BAILII is celebrating 24 years of free online access to the law! Would you consider making a contribution?
No donation is too small. If every visitor before 31 December gives just £1, it will have a significant impact on BAILII's ability to continue providing free access to the law.
Thank you very much for your support!
 |
[Home] [Databases] [World Law] [Multidatabase Search] [Help] [Feedback] | |
First-tier Tribunal (Tax) |
||
|
You are here: BAILII >> Databases >> First-tier Tribunal (Tax) >> N Brown Group Plc & Anor v Revenue & Customs (VALUE ADDED TAX - partial exemption : costs associated with marketing) [2019] UKFTT 172 (TC) (06 March 2019) URL: http://www.bailii.org/uk/cases/UKFTT/TC/2019/TC07022.html Cite as: [2019] UKFTT 172 (TC) |
||
[New search] [Contents list] [Printable PDF version] [Help]
TC07022
Appeal number: TC/2011/01927
VALUE ADDED TAX – partial exemption – costs associated with marketing used for the purposes of taxable supplies – whether also used for the purposes of exempt supplies – basis of apportionment to the extent input VAT is “residual”
FIRST-TIER TRIBUNAL
TAX CHAMBER
|
|
N BROWN GROUP PLC JD WILLIAMS & COMPANY LTD |
Appellants |
|
|
|
|
|
|
- and - |
|
|
|
|
|
|
|
THE COMMISSIONERS FOR HER MAJESTY’S |
Respondents |
|
|
REVENUE & CUSTOMS |
|
|
TRIBUNAL: |
JUDGE JONATHAN RICHARDS |
|
|
JOHN ADRAIN |
Sitting in public at The Rolls Building, London from 21 May 2018 to 8 June 2018 and having reviewed further written submissions from the parties following a circulation of a draft decision on 31 October 2018
Kevin Prosser QC and Valentina Sloane, instructed by Eversheds Sutherland (International) LLP for the Appellants
Hui Ling McCarthy QC and Natasha Barnes, instructed by the General Counsel and Solicitor to HM Revenue & Customs, for the Respondents
© CROWN COPYRIGHT 2018
DECISION
Note: This is a redacted version of the decision released to the parties pursuant to a direction dated 4 March 2019 made under Rule 14 of the Tribunal Procedure (First-tier Tribunal)(Tax Chamber) Rules 2009.
Throughout this decision, we express our findings and analysis largely in the present tense. However, the dispute in this appeal relates to the tax treatment of transactions entered into between 2006 and 2016. Readers should not infer from the fact that we use the present tense that we are making findings on the current nature of the appellants’ business or related matters unless the context makes this clear. Rather, our general focus is on the position between 2006 and 2016.
1. N Brown Group Plc and JD Williams & Company Limited (“JDW”) are members of the same VAT group (the “Group”). JDW has, at all material times, been the representative member of the Group. The activities of the Group include the sale of goods, which is standard-rated for VAT purposes, and the provision of finance, which is exempt for VAT purposes. JDW is appealing against assessments that HMRC have made on the footing that input VAT that the Group[1] has incurred is “residual” for VAT purposes, so that only a proportion of that input VAT is recoverable.
2.
The dispute between JDW and
HMRC relates to VAT periods from, and including, 08/06 to, and including,
05/16. The dispute has been going on for over ten years and covers a wide range
of input VAT that the Group has incurred. Following a case management hearing before
Judge Sinfield on 5 July 2016, the First-tier Tribunal (“FTT”) directed that
there should be a determination of a number of preliminary issues that focus on
the parties’ dispute on input VAT associated with marketing specifically. HMRC’s
position is that input VAT that the Group incurs on goods and services it
receives that are connected with marketing (which we will refer to as
“marketing input VAT”) is residual because, although they do not seek to deny
the existence of a “direct and immediate link” between the relevant goods and
services and taxable supplies that the Group makes, they consider that there is
also a direct and immediate link with exempt supplies.[2] The Group’s
position is that the vast majority of goods and services received in connection
with marketing have a “direct and immediate link” only with taxable
supplies that it makes and so the related marketing input VAT is not residual
and is recoverable in full.
(1) Issue 1 (Attribution) For each of the 15 marketing categories identified for the purpose of these proceedings:
(a) Are some or all of the goods and services that the Group receives in connection with that marketing “wholly taxable” in the sense that they are used exclusively for the making of taxable supplies? If only part of those goods and services are wholly taxable, how should that part be identified?
(b) Are some or all of the goods and services that the Group receives in connection with that marketing used “wholly exempt” in the sense that they are used exclusively for the making of exempt supplies? If only some of those goods and services are wholly exempt, how should that part be identified?
(c) Are some or all of the goods and services that the Group receives in connection with that marketing “residual” in the sense that they are used both for the making of taxable supplies and exempt supplies? If only some of those goods and services are residual, how should that part be identified?
(2) Issue 2 (Apportionment) In relation to those goods and services determined as residual under Issue 1, is the relevant input tax to be attributed to taxable supplies by the use of:
(a) a turnover methodology (as HMRC argue and as applied when making the assessments); or
(b) an alternative methodology that JDW has proposed?
If the FTT finds that both HMRC’s methodology and JDW’s methodology reflect use, which methodology reflects use in making taxable supplies more accurately?
(3) Issue 3 (Alternative apportionment methodologies) If the Tribunal:
(a) considers itself unable to determine the question of apportionment in the light of its determination on attribution; or
(b) determines that neither HMRC’s methodology nor JDW’s methodology accurately reflect the use of residual input VAT in the making of taxable supplies; or
(c) determines that there is a more accurate methodology than either HMRC or JDW have proposed
then either:
(d) what is the methodology that most accurately reflects the use of residual input VAT in making taxable supplies; or
(e) what are the principles that should be applied to determine a more accurate methodology?
4. The following witnesses gave evidence for the appellants (and we have indicated with an asterisk those witnesses who were cross-examined):
(1) Angela Spindler*, who has been the Chief Executive Officer of the Group since July 2013
(2) Ann Steer*, who has been employed in the Group’s marketing department since 2007 and has been the Group’s Marketing Director since May 2014
(3) Anthony Davies, who has been employed by the Group for 20 years and has been the Group’s Head of Merchandise Operations for over 12 years
(4) Caroline Pettifor-Robson*, who was First Line Control and Operations Manager from 2012 to October 2016. One of her main roles at that time was to help the Group to market its credit better
(5) Craig Lovelace*, who joined the Group in May 2015 and has been its Chief Financial Officer since then
(6) Gareth Powell*, who has been employed by the Group since March 2002 and is its Head of Customer Data and Analytics
(7) Katie Ascott, who has been employed by the Group since 2009 and is Customer Retention Manager for some of the Group’s brands[3]
(8) Jennifer Morris* who joined the Group in 2009 and is a Customer Retention Manager and Marketing Communications Manager for some of the Group’s brands
(9) John Guest*, who joined the Group in 1990 and was its Financial Controller (with some responsibility for the Group’s VAT returns) from around 2004 to 2016
(10) Laura Robertson, who has been employed by the Group since 2008 and is Customer Retention Manager for some of the Group’s brands
(11) Ralph Tucker, who has been employed by the Group since July 2014 and is the Group’s Product Director
(12) Stephen Johnson*, who joined the Group in February 2016 as Financial Services Director
(13) Theresa Casey, who joined the Group in January 2015 and is its Company Secretary.
5. HMRC relied only on the witness evidence of Philip Todd, an HMRC officer. He was cross-examined.
6. We were satisfied that all witnesses who gave evidence were reliable and honest.
7. The preliminary issues involve an analysis of two distinct issues:
(1) Preliminary Issue 1 (the “Attribution Issue”) is to determine whether particular goods or services are “used for” taxable supplies, exempt supplies, or both.
(2) To the extent that goods or services are “used for” both taxable supplies and exempt supplies, the next question, which is raised in Preliminary Issues 2 and 3 (the “Apportionment Issue”) is to determine the methodology to be applied to calculate how much marketing input VAT is to be attributed to taxable supplies (and so be creditable).
8. We will divide this decision into four parts:
(1) In Part I, we will make detailed findings as to the nature of the Group’s business, the types of marketing that it employs and the nature of the expenses that it incurs in connection with its marketing.
(2) In Part II, we will set out relevant UK and EU law provisions. We will also survey relevant authorities, both of UK courts and tribunals and the Court of Justice of the European Union (“CJEU”) and set out our conclusion on the relevant legal principles that we must apply in order to determine the preliminary issues.
(3) In Part III, we will explain our conclusions on the Attribution Issue which will involve us making further factual findings (such as whether there is a “direct and immediate link” between input transactions and taxable or exempt supplies that the Group makes).
(4) In Part IV, we will explain our conclusions on the Apportionment Issue. That too will involve us making further factual conclusions based on the findings we made in Part I.
9. The Group’s business was established in 1859 by James David Williams who, in 1882, was the first person to use the UK’s parcel post service to send products direct to customers. It is now a large and successful corporate group which employs some 3,000 people in the UK. Its ultimate parent company, N Brown Group Plc is listed on the London Stock Exchange and is part of the FT-SE 250 index. At all material times, JDW has been registered for VAT and been the representative member of the Group for VAT purposes. The Group draws up accounts to the end of February in each year (although the precise date in February can vary). Therefore, in this decision when we refer, for example, to the Group’s “2008-9 financial year”, that is a reference to the year ending on the relevant date in February 2009.
10. The Group specialises in fashion retailing. It owns a number of brands (such as Jacamo, Simply Be, Fashion World and House of Bath). The Group’s fashion business is differentiated from that of other fashion retailers in at least the following two respects:
(1) It offers wide choice and good credentials in terms of making “fashion that fits” for plus size customers (regardless of age). It follows a special approach to do this: for example, it has developed partnerships with specific suppliers who are able to focus on fitting and producing larger sizes rather than simply fitting products to suit a smaller size and “scaling up”.
(2) It designs products for customers aged 50 and over (regardless of size) and has a good understanding of what fashion products work well for that age of customer (for example what colour is better for that age group’s skin tone).
11. Although the Group now has 23 stores, it primarily sells its clothes over the internet or by mail-order. It therefore has a different business model from a “traditional” high street fashion retailer that displays clothes in shops and offers prospective customers the ability to try clothes on in those shops. Since people buying clothes tend to like to see them and try them on before they buy them, at all material times, the Group has offered its distance-selling customers the ability to “try before you buy”. Until 2014, the facility to “try before you buy” was offered almost exclusively via the “personal account” described in the next section. Since 2014, customers of the Group have been able to order goods and pay for them in cash (i.e. by means of a credit or debit card) without first opening a personal account, and, as we will explain, the “try before you buy” facility operates slightly differently for customers who pay for their goods in this way.
(1) In general, the customer would be able to order goods (i) over the internet, (ii) by filling in an order form and sending it by post or (iii) by telephone[4].
(2) Provided that the price of the goods ordered, plus any charge for delivery, did not exceed the available credit limit on the customer’s personal account, she would not need to make any payment at all at the time of placing the order. However, she could, if she wished, choose to pay for the goods and delivery in full (with a credit or debit card) at the time of ordering.
(3) After processing the orderthe Group would debit the customer’s personal account with both the cost of the goods and a charge for postage and packing (assuming, of course, that she had not already made payment when placing her order) and send the goods by post to the customer.
(4) The period ending 14 days[5] after the customer received the goods was described in the Group’s terms and conditions as a “home approval period”. If a customer returned the goods in new condition within that period, she was under no obligation to pay for the goods, and the price of the goods, but not the charge for postage and packing, would be credited to her personal account[6].
(5) Every 28 days, the Group would send its account holders a statement summarising activity on the account. That statement would set out the balance on the account and the minimum payment that the Group required to be made (which would, for new customers, be 1/12th of the cash price of the goods purchased but, for more established customers with a good payment history, could be just £5). Payment of the amount due was required within 21 days of the statement date.
(6) If the customer paid the entirety of the balance shown on the statement within 21 days, the Group would charge no interest.
(7) If a customer failed to make the minimum payment within 21 days, the Group would charge interest as noted at [(8)] below and would also charge an administration charge of around £12. Both interest and any administration charge would be added to the customer’s account[7].
(8) If a customer made the minimum payment, but this was not enough to clear the balance on the account, she would be charged interest on the outstanding balance, but no administration charge.
(9) The Group charges interest, calculated monthly, on the balance shown on a customer’s account. Over the period relevant to this appeal, the APR charged by the Group varied between 39.9% and 58.7%. Any interest charged would be added to the customer’s account.
14. Therefore, customers holding a personal account fell into two categories. Those who did not pay off the full balance on their account within 21 days of receiving their statement were described during the hearing as “rollers” (on the basis that they were “rolling” forward their credit balance). All other customers were described as “non-rollers”.
15. The personal account offers customers of the Group the following benefits (in no particular order):
(1) Customers could “try before they buy” by ordering goods and returning them within the home approval period. Feedback that the Group receives indicates that some larger customers, or customers with mobility problems, welcome the ability to try goods on at home rather than in a shop fitting room. Customers frequently use this facility by, for example, ordering multiple versions of the same product, in different sizes and colours, returning those that they do not want.
(2) Customers ordering goods obtained up to seven weeks’ interest-free credit[8] after the home approval period expired.
(3) Customers ordering goods could also choose to take paid-for credit. A customer did not need to take any particular steps to obtain the Group’s paid-for credit other than not paying the full amount of the balance shown on their statement within 21 days.
16. From early 2014, the Group changed its business model and allowed customers to order goods without first opening a personal account. Customers without a personal account were required to pay the full price of the goods, together with the charge for postage and packing, by credit card or debit card at the time of placing their order. Such customers still obtained the benefit of the home approval period inasmuch as, if they returned their goods within 14 days of receipt, the price of those goods would be refunded (onto their credit or debit card). However, such customers could not obtain either interest-free credit, or paid-for credit, from the Group. (Of course, customers paying by credit card might obtain a period of interest-free credit and/or paid-for credit from their credit card provider, but they could not obtain it from the Group without making an order through a personal account). At the hearing, customers who chose to order goods otherwise than through a personal account were described as “cash customers” or as “paying cash”, even to the extent that they paid by credit card and we will adopt the same terminology in this decision.
17. The change described above gave customers an additional facility: they could order goods from the Group otherwise than through a personal account[9]. It remained possible for customers to order goods from the Group via a personal account.
19. As we note at [166] below, the questions in this appeal are concerned primarily with how goods and services received in connection with marketing are, viewed objectively, used. The subjective intentions of individuals as to why the Group makes particular statements to its customers cannot, therefore, determine matters one way or the other although they may shed a light on the objective question of “use”. We will, therefore, simply record here Ms Spindler’s evidence that the reason why the Group chose to promote the personal account in this way was because it believes that customers who hold a personal account are more loyal (in the sense that they order more retail goods more frequently). Later in this decision, we will explore the concept of customer “loyalty”, and its link with the Group’s credit offering, in more detail.
20. The Group’s decision to allow customers to pay for goods in cash represented a significant change for its business. It is clearly still working through the full implications of that change. However, this appeal is concerned with a period that ended on 31 May 2016. By 31 May 2016, for reasons set out at [21] to [24] below, we consider that the introduction of the facility to pay cash had no material effect on the Group’s overall business.
22. Second, contemporaneous evidence suggests that by 31 May 2016, the impact of the change was still not significant. In the Group’s Interim Report in 2015 (which covered the 26 weeks ended on 29 August 2015 and was prepared on 14 October 2015)Ms Spindler wrote:
It is still relatively early days in terms of welcoming cash customers … but we continue to be pleased with our performance with cash customers. The introduction of cash customers has not had a negative impact on the population of our credit customers who roll a balance. Instead, cash customers are either incremental or at the expense of credit customers who immediately paid off their balance, not incurring interest charges.
23. Ms Spindler’s optimism was, at the time, echoed by UBS in an analyst report prepared on 2 October 2015 which predicted that, despite the facility to pay in cash:
… the utilisation of credit will not significantly change amongst the core customer base and arguably there is demand for increased credit as economic conditions brighten.
24. With both the Group and analysts making such statements in October 2015, we do not consider the position would have changed materially by May 2016. We agree that after May 2016, changes in customer behaviour that had started before May 2016 started to have more significant effects. For example, we accept Mr Johnson’s evidence that the number of roller accounts fell by around 20% between August 2013 and August 2016 (although this statistic of itself tells us little about the total revenue realised from rollers). Moreover, the Group came to realise that over 50% of new customers were, despite the Group’s promotion of the personal account referred to at [18], choosing to pay cash rather than open a personal account. However, the effect of such changes had not been felt to any material extent by May 2016.
25. Since the Group has comparatively few retail stores, it does not have the facility that a “bricks and mortar” retailer would have of displaying goods in its stores. It therefore must incur significant marketing costs to bring its products to the attention of its customers and present them in an attractive way that encourages sales. Therefore, while the Group does not have the premises costs that a “bricks and mortar” retailer would have, it spends much more on marketing than a “bricks and mortar” retailer would.
27. Given the Group’s heritage as a mail order business, in the periods under appeal, it incurred significant costs in sending physical printed material to its customers. We will start by outlining the broad general categories of physical marketing material and will give more detail where appropriate in the section below relating to that category.
28. The Group’s physical marketing material falls into the following broad categories:
(1) catalogues and leaflets;
(2) small “solus” mailings to existing customers, typically offering some kind of incentive to order;
(3) parcel packs and statement inserts, consisting of small mailings sent to existing customers either with a parcel or with a statement;
(4) “mini-offers” and inserts, which typically consist of loose-leaf inserts in a newspaper or magazine advertising products or websites;
(5) “direct mailings” which broadly consists of smaller versions of catalogues or leaflets sent to potential new customers.
29. Catalogues and leaflets are largely sent to existing customers and, until recently, made up the largest part of the Group’s marketing spend. The Group operates by reference to two “seasons” each year (Spring/Summer and Autumn/Winter) and historically it would send catalogues to its customers at the start of a season to inform them of new products for that season but, as customer behaviour is changing, with more customers wanting to order clothes during a season, rather than before the season has started, the Group has started to send out smaller, more frequent, mailings to its existing customers.
30. Catalogues typically include a “roll fold” which is essentially a four-page document that includes a personalised cover with the name and address of the customer, a product offering on the inside and an order form on the back. Since the Group is incurring the cost of postage of these catalogues, it sometimes seeks to achieve a “free ride” on these postage costs by including, with the catalogue other small mailings or third party advertising materials.
31. We saw several examples of catalogues and were taken, in particular, to a sample catalogue relating to the JD Williams brand. Our overall impression of this catalogue was as follows:
(1) It was over 500 pages long and the overwhelming majority of it consisted of pictures of the Group’s products together with descriptions and prices. However, every now and then one or more pages would be included that did not feature any details on products but, instead, contained information on how a customer could order on the personal account. Out of 500 pages (and excluding for these purposes a 10 page “small print” section that set out all terms and conditions) perhaps 10 pages or so contained something other than pictures and descriptions of products.
(2) On page 113, a full page was devoted to explaining how easy it was to order products online and outlining the process for doing so.
(3) The sample catalogue included at most one or two pages (outside the “terms and conditions” section at the back) that outlined the Group’s credit offering in detail. For example, on page 209[10] the catalogue explained how a customer could “make even more of your budget with a personal account” and explained how a particular outfit could, by means of the personal account, be purchased for just £8.01 a month for 24 months. That page included an explanation of the interest that would be charged if a customer spread the cost in this way.
