South Africa: Competition Tribunal
You are here:
SAFLII >>
Databases >>
South Africa: Competition Tribunal >>
2009 >>
[2009] ZACT 46
| Noteup
| LawCite
Competition Commission and Another v British American Tobacco South Africa (Pty) Ltd (05/CR/Feb05) [2009] ZACT 46 (25 June 2009)
Download original files |
PDF format
RTF formatCOMPETITION TRIBUNAL OF SOUTH AFRICA
Case No: 05/CR/Feb05
In the matter between:
The Competition Commission 1st Applicant
JT International South Africa (Pty) Ltd 2nd Applicant
And
British American Tobacco South Africa (Pty) Ltd Respondent
Panel : D Lewis (Presiding Member), Norman Manoim (Tribunal
Member) and M Mokuena (Tribunal Member)
Heard on: 6-10,13-15,17,27-29 and 30 August 2007; 19,21, and 25-28 September 2007; 27- 29 November 2007; 14-18, 21-24, and 31 January 2008; 1 February 2008; 5-9, and 12-14 May 2008; 20, 23, 25,26, 27, and 30 June 2008; 1-3 July 2008; and 30 and 31 October 2008.
Order Issued : 25 June 2009
Reasons Issued: 25 June 2009
REASONS FOR DECISION
BACKGROUND
Parties
The first applicant is the Competition Commission (“the Commission”), a body established in terms of section 19 of the Competition Act 89 of 1998 (“the Act”).
The second applicant is JT International South Africa (Pty) Ltd (“JTI”), a private company incorporated in terms of the company laws of the Republic of South Africa.1 JTI is the South African subsidiary of the multinational tobacco company, Japan Tobacco International, which is in turn a subsidiary of Japan Tobacco Inc.2
The respondent is British American Tobacco South Africa (Pty) Ltd (“BATSA”), a private company incorporated in terms of the company laws of the Republic of South Africa. BATSA, is a wholly owned subsidiary of British American Tobacco plc, the UK-based tobacco multinational.3 BATSA was formed out of the merger in 1999 between British American Tobacco (BAT) and Rothmans International.4 The latter was a South African-based company and a pillar of the Rupert family empire. Its brands accounted for an extremely significant share of South African cigarette sales, a dominant position which was inherited and extended by the merged entity.
The allegations
On 1 October 2003 JTI laid a complaint before the Competition Commission against its rival, BATSA. JTI alleged that BATSA is an ‘overwhelmingly dominant manufacturer of cigarettes in South Africa” and that its marketing practices constituted contraventions of sections 5(1), 8(c), 8(d)(i), and 8(d)(iii)’.
The Commission investigated the complaint and brought a complaint referral on 10 February 2005. The Commission did not refer the complaint in respect of the alleged section 5 and section 8(d)(iii) contraventions, but only in respect of alleged contraventions of sections 8(c), and 8(d)(i) of the Act which provide:
8. Abuse of Dominance prohibited
It is prohibited for a dominant firm to:
(a)
(b)
(c) Engage in an exclusionary act, other than an act listed in paragraph (d), if the anticompetitive effect of that act outweighs its technological, efficiency or other precompetitive gain; or
(d) engage in any of the following exclusionary act, unless the firm concerned can show technological, efficiency or other pro-competitive gains which outweigh the anticompetitive effect of its act-
(i)Requiring or inducing a supplier or customer to not deal with a competitor
The Commission’s referral alleged that certain agreements concluded between BATSA and selected cigarette retailers (BATSA’s “trade investment” agreements) and certain of BATSA’s retailer incentive programmes (“Vecta” and “Sprint” programmes) incentivised the retailers to market and sell BATSA brands irrespective of the price and/or quality advantages that competitor brands may offer consumers over BATSA brands and irrespective of customer demand. It was also alleged that these agreements and programmes ensure that BATSA secured preferential, alternatively exclusive, access to the point of sale for promotional purposes. In its Founding Affidavit, the Commission asked that the Tribunal grant the following orders:
“Determining that the conduct of the respondents is prohibited conduct in contravention of sections 8(d)(i) and/or 8(c) of the Act.
Declaring the trade investment agreements and/or the Vecta Rewards Programme and/or the Sprint Programme as void.
Ordering the respondent to pay an administrative penalty in terms of section 59(1)(a) of the Act up to 10% of the turnover generated during the firm’s preceding financial year.
Further and/or alternative relief.”
On 17 June 2005 JTI referred a complaint to the Tribunal alleging that BATSA had, by this conduct, also contravened Section 5(1) which provides that:
“An agreement between parties in a vertical relationship is prohibited if it has the effect of substantially preventing or lessening competition in a market, unless a party to the agreement can prove that any technological, efficiency or other pro-competitive, gain resulting from that agreement outweighs that effect”.
In its complaint referral, JTI alleges that BATSA is in a vertical relationship with the various retail outlets, the routes to market for cigarette manufacturers. These outlets, in addition to being sites of distribution, are also alleged to be important promotional sites, that is, sites for promoting particular cigarette brands through the use of, inter alia, point of sale (POS) signage at or near cigarette dispensing units (CDU’s) and the space allocated and positioning assigned to the various brands in the cigarette dispensing units located at the POS. Note that it is the ability of rival manufacturers to promote, rather than to distribute, their brands that BATSA’s various agreements with retailers are alleged to foreclose. JTI asked that the Tribunal make the following order:
“In terms of section 58(1)(a)(i) of the Act, interdicting BATSA from concluding restrictive agreements with retail outlets, the operators of “horeca” venues, as well as the operators of cigarette vending machines and the owners of venues where CVM’s are installed;
In terms of section 58(1)(a)(vi) of the Act, declaring the restrictive agreements between BATSA and retail outlets, the operators of “horeca” venues, as well as the operators of cigarette vending machines and the owners of venues where CVM’s are installed, to be void;
In terms of section 58(1)(a)(v) of the Act, declaring the restrictive agreements between BATSA and retail outlets, the operators of “horeca” venues, as well as the operators of cigarette vending machines and the owners of venues where CVM’s are installed, to constitute prohibited practices for the purposes of section 65.”
JTI also applied for and was granted permission to intervene in the proceedings launched by the Commission’s referral to the Tribunal. On 14 October 2005, JTI filed its “intervenor’s statement of complaint” alleging a breach of section 8(d)(i), alternatively a breach of section 8(c). Neither the Commission nor JTI persisted in the allegation of a breach of section 8(d)(iii), as had been alleged in JTI’s complaint to the Commission. In its intervenor’s statement of complaint JTI asked that the Tribunal issue the following order:
“In terms of section 58(1)(a)(i) of the Act, interdicting BATSA from inducing cigarette resellers (both retail outlets and horeca venues), to remove JTI signage and CDUs, conceal JTI products at CDUs and continuing with the loyalty incentive programmes initiated by BATSA;
Declaring that BATSA’s conduct is in contravention of sections 8(d)(i), alternatively, 8(c) of the Act;
Declaring in terms of section 58(1)(a)(v) of the Act, that BATSA’s conduct continues(sic) prohibited practices for the purposes of section 65;
Declaring the exclusive arrangements concluded between BATSA and cigarette resellers and the Vecta Awards Programme void in terms of section 58(1)(a)(vi) of the Act.”
In short then the burden of the allegations made by both the Commission and JTI is, firstly, that BATSA is dominant in the market for the supply of manufactured cigarettes in South Africa. They further allege that BATSA has abused its dominance by engaging in general exclusionary conduct prohibited by Section 8(c) of the Competition Act as well as the specific exclusionary abuse of inducement that is prohibited by Section 8(d)(i). This conduct, insist the applicants, redounds to the ultimate detriment of cigarette consumers whose choice is restricted and who, in consequence of this reduced competition, are obliged to pay higher prices than would otherwise be the case.5 As noted JTI additionally alleges a violation of Section 5(1) of the Competition Act, claiming that certain agreements concluded between the cigarette manufacturers and retailers – essentially the identical agreements that are the subject of the Section 8 allegations – have the effect of substantially preventing or lessening competition without any countervailing pro-competitive gains emanating from the conduct in question.
The actual conduct complained of does not emerge clearly from a bald recitation of the formal notices. Moreover the main thrust of the complaint, and certainly the nature of evidence relied upon by the applicants, has shifted since the original filings. Initially, the Commission and JTI’s complaints were bolstered by evidence of acts of sabotage that BATSA was accused of perpetrating, the so called “dirty tricks”.6 These incidents included removal of competitors’ stock and promotional material in store, exclusion from the store, and hiding competitors’ stock, among others.7 Affidavits from 49 deponents were used in support of these allegations. However, none of the deponents to these 49 affidavits was called to testify at the hearing. Indeed only one retailer – and a rather minor one at that – was called by the applicants. The Commission and JTI then sought to make their case by using the agreements between BATSA and the various retail channels and the testimony of economic and marketing experts. In the hearings before us the Commission and JTI now essentially allege that through a series of agreements concluded between BATSA and the retail channels – including the HORECA venues - and through conduct unilaterally perpetrated by BATSA, BATSA, relying on its dominant position in the market, has appropriated the promotional opportunities which, it is alleged, are uniquely available through these channels.
The principal allegations ventilated at the hearings are then that BATSA incentivised the various retail outlets to allocate space and position in the various cigarette display units on a preferential basis to BATSA. These included the variety of display units utilised in the grocery, convenience and forecourt stores as well as the vending machines used in the HORECA channel. In fact, although JTI has continued to insist that the question of the quantity of space allocated to the respective manufacturers’ brands remains a central element of its case, the Commission acknowledged that its case is principally concerned with the preferential positioning of BATSA brands in the display units, a preference that, it alleges, is secured by the agreements between the various participants in the retail channels and BATSA.8 In addition it is alleged that the agreements limited the ability of BATSA’s rivals to place secondary display units and advertising material at the POS. Further it is alleged that where BATSA’s rivals did succeed in placing promotional materials in the stores these were unilaterally removed by or at the instance of BATSA trade representatives. However, as we shall elaborate below, although these were central to the initial case and to the evidence submitted they were not rigorously pursued in the case that was presented to us.
It also appears that certain other of the allegations germane to the agreements have not been pursued. Hence while certain of the agreements did for a limited period provide for market share incentives – incentives which are generally intended to actively discourage retailers from selling non-BATSA brands – the unchallenged evidence is that these were not implemented and were quickly withdrawn by BATSA acting, it seems, on advice that they may be in contravention of the Competition Act.
It
The Hearings
The matter was heard over a marathon period of 50 days, these being the 6-10,13-15,17,27-29 and 30 August 2007; 19,21, and 25-28 September 2007; 27- 29 November 2007; 14-18, 21-24, and 31 January 2008; 1 February 2008; 5-9, and 12-14 May 2008; 20, 23, 25,26, 27, and 30 June 2008; and 1-3 July 2008. The Tribunal heard closing arguments on 30 and 31 October 2008. Various witnesses were called by the parties. The Commission called Mr Cornelius Potgieter Terblanche, the Key Account Manager of BATSA, Mr Piet Botha, a former Director of Trade Marketing and Distribution for BATSA who unsuccessfully tried to enter the cigarette market with a super premium brand called Matrix; Mr Ari Zelezniak, a vending machine operator who controls Vendomatic and a web of other companies in this distribution channel; Mr Richard Cuthbertson, an expert in customer strategy, loyalty marketing and consumer insight; and Ms Helen Jenkins, an Economist at Oxera. JTI called Mr Dean Doyle, the National Trade Marketing Manager of JTI, Mr Oscar Makola, the Portfolio Strategy manager of JTI’s parent company in Geneva; Mr Allan Denby, a former SPAR owner in Stellenbosch; Mr Mlamli Papu, a former Area Sales Manager of BATSA who is currently working at JTI, Mr Siemon Scammel Katz, a marketing expert who specialises in shopper research; Mr Andre Van Vuuren, a category management expert; and Mr. Simon Baker, an Economist with RBB Economics. BATSA called Mr Gerhard Ackerman from Shoprite, Stephen Van Vuuren from Pick n Pay Western Cape, Mr Eric Hatz from Pick n Pay Hyper, Mr Martin Potgieter, the Head of Trade for BAT in the UK, Mr Ralf Wittenberg, the General Manager of BAT, formerly the Marketing Director of BATSA; Mr Dale Parker, the operations manager of Ignite, a HORECA venue; Mr. Alan Van Der Westhuizen from News Cafe; Ms. Sally Talbot, the owner of a Sasol Forecourt in East London; Mr. Robert Brink of Dros Stellenbosch; Mr. Richard Goodwin of Spar in Centurion, Mr. “Chico” Chikki of Ninos in Rosebank; Mr. Tremayne Doffay of JB’s Corner in Melrose Arch; Mr Boshielo from Xai Xai restaurant in Melville, Mr. Le Beau Taljaard from the Smollens Group; and Dr Nicola Theron, an economist from Econex.
