The Commissioners for HM Revenue and Customs v Lok'nStore Group Plc: [2014] UKUT 0288 (TCC) [2014] UKUT 0288 (TCC)

UPPER TRIBUNAL
TAX AND CHANCERY CHAMBER
[2014] UKUT 0288 (TCC)Case No FTC/05/2013
Between
The Commissioners for HM Revenue and CustomsAppellantLok'nStore Group PlcRespondent
Before
MR JUSTICE HENDERSONDate 23 June 2014Category: Tax
[1]This is a case about the methodology to be applied in apportioning residual input value added tax (“VAT”) incurred on supplies of goods or services which are used by a taxable person for the purposes of making both taxable and exempt output supplies. The input tax in question is “residual” because it cannot be exclusively attributed to either taxable or exempt supplies made by the taxable person. Where input tax is exclusively attributable to taxable supplies, a trader is entitled to deduct it in full from the output tax due on his taxable supplies. Conversely, where input tax is exclusively attributable to exempt supplies, none of it is deductible. Where, however, a trader incurs input tax on supplies (typically overheads) which are used, or to be used, by him in making both taxable and exempt supplies, the input tax has to be apportioned. Only the portion of this residual input tax which is apportioned to the taxable supplies is deductible: the balance, apportioned to the exempt supplies, is not.[2]The default, or “standard”, method of apportioning residual input tax is laid down by regulation 101(2)(d) of the Value Added Tax Regulations 1995, SI 1995/2518 (“the VAT Regulations”). In its current form, this provides that:
“(d) … there shall be attributed to taxable supplies such proportion of the residual input tax as bears the same ratio to the total of such input tax as the value of taxable supplies made by [a taxable person] bears to the value of all supplies made by him in the period.” “Residual input tax” is defined in regulation 101(10) as meaning: “input tax incurred by a taxable person on goods or services which are used or to be used by him in making both taxable and exempt supplies.”
It can be seen, therefore, that the standard method is based on turnover. The proportion of residual input tax which is deductible is the same as the proportion which the value of the taxable person’s taxable supplies bears to the value of all the supplies (both taxable and exempt) which are made by him in the relevant period.[3]The standard method must be used unless, pursuant to regulation 102 of the VAT Regulations, the Commissioners for Her Majesty’s Revenue and Customs (“HMRC”) either approve or direct the use of a different method. A different method of this kind is usually called (although not in the legislation 4 itself) a partial exemption special method, or “PESM”. By virtue of regulation 102(9), HMRC shall not approve the use of a PESM:
“… unless the taxable person has made a declaration to the effect that to the best of his knowledge and belief the method fairly and reasonably represents the extent to which goods or services are used by or are to be used by him in making taxable supplies.”
This requirement brings out the fundamental point that the function of a PESM, as of the standard method itself, is to produce a fair and reasonable apportionment of input tax which reflects the use made by the taxable person of the relevant goods or services in making taxable supplies. In the case of residual input tax, where direct attribution to supplies which are either exclusively taxable or exclusively exempt is impossible, the search is always for an apportionment which captures, as fairly and reasonably as possible, the actual use of the relevant goods or services in making taxable supplies.[4]In the present case, the taxable person is a company called Lok’nStore Group Plc (“LnS”). As its name suggests, LnS provides self-storage services to businesses and the general public at purpose-built stores. It operates 21 such stores in the south-east of England. The supplies of storage space are taxable at the standard rate of VAT, as are certain ancillary supplies made by the company (the hire of vans to storage customers, and the sale of storage-related products such as bubble-wrap, tape and boxes). These taxable activities typically account for well over 90% of the company’s turnover.[5]Customers who store goods with LnS are required to declare their maximum value and to insure them while they are in storage. Some customers already have suitable insurance, or prefer to arrange it elsewhere, but those who do not are obliged to buy it from LnS. The insurance cover is sold pursuant to a block insurance policy taken out by LnS with Brit Insurance Limited for a fixed annual premium. The block policy entitles LnS to offer cover to customers up to a pre-set limit. The price charged by LnS for insurance is currently £1 per week per £1,000 of goods insured; before November 2008, the price was 75p per week. There is a fixed excess of £100, and the terms of the insurance policy are non-negotiable.[6]These supplies of insurance by LnS are agreed to be exempt from VAT. They typically account for between 4 and 7 % of the company’s total turnover, and make a significant contribution to the company’s gross and net profit.[7]Because LnS makes both taxable and exempt supplies, an apportionment of its residual input tax is necessary. HMRC have always taken the view that the 5 standard method of apportionment should be applied, but in 2007 and 2008 LnS put forward various proposals for PESMs which it argued would achieve a fairer and more reasonable attribution than the standard method. The matter was debated exhaustively in correspondence, but HMRC remained unpersuaded. Finally, in a letter dated 10 June 2009, HMRC rejected the PESMs proposed by LnS, and confirmed an assessment for VAT in the sum of £140,899, relating to the quarterly periods from 04/05 to 04/07, which HMRC considered was not recoverable in those periods pursuant to the standard method.[8]The company asked HMRC to review the decisions, but they were upheld in a letter dated 21 August 2009. LnS then appealed against the review decision. The appeal was heard by the First-tier Tribunal (Tax Chamber) (Judge Greg Sinfield and Nigel Collard), sitting in London, on 2 and 3 July 2012. By this stage, the company relied on only one of the four PESMs which it had previously proposed, namely a method which depended on a mix of floor space and values. This method had first been submitted in June 2008, and had been amended by a letter dated 15 December 2008. The company was represented at the hearing by counsel, Mr Andrew Hitchmough (now QC) and Mr Thomas Chacko, instructed by the company’s accountants, Baker Tilly. HMRC were also represented by counsel, Mr Sarabjit Singh.[9]By their written decision released on 14 September 2012 (“the Decision”), the FTT decided that the PESM proposed by the company produced an attribution which was fairer and more reasonable than the attribution that would result from the standard method. The company’s appeal was therefore allowed. The FTT refused HMRC permission to appeal, for reasons given in a further decision by Judge Sinfield released on 21 November 2012. HMRC then renewed their application for permission to the Upper Tribunal (Tax & Chancery Chamber), relying on detailed grounds of appeal settled by counsel, with Mr Owain Thomas now leading Mr Sarabjit Singh. Permission was granted by Judge Colin Bishopp of the Upper Tribunal on 11 January 2013. In his short decision notice issued on 15 January 2013, he said he had come to the conclusion that, taken together, the grounds advanced were arguable.[10]An appeal to the Upper Tribunal from a decision of the FTT can only be made on a point of law: see section 11 of the Tribunals, Courts and Enforcement Act 2007. Facts[11]The FTT heard oral evidence from two witnesses, Mr Raymond Alan Davies for LnS and Mr Alexander James Sherwood for HMRC. Mr Davies is a 6 chartered accountant who has been the company’s finance director since he joined it in January 2004. Mr Sherwood is a senior compliance accountant with HMRC, for whom he has worked since 2002. He is a member of the Association of Chartered Certified Accountants, with particular expertise relating to the application of UK and international accounting standards and company law. He is not, however, a VAT specialist, and his evidence was mainly directed to an examination of the way in which the company had presented its business in its financial statements and management accounts, and what the accounting evidence showed about the economic reality of the business.[12]The witness statements of Mr Davies and Mr Sherwood were admitted as evidence in chief, and are included in the bundles for the present appeal to the Upper Tribunal. Each witness was also cross-examined at some length, but I have not been provided with any transcript or record of the cross-examination. On the basis of the witness and documentary evidence before them, the FTT found the material facts as follows (paragraphs 8 to 16 of the Decision): “8. LnS operates 21 self-storage facilities or stores in the southeast of England from which it provides self-storage services to the general public and to businesses. The stores have been purpose-built by LnS. They are, typically, large buildings with rather plain exteriors distinguished by panels painted in the company colour of bright orange. They have a ground floor and, usually, two or more other floors of storage space, divided into steel windowless rooms with wire mesh ceilings. On the ground floor of each floor is a reception area where staff deal with existing and potential customers. All stores have CCTV monitoring and a secure perimeter. This appeal is concerned with the deduction of VAT on the overhead costs associated with the construction, maintenance and operation of LnS’s stores.[9]LnS grants customers licences to store goods in the stores, usually in lockable steel containers but sometimes in open covered containers and (for very large items) on pallets in storage areas. Some of the stores have outside storage. Customers either provide their own padlock to secure the units or they can buy a padlock and key from the store. With rare exceptions, LnS does not keep keys to the customers’ units. They units range in size from 25 to 10,000 square feet. LnS staff advise potential customers how much space they are likely to need and there is also information on this on LnS’s website. LnS charges different amounts per square foot in different stores and ground floor space is often charged at a higher rate. It was common ground that the provision of storage by LnS is a supply of services chargeable to VAT at the standard rate. 7[10]LnS also hires vans to its storage customers. The van hire operates at a loss and is used to encourage people to rent storage space. The vans are hired out at £10 per day for those moving in, and £40 per day for customers who have already moved in. It is common ground that the van hire is chargeable to VAT at the standard rate. LnS also sells products related to storage, such as bubble-wrap, tape and boxes. These products are sold to anyone who walks into the store, not just those renting storage space. Any walk-in customers are asked about their storage needs and encouraged to lease storage space where appropriate. These supplies of storage-related products are also standard rated for VAT.