(4) As well as the full page referred to at [(3)] above, the catalogue contained less detailed allusions to credit. For example, on page 407 (on a contents page for the Group’s furniture and homeware section), it was explained that customers could “make life easier with your personal account … by spreading the cost of your favourite items with your existing personal account”. On another page that directed the Group’s customers to one of its brands (The Brilliant Gift Shop) customers were told that they could “spread the cost” using their personal account.
32. Leaflets are smaller versions of catalogues that are sent to existing customers.
33. The Group sends small mailings to existing customers often with an incentive to order. We were shown an example mailing relating to the Group’s Fashion World Brand that consisted of around 8 pages predominantly consisting of colour pictures of the Group’s products with details of their prices. The mailing promised a customer a free scarf with her next order and contained detail on how to order. On at least one page of that mailing, customers were told that the Group offered “easy returns within 14 days” and the ability to “spread the cost with a personal account” and wording in small print advised customers that, if they chose to “spread the cost”, the item would cost more than if bought for cash which served as a clear indication that interest would be charged. However, no information was given on the precise terms of the Group’s interest offering or of the APR that would be charged.
34. When the Group sends statements or parcels to its customers, it incurs a postage cost. It seeks to make the most of that cost by including small mailings together with those statements or parcels. We were shown a few examples of these. Each ran to around 30 or 40 pages and consisted primarily of pictures of products together with prices. They tended to contain less detailed descriptions of the goods than was contained in catalogues. Information on how to order was presented in abbreviated form (in the form of a banner footer setting out the email address of the relevant website or a telephone ordering number). There were few, if any, references to the Group’s credit offering.
35. The Group pays for loose-leaf inserts to be included in magazines or newspapers to advertise its products. Given the way in which this marketing material is distributed, it could be received by existing customers or non-customers.
36. A “mini-offer” would consist of around 24-48 pages and would consist primarily of photos of products, descriptions and prices. The document would explain how a product could be ordered, for example over the website or by phone. Up until 2014, when the Group’s products could generally be ordered only through a personal account, the document might explain briefly that, if someone who was not an existing customer of the brand concerned sought to order a product, that would involve that person opening a personal account that would have a credit facility attached to it. As a general matter, references to the detail of the Group’s credit offering in the “mini-offers” we saw were either fleeting or non-existent. For example, a mini-offer for the Fashion World brand included the occasional banner on a page saying “Send no money now – open a personal account today”. A mini-offer for the “Fifty Plus” brand included, on the first page, an invitation to “Spread the Cost – Open a personal account and pay off your balance in easy-to-budget amounts at your leisure”.
37. An “insert” is a small glossy flyer, typically of around 4 to 6 pages whose function is to encourage customers to request a catalogue or visit a website for a brand to order products.
38. The Group buys lists of people who are not currently customers of the Group, but who are thought might be interested in becoming customers. It sends direct mailings to those customers consisting of smaller versions of catalogues and leaflets.
39. We were shown examples of such direct mailings for the Group’s Fifty Plus, Fashion World and Julipa brands. All of the documents that we saw consisted primarily of pictures of products, descriptions and prices. There were few, if any references, to the detail of the Group’s credit offering in these documents. Perhaps unsurprisingly, given that these mailings were aimed at inducing people to become customers of the Group, they offered inducements to order, such as a discount on the first order or free packaging and returns with the first order.
40. The Group incurs the following costs associated with the production and distribution of physical marketing material.
(1) The costs of paper on which the material is printed.
(2) The costs of third parties that the Group engages to print the materials. (The combined costs of printing and paper in 2013 was around £34.4m.)
(3) Royal Mail charges for postage (£31.2m in 2013).
(4) Third parties’ fees for personalisation of printed matter (including laser-printing customers’ names and addresses and packaging the printed matter). This cost around £2.7m in 2013.
(5) The Group also incurs the costs of creative services associated with the preparation of physical material. For example, the Group pays external agencies to photograph products, it pays models and makeup artists. These costs amounted to £8.2m in 2013 and we deal with these costs in more detail in Category 15 below.
41. The Group places adverts for products in newspapers or magazines to induce readers to order those products “off the page”. We were shown an example for the Julipa brand that consisted of a full-page advert consisting largely of a photograph of a person wearing a “Julipa Duffle Showerjacket”. The advert informed readers that they could order that product for £18 but that they need “send no money now”. The advert contained an order form and the small print informed readers that, if an order was accepted, that would also constitute a request by the customer to open a personal account (described as a “credit account”) and that the Group would perform a credit check on the customer for this purpose. The small print also informed the reader that, once the credit account was open, it was possible to “spread the cost using our easy repayment facility”.
42. The advert for the “Julipa Duffle Showerjacket” was an example of a “one stage” advertisement: the advert itself contained an order form that the reader could use to order the product advertised. The Group also places “two stage” adverts. These do not focus on a particular product, but rather showcase one of the Group’s brands in more general terms in the hope that a reader will order a catalogue. We were shown an example for the Group’s Fabrizi brand. That consisted primarily of a photograph of a model wearing an (unspecified) Fabrizi product. The advert informed readers that Fabrizi offers fashion in sizes 12 to 32 and featured the strapline “The look. The feel. The fit.” Readers were given details of Fabrizi’s website and a telephone number and invited to order a catalogue. The small print informed readers that, if they ordered a catalogue, that would act as a request for the Group to open a personal account (described as a “credit account”) and that a credit search would be undertaken for that purpose.
43. The costs that the Group incurs in connection with this category of marketing consist primarily of fees payable to the newspaper or magazine for the advertising space and the fees of advertising agencies who develop the advertisements.
44. We were shown a selection of TV adverts that the Group used during the periods under appeal to promote its Jacamo, JD Williams and Simply Be brands. These adverts tended to feature models wearing the Group’s clothing in short scenes designed to appeal to the target market for the brand concerned. The adverts that we saw tended to show models and actors in a “lifestyle” setting wearing a variety of the relevant brand’s clothes. So, for example, the Jacamo advert featured a group of young men chatting and kicking a football. A JD Williams advert featured characters dancing spontaneously in a coffee shop with a “retro” feel. Another such advert featured a group of friends cooking a meal together and having fun chatting and enjoying a glass of wine.
45. Overall, we considered that the focus of the adverts we saw was on associating the Group’s clothing with an attractive lifestyle. The adverts tended to refer directly or obliquely to the “fashion that fits” selling point. So, for example, one of the adverts for JD Williams clothing featured models and actors who were in early middle age with a voice-over saying “At JD Williams you’ll find clothes that fit to flatter. Looking good has never felt so good.” Other JD Williams adverts emphasised that clothes were available in sizes 10-32. A Simply Be advert featured larger models and noted that Simply Be clothes were available in sizes 14-32. One of the adverts showed the prices of the clothes featured.
46. In four of the five adverts we saw, there was no reference at all to the Group’s credit offering. In one of the adverts, for Jacamo clothing, at the foot of the screen was prominent wording saying “18 yrs + T&Cs apply. Credit subject to status.” Overall, it was quite clear that the adverts referred almost exclusively to the Group’s fashion offering, and the lifestyle the Group was seeking to associate with that offering. In this respect, there was a clear contrast between the adverts and an advert by Littlewoods. The Littlewoods advert was for swimwear and contained an express inducement to buy Littlewoods goods on credit in the form of a subtitle saying “Representative 0% APR. Credit subject to status”. The Littlewoods advert also concluded with a voice-over saying “Spread the cost and everyone can have a beautiful summer. That’s the Littlewoods touch.”
47. We also saw a sample brief that the Group prepared for advertising agencies creating the Group’s adverts. That brief focused on key messages associated with the clothes and the Group’s brands and did not mention credit at all.
48. The Group spent some £10.6m in 2013 on such adverts. Those costs included the costs of shooting the adverts (including fees paid to models, makeup artists and set designers), disbursements (for travel and expenses) as well as fees paid to the creative agencies who designed the adverts and the cost of TV airtime.
49. The Group pays online search engines (such as Google and Bing) for advertising the Group’s different brands in search results if a user searches relevant terms.
50. The Group sends the search engines a list of key words in which it is interested. There might be several thousand key words for each brand, but it is instructive to explain how the service works in relation to the key words “black dress”. Search engines advertise the Group’s products to someone who keys the words “black dress” into the search engine in the following two principal ways:
(1) Links to the Group’s products might appear in the “Shopping” section of search engine results; or
(2) Links to the Group’s products might appear as a “sponsored link”.
52. Similarly, a customer keying “black dress” into the Bing website might find links to the Group’s websites within the “sponsored links” in the search result. Unlike the “Google Shopping” results referred to at [51], these results are not just presented as a photograph of a product with an associated price. Rather, the first line of the search result would be some clickable advertising copy that the Group prepared (in an example we were shown: “Black Dresses – Work Those Curves & Look Fabulous”). Clicking on that phrase would take the user to a website for the brand concerned (for example the Simply Be brand). Underneath that advertising copy were further, non-clickable phrases, also generated by the Group (referred to as the “call out extension”) that highlighted particular messages that the Group wanted to deliver (for example “next-day delivery”).
53. The search engines are paid by reference to the number of times that a search on a key word results in a “click through” to one of the Group’s websites.
55. The Group’s costs associated with “pay per click” marketing activity consist primarily of fees payable to search engines. In 2013, these amounted to some £17.4m.
56. The Group pays fees to external agencies to optimise its brands’ rankings in the non-sponsored results returned by search engines such as Google and Bing. Fees for these services amounted to around £400,000 in 2013. The third party agencies achieve this by assisting the Group to make their brands’ websites more relevant to particular search terms, and to optimise links to the website by, for example, ensuring that they contain appropriate references to particular products.
57. We accept Ms Steer’s unchallenged evidence that in the periods in dispute, the costs incurred related to the optimisation of the Group’s brands’ rankings by reference to search terms relating to physical products (e.g. “black dress”) and not search terms relating to credit.
58. The Group pays fees to operators of third party websites who direct users of those websites to the Group’s brands’ websites. The third party websites concerned tend to be websites such as Quidco or Vouchercodes which offer promotional discounts (such as 5% off an order) or incentives such as cash back when a user clicks a link from their website onto one of the Group’s websites.
Expires in 5 days
Exclusive 25% off Plus Free Delivery on Orders over £50.
A user of the Vouchercloud website who clicked through on the link would be able to use an electronic voucher on the Fashion World website to obtain the benefit advertised.
60. HMRC also showed us some pages on the Vouchercodes website (albeit relating to a period outside that with which this appeal is concerned). These pages followed much the same format as is outlined at [59]. However, as well as the link to the Group’s website and bare details of the offer in question, the Vouchercodes page contained a much more detailed “Hints and Tips” section that read as follows:
Hints and Tips
Fashion World knows how to maximise your pennies and here are some handy tips to make them go even further.
- Spend £40 or more and get free delivery.
- Don’t feel pressure to pay the full amount up front. You can spread the cost of the purchase to suit your budget.
…
61. Although that example came from a website page dating from after the period in dispute, we have concluded that in the period relevant to this appeal similar references to credit, or “spreading the cost”, appeared on some Affiliates’ websites. Ms Steer said in her witness evidence that
In previous periods (up to 2017) N Brown had a number of Affiliates who included credit messaging on their adverts.
We have also concluded from Ms Steer’s evidence that, although the Group has to determine what offers it is prepared to make via Affiliates’ websites (for example that it is indeed prepared to offer free delivery on orders over £50), it was entirely up to Affiliates whether they included references to the Group’s credit offering in adverts appearing on their own websites.
62. The Group uses Affiliate advertising primarily to recruit new customers. Costs that the Group incurs in connection with this kind of marketing consist almost exclusively of fees payable to the Affiliates and amounted to around £1.2m in 2013.
64. The Group regards the purpose of this form of advertising as being to recruit new product customers or to encourage existing customers to shop more. The costs that it incurs consist primarily of fees payable to the third party websites and amounted to £1.7m in 2013.
65. The Group uses telemarketing to contact existing customers that the Group’s own statistical analysis indicates are likely to make orders in sufficient quantities to justify this use of a focused marketing technique. The Group’s telemarketing is outsourced to third party providers. Very broadly, two types of telemarketing teams are used: a traditional “outbound” team will simply make telephone calls to the Group’s customers and read a generic script and offer an incentive (such as a discount) to make an order. A “blended” team would make similar marketing calls but would also have sufficient experience with the Group’s offerings to answer basic questions for example as to the credit limit on a customer’s personal account. If a call made by an “outbound” team results in the customer asking such questions, or if a call made by a “blended team” gave rise to a question that could not be answered, the customer would be transferred to a more specialised enquiry team.
66. We have accepted Ms Steer’s evidence that, until April 2016, the generic scripts that telemarketing teams read out made no reference to the Group’s credit offering at all. From April 2016, a new additional script has been produced which is aimed at customers whose credit limit on their personal account has been increased and we accepted Ms Steer’s evidence that, in the Group’s 2017 financial year (up to 20 February 2017), a total of 3,774,500 outbound calls were made with just 207,657 (or 5.5%) using the new credit-led script.
67. In cross-examination, the Group’s witnesses were asked questions about the possibility of “deviations from the script”. We have concluded that the Group did not expect such deviations (except to the extent that a blended team departed from the script to answer a basic question). The Group was not paying the telemarketing teams to use initiative; it was paying for them to follow the script. In addition, we have concluded that unauthorised departures from the script would be rare since calls are monitored and a process is in place to ask questions of operatives whose calls were lasting too long. If it looked as though a particular operative was going “off-script” in telephone calls, we accepted Mr Powell’s evidence that the third party would train the operative to avoid this recurring. We agree with Ms McCarthy, however, that it could not be said with certainty whether any particular call involves an operative going “off-script”.
68. The Group pays third party agencies for advice about what its brand positioning and strategy should be for particular brands. This includes, for example, producing brand toolkits and guidelines which set out the “spirit” of a brand, the message that the brand is trying to promote to its customers and what the brand’s target market is.
71. The Group pays third parties to provide them with lists of potential new customers with a view to sending marketing material to them. In 2013, the Group paid £700,000 for such lists.
72. The Group incurs costs in connection with public relations events. For example, we were shown a contract between JD Williams & Co Limited and Lee Publicity under which Lee Publicity agreed, among other matters, to position the Simply Be brand in a particular way in the press, to make proposals for public relations activity, to prepare public relations material for distribution to the press (for example product releases) and to support the Simply Be brand’s public relations generally.
73. The combined costs of public relations activity and celebrity costs (described in Category 12 below) amounted to £5m in 2013.
74. The Group identifies celebrities to act as “brand ambassadors” for some of the Group’s brands. Celebrities are identified by reference to their “fit” for particular brands. For example, the Group selected Andrew Flintoff, a former professional cricketer, to be a brand ambassador of the Jacamo brand as he is young, athletic, well-built and perceived to be down-to-earth and plain-speaking: attributes that the Group considers to fit well with the overall message of the Jacamo brand. Similarly, Lorraine Kelly, the television presenter, has been selected as brand ambassador for the JD Williams brand as the Group considers that she shares the values of that brand.
75. We reviewed a sample agreement with a celebrity and we concluded that the celebrity’s duties under the agreement involved the promotion of clothing and fashion products[11]. The celebrity had no obligation to promote the Group’s credit offering. Celebrity ambassadors get involved in the design of the relevant brand’s clothing, and they take part in photo shoots and public relations events generally. In return, the Group pays celebrity ambassadors a fixed fee together with commission calculated by reference to the sales of the relevant brand’s products (including clothing, footwear and accessories). Celebrity ambassadors are not paid commission by reference to the Group’s sales of credit.
76. The Group commissions third party companies to undertake a wide range of customer research. To give an indication of the scale of this research, the Group provided some 1,139 items of market research to HMRC as part of their enquiries. The overwhelming majority of the market research that the Group commissions relates to its clothing and fashion products.
77. Given that volume of material, it is not possible for us to make general findings as to what the market research covers and we will simply note some examples that we saw. Market research on clothing and fashion products covered topics such as barriers to customers buying footwear from the Group, customers’ response to the Group’s underwear offering, why the Group’s best-selling corset was the subject of higher than expected returns and how the Group’s customers respond to the size and fit of the Group’s, and its competitors’, clothing.
78. The Group has on occasion commissioned research on credit specifically. For example, in September 2009, it commissioned an external provider, Bamboo Research Limited, to provide a report on the attitudes of the Group’s customers to credit with the stated business objective of that research being:
To increase demand, sales and credit through appropriate provision of credit.
79. This category covers the expenses of digital marketing that do not fall into any other category and includes the following:
(1) The Group pays third party websites to host banner advertisements for the Group’s brands on their websites.
(2) The Group pays to advertise via social media.
(3) The Group incurs costs associated with maintaining its domain names.
(4) The Group paid a third party to install software on the Group’s websites to show customers looking at a particular product on that website what other products purchasers of the first product typically bought.
80. Given that the Group does not have many physical stores, it takes a large number of photographs of its products in order to advertise them to its customers. The Group applies a “pyramid” structure to these photographs. At the top of the pyramid are more expensive photographs designed to communicate a particular “brand story”. These photographs would be used on billboards or the front cover of a catalogue, will be bespoke for each brand and may be taken on location (although, increasingly, to reduce cost the Group is taking such photographs in studios). In the middle of the pyramid are photographs communicating seasonal stories, trends or themes. For example, the Group might take photographs of people wearing its clothing in a sunny climate with the photos being used in a catalogue or booklet to market that category of clothes. At the bottom of the pyramid are trade photos which are taken primarily for the purpose of displaying products on the Group’s website or in catalogues and so focus on giving a good view of the product’s features.
81. The Group stores all of its photographs in a central database and uses them as and when required. If it wished to include photographs of products as part of an advertisement or promotion of either the Group’s personal account, or its credit offering, it would choose photographs from its central photo library. It cannot, therefore, be determined, at the time any particular photograph is taken (or charged for) how that photograph will be used.
82. The Group’s senior management manages the Group’s business to optimise the profit that the Group obtains from all sources. We have used the word “optimise” rather than “maximise” for a reason. It was put to a number of the Group’s senior executives in cross-examination that, since the Group’s most profitable customers were the “medium risk rollers”, it necessarily follows that the Group’s strategy is to seek to recruit such customers, increase the number of such customers and focus their sales and marketing efforts on those customers. Our conclusion, however, is that this categorisation of the Group’s strategy is not entirely accurate and the position is more complex.
… Our objective is to increase traffic by creating a proposition that has broad appeal, that brings in many customers. The fact that some of them go on to be more profitable is of course helpful financially and we will continue to work to perpetuate that, but if we didn’t work very hard to bring many, many, many customers in, the efficiency of our model would actually deteriorate.
84. However, in reaching the conclusion at [83] above, we should not be taken as finding that the Group did not care about the income from its finance business or was content for that business to generate whatever income it could without much intervention at a strategic level. Ms McCarthy and Ms Barnes put together a highly detailed set of extracts from the Group’s board minutes and published accounts that demonstrate that the Group’s senior management was acutely aware of the importance of its financial services income, was well aware of the strategies at their disposal to increase that income (and of the knock-on effects that those strategies might have) and took the kind of sensible steps that one would expect a FT-SE 250 board to take in order to optimise that income. The clear picture that emerges from records of its board meetings, presentations to analysts and statements in its accounts is that the Group’s business was managed strategically as a whole. Often the Group would see that a proposed action in the financial business could have implications for the retail business or vice versa. In those situations, the Group would carefully consider the implications for the business as a whole before deciding what actions to take. A few examples of this behaviour will suffice.