Key features of the cigarette market
The cigarette market exhibits some unique features that are central to the matter before us. For ease of exposition we will, at the outset, identify the most pertinent of these characteristics.
This case is concerned with pre-rolled cigarettes but not with other general tobacco products such as loose tobacco, pipe tobacco and snuff. Cigarettes are branded consumer products with the various brands marketed in one of four categories. These are the prestige or super premium category which accounts for approximately 1,1% of all cigarettes sold in South Africa; the premium category which accounts for some 8,5% of cigarette sales; the popular category which accounts for 78,6% of cigarette sales; and the value for money (“VFM”) category that accounts for approximately 11% of cigarette sales.10
Each brand is typically ‘extended’ into a number of ‘stock keeping units’ (‘SKUs’) which are predominantly distinguished by strength (e.g. full flavour, mild, extra mild, etc), taste (e.g. plain or menthol) or by the presence or absence of a filter. So, for example, BATSA’s brand portfolio consists of some 27 different brands, comprising 142 different SKUs, while JTI markets 5 brands and 27 SKUs in South Africa and Philip Morris International (‘PMI’), a US based tobacco multinational, only markets its famous global brand, Marlboro, in South Africa, although the single brand comprises some 3 SKUs.11
While there appears to be some relationship between, on the one hand, the ‘intrinsic’ and ‘extrinsic’ qualities of a particular brand – that is, respectively the quality of the product itself and the quality of the packaging – and, on the other, the category to which it is assigned, these considerations are clearly not the most important bases for categorisation. Hence it is not unknown for a brand owner to shift a cigarette brand’s categorisation without any accompanying change in the intrinsic or extrinsic qualities of the product. For example, Camel, a JTI-owned brand which is at the centre of this dispute, appears, at various stages, to have been marketed as either a premium or popular brand with no suggestion that there had been a predictable accompanying change in the intrinsic or extrinsic qualities of the product.12
It appears then that, for the most part, a marketing decision of the cigarette manufacturer determines the category in which a given brand is placed. The manufacturer then devises a campaign which will seek to identify the brand with a particular demographic and lifestyle. While this decision regarding categorisation may well exercise some degree of influence on the packaging and, possibly, certain of the intrinsic features of the product, there is no clearly evident relationship between these factors and the category in which the brand is placed. The building of a brand and its placement in one or other of the categories listed above appear to be mostly concerned with the construction, through various mechanisms of promotion, of a desired image and, therefore, association with a class of consumers.
However the decision as to the category in which a given brand is to be marketed determines the price band into which it is placed. Indeed price is the most reliable signifier of the category in which a brand is placed. While the pricing of each category is located within a band rather than at a fixed point and there is clearly ‘room’ for price competition within these bands, it was never suggested that a premium brand would compete for a sustained period within the popular price band or a popular brand within a VFM price band. Indeed, one of JTI’s key grievances is precisely that Camel, at various times categorised as either a premium or popular brand, is physically displayed at the point of sale alongside the VFM brands where its higher popular or premium price causes it to appear relatively expensive. This physical positioning in the cigarette display units is, alleges JTI, a direct outcome of BATSA’s trade marketing strategies and is precisely intended to create this confusion.
This having been said, additional and peculiar factors influence the pricing of cigarettes. In particular taxation plays a very significant role in influencing absolute price levels and movements in the price of cigarettes. This is evidenced by a corresponding increase in the price of cigarettes when excise duties go up.13
Although brand development and promotion and other aspects of marketing are extremely costly activities, the prices commanded by the brands in the higher categories generate the highest margins for the brand owner and others in the value chain, namely wholesalers and retailers, notwithstanding the considerable cost of brand promotion. Thus encouraging consumers to ‘trade up’ – or dissuading them from ‘trading down’ – is an important consideration in the marketing of cigarettes.
It is generally accepted – including by all the parties involved in this matter – that cigarette smokers are unusually loyal to their chosen brand, although there is considerable debate, between the parties, over the precise extent of this loyalty. Recognition of high degrees of brand loyalty dictates the importance attached by the marketing and merchandising departments of a cigarette manufacturer to a new smoker’s initial choice of brand and to that category of smokers – essentially young smokers – most susceptible to switching their brand in response to marketing and promotional initiatives. Hence the focus of marketing activity on adult smokers under 30 years old (‘ASU30s’) or under 24 years old (‘ASU24s’). Certainly intense brand loyalty dictates that the market shares of cigarette brands change only slowly and incrementally – indeed it is common cause that it is not possible to track short term responses to marketing initiatives. Indeed, for many – and this, as we shall see, includes many sophisticated and experienced retailers – the unusual strength of the loyalty factor underpins a belief that merchandising and promotional initiatives will have little impact on the fortunes of any particular brand.
The strict separation that has hitherto existed between, on the one hand, the distribution of cigarettes and, on the other, the marketing and brand promotion functions, also bears mention. A well established network of independent wholesalers physically distributes the products of all manufacturers to the various retail outlets. Although, as is implied by the wholesale model of distribution, the wholesalers take title of the products that they distribute, that is they purchase cigarettes from the manufacturers and on-sell them to the retailers, the manufacturer re-enters the value chain at the retail level by assuming responsibility for the full range of retail marketing and merchandising functions, including brand building and development. This involves all advertising and sponsorship activity, direct marketing initiatives, that is activities which entail a direct interface between the manufacturer and the end-consumer, and promotional activity at the point of sale, that is, promotional activity in the premises of the retailer. This latter function naturally requires close interaction between the manufacturer and the retailer. It is promotion at the POS and, in particular, the nature of the interface between the manufacturer and the retailer that is central to this matter and which brings us to arguably the most particular and, for present purposes, pertinent feature of the cigarette market.
As is well known, the marketing, sale and, indeed, the consumption of cigarettes is, largely at the instance of health advocates, highly regulated. The relevant statute is the Tobacco Products Control Amendment Act 12 of 1999 which came into effect in October 2000. It amended an Act passed in 1993 that prohibited smoking on public transport and that required health warnings on cigarette packs and advertising material for the first time. The 1999 amendment took aim at marketing and promotion. It prohibited media advertising – print media, bill board advertising, radio, television and cinema advertising – and other forms of public sponsorship. The sale of cigarettes to persons under the age of 18 is prohibited. Smoking in places to which the public have access is highly regulated and limited. However, the regulations do not provide for an outright prohibition on the sale and consumption of cigarettes and so specific provision is made for the selling of cigarettes in retail outlets. As with the general approach to the trade in cigarettes, many aspects of the retail point of sale are also strictly regulated.
The relevant regulations imposed upon the cigarette point of sale in retail outlets principally concern display and advertising signage. The amount of signage is limited as to its size and the maximum distance from the actual point of sale itself. It appears that advertising signage at the point of sale may not exceed one square metre in size and must be placed within one metre of the point of sale. Although the regulations do not appear to prohibit the sale of cigarettes from the open shelves of the retailer, it seems that the danger of pilferage of these products dictates that cigarettes are always sold from a kiosk or from the till effectively obliging the customer to engage with an attendant when purchasing cigarettes. As we shall see this practice assumes significance in this case. In similar vein, while we are not aware of specific restrictions on the amount of space that may be devoted to the sale of cigarettes, we presume that this is self-regulating to the extent that the retailer would not want the cigarette POS to impinge unduly on the space available for the promotion and stocking of other items.
It was not always thus. In recent memory cigarette advertising arguably constituted amongst the most prominent media advertising. Figures such as the strong, silent ‘camel man’ and the glamorous couples on alpine ski slopes assumed iconic proportions as did a variety of advertising copy – ‘your passport to international smoking pleasure’ or ‘for after action satisfaction’. Major sporting events – the Rothman’s Durban July – were inextricably associated with their cigarette brand sponsors. The new restrictive regulatory regime outlined above was the product of intensive lobbying by health activists. It heralded the coming of the ‘dark market’ which, as we have just briefly outlined, significantly reduced the opportunities for cigarette brand promotion. Indeed, contend the applicants, it reduced the opportunity for brand promotion to the point of sale in retail outlets and to private events held at places of entertainment. It is alleged that this limited promotional space has been appropriated by BATSA thereby constraining the growth prospects of the brands of its rival manufacturers in the retail market for the sale of cigarettes.
The retailing of cigarettes takes place through a well established set of channels. These will be examined at length but, in order to complete this preliminary description of the institutions and practices that characterise the cigarette market, it is as well to note that the major channels comprise, firstly, ‘organised retail’ which include the large national grocery chains (specifically Pick n Pay and Shoprite/Checkers), the large national convenience store chains (the Spar group, OK Franchises, Seven Eleven and others) and the national chains of garage forecourt stores which are either franchised or directly owned and operated by the large oil companies including Shell, Caltex, BP, Sasol and Engen.
Secondly, cigarettes are sold through the myriad of convenience stores which do not belong to one or other of the national chains. These are referred to as ‘independent convenience’ and include a number of unaffiliated garage forecourt stores, a number of large stores serving diverse customer demographics, as well as the ubiquitous South African ‘cafes’ and spaza shops.
Finally, there is ‘HORECA’, which in other national markets is an abbreviation for Hotels, Restaurants and Cafes, although it appears that here the ‘ca’ refers to ‘captive audiences’, so as not to confuse it with the traditional South African ‘café’ which would fall into the independent convenience channel. HORECA refers to a range of entertainment venues at which cigarettes are sold and, subject to the relevant regulations, consumed. These may range from top end bars and clubs to lower end entertainment establishments and includes recreation clubs, sports venues and the like. These retail channels, which, it appears, were largely considered to be sites of distribution, have, particularly since the advent of the dark market and the regulation of above the line advertising and sponsorship that this has entailed, become important sites of brand promotion. It is the relationship between these various channels and BATSA that is at the centre of the matter before us.
The relevant market
On the face of it the various parties are in agreement regarding the boundaries of the relevant market – they all appear to contend for a national relevant market in the sale of manufactured cigarettes. Some consideration has been given in the arguments of the contending parties to whether the geographic market may not be regional and to whether the category segments referred to above may not constitute distinct product markets. Ultimately these contentions have not been seriously pursued even by those who advanced them in the first place. BATSA’s suggestion that acceptance by its opponents of market segmentation – and, more particularly, ascribing analytical significance to these segments in the competition analysis - implies acceptance of segmented relevant markets is without merit.