[11]LnS provides some of its customers with insurance for their goods while they are stored with LnS. It was common ground that the supplies of insurance are exempt. LnS takes out a block insurance policy with Brit Insurance Ltd for a fixed premium payment, which entitles LnS to offer a single insurance cover product, up to a pre-set cover limit, to storage customers. Insurance is only provided to customers and only covers the goods (not the container or space) while they are in storage and not while they are in transit or elsewhere.[12]Insurance is currently supplied at a fixed price of £1 per week per £1,000 of goods insured, with a minimum of £2,000 worth of insurance. Prior to November 2008, the price was 75p per week per £1,000 of goods. The price increase was driven by market forces rather than by any other factor. LnS set the price of its insurance by reference to the amounts charged by its competitors for similar insurance.[13]Customers taking out insurance are required to declare the maximum value of goods they are storing. There is a fixed excess of £100 and the terms of the insurance policy are not negotiable. It is a requirement of the Self-Storage Association, of which LnS is a member, that all customers must insure their goods. Customers cannot move their goods into storage at an LnS store until they either show that they already have suitable insurance or buy it through LnS.[14]Insurance is discussed with customers when they discuss their storage requirements and before any documents are signed. It is possible that insurance may be discussed when a customer is being shown round the storage units but we find that this was not what usually happened and, in any event, insurance agreements were always concluded in the reception areas. 8[15]Not all customers purchase insurance from LnS. In the Annual Report and Accounts for the year ended 31 July 2011, the chief executive’s review stated that, during the year, over 86% of new customers took LnS’s insurance. The review also stated that ancillary sales accounted for 9.9% of storage revenues in the year and were increasingly focused on insurance which increases overall margin. In each of the last seven years, insurance turnover was between 4.1% and 6.8% of LnS’s total turnover. The insurance sales contribute significant profit to LnS but the business would still be profitable and sustainable without it.[16]LnS’s management accounts and financial statements do not allocate costs to insurance or other ancillary sales.” The PESM proposed by LnS[13]The FTT described the PESM proposed by LnS as follows (paragraphs 25 to 27 of the Decision): “25. The PESM proposed by LnS replaces the standard method’s turnover-based calculation with a method that uses floor space as the proxy for the use of VAT-bearing costs together with a turnover element for those parts of the stores used for taxable and exempt supplies (i.e. the reception areas). The PESM was set out in detail in a letter of 15 December 2008 to HMRC.[26]Under the proposed floor space and values PESM, input tax is directly attributed to taxable and exempt supplies as far as possible and deducted or not accordingly. Input tax that is not directly attributable to either taxable or exempt supplies is attributed to taxable supplies in the proportion which “taxable floor space” in LnS’s stores bears to “total floor space”. “Taxable floor space” for this purpose means areas of the stores used for making taxable supplies of storage space to customers. The PESM states that the only areas that are used for making both taxable and exempt supplies are the reception areas. The floor space of the reception area is further reduced to reflect the area used exclusively for making taxable supplies. The remaining mixed use floor space of the reception area is apportioned between taxable and exempt use in accordance with the ratio of the value of LnS’s taxable supplies to the value of all its supplies. 9[27]On this basis, LnS calculated that only 0.02% of its income (sic, but the FTT clearly meant “input tax”) was attributable to insurance sold through the reception areas. The result of the PESM is that LnS would be entitled to deduct 99.98% of VAT incurred on the construction, maintenance and operation of the stores.”[14]Baker Tilly’s letter of 15 December 2008 to HMRC included these general comments under the heading “Cost components and direct attribution”:
“As we have previously stated, there is no link between the storage space and the sale of insurance. The costs relating to the storage area are not cost components or in any way attributable to the insurance commission. The insurance is sold independently of the storage and many customers use their own insurance. … The insurance element of our client’s business is ancillary to its core business of self-storage. Insurance is not sold separately or in its own right. As stated above, a customer will purchase insurance from Lok’nStore only if their goods are not covered by their own domestic or business insurance. Insurance is sold only to customers storing their items in Lok’nStore and will not be sold to other people or to customers who have already obtained insurance cover elsewhere. The activity is purely subsidiary and subservient to the storage business.”
[15]The relevant PESM was described thus in the body of the letter, under the heading “Mixture of Floor Space and Values Method”:
“… The first step is attribution of directly attributable costs. The second step would be to identify the floor space directly attributable to the four areas of the business, being:  Self storage  Retail/packing materials  Van hire  Insurance The bulk of floor space will be “back of office” i.e., where the actual storage takes place. Floor space that cannot be attributed solely to one of the income generating activities would 10 represent a “non-attributable space”
. This would, in practice, be the front office of the storage units. Residual VAT would be allocated to the various areas by looking at the proportion of floor space allocated to each income generating activity. VAT allocated to the nonattributable space would be recoverable by reference to a values based calculation.”[16]The proposed method was described in more detail in appendix D to the letter, from which I quote the following extract: “7. Taxable floor space means floor space that is used for making taxable supplies of storage space to customers.[8]For the purposes of calculating the proportion of total floor space that relates to both taxable and exempt supplies, the reception area of each store is the only area that makes both taxable and exempt supplies.[9]The floor space in the reception will be further reduced to reflect the area that is used exclusively for making taxable supplies of goods [I was told that this referred to the part of the reception area used for the display of retail goods].[10]The remaining reception floor space (the “residual floor space”) will then be apportioned between the ratio of taxable supplies to the ratio of exempt supplies using the following formula …” The formula used was the turnover-based formula of the standard method; but whereas the standard method would apply the formula to the entirety of the residual input tax, the PESM applied it only to the minute proportion of the input tax yielded by the floor space calculation. The justification advanced by Baker Tilly for the floor space calculation was that the reception area was the only part of LnS’s premises which generated insurance income; and even the reception area had to be sub-divided, so as to exclude the retail display area. In this way the amount of non-deductible input tax was reduced to approximately 0.02%, as compared with between 4 and 6% under the standard method. This may at first sight seem a surprising result, bearing in mind that customers who did not have their own insurance were obliged to purchase it from LnS, and the sale of such insurance formed a profitable part of the company’s business, typically yielding between 4 and 7% of its turnover. 11 The law[17]This is not the occasion for a detailed review of the law on the attribution of input tax incurred by partially exempt taxable persons. The general principles have been clearly and consistently stated by the Court of Justice of the European Union (“the ECJ”) in a long line of cases, and the Court of Appeal has given guidance on the application of those principles in at least four significant domestic decisions: Customs & Excise Commissioners v Southern Primary Housing Association Ltd [2003] EWCA Civ 1662, [2004] STC 209 (“Southern Primary”); Dial-a-Phone Ltd v Customs & Excise Commissioners [2004] EWCA Civ 603, [2004] STC 987 (“Dial-a-Phone”); Mayflower Theatre Trust Ltd v Revenue & Customs Commissioners [2006] EWCA Civ 116, [2007] STC 880 (“Mayflower”); and Revenue and Customs Commissioners v London Clubs Management Ltd [2011] EWCA Civ 1323, [2012] STC 388 (“London Clubs”). Furthermore, the authorities (both European and domestic) have recently been submitted to a valuable chronological review by the Upper Tribunal (Vos J and Judge Herrington) in Revenue and Customs Commissioners v Volkswagen Financial Services (UK) Ltd [2012] UKUT 394 (TCC), [2013] STC 716 (“Volkswagen Financial Services”) at [48] to [78]. The Upper Tribunal gave its decision in Volkswagen Financial Services on 12 November 2012, some three months after the Decision of the FTT in the present case.[18]The FTT referred to the relevant provisions of the Principal VAT Directive (Directive 2006/112/EC), the Value Added Tax Act 1994 and the VAT Regulations in paragraphs [17] to [24] of the Decision. I will not repeat this material, none of which is controversial. I would merely comment that the FTT quote from regulations 101 and 102 of the VAT Regulations in their current form, as amended by SI 2009/820. The amendments have effect only in relation to input tax incurred on or after 1 April 2009, and therefore do not apply to the periods in issue in the present case. However, nothing of any substance turns on the amendments, so far as I can see, and both sides were content to argue the case by reference to the VAT Regulations in their current form.[19]The FTT went on to review the authorities in paragraphs [28] to [36] of the Decision. Like the FTT, I find it convenient to begin with the principles set out by Carnwath LJ in Mayflower at [9]:
“(i) input tax is directly attributable to a given output if it has a “direct and immediate link” with that output (referred to as “the BLP test”) [a reference to the decision of the ECJ in Case C- 4/94, BLP Group Plc v Customs and Excise Commissioners [1995] ECR I-983, [1996] 1 WLR 174]; 12 (ii) that test has been formulated in different ways over the years, for example: whether the input is a “cost component” of the output; or whether the input is “essential” to the particular output. Such formulations are the same in substance as the “direct and immediate link” test; (iii) the application of the BLP test is a matter of objective analysis as to how particular inputs are used and is not dependent upon establishing what is the ultimate aim pursued by the taxable person. It requires more than mere commercial links between transactions, or a “but for” approach; (iv) the test is not one of identifying what is the transaction with which the input has the most direct and immediate link, but whether there is a sufficiently direct and immediate link with a taxable economic activity; and (v) the test is one of mixed fact and law, and is therefore amenable to review in the higher courts, albeit the test is fact sensitive.”