85. The Group’s senior management recognised that there was a complicated trade-off between credit policies, retail sales, bad debts and overall profit. Put simply, the Group could, in theory, loosen its credit criteria and offer more credit to customers. That would increase retail sales and finance income but would, in turn, lead to higher bad debts. By contrast, if the Group restricted its customers’ access to credit, bad debts would reduce but retail sales and finance income would suffer as well. Therefore, there was a balance to be struck. At different points, the balance was struck in different ways. For example, in the Chairman’s Statement accompanying the Group’s announcement of interim results in October 2007, the Chairman stated:
Changes to our credit scoring and credit limit policies have resulted in a planned increase in the rate of bad debts over the last 18 months because they produce incremental sales for little additional marketing cost thereby boosting profitability.
At a board meeting on 24 January 2008 the Finance Director reported on an increase in bad debt provisions arising as a result of “deliberate taking on of higher risk customers to generate profit”.
However, at other points in the economic cycle, the balance has been struck differently. In the midst of the economic crisis, the Group’s Preliminary Results Announcement for 2009-10 sounded a very different note as follows:
We have implemented a number of changes to our credit policy during the year, raising the threshold for new account applications, changing payment terms and selectively reducing credit availability, particularly for high value items ordered by our younger customers. The consequence of these actions has been to reduce sales growth but has also reduced the bad debt charge since implementation.
86. The Group was not content simply to sit back, sell retail goods and derive whatever financial turnover it happened to obtain off the back of those retail sales. Instead, as would be expected of a FT-SE 250 group, the Group sought to optimise both its retail and financial turnover. At a board meeting in 2010, the Finance Director observed that:
… profit is not only dependent on core growth. New brands contribution mix changes and further internet penetration are key but need support from financial income development and lower bad debt charges.
87. The Group was alive to both the benefits and risks of having a core portfolio of profitable “roller” customers. On one hand, the level of income and profit those customers generated was welcome. However, the Group often asked itself whether it was too reliant on these core customers and what opportunities for growth existed in this constituency. For example, at a board meeting on 28 September 2011, it was noted:
The data suggests that the company’s profits are becoming more heavily reliant on fewer customers who use credit. The trend becomes more concerning when we consider that the average balance is growing therefore increasing the probability that customers will slow down their ordering frequency…
A similar theme emerges in a board meeting of 22 May 2015.
88. The Group also recognised that a strong retail product offering was important to the finance business, with Ms Spindler saying, in her published Review of the Half Year Results for the 26 weeks ended 29 August 2015:
For the first time in three years we saw an increase in new credit customer recruits, up 15%, a very encouraging trend. We believe that this was driven by our improved product proposition.
89. Mr Prosser and Ms Sloane have invited us to conclude that the Group’s “core business”, its “driver” and its “unique selling point” is its retail proposition that is based around its offering to plus-sized and older customers. However, we think that categorisation is too simplistic. Given that the overall strategy of the Group is to optimise its revenue from all sources and manage the complex relationship between income from retail services and turnover from financial services, we do not think it is correct to refer to either the retail operations or the financial services operations as a “core business”. Of course, as we will consider in more detail below, the two sides to the business generate differing levels of income. However, if only the retail business were regarded as “core”, it would necessarily follow that the financial services operations are “non-core” but we would regard that as inconsistent with the fact that the finance and retail operations are highly dependent on each other and symbiotic. We therefore consider that neither the retail operations, nor the financial services operations can fairly be described as “core”[12].
90. To emphasise the point at [89], one of the top 20 institutional shareholders in the Group commented in investor feedback in 2016 (after the Group had decided to allow customers to order goods other than through a personal account):
…My other concern is that I don’t fully understand the credit book. It seems like this used to be a credit business: they would sell goods in order to get people on their credit book and make money in this way. Depending on how you cut it up you could argue that their profits still come from the credit book. Are they looking at the business in the wrong way as they move towards being a fashion and internet focused retailer and away from the credit?
Mr Prosser and Ms Sloane invited us to express caution at anonymous expressions of opinions by analysts, but this was a statement on behalf of a shareholder who holds a significant stake in the Group and might be expected to have an economic incentive to understand it well. We do not ourselves share the view that the Group is a “credit business” that uses the sales of goods to enable credit to be earned. As we have said, we view the credit and retail sides of the business as symbiotic, with both depending on each other. However, the fact that someone apparently well acquainted with the business could view it as a “credit business” points against the conclusion that the Group’s “core business” is the provision of goods.
91. Mr Prosser and Ms Sloane were on stronger ground in referring to the “unique selling point” of the Group’s offering being its “fashion that fits” proposition. That, point was neatly brought out in Mr Tucker’s unchallenged evidence to the effect that:
The business is focused on and driven by the idea of bringing fashionable choices to consumers who were, and still are, not well served by the traditional high street stores. In particular, N Brown’s ability to succeed in these niche markets lies in its ability to produce “fashion that fits” which requires particular focus on the retail product itself in terms of its initial design and development.
92. That, however, does not demonstrate that the Group’s retail business is “core” (and its financial services business non-core). Rather it simply demonstrates that customers of the Group are likely, in most cases, to be attracted to the Group because of the fashion that it offers. It is therefore a statement as to how customers respond to the Group, not how the Group is operated. Having attracted customers to it with a strong niche fashion offering, the Group seeks to optimise income from both retail and finance sides of the business and manage the complex relationship between those sides.
93. We have accepted the evidence of Mr Tucker that, when setting a price for its retail goods and credit, the Group takes into account the costs of producing the product and what its competitors charge. It does not take overheads into account when setting the price of its retail goods.[13] It does not adjust the price of its goods by reference to credit income that it receives. So, for example, the retail price of the Group’s goods is the same whether a customer pays in cash or uses the Group’s credit. Similarly, the Group does not increase retail prices in a particular financial year simply because it perceives that it has received insufficient interest income in that year. Nor, on the other hand, does the Group have a policy of increasing charges for credit to enable it to charge less for retail goods.
94. In the periods relevant to this appeal, the Group’s systems have enabled it only to charge all of its customers the same rate of interest[14]. The Group’s APR has been set at different rates at different periods. It has ranged from around 39.9% to 58.7%. That APR was, at material times, considerably higher than rates charged by “mainstream lenders” such as high street banks which would have ranged from around 14% to 20% in the relevant periods (although we accept Mr Johnson’s evidence that some specialist lenders might issue credit cards with a higher APR). We accept Mr Johnson’s evidence that store cards, of the kind operated by large “bricks and mortar” retailers charged an APR that, in percentage terms, was “in the 30s”. At the date of the hearing, the Group is at an advanced stage in updating its systems so that it can charge different customers different APRs, but those system changes were not in place in the periods relevant to this appeal.
96. As we have noted, marketing is an important part of the Group’s business since, unlike a retailer with a large high street presence, it tends not to display its products in stores and relies on a variety of media to disseminate a marketing message. The importance of marketing is emphasised by the size of the marketing department: at the time of Ms Steer’s witness statement, it employed around 260 people. The Group has a marketing budget of around £60 million per year.
97. Because marketing is so important to the Group, and because it spends so much money on marketing, it has a sophisticated marketing strategy and it monitors the effects of its marketing closely. By far the most widely used measure of the success of the marketing department is the “accepted demand” of the Group. That is defined as the value of product orders[15] which are to be despatched to customers, before product returns or credit income, but after credit rejections. The rationale for excluding product returns from the calculation is that returns are considered to be affected by factors outside the marketing department’s control (for example a customer may not like the product, or it may not fit). For a similar reason, goods which a customer orders, but which the Group cannot fulfil (for example because of a lack of stock) are counted towards accepted demand (because the marketing department has generated demand for that product even though, for reasons outside the marketing department’s control, the Group could not fulfil the order). There are other key performance indicators (“KPIs”) which the Group uses to measure the success of its marketing department, but none of those refer to the amount of credit income that the Group generates.
98. A number of witnesses in their evidence, spoke about the function of marketing being to build “the brand”. Mr Lovelace, for example, said:
The ultimate aim of all your marketing is to increase brand awareness … to increase your customer base … and to generate loyalty insofar as it is possible throughout your customer base.
We consider that, viewed in context, Mr Lovelace and the other witnesses were talking about the “brand” primarily in terms of the retail products that the Group sells. We do not consider that they were using the word “brand” in the sense of the totality of products (including financial services) that the Group offers. However, that is subject to an important qualification. The Group’s retail brands are, with the exception of the relatively few high street stores that the Group maintains, built around its business model of a distance selling retailer. An aspect of the Group’s brand is that it offers customers the ability to “try before you buy” by having clothes delivered to them at home without obligation to pay for them and indeed a number of the Group’s customers prefer to try on clothes at home rather than travel to a shop to do so. To a large extent, the ability to “try before you buy” is provided by the personal account[16] and, as noted at [13], a personal account also offers customers the ability to access credit. Therefore, while we would not go so far as to say the Group’s brands focus on its financial services offering, we certainly do not consider that those brands are separate from that offering.
99. We have set out above the generic types of marketing that the Group uses and an indication of the extent to which it refers to credit on its face (noting that references to “spreading the cost” are necessarily references to credit – see [31(5)] above). The Group’s marketing managers determine how best to market particular brands on a “brand by brand basis”. As we have noted, the Group’s printed marketing material occasionally refers to the Group’s credit offering, either obliquely (by, for example telling readers that they can “spread the cost”) or explicitly (by means of a full credit advertisement that complies with regulatory requirements and informs customers of the applicable APR). The decision on how to refer to the Group’s credit offering in marketing material sent to customers is taken by the relevant brand manager, in consultation with colleagues. There are no central guidelines that prescribe the extent to which particular marketing material must refer to credit.
101.Once the marketing department has prepared a draft budget following the process set out at [100], that budget is presented to the Group’s board of directors, along with budgeted expenditure in other areas. As well as considering the marketing department’s projections as to customers and accepted demand, the Group’s board also considers the overall affordability of the marketing budget in the context of the Group’s business as a whole and other expenditure that the Group needs to incur. Therefore, while the Group does not explicitly include financial services income when setting the marketing budget, financial services income does have some indirect bearing on the marketing budget in the following two ways:
(1) If financial services income declined to an extent that the Group’s board considered that the marketing budget, viewed in the light of the Group’s other spending priorities, was unaffordable, the marketing budget would be reduced.
(2) In the first instance, the marketing budget is designed to secure a particular level of “accepted demand”. However, as noted from the table at [123], in the period 2010 to 2016, around 60% of the Group’s retail sales (by value) come from rollers. Therefore, whether it is expressly recognised or not, the majority of accepted demand that the marketing budget is seeking to create has to come from customers who are buying the Group’s goods on credit and will therefore also be producing financial services income for the Group.
102.Mr Prosser and Ms Sloane have submitted that the evidence demonstrates that, even if the Group did not provide credit, the vast majority of the marketing costs would need to be incurred in order to enable the Group to sell goods. We do not accept that submission. We have certainly accepted, from Mr Lovelace’s witness evidence that, if the Group sold its finance business, it would need to review all of its operating expenses and that, given that marketing expenditure represents the Group’s “shop front”, it would be one of the last categories of expense to be reduced. However, that is not the same as saying that marketing expenditure would remain the same. Indeed, given our finding at [121] that the Group’s retail business would be materially smaller if it had no credit offering, it follows that the Group’s accepted demand would be materially smaller (since accepted demand is driven by retail sales). From that it must follow that the marketing budget would be materially smaller since, as noted at [100], the marketing budget is itself driven by accepted demand.
103. The Group places a high premium on customer loyalty as measured by how recently a customer has ordered, the frequency with which she orders and the quantities that she orders. It costs the Group significant sums of money to attract new customers, so one of the focuses of the marketing department is on increasing the amount of money that its existing customers spend. The Group uses data that it collects on its customers both to decide which marketing material to send to which customers and the marketing opportunities that exist.
104.The Group classifies its customers by reference to the “binary system” which allocates a score (between 0 and 15) to each customer with a higher score indicating higher “loyalty”. That system considers whether a customer has ordered in any of the preceding four seasons and uses binary notation to give that customer a number between 0 and 15 to measure the “recency and frequency” of that customer ordering. For example, a customer who has ordered in each of the four preceding seasons would be given the number “1111” in binary which, when converted into base ten would give that customer a binary score of 15. A customer who placed their first and only order in the previous season would be given the binary number “1000” which equates to a binary score of 8 when converted into base ten. A customer who was newly recruited in the previous season but did not make an order then would be given the number “0000” and would have a binary score of zero. A customer’s binary score is affected only by orders of goods; whether she has rolled a credit balance and incurred an interest charge is not relevant to her binary score.
105.The Group also allocates its customers a “propensity to buy score”. While a customer’s binary score is generated by looking backwards at behaviour over the past four seasons, the “propensity to buy” score uses statistical analysis to seek to predict the value of goods that the customer will order over the next six months. While the propensity to buy score is seeking only to predict the value of goods (not credit) that a customer will order, the variables taken into account in producing that score may take into account “financial data” such as the amount of the customer’s credit balance and the amount of interest charges incurred.
We would spend more on sending leaflets and brochures out to binary 15s than 1s because they are more likely to respond.
107.However, decisions to send mailings to customers are by no means exclusively determined by reference to customers’ binary scores. Moreover, the Group sometimes considers it to make business sense to send mailings to customers with a lower binary score rather than a higher binary score. For example, someone with binary score 12 (1100 in binary) has ordered in the two most recent seasons, but not in the two seasons before that. Someone with binary score 8 (1000 in binary) has ordered in the most recent season, but not in the preceding seasons. Those are important customers because they offer the Group the prospect of a “good thing starting up” (to paraphrase Ms Steer and it sometimes makes sense to target those binary categories (in preference to binary 15s).
109.There is a further correlation. Although binary scores do not measure the volume of orders that a customer makes (and instead, only measure how recently and frequently the customer has ordered goods), the Group’s most loyal customers (its binary 15s) generate a disproportionate amount of all retail sales. In the 2015-16 half-year mentioned in paragraph [108] above, binary 15s accounted for 11% of total customers, but generated 44% of total sales of goods[17]. 78.4% of retail sales (by value) made to binary 15s in 2015-16 were made to rollers even though rollers accounted for only 65.3% of binary 15s. In summary, in 2015-16,65% of the Group’s most loyal customers (its binary 15s) were rollers. Moreover, even when judged against the cohort of other, highly loyal, binary 15s, rollers generate proportionately more retail sales than non-rollers[18].
110. There is also a correlation between being a roller and receiving proportionately more physical marketing material. We accept the submission that Mr Prosser and Ms Sloane made to the effect that the Group does not generally discriminate between rollers and non-rollers when sending marketing material to existing customers in the sense that a roller and a non-roller with the same binary score could generally expect to receive the same volume of physical marketing material.[19] However, since there is a general correlation between having a high binary score and being a roller, and since the Group’s most loyal binary 15s consist to such a large extent of rollers, it follows that there is a correlation between being a roller and receiving higher quantities of physical marketing material.
111.The overwhelming majority of the marketing material that the Group sends to its customers refers primarily to the Group’s retail products, with references to credit being at most tangential. However, the Group does occasionally send material to its customers that focuses on the Group’s credit offering. We saw, for example, a leaflet sent to customers in or around October 2016. The key message of that leaflet was “Interest Free Until 2017” and informed customers of the Jacamo brand that they could “plan ahead for Christmas” and pay no interest until 2017.
112.As we have noted, the Group’s business involves it both selling goods to retail customers and providing finance to its customers. At the hearing, those aspects of the Group’s business were referred to as the “retail business” and the “credit business” (or the “financial business”) respectively and we will use the same terminology. However, it was not suggested that the Group was operating two separate businesses (although, as noted below, the appellants do argue that the retail business could be run as a separate business even if the Group offered no credit). Therefore, the difference between the parties was not whether the Group was carrying on two separate businesses, but on the degree of links between the two aspects of the Group’s single business.
113.In its Group accounts, the Group tends to describe itself as a retailer: its accounts for the year ending February 2016 describe it is a “leading digital specialist fit fashion retailer” and in accounts for previous years its principal business activity has been stated as “multi-channel retailing” and “retailing through direct home shopping”. Throughout the period in dispute, the Group’s accounts have split turnover into retail and financial services categories. However, only from October 2015 has the Group, in public financial data, split out profit into a figure for the retail business and a figure for the credit business.
114.For the financial years ended February 2007 to February 2016, we have drawn the following conclusions as to the financial contributions of the two sides to the Group’s business:
(1) The Group’s total turnover from all sources is significant and has increased continuously over that period. Total turnover in the Group’s 2006-07 financial year was around £487.2m. By the 2015-16 financial year that figure had increased to £826.2m.
(2) Turnover from the retail business has, in all financial years in that period, been between 71% and 74% of total turnover. Turnover from the financial business therefore, has been between 26% and 29% of total turnover. Financial services turnover therefore represents a sizeable minority of the Group’s total turnover and is significant in absolute terms: in the 2006-07 financial year, the Group generated £127.2m of financial services turnover. In the 2015-16 financial year, it generated £242m of financial services turnover.
(3) It is much less straightforward to determine the proportion of the Group’s profit that is generated from its retail and financial businesses respectively. Prior to October 2015, the Group’s accounts reported only total profit and did not split that total into profits for the retail and financial businesses respectively. The Group has, for the purposes of its appeal prepared figures as to the “net contribution” of both the retail and finance businesses which are summarised below. Necessarily, those calculations depend on how expenses are allocated to the various sides of the business (which is at the heart of the dispute between the parties in this appeal). We also accept Mr Lovelace’s evidence that this information was prepared solely for the purposes of this litigation and the Group does not actually split its cost base in the way set out in the table. With that note of caution, the Group’s own calculations set out in the table summarised at [117] below suggest that the net financial contribution of the financial business was, on average, 69.1% of the total contribution of both businesses[20].
115.There is a clear and close correlation between the amount of the Group’s retail sales and the amount of its finance turnover. The nature of the correlation is best illustrated by the following figures that Mr Lovelace included in his witness evidence.[21] [Note: the figures in the table have been redacted]
|
Year ending Feb |
2007 |
2008 |
2009 |
2010 |
2011 |
2012 |
2013 |
2014 |
2015 |
2016 |
|
Retail turnover (£m) |
xx |
xx |
xx |
xx. |
xx. |
xx. |
xx. |
xx. |
xx |
xx. |
|
Annual increase (%) |
|
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
|
FS turnover (£m) |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
|
Annual increase (%) |
|
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
|
Total turnover (£m) |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
|
FS turnover as % of total |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
116.The clear implication from this table is that, turnover from financial services has been predictable over the period, at between 26% and 29% of total turnover. Moreover, the rate of increase in financial turnover broadly mirrors the increase in retail turnover, although there some exceptions. For example in the financial year 2012, retail turnover increased by 4.7%, but financial turnover increased by 14.2%. Mr Lovelace accepted in cross-examination that this broad correlation exists in relation to similar figures that he had prepared focusing on JDW alone, but pointed out, and we accept, that an increase in retail turnover in one financial year would not necessarily translate into an immediate increase in financial turnover in the same financial year. Rather, there is a “lag” effect since interest income received from increased sales might come in over subsequent financial years.