BATSA has however raised a rather more trenchant objection to its opponents’ treatment of the relevant market. It insists that the applicant’s analysis effectively posits a market in promotional opportunities in the sale of cigarette brands and that it is here – in this market for promotional opportunities - that the alleged foreclosure must be established. It is acknowledged that the anti-competitive harm, if any, must be manifest in the retail market for the sale of cigarettes but the foreclosure must be demonstrated to occur in the market for promotional opportunities. This appears to suggest, to BATSA at any rate, that the case should properly be brought in terms of Section 8(d)(v) which prohibits a dominant firm from ‘buying-up a scarce supply of intermediate goods or resources required by a competitor.’
In our view while much is made of this, little hangs on it. All agree that harm to competition must be established and all agree that it is in the retail market for the sale of cigarettes that such harm must be demonstrated to occur. It is agreed that BATSA’s share of total cigarette sales in South Africa significantly exceeds the threshold for ascribing dominance as provided for in Section 7 of the Act. The conduct whereby BATSA is alleged to cause anti-competitive harm largely takes the form of agreements with retailers – lubricated by the payment, in kind and in cash, of incentives to these retailers – that give it preferential access to a range of promotional opportunities in retail outlets and HORECA which effectively forecloses access by its rivals to promotional opportunities. And all parties accept that the existence or otherwise of promotional opportunities other than those alleged to have been appropriated by BATSA has bearing on the outcome of this complaint.
All of the above is common cause. Our task is to determine whether the conduct complained of indeed conforms with that proscribed by the relevant provisions of the Act – whether it constitutes a vertical restrictive agreement, that is, an agreement between parties in a vertical relationship that causes a substantial lessening of competition without sufficient countervailing pro-competitive gains; whether it is an exclusionary act, that is an act that prevents competitors from entering and growing in the market and causes harm to competition; and whether it is conduct that induces a customer or supplier not to deal with a competitor and causes competitive harm. If these are the agreed elements of the enquiry the resolution of a formalistic dispute concerned with a ‘proper’ identification of the relevant market – which in this case essentially involves disentangling the alleged conduct in the market from the market itself – take us nowhere.
Market shares
It is however necessary to outline the positions enjoyed by the various cigarette manufacturers and their key brands in the retail market as well as the role of brand development and promotion that is at the centre of this dispute.
As already noted BATSA was formed out of the merger in 1999 of the UK based multinational, the British American Tobacco plc (BAT) and the South African based, Rothmans International. At the time of the merger the local subsidiary of BAT enjoyed a market share of approximately 10%14, based on several brands including Benson & Hedges, JPS and Embassy.15 While BAT was and remains a major international cigarette manufacturer its strength appears to be based on several strong national brands but its portfolio does not include a strong global brand.
At this time Rothmans was overwhelmingly the dominant cigarette manufacturer in South Africa with a market share of some 86% in the South African cigarette market.16 It owned a number of important brands including Rothmans and, Peter Stuyvesant, but none so important as Peter Stuyvesant which enjoyed a market share of approximately 40%. However, as with the rest of BAT’s brand portfolio, Peter Stuyvesant’s strength was – and has remained – local. The advertising claim notwithstanding, it was not an international brand, but rather an extremely strong national player.
Largely on the back of Rothmans’ dominant position, which in turn relied on the strength of the Peter Stuyvesant brand, the merged entity enjoyed a position in the South African market that few manufacturers could even aspire to in other national markets. Its share was approximately 96%. Other manufacturers present in the South African market followed, at great distance, in BATSA’s wake, with a total market share of less than 5%.
Japan Tobacco International entered the South African market in 1995. It had purchased the international tobacco interests of the US Company, RJ Reynolds, which included two significant international brands, Camel and Winston. Camel had previously been imported into the South African market. Conspicuously absent from the South African market was Marlboro, owned by the US multinational, Philip Morris International (PMI), and the most valuable cigarette brand in the world, which enjoyed a position in the giant US market somewhat akin to that enjoyed by Peter Stuyvesant in the South African market. PMI, through Marlboro, only entered the South African market in 2004.
Brand promotion in a dark market
It is brands, not manufacturers that attract consumers. Indeed most smokers would be hard pressed to identify the manufacturers of their brands. But it is the manufacturers who develop and promote their various brands. It appears that Rothmans’ promotion of its brands relied heavily upon above-the-line advertising. Rothmans does not seem to have engaged in active brand promotion at the point of sale which it seems to have viewed purely as a site of retail distribution rather than brand promotion. Our impression is that JTI, largely through the medium of its Camel brand, competed with BATSA on the same above-the-line advertising terrain.
This underwent a fundamental transformation for two inter-related reasons. The first was a major transformation of the domestic regulatory environment essentially aimed at significantly reducing cigarette consumption partly by restricting – indeed eliminating – the ability of cigarette companies to promote their brands through the medium of above-the-line advertising. This is described above.
The second was the presence, through the merger with Rothmans, of a powerful international manufacturer, BAT, which was already familiar with operating in dark market conditions elsewhere, which actively contemplated the advent of the dark market in South Africa, and which accordingly planned to redeploy its promotional resources from above the line advertising to active merchandising at the point of sale.17 JTI’s counsel, on the basis of a detailed description of the strategic advice received by BATSA and the action taken, excoriates BATSA for its aggressive and allegedly anti-competitive conduct.18
However this is by no means the only possible reading of the documents cited. To be sure, the language is aggressive, filled with the bellicose terminology favoured by strategy consultants. And, of more interest, the action contemplated – and, as we shall see, for the most part taken – is extremely robust. However the description of the strategy process and its content may well be read as an outstanding example of a firm examining the environment in a national market of considerable importance to its fortunes, taking note of the imminent, radical transformation of the regulatory environment surrounding that market, and making its plans, critically including the deployment of its brand promotion resources, accordingly.19 In apparent support of its various allegations, JTI cites a key JTI training manual explaining that the new regulatory environment ‘has prevented BATSA from communicating to the consumer in the traditional manner’ and, consequently, that BATSA has been ‘compelled to communicate with the consumer using new and innovative methods.’20 But what, on the face of it, could be more pro-competitive than responding to a changed environment by seeking out ‘new and innovative methods’ with which to communicate with one’s customers.
With the merger of Rothmans and BAT, BATSA’s obvious strategic imperative was to defend its newly acquired massive market share and that this should be clearly acknowledged in its strategy documents is unremarkable. However aggressive language and massive market shares notwithstanding, we must remind ourselves that these objectives and their aggressive pursuit are by no means necessarily anti-competitive even when pursued by a dominant firm, which, although uniquely subject to the Act’s abuse of dominance provisions, is as entitled, is indeed expected, to compete as vigorously to increase or defend its market share as any of its aspirant smaller rivals. JTI’s counsel appears to argue that having identified what it believes to be an anti-competitive strategy – but may simply be a strategy directed at besting its competitors - the onus should then shift to BATSA to prove that it did not engage in anti-competitive conduct. Even if this rather unusual proposition had found favour with us – and it does not – we would not have found the strategic direction apparently favoured by BATSA to be necessarily anti-competitive.
Indeed what the documents certainly confirm – and this is well captured in the lengthy extracts quoted in the JTI heads of argument – is BATSA’s close appreciation of the added significance that a dark market accords to the promotional function of the retail channels. Under dark market circumstances there is clearly no warrant, if indeed there ever was, for continuing to treat the POS as merely a distribution site – it comes into play as an important site of promotion. It is recognition of the centrality of the POS as a promotional site – which, as we shall outline below, we find to be significantly overstated - that resonates through the BATSA strategy documents and it is the conduct arising from this strategy that is so roundly condemned by the Commission and JTI.
As we shall demonstrate, this unremarkable insight regarding the elevated importance of the POS under dark market conditions has not escaped JTI’s attention either, also a major multinational with the range of experience possessed by BATSA. As we will elaborate below, certainly in those retail channels which JTI viewed as central to its brand building efforts – notably the organised forecourts and HORECA – the evidence shows that it invested heavily in securing its preferred allocations of space and position in the CDUs and in gaining access to preferred HORECA venues for the holding of promotional events. However in respect of a number of retail channels JTI appears to have adopted a relatively supine approach to the brand building role of the POS. As already noted it has rather chosen to insist, with the support of the Commission, that in these channels and with respect to certain promotional conduct, rules or ‘principles’ derived from a marketing concept, referred to as ‘category management’, must be applied. It has sought to establish – not, in our estimation, very persuasively – that these category management principles embody competition objectives and rules insofar as they ensure, largely through conveying ‘objectively correct’ information, an improved ‘shopping experience’ for consumers. Proceeding from this rather fuzzy proposition, they then make a truly heroic conceptual leap when they appear to argue that failure to apply category management principles is equivalent to a contravention of competition law.
At no level is this contention easy to defend at competition law. The central proposition advanced effectively argues against suppliers competing for space and position at the POS precisely because the competitive process will result in the allocation of these promotional opportunities being determined by the commercially driven preferences of those who prevail in the competition rather than on ‘objective’ historical data and allegedly widely accepted merchandising ‘principles’. Simply to state the proposition is to recognise how distant it is from fundamental principles of competition law and economics.
The case is presented at two levels: a theory of harm is developed, and then evidence is adduced to support the theory argued for. The former – the theory of harm rather than the evidence – has assumed pride of place in the applicants’ presentation of its case.
Theoretical propositions loom so large in this matter because, firstly, the central approach of the Commission and JTI has been to contrast BATSA’s alleged conduct at the POS with an idealised set of principles – ‘category management’ – on the basis that the latter principles purport to advance consumer interests by basing POS shelf space allocation on historical sales data and shelf positioning on established consumer preferences and retail practices. We are effectively being asked to void actual practices widely employed by highly sophisticated, profit maximising South African retailers on the say-so of marketing experts who insist that these retailers have been induced or incentivised by a single supplier to depart from efficient merchandising practices, in the process undermining their own profitability and the best interests of their customers and of many of their suppliers. It is little wonder then that JTI contends that
“…the sheer extent of the coverage of BATSA’s incentive schemes, along with the extent of BATSA’s dominance, suggests, by way of theoretical inference, that significant foreclosure would be inevitable, and that the cost to its rivals of meeting the exclusionary effect would be prohibitive.”21
Secondly, these contrasting theoretical propositions – effectively one predicated on hard-nosed competition for retail space and position, the other on a set of merchandising propositions regarding the optimal organisation of space and positioning of brands in a category from the perspective of the effective management of that category – are relied upon because the evidence submitted by the Commission and JTI is so sparse. The sheer volume of the record and the inordinate length of the hearings may have served to obscure this point despite it being so loudly and frequently insisted upon by BATSA. But a reading of the lengthy witness statements and the transcript bears out BATSA’s contention.
BATSA has been guilty of somewhat overstating its case regarding the type and level of detail of evidence required. We are in sympathy with the view expressed by the court in Conwood to the effect that it is impossible in a case of this kind to verify claims and counter-claims on a store-by-store basis because, given the number of retail establishments involved, this would ‘have been so costly as to have effectively ended this suit, despite substantial evidence of anti-competitive activity.22 We are also at one with the standard articulated in Lorain Journal – and cited approvingly in Conwood – that in an anti-trust case ‘the trier of the facts may make a just and reasonable estimate …based on the relevant data and may act upon probable and inferential …proof’.23
However these admirably pragmatic approaches to evidence cannot disguise the reality that the case before us essentially concerns allegations of actual practices in a range of retail outlets. These are practices allegedly devised by predatory manufacturers and implemented with the active connivance of a range of retailers who stand accused of violating fundamental marketing principles, and so compromising their customers (and ultimately their own) interests, in exchange for free merchandising architecture and cash payments. It is alleged – as for present purposes it must be – that the welfare of their customers is thereby undermined, that their ‘shopping experience’ is compromised by the limitations that these arrangements have imposed on competition in the retail market for cigarettes. And yet the applicants have chosen to rest their case on oral evidence from several manufacturers or erstwhile manufacturers (that is, competitors of BATSA), on several marketing and economic experts and, at the instance of JTI, on a single, decidedly minor player in the retail trade. It is little wonder then that the applicants have chosen to rely almost entirely on inferences drawn from theoretical arguments. BATSA, on the other hand, has introduced a wide range of testimony from witnesses in the retail trade.