[20]The Mayflower Theatre Trust was a charitable trust which operated the Mayflower Theatre in Southampton. The Trust bought in performances from production companies under separate production contracts, on which it paid VAT at the standard rate. The Trust derived most of its income from ticket sales for performances, which were exempt from VAT. However, it also carried on various other activities, such as the sale of confectionary, drinks and programmes, which were taxable and either standard or zero-rated. The basic issue was whether a direct and immediate link could be established between the input tax paid to the production companies and any of the Trust’s taxable supplies, or whether the only direct and immediate link was with the sale of tickets for the performances. If the input tax were solely attributable to the exempt supplies, none of it would be recoverable; but if the necessary link could be established with any of the taxable supplies, regulation 101(2)(d) of the VAT Regulations would be engaged and the input tax would have to be apportioned between the Trust’s exempt and taxable output supplies. The Court of Appeal held that there was a sufficiently close objective link between the purchase of productions and the sale of programmes (albeit zero-rated) to satisfy the BLP test, with the result that the input tax had to be apportioned. Because of the unusual factual circumstances in which the issue arose, there was no dispute that the apportionment had to be performed on the standard basis, even though the sale of programmes constituted only a tiny proportion of the Trust’s taxable activities: see at [19] and [20] per Carnwath LJ, and [68] and [69] per Chadwick LJ. 13[21]There are two points about the decision in Mayflower which it is worth emphasising. First, as I have just explained, it was not a case about the apportionment of residual input tax, but rather about the logically prior question whether the input tax was exclusively linked to the Trust’s exempt supplies (in which case it was common ground that none of it could be deductible). Secondly, the case illustrates that there are two different types of situation in which residual input tax can arise. The first type of case, exemplified by Mayflower itself and Dial-a-Phone, is where a direct link can be established between the relevant input tax and specific taxable and exempt supplies made by the taxable person. These are cases of what Carnwath LJ called “specific attribution”, where the input tax is directly attributable to both taxable and exempt supplies. A second type of case comprises overheads properly so-called, where no direct link can be established with any specific supplies made by the taxable person, but the expenditure in question is incurred in the course, and for the purposes, of the business viewed as a whole. As Carnwath LJ explained at [27] to [34], the ECJ caters for the practical need to accommodate overheads within the partial exemption rules by the expedient of saying that they have a direct and immediate link with the whole economic activity of the taxable person, even though they cannot be attributed to particular supplies. Carnwath LJ commented at [33]:
“The special treatment of “overheads” or “general costs” serves a particular and limited purpose in the VAT system, for those inputs which would not otherwise be brought within the calculation. It should not be extended beyond that purpose.”
[22]Although not explicitly stated in the Decision, I understand it to be common ground in the present case that all of the disputed input tax was incurred on overheads in the usual sense of that term, and none of it was specifically attributable to both exempt and taxable supplies. The only description given by the FTT of the nature of the relevant overhead costs is that they are “associated with the construction, maintenance and operation of LnS’s stores”: see paragraph [8] of the Decision, quoted above.[23]After Mayflower, the next case referred to by the FTT was Southern Primary which they cited for the proposition that the “direct and immediate link” test requires more than satisfaction of a “but for” test of causation. Counsel for LnS had submitted that there was “a direct read across” from Southern Primary to the present case, whereas Mr Sarabjit Singh for HMRC had submitted that the case could not be read across “because it was about direct attribution to specific outputs whereas LnS’s appeal concerned overheads which could not be attributed to particular outputs”. The FTT continued, at [31]:
“It is correct that Southern Primary did not concern overheads but, as Skatteverket v AB SKF Case C-29/08 [2010] STC 419 14 (discussed further below) shows, the “direct and immediate link” and “cost component” tests are also relevant when considering overheads. We accept Mr Hitchmough’s submission that “but for” is not the test for attribution of VAT on overheads. It follows from Southern Primary that the fact that LnS would not have made supplies of insurance if it did not have facilities to store the insured goods is not the correct test.”
[24]The FTT would in my judgment have been entirely correct to direct themselves that a “but for” test of causation was insufficient at the attribution stage of the analysis, had attribution to specific outputs been in issue. But it was common ground that the relevant input expenditure was on overheads, which ex hypothesi were not attributable to specific outputs but only to the business as a whole. The relevance of the FTT’s comment about Southern Primary is therefore not immediately apparent to me. I will also need to return to their view that the decision of the ECJ in the SKF case shows that the “direct and immediate link” and “cost component” tests are also relevant when considering overheads. It is not obvious that those tests should have any role left to play once the stage of apportioning residual input tax has been reached. Where (as in the present case) the residual input tax is attributable to overheads properly so-called, those tests will necessarily have been satisfied, but only in the somewhat artificial sense laid down in the ECJ’s jurisprudence, namely that the direct link is with the business viewed as a whole. Once that stage has been reached, an apportionment of the relevant input tax must be made, either in accordance with the standard method or with a method which better reflects the use actually made by LnS of the relevant overheads.[25]The leading English authority on the principles to be applied at the stage of apportioning residual input tax is the decision of the Court of Appeal in London Clubs, where the only reasoned judgment was delivered by Etherton LJ (with whom Pitchford and Ward LJJ agreed). The taxable person in London Clubs was the representative member of a VAT group which operated 11 casinos in the United Kingdom. Following legislative changes, the group acquired premises with greater floor space which it used to provide catering as well as gaming facilities. Under its existing PESM, residual input tax was apportioned on a turnover basis, with an adjustment to take account of the fact that some customers were provided with food and drink free of charge. The group then proposed a new PESM, which would move from a turnover based method to a floor space method. As Etherton LJ said, at [17]:
“The fraction to be applied to the residual input tax under the proposed PESM is, in simple terms, the area of floor within the respondent’s premises occupied to make taxable supplies over the area of floor occupied to make taxable and exempt supplies, 15 again with an adjustment to take account of residual costs associated with non-charged food and drink.”