117.In his witness evidence, Mr Lovelace produced a table which drew on figures set out at [115] to illustrate the respective contributions of the retail business and the financial business to the Group’s overall business. We have already explained, at [114(3)], the notes of caution Mr Lovelace sounded as to his figures, which were understandable. However, Mr Lovelace provided his figures in support of his conclusion that, even if the Group had no financial business, its retail business would still operate as a standalone, viable business. HMRC did not accept Mr Lovelace’s conclusion and therefore, despite his notes of caution, it is appropriate to examine the figures in some detail to see what conclusion emerges from them. [Note: the figures in the table have been redacted]
|
|
2007 |
2008 |
2009 |
2010 |
2011 |
2012 |
2013 |
2014 |
2015 |
2016 |
|
Net contribution – financial business |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
|
Net contribution – retail business |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
|
Total contribution |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
|
Net retail contribution as % of total contribution[22] |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
|
Overhead costs |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
|
Group profit |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
|
Difference between retail contribution and overhead costs |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
xx |
118.Before setting out our conclusions on whether the retail business could exist separately from the financial business, we will simply explain the figures set out in the table as follows:
(1) All monetary figures in the table are in millions of pounds.
(2) Because the retail businesses and financial businesses are not actually separate businesses, to determine the “net contribution” of the retail and finance businesses, Mr Lovelace had to extract the above figures from other figures that the Group maintains that relate to the whole of its business. The process of extracting those figures necessarily involves making assumptions and judgements as to whether particular costs are allocated to the retail business or to the financial business.
(3) In making his calculations, Mr Lovelace took the Group business during the period in dispute and sought to allocate income and expenses as between the retail business and the financial business. The figures do not, therefore, estimate the effect on the retail business if the Group had no financial business. In particular, the figures do not hypothesise whether sales of retail goods might reduce if the Group was not able to offer its customers credit.
(4) In determining the “net contribution” of the retail business, Mr Lovelace has allocated all costs of marketing (except the costs of sending statements) and costs associated with the distribution of goods to the retail business. Mr Lovelace allocated all of the cost of bad debts (net of VAT bad debt relief) to the financial business. Payroll costs that relate directly to the retail business or finance business have been allocated accordingly and Mr Lovelace has sought to pro-rate payroll costs that relate to both the retail and finance businesses.
(5) The process at [(4)] above leaves a core of overhead costs that have not been specifically allocated to either side of the business. Mr Lovelace’s overall conclusion was that the contribution made by the retail business was sufficient to cover those costs in “most, albeit not all, years”.
(6) We do not consider that the figure set out in Row 4 of the table sets out the profit generated by the retail business expressed as a proportion of group profit. Group profit can only be determined after all expenses have been taken into account. Yet, in Mr Lovelace’s table, the “overhead” expenses were not taken into account in calculating either the net contribution of the financial business or the retail business (because the purpose of the table was to demonstrate that the retail business was viable without the financial business).
119.Mr Lovelace’s overall conclusion was that the Group could still function as a viable retail business without a credit offering (because it had core unique selling points associated with the fit and size of the clothes that it sells). However, he accepted that the Group would be a smaller retailer because the Group’s finance offering makes “the purchase proposition more attractive for some customers”. He also accepted that, if the Group did not have a financial services offering, there would be a “cohort of customers who may choose not to shop with us”. He was not able to say, however, whether he thought that income from customers who currently roll a balance on a personal account would be “significantly depressed” because he regarded it as a hypothetical question.
120.HMRC challenged Mr Lovelace’s overall conclusion that the Group could function as a viable business even without a financial services offering. They pointed out that, in five out of the ten years covered by his table, the contribution of the retail business would not have been sufficient to meet overhead costs. Moreover, they pointed out that, over the ten years in total, even on Mr Lovelace’s figures, the contribution of the retail business would have been £[figure redacted] less than the overhead costs. In response, the Group submits that there were particular reasons for some of the heavy shortfalls. For example, in 2008, the Group paid some £15m into its pension scheme and the years of 2017/18 have seen the Group “investing” heavily in its retail operations (by, for example, reducing prices) so the shortfall in 2015 and 2016 is not representative or typical.
(1) A number of the Group’s witnesses, including Ms Spindler and Mr Lovelace, accepted that the Group’s ability to offer finance enabled some retail sales to take place. Ms Spindler accepted that the provision of finance is a “key enabler” of retail sales. Mr Lovelace accepted that the absence of a financial business would cause the retail business to reduce in size. Therefore, the question is not whether the retail business would be smaller if the Group had no financial business but by how much it would be smaller. Since Ms Spindler described finance as a “key enabler” of retail sales, we have inferred that the reduction would not be small.
(2) Mr Lovelace’s table does not purport to address the scale of the reduction, as he accepted in cross-examination. His table, therefore, does not support an assertion that any reduction in scale would be only modest.
(3) As noted at [122] to [124] below the Group generates around 60% by value of its retail sales from rollers. Between 2010-11 and 2015-16, the Group generated well over £300m per year from retail sales to rollers. If the Group ceased providing finance, a significant fall in retail sales could be averted only if those rollers were in general able to, and chose to, obtain finance from other sources while still continuing to buy goods from the Group in similar quantities as before. No doubt some rollers could, and would, do this. However, we were not satisfied on the evidence that rollers would do so in sufficient quantities to avert a material reduction in the Group’s retail income.
(4) Even on Mr Lovelace’s figures (which assume no reduction in retail sales), a hypothetical standalone retail business would fall £[figure redacted] short of covering overheads over the 10-year period. Moreover, the retail business has evidently required significant “investment” over the past two years and, without the profit generated from the financial business, and factoring in some reduction in retail sales, we do not consider that “investment” could have been afforded. In short, if the Group had no financial services income, it would, over a 10-year period, have fallen £[figure redacted] short of meeting its (current) overheads and would not have been able to afford “investments” in its retail business it thought were necessary. Those factors strongly suggest to us that any standalone retail business would be materially smaller than the retail operations currently carried on by the Group.
122.Rollers represent a minority of the Group’s customers. Between the Group’s financial years 2010-11 and 2015-16, the proportion of active accounts[23] that were roller accounts varied from around 26.5% to 29.5%. However, despite forming a clear minority of the Group’s customers, rollers account for the majority of the Group’s retail income and the overwhelming majority of its total income (including interest income).
123.The central importance of rollers to the Group’s business is shown in the following table. This table does not contain data for all financial years relevant to the appeal. However, since none of the Group’s witnesses suggested otherwise, we have inferred that the figures are broadly representative of the position over the whole period under appeal.
|
|
2010-11 |
2011-12 |
2012-13 |
2013-14 |
2014-15 |
2015-16 |
|
Proportion of active accounts that are roller accounts |
26.7% |
28.1% |
29.5% |
28.7% |
26.8% |
26.5% |
|
Proportion of retail sales[24] - rollers |
59.3% |
60.1% |
61.8% |
62.8% |
62.3% |
64.0% |
|
Proportion of total income - rollers |
72.3% |
73.9% |
75.1% |
75.6% |
74.1% |
73.9% |
124.As noted in more detail at [108] and [109], the “loyalty” of the Group’s customers, as measured by the frequency with which they order, is a key metric of the Group’s business and one which the Group monitors closely. Viewed as a whole, customers of the Group who are rollers are more “loyal” in the sense that they order retail products from the Group with greater frequency.
|
Marketing costs per £ of product sales |
Non-roller |
Roller |
|
2010-11 |
29.3p |
15.5p |
|
2013-14[25] |
22.8p |
10.2p |
|
2015-16 |
21.9p |
10.5p |
126.To calculate the figure of 29.3p for the year 2010-11, for example, Mr Lovelace took the total cost of marketing material that could be identified as sent to non-rollers in that year and divided that by the total value of sales of goods made to non-rollers in that year. Ms McCarthy submitted that Mr Lovelace’s evidence was simply “assuming what it sought to prove” (namely that marketing material is used more to generate retail sales than it is to generate finance income). We do not agree with that general criticism. Mr Lovelace’s figures were calculated by tracking (where possible) how much marketing material was actually sent to rollers (and the cost of that marketing material) and comparing it with the cost of marketing material sent to non-rollers. Therefore, Mr Lovelace’s evidence was not circular in its reasoning and we accept his figures as accurate. However, we would note that the figures he gives apply only to marketing material which can be identified as sent to specific addressees. Not all material (particularly material that is designed to attract new customers) can be identified in this way. For example, the Group will not be able to tell who watches a television advert or sees an advertisement in a newspaper. Therefore, we do not accept that Mr Lovelace’s figures of themselves demonstrate that, taking all marketing material together, it costs the Group more to generate £1 of sales from a non-roller than it does from a roller.
[The Group’s] core customer base has been found to be particularly susceptible to relying on credit to make purchases, sometimes in circumstances where they would not otherwise be able to afford to.
In this section, we will consider this argument and, in doing so will consider the following points:
(1) Characteristics of the Group’s customer base as a whole.
(2) Whether the Group has a “core customer base” and, if so, what the characteristics of that “core customer base” are.
(3) Whether, in the light of our findings, HMRC’s analysis of that core customer base is correct and, if not, how it should be adapted.
128.The first point to note is that the Group has some 6 to 7 million registered customers. In addition, the Group’s brands are each aimed at different sectors of the market: for example at older or younger customers or at more or less affluent customers. In those circumstances, it is simply not possible to make generalisations about the Group’s customer base as a whole.
129.For the purposes of this litigation, the Group commissioned some research as to the composition of its customer base (in terms of the commonly used National Readership Survey of social grades) as compared with that of the UK population as a whole. The conclusion of that research is set out in the following table:
|
NRS Social Grade |
Percentage of UK population aged 15+ |
Percentage of the Group’s customers |
|
AB |
27.2% |
22.59% |
|
C1 |
26.9% |
25.37% |
|
C2 |
21.7% |
19.41% |
|
DE |
24.3% |
32.63% |
From that research, attributable to the period 2013-2015, we have concluded that the profile of the Group’s customers, in terms of NRS social grade broadly matches that of the UK as a whole. However, the proportion of the Group’s customers in the AB social grade is lower (by around 4.6%) than the proportion of AB individuals in the UK populations. Similarly, the proportion of the Group’s customers in the DE social grade is higher (by about 8.3%) than the proportion of DE customers in the UK population as a whole.
130.We have concluded that the Group’s “roller” customers can fairly be described as its “core customer base”. At [131] to [135] below, we explain what we mean by this finding.
133.Our conclusion at [132] is based on objective inferences that can be drawn from the Group’s financial performance. However, we are reassured to note that there are indications that our conclusion was, at least in 2011, shared by the Group. In a board meeting on 28 September 2011, the Group’s board was told:
The data suggests that the company’s profits are becoming more heavily reliant on fewer customers who use credit. The trend becomes more concerning when we consider that the average balance is growing therefore increasing the probability that customers will slow down their ordering frequency.
134.However, in concluding that rollers are the Group’s core customer base, we should not be taken as saying that non-rollers are unimportant to the Group. First, the Group clearly receives material income from non-rollers. Moreover, as Ms Spindler noted, the Group’s strategy of recruiting and retaining large numbers of customers means that the Group could not have one roller customer without having around 3 non-roller customers (since rollers represent around 26% to 28% of active accounts).
135.In their submissions, Ms McCarthy and Ms Barnes also highlighted the importance of the Group’s “binary 15” customers. We do not consider that the Group’s “core customer base” is the same as its customer base that consists of binary 15s. The Group uses the binary system to decide on how best to deploy its marketing resource efficiently. To be classified as binary 15, a customer must have ordered retail goods from the Group in each of the preceding four seasons. Such customers are, therefore, the Group’s most loyal customers and, as noted at [106], the Group understandably ensures that such customers receive high quantities of marketing material. However, characterising binary 15 customers as the Group’s “core customers” amounts simply to an assertion that customers who order frequently are “core” which reveals little about the objective characteristics of core customers. Of course, as we have noted at [108], binary 15 customers consist disproportionately of rollers and to the extent they consist of rollers, binary 15 customers are core customers (since rollers are core customers). However, we do not consider that binary 15 customers are “core customers” in their own right although they are clearly loyal customers to whom the Group sends a large volume of marketing material.
136.Having established that the Group’s core customer base consists of its roller customers, we will now make some findings on the characteristics of rollers. We will start by summarising relevant evidence (though, in the interest of brevity we will not attempt to summarise all of that evidence) and expressing some conclusions on that evidence and will then express our overall conclusions on the characteristics of the Group’s core customer base.
137.HMRC argue that, even though the Group’s customer base as a whole consists of a large proportion of AB individuals, the high APR that the Group charges to customers who roll a balance would deter AB individuals from using a personal account as a source of credit. Therefore, they argue that, of those individuals who roll a balance, a large proportion are represented by individuals in social grade DE who have difficulty in obtaining credit elsewhere.
138.Mr Johnson gave evidence to the effect that, from his experience in the retail credit industry:
Those who use retail credit more frequently tend to be from higher socio-economic groups as they are more easily able to pay off the retail credit which they take.
He pointed to some research that the Group had performed on the UK credit market that suggested that 29% of the UK credit market was taken up by individuals in social grade AB (even though those individuals make up only around 27.2% of the UK population). We also saw a report by the UK Competition Commission prepared on 7 March 2006 into the market for store cards in the UK. That report suggested that, although store cards often charge high APRs, around 60% of those individuals who “rolled” a balance on their store cards fall into the AB social grade. That report suggested that APRs were less of a motivation in the decision whether to roll a balance than might be expected: users of store cards often attach more significance to the flexibility that store cards offer to access short-term credit than on a comparison with APRs charged on other forms of credit.
139.
The Group’s APRs
throughout the period in dispute ranged from between 39.9% to 58.7% and, as
noted at [95], we consider that this indicates that the Group realised that it
was taking more risk in lending to its customers than other mainstream lenders
were taking in lending to theirs. Ms Spindler, in cross-examination accepted
that the Group’s high APR would not be attractive to customers in the AB social
category[26].
140.In support of their submission that the Group’s credit customers were unlikely to consist, to any material extent, of AB customers, HMRC relied on aspects of the Group’s own research. For example, internal market research that the Group prepared in 2013 entitled “UK Credit Marketing” suggested that 7.18% of the UK population felt that they had “stretched finances”. However, around twice this percentage (14.26%) of a sample of the Group’s customers who were identified as “high rollers”[27] thought that they had “stretched finances”. Mr Powell, in cross-examination accepted that AB customers were unlikely to have “stretched finances” and therefore, HMRC suggested that the Group’s own research indicated that customers using its credit offering were less likely to be from the AB social category.
141.Our overall conclusion was that, as a general matter, the AB social category is under-represented in the Group’s rollers. We had no evidence as to what precise percentage of rollers are in social category AB, but we consider that the proportion would be considerably less than the 22.59% of the Group’s total customers that fall in category AB. Ms Spindler accepted this in cross-examination and, as chief executive of the Group, she could be expected to know. The Competition Commission report caused us to question this conclusion somewhat, but ultimately, we concluded that the situation of store cards and the Group’s own credit were not necessarily comparable for the following reasons:
(1) The Group’s APRs (at between 39.9% and 58.7%) were, during the period relevant to this appeal, higher than the store card APRs (of around 30%) referred to in the Competition Commission report.
(2) The Competition Commission report indicates that store cards are “slightly more” likely to be held by those on higher incomes and that holders of store cards tend to have a credit card and, on average 1.6 other store cards. By contrast, the proportion of the Group’s customers in social category AB is less than is found in the UK population. The combination of store card holders being “slightly more” likely to have higher incomes and the proportion of the Group’s customers in social category AB being lower than found in the population as a whole could mean that there is a material difference in the demographic of the Group’s customer base as compared with holders of store cards. We were not, therefore, satisfied that the characteristics of those rolling balances on store cards were necessarily similar to those rolling balances on the Group’s personal account.
(3) The Competition Commission report demonstrates that 80% to 85% of store card holders have at least one other credit card. Therefore, a typical store card holder who wants to buy goods on credit had a choice between using a credit card or the store card. By contrast, until 2014, a customer of the Group had no option to use a credit card. In those circumstances, we do not consider it necessarily follows that behaviour involving store card use is necessarily a reliable guide to customer behaviour involving the Group’s personal account.
(4) Research that the Group commissioned from Bamboo Research Limited on its customers’ attitudes to credit suggested that the Group’s customers view store card credit differently from the Group’s credit. Store cards were perceived as a more “serious” and “harder” form of credit than rolling a balance on the Group’s personal account. We do not know what, if any, consequence flows from that perception but it did suggest to us that it might suggest that the behaviour of AB customers in relation to store cards would not necessarily translate into the same behaviour with a personal account.
142.A separate, but related question, was whether the Group’s roller customers “needed” the Group’s credit to make retail purchases. Mr Prosser and Ms Sloane invited us to conclude that rollers generally could be expected to have access to other sources of credit, that balances on personal accounts tended to be low (reflecting the fact that the Group’s goods are not expensive) and that there was nothing exceptional about the level of customer payment defaults. All those factors, they submitted, indicated that rollers do not “need” the Group’s credit.
143.Below, we will make findings on the factors Mr Prosser and Ms Sloane highlighted. However, we consider that they represent an unduly restrictive approach to the question of “need”. In cross-examination, Ms McCarthy put to Ms Spindler the example of a holder of a personal account, before 2014, who had £65 available until pay-day and who wanted both to buy a coat costing £65 and take her family to the cinema which would cost £65. Ms Spindler accepted, in our view quite rightly, that if the customer liked the coat, the credit facility associated with the personal account enabled her both to buy it and take her family to the cinema. In a narrow sense the customer might not “need” the Group’s credit (not least since she does not “need” either to buy the coat or to go the cinema). She may be able to pay for the cinema tickets by credit card and use her £65 to pay off the balance on her personal account in full when the statement arrives. However, in a broader sense the customer above could be said to “need” the Group’s credit if she wants to buy the coat since it provides her with a means of doing so and still being able to go to the cinema. We therefore think that Ms Spindler was also correct to accept that the Group’s credit offering based around the “spread the cost proposition”:
Enables sales today that might not otherwise happen.
We have a broad customer base, but we recognise that a proportion of our customers are from a low-income demographic who may find their access to mainstream credit restricted[28].
145.The Group takes steps, before advancing credit, to minimise the number of its customers who will have difficulty repaying amounts borrowed. It undertakes credit checks before advancing credit to a customer and, in the 12 months prior to September 2015, it rejected 40-45% of all applicants. Moreover, when it takes on a new customer, the Group typically offers a low credit limit and will increase the credit limit when it sees a pattern of payments being made on time. The Group therefore has a rigorous approach to credit checks and, if a customer can satisfy the Group’s credit checks, it is reasonable to suppose that she could satisfy another lender’s credit checks.
146.However, the impact of the point at [145] is diminished by the fact that the Group only undertakes credit checks on new customers. It does not continually refresh credit checks on existing customers and indeed it would be difficult to do so without asking customers to provide periodic updates on their financial positions. Therefore, the fact that a new customer passed the Group’s credit checks (before taking on any debt with the Group) does not demonstrate that she would necessarily be able to obtain credit elsewhere once she has become a roller and incurred debt with the Group.