We will examine the allegations of anti-competitive conduct along four dimensions. Firstly, we will examine the theory of harm advanced by the applicants. This effectively seeks to contrast the competitive conduct allegedly employed by BATSA and the category management principles favoured by the Commission and JTI and their marketing experts. We will examine the general claim that adherence to category management principles effectively be treated as a proxy for pro-competitive conduct, or, at least, that contravention of category management principles be regarded as evidence of anti-competitive conduct. In particular we will examine the contention that the payment of incentives to secure additional space and a preferential position on the cigarette dispensing units, or, indeed, on retail shelves in general, is indeed anti-competitive. This is effectively the applicants’ theory of harm. In assessing this theory of harm, we will largely accept the factual averments relied upon by the applicants. That is, we will accept that retailers were paid incentives in cash and kind in order to secure an allocation of space for BATSA brands on the CDU that accorded with BATSA’s national market share and that its rival brands were assigned the bottom left hand slots on the CDU and that these requirements were adhered to by the retailers.
While our view on the legitimacy, from a competition perspective, of the payment of incentives of this kind, holds good for the payment of incentives to the VMOs and the HORECA venue owners, we will generally examine these separately from the other channels, in substantial part because category management principles have not been asserted as strongly in this channel.
Secondly, we will examine the factual evidence on foreclosure and the coverage of the arrangements that brought about the alleged foreclosure. Relying upon the Tribunal’s decision in South African Airways, the applicants allege that BATSA’s conduct has given rise to significant foreclosure.
Although, as will be elaborated at length, we do not accept the applicant’s contention that significant foreclosure of promotional opportunities has occurred as a result of BATSA’s conduct, we will nevertheless briefly examine certain of the origins of the difficulties that BATSA’s rivals have in growing their market shares. We will show that the reasonable probability is that their travails are attributable to the regulatory interventions rather than to BATSA’s conduct.
Finally, we will investigate whether the conduct complained of has given rise to cognisable efficiency gains.
Category management and competition
JTI avers that category management ‘involves deciding on which products within the category should be stocked’, the purpose being ‘to maximise the sales of the category within a retail outlet for the benefit of the retailer and suppliers whose products are stocked within the category.’24
For its part, the Commission’s heads define category management as the ‘effective use of a retailer’s shelves to maximise profit to the retailer and value offered to the customer….to maximise retailer’s sales from the category as a whole, relative to other retailers, rather than to advance the cause of any particular brand or supplier’.25
Note that common to these definitions is the assertion that category management consists in the organisation of retailing shelf space for the benefit of the retailer – indeed the Commission boldly asserts that its effect will be to ‘maximise profit to the retailer’. JTI also asserts the interest of other suppliers while the Commission makes mention of the ‘value’ that accrues to ‘the customer’ in consequence of the ‘effective use of a retailer’s shelves’. The claimed retailer benefit will be interrogated at length. However, we will show that the evidence is clearly that the retailer’s own conception of profit maximisation in relation to shelf space is focused on the sale of that space, rather than on the manner in which brands are displayed on the shelf. This is not without qualification: the retailers are indeed keen to ensure that their customers’ range of preferred brands is available and that no ‘out of stock’ situations arise. However, these qualifications notwithstanding, there can be little doubt that the retailers themselves view the sale of space as the major contribution that their shelving offers to their bottom line.
Although of relatively recent provenance, the concept of category management is extensively reviewed in the marketing literature, which would generally associate it with the following practices and principles:
One of the manufacturers, usually, although not inevitably, one of the larger manufacturers, is appointed ‘category captain’ with responsibility for the ‘management’ of the category.
The category manager’s task is to promote the category as opposed to merely promoting its own brands. This it will do by ensuring that the retail display ‘communicates’ effectively with the customer by conveying accurate and accessible information. This will enhance the consumers’ ‘shopping experience’ which will redound to the benefit of the satisfied customer, to the retailer whose store is the source of his satisfaction and to all the manufacturers who supply the well managed category.
While category management incorporates a number of functions at the point of sale – merely ensuring the orderliness of the display being an important one – the critical functions are the allocation of shelf space and position amongst the various competing brands within the category.
Where the allocation of shelf space is concerned it appears that the governing principle to which the category manager is expected to adhere is the rate of sale of each brand – indeed each SKU - within the category. This will not only minimise ‘out of stocks’, a circumstance that the retailer is intent upon avoiding, but will also represent a ‘fair’ allocation of space. The rate of sale used is expected to approximate as closely as possible to that prevailing in the individual store and while the necessary data for store-based space allocation may not always be available, the more finely disaggregated the data, the better from a space allocation perspective. That is to say store-level data is preferred to regional data and regional data is preferred to national data. The most accurate and disaggregated data is nowadays procured from the retailers’ scanners when available – ‘electronic point of sale’ or ‘epos’ data – and then sold by the retailer to the category captain and the other manufacturers who wish to participate in the space allocation function.
While the principles governing the allocation of shelf position are not as clear as those governing space, ‘principles’ mentioned include the notion that premium brands (that is, the most expensive brands) be positioned in the higher shelves with ‘popular’ brands in the middle and bargain brands at the bottom. A generally accepted viewpoint is that the preferred position for promoting sales is in the middle of the display, generally referred to as the ‘hot spot’, because this is where the eye of the average shopper – by many accounts a highly unobservant species – tends to alight.
Because it appears to be recognised that the category captain may be beset by conflicting incentives – of which more anon – it was stressed that in performing its duties, in particular drawing up ‘planograms’ which are essentially maps of space and position, the category captain should be obliged to canvass the views of the ‘partner’ manufacturers (that is, those suppliers who are not assigned category captaincy).
The retailer owns the space in which the space and position is being allocated and retains, at all times, the final word on the exercise of all of the category captain’s functions including the role assigned those suppliers who are not responsible for managing the category.
Finally, there was much equivocation around the critical question of whether it is appropriate, and indeed whether or not it is established practice, for the manufacturer to pay the retailer a fee for the privilege of undertaking the category captaincy function.26 This is, of course, a revealing way of posing the issue because all acknowledge that category captaincy is a time consuming and costly function and yet it was never suggested that the retailer pay the category captain a fee for performing this service despite the benefits that allegedly accrue to the retailer. In our view – and contrary to the ‘principles’ outlined above -this clearly reveals the implicit understanding that category captaincy, properly construed, is intended to privilege the brands of the category captain who, after all, will, at the very least, be bearing the burden of providing a costly service. The category captain benefits – as do all other manufacturers – from a well-managed category and, in the rarefied world of high category management principles, this is presumably thought to constitute payment enough. What is clear, as noted above, is that it is established practice for the category captain, and, to varying degrees, the other manufacturers, to pay the retailers sometimes considerable sums for the EPOS data which is the empirical basis – in principle at least - for space allocation.
While it is not clear that BATSA has ever been formally appointed to the category captaincy by any of the retail outlets, all agree that it exercises the key elements of that function and this is the basis on which we shall approach our assessment of BATSA’s conduct. BATSA purchases the EPOS date from the retailer. It enters into agreements with the various retail channels that purport to regulate the display on that unit and at the point of sale in which the unit is located, including, of course, the display of its rivals’ products and advertising material. These agreements provide for the payment of a range of cash incentives in exchange for the retailers’ adherence to the terms of the agreements, including those governing space and positioning. BATSA also provides free merchandising furniture to a significant number of retailers. These are subject to separate agreements.
Assuming with JTI and the Commission that BATSA is, at least, the de facto category captain in the retail channels (we are not considering HORECA for the time being) it is alleged that BATSA, in performing this role, deviates from accepted category management principles and practices in the following manner:
The general allegation levelled at BATSA is that, assuming the guise of category captain, it promotes its own brands rather than the category as a whole. It allegedly achieves this through payments, in cash and kind, which induce the retailers to give BATSA preferential space allocation and positioning. It also allegedly seeks to dominate secondary displays at the point of sale.
While there is a great deal of factual dispute regarding actual space allocation, the Commission and JTI allege that BATSA’s avowed objective is to achieve a space allocation equivalent to its national market share regardless of the shares that it actually achieves in the various channels, regions and stores. This objective is said to conflict with category management principles and best practice which aims to achieve a space allocation that reflects the greatest level of disaggregation of sales data as possible.
BATSA is alleged to abuse its role as category captain, and, in particular, the planogramming function, by its insistence on positioning opposition brands in the bottom left hand corner of the CDU. This not only prejudices non-BATSA brands with respect to visibility but it also imparts confusing and inaccurate information insofar as it locates non-BATSA premium and popular brands alongside VFM brands. This not only sends confusing signals about the quality of the non-BATSA brands but it emphasises the price differential between the key BATSA opposition popular and premium brands and the immediately proximate VFM brands.
In the performance of its category captaincy functions, BATSA does not consult with other participants in the category. Moreover, the retailers have abdicated responsibility for the management of the cigarette category to BATSA, that is, they fail to exercise the oversight role that is a critical principle of effective category management.
Finally BATSA secures these skewed space and positioning outcomes against the payment of incentives in both cash and kind, including the provision of free merchandising furniture. It is suggested that the payment of incentives is in conflict with category management principles.
This conflict between the principles of category management and the practices in which BATSA is alleged to engage do not strike us as surprising. There are clearly powerful and strikingly obvious incentives that pose severe challenges to the high minded principles of category management. In a nutshell: in an oligopolistically structured market for the manufactured product, the category captain is more likely than not to use category captaincy as the organising device for a collusive arrangement between the manufacturers and in an oligopolistically structured retail market it may provide similar opportunities for collusion between the retailers. On the other hand, in a market dominated by a single firm, the dominant firm, which is likely to be the firm tasked with responsibility for undertaking category captaincy, will be well positioned to use its category captaincy in order to defend its dominant interest. This is, of course, precisely what BATSA stands accused of.
Ironically – or perhaps predictably – while the Commission and JTI insist that category management principles be applied, JTI has emphatically not offered to take on the role of category captain itself. Indeed the evidence is that JTI has until as recently as 2007, in both Pick n Pay and Shoprite Checkers, refused to purchase the EPOS data, the minimum requirement that the retailer demands for the right of suppliers to participate in category management.
JTI effectively argues that, given its low market share, were it to assume the role of category captain – or even devote resources to participating in category management - it would be obliged to spread the costs of its participation over too small a revenue base. Alternatively expressed, JTI argues that the category advantage that will accrue to BATSA as a result of JTI undertaking category management would – the more so if strict category management principles were adhered to - dwarf the category advantage that accrues to JTI, that is, that BATSA will, by virtue of its dominant share of sales, receive an excessive free ride. In other words, BATSA’s volume of sales and market share would ensure that its brands would dominate the space allocated on the CDU and that it would secure desirable positions regardless of the identity of the category captain thus receiving a substantial free ride on the back of a smaller rival’s category captaincy.