The proposal was rejected by HMRC, who considered that it was not fair and reasonable, and certainly not more fair and reasonable than the existing method. The proposed PESM was, however, upheld by the FTT, which made a crucial finding of fact that the group’s catering activities were businesses in their own right and not merely ancillary to the gaming business. This decision was upheld both by Proudman J in the Upper Tribunal and by the Court of Appeal, on the basis that it involved no error of law.[26]The relevant legal principles were stated as follows by Etherton LJ at [33] and [34]: “33. The need for a process of attribution only arises where an item is a cost component (within art 2 of the First Directive) of two supplies, one taxable and one exempt … If the standard (turnover) method does not result in a fair and reasonable attribution of the cost component, the search is for a more fair and reasonable method of attribution. The onus is on the taxpayer to show that the proposed PESM is more fair and reasonable, that is to say, more accurate: Royal Bank of Scotland Group Plc v Revenue and Customs Commissioners (Case C-488/07) [2009] STC 461, para 24 of the judgment.[34]A fair and reasonable attribution to a taxable supply must, for the purposes of art 17(2) and (5) of the Sixth Directive and reg 101(2)(d) of the Regulations, reflect the use of a relevant asset in making that supply. In assessing that use, and its extent, consideration is not limited to physical use. The assessment must be of the real economic use of the asset, that is to say having regard to economic reality, in the light of the observable terms and features of the taxpayer’s business.”[27]Etherton LJ went on to say that these principles had been “well captured and applied” by Warren J in St. Helen’s School Northwood Ltd v Revenue and Customs Commissioners [2006] EWHC 3306 (Ch), [2007] STC 633 (“St. Helen’s School”). Etherton LJ then explained why he agreed with Warren J’s approach and analysis in that case, in a passage which is of such central relevance to the present case that I will quote it in full: “35. … In that case the taxpayer school was granted planning permission to build a new swimming pool and sports hall. It envisaged commercial use of the complex as well as school use. It set up a company, which was to use the complex for club and community purposes outside the school hours. The school 16 was registered for VAT in order to recover VAT on the building of the sports complex. The school proposed a PESM, namely a percentage recovery of VAT based on the number of hours of actual use by the company as a proportion of the total hours of use. The Commissioners refused the school's proposal. The school's appeal to the Value Added Tax and Duties Tribunal ("the Tribunal") was dismissed. Warren J dismissed the school's appeal to the High Court. At [60] he referred to the decision of Patten J in Customs and Excise Commissioners v Yarburgh Children's Trust [2002] STC 207, in which it was held that the motive of a person in making a supply is not relevant to, and cannot dictate, the correct VAT treatment of a transaction. Warren J said (at [60]) that the exclusion of motive or purpose did not allow the Tribunal to disregard the observable terms and features of the transaction and the wider context in which it came to be carried out. He said that applied in the context of establishing the use (for VAT purposes) to which an item of property is put and in deciding whether a proposed PESM is fair and reasonable when determining what is or is not a valid proxy for that use. I agree.[36]Warren J accepted (at [75]) the submission of counsel for the Commissioners (at [63]) that physical use may reflect economic use, but does not necessarily do so, and that any allocation or special method must give a credible result in economic terms.[37]Warren J applied that approach, and the concept of economic use, in his analysis of the facts. He said: [75] I agree with Mr [Roger] Thomas [counsel for the school] that the search in the present case is for a fair and reasonable proxy for the 'use' of the sports complex in making the exempt and taxable supplies made by the School. However, I also agree with Miss Simor that the physical use of the complex is not necessarily a fair and reasonable proxy for that use. I consider that her use of the phrase 'economic use' is a helpful approach to establishing what the search is for. [76] In that context, it is instructive, I consider, to look at the position had the School not granted the licence at all and had not allowed any out-of-hours use. In those circumstances, there would have been no taxable supply at all. In consequence, none of the input tax would fall to be attributed to taxable supplies as a result of regs 101(2)(b) and (c), reg 101(2)(d) not applying. However, the sports complex is used for the purposes of the School's (exempt) business. It is so used not because there is a supply to parents of the physical use (by their daughters) of the 17 sports complex to their children, but because the availability of the complex is part of the package of benefits which is acquired by parents for the fees they pay and which constitutes the exempt supply by the School. The use made by the School, for VAT purposes, of the sports complex is its use in providing that package of services, a single supply. There is, of course, no need to identify a proxy for use when there is only an exempt supply since questions of allocation under reg 101(2)(d) do not then arise. Nonetheless, one can see that the 'use' referred in reg 101 (as elsewhere) is not physical use but some special VAT use. It is, I think, the same as what Miss Simor terms 'economic use'. [77] On the facts of the present case, it seems to me that the overwhelming economic use of the sports complex by the School is in relation to the provision of educational services. In that context, I agree with Miss Simor that the source of funds and the purpose of constructing the sports complex are relevant considerations. To regard those factors as relevant is not, in my judgment, to fall into the error, as Mr Thomas would say it is, of categorising the nature of a supply by reference to the purpose or motive in making it. There is no doubt that in the present case, the supplies are distinct and readily identifiable, that is to say the taxable supply of the licence to [the company] and the exempt supply of education. Nor, in my judgment, is there any question, in taking those factors into account of treating a taxable supply as an exempt supply or vice versa. The question is what 'use' is being made of the inputs in producing the outputs. It seems to me that the purpose of the School, objectively ascertained, in constructing the sports complex is a highly relevant factor in attributing cost components between the relevant outputs and is an entirely different issue from identifying the nature of the output by reference to purpose or motive (which is inadmissible), the issue addressed by Patten J in Customs and Excise Comrs v Yarburgh Childrens Trust [2002] STC 207.” (Warren J’s emphasis).[38]I agree with Warren J's approach and analysis. He went on to say (at [78]) that, on the evidence, it was clear that, objectively assessed, the principal purpose of the school in building the sport complex was the furtherance of its educational activities and was carried out in connection with its business of making exempt supplies of education; and, further, the capital cost of the complex was met out of funds which were either charitable funds or derived from a fund-raising exercise and which were clearly dedicated to the educational purposes of the school. He also concluded (at [79]) that the income generated by the licence to the company was never 18 intended or expected to meet a share of the capital cost proportionate to the physical use of the sports complex by the company. The licence to the company was simply putting to productive use that which had been acquired for a different main purpose. In Warren J's judgment (at [80]) the standard method produced an allocation which was more fair and reasonable than the school's proposed PESM.[39]Warren J's endorsement of a test of economic use anticipated the emphasis of the [ECJ] on "economic reality" in Joined Cases C-53/09 and C-55/09 Revenue and Customs Commissioners v Loyalty Management UK Ltd, Baxi Group Ltd v Revenue and Customs Commissioners [2010] STC 2651, which concerned the VAT treatment of supplies under customer loyalty reward schemes. The ECJ said at [39]:
“It must also be recalled that consideration of economic realities is a fundamental criterion for the application of the common system of VAT ...””