147.Similarly, in 2011, some 13.2% of the balances owed to the Group were in arrears by more than 28 days (with this figure falling to 10.3% in 2015). We accept Mr Johnson’s evidence that this figure is a percentage of balances owed to the Group (and does not demonstrate that 13.2% of the Group’s customers are in default[29]). We also accept that there are a number of reasons why accounts might be in arrears ranging from fraud to the fact that the Group might be regarded as a non-priority debt and customers might regard it as more important to pay their rent or mortgage payments than pay the Group. However, a customer choosing to prioritise payment of rent over payments to the Group would clearly be suffering a degree of financial distress since she would not be able to pay both her rent and sums owed to the Group.
148.HMRC also referred extensively to a report prepared by Anna King (the “Anna King Report”), a customer insight specialist employed by the Group, in 2008. One of the stated objectives of that report was to produce a profile of body mass index (“BMI”) categories for customers of the Group and establish the prevalence of obesity within the Group’s customer base. The report also sought to compare the BMI profiles of customers of the Group with profiles of the population of Great Britain as a whole. Finally, the report sought to examine attitudes of N Brown customers who are obese and compare those with attitudes of obese women in Great Britain generally. Unsurprisingly, given that the Group is focused on providing fashion to “plus size” customers, the report concluded that customers of the Group are more likely to be obese than the average woman in Great Britain. The report also set out a set of impressionistic conclusions as to attitudes of obese customers including:
Women who are obese and particularly obese N Brown customers appear to be less disciplined than average GB women in various areas of their lives including eating habits, their choice of hobbies and financially. They spend money without thinking and use credit or cards to buy things they would otherwise be unable to afford.
149.We concluded, however, that the Anna King Report was of little assistance in building a picture of the use of the Group’s credit offering by its customers or categories of them. The Anna King Report struck us as impressionistic, and not rooted in rigorous statistical analysis. When it came to considering attitudes to credit, the Anna King Report simply took some attitudes to credit that were said by a “TGI 2007 2006-7 Battery of Lifestyle/Attitude Statements” to be “more likely”[30] to be expressed by an obese woman than an average woman in Great Britain. Having concluded (no doubt accurately) that the Group’s customer base contained a higher proportion of obese women than the proportion of the Great Britain population as a whole, the report concludes that the Group’s customers are more likely than the Great Britain population as a whole to have the stated attitudes to credit. Overall, the purpose of the Report was to consider what high-level conclusions could be drawn about attitudes of the Group’s customers to a variety of matters taking into account that the Group’s customers consist of a greater proportion of obese women than would be found in the population as a whole. It did not purport to be a scientific analysis of attitudes to credit generally and we do not consider that it adds much to an analysis of the characteristics of the Group’s roller customers.
150.Our overall conclusion is that a good proportion of the Group’s roller customers would have access to other sources of credit. In particular, the fact that roller customers are disproportionately from a DE social background did not satisfy us that they would necessarily have restricted access to credit elsewhere as we concluded from Mr Johnson’s evidence that such customers could still in many cases obtain credit, even if that was more expensive. We reject the submission that rollers generally tend to show a “reckless” attitude to credit of the kind described in the Anna King Report. (Of course, given the Group’s large customer base there will doubtless be some customers who adopt such an attitude, but we do not consider that to be reflective of rollers’ attitudes generally.) However, equally, as the Group has acknowledged, a fair proportion of the Group’s roller customers may struggle to obtain credit elsewhere[31] and may struggle to keep up repayments on the totality of their household debt.
151.With this evidence, we have reached the following broad conclusions:
(1) The Group’s “core customer base” could, in the period in dispute, fairly be categorised as its roller customers.
(2) It is not possible to make meaningful generalisations about rollers as a whole since there are so many of them and they will have different reasons for choosing to roll a credit balance. However, we consider that customers in the AB social grade are under-represented among rollers and that customers in the DE social grade are over-represented.
(3) A good proportion of the Group’s roller customers consist of individuals who have access to other sources of credit. However, rollers have, by definition, chosen to access the Group’s credit rather than other credit that may be available to them. Therefore, while it overstates matters to say that, as a general matter, the Group’s roller customers had limited access to other credit, we still consider that a customer wishing to purchase goods from the Group prior to 2014, and not pay immediately, would, in many cases “need” the Group’s credit in the sense that the Group’s credit enabled them to make a purchase without needing to make finely calibrated decisions on how to pay for it at the time of ordering.[32]
(4) After 2014, the Group’s customers could choose to access credit offered by other providers by using a credit card for their purchase without opening a personal account. However, as we have concluded at [20], by 31 May 2016, the introduction of this choice had yet to have any significant effect on the Group’s business.
(5) We agree with HMRC that a fair proportion of the Group’s roller customers consist of individuals who might struggle to obtain credit elsewhere. On any view, such customers “need” the Group’s credit to purchase goods.
(6) We reject HMRC’s submission that the Group’s roller customers tend, on average, to exhibit a reckless or impulsive attitude to credit, though we accept that there will be some individual rollers who do so.
152.In 1999, HMRC approved the use of a partial exemption special method (the “1999 PESM”) which provided a special method of apportionment to calculate the deductibility of VAT incurred on certain “pots” of costs used in making both taxable and exempt supplies. The special method in respect of residual marketing costs was a “page count” method, whereby the number of catalogue pages which advertised or referred to exempt activities divided by the total number of catalogue pages represented the extent of exempt use.
153.However, in 2005-06, HMRC expressed concerns that the 1999 PESM did not secure a fair and reasonable attribution of residual input tax to taxable supplies for the various pots. In relation to residual marketing costs, for example, HMRC criticised the page count method because it took no account of the exempt use of non-catalogue marketing material such as leaflets and did not work well in the context of TV, radio and website advertising. HMRC therefore invited the Group to submit an alternative partial exemption special method for approval by HMRC before 29 May 2006, the start of the Group’s next tax year. The Group submitted an alternative PESM, but HMRC did not approve it.
154.In those circumstances, and in order to protect HMRC against the risk of a loss of tax while a new PESM was being approved, on 26 May 2006, HMRC served on the Group a “special method override notice” (a “SMON”) effective from the start of the Group’s next tax year under the provisions of Regulation 102A of the VAT Regulations. In consequence of the service of the SMON, the Group was required to show in its returns, the difference between the attribution of input tax to taxable supplies under the 1999 PESM and “an attribution which represents the extent to which the goods or services used by [the Group] are used … or are to be used… in making taxable supplies”.
156.Officer Todd was clear in his evidence, and we accept, that the reservations HMRC had at [155] related to the use of a “one pot” standard method. They had no problem at all with the idea that the Group’s input tax recovery on marketing costs should be calculated on a turnover basis by calculating a fraction equal to the Group’s exempt finance income divided by taxable turnover from its fashion business.
157. Once HMRC had issued the SMON, discussions continued between the Group and HMRC with a view to agreeing a new PESM. However, ultimately those discussions broke down. Having reviewed the VAT returns that the Group submitted in response to, and following the issue of the SMON, HMRC decided that the input tax recovery that the Group was claiming did not represent the extent to which goods and services were used in making taxable supplies and made VAT assessments. Those assessments are summarised in Appendix One to this decision.
158.Article 1(2) of Council Directive 2006/112/EC[33] (the “PVD”) sets out the overall scope of VAT as follows:
(2) The principle of the common system of VAT entails the application to goods and services of a general tax on consumption exactly proportional to the price of the goods and services, however many transactions take place in the production and distribution process before the stage at which the tax is charged.
On each transaction, VAT, calculated on the price of the goods or services at the rate applicable to such goods or services, shall be chargeable after deduction of the amount of VAT borne directly by the various cost components.
159.Pursuant to Article 135(1)(b) of the PVD, Member States are obliged to exempt from VAT:
(b) the granting and the negotiation of credit and the management of credit by the person granting it.
160.Article 167 and Article 168 set out the right to deduct input tax incurred on supplies of goods and services to a taxpayer where those goods and services are used for the purposes of taxable transactions as follows:
Article 167
A right of deduction shall arise at the time the deductible tax becomes chargeable.
Article 168
In so far as the goods and services are used for the purposes of the taxed transactions of a taxable person, the taxable person shall be entitled, in the Member State in which he carries out these transactions, to deduct the following from the VAT which he is liable to pay:
(a) the VAT due or paid in that Member State in respect of supplies to him of goods or services, carried out or to be carried out by another taxable person;
Chapter 2
Proportional deduction
Article 173
1. In the case of goods or services used by a taxable person both for transactions in respect of which VAT is deductible pursuant to Articles 168, 169 and 170, and for transactions in respect of which VAT is not deductible, only such proportion of the VAT as is attributable to the former transactions shall be deductible.
The deductible proportion shall be determined, in accordance with Articles 174 and 175, for all the transactions carried out by the taxable person.
2. Member States may take the following measures:
…
(c) authorise or require the taxable person to make the deduction on the basis of the use made of all or part of the goods and services;
…
Article 174
1. The deductible proportion shall be made up of a fraction comprising the following amounts:
(a) as numerator, the total amount, exclusive of VAT, of turnover per year attributable to transactions in respect of which VAT is deductible pursuant to Articles 168 and 169;
(b) as denominator, the total amount, exclusive of VAT, of turnover per year attributable to transactions included in the numerator and to transactions in respect of which VAT is not deductible.
Member States may include in the denominator the amount of subsidies, other than those directly linked to the price of supplies of goods or services referred to in Article 73.
Article 175
1. The deductible proportion shall be determined on an annual basis, fixed as a percentage and rounded up to a figure not exceeding the next whole number.
162.For the purposes of this appeal, there was no dispute that the relevant provisions of EC law had been properly incorporated into UK law. We will not, therefore, set out in detail the UK domestic provisions in s25 and s26 of VATA 1994 governing input tax deductibility.
163.The provisions dealing with partial exemption in Article 173 to Article 175 of the PVD are largely reflected in the VAT Regulations. Regulation 101 provides, relevantly, as follows:
(1)…the amount of input tax which a taxable person shall be entitled to deduct provisionally shall be that amount which is attributable to taxable supplies in accordance with this regulation.
(2) … in respect of each prescribed accounting period—
(a) goods imported or acquired by and goods or services supplied to, the taxable person in the period shall be identified,
(b) there shall be attributed to taxable supplies the whole of the input tax on such of those goods or services as are used or to be used by him exclusively in making taxable supplies,
(c) no part of the input tax on such of those goods or services as are used or to be used by him exclusively in making exempt supplies, or in carrying on any activity other than the making of taxable supplies, shall be attributed to taxable supplies,
(d) where a taxable person does not have an immediately preceding longer period and subject to subparagraph (e) below, 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 him bears to the value of all supplies made by him in the period,
(e) the attribution required by subparagraph (d) above may be made on the basis of the extent to which the goods or services are used or to be used by him in making taxable supplies…
(10) In this regulation “residual input tax” means 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.
164.Pursuant to Regulation 102(1) of the VAT Regulations, HMRC may approve or direct the use by a taxable person of a method other than the “standard method” set out in Regulation 101. We will refer to that as a “partial exemption special method” or “PESM”. Where a taxable person is using a PESM, it must continue to do so unless HMRC approves or directs that its use is to terminate.
165.Regulation 102A allows HMRC to serve a notice (a “special method override notice”) requiring a taxable person to make adjustments to the results of its partial exemption calculation under a PESM. Regulation 102A (and Regulation 102B which prescribes the effect of such a notice) provide as follows:
102A—
(1) Notwithstanding the Commissioners' powers to serve a notice under regulation 102, where a taxable person—
(a) is for the time being using a method approved or directed under regulation 102, and
(b) that method does not fairly and reasonably represent the extent to which goods or services are used by him or are to be used by him in making taxable supplies,
the Commissioners may serve on him a notice to that effect, setting out their reasons in support of that notification and stating the effect of the notice.
(2) The effect of a notice served under this regulation is that regulation 102B shall apply to the person served with the notice in relation to—
(a) prescribed accounting periods commencing on or after the date of the notice or such later date as may be specified in the notice, and
(b) longer periods to the extent of that part of the longer period falling on or after the date of the notice or such later date as may be specified in the notice.
102B—
(1) Where this regulation applies, a taxable person shall calculate the difference between—
(a) the attribution made by him in any prescribed accounting period or longer period, and
(b) an attribution which represents the extent to which the goods or services are used by him or are to be used by him in making taxable supplies,
and account for the difference on the return for that prescribed accounting period or on the return on which that longer period adjustment is required to be made, except where the Commissioners allow another return to be used for this purpose.
166.In C-4/94 BLP Group v Customs & Excise Commissioners [1995] STC 424, the CJEU concluded that goods and services would be “used for” taxable or exempt transactions only if there was a “direct and immediate link” between the goods and services and those transactions. Therefore, the entitlement to deduct input tax is not determined by a taxpayer’s subjective motivation or purpose but by reference to objective factors. Moreover, in C-126/14 'Sveda' UAB v Valstybine mokesčių inspekcija prie Lietuvos Respublikos finansų ministerijos [2016] STC 447, the CJEU confirmed that is necessary to take into account all objective factors surrounding the transactions concerned.
167.In case C-98/8, Midland Bank plc v Customs & Excise Commissioners [2000] STC 501, the CJEU provided a further gloss on the concept of a “direct and immediate link” stating, at [30] of their decision that, for the right to deduct to arise, the expenditure in which the input tax is incurred must be a “cost component” of the relevant output transaction. Moreover, it is a matter for the national courts to determine, in any particular case, whether the requisite “direct and immediate link” is present.
168.Mr Prosser and Ms Sloane submitted that the CJEU’s decision in Midland Bank plc demonstrated that, in order for a “direct and immediate link” to exist between an input transaction and an output transaction, there must necessarily be a “close temporal link” between those transactions. There was some support for their submission in [29] of the Advocate General’s opinion in that case. However, we do not consider that this aspect of the Advocate General’s reasoning formed part of the CJEU’s conclusion. Indeed, at [31] to [33] of the decision, the CJEU avoided giving detailed guidance on the nature of a “direct and immediate link”, stating that this was a question for the national courts to determine in each case (although the CJEU did indicate at [31] that there is in general no direct and immediate link between an output transaction and services used by a taxable person after that output transaction has taken place). Therefore, while we consider that the existence or otherwise of a “close temporal link” may be relevant to the question of whether a “direct and immediate link” exists, it is not determinative of that question.
169.The question posed by the PVD at the attribution stage is whether or not goods and services received are “used for” taxable transactions, exempt transactions or both. The Midland Bank and BLP authorities set out the CJEU’s approach as to how that question of “use” should be determined. Since the concepts of a “direct and immediate link” and a “cost component” as a means of determining how inputs are used first made an appearance in the jurisprudence of the CJEU, they have been explained and applied in a number of separate decisions. In Mayflower Theatre Trust Ltd v Revenue & Customs Commissioners [2007] STC 880 the Court of Appeal gave a succinct summary of the relevant principles with Carnwath LJ (as he then was) saying:
The main principles … are not controversial… (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'); (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.
170.The “use” that must be identified for the purposes of Article 173 of the Principal VAT Directive is “economic use” (see, for example HMRC v London Clubs Management Ltd [2012] STC 388). The mere physical use to which goods or services are put is not necessarily determinative of how they are “used” for VAT purposes. However, physical use is not irrelevant: as it may shed a light on the economic use of goods or services. The focus on “economic use” can make it appropriate to consider the source or potential source of profit of a business. For example, in the London Clubs case, the Court of Appeal referred with approval to a decision of the VAT Tribunal in Aspinall’s Club Ltd v Revenue and Customs Commissioners (2002) VAT Decision 17797. In that decision, the VAT Tribunal considered a business that made exempt supplies of gaming and taxable supplies of catering. The question arose as to the extent to which certain “residual” services were used for the purposes of the taxable catering services[34] with the taxpayer arguing that the apportionment should be made by reference to a comparison of floor space used for gaming and floor space used for catering, with HMRC arguing that a standard “turnover” based apportionment should be followed. The VAT Tribunal carefully considered the economics of the taxpayer’s business, including the question of whether the taxpayer’s catering business could make a profit without the gaming business in reaching their conclusion.
171. As we have noted at [167] one approach to the question of “use” as endorsed by the CJEU is to ask whether the input is a “cost component” of the output. However, the Court of Appeal in Volkswagen Financial Services v The Commissioners for HM Revenue & Customs [2016] STC 417, echoing comments made by Advocate General Kokott in Sveda, has confirmed that this does not involve an examination of whether, subjectively, a trader incorporates the cost of an input transaction in the price of an output. For example, a trader carrying on an economic activity who chooses for business reasons to sell particular goods for less than their cost price is not to be denied input tax recovery for that reason alone. In Volkswagen Financial Services, at [55], the Court of Appeal expressed this point as follows:
As cases like Rompelman and Abbey National demonstrate, the ability of the taxable person to deduct input tax depends on its use for the purpose of the taxable transactions which he makes, not on whether that expenditure is actually built into the price charged for the supply. The way in which he chooses to attribute those costs to the supplies he makes and so recover them from his consumers is likely to be based on a range of factors including tax considerations. It may be highly material to the apportionment of the costs and therefore the input tax between the different supplies which are made. But non sequitur in my judgment that the inclusion of the costs in the price of a particular supply is in itself a pre-condition to the recovery of the input tax. There is simply no authority to justify such a rule.
172.Finally, we were shown a number of authorities dealing with the VAT treatment of marketing expenses incurred by taxpayers running partially exempt businesses. We will not refer to those authorities in detail in this decision since, while we found it helpful to consider how different courts and tribunals have approached the Attribution Issue in different circumstances, we do not consider that those authorities fundamentally alter the statements of principle that we have set out above. However, one important general point emerges from those authorities. The Attribution Issue is focusing on how the Group uses goods and services received in connection with the Group’s marketing. The focus, therefore, is on use by the Group. It is not focusing on how the Group’s customers will read or interpret the marketing material. Of course, the way that a customer might respond to marketing material may well shed a light on how the Group is using the goods and services that it receives and so the response of a customer is of potential relevance. However, the Attribution Issue cannot be determined simply by deciding how a customer would view the material in question.
173.As noted at [161], the PVD provides, in the first instance for the proportion of residual VAT that is creditable to be calculated by applying a standard methodology based on turnover. However, Member States are permitted to require taxpayers to determine the amount of their credit for input tax on the basis of, among others, the “use” of the relevant goods or services. In C-332/14 Wolfgang und Dr Wilfried Rey Grundstücksgemeinschaft v Finanzamt Krefeld [2017] STC 1668, the CJEU determined that Member States could exercise that right provided the alternative method “guarantees a more precise determination of the deductible proportion of the input VAT than that arising under the [turnover-based] method”. That did not mean, however, that the alternative method has to be the most precise possible.
174.In Revenue and Customs Commissioners v Lok’nStore Group plc [2015] STC 112, the Upper Tribunal considered the apportionment question in the context of a business’s overheads[35]. The Upper Tribunal noted that the “attribution” and “apportionment” stages are logically distinct. In the context of “overheads” of a partially exempt business, the “attribution” stage is concerned with the question of whether there is a direct and immediate link between goods and services and the taxpayer’s whole economic activity. If there is such a direct and immediate link, the apportionment stage is then concerned with identifying the extent to which the overheads are used in making taxable supplies. It follows that questions of whether there is a “direct and immediate link” between inputs and outputs is not relevant at the apportionment stage. However, considerations that were relevant at the attribution stage might also be relevant at the apportionment stage as well.