In short it appears that JTI believes that BATSA should be willing to provide its rivals with the relatively ‘small’ free ride that will accrue to players with a small market share should BATSA adhere to pure category management principles. But why should BATSA provide any free service to its rivals? We are at one with the Courts dictum in Olympia Equipment Leasing Cc. V Western Union Tel. Co:
‘It is clear that a firm with lawful monopoly power has no general duty to help its competitors, whether by holding a price umbrella over their heads or by otherwise pulling its competitive punches…Advertising a competitor’s product free of charge is not a form of cooperation commonly found in competitive markets; it is the antithesis of competition’.27
It would be surprising if even JTI or its Counsel disagreed with this unimpeachable viewpoint. The truth is that JTI probably calculates that because BATSA has, by virtue of its dominant market share, such a major interest in the development of the category that it will continue to provide category management even were we to order it to adhere to ‘pure’ category management principles thus providing its rivals with a legally sanctioned free ride. That is, it will continue to provide category management even if we were to prevent it from appropriating any private return from performing the role of category captain. The category will continue to be neatly displayed, the cigarette display unit will be unsullied by the presence of non-tobacco products, effective display units will be provided, and these will be plannogrammed in terms of category management principles which will ensure that BATSA rival brands will be accorded the shelf space allocation and positioning prescribed by category management. All of this will be the responsibility of the category captain.
While JTI would have us order that BATSA adhere to category management principles, it appears that the Commission would have us prohibit category management. At least the Commission would have us order that BATSA be prohibited from playing the role of category captain.
This raises the question of the retailer’s interest in the services of a category manager? Why do the retailers’ own merchandisers not manage the various categories themselves particularly in the face of clear incentives for the designated category manager to promote its own brands in conflict not only with the interests of its competitors but also contrary to the alleged interests of both the retailer itself and its customers? After all the strongly stated ‘principle’ of category management which holds that all participants in the category should be consulted by the category captain and that the retailer should supervise the activities of the category manager and retain ultimate decision making power over the management of the category, suggests that the existence and strength of the incentives for the category captain to utilise the role to capture a private return (particularly when the costs incurred by the category manager are acknowledged) is clearly appreciated by the proponents of category management. Why then would the retailer ‘put out’ the function of category management to such a deeply conflicted market participant?
We are told that the retailer’s interest in having one of its suppliers manage its product categories is to be found in the superior category knowledge possessed by the manufacturers relative to the retailers. This claim is asserted despite the fact that the category captain is obliged to rely on data generated by the retailer in order to carry out its key function, namely the allocation of space, particularly if the avowed principles of category management are to be applied to the decisions regarding space. However because retail outlets cater for literally thousands of categories of products, the retailer cannot be expected to understand detailed aspects pertinent to the marketing of each category. This detailed know-how is said to be the preserve of the manufacturers whose interest is, for the most part, confined to a small number of product categories, indeed in the case of most cigarette manufacturers to a single category. For example, included in the knowledge of the manufacturer would be an understanding of the various sub-categories that are present in the cigarette market and, indeed, in many other fast moving consumer good markets as well as the demographics of consumers who are drawn to one or other sub-category.
Although we understand that all the manufacturers may have an interest in a well managed category – at least to the extent that the category is neatly presented and receives the requisite space allocation and positioning relative to other product categories within the store – as competition authorities we remain acutely cognisant of each manufacturer’s interest in promoting sales of its own products in relation to those of its competitors’ products. Indeed even were we to accept the vague claims made for category management’s ability to promote consumer interests through the provision of accurate and accessible information, we would be hard pressed to privilege these claims above those that flow from vigorous competition between brands. And, as we shall elaborate at length, it is our firm view that robust competition for space and positioning at the POS is a critical element in the competitive interface between competing brands and, conversely, that the appeal to ‘pure’ category management principles is little other than an attempt to place an entire terrain of competition beyond the ambit of legitimate competitive conduct.
We turn then to the central issue of competition for the key offerings of the POS, namely the quantity and quality of space allocations, what we shall refer to as space and positioning. This issue is central, firstly, because, a principal form that BATSA’s much maligned incentive payments take is precisely that of a cash payment made by BATSA to the retailers for an agreed allocation of shelf space and position. Secondly, as already observed, it is noteworthy that from the perspective of the retailer the profit maximising potential of the shelf space clearly derives from the sale of the space rather than from its function as a category promoter.
Retailers are clearly willing to sell preferred positions on their shelves and, within reasonably clear limits, they are willing to permit space allocations to be utilised for promotional purposes. In the present case, the evidence is that the various retailers have, to varying degrees, permitted space allocations that exceed the rates of sale of the manufacturers, the brands and the various SKUs. These advantages have largely been procured by BATSA utilising the incentive payments which the complainants seek to impugn. We note – and this is discussed further below - that where a BATSA rival has been able to outbid BATSA, it too has secured space in excess of its market share.
The retailers have also permitted space to be allocated on the basis of manufacturers’ market shares rather than on the basis of the market shares of individual brands or SKUs, and have then permitted the manufacturers to distribute their brands and SKUs within their allotted space as they see fit. Again BATSA with its relatively large portfolio of brands and SKUs has been the principal beneficiary of this deviation from ‘pure’ category management principles and which accounts for the claimed anomaly of BATSA brands with very low market share sometimes occupying significant space and preferred positions. The intra-BATSA brand allocation has clearly been driven by promotional considerations. Lucky Strike, a BAT drive brand but with a very low South African market share, is the most frequently cited case in point, although Dunhill also appears to have benefitted from this practice.
Although retailers are clearly willing to sell preferential space and position on their shelves, the evidence does not suggest that any participant in the organised or independent retail channel has contemplated assigning their shelf space to BATSA on an exclusive basis or that BATSA itself has entertained this prospect. That is, no-one has contemplated preventing the physical distribution of non-BATSA brands through the retail outlets that sell cigarettes. Indeed our distinct impression is that the grocery and convenience trade does not generally offer exclusives to its suppliers. However we concur with BATSA – and with persuasive scholarly work on this question – that the practice of selling preferential allocations of space and position should be viewed as a form of limited exclusive. Indeed we concur with Klein and Wright’s contention, that the entire practice of category management be viewed as a form of limited exclusive:
‘Category management contracts, which shift control of the retailer’s shelf space within a product category to a manufacturer, are another form of limited exclusive, where the manufacturer determines the quantities of other, highly demanded products to be stocked.’28
We should clarify at the outset that treating category management – and particularly the function of category captaincy – in this way does not, of necessity, presuppose the conflation of distribution and promotion, although in practice it has frequently done so. Certainly, for the purpose of measuring the extent of foreclosure it would be erroneous to elide the distinction between distribution and promotion and we are cognisant that BATSA stands accused of foreclosing promotional opportunities while no such allegation is made with respect to distribution.
However, even when focusing on access to promotional opportunities, the pertinence of the existence of universal distribution should not simply be ignored. That the products of rival brands are widely available – that is, that they are actually freely distributed into and by the retail trade – and that they are generally known to be universally available at outlets where cigarettes are sold, is pertinent and is indeed, on its own, sufficient to characterise category management as a form of limited exclusive. That is to say, even if were to accept all of the applicants’ evidence regarding the limitations imposed on its promotional opportunities by its rival’s incentive payments, the mere fact that all but the proverbial Martian knows with a powerful degree of certainty that simply by asking the kiosk or till attendant she will receive her packet of Camel limits the impact of the degree of exclusivity over promotional opportunities. In other words widespread and longstanding practice has promoted the knowledge that all cigarette brands are available at all outlets at which cigarettes are sold. This widely known fact vitiates the promotional function of the POS display, it is a limitation on the impact of the exclusive – a blind person may confidently approach the cigarette POS and ask for his brand of choice, confident that it will be made available to him.29 Promotion, the exercise of influence over the brand chosen, clearly comes from elsewhere.
Of course the fundamental limitation on the degree of exclusivity resides in the fact that the category captain is actually obliged to place rival brands on its planograms, and that the final decisions regarding listing and the allocation of space and position belong to the retailer and not the category captain. In this regard our task is to determine the extent of foreclosure of promotional opportunities, but, we repeat, the extent of foreclosure from cigarette promotion, indeed the significance of the POS as a site of promotion, is significantly vitiated by the absence of any limitation on distribution and by the fact that cigarette consumers know that their brand, unless of a particularly exotic or recently introduced variety, will be carried by all retailers who sell cigarettes. Moreover, in our view the fact that the cigarette purchaser does not have to find her brand on the shelf, but has simply to ask an attendant to find it, enhances the importance of the lack of limitation on distribution and concomitantly reduces the importance of visibility and POS promotion in general. One of the least convincing submissions made to the Tribunal came from Mr. Scammel Katz, the JTI marketing expert, who attempted, without much apparent conviction, to persuade us that a customer who could not actually see her brand on display would be too ‘embarrassed’ to request it of the assistant.30 We can give no credence to this claim particularly, we repeat, because smokers know that retail outlets carry all brands.
These conclusions regarding the interplay between universal distribution and limited promotion may have to be qualified somewhat with respect to first time smokers. We will deal with this below.
Thus to the extent that category management provides exclusive promotional rights to a single manufacturer it is a limited exclusive, limited by universal distribution and by the very substance of the right which, in Klein and Wright’s words, permits the category captain to ‘determine the quantities of other, highly demanded products to be stocked’. In other words the category manager’s right is to determine the level of stocking of other manufacturers’ products, it is not to exclude them from being stocked.
In fact this right to determine stocking is, as we have already seen and will elaborate further, significantly more limited than is suggested by Klein and Wright. Certainly in this matter it has not been suggested that BATSA’s right to determine the stocking levels of its rivals is unfettered. The retailers’ anxiety to avoid out-of-stock situations is acknowledged and therefore stocking levels have to approximate the rate of sale regardless of any wish to the contrary that the category captain may entertain.31 Moreover all agree that the ultimate right to determine the quantities to be stocked belongs to the retailer. The complainants aver that the retailers have effectively abdicated this right. We will examine this claim below. However it is common cause that it is an ultimate right that no retailer has actually signed away.
But limited though this exclusive may be, we must now focus on the exclusive dimension itself rather than on the limitations. Because certainly no other manufacturer matches BATSA’s sway – whether de jure or de facto or both – over the POS. There are rights extended in the trade merchandising agreements between BATSA and important participants in the retail channels that belong to BATSA, and BATSA alone. In order to determine the extent of foreclosure we will, in due course, examine the mass of evidence dealing with, inter alia, the coverage of the agreements and the extent to which they were implemented. However for present purposes we will treat the agreements at face value: as the extension by the retailers of limited exclusive rights to BATSA in exchange for cash incentives paid against compliance. For present purposes – that of examining the theory of harm that derives from limited exclusives - we can assume full compliance and extensive coverage.
Why is it important to dwell on the exclusive element in the agreement? Because it is this that establishes the character and purpose of category management. The idealism – or is ‘disingenuous’ a more accurate characterisation? - of Mr. Scammel Katz, Dr. Cuthbertson and the other marketing experts notwithstanding, the notion that a profit maximising manufacturer assumes the role of category captain with all its attendant costs for the good of ‘the category’, the retailers, the consumers as well as all the manufacturers, including its competitors, simply does not hold water. We may accept for present purposes that all manufacturers derive some benefit from a growing category. Indeed, because of BATSA’s dominance and therefore the extent to which it, in particular, benefits from a well managed and growing category, the structure of the cigarette market may well be best suited to the category captain adopting category management principles.32 However when its market is under attack from brands of the strength of Camel and Marlboro, when it’s share has only one way to go and that is down, and when its major brand – Peter Stuyvesant – is mature and vulnerable to attack, then it is strongly incentivised to use category management for its truly intended purpose: to defend its market share, to advance its private interests.