[28]Etherton LJ then distinguished the decision of the VAT and Duties Tribunal in the superficially similar case of Aspinall’s Club Ltd v Revenue and Customs Commissioners (2002) VAT Decision 17797, where a floor space PESM had been rejected by the tribunal on the basis (shortly stated) that the catering activities were not themselves conducted with a view to profit, and were truly ancillary to the gaming which was the foundation of the business. Etherton LJ continued:
“41. That case and the reasoning of the tribunal, with which I agree, is illustrative of three points of principle. First, it shows the importance in these cases of close attention to the facts in order to understand the economic or commercial reality underlying the use of the relevant VAT inputs. Secondly, identification of the source or potential source of profit in a business may be an important feature of a business throwing light on whether or not the standard method or a PESM is a more fair, reasonable and accurate method of attribution. It all depends on the facts of each case … Thirdly, depending again on the precise factual situation under consideration, the approach of the tribunal in Aspinall’s Club (see para 49) may well be appropriate in a case where the taxable supplies are not, in themselves, a source of profit: “49. … Those costs are funded by the gaming. That in itself does not make them cost components of those exempt supplies. But in this case it is additional proof, if any is needed, that gaming is the foundation of the business and it 19 is the furtherance of that gaming which causes and is seen as justifying commercially the decisions to incur the expenditure …”
[42]As both St. Helen’s School and Aspinall’s Club show, and as was emphasised in [Dial-a-Phone] at [72] by Parker LJ (with whom the other members of the court agreed), analysis of attribution for the purposes of art 2 of the First Directive, art 17 of the Sixth Directive and reg 101 is highly fact sensitive.”[29]Having considered the facts in detail, and the submissions of the parties, Etherton LJ concluded that the decision of the FTT could not be disturbed, although he said at [71] that the critical finding of fact by the FTT (referred to in paragraph 25 above) struck him as “remarkably benign, that is to say surprisingly favourable to the respondent”. He further said at [73], in relation to the implicit finding by the FTT that catering was a potential source of future profit, even though it had been significantly loss-making to date and the court had been shown no material to indicate that there was any realistic prospect of profit from catering in the foreseeable future: “73. … That is a specific finding of primary fact on the evidence. It is a fact which feeds into the enquiry as to the economic use of the relevant overheads: it is not a conclusion which results from the test itself. If it is to be challenged as a perverse finding of fact, then the perversity must be raised as a distinct ground of appeal. That is particularly important in the case of an appeal from a specialist tribunal, with whose expertise an appellate court should only interfere with caution: Procter & Gamble UK v Revenue and Customs Commissioners [2009] EWCA Civ 407, [2009] STC 1990.[74]Further, as Mr Hitchmough rightly emphasised, in order to ascertain the reasoning of the FTT, the decision must be read as a whole. It is not to be interpreted like a statute drafted by Parliamentary Counsel. Its reasoning and sense are to be gathered by a fair reading of its entirety. This is true of every judgment, but particularly so an expert tribunal which, like the FTT, includes non-lawyers.”[30]Reverting to the present case, the FTT discussed, and quoted extensively from, both St. Helen’s School and London Clubs at [33] to [35], before directing themselves as follows at [36]:
“36. It is clear from the passage cited above [from London Clubs] that the task for the Tribunal is to determine the use of the supplies on which the VAT is incurred by reference to economic or commercial reality. We bear in mind that the 20 profit which is derived from an activity may be relevant in determining whether a method produces a fair and reasonable attribution but that is not necessarily the case. As Etherton LJ observed in London Clubs Management at [84], “profit may be an important factor, but it is not necessarily so, and in some cases it may be entirely irrelevant”.”
I can detect no error of law in that self-direction, which appears to me firmly based on the authorities. The decision of the FTT[31]The FTT began their analysis by recording the main submissions on each side. For present purposes, the following brief summary will suffice. On behalf of LnS, it was submitted that the standard method does not work because it assumes that exactly the same amount of residual input tax is used in order to generate £1 of exempt income as to generate £1 of taxable income. By contrast, the proposed PESM reflected the true nature of the business (selling storage space) and the economic use of the overheads. The sale of storage space, not insurance, was the “driver” of the business, just as education was in St. Helen’s School and gambling was in Aspinall’s Club.[32]On behalf of HMRC, the proposed PESM was said to be flawed because it allocated all storage space exclusively to a taxable use, whereas the reality was that the storage space was attributable to both taxable and exempt supplies. The error was similar to that made by the school in St. Helen’s School. The premises were used for the purposes of the business as a whole, which included making exempt supplies of insurance. The sales of storage and insurance were negotiated at the same time, and were “inextricably intertwined”. Physical use of space was therefore not an appropriate proxy, as it failed to reflect economic use. Furthermore, the lack of allocation of costs in the management accounts showed that the overheads were used for the purposes of the business as a whole. The standard method was fair and reasonable, because it changed with the levels of turnover for exempt and taxable supplies.[33]The FTT then described its task: “40. The task for this Tribunal is to determine whether the standard method and the proposed PESM produce a fair and reasonable attribution of the supplies on which LnS has incurred VAT to taxable supplies by LnS. That requires us to form a view on whether the methods are accurate proxies for 21 apportionment according to use. If we conclude that both do so, then we must determine whether LnS has established that its proposed PESM is fairer and more reasonable, i.e. a more accurate proxy, than the standard method.[41]The starting point is use. This appeal concerns the deduction of VAT on the overhead costs associated with the construction, maintenance and operation of LnS’s stores. We must determine the extent to which the goods and services supplied to LnS in connection with the construction, maintenance and operation of its stores are used for transactions in respect of which VAT is deductible i.e. taxable supplies.[42]The meaning of “use” and the way it should be measured for VAT purposes was discussed by Warren J in St. Helen’s School. Warren J observed, at [75] that “physical use of the complex is not necessarily a fair and reasonable proxy … [and] … the phrase “economic use” is a helpful approach to establishing what the search is for.”[34]So far, the FTT’s approach to the issue cannot in my judgment be faulted, although it might have been useful to have a fuller explanation of the general nature of the relevant overhead costs, particularly those associated with the construction of the stores. As I have already explained, however, the agreed status of the costs as overheads must mean that they were not directly and exclusively linked with the taxable supply of storage, but were instead linked with the business as a whole, including the exempt supply of insurance.[35]The FTT then turned to the decision of the ECJ in the SKF case, to which they evidently attached considerable significance. I shall begin by quoting what the FTT said: “43. The term “economic use” is consistent with the analysis of the CJEU in the SKF case. SKF was the parent company of an industrial group which made taxable supplies. SKF proposed to sell shares in two of its subsidiaries in order to raise funds to finance other activities of the group. The SKF case concerned the deductibility of VAT incurred on services relating to the sale of shares. The issue was not simply whether the services were attributable to the sale of shares but also whether they were attributable to SKF’s business generally i.e. were overheads. At [57] – [58], the CJEU said: “57. According to settled case-law, the existence of a direct and immediate link between a particular input transaction and a particular output transaction or transactions giving rise 22 to entitlement to deduct is, in principle, necessary before the taxable person is entitled to deduct input VAT and in order to determine the extent of such entitlement … The right to deduct VAT charged on the acquisition of input goods or services presupposes that the expenditure incurred in acquiring them was a component of the cost of the output transactions that gave rise to the right to deduct …[58]It is, however, also accepted that a taxable person has a right to deduct even where there is no direct and immediate link between a particular input transaction and an output transaction or transactions giving rise to the right to deduct, where the costs of the services in question are part of his general costs and are, as such, components of the price of the goods or services which he supplies. Such costs do have a direct and immediate link with the taxable person’s economic activity as a whole …”[44]There is no dispute in this case that the VAT incurred on construction, maintenance and operation of LnS’s stores are part of its general costs i.e. are overheads. In the passage above, the CJEU makes clear that the “direct and immediate link” and “cost component” tests are also relevant when considering overheads. At [60], the CJEU set out how to apply the tests:
“It follows that whether there is a right to deduct is determined by the nature of the output transactions to which the input transactions are assigned. Accordingly, there is a right to deduct when the input transaction subject to VAT has a direct and immediate link with one or more output transactions giving rise to the right to deduct. If that is not the case, it is necessary to examine whether the costs incurred to acquire the input goods or services are part of the general costs linked to the taxable person’s overall economic activity. In either case, whether there is a direct and immediate link will depend on whether the cost of the input services is incorporated either in the cost of particular output transactions or in the cost of goods or services supplied by the taxable person as part of his economic activities.”