175.In Lok’nStore Group plc the FTT concluded that the cost to the taxpayer of making taxable supplies was materially higher than the cost of making exempt supplies. They noted that such a disparity pointed against the use of a turnover based system of apportionment in the following paragraph:
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 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 VAT-bearing overheads.
Having reached that conclusion, and having considered the way in which the taxpayer used its overhead costs, the FTT concluded that the PESM proposed by the taxpayer produced an attribution that was fairer and more reasonable than an attribution based on turnover. That conclusion was not disturbed on appeal.
176.In this section, we will answer the Attribution Issue separately for all 15 categories of marketing expenditure. As will be seen, much of our analysis on one category of expenditure can be applied to other categories as well. For that reason, we will analyse early categories in considerable detail but, where that analysis can simply be read across to later categories, we will simply say so with the result that the analysis of later categories will be more abbreviated than that of earlier categories.
177.The Group considers that the costs of producing physical marketing material are only used for the purposes of making exempt supplies where that marketing material contains a “reference to credit”. Similarly, the Group argues that physical marketing material that refers only to the Group’s credit offering is necessarily attributable only to exempt supplies so that none of the associated input tax is recoverable.
178.HMRC disagree with both of the Group’s propositions. Their position is that it does not matter at all whether the physical marketing material refers to credit or not. Their “overarching submission” (which they consider applies to all categories of marketing material) is that the Group has a “unitary business” with the retail and finance businesses being inextricably linked. The Group’s business, therefore, involves the opportunity to earn income from the making of exempt supplies of credit to a predictable percentage of customers (i.e. rollers). Viewed in this way, in HMRC’s submission, it necessarily follows that all the Group’s marketing has a dual function. Therefore, HMRC criticise the Group’s focus on “references to credit” as being unduly focused on physical use of the materials and as ignoring the part played by physical marketing material in the generation of credit income.
(2) The link between the marketing material and a supply of credit could be said to be indirect. Most fundamentally, having received physical material that refers only to goods, a recipient may simply decide to order no goods at all in which case there will be no supply of goods or credit. Even if the customer decides to buy some clothes that have been advertised, she has real choices to make, some of which may result in no supply of credit taking place at all. From 2014, she could decide to pay for the goods on her own credit or debit card without opening a personal account, in which case the Group would make no supply of credit to her. Before 2014, she would, in the overwhelming majority of cases, have to make her order through a personal account. However, even then, she would have up to 7 weeks, depending on when precisely in the month she placed her order, to decide whether to pay her balance in full or whether to roll the balance and take credit from the Group. The Group would have no control over whether she decides to take credit or not.
(4) The way that the Group measures the success of its physical marketing material is primarily by analysing “accepted demand” (see [97]), a metric that measures demands for goods, not credit. Moreover, the marketing budget is, at least primarily, determined by reference to the level of “accepted demand” that the Group is targeting. Those factors suggest that the Group regards itself as using physical marketing material to sell goods. If the Group thought that it was using the material to sell credit as well, it might be expected to take supplies of credit into account either when fixing its marketing budget, or measuring the success of its marketing material.
180.We have not found the Attribution Issue easy. We understand why the parties have been apart on it for so long: the issue is genuinely debatable. However, on balance, we have concluded that, strong though the points set out at [179] are, there is nevertheless a direct and immediate link between goods and services that the Group receives associated with physical marketing material and exempt supplies of finance that the Group makes.
181.Central to our conclusion is that the Group’s business as a whole demonstrates that there is a two-way relationship between sales of goods and sales of credit. One aspect of that two-way relationship is obvious: since the Group provides its customers with credit only where that customer is purchasing goods from the Group, there can be no supply of credit unless there has first been a supply of goods. However, the relationship between goods and credit does not end there as the Group’s offering of credit to its customers supports its ability to sell goods. More specifically:
(1) As we have concluded above, the Group’s core customer base consists of its rollers. They tend to be the Group’s most loyal customers, they generate the majority of the Group’s retail sales measured by value and the overwhelming majority of its turnover.
(2) There was much discussion at the hearing about why rollers choose to take up the Group’s credit. However, there can be little doubt that in fact, the Group’s core customer base does take up that credit. In those circumstances, it is too simplistic to say that the Group sells retail goods and hopes that its customers will exercise a choice to take up the credit. The Group’s rollers, its core customers, take up the credit to buy the goods and for those customers it is just as accurate to refer to the decision to buy the goods being facilitated by the availability of credit as it is to refer to the choice to take the credit being enabled by the Group’s provision of desirable goods.
182.Until 2014, it was only possible to order goods from the Group through the medium of a personal account. As we have noted at [12], a customer would not even be allowed to open a personal account until she had satisfied the Group’s credit checks. Therefore, until 2014, any inducement to buy goods from the Group was necessarily an inducement to buy goods through the medium of a personal account specifically. Of course, the personal account was not just a facility for the delivery of credit: it enabled customers to “try before you buy” which was very important given that it was a distance selling retailer of clothes. However, nevertheless, up until 2014 the end result that, viewed objectively, the Group’s marketing material sought to achieve was a sale through a personal account that was already set up to provide the customer with credit if she wanted to take it. That is a further link between the marketing material and exempt supplies of finance.
183.Since 2014, it has been possible for customers to order goods from the Group without opening a personal account with the result that, since 2014, the force of the additional link between goods and credit referred to at [182] was somewhat diminished. However, that additional link was by no means eliminated. First, by the end of the period to which this appeal relates, the introduction of cash customers had yet to have a significant effect on the Group’s business. In addition, as we note at [18], the Group actively sought to encourage people to open a personal account even after 2014. Therefore, even after 2014, the Group was using physical material to induce customers to make a purchase through the medium of a personal account even though, by then, it accepted that it had to allow customers to pay cash if they wanted to.
184.The existence of the “two-way relationship” between goods and credit diminishes the force of the point at [179(2)]. Certainly, there will be many customers who order goods and then make a separate and independent choice about whether to use the Group’s credit to pay for them. However, the Group’s core customers, its rollers, do use the Group’s credit facility. Without the Group’s credit business, the retail business would be significantly smaller, as noted at [121]. As we have explained at [151], the Group’s credit therefore enables and facilitates retail sales. It is not simply something that customers think about after they have decided to make a retail purchase.
185.The two-way relationship also has a specific effect on the volume of physical material received by rollers. As we have noted at [106], the Group understandably sends most physical material to its binary 15 customers who are its most loyal customers. Since a roller is more likely than a non-roller to be a binary 15 customer (see paragraph [108] above), it follows that, on average, a roller can expect to receive more physical marketing material of the kind falling in sub-categories (a) to (c) from the Group than a non-roller[36].
186.The existence of the two-way relationship also diminishes the force of the point at [179(1)]. We quite accept that a customer receiving a catalogue that does not mention credit at all might not view that catalogue as an inducement to purchase credit. However, the question is not how a customer reacts to the Group’s physical marketing material, but how, viewed objectively, the Group “uses” it. Once it is appreciated that, (i) viewed objectively, a significant proportion of retail sales can take place only if the customer concerned also takes the Group’s credit and (ii) viewed objectively, there is a predictable correlation between retail sales and finance income and (iii) the Group’s core customer base consists of rollers who do take the Group’s credit, a conclusion that the Group uses its physical marketing material in making supplies of both goods and credit becomes less surprising. Mr Prosser and Ms Sloane submitted that this approach focuses on the “results” of the deployment of marketing material and not its “use” but we do not accept that submission. The “two-way” relationship between goods and credit suggests that, when the Group sends out physical marketing material that refers only to goods, it is nevertheless using that material to promote credit as well. Such an approach does not involve brushing aside, or ignoring, the fact that the material refers only to goods, as Mr Prosser and Ms Sloane submitted. Rather it involves a conclusion that, despite referring only to goods, the two-way relationship between goods and credit means that the material is also used to promote credit.
187.That in turn leads to a further point. The logic of the Group’s case that focuses on whether a particular document contains a “reference to credit” is that comparatively little of its physical marketing material is used to promote its finance offering even though finance income accounts for some 30% of its turnover. Effectively, that amounts to an argument that the Group’s credit largely “sells itself”. The Group has few physical stores and so its marketing material represents an important medium for displaying its goods and services. In those circumstances, we consider it unlikely that the Group could obtain such significant revenues from its financial business without marketing credit to a significant extent. Rather, we consider that an analysis of “economic reality” supports a conclusion that, when the Group sends out physical material that refers only to its goods, it is still using that material to promote both goods and credit. Given the two-way relationship between goods and credit, physical marketing material that is used to stimulate demand for goods must inevitably also be used to stimulate demand for credit.
188.A similar point can be made about the matters referred to at [179(4)]. Of course the subjective ways in which the Group chooses to measure the success of its marketing, or set its marketing budget, cannot determine the objective question of how its physical marketing material is “used”. However, even putting that point to one side, the two-way relationship between sales of goods and sales of credit means that, viewed objectively, there is no compelling economic reason why the Group needs to monitor the effect of its marketing material on sales of credit specifically. The Group knows that there is a predictable correlation between sales of goods and sales of credit. Therefore, if marketing is successful in stimulating demand for goods, it will necessarily be successful in stimulating demand for credit to a predictable extent.
189.Similarly, as we have noted at [100], the fact that the marketing budget is set by reference to targeted “accepted demand” does not mean that the marketing budget does not depend on sales of credit. The two-way relationship between goods and credit means that some accepted demand can be generated only if there is a corresponding supply of credit.
190.We have considered carefully whether the additional period, referred to at [179(3)] which a customer has to make up her mind whether to take the Group’s credit prevents the link between the marketing materials and any supply of credit from being direct or immediate.[37] We do not, however, consider that it does. First, given the two-way relationship we have referred to, in many cases, the decision to take credit will be made before the goods are even ordered. Second, the delay is relatively short, at most a few weeks. Therefore, while the delay is a factor to be taken into account, we do not consider that it would prevent any link between marketing materials and credit being either direct or immediate.
191.Pausing there, the conclusions that we have expressed above point us to the conclusion that there is a direct and immediate link between the Group’s physical marketing materials and supplies of credit and that accordingly, the costs of producing those marketing materials are “used for” both supplies of goods and supplies of finance. However, before confirming that we should deal with an important submission that Mr Prosser and Ms Sloane made which gave us much pause for thought. We were invited to consider the expense that the Group incurs in purchasing a particular black dress from a supplier. If it is the case that the link between the retail and finance sides of the business means that all the Group’s physical marketing costs are used in making both taxable and exempt supplies, Mr Prosser and Ms Sloane submitted that the same must necessarily be true of the costs of the black dress[38]. Therefore, it was submitted, that if marketing input VAT is residual, so input VAT on the black dress must also be residual. Since input VAT on the black dress was, in the submission of Mr Prosser and Ms Sloane a paradigm example of VAT attributable only to taxable supplies, it followed that there was a contradiction in HMRC’s “overarching submission” which demonstrated it could not be correct.
192.Ms McCarthy and Ms Barnes answered the apparent conundrum by referring to evidence which, in their submission, demonstrated that the Group did not price its credit by reference to the cost of black dresses that it purchases whereas, self-evidently, it incorporates the cost of purchasing black dresses into the retail price for those dresses. Therefore, they argued, the cost of the black dress could not be a “cost component” of supplies of credit.
194.Of course, another possible answer to Mr Prosser’s and Ms Sloane’s apparent conundrum is that the Group’s marketing costs are “overheads” that have a direct and immediate link with the Group’s business as a whole but do not have a direct and immediate link with either taxable supplies or exempt supplies, whereas the costs of the black dress are attributable only to taxable supplies of goods. As we have noted above, without “conceding” the point, HMRC did not make any such argument at the hearing, considering that they are precluded from making it by the Court of Appeal’s decision in Dial a Phone. Since that was HMRC’s clear position in the hearing before us, we think it would be quite wrong of us to express any view on whether the Group’s marketing costs are, or are even capable of being, overheads and we will not do so.
196.By virtue of Regulation 90(1)(a) of the VAT Regulations, each payment of interest that the Group receives is consideration of a separate supply of finance. In reaching the conclusion at [195], we should not be taken as concluding there is a “direct and immediate link” between the costs associated with producing physical marketing material and every separate and successive supply of finance that the Group receives. Rather, we are concluding that there is a direct and immediate link between those costs and some exempt supplies of finance (which is sufficient to make input VAT on those costs “residual” for VAT purposes).
197.Finally, we note that in the course of the hearing, and in oral argument, we were referred to a number of decisions of this Tribunal (and the predecessor VAT Tribunal) dealing with whether marketing expenses used by a partially exempt business were either (i) used only for taxable supplies; or (ii) used only for exempt supplies only; or (iii) used for both taxable and exempt supplies or (iv) were overheads (on the basis that there was no “direct and immediate link” with taxable or exempt supplies, but there was a link with an economic activity as a whole). We agree with the parties that the question in this case is essentially factual and must be applied in the context of all relevant circumstances. It would unduly lengthen this decision if we explained each point of difference with each previous Tribunal decision. We will therefore say only that we have considered all the authorities to which we referred and, where we are reaching a different conclusion in the context of this appeal, it is because we think that the relevant underlying facts are different.
198.There are some differences between this category of marketing costs and those associated with physical material that we have considered in connection with Category 1. For example, whereas much of the physical material described in Category 1 is sent to existing customers of the Group[39], Category 2 marketing material appears in mainstream media and therefore the Group cannot control whether it is read by existing customers. However, the similarities between Category 2 and Category 1 material are much more significant than the differences. Although Category 2 material may be aimed at a wider category of reader (as it is not restricted to existing customers of the Group), like Category 1 material, it is used to stimulate demand. For the reasons we have given in connection with Category 1 material, the demand being stimulated is for both taxable supplies of goods and exempt supplies of finance. Our overall conclusion, therefore, for reasons similar to those set out in the context of Category 1, is that there is a direct and immediate link between the Group’s costs associated with Category 2 marketing material and both taxable and exempt supplies. Marketing input VAT associated with Category 2 marketing material is therefore “residual” for VAT purposes.
199.For the purposes of the Attribution Issue, there is no material difference between TV adverts and advertising falling within Category 2. Both are intended to stimulate demand and neither are targeted at either existing customers or potential customers (since the adverts are distributed by mass media and the Group cannot control who will see the adverts). We acknowledge that the TV adverts that we saw scarcely mentioned the Group’s credit offering at all. However, for reasons that we have given in relation to Category 1, viewed in the light of the Group’s business as a whole, we do not consider that this prevents the costs associated with producing the adverts from being “used for” the making of exempt supplies. Our overall conclusion, therefore, is that marketing input VAT associated with Category 3 marketing material is “residual” for VAT purposes.
200.Pay per click marketing is used in a similar way to the marketing materials set out in Categories 1 to 3. It results in a promotional message being delivered to customers, or potential customers of the Group. That message is not in the form of a paper catalogue or television advert, but it remains a message that is used to promote the Group’s economic offering. Therefore, as with Categories 1 to 3, the question to determine is whether “pay per click” marketing is used only for the purposes of sales of goods (as the Group argues) or for the purposes of both selling goods and selling credit (as HMRC argue).
201.As we have noted at [54], the vast majority of the keywords that the Group chose to designate related to goods that the Group sells, rather than credit. That, therefore, was the digital analogue of the Group’s physical material being devoted overwhelmingly to pictures of, and descriptions of, goods that the Group sells as opposed to its credit offering. For reasons that we have explained in the context of the Group’s physical marketing material, we do not consider that the “use” of pay per click advertising can be determined solely by reference to the search terms that the Group chooses to specify. Rather, the points that we have made about the two-way relationship between goods and credit apply to “pay per click” marketing in much the same way as they apply to physical marketing materials. We note that “pay per click” marketing is targeted indiscriminately. Therefore, the point we make at [185] does not apply to “pay per click” marketing and it cannot be said that rollers, as a group, receive a greater quantity of pay per click marketing than non-rollers. However, that minor difference has not caused us to reach a different conclusion on the costs of pay per click marketing and we have concluded that VAT on such costs is “residual” for VAT purposes.
202.We consider that, for the purposes of these appeals, costs falling in Category 5 are in all material respects the same as those falling in Category 4. In particular, the fact that the costs incurred ensure that the Group’s websites are optimised by reference to searches for the Group’s goods (as opposed to its credit offering) does not alter our overall conclusion that marketing input VAT incurred on costs connected with Category 5 material is “residual” for VAT purposes.
203.This category of cost is in substance the same as costs associated with catalogues or “pay per click”. The Group incurs costs in order to stimulate demand for its economic offering. Given the nature of the Group’s business, and the two-way relationship between goods and credit, we are not satisfied that the costs are used only to stimulate demand for goods. We have therefore concluded, for reasons similar to those set out above, that the costs are used for the purposes of both supplies of goods and supplies of credit and VAT on those costs is therefore “residual” for VAT purposes.
204.Again, costs associated with Category 7 marketing are used to stimulate demand. Even though the banner adverts referred to at [63] feature pictures of the Group’s goods (and no “reference to credit”), we consider that they are nevertheless used to promote both goods and credit for reasons that we have given in connection with the Group’s physical marketing material falling into Category 1. Accordingly, VAT on the costs associated with this category of marketing expense is “residual” for VAT purposes.
205.Telemarketing results in the Group giving its customers, or potential customers, a promotional message. In most cases, that promotional message referred only to goods although, from April 2016, a new “credit-led” script was used in a minority of cases. However, given that Category 8 material is, like Category 1 material designed to stimulate demand, the points that we have made in relation to Category 1 costs mean that, whether the telemarketing refers to goods, credit or both telemarketing is nevertheless used in order to sell both goods and credit. It follows that input VAT associated with Category 8 telemarketing costs is “residual” for VAT purposes.
206.The parties were agreed that marketing input VAT on Category 9 costs is “residual” for VAT purposes. Nevertheless, they both asked the Tribunal to express its own conclusion, and reasons, on the issue.
207.By contrast with Categories 1 to 8 above, Category 9 marketing does not involve the actual delivery of a promotional message. Rather, it involves the Group obtaining advice as to how it might best formulate a promotional message. Therefore, Category 9 shares some similarities with Category 13 (market research). As noted at [69], brand development costs are used to enable the Group to target the entirety of its economic offering (both goods and credit) by taking into account the attributes of a typical customer. The “wants” and “needs” of the customer referred to at [69] included an analysis of her likely attitude to credit. For that reason, even though HMRC are not seeking to argue that marketing input VAT associated with Category 9 is an overhead, we consider that the parties were right to agree that VAT on this category of marketing cost is “residual” for VAT purposes.
208.Again the parties were agreed that VAT on this category of expense was “residual” for VAT purposes. In his oral submissions, Mr Prosser explained that the Group accepted that this category of VAT is “residual” because the lists of potential customers that it obtains could be used just as much to send marketing material that references credit as marketing material that makes no mention of credit. The logic, therefore, of Mr Prosser’s position was that, if the Group could demonstrate that the lists it purchased were used only to send potential customers information on the Group’s goods, the costs would be attributable only to taxable supplies and would not be residual.