We do not then accept the view that category management is properly regarded as a selfless promotion of the category, an activity that expresses a shared interest in which the returns accrue in unspecified shares to the retailers, the consumers and all the manufacturers with all the costs borne by the category captain. While the ordinary operation of markets clearly incur and distribute social and private costs and generate and distribute social and private benefits, the level of coordination that is required to share the costs and distribute the benefits that is implied by the category management ideal – and that would thus serve to incentivise one of the manufacturers to play the role of category captain - would require an explicitly bargained and managed outcome. It cannot be achieved by the operation of market forces alone.33 In the course of a discussion on remedies Mr. Baker clearly recognises these difficulties:
‘In framing the remedies some accommodation would need to be reached about bearing the burden of category management and sharing the benefits yes. One wouldn’t necessarily expect one party to pay all of the costs and then split the benefits. I mean even though most of the benefits accrue to BATSA, but that would be a discussion wouldn’t it as to how to accommodate fair category management and how to split the funding of that relative to the benefits accrued.’34
Can one possibly imagine the retailers entering negotiations of this sort and at this level of detail with their cigarette suppliers, not to mention their suppliers in all of the other ‘managed’ categories? Is it desirable, from a competition perspective that competitors be encouraged to co-operate with each other in this manner? The difficulties outlined by Mr. Baker are insurmountable and the proposition is risible. ‘Socialism in one country’ proved sufficiently elusive; ‘socialism in one grocery category’ is simply not worth the candle.
In the real world of profit maximisation, category management – and, in particular, the category captaincy role – is then better understood as the purchase by the category captain manufacturer of a limited exclusive designed precisely to generate private returns through preferred allocations of space and position at the POS for the brands of the purchaser vis a vis competitor brands. Each of the owners of the competitor brands - in this instance JTI and PMI – is also a potential purchaser of the limited exclusive.35 There is a well established scholarly literature and jurisprudence devoted to identifying the impact of exclusive dealing arrangements and this is the appropriate empirical context for examining the limited exclusives concluded between BATSA and the various retail channels, the arrangements which the marketing experts and practitioners dub ‘category management’.
The manufacturer’s objective in purchasing a limited exclusive agreement of the category management type is easily discernible and clearly outlined by Klein and Wright. In essence, where significant inter-retailer effects do not arise from displaying one brand more prominently than another, the retailer may have no interest in anything other than a competitively neutral display of the idealised category management variety. However because of the prospect of inter-manufacturer effects or inter-brand effects arising from the allocation of preferential space and position to a particular brand or manufacturer, the manufacturer has an interest in achieving a better allocation of space and position than the relatively disinterested retailer – disinterested because of the absence of inter-retailer effects – is prepared to concede.36 The purchase by the manufacturer of a preferred allocation of space and position – a limited exclusive – from the retailer is the mechanism for solving the incentive incompatibility that we have just described and that characterises the relationship between manufacturer and the retailer. The argument is clearly captured in the following passage from Klein and Wright:
‘Since consumers are not willing to pay for promotional shelf space, but it induces profitable incremental sales for manufacturers, manufacturers will generally want greater retail space than retailers are willing to supply on their own. Retailers deciding how much promotional shelf space to supply will not take account of the manufacturer’s profit margin on the incremental sales produced by the promotional shelf space. The incentive incompatibility should therefore be especially significant when the manufacturer supplies differentiated products where the wholesale price is much greater than its marginal cost of production…it is reasonable to assume that a retailer’s decision to prominently display a particular brand does not generate significant inter-retailers effects in the form of consumers switching to competing retailers because their desired brand is not prominently displayed. Therefore retailers will supply less than the jointly profitable amount of promotional shelf space’37
We have noted that no consideration appears to have been given – either by the retailers or by BATSA itself – to a full exclusive. Again there is a straightforward explanation for this:
‘In the common case when consumer demand for variety within the product category or for a particular brand is high, this analysis explains why the efficient shelf space contract involves a limited exclusive.’38
This accurately describes the cigarette market. Variety is clearly important for the consumers as is verified by the large number of brands with quite low market shares. In fact, the importance of maintaining variety and choice for its customers was confirmed by Mr. Ackerman of Shoprite Checkers. 39 And, considerations of variety aside, there are certain international brands possessed of a cachet – which is more often than not reflected in a respectable market share – that a retailer would not want to exclude from his shelves. Camel and Marlboro are very clear cases in point.
It is always possible to write a contract that ensures variety. To this end certain of the ‘principles’ of category management – to wit the requirement to consult other manufacturers and the overriding decision making power of the retailer – may be specified. But this is probably not necessary. If one appreciates that the assignation of category captaincy implicitly assumes that the category captain’s brand will receive preferential allocations of space and position but will not totally exclude rival brands, then the retailer can rely on the category captain’s desire to retain his captaincy as a means of ensuring that he respects the retailer’s desire to reserve a requisite degree of space and, at times, preferred positions for the opposition brands:
‘Category management contracts, like limited exclusives, therefore offer retailers the ability to commit its marginal consumers to manufacturers and increase the value of its shelf space when consumers value specific rival brands. The manufacturer, therefore, is assured of receiving the limited exclusive it has paid for and the retailer relies on the manufacturer’s brand name to assure manufacturer performance, that is, to assure that overall retailer profit is not reduced by the category captain’s decisions that make rival offerings restrictive’.40
BATSA’s trade investment agreements are the purchase of a limited exclusive which, to be sure, extends privileges – largely in the shape of preferential allocations of space and position – to its brands. However, our examination of the extent of foreclosure will demonstrate that the impact on the promotional opportunities available to ‘rival offerings’ has not thereby been made unduly ‘restrictive’, thus confirming our theoretical propositions.
How are the consumers’ interests implicated in this attempt to resolve the incentive incompatibility between manufacturer and retailer? We should underline that there is no onus on the respondent to prove consumer gain, but rather on the applicant to prove harm. Nevertheless, consumer gain can be inferred if category management is treated as a limited exclusive. Indeed in order to demonstrate consumer gain from a limited exclusive one does not have to rely on vague notions like the ‘more satisfying shopping experience’ which the marketing experts characterise as the gain deriving from the application of idealised category management principles.
The incredulity expressed by JTI’s counsel at the very thought of possible consumer benefit deriving from BATSA’s planograms misses the point entirely.41 The consumer gain does not derive from lowering the space allocation accorded to BATSA’s rival brands or from their inferior positioning in the CDUs. It rather derives from the pass through to the consumer of a portion of the incentive payment that BATSA had to make to achieve its limited exclusive. We were not presented with direct evidence of this. However there is US evidence indicating that since the period of the introduction of slotting payments and the like – including category management – the margins of US retailers have not increased, thus inferring that these payments have not gone to the retailers’ bottom lines but rather to their customers and thus represent a re-distribution from manufacturer to consumer.42
This is precisely what one would expect from retailers trading in a competitive market. We have been given no reason to believe that this does not accurately describe the South African grocery and convenience store market and so we infer that a portion of the considerable payments made by BATSA in cash and in kind filters down to consumers. This need not have taken the form of lower cigarette prices. Indeed were the retailers to believe that the price elasticity of demand for cigarettes is low – a reasonable belief we would contend – then they may well apply the cash incentives received from BATSA to other products sold in their stores. Be that as it may, the inference of consumer gain from the pass through of the incentive payments to competing retailers is a considerably more compelling proposition than the notion that a ‘more satisfying shopping experience’ will follow the application of ‘pure’ category management principles.
In summary then the joint profit maximising strategy employed by BATSA and the retailers does not rely on the application of category management principles espoused by the applicants and their marketing expert. As already elaborated, the applicants based their case for the imposition of ‘pure’ category management principles on the contribution that this would make to the growth of the category and the consequent benefits to be derived by the manufacturers and the retailers as well as the consumers. However, these arguments for category management marketing principles notwithstanding, the retailers and BATSA entered into a strategy intended to maximise joint profit maximisation through agreeing a form of a limited exclusive agreement that gave the manufacturer, BATSA, privileged, if not exclusive, access to the POS in exchange for the payment of cash incentives as well as free merchandising equipment.
This bears out the essential wisdom in the contention of Mr. Van Vuuren, a senior Pick n Pay manager, who testified before the Commission that ‘the best form of category management depends on the strategy of the retailer’, and serves to underline some of the dangers of seeking to impose idealised profit maximising strategies devised by marketing experts and economists on actual profit maximising enterprises However, it does not serve to immunise the chosen profit maximising strategy from anti-trust scrutiny. As with any exclusive agreement, a limited exclusive implies a degree of foreclosure, in this case foreclosure of promotional opportunities at the POS. We will examine the extent of foreclosure consequent upon the limited exclusive agreed between BATSA and the retailers in the following section of this decision. However the theory of harm advanced by the applicants – essentially predicated on the contravention of category management principles – does not survive scrutiny. Far more persuasive is the view of Klein and Wright in which category management is viewed as the purchase by the category manager of a limited exclusive, a purchase which may, to a lesser or greater degree, be injurious to the interests of one or another competitor but which maximises the joint profitability of both retailers and manufacturers.
However, before turning to an examination of the foreclosure evidence, we should dispose of an important issue, one that must largely be determined by an examination of the facts before us, but that serves to further elucidate the interplay between the manufacturer and the retailers in the construction of the limited exclusive. It concerns the issue of countervailing power.
Much of the applicants’ case is predicated on the strictly limited universe that the regulatory environment allows for cigarette promotion. We are told – and we will examine certain of the bases for this claim when we discuss foreclosure – that regulation has confined cigarette promotional opportunities, indeed marketing opportunities, to the retail POS. We are further told that BATSA has been able to impose its requirements upon the retailers because the latter lack the countervailing power necessary to resist BATSA. The basis for this claim appears to lie in BATSA’s huge market share which allegedly makes BATSA a must-have supplier for all cigarette retailers. That is to say there is no point in making a retail cigarette offering if the brands of the manufacturer that accounts for some 90% of retail sales are not available.
While the latter point is superficially compelling, a cursory glance at the argument immediately suggests that the issue is significantly more complex than that suggested by the applicants. If, as the applicants insist, cigarette marketing, and certainly brand promotion, possibilities have been reduced to the POS, then certainly the immediate implication is that the gatekeeper of the POS is placed in an enormously powerful position vis a vis those who allegedly rely on this facility to pursue their brand promotion programmes.43 Moreover, although BATSA brands may well be must-have products in a cigarette retailing outlet, cigarettes themselves are not must-have products in all retail channels. At least one large grocery chain, Woolworths, does not sell cigarettes at all and another, Pick n Pay, indicated that the possibility of excluding cigarettes from their product portfolio has been entertained at a high level in the grocery chain whose witness indicated that they only sold cigarettes because their major rivals did so.44 While this may indeed lend credence to one of the applicants’ other contentions, namely that a lack of commitment to the cigarette category caused the retailers to concede more decision making authority to the category captain than may ordinarily be the case, it certainly does not support the contention that BATSA enjoyed a significant bargaining advantage over allegedly dependent retailers, certainly those large retail chains belonging to the organised grocery and convenience channels.
While we acknowledge that those participants in organised grocery and convenience who sell cigarettes do so in order to ensure that their customers are provided with a full service, cigarettes are clearly a classic convenience purchase. That is to say, customers do not enter a large grocery or convenience chain store for the purpose of purchasing a pack of cigarettes and then elect to make their monthly or weekly grocery purchases as an incident to purchasing cigarettes. The converse is far more likely to be true – while purchasing his groceries, the customer may decide to pick up a pack or carton of cigarettes. However, in general we would expect to find that the regular smoker who purchases cigarettes more frequently than he purchases groceries would purchase his cigarettes at an independent convenience store or a garage forecourt. These remarks are of course particularly apposite to young adult smokers, the agreed principal target of cigarette promotions, who, it is fair to generalise, represent a far smaller share of sales in organised grocery and convenience than their more aged counterparts. The data, which will be considered in our discussion of foreclosure, bears out these contentions.