[45]At [62], the CJEU showed the national court how it should approach the issue: “In order to establish whether there is such a direct and immediate link, it is necessary to ascertain whether the costs incurred are likely to be incorporated in the prices of the shares which SKF intends to sell or whether they are only among the cost components of SKF’s products.”” 23[36]It is clear from this passage that the FTT regarded SKF as authority for the proposition that the “direct and immediate link” and “cost component” tests are relevant when considering the apportionment of residual input tax attributable to overheads. In my respectful opinion, however, that is a misreading of the ECJ’s decision in that case. The principles recited by the ECJ in paragraphs 57 to 60 of its judgment do little more than repeat the familiar learning which has to be applied in order to determine whether there is a right to deduct input tax at all, either because it is directly and immediately linked with a taxable output supply, or because it is directly and immediately linked with the taxable person’s economic activity as a whole. The Court was not concerned with the subsequent question of apportionment of residual input tax, because it did not even have sufficient information to determine whether the costs in question were properly to be characterised as overheads having the necessary link with SKF’s overall economic activity: see paragraphs 62 and 63 of the judgment. Thus the case is no authority at all on the principles to be applied at the apportionment stage of the exercise, which is the only stage in issue in the present case. I conclude, therefore, that in this respect the FTT made an error of law. Whether this error of law vitiated their conclusions is among the questions which I will have to consider.[37]The remainder of the FTT’s reasoning is contained in paragraphs [46] to [54] of the Decision, as follows: “46. Applying the CJEU’s guidance in SKF, in determining what the goods and services supplied to LnS in connection with the construction, maintenance and operation of its stores are used for, it is necessary to ascertain whether and, if so to what extent, the costs of such supplies are likely to be incorporated in the prices of LnS’s supplies to its customers. In our view, the actual or likely impact of the costs of overheads on the prices of LnS’s supplies not only establishes whether there is a direct and immediate link with those supplies but is also a useful measure of the extent of the economic use of the overheads.[47]First, we consider whether and the extent to which the overhead costs are incorporated in the price of the insurance. The evidence showed that LnS set the price of its insurance by reference to the amounts charged by its competitors for insurance rather than in response to any costs (not even the cost of the block policy). We find that the costs of constructing, maintaining and operating the stores did not materially affect and were not incorporated in the price of the insurance. We consider that there is some link between overhead costs and the sale of insurance simply because the insurance is sold in the reception areas of the stores and the overheads relate, in some part, to those areas. We could not determine the impact of such costs on the price of the insurance from the evidence before us 24 but, for the reasons given above, we consider that the impact of the cost of general overheads on the price of insurance must be very small. Accordingly, we conclude that LnS uses the goods and services supplied to it in connection with the construction, maintenance and operation of its stores in relation to the exempt supplies of insurance only to a very small extent.[48]The link between overhead costs associated with the construction, maintenance and operation of LnS’s stores and the taxable supplies of storage is easier to discern. In our view, if LnS opens a new store or enlarges or refurbishes a store then the overhead costs will increase. Not all customers purchase insurance from LnS and it follows that, if LnS is to recover them, the costs of the new or improved space are likely to be incorporated in the prices of the storage. We consider that costs of constructing, maintaining and operating the stores are linked to the price of the supplies of storage because expenditure on new stores and valuation of development projects is assessed in the LnS annual reports in terms of projected space rental levels and levels of occupancy and not by reference to projected sales of insurance. In our view, if the overhead costs increased then that would be likely to lead to an increase in the charges per square foot for storage.[49]Our conclusion is that LnS uses the goods and services supplied to it in connection with the construction, maintenance and operation of its stores almost exclusively for the purpose of making supplies of storage. This conclusion does not determine the appeal. Next we consider whether the methods provide a fair and reasonable determination of the amount of the VAT that is attributable to LnS’s taxable supplies and whether one method is fairer and more reasonable than the other.[50]As Etherton LJ stated in London Clubs Management at [34]: [The citation is set out][51]The standard method, found in regulation 101(2)(d) of the VAT Regulations, involves dividing the value of taxable supplies by the value of all supplies to arrive at a percentage figure, which is treated as the percentage of residual input tax that is attributable to taxable supplies. The application of the standard method in this case would result in 94% to 96% of LnS’s residual input tax being attributed to taxable supplies. The proposed PESM produces a level of taxable use of 99.98%. 25[52]HMRC contend that the level of taxable income to total income is generally a good measure of the economic use of goods and services. The greater the level of taxable income the greater the economic use of the overhead costs in making taxable supplies. Equally, the greater the level of exempt income, the greater the use of the overhead costs in making exempt supplies. In our view, that proposition only holds good where the relationship between the overhead costs and the income from the taxable and exempt supplies is, broadly, the same. If the costs of goods and services used to make exempt supplies are far greater than the costs of the goods and service[s] used to make taxable supplies then the use of a turnover method would lead to an over recovery of VAT on those costs. In such a case, the economic reality is that the use of goods and services is weighted towards the exempt supplies which cost more to make and consume more of the VATbearing overheads.[53]Further, we do not consider that the contribution to LnS’s profitability made by insurance sales and the, understandable, focus on increasing the volume of such a profitable line of business are relevant in determining the extent to which supplies relating to the construction, maintenance and operation of its stores are used by LnS to make supplies of insurance. The fact that a supply generates a large turnover or profit does not, by itself, indicate that the activity uses a high level of overheads.[54]In LnS’s case, we have found that the goods and services on which the residual VAT is incurred are used almost exclusively for the purpose of making taxable supplies of storage which is the main focus of its business. We consider that a fair and reasonable attribution of the residual input tax would show that the overheads were almost exclusively attributable to taxable supplies of storage. Although both methods attribute the majority of the overheads to taxable supplies and both might be considered to be fair and reasonable, the PESM proposed by LnS better reflects the economic use of the overheads by LnS and is, accordingly, a more accurate proxy than the standard method.”[38]The critical steps in the line of reasoning which led the FTT to their conclusion may, I think, be fairly summarised as follows:(a) The relevant overhead costs did not materially affect, and were not incorporated in, the price of the insurance (paragraph [47]). 26(b) There was, nevertheless, some link between the overheads and the sale of insurance “simply because the insurance is sold in the reception areas of the stores and the overheads relate, in some part, to those areas” (ibid.).(c) The impact of such costs on the price of insurance must have been very small, and LnS therefore used the relevant overheads in making its exempt supplies of insurance “only to a very small extent” (ibid.).(d) Conversely, the link between the overheads and the taxable supplies of storage is easier to discern, and is likely to be reflected in the prices charged for storage (paragraph [48]).(e) Accordingly, LnS uses the overheads “almost exclusively for the purpose of making supplies of storage” (paragraph [49]).(f) The fact that sales of insurance make a significant contribution to the company’s turnover and profit does not, by itself, show that the insurance sales use a corresponding level of overheads (paragraph [53]).(g) A fair and reasonable attribution of the residual input tax would allocate it “almost exclusively … to taxable supplies of storage”. Although both methods might be regarded as fair and reasonable, the proposed PESM “better reflects the economic use of the overheads by LnS and is, accordingly, a more accurate proxy than the standard method” (paragraph [54]). The first ground of appeal: did the FTT place erroneous reliance on the “direct and immediate link” test?[39]HMRC’s first ground of appeal is that the FTT erred in law by relying on the “direct and immediate link” test in deciding how the residual input tax should be apportioned. The nub of HMRC’s argument on this point is that the direct link test (as I shall call it for short) is relevant only at the prior stage of attribution of input tax to exclusively exempt or taxable supplies, or (in the case of overheads) to the business as a whole. The test therefore has no part left to play where (as in the present case) it is common ground that the input tax in question is referable to overheads which are directly linked to the business as a whole, and the only issue is how such input tax should be 27 apportioned between the taxable person’s exempt and taxable output supplies. Since the overheads are inevitably reflected in the prices which LnS charges for its supplies, and since (by definition) they are not exclusively attributable to the company’s taxable supplies of storage, it must follow that they are also attributable in part to the exempt supplies of insurance. It is therefore a contradiction in terms, so the argument runs, to apply the direct link test again at the apportionment stage, and to use it as a tool for apportioning all but 0.02% of the input tax to the taxable supplies of storage.[40]I have already concluded that the FTT made an error of law when it treated the decision of the ECJ in SKF as authority that the direct link test is relevant at the apportionment stage: see paragraph 36 above. I have also noted that the FTT’s comments on Southern Primary in paragraph [31] of the Decision may suggest some confusion between the attribution and apportionment stages of the exercise which has to be performed: see paragraph 24 above. But it does not follow from this, in my judgment, that the two stages always have to be treated as rigidly distinct from each other. Depending on the precise facts, considerations which are relevant at the first (attribution) stage may also be relevant when examining the economic use made of the overheads at the second (apportionment) stage. An examination of the economic use made of particular overheads in the business may show that it is fairer to apportion a larger proportion of them than the standard method would allow to either exempt or taxable supplies; and in some cases it may be right to conclude that the apportionment should be 100% one way or the other. None of this, in my view, is necessarily incompatible with the prior analysis at the attribution stage which led the expenditure in question to the classified as overheads in the first place. A further reason why, always depending on the facts, it may be appropriate to proceed in this way is that it is only in rather an artificial sense that the direct link test is taken to be satisfied in respect of overheads at the attribution stage. The FTT should therefore not be inhibited from examining the economic use made of particular overheads at the apportionment stage, even if it leads to the conclusion that they are largely, or sometimes entirely, used for the purposes of generating particular types of supply. If the facts justify such a conclusion, it would not be a misdirection of law to say that the direct link test is or is not satisfied to the relevant extent, although it would in my respectful opinion promote clarity of analysis and expression if that test were reserved for the earlier attribution stage of the exercise.[41]A good example of the kind of case which I have in mind is provided by Volkswagen Financial Services. The taxpayer company, which was part of the Volkswagen group, entered into hire purchase agreements with customers for group brand vehicles. The hire purchase transactions involved both a taxable supply (sale of the vehicle) and an exempt supply (the provision of finance); but the vehicle was always sold at cost, with the consequence that in economic terms the company’s overheads of its hire purchase business were loaded entirely onto the exempt supply of finance, and entirely reflected in the 28 finance charges made. The FTT considered that, because the residual input tax related to overheads, it was necessarily attributable to both the chargeable and the exempt supplies, which formed part of a single indivisible transaction for the supply of vehicles on hire purchase terms. Accordingly, the input tax should be apportioned on a 50/50 basis, on the footing that each element of the composite transaction had equal weight. Allowing HMRC’s appeal, the Upper Tribunal held that the FTT’s approach was wrong in law, because as a matter of economic reality the overheads were exclusively reflected in the finance charges made to customers. As the Upper Tribunal explained at [100]:
“100. It is not the case, in our view, that residual input tax can never be deductible when the taxable part of the trader’s business is loss-making or cost-neutral, but in this case it seems really quite obvious to us that a proper application of the correct tests shows that there is no direct or immediate link between the residual input costs in question and the taxable sales of vehicles by VWFS. The direct and immediate link is between the residual input costs and the finance supplies which are predominantly exempt outputs. Likewise, the residual input costs are not, properly regarded, cost components of the taxable part of VWFS’s entire economic activity. They are cost components, as the FTT correctly found, of the financing part of VWFS’s business. That is the economic reality of VWFS. Its overheads are used for its financing business, which is exempt from VAT.”