209.We consider that the parties were right to agree that this category of input VAT is “residual”. We would, however, give slightly different reasons from Mr Prosser. Even if the lists that the Group purchases were used only to send potential customers information on goods, we consider the input VAT incurred would still be residual. In such a case, the lists would still be “used” in order to deliver a promotional message to stimulate demand. For reasons that we have given in relation to Category 1, given the two-way relationship between goods and credit, that promotional message stimulates demand for both goods and credit. Therefore, the points we have made in relation to Category 1 lead us to the conclusion that marketing input VAT associated with Category 10 is “residual”.
210.The parties were agreed that VAT on this category of expense is “residual” for VAT purposes. We consider that they were correct to do so. Category 11 material is similar to Categories 1 to 8: it is used to stimulate demand and, given the two-way relationship between goods and credit it is necessarily used to stimulate demand for both goods and credit.
211.Viewed objectively, the Group pays celebrities because it hopes that, by presenting sympathetic and popular figures with whom customers and potential customers will identify, demand for the Group’s economic offering will be stimulated.
212.As we have noted at [75], celebrity ambassadors’ duties under their agreements with the Group are expressed by reference to the Group’s retail products. Moreover, celebrity ambassadors receive commission by reference to sales of the Group’s goods, and not by reference to sales of credit. Those factors certainly point away from a conclusion that the costs of engaging celebrity ambassadors are “used for” the Group’s exempt supplies.
213.However, for reasons we have set out above, the mere fact that celebrities’ duties and remuneration are expressed by reference to the Group’s goods, and not its credit offering, is not determinative. Rather, given the two-way relationship between goods and credit, we have concluded that the costs of celebrities are used to stimulate demand for both goods and credit for reasons explored in more detail in relation to Category 1 above. We have therefore concluded that VAT on this category of cost is “residual” for VAT purposes.
214.We consider that this category of marketing expense is somewhat different from the other categories that we have considered. First, it is much broader than other categories and expenses incurred in Category 13 are not united by a single common theme. As we have noted, the topics on which the Group commissioned research ranged from reasons for high levels of returns of a particular corset to customers’ attitudes to credit. Second, the market research is much less directly related to a communication to customers that is used to stimulate demand. Market research does not involve a direct message being given to customers or potential customers (unlike Categories 1 to 8, 11 or 12 referred to above). Market research may not even be used in formulating a message to customers (unlike Category 9) or in providing information that facilitates the delivery of a message to customers (such as Category 10). Of course, market research as to why customers return a particular corset might lead to changes in the way that the Group promotes that corset. For example, if that market research indicated the corset was not suitable for particular customers, it might be marketed differently to customers for whom it is more suitable. But, the market research will not inevitably lead to a promotional message being delivered. For example, if market research revealed that a corset is frequently returned because of a particular design flaw, no promotional message to customers might result: the Group might simply ask its suppliers to redesign the corset.
215.Those distinctions are important. As we note in our analysis of Category 1 expenses, the nature of the Group’s business, including the “two-way” relationship between goods and credit means that it uses promotional material that refers only to goods to stimulate demand for both goods and credit. However, while that conclusion holds good in relation to promotional material, it is not of universal application.[40] Therefore, it by no means follows from our analysis above that input VAT on all market research costs is necessarily residual. For example, when the Group asks its customers to provide it with thoughts on its underwear offering, it is not delivering those customers a promotional message that must necessarily serve to stimulate demand for both underwear and credit. Rather, like the example of the black dress set out at [193], the costs are attributable solely to taxable supplies.
216.We therefore consider that, in relation to market research costs, Mr Prosser and Ms Sloane were substantially correct to submit that VAT on market research costs is residual only where the matters being researched include matters relating to credit. More specifically, the correct approach to the Attribution Issue is as follows:
(1) First, it should be ascertained whether the cost truly relates to “market research” (i.e. recording data on particular matters) or is more in the nature of advice as to how the Group should formulate a promotional message. As we have noted, “brand development” costs are the province of Category 9 and, because those costs relate to how the Group’s offering of both goods and credit should be positioned, VAT on those costs is “residual”.
(2) Since the Group has not sought to argue that any costs associated with market research are “overheads”, we consider that input VAT associated with market research whose sole subject matter is credit, such as that the Bamboo report referred to at [70], is attributable only to exempt supplies and is irrecoverable.
(3) Input VAT associated with market research relating solely to goods is attributable only to taxable supplies and is recoverable in full.
(4) Market research that relates to both goods and credit (including market research that is focused on the Group’s brands generally) is attributable to both taxable supplies and exempt supplies[41].
217.In substance, Category 14 involves promotional material that is sent to customers or potential customers to stimulate demand. It is not in substance any different from Category 1 material although, since it is not focused on existing customers of the Group, it cannot be said that rollers will, as a group receive more of this material than non-rollers. However, that is a relatively minor point of distinction. For reasons that we have set out in relation to Category 1, we consider that input VAT on this category of material is “residual”.
218.These costs, like those falling within Categories 2, 3 and 14 (for example) are used to stimulate demand for the Group’s products. For reasons that we have already given in relation to Categories 2, 3 and 14, the associated costs are used for the purposes of both taxable and exempt supplies.
219.As we have noted, prior to the hearing before us, the text of the preliminary issues that the parties wanted the Tribunal to determine was different from those set out at [3] above. However, albeit in different form, those preliminary issues raised both the Attribution Issue and the Apportionment Issue.
220.In its skeleton argument served prior to the hearing, the Group set out the following approach:
(1) On the Attribution Issue, the Group argued that, for most of the 15 categories of marketing expense, the relevant question was whether the relevant marketing material referred in some way to the Group’s credit offering. If there was no such “reference to credit” then costs incurred in connection with that marketing material were used only for the purposes of its taxable supplies so that VAT on those costs was fully recoverable. If the marketing material contained a “reference to credit” then, in most cases, the Group accepted that it was used for the purposes of both taxable and exempt supplies so that the associated VAT was residual and the Apportionment Issue needed to be determined.
(2) Having put its case on the Attribution Issue in that way, the Group argued that the Apportionment Issue was conceptually straightforward in most cases: it was necessary to determine the proportion of the marketing material that contained a “reference to credit”. So, for example, in the context of physical material falling within Category 1, a “page count” method should be followed. If a 500-page catalogue contained 10 pages that had a “reference to credit” and 490 pages that did not, 2% of the costs of producing the catalogue (i.e. 10/500) were used for the purposes of exempt supplies and it followed that 98% of the VAT on those costs should be recoverable. A “page count” method of apportionment clearly could not be applied to marketing material that did not have pages (such as, for example, the costs of telemarketing falling within Category 8). However, in many cases the principle could be applied by analogy. So, for example, the Group might receive an invoice for 10,000 telemarketing calls of which only 300 involve the use of a script which refers to credit as well as goods. In such a case 3% of those costs should be treated as used for the purposes of exempt supplies so that 97% of the VAT should be recoverable.
(3) The Group acknowledged that there would be some cases in which the approach outlined at [(1)] or [(2)] above might be more difficult. For example, it was not straightforward to see how it could be applied to invoices for television adverts. On that issue, the Group made some general points about why a turnover basis of apportionment would not produce a precise result and outlined a possible modification to the turnover basis that it considered would lead to a more precise result.
221.Therefore, in large part, the Group’s submissions on the Apportionment Issue proceeded on the basis that its submissions on the Attribution Issue were correct with the result that there was a meaningful distinction between marketing material containing a “reference to credit” and material containing no such reference. Mr Prosser accepted that there was a “lacuna” in the Group’s skeleton argument in this regard but submitted that it simply was not practicable to outline a comprehensive case on the Apportionment Issue if its position on the Attribution Issue was incorrect without knowing why its position on that issue was incorrect. He therefore submitted that the fairest course for the Tribunal to adopt if it did not accept the Group’s case on the Attribution Issue would be to invite further submissions from the parties limited to the Apportionment Issue. However, at the Tribunal’s request, Mr Prosser and Ms Sloane have made some submissions on how the Apportionment Issue should be approached if the Group’s case on the Attribution Issue is rejected.
222.In their skeleton argument served prior to the hearing, HMRC took the following approach:
(1) On the Attribution Issue, they made the “overarching submission” that because of the close relationship between goods and credit in the Group’s business, it necessarily followed that all of the Group’s marketing expenses were used for the purposes of both taxable and exempt supplies whether or not that marketing material contained any “reference to credit”.
(2) On the Apportionment Issue, residual input VAT should be determined by reference to a turnover basis of apportionment.
(3) Applicable VAT legislation demonstrated that there was a “hierarchy” of apportionment methodologies with the turnover methodology set out in Article 174 of the PVD being at the top of that hierarchy. Accordingly, in HMRC’s submission, a methodology other than a turnover methodology should be applied only if the turnover methodology did not fairly and reasonably reflect the “use” of the relevant inputs and the Group has the burden of proof in that respect.
223.Therefore, HMRC’s arguments on the Apportionment Issue were not as reliant on their case on the Attribution Issue succeeding as, irrespective of the answer on the Attribution Issue, HMRC were arguing for a turnover basis of apportionment.
224. As will be seen from Part III of this decision, we have largely rejected the Group’s case on the Attribution Issue which leads to the question how we should approach the Apportionment Issue. The parties had different perspectives on the approach we should take:
(1) The Group argues that the best course would be to ask the parties for further submissions on the Apportionment Issue that are based on our conclusion on the Attribution Issue. However, it has also made the submissions on the Apportionment Issue referred to above.
(2) HMRC argued that case law of the CJEU referred to at [173] demonstrates that a methodology other than the turnover methodology should be applied only if it “guarantees a more precise result”. The burden is on the Group to establish that an alternative method does this. Therefore, if the Tribunal is satisfied that the turnover methodology that HMRC are proposing produces an accurate calculation of the extent to which residual input VAT is used for the purposes of making taxable supplies and the Group fails to demonstrate an alternative method that guarantees a more precise result, the Tribunal should determine the Apportionment Issue in favour of HMRC and uphold a turnover methodology. Both parties have incurred significant costs in connection with a lengthy hearing assuming that both the Attribution Issue and the Apportionment Issue would be fully disposed of. Accordingly, Ms McCarthy and Ms Barnes submitted that it would be disproportionate for the Tribunal to subject the parties to a lengthy iterative process that seeks to identify the most accurate method of apportionment.
(3) Ms McCarthy did, however, accept that if the Tribunal was not satisfied that HMRC’s turnover methodology produced an accurate calculation of creditable residual input VAT but was not satisfied that the Group’s alternative method produced a more precise result, the Tribunal should consider giving some guidance and/or requesting further submissions.
225.If the hearing before us had been a final determination of the Group’s appeal against the assessments that HMRC had made, that appeal could certainly be determined by reference to the Group’s burden of proof. Therefore, if in a final appeal against assessments, the Group could not persuade the Tribunal that it has a basis of apportionment that guarantees a more precise result than HMRC’s turnover methodology, the Tribunal might well decide simply to dismiss the Group’s appeal insofar as relating to the Apportionment Issue[42]. However, this is not a final appeal against assessments: it involves the Tribunal determining preliminary issues which, it is hoped, may facilitate settlement between the parties or, at least, make future Tribunal hearings relating to this dispute briefer and more focused. In those circumstances, considerations of the Group’s burden of proof alone should not determine how those preliminary questions should be determined.
226.Given that we have largely determined the Attribution Issue in favour of HMRC, we consider that a fair, just and proportionate method of dealing with the Apportionment Issue is as follows:
(1) The preliminary issues in this appeal all arise in the connection with a special method override notice that the Group has received under Regulation 102A of the VAT Regulations requiring the Group to calculate its creditable input tax on the basis of “use”. HMRC have concluded that the calculation of input tax in the Group’s VAT returns for the relevant period does not reflect the “use” of those inputs and, accordingly, have made assessments that apply a turnover methodology at the apportionment stage.
(2) Therefore, we will start by considering whether the turnover methodology that HMRC have applied results in the Group obtaining credit for input tax on a basis that reflects “use” of the relevant input tax. As noted below, our conclusion is that HMRC’s turnover methodology does not properly reflect the use of the relevant inputs (in at least some cases).
(3) Having explained why we do not consider that HMRC’s turnover methodology properly reflects use of the relevant inputs, we will consider the Group’s competing methodologies. We will explain why we do not consider those properly reflect use either.
227.The turnover methodology that HMRC propose involves the calculation of a pure fraction that expresses the Group’s taxable income as a proportion of its total income. We respectfully agree with the following statement on the turnover methodology of this Tribunal at paragraph 52 of Lok’nStore Group Plc v HMRC [2012] UKFTT 589:
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 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 VAT-bearing overheads.
228. We do not consider it matters that Lok’nStore was concerned with the treatment of overheads, whereas this appeal, given the way HMRC have set out their case, does not. Even though we have concluded, on the Attribution Issue that, in most cases, marketing expenses that the Group incurs are used for the purposes of both taxable and exempt transactions, it does not necessarily follow that they are used to the same extent in exempt transactions as they are in taxable transactions.
229.Nor do we consider that the First-tier Tribunal’s conclusions set out above are expressly or impliedly qualified by their earlier findings as to the way in which the taxpayer’s overheads were incorporated into the prices that it charged for its services. Rather, we consider that the First-tier Tribunal is making the simple point that, if costs that are categorised as “residual” for VAT purposes are not used rateably in making both taxable and exempt supplies, the standard turnover method of apportionment may not produce an accurate determination of the extent to which those costs are used in taxable or exempt transactions. It will be for the fact-finding tribunal to decide whether costs are used rateably by reference to all of the evidence: in some cases considering how costs are incorporated into prices that a business charges may provide the answer and in other cases, different considerations will be relevant.
230.Therefore, we consider that there is force in the core submission of Mr Prosser and Ms Sloane that if it costs the Group more in terms of marketing expense to generate £1 of goods income than it does to generate £1 of finance income, the pure fraction that the standard methodology produces may not accurately reflect the extent to which marketing costs are “used” to make taxable supplies.
231.Mr Prosser and Ms Sloane then submitted that the evidence demonstrates that it does cost the Group more in terms of marketing expense to generate £1 of sales income than it does to generate £1 of finance income for the following broad reasons:
(1) There is a clear link between the cost of marketing and taxable revenues (not least since the Group’s marketing budget is set at between 16% and 18% of “accepted demand”, a metric that addresses taxable sales of goods). There is no such clear link between the costs of marketing and exempt revenues[43].
(2) The Group would still have to incur the vast majority of its marketing costs in order to sell goods even if it did not also provide credit. Admittedly, the Group might well make fewer sales to rollers if it did not provide credit, but there is no reason to think that such customers would not purchase the Group’s goods if they could not also obtain its credit.
(3) As noted at [26], costs associated with recruiting new customers are higher than costs of retaining existing customers and inducing them to order more goods.
232. We do not consider that the arguments at [231(1)] or [231(2)] are borne out by the evidence. As we have noted at [101], the two-way link between the Group’s sales of goods and its sales of credit means that achieving a particular level of “accepted demand” necessarily requires the Group to achieve a particular level of sales of credit. Therefore, while it is correct to say that finance income is not taken into account expressly, as a separate item, when the marketing budget is set, it is wrong to say that there is no link between the level of finance income and the amount of the marketing budget. Moreover, as we have concluded at [102], if the Group had no credit business, its retail business would be materially smaller than it is currently. It follows that the level of accepted demand generated by the retail business and so the marketing budget, which is linked to accepted demand would also be materially smaller.
233.The parties were agreed that “recruitment” marketing is more expensive than “retention” marketing. However, we do not consider that of itself demonstrates that HMRC’s turnover methodology fails to reflect the use that the Group makes of inputs in making taxable supplies. Customers of the Group need to be recruited as customers before they can purchase either goods or credit. Therefore, the fact that recruitment advertising is more expensive than retention advertising says relatively little about the use that the Group makes of its marketing costs in making taxable or exempt supplies respectively.
234.We do, however, accept the point that Ms Prosser and Ms Sloane make at [231(4)]. For example, suppose that a customer orders £200 of clothing through a personal account and rolls her credit balance. Assuming an APR of 39.9%, and that she pays only enough into her account to enable her to keep pace with interest charged, she would be paying interest of around £79.80 per year (or around £6.65 per month). Given the two-way relationship between goods and credit that we have explained in our analysis of the attribution issue, we have no difficulty with the proposition that the costs of marketing to this customer were “used for” the purposes of making the taxable sale of the goods and the exempt supply of finance. However, unless the customer makes further orders of goods, over time the extent to which costs of marketing are “used for” the purposes of exempt supplies of finance will get less and less. To give an extreme example, suppose that three years after buying the goods, the customer is still paying £6.65 per month by way of interest, but is not making any new orders of goods. It might be said that the finance income is largely received because of the customer’s personal circumstances rather than because of any significant use of costs that the Group incurred in preparing marketing material in year three. This is not a purely academic point. Mr Johnson’s unchallenged evidence was that in 2012, some 15.4% of active customers made the minimum payment then required on their account or less. He also said that a new customer (whose monthly minimum payment is set at 1/12th of the balance on their account) would take 17 months to clear a balance if they make minimum payments only. More established customers have a minimum payment of £5 per month and, on average, it would take such a customer 34 months to pay off their balance. There is clearly scope for the Group to receive material amounts of interest from a customer years after that customer last placed an order and in circumstances where the customer is showing no signs of placing new orders.
235.Ms McCarthy and Ms Barnes argued that the point at [231(4)] does not call into question the accuracy of the standard method of apportionment in the context of this particular business. They pointed to the predictable link between the Group’s interest income and the sales that it makes. They also submitted that, over multiple years, the standard turnover methodology would produce the correct result. For example, if the Group received comparatively low levels of interest income in Year 1 (for example, because it made fewer sales, or made more sales towards the end of a year thereby generating less “in year” interest from those sales) the standard method would give the Group a relatively higher level of VAT recovery on marketing costs (as taxable turnover in Year 1 would be high relative to exempt turnover). Therefore, in Year 2, when the Group receives the benefit of interest income derived from sales in Year 1, it was right, they submitted, that the standard method should result in a lower level of VAT recovery.
236.However, we do not consider that this submission adequately deals with the point at [226(4)]. Although, as a whole, we have concluded that there is a predictable link between the Group’s sales of goods and the amount of interest that it received, it does not follow that marketing material is “used for” the purpose of interest received several years after a sale was made (in circumstances that may be different from those prevailing when the sale was made) to the same extent as it is “used for” the first few payments of interest that immediately follow the sale of the goods. Moreover, the fact that, in particular circumstances, a higher than usual recovery of VAT in Year 1 should fairly be balanced by a lower than usual recovery in Year 2 does not diminish the force of the point that, if marketing material is not “used for” later receipts of interest to the same extent as it for earlier receipts, the turnover methodology may not operate fairly.
237.We recognise that the Group has several million customers. Therefore, we are not suggesting that the example set out at [234] is typical or representative. However, we do consider that it is an issue that is simply not addressed by the use of an unadjusted fraction based purely on the application of a turnover methodology. The fact that it is not addressed means that we are not satisfied that HMRC’s turnover methodology accurately determines the extent to which residual input tax is “used for” the purposes of making taxable supplies.
238.For completeness, we are not satisfied that the “page count” basis that the Group put forward[44] accurately determines the extent to which residual input VAT is “used for” the purposes of the Group’s taxable supplies. Indeed, we did not understand Mr Prosser or Ms Sloane to be arguing that such a basis would be appropriate in circumstances where the Tribunal rejected their primary case on the Attribution Issue. Our reasons can be shortly stated.
