunals Courts and Enforcement Act 2007 provides that if the Upper Tribunal is satisfied that the First Tier Tribunal made an error on a point of law in reaching its decision then the Upper Tribunal “may (but need not) set aside the decision”. For Rank, Mr Lasok QC submits that 45 notwithstanding my decision that the Tribunal erred in law I should not set aside its decision because the evidence before the Tribunal did not establish any relevant or significant elements or circumstances liable to have a 9 considerable influence on the consumer’s decision whether to play a Part III gaming machine or a FOBT. For HMRC Mr Peretz responds by submitting this requires me to conclude that even if the correct legal tests had been applied HMRC must have failed on the evidence adduced. I agree. Only if 5 there was no evidence fit for consideration by the fact-finding tribunal when applying the correct legal test should I leave the present decision in place. If there is evidence relevant to the correct legal test which demonstrates an arguable case then the appeal should be allowed and that arguable case considered. 10[21]In my judgment there is evidence relevant to the correct legal test which is fit for consideration by a fact-finding tribunal. This is not an indication of the conclusion which I consider ought to be drawn from that evidence: I am careful to say no more than that the evidence is worthy of consideration. It is 15 now known that the correct legal test is whether a differential feature is (alone or in combination with others) liable to have a considerable or significant influence on the decision of the average consumer to use one service or the other. There are at least five types of evidence which are worthy of consideration by the Tribunal. (In identifying five I am seeking to identify a 20 sufficient body of evidence to justify the view I have reached: I am not seeking to give an exhaustive list, and I anticipate that the whole of the evidence (documentary, written and oral) will need to be reviewed in the light of the law as now clarified by the ECJ.) 25 22. First, the evidence of Mr Appleton dealt with the impact of the existence of FOBTs in licensed betting offices upon the income from the Part III gaming machines located on his own company’s premises. It cannot be read as unequivocally demonstrating that FOBTs were comparable to Part III gaming machines from the point of view of the average consumer and met the same 30 needs of that consumer (although that might be its true effect). It might also be read as demonstrating that FOBTs made a significantly different appeal to the same market because of the then unlimited stakes and prizes on FOBTs.[23]Second, the evidence of Mr Kavanagh was that whilst both FOBTs and Part III 35 machines offer gambling activities and both might appeal to players who liked such gambling there were differences from a player’s perspective; and that whilst he could not say with certainty to what extent players regarded them as interchangeable there must have been something about FOBTs which made them very different (because of the rapid growth in their popularity). The 40 Tribunal found that to a player they were “all just gambling machines”. But this finding was made in the context of holding that differences in stakes and prizes and the format of games were not legally relevant. Mr Kavanagh’s evidence is fit for consideration in the context of the correct legal test. 45 24. Third, the evidence of Mr Simon Thomas was that from the player’s point of view “it’s gambling and getting value for money”. He said:- 10 “The importance to the player is value for money, and the players will move between different games within an arcade within the gaming space, move from the arcade to the bookies next door, to the bingo hall, dependent on what they want at the time. I suppose in the same 5 way if you like sweets, you don’t always want to eat chocolate. Sometimes you want jellies or dark chocolate or your taste of the moment”. Elsewhere he said:- 10 “From the players point of view, all these machines are gambling machines and what the player wants is value for money, and the value for money is the amount of entertainment they get out of it, and people have different measures of entertainment. It might be the amount of 15 time they can spend for their money, or the thrill they get, or whatever, for the amount of money they spend. And it is a combination of the speed of play, the actual percentage return, and the amount of money a person is playing each game. From that you’ll get a cost of play, and against that the player bases their enjoyment. If you don’t give them 20 value for money, they don’t play”. That may or may not support a conclusion that differences relating to the minimum and maximum stakes and prizes, the chances of winning, the formats available and the possibility of interaction between the player and the 25 slot machine do have a considerable influence on the decision of the average consumer: I express no view, save that the evidence is fit for consideration in that connection.