[42]HMRC had a second ground of appeal in Volkswagen Financial Services, to the effect that even if their main argument failed, the 50/50 apportionment proposed by the company was not fair and reasonable, and a lesser figure than 50% should have been attributed to the taxable supplies. In view of their decision on the main argument, the Upper Tribunal’s observations on this part of the case were obiter. Having said at [106] that “[t]he PESM adopted must fairly and reasonably represent the extent to which goods or services are used by or are to be used by the taxable person in making taxable supplies”, the Upper Tribunal continued: “107. The FTT was right to find that there is no rule to the effect that, where residual input costs are in fact a cost component of only an exempt output, the input tax will never be deductible. That will normally be the case, but on authority, a twin approach is appropriate in the case of overheads: one looks to see whether the residual cost inputs have a direct and immediate link with the taxable transactions, and whether the residual cost inputs are a cost component of the taxable transactions. The concepts of asking whether residual inputs are a cost component of the taxable outputs, and asking whether they are a cost component of the price of the taxable 29 outputs are substantially identical. These twin approaches are alternative ways of expressing the same basic test.[108]When it cannot properly be said (as is normally the case with overheads properly so-called) that a residual cost input has a direct and immediate link with any particular output, these twin tests are to be applied objectively from the broader economic standpoint. The question is whether the residual cost inputs have a direct and immediate link with or are cost components of the taxable part of the taxable person’s entire economic activity.”[43]I respectfully think that this passage has the potential to cause confusion, if it were interpreted as suggesting that the standard tests at the initial attribution stage must always be applied again at the subsequent apportionment stage, albeit from “the broader economic standpoint”. While such an approach may sometimes be justified on the facts, I think it is important to maintain the distinction between the basic tests to be applied at each stage. In particular, it always needs to be kept firmly in mind:(a) that once overheads have been identified as such, at the attribution stage, they are necessarily taken to have satisfied the test of a direct link with the business as a whole; and(b) at the second apportionment stage, the search is for the methodology which best reflects the actual economic use made of the overheads in making taxable supplies. Although I agree with the conclusion reached by the Upper Tribunal in Volkswagen Financial Services, I do feel, with the greatest respect, that their reasoning is sometimes in danger of blurring the distinctions between the two stages.[44]Returning to the present case, something of the same confusion may be seen in the FTT’s mistaken reliance on SKF as a source of guidance at the second apportionment stage. But on a fair reading of the Decision as a whole, I do not consider that this error materially affected their consideration of the second stage question. As I have explained, they clearly had the detailed guidance given by the Court of Appeal in London Clubs well in mind, and their selfdirection at paragraph [36] of the Decision was impeccable: see paragraph 30 above. Furthermore, it was clearly relevant for the FTT to ask themselves whether, and if so to what extent, the costs of the overheads were likely to be incorporated in the prices which LnS charged to its customers (paragraph [46]). As Volkswagen Financial Services shows, such an enquiry goes to the 30 heart of the economic use test, and may be determinative. It is thus worthy of particular note that, immediately after asking themselves this question, the FTT said:
“In our view, the actual or likely impact of the costs of overheads on the prices of LnS’s supplies not only establishes whether there is a direct and immediate link with those supplies but is also a useful measure of the extent of the economic use of the overheads (my emphasis).”
This clearly shows, to my mind, that the FTT had not lost sight of the economic use test, and that their mistaken reliance on SKF did not in fact cause them to set off in the wrong direction, or vitiate the analysis which they undertook.[45]For these reasons, I would reject the first ground of appeal. The second ground of appeal: did the FTT err in their application of the “direct and immediate link” test?[46]The second ground of appeal assumes that the FTT were in principle correct to apply the direct link test, but contends that they erred in their application of it. Since I have now held that the FTT were wrong in so far as they sought to apply the direct link test as a separate test, but that this error made no material difference to their application of the correct economic use test, the second ground does not strictly arise. Nevertheless, I will briefly examine it, in order to see whether the points relied upon by Mr Thomas in support of it might show that the FTT erred in law in their application of the correct test.[47]The main focus of Mr Thomas’ submissions was on paragraphs [47] and [48] of the Decision, and the FTT’s examination of the extent to which the overheads were incorporated in the price of insurance on the one hand, and storage on the other hand. He complained that it was illogical for the FTT to say that the price of insurance was set by LnS by reference to the amounts charged by its competitors, rather than in response to any costs associated with it, but to ignore the fact that the charges for storage were also affected by market forces, and were likewise not explicitly fixed by reference to the cost of the overheads. In my view, however, these were matters for the FTT to evaluate, having heard and considered all the evidence. It is impossible to say that the FTT erred in law in approaching the question as they did. The sale of insurance was, on any view, ancillary to the sale of storage space, which was the company’s principal activity. The FTT found as a fact that LnS “set the 31 price of its insurance by reference to the amounts charged by its competitors for insurance rather than in response to any costs (not even the cost of the block policy)”. They also found that the costs of the overheads “did not materially affect and were not incorporated in the price of the insurance” (a finding to which I will return later in this judgment: see paragraphs 54 to 55 below). It is true that these findings do not exclude the possibility that the overheads were, to some extent, still reflected in the price of the insurance; but the FTT then allowed for this possibility in their recognition that there is some link between the overheads and the sale of insurance “simply because the insurance is sold in the reception areas of the stores and the overheads related, in some part, to those areas”. They said they were unable to determine the impact of such costs on the price of the insurance from the evidence before them, but they thought it “must be very small”. Conversely, in paragraph [48] they gave reasons for saying that the cost of overheads was more likely to be factored into the price of storage. In the first place, not all customers buy insurance from LnS; and secondly, in its annual reports LnS assesses projected expenditure on new stores, and values development projects, by reference to space rental and occupancy levels, not by reference to projected insurance sales.[48]The conclusion drawn by the FTT, in paragraph [49], is that LnS made use of the overheads “almost exclusively for the purpose of making supplies of storage”. I can find no indication that, in reaching this conclusion, the FTT failed to apply the economic use test and the guidance given by the Court of Appeal in London Clubs. Mr Thomas also sought to draw distinctions between some of the findings made by the FTT in the present case and the findings made in certain other cases, particularly Dial-a-Phone and Volkswagen Financial Services. But it is elementary that every case turns on its own facts, and as Etherton LJ emphasised in London Clubs at [42] the analysis of attribution of expenditure for the purposes of regulation 101 “is highly fact sensitive”. It is therefore a hopeless endeavour to try to establish an error of law by pointing to differences in the facts found, or the conclusions drawn, in other cases.[49]I would therefore dismiss HMRC’s second ground of appeal. The third and fourth grounds of appeal: did the FTT fail to recognise the real economic use of the VAT bearing costs, or did they adopt a method based on physical use which does not reflect economic use?