239. Central to the “page count” basis of apportionment is the argument that the Group’s material should be categorised depending on whether it contains a “reference to credit” or not. Exempt use of material containing a “reference to credit” would then be determined by calculating the proportion of the material that contains a “reference to credit”. For physical material, in the Group’s submission, that would be achieved by counting pages that contain a reference to credit. For other material some other kind of itemised basis would be used whose purpose would be to measure the extent of the “references to credit”.
240. Our conclusion on the Attribution Issue means that we do not accept that there is a fundamental distinction between marketing material that contains a “reference to credit” and material that does not. Rather, subject to the limited exceptions we have identified in our analysis of the Attribution Issue, marketing material is “used for” sales of goods and sales of credit whether or not it contains an express “reference to credit”. Given that conclusion, we consider that a basis of apportionment that seeks to measure the extent of “references to credit” would be contrary to the economic reality of the Group’s business.
241.Mr Prosser and Ms Sloane outlined an alternative basis of apportionment that is based on Mr Lovelace’s evidence outlined at [125] above. To make that calculation easier to follow, they proposed that we accept that it costs 10p to generate £1 of product sales from a roller customer and 20p to generate £1 of product sales from a non-roller customer.
242.Building on the fact, referred to at [115] above that approximately 70% of the Group’s turnover consists of revenue from the sales of goods and 30% of turnover comes from interest income, Mr Prosser and Ms Sloane outlined an example calculation that could apply to the following hypothetical facts:
(1) The Group makes turnover of £100 of which £70 is turnover from sales of products and £30 is interest income.
(2) £50 of the £70 sales of products comes from customers who are rollers[45]. The remaining £20 comes from customers who are not rollers. (All of the £30 interest income naturally comes from rollers).
(3) Applying Mr Lovelace’s evidence, marketing costs associated with the £50 product sales to rollers are £5 (10% of £50) and the marketing costs associated with the £20 product sales to non-rollers are £4 (20% of £20). The input VAT on those costs would be £1.80 (20% of the total costs of £9) of which £1 would relate to costs of marketing to rollers and 80p would relate to the costs of marketing to non-rollers[46].
243. Mr Prosser and Ms Sloane said that, applying HMRC’s turnover basis to the above figures, the proportion of input tax disallowed would be 30% (the £30 exempt interest income divided by total turnover of £100). Accordingly, 30% of the total input VAT of £1.80 would be disallowed, a disallowance of 54p.
244.By contrast, Mr Prosser and Ms Sloane submitted that, if a turnover methodology were to be applied, it would be fairer to apply it only to rollers and the revenue generated from them. Applying that alternative turnover methodology, they submitted that the total input VAT associated with marketing to rollers was £1. £50 of revenue from rollers comes from taxable sales of products and £30 from exempt supplies of finance. Therefore 30/80 of the £1 input VAT should be disallowed, a total disallowance of 37.5p.
245.We are not satisfied that this alternative method of apportionment accurately reflects use of residual costs in making taxable supplies. Our principal reservation is that the alternative method involves treating marketing material differently for VAT credit purposes depending on whether it is sent to a roller or to a non-roller, with input tax associated with material sent to non-rollers being fully recoverable and input tax associated with material sent to rollers being restricted. We regard that distinction as artificial. As we have concluded in our conclusion on the Attribution Issue, most if not all of the marketing material that the Group dispatches is “used” to stimulate demand for both goods and financial services. Therefore, even if a particular catalogue is sent to a customer who happens, at a particular point in time, to be a non-roller, that catalogue is still “used” in part to stimulate demand for credit even if the customer has not taken credit in the past or does not take credit shortly after receiving the catalogue. Since we do not accept a fundamental principle underpinning the alternative methodology (that input tax on material sent to non-rollers should necessarily be recoverable in full), we do not accept that the alternative calculation accurately determines creditable input tax.
246.A further reservation is that the figures that Mr Lovelace produced assume that it is possible to track whether marketing material has been sent to a roller or a non-roller. That assumption may be correct in relation to certain physical material that the Group sends to its existing customer base (such as catalogues). However, it is less clear how it could be adapted to other important categories of marketing cost which are not directed at specific individuals (for example pay per click advertising or television adverts).
249.The rationale for this proposed methodology was similar to that outlined at [234] above as Mr Prosser and Ms Sloane argued that, if the Group made a sale in Year 1, and received interest from that sale in Year 2, that interest had nothing to say about how marketing material was “used” in Year 2. By contrast, interest received in Year 2 from sales also made in Year 2 was relevant to the determination of the use of marketing costs in Year 2.
250.As we have noted, we accept the principle that the Group does not use marketing material in connection with interest that it receives a long time after the related sale of goods to the same extent as it uses that material in connection with interest that it receives shortly after the sale of goods. In that respect, we thought that there was some merit in the proposal.
251.However, we consider that the proposal contains a fundamental flaw. The Group’s year end falls at the end of February, or in the first few days of March in each year. For a retailer such as the Group, Christmas is a particularly busy time and the Group engages in particularly heavy marketing activity to try to stimulate demand. As we have found, that marketing is used for the purposes of both taxable supplies of goods and exempt supplies of finance. However, on the Group’s approach, only around two months’ interest attributable to sales taking place over the Christmas period (i.e. interest receivable in January or February) would be included within the standard turnover fraction in Year 1. In addition, any resulting interest would be excluded from the calculation in Year 2 (since it would arise in respect of sales made in Year 1). That anomaly of itself is such that we cannot conclude that that it produces an accurate calculation of use.
252.Our overall conclusion on the Apportionment Issue is as follows:
(1) The relevant input tax should not be apportioned using the standard turnover methodology that HMRC have used when making the assessments. However, subject to the modification outlined at [(3)] below, an apportionment based on turnover is capable of producing an accurate reflection of the extent to which residual input VAT is used in making taxable supplies.
(2) The relevant input tax should not be apportioned using any of the alternative methodologies that the Group has proposed.
(3) In principle, a more accurate methodology for apportioning relevant input tax could be achieved by adapting the turnover methodology that HMRC have proposed so as to take account of the point at [226(4)]. Some finance income should be excluded from the denominator of the turnover fraction to reflect this point.
(4) We are not in a position to explain how an appropriate exclusion of some finance income should be determined (beyond noting that the Group’s proposal of excluding all in-year interest as outlined at [248] does not produce an accurate result). However, the principle that should be applied is reasonably straightforward. The more time that goes by during which a customer makes no new orders for goods but still pays interest on her credit balance, the less the extent to which marketing material she received in the past (and so expenses that the Group incurred in connection with that material) is “used for” exempt supplies. An accurate calculation based on turnover needs to take that factor into account without producing anomalous results in the first year of operation of that calculation (or subsequent years).
(5) In reaching the conclusions set out above we are not excluding the possibility that there are other methods of apportionment that produce an accurate calculation of creditable input VAT. For example, given the central importance of rollers to the Group’s business (which has underpinned many of our conclusions on the Attribution Issue), it is possible that there are methods of apportionment that compare turnover from rollers with turnover from non-rollers[47]. However, no such alternative method has been presented to us and we consider it would be unfair for us to seek to formulate such an alternative method ourselves and without submissions from the parties.
253.Our answers to the Attribution Issue are set out separately in relation to each category of marketing expenditure above. Our answers to the Apportionment Issue are set out at [253].
254.We will separately make directions relating to the future case management of this appeal in the light of our decisions on the preliminary issues.
255.This document contains full findings of fact and reasons for the decision. Any party dissatisfied with this decision has a right to apply for permission to appeal against it pursuant to Rule 39 of the Tribunal Procedure (First-tier Tribunal) (Tax Chamber) Rules 2009. The application must be received by this Tribunal not later 30 April 2019 (a longer deadline than would usually apply as the parties have told us that they wish to explore settlement). The parties are referred to “Guidance to accompany a Decision from the First-tier Tribunal (Tax Chamber)” which accompanies and forms part of this decision notice.
Amended pursuant to Rule 37 of the Tribunal Procedure (First-tier Tribunal)(Tax Chamber) Rules 2009 on 4 March 2019
|
VAT quarters |
Amount of assessment (£) |
Date of assessment |
Date of review |
Date of appeal |
|
08/06 to 05/09 |
5,281.076 |
9 July 2010 |
2 February 2011 |
3 March 2011 |
|
08/09 to 05/11 |
5,948,804 |
10 January 2013 |
7 June 2013 |
4 July 2013 |
|
08/11 to 05/13 |
9,203,504 |
1 May 2014 |
8 July 2014 |
4 August 2014 |
|
08/13 to 05/14 |
6,088,164 |
6 March 2015 |
none |
17 April 2015 |
|
08/14 to 05/15 |
7,546,976 |
23 October 2015 |
7 January 2016 |
4 February 2016 |
|
08/15 to 05/16 |
8,354,756 |
12 September 2016 |
25 November 2016 |
15 December 2016 |
|
Total |
42,423,280 |
|
|
|
[1] Section 43 of the Value Added Tax Act 1994 treats all supplies made by, or to, members of the Group (other than intra-Group supplies which are disregarded) as being made by, or to, JDW as representative member. In this decision, as a shorthand, we will refer to supplies being made by, or to, “the Group” as a shorthand.
[2] Ms McCarthy and Ms Barnes emphasised at the hearing that HMRC were not “conceding” that there was a direct link between marketing costs and taxable supplies of goods. However, HMRC consider that the Court of Appeal’s decision in Dial a Phone v Customs & Excise Commissioners [2004] STC 987 prevents them, as a matter of law, from arguing that the marketing expenses are “overheads” that are attributable to the entirety of the Group’s economic undertaking. Therefore, in this preliminary hearing, HMRC seek to argue that there is a direct and immediate link between marketing costs and both taxable supplies of goods and exempt supplies. They are not seeking to argue that the costs are overheads (which would be the case if there was no direct and immediate link between marketing costs and either taxable or exempt supplies, but instead only a direct and immediate link with the Group’s undertaking as whole).
[3] Ms Ascott was not able to attend the hearing. However, her witness statement was adopted by Jennifer Morris, who was available for cross-examination. HMRC therefore did not object to the admission of Ms Ascott’s witness evidence and we have admitted it.
[4] Different brands might offer different ways of ordering. Some of the Group’s brands are “internet only” brands. In addition, exceptionally there might be different ways of ordering a particular product.
[5] This was 28 days for some of the period under appeal.
[6] There were some exceptions for, for example, personalised or perishable goods, or goods that could not sensibly be returned used for hygiene reasons.
[7] The Group exercises discretion to waive interest and late payment charges for customers who, having defaulted on their obligations, enter into approved payment plans where the Group considers that the customer could not afford to pay interest and late payment charges and still reduce the principal amount due.
[8] They would receive their statement up to 28 days after the home approval period expired and then had a further three weeks to clear their balance without incurring an interest charge.
[9] So, a customer who had an existing personal account was free to choose to “bypass” the personal account and pay for goods with her own credit or debit card.
[10] It was not clear to us whether this “page” of the catalogue was actually a section that had separately been “stitched in”, but we do not think much turns on this since it was clearly intended to form part of the overall message that the catalogue was conveying.
[11] Defined as “Products” in the relevant agreement.
[12] Without the Group’s retail business, there could be no finance business (since the Group provides finance only to customers seeking to buy its retail products). As explained in more detail at [121], without the finance business, the Group’s retail business would be much smaller in size.
[13] We do not think that Mr Tucker was using the expression “overheads” in a technical VAT sense in his evidence. We have therefore concluded that he was saying that the Group determines the sale price of its retail goods without taking into account marketing costs, even though neither party is arguing that the Group’s marketing costs are “overheads” in a VAT sense.
[14] Although the Group’s systems could accommodate promotions under which customers are offered a 0% interest rate for a period of a few months.
[15] i.e. physical “products” such as clothing and fashion items. Sales of credit are not included within the calculation of “accepted demand”.
[16] Until 2014, as noted at [12], it was only possible to “try before you buy” by opening a personal account. Since 2014, customers can “try before they buy” by ordering multiple items, paying for them via a credit or debit card and, when they return unwanted items, obtain a refund. However, as Mr Johnson noted in his evidence, that is a less attractive proposition since it involves customers paying for all goods up front. Moreover, as noted at [18], the Group actively promotes the personal account.
[17] These percentage figures do not take into account “binary zero” customers referred to at [104]. However, we do not consider that affects the general point that binary 15 customers are highly loyal and generate disproportionate quantities of retail sales.
[18] The statistics in this paragraph are distilled from figures that the Group prepared for 2015-16. None of the Group’s witnesses suggested the figures were not representative of the whole period under appeal. We have, therefore, concluded that the position was at least broadly the same throughout that period.
[19] A roller might receive the occasional “credit only” leaflet of the kind referred to at [110] which a non-roller would not receive but since such mailings are infrequent it does not alter our conclusion.
[20] Therefore, the Group considers that, while the finance business generates only, on average, 27.5% of total income, it generates, on average, 69.1% of total “net contribution” because, in the Group’s view, the expenses associated with the finance business are much lower than the expenses associated with the retail business. As we will note, we do not consider that this is the same as saying that 69.1% of the Group’s profit comes from financial services.
[21] The calculation of the annual increase in financial services and retail turnover is our own. However, Mr Lovelace performed a similar calculation, in his first witness statement, of similar figures limited to JDW (and not the Group as a whole).
[22] Mr Lovelace did not himself calculate this percentage, but HMRC have correctly calculated it from his figures. As discussed below, we do not consider that this figure is determining the profit generated by the retail business as a percentage of total profit.
[23] An “active” account was defined as one where the customer concerned had placed an order that year. For the purposes of this table, a “roller” was defined as someone who had incurred more than £5 of interest in the given year (so as to exclude “accidental” rollers who incurred a small amount of interest when they forgot to pay their balance on time).
[24] The calculation is by value of sales, not by number. “Sales” for these purposes referred to gross despatches made by the Group (so that, for example, where a customer sought to open a personal account to order goods, but the order was not completed because the customer failed a credit check and the personal account was not opened, this does not count as a “sale”).
[25] For some reason, Mr Lovelace gave data only on the three specific financial years mentioned. He gave no data for financial years in between.
[26] In cross-examination, Mr Johnson said that, in this respect, his view was different from that of his chief executive.
[27] It was not clear to us how large this sample was or how the Group had sought to ensure that it was representative of its customers as a whole. Nor was it clear to us what the term “high roller” meant as it was not clearly defined. We have concluded that it is referring to a customer of the Group who is “rolling” a credit balance considered large in the context of credit balances generally.
[28] Ms Spindler also accepted in cross-examination that “we certainly offer credit to people who may find it difficult to get more traditional forms of credit”.
[29] Mr Johnson said in his evidence that 7% of “customers” are in arrears. We assume that he meant that 7% of customers with credit balances were in arrears and was not intending to say that 7% of all customers were in arrears. Since only rollers have credit balances and since rollers represent 26% of the Group’s total customers, if 7% of total customers were in arrears then 27% of rollers (i.e. 7/26) would be in arrears. That struck us as unlikely since, if 27% of rollers were in arrears, but only 13.2% of balances were in arrears that would tend to suggest that customers with lower balances were more likely to have arrears, a conclusion we regarded as counter-intuitive.
[30] The Anna King Report does not say how much more likely.
[31] As noted at [144] the Group in its Regulatory Business Plan acknowledged that some proportion of the Group’s customers fall into this category without saying what that proportion was. Since it was not suggested that the proportion was small, and given the evidence on arrears rates and “high rollers” having “stretched finances”, we have concluded that a fair proportion of the Group’s customers fall into this category.
[32] The hypothetical customer referred to at [143] might, therefore, be able to pay for the cinema tickets using her credit card and clear the balance of her personal account using £65 in cash. She could also pay £65 cash for the cinema tickets, borrow £65 from a pay-day lender and use that to clear the balance on her personal account. She could, having ordered goods through her personal account, choose to pay for them in full at the point of order with a credit card and borrow £65 from a pay-day lender to clear her credit card balance. But if she took any of these steps, she would not be a roller so we do not consider that these theoretical alternatives alter the conclusion that, in large part, until 2014 rollers “needed” the Group’s credit in a practical sense. Moreover, the availability of the Group’s credit means that the customer would not need to worry about how to pay for the coat when placing her order.
[33] Throughout this decision, we will refer to provisions of the Principal VAT Directive that were in force in respect of the majority of the VAT periods that are under appeal. The parties were agreed that the predecessor provisions set out in the First and Sixth VAT Directives were not materially different.
[34] So, strictly, this case related to the “apportionment” stage and not the “attribution” stage. However, both parties were in our view rightly agreed that the VAT Tribunal’s approach of ascertaining “economic use” was applicable at the attribution stage as well.
[35] As noted in more detail below, both parties have approached this appeal on the basis that the marketing costs at issue are not overheads (because there is a direct and immediate link between those costs and taxable supplies of fashion goods).
[36] Sub-categories (d) and (e) of Category 1 relate to physical marketing material that is sent to people who are not customers of the Group and therefore, it is not meaningful to examine the extent to which category (d) and (e) material is sent more to rollers than it is to non-rollers.
[37] We do not accept HMRC’s submission, based on Midland Bank that a time delay is necessarily completely irrelevant in cases where the input transaction precedes the output transaction as we consider that an analysis of periods of time is capable of being relevant when deciding whether a link is “direct” or “immediate”.
[38] If anything, it was argued, the link between the cost of the black dress and supplies of finance would be even more direct and immediate than the link between marketing costs and supplies of finance. The argument proceeded as follows: (i) a supply of finance could take place only if the black dress is sold; (ii) marketing costs only to induce a customer to buy the black dress and may not achieve their desired aim; (iii) by contrast, the cost of the black dress is closely related to the sale of that dress and so is more closely related to any supply of finance that takes place as a result.
[39] Though not the “mini-offers” and “inserts” and “direct mailings” set out as sub-categories (d) and (e).
[40] For example, as noted at [193], costs that the Group incurs in purchasing a black dress for resale do not have a direct and immediate link with exempt supplies of credit.
[41] As we have noted, HMRC have expressly declined to make an argument that any category of marketing costs, including market research costs, are overheads for VAT purposes.
[42] Even in a final hearing of an appeal against assessments, the Tribunal would retain a case management discretion to request further submissions or evidence if it wanted to. Any such discretion would need to be exercised judicially taking into account the overriding objective set out in Rule 2 of the Tribunal Rules which include dealing with cases in a proportionate manner.
[43] So, for example, it would cost the Group the same amount to send a catalogue to a non-roller customer as it would to send it to a roller customer. A similar point applies to other categories of cost. Therefore, there is a link between the number of the Group’s customers and its marketing costs, but no such link between marketing costs and exempt revenues.
[44] In which we include the “itemised” basis of apportionment that the Appellant proposed for material which, unlike physical material, does not have pages.
[45] £50 might be considered to be high since, using the figures at [123] above, some 59.3% to 64% by value of retail sales are made to rollers, but Mr Prosser and Ms Sloane were providing an indicative calculation only.
[46] In fact, Mr Prosser and Ms Sloane’s submissions assumed that the total input VAT was £9, but that was the amount of the assumed total cost, not the VAT element of that cost. We have corrected that anomaly in our explanation of, and analysis of, their submissions.
[47] The method analysed at [241] to [247] sought, at least in part, to do this but we rejected it as it had flaws.