[25]Fourth, the body of evidence from various witnesses that users of FOBTs also 30 played Part III gaming machines does not compel the conclusion that the two types of machine were comparable from the point of view of the average consumer and met the same needs of that consumer. It might support that conclusion if the evidence was that playing the two types was indiscriminate and unthinking. It might not support that conclusion if the evidence was that a 35 deliberate or discriminating choice was made to play the one and then the other.[26]Fifth, in its submission to be Gaming Board in relation to the 2004 Triennial Review of stake and prizes limits for Part III machines the British Amusement 40 Catering Trades Association sought an increase in the stake and prize limits for Part III machines on the grounds that “customers will further migrate to the perceived superior offer of [FOBTs] which could result in the closure of many pubs clubs etc”. This is material fit for consideration when addressing the question whether stakes and prizes were liable to have a considerable 45 influence on whether a consumer chose one machine over another. 11[27]This evidence must be considered along with all of the other evidence: and if it is I am not certain that the outcome would be the same as that reached by Tribunal when excluding some material as legally irrelevant and when applying a differently expressed test from that required by the ECJ. 5 Accordingly I shall allow the appeal and set aside the decision of the Tribunal.[28]The question then is whether I should “re-make” the decision, or remit it to the Tribunal. Rank urged that I should re-make the decision on the footing that all the relevant evidence had been adduced and that the Tribunal (which had 10 considered the evidence nearly 3 years ago) possessed no advantage over the Upper Tribunal. HMRC urged that I should remit the matter and to a differently constituted First Tier Tribunal (taking into account the factors suggested as relevant in Sinclair Roche & Temperly v Heard [2004] IRLR 763 at paragraph [46]). 15[29]I propose to remit the case to the First Tier Tribunal but to give no direction that the Tribunal be differently constituted. (It may, because of retirement of members, have to be differently constituted: and if that is so then I would still remit rather than myself re-make the decision). 20[30]First, as a matter of policy I think that in general the facts should be found by the specialist tribunal created for that purpose. Second, the evidence in this case requires careful analysis and interpretation, and seems to me dependent on inferences to be drawn from primary fact (given that the task is to assess 25 what is liable to be of considerable significance to the average consumer). It is therefore particularly a case for the specialist tribunal to bring to bear its particular skills and expertise. Third, whilst memories will be faint (given the passage of time) and whilst the constitution of the Tribunal may not through force of circumstance be the same, the case is of such significance that any 30 advantage to be gained from the Tribunal’s having considered the documents in depth and heard evidence and argument over some 9 days (of which a transcript is available) should be exploited. Fourth, the case is of such significance that preservation of the full appeal route is appropriate. Fifth, this is not a case where the hearing before the Tribunal was wholly flawed or 35 completely mishandled. The Tribunal made an error of law in its analysis of the decision in Linneweber and faithfully applied a test which has now been reformulated by the ECJ. The terms of the reformulation now make it clear what has to be considered, decided and explained. I share the view expressed by Burton J in Sinclair Roche (supra) at paragraph 46.6: 40 “…where a tribunal is corrected on an honest misunderstanding or misapplication of the legally required approach (not amounting to a “totally flawed” decision…..), then, unless it appears that the tribunal has so thoroughly committed itself that a rethink appears impracticable, there can be a presumption that it will go about the tasks 45 set them on remission in a professional way, paying careful attention to the guidance given to it by the appellate tribunal”. 12[31]I will therefore remit the case to the First Tier Tribunal for rehearing in the light of the ruling of the ECJ on the Reference. So that all parties know where they stand I will direct that all evidence tendered and led in connection with the first hearing that resulted in the decision of 11 December 2009 (including 5 the transcript of the oral evidence) (“the original evidence”) shall stand as evidence at the rehearing: and that subject to any further direction of the First Tier Tribunal the evidence at the rehearing shall be limited to the original evidence. As to any further direction of the Tribunal I would emphasise that the Tribunal has its full powers under the Rules to do what is just in the case. 10 The direction I have given is not intended to fetter that power, nor is this caveat intended to encourage any particular attitude to the admission of evidence beyond the original evidence. 15 TRIBUNAL JUDGE: Mr Justice Norris RELEASE DATE: 04 October 2012
Cited in 4 later judgments