[50]I propose to take the remaining two grounds of appeal together, because they both allege that the FTT erred in their application to the facts of the correct 32 economic use test. In other words, they are further attempts to discredit the Decision by inviting the Upper Tribunal to discern an error of law in the FTT’s findings of fact, or their evaluation of the facts, even though (on this hypothesis) they directed themselves correctly about the test to be applied.[51]The need for caution and restraint by an appellate court or tribunal when faced with a challenge of this nature has often been emphasised, not least by Etherton LJ in the passages from London Clubs at [73] and [74] which I have already cited (see paragraph 29 above). Mr Hitchmough QC also reminded me of what Mummery LJ said in Procter & Gamble UK v Revenue and Customs Commissioners [2009] EWCA Civ 407, [2009] STC 1990 (the wellknown case about the classification for VAT purposes of Regular Pringles, the savoury snack product), at [74]: “For such an appeal to succeed it must be established that the tribunal’s decision was wrong as a matter of law. In the absence of an untenable interpretation of the legislation or a plain misapplication of the law to the facts, the tribunal’s decision that Regular Pringles are “similar to” potato crisps and are “made from” the potato ought not to be disturbed on appeal. I cannot emphasise too strongly that the issue on an appeal from the tribunal is not whether the appellate body agrees with its conclusions. It is this: as a matter of law, was the tribunal entitled to reach its conclusions? It is a misconception of the very nature of an appeal on a point of law to treat it, as too many appellants tend to do, as just another hearing of the selfsame issue that was decided by the tribunal.” (Mummery LJ’s emphasis) See too the observations of Jacob LJ at [9] to [11] and [19] and Toulson LJ at [48] and [60] to [62].[52]Of equal importance is the principle that, where an appeal lies only on law, and the tribunal has not made an overt error of law, a finding of primary fact, or an inference drawn from the primary facts, may only be challenged on the limited grounds explained by the House of Lords in Edwards v Bairstow [1956] AC 14: see in particular the speech of Lord Radcliffe at 35-36. It was in relation to such challenges that Evans LJ (with whom Saville and Morritt LJJ agreed) said in Georgiou and Another (trading as Marios Chippery) v Customs and Excise Commissioners [1996] STC 463 at 476:
“It follows, in my judgment, that for a question of law to arise in the circumstances, the appellant must first identify the finding which is challenged; secondly, show that it is significant in relation to the conclusion; thirdly, identify the evidence, if any, which was relevant to that finding; and, 33 fourthly, show that that finding, on the basis of that evidence, was one which the tribunal was not entitled to make. What is not permitted, in my view, is a roving selection of evidence coupled with a general assertion that the tribunal’s conclusion was against the weight of the evidence and was therefore wrong.”
[53]In the light of these principles, counsel for HMRC faced an uphill struggle in trying to establish a demonstrable error of law in the FTT’s findings of fact and the evaluative conclusions which they drew from those findings. Although there is apparent force in some of HMRC’s individual points taken in isolation, I remain unpersuaded from a reading of the Decision as a whole that the FTT reached a conclusion which it was not in law open to them to reach. Whether I would have reached the same conclusion myself is, of course, irrelevant; and in any event I lack much of the material which the FTT had to consider, including the oral evidence of Mr Davies who (I am told) was crossexamined for some two hours. In the circumstances, I do not propose to go through each and every point raised by counsel for HMRC in their written and oral submissions. It is enough to say that they have in my judgment failed to make good any material error of law on the part of the FTT. I will, however, illustrate the problems faced by HMRC with two examples.[54]The third ground of appeal is primarily focused on paragraphs 47 to 49 of the Decision, which according to counsel for HMRC’s skeleton argument “reveal an erroneous approach to the question of assessing economic use of the costs in dispute”. Thus, it is submitted, the fact that the insurance prices were set at a market rate by reference to competitors does not logically support the proposition that the prices do not reflect any (or only minimal) overhead costs. It is said to have been common ground before the FTT that both the insurance and the storage charges were set at competitive market rates, but this did not lead the FTT to conclude that the overheads were not included in the price of storage. Further, Mr Sherwood in his statement referred to a sample statement of the price of insurance which Mr Davies had supplied to HMRC during a meeting in January 2009, according to which the former price of £0.75 per £1,000 of goods insured included £0.45 in respect of “cost of sale”. How then, it is asked, could the FTT have concluded that “the costs of constructing, maintaining and operating the stores did not materially affect and were not incorporated in the price of the insurance” (paragraph [47])?[55]At first blush, these may appear to be telling points; but in my view this is just the kind of cherry-picking exercise which it would be wrong in principle for an appellate tribunal to indulge in. There are a number of possible ways in which the FTT might have reached the conclusion which I have just quoted without any error of law, and it is impossible for me to conclude on the balance of probabilities that they must have misdirected themselves. For 34 example, Mr Sherwood goes on in his statement to record that in a later letter of 8 February 2012 LnS stated that “there are no documents or calculations underpinning the costs of sale allocation” in the sample statement of price, and LnS “has assumed that the costs of sale included in the statement were an estimate of the staff costs”. But staff costs would have been composed principally, if not entirely, of salaries, which are not VAT-bearing and therefore could not have generated any residual input tax. Since the FTT were only concerned with the apportionment of input tax which had been actually incurred on chargeable overheads, it seems to me very probable that they simply left salaries out of account when considering and applying the economic use test. This may or may not be an accurate explanation of how the FTT came to make the finding challenged by HMRC, but it serves to make the point that an error of law cannot safely be deduced from an incomplete review of the evidence or from passages in the Decision read out of context.[56]In relation to the fourth ground of appeal, HMRC’s basic complaint is that the FTT adopted a method based on physical use of the premises which did not reflect their economic use. In reliance on cases such as St Helen’s School and London Clubs, HMRC submit that physical use is not necessarily an accurate proxy for the real economic use of the VAT-bearing costs, in view of the economic reality of the business as a whole. It is said that the business was in effect a unitary one, which inevitably provided the opportunity to earn income from the making of exempt insurance supplies to a predictable percentage of customers. Viewed in this way, the storage space has a dual function to perform in the economy of the business, and it is unrealistic to ignore the part played by the storage space in the generation of insurance income.[57]This is, to my mind, an attractive way of looking at the problem, and (if adopted) it would prima facie justify use of the standard method for the apportionment of overheads. But I feel quite unable to say that it is the only reasonably possible way of looking at the matter, or that the taxable floor space PESM proposed by LnS could not legitimately be preferred to it. For example, a powerful point which might have weighed with the FTT is the fact that the insurance charge related only to the value of the goods stored. It bore no relation to the amount of storage space occupied by the goods, nor did it involve any element of insurance of the premises. If the matter is viewed in this way, there is no obvious economic link between the insurance charges and the relevant storage space, and it is arguably more realistic to concentrate on the part of the premises in which the insurance is actually sold. On that approach, there is in my view no obvious flaw in the PESM proposed by LnS, and although the percentage (0.02%) of chargeable overheads apportioned to exempt supplies of insurance does at first sight look remarkably low, it must again be remembered that the bulk of the overheads attributable to the sale of insurance was composed of non-chargeable salaries. 35 Conclusion[58]For the reasons which I have given, despite Mr Thomas’ well sustained and attractively presented arguments HMRC’s appeal will be dismissed. TRIBUNAL JUDGE: MR JUSTICE HENDERSON RELEASE DATE: 23 JUNE 2014

Cited in 1 later judgment