Why is it important to correct the applicants’ erroneous claims regarding countervailing power? Firstly, it goes to the underlying contention of the applicants’ case which asserts that an all powerful monopolist supplier is capable of imposing its will upon relatively weak and fragmented retailers. This is clearly not the case. The retail buyers who gave oral evidence at the hearings and those who provided sworn testimony to the Commission did not exactly strike us as wilting violets. Far from it. Many of them are probably more accurately described as bullying tyrants who thoroughly dominated the unfortunate sales personnel of their suppliers, including their cigarette suppliers. And so, while as noted, we are prepared to concede a certain lack of attention on the part of the retail management to the cigarette category, the claim that the large organised retailers would have been prepared, much less obliged, to accept, at BATSA’s behest, an arrangement that significantly undermined category sales in order to privilege BATSA brands is thoroughly unpersuasive.
Secondly, the fact that countervailing power is possessed by the retailers, lends weight to an important and disputed element of BATSA’s case and that is its contention that it did not enforce its limited exclusive in an inflexible manner. BATSA has persistently claimed that as long as the space allocations and, particularly, the allocation of position, approximated its agreements with the retailers and its stated preferences, it would continue marking stores compliant. Where individual stores deviated, these deviations would either have been tolerated on good cause – as in the decision of Pick n Pay’s Cape Town Waterfront’s store to foreground Marlboro – or it would have attempted to secure compliance through persuasion and consultation before withholding the incentive payments. We believe this to be the case and the basis for our view is the evidence – which we will turn to imminently – which is bolstered by our conclusions regarding the countervailing power of the retailers. Certainly we believe it extremely unlikely that BATSA would have removed one of its CDU’s in order to punish non-compliance. This is why, despite the fact that it retained the power to terminate the agreement governing the provision of merchandising furniture, and despite the suggestions that it may have utilised this power or threatened to utilise it, we do not believe this to be at all likely. BATSA paid handsomely for its limited exclusive and, to be sure, where it was faced with comprehensive non-compliance – as in Pick n Pay’s Northern Gauteng region which we examine below – it would have withheld its incentive payments, but our general view of the balance of power between the retailers and BATSA leads us to conclude that they are unlikely to have withheld the incentive payments in the face of lesser instances of non-compliance.
Thirdly, the countervailing power of the retailers accounts for the high level of incentive payments which BATSA was compelled to make in exchange for a limited exclusive which generated returns that, as we shall demonstrate, are uncertain at best. The applicants have made much of the scale of the incentives that BATSA is willing to pay. It is suggested that the bases for these payments are compensation for the lost cigarette sales incurred by the retailers as a result of the category-damaging consequences of BATSA’s conduct.45 Alternatively – and rather eccentrically – it has been suggested, certainly where the payments made to the HORECA venues are concerned, that BATSA’s level of payment are driven by a raising rivals cost strategy in which is purposefully pays a larger incentive than that which the retailer demands in order to raise the cost of outbidding it.46 In our view neither of these explanations is persuasive. The more prosaic explanation is however. The scale of the payments represents the power of the retailers and the key HORECA venues vis a vis the manufacturers, including BATSA, and is inconsistent with the notion that BATSA enjoyed the freedom to do with the cigarette category what it wanted.
The upshot of our view on countervailing power is that the constraints that are characterised as core category management principles and that are needed to maintain a limited exclusive are likely to be in place regardless of whether or not they are actually reflected in a formal contract – in particular the core requirement that the retailer retains the power to impose its interests over that of the category manager. Indeed we are able to conclude that an experienced supplier managing a category on the part of retailers who are clearly cognisant of their power would, in its own interest, have internalised the requirements of the retailers when performing its category management tasks, without the expedient of a contractual requirement specifying the precise limits of the exclusive. The aggressive testimony of Mr. Van Vuuren and his account of his interactions with his suppliers, including BATSA, may well have been interpreted as bluster. On closer reflection they are likely to accurately reflect a very real power relationship.
In summary then we do not accept the argument that holds out a set of abstract, idealised and thoroughly implausible marketing principles as the standard against which BATSA’s conduct must be judged. Properly construed, the BATSA trade investment agreements, the hallmark of which is the constrained right that it has purchased over promotional opportunities at the POS, is a species of exclusive agreement. Critically, this includes the right to determine, within limits, the amount of space allocated to competitor brands and the positioning of those brands within the CDUs. The agreements also frequently include the right to determine the display of secondary promotional material at the POS. The right is granted against the payment of cash incentives, bolstered by the provision of free merchandising architecture at the retail POS.
However, this right, and therefore the exclusive itself, is limited. The retailer retains ultimate control over the POS. This includes control over the listing of new products and also over the form of the planogram, the map that lays out allocations of space and position on the CDU. No retailer has granted BATSA an exclusive right to the POS, and nor, as far as we know, has this been contemplated by BATSA itself. The nature of the product (characterised by a strong demand for variety and therefore a large number of brands with low market shares) and the consequent interests of the retailers strongly militate against granting a full exclusive to any manufacturer even one with as large a market share as BATSA. We will examine the trade investment agreements, particularly as they pertain to the retail outlets, including the vending machine operators, from the standpoint of a limited exclusive dealing arrangement.
There are well-established anti-trust methodologies for examining exclusive dealing arrangements. From an anti-trust perspective, the sensitivity that attaches to these arrangements derives from the likelihood of foreclosure and the anti-competitive harm that may be generated by significant foreclosure. It is however also widely established that exclusive agreements may well generate pro-competitive outcomes.
We have held previously that in order to sustain an allegation of abuse of dominance anti-trust harm must be demonstrated. We have also held that such harm may be inferred from a direct loss of consumer welfare or from a significant degree of foreclosure.47 However, not only must foreclosure be shown, the foreclosure must also be shown to have derived from the alleged anti-competitive conduct. In other words where elements of the foreclosure may reasonably be inferred to have occurred in consequence of conduct beyond the reach of competition law, then we cannot necessarily conclude that the foreclosure amounts to anti-trust harm. This observation is particularly pertinent in this case, because there is, as we shall show, an eminently reasonable inference to be drawn that harm to the competitive structure of the cigarette market has been caused by the comprehensive regulatory interventions that gave rise to the existence of a ‘dark market’. In this unusual instance, when there are two sources of likely foreclosure, we would require a showing that the extent of foreclosure generated by the anti-trust contravention was, on its own, and independently of the foreclosure attributable to the second source, sufficiently significant to infer anti-trust injury.
Furthermore, as already intimated, it is eminently possible to draw an inference of pro-competitive consequences arising from the limited exclusive agreements that BATSA have concluded with the retailers. It is likely that at least part of the incentive payments used to purchase the limited exclusive are passed through to consumers. And then there is the very significant investment by BATSA in merchandising architecture. Although the precise interface between the provision of merchandising architecture and BATSA’s privileged POS promotional rights is difficult to establish conclusively, we simply note here – and will elaborate further – that these inferred pro-competitive gains must be weighed against any inferred harm that may arise from the foreclosing effects of the limited exclusive.
Foreclosure
As we have already intimated, we believe that distribution and promotion must be carefully distinguished when measuring the extent of foreclosure. The extent of foreclosed distribution can be measured relatively easily. If BATSA enters into an exclusive distribution agreement with a chain of grocery stores – that is, an agreement that provides that only BATSA products will be available in the chain’s cigarette category - the extent of foreclosure will simply be measured by the chain’s share of cigarette sales. Promotion, on the other hand, embodies qualitative dimensions that cannot be so easily measured.
That having been said it is, at this juncture, appropriate that we restate our view that there is, when assessing the extent of foreclosure of promotional opportunities, a definite interface with the fact of universal distribution. It is, through long established practice, well appreciated and understood that all brands and SKUs of any significance, and a great many without any significance, are available at all outlets from which cigarettes are sold. This is certainly true of organised retail and of much of independent convenience. It is, for space reasons, not true of the vending machines at the HORECA venues but these will be dealt with separately.
Moreover because cigarettes are not sold from the open, self service shelves but have rather to be requested from an attendant at a dedicated kiosk or at the till, the purchaser does not have to browse the shelf in order to find her brand.
In short, a cigarette purchaser knows full well that her brand will be available at the POS, and, because of the loyalty element that characterises cigarette purchases, she knows full well what brand she will be purchasing before she even approaches the POS, and, even if she does not spot her brand on the CDU, she will merely have to ask the attendant to supply it. All of this serves to reduce significantly the influence of the retail shelf or POS over cigarette brand purchasing, certainly where the established smoker is concerned.48 If the purchaser was not certain of the availability of her chosen brand, then indeed a limited amount of shelf browsing may take place before the selection of the brand takes place and relative space and positioning may then assume more importance over brand selection. However, that the consumer knows full well that distribution is universal is a key factor that reduces the importance of the POS as a site of promotion, and this is further reduced by the intermediation of an attendant and the fact of brand loyalty which, for most smokers, predetermines the choice of brand.
As noted above, this is unlikely to apply with the same strength to a first time smoker. That is, someone who has little or no pre-conceived idea of the brand that she wants to purchase is more likely to be influenced by space and positioning on the shelves than would the experienced smoker. However, given the overriding importance that the applicants have attached to peer pressure and social signals in the selection of a cigarette brand, it is likely, on the applicants’ own argument, that a significant portion of even first time or inexperienced smokers will approach a retailer’s cigarette kiosk with a reasonably clear idea of the brand that she wishes to purchase. As already noted we give little credence to the contention of Mr. Scammel Katz, JTI’s marketing expert, that cigarette purchasers would be too ‘embarrassed’ to ask for a brand that they could not immediately see on the CDU.
In brief then our approach is that while distribution foreclosure and promotional foreclosure are to be measured using distinctive criteria – the former purely quantitative, the latter with a strong qualitative dimension – there is, certainly as far as the established smoker is concerned, a strong inverse relationship between the known fact of universal distribution (that is, the complete absence of distributional foreclosure) and the importance of the POS as a site of promotion.
The huge edifice of this case then comes to rest on an extremely narrow platform. The conduct alleged relates to plannogramming (space and position in the CDUs and vending machines), exclusivity over secondary displays and sundry acts of sabotage (removal by BATSA of rivals’ secondary display material). Also at issue are exclusive opportunities allegedly purchased by BATSA for hosting promotional events at HORECA venues. While JTI continues to hold both space and positioning important to its case, the Commission’s counsel tells us that his client views position as the critical element. Both the Commission and JTI argue vehemently that independent convenience is an ineffective site for promotional opportunities and so, they insist, are non-metropolitan areas. It is agreed that promotion is directed overwhelmingly at young adult smokers or first time smokers but it appears that the ‘main market’, the cigarette industry’s particular euphemism for black smokers, is not considered fertile ground for the promotion of popular and, certainly, premium brands.49 And it is agreed that promotional activity is directed as supporting premium and popular brands and not at VFM brands.
The core of the Commission and JTI’s case is then narrowed to the influence of position in the CDUs at organised retail and in vending machines, with JTI still arguing that allocation of space at these points of sale is important, to sundry acts of sabotage (although this has generated more atmosphere than evidence), and to promotional events at selected HORECA venues. Moreover the promotional activity with which the applicants are concerned is overwhelmingly that directed at influencing young, predominantly white, adult smokers who reside and entertain themselves in the upmarket parts of the metropolitan areas, a small sub-set of the smoking population, in their choice of premium and popular brand segments of the cigarette market.
This all adds up to a very slender basis upon which to claim significant foreclosure, not to mention far-reaching remedies. There is clearly some substance in the observation of BATSA’s counsel that its opponents have narrowed the basis of their concern precisely in order to ensure that the scale of foreclosure increases relative to the ever diminishing universe identified as viable for promotion.50
Indeed, the approach of the applicants serves mainly to draw attention to the really fearsome gorilla in the room and that is the regulations that have prohibited above-the-line advertising and sponsorship. In other words if this is all that is left of the once ubiquitous juggernaut of cigarette brand promotion, then it speaks volumes about what was removed by the regulations. And so even if the applicants had been able to prove all of the elements of their case – and we do not believe this to be so - it would be difficult to conclude that the significant element of foreclosure, and therefore the cause of any harm, inferred or otherwise, comes from the conduct of BATSA rather than from the decisions of the legislature whose manifest intent was to limit, indeed to eliminate, the promotion of cigarette sales.
The promotional theory that is advanced places peer and social pressure at the centre of a smoker’s brand choice. It is accepted that young adult smokers are the principal target of promotional activity. These are either first time smokers or smokers relatively uncommitted to ‘their’ brand and therefore susceptible to switch brands as a result of promotional activities. The sites at which young adults congregate are HORECA venues. The extent and character of social pressure generated by HORECA venues varies from venue to venue. Only a select number of HORECA venues have, we are told, the cachet which generates social pressure to conform to the brands with which the venue has associated itself and the choices of the venue’s clientele.
Having been influenced by the association of a desirable HORECA venue and its clientele with a particular brand, the young smokers or would-be smokers will endeavour to purchase their chosen brands at one or other of the retail outlets. It is agreed that very few of the actual purchases take place at the vending machines in the HORECA venues (of course only a very small proportion of the target demographic of the cigarette manufacturer’s promotional efforts actually attend these venues and those who do would tend to bring their cigarettes with them).51 Having been introduced to a particular brand they will make their subsequent purchases at the various outlets that retail cigarettes. In fact, garage forecourts stores are accorded particular significance as sites of purchase in a hub (horeca venue) and spoke (retail outlet) theory of cigarette promotion and purchase.
While associating a desirable HORECA venue and its clientele with a cigarette brand is generally considered to be a particularly effective form of brand promotion through peer pressure, the general impression gained from the mass of evidence is that, in one way or another, it is peer suggestion in general (which may take place at a HORECA event or venue or may simply come via friends and colleagues), that is the most common mechanism whereby a smoker or would-be smoker is introduced to a new brand. The evidence suggests that a small, but not insignificant, proportion of young smokers are, at any given point in time, susceptible to switching cigarette brands.
The purchase then takes place at a retail outlet. Clearly the POS possesses some degree of independent power of communication or persuasion. That is, the POS is not merely a site of distribution from which the new smoker or the adherent to a new brand, acquires his cigarettes having been persuaded of the newly chosen brand’s virtues by the peer pressure exercised in another setting. We do not however understand that a strong claim was made to the effect that browsing the CDU at the retail outlet is an important mechanism for choosing a new brand. Rather the high water mark claimed for the role of the POS is that a relatively uncommitted smoker approaching the cigarette kiosk, possibly with a brand in mind, may not, at a glance, see his brand on the CDU. Embarrassed at having to stand at the till or the kiosk searching through the CDU and embarrassed at having to request the attendant to supply a choice which he may believe is not supplied by the store – a claim that we find hard to credit - his eye will alight upon the ‘hot spot’ shelves, resplendent with one or other BATSA drive brand, and, being relatively uncommitted to his new brand, he may then select the brand best positioned and allocated the most space. This is the basis for the professed anxiety of the applicants regarding the allocation of space and position in the retail outlets.
Indeed even this may overstate – or, indeed, may misstate - the claim made on behalf of the POS. Possibly in recognition of the highly limited influence of the POS on short term purchasing decision, the claim for its promotional significance is more frequently grounded in its long term influence - the constant repetition of the communication received from the POS in the form of the greater visibility of a particular brand is postulated to impact on longer term decisions regarding the selection of a cigarette brand. This is, of course, a convenient claim for the applicants not least because it makes it extraordinarily difficult to measure the impact of promotional efforts at the POS. It does though weaken the claim regarding peer suggestion which we would have thought is, by its very nature, pressure that is manifest in a short term, indeed immediate, decision to experiment with a new brand or to savour, for the first time, the delights of smoking.
These then are the parameters within which we will consider whether BATSA’s conduct with respect to the allocation of space and position and the other limited exclusives it has acquired at the POS has significantly foreclosed the promotional opportunities of BATSA’s rivals. However before doing so we should dispose of an important issue regarding the allocation of space and position on the basis of manufacturer share rather than brand or SKU share.
The applicants have made much of the fact that, because manufacturer, rather than brand or SKU, share determines the allocation of space on the CDUs, certain of BATSA’s brands have received allocations of space and position significantly disproportionate to their market shares because BATSA reduces the space effectively commanded by the market share of Stuyvesant and increases the space commanded by the market shares of its lower selling brands. In particular it has been pointed out that Lucky Strike and Dunhill have benefitted from these skewed allocations. BATSA acknowledges that its own brand development strategies dictate the allocations of its brands within its overall manufacturer allocation. While it is not clear whether the applicants actually contend for space and position to be allocated on the basis of brand or SKU, they nevertheless object strenuously to the ‘excessive’ allocations accorded to Lucky Strike and Dunhill in particular and advance this as evidence of the extent of BATSA’s deviation from pure category management principles and, by extension, as evidence of its anti-competitive conduct.
However we can find no fault with BATSA’s approach on this score. On the contrary, it would be an intolerable limitation on competition in the cigarette market were a manufacturer compelled to concentrate its promotional resources (and certainly on the applicants’ version this incorporates allocations of space and position at the POS) on a brand that, although with a large current market share, it believed to be in long term decline or, for one reason or another, less well suited to competing with a rival brand than another brand in its portfolio, albeit one with lower market share.52 We can understand why, particularly in dark market conditions, BATSA may wish to focus its promotional efforts on international brands like Dunhill and Kent rather than on big, well established national brands like Peter Stuyvesant.53 It is certainly not the place of competition authorities to stand in the way of decisions of this sort. In any event, the extent of ‘excessive’ space accorded a particular brand is limited by the necessity to avoid out-of-stocks of under-represented brands and holdings of unsold stock of over-represented brands.
There is then no question in our mind regarding BATSA’s entitlement to distribute its allocations of position and space between its various brands as it sees fit. We will not comment further on this issue save to point out that after lengthy periods of ‘excessive’ allocations of space and position Lucky Strike remains mired in very low market shares, an outcome which is starkly inconsistent with the promotional power – in either the short or long term - that the applicants ascribe to the retail POS.
We now proceed to examine the foreclosing effect of BATSA’s trade investment agreements on the points of sale in each retail channel.
BATSA’s trade investment agreements – organised retail
As noted organised retail refers to two of the large national grocery chains (Pick n Pay and Shoprite Checkers - the third, Woolworths, does not sell cigarettes), the large chains of mostly franchised convenience stores, including Spar, and the organised garage forecourts stores, some of which are franchised stores while others are operated by the various oil companies. Certain of the chains – particularly the grocery chains – are sub-divided into separately managed formats such as supermarkets, hypermarkets and franchised stores.
In 2003, organised retail contributed approximately 19.5% of total cigarette sales in South Africa with the grocery channel contributing 5.9%, compared to 10% contributed by organised convenience, and 3.6% by organised forecourts. In the organised convenience channel, Spar accounts for more than half of the cigarette sales and Pick n Pay franchise stores contributed approximately 20% of the total cigarette sales. 54 The proportion that cigarettes contribute to the total sales or turnover in a channel is lowest in grocery followed by organised convenience, with the highest proportion in organised forecourts where cigarettes account for between 20-30% of an outlets turnover.55
Grocery
The grocery channel, essentially comprising Pick n Pay and Shoprite Checkers, contributed approximately 5,9% to total market sales of cigarettes in 2003. The largest contributors were the Pick n Pay supermarkets which accounted for the 30,5% of the channel’s sales (1,8% of total cigarette sales) and the then 343 Shoprite Checkers’ supermarkets which accounted for 55,9% of the channel’s sales (3,3% of total cigarette sales), with the remainder accounted for by the hypermarkets belonging to each chain. 30% of the adult cigarette purchasers in this channel are under the age of 30 and 14% under the age of 24. 10% of smokers and only 8% of ASU30s record buying their cigarettes most often in this channel.56
Although there are important differences in the agreements concluded between BATSA and the grocery chains and in the manner in which each of the grocery chains translates the agreements into actual merchandising practices, key general features are common. The agreements are all characterised by the specification of cash incentive payments to be made against performance on a range of ‘drivers’. In the case of Shoprite Checkers these drivers are in-store merchandising architecture, ranging and availability and, critically for these proceedings, planograms, the driver which governed the allocation of space and position.57 However, it appears that for the most part – and this appears to be so in all the channels in which space and position are one of the drivers - the agreements governing BATSA’s right with respect to space and position on the CDU are not generally accompanied by formally approved planograms, although versions of planograms of uncertain status proliferate and were filed. Note too that Shoprite Checker’s stores all utilised CDU’s provided free of charge and at considerable cost by BATSA. These were provided in terms of separate agreements, the Equipment Loan Agreements (ELAs), concluded between the parties. We will discuss the interface between the trade investment agreements and the ELAs below.
Two other pertinent general features relating to the agreements are noteworthy. Firstly, the trade investment agreements are of short duration. The initial agreement concluded between Shoprite Checkers and BATSA in July 2002 ran for 18 months. Subsequent agreements were for a 12 month duration. The duration of the agreements or understandings with Pick n Pay supermarkets – as we shall see the status of the unsigned agreements is uncertain - range from six months to 15 months. Thus to the extent that the existence of an agreement shuts out the possibility of competition for promotional space – and in our view it does so to only a limited extent – the duration of the agreements ensures that the competitive arena is re-opened at short intervals.
Secondly, and related to the above point, JTI’s failure to engage with the grocery chains in the competition for space and position on the CDUs and at the POS generally is conspicuous. Indeed, far from attempting to offer, on any basis whatsoever, a competing or additional incentive, JTI did not until 2007 even purchase Shoprite Checkers or Pick n Pay sales data, the threshold requirement imposed by the retailers for participating in category management. By contrast, in 2002, the year in which the first agreement was concluded between Shoprite Checkers and BATSA, the latter paid R120 000 for the sales information generated by the grocery chain.58 The Pick n Pay witness, Mr. Stephen van Vuuren, is very clear on the requirement to purchase the retailer’s EPOS data as a pre-condition for participating in planogramming. van Vuuren contends, and he was not challenged on this, that since 2007 when, JTI purchased its EPOS data for the first time it and PMI, which had agreed to purchase this data some time earlier, have provided input into the category management process.
JTI argues that a key factor, particularly if not exclusively applicable to the grocery chains, that constrained it from bidding against BATSA is the ‘all or nothing’ nature of the incentive. In other words, alleges JTI, it was compelled either to bid against BATSA for the full package of incentives, for the full category of items, agreed between BATSA and the retail chains, or to accept nothing at all. This issue is examined below.
As to the ever contentious matter of planogramming, Mr. Ackerman, Shoprite Checker’s senior buyer, testified that the rate of sale determined the quantum of space allocated and that in the management of the category, particularly in the allocation of space and position, divergent demographics of the various stores had to be factored in.59 Ackerman also testified to the elaborate procedure that had to be followed in the event that BATSA claimed non-compliance that may result in the withholding of part of the incentive payment. In essence BATSA was only entitled to withhold compliance payments if it had notified specified managers within the retail chain, including ultimate escalation to Ackerman himself, of the alleged non-compliance and only then if the non-compliance persisted for a period of four weeks. Ackerman suggested that the cumbersome and high-level nature of this process ensured that it was only activated in the face of serious allegations of non-compliance. He also testified that none of the other cigarette manufacturers approached him with alternative plannogramming proposals.
We found Ackerman to be a reliable and informative witness. Admittedly he was clearly sceptical of the claims made for the brand promotional power of the CDU display, indeed of the entire POS display. He also emphasised the overriding importance of payments for shelf space in the Shoprite Checkers business model. This combination of considerations undoubtedly gave rise to a certain degree of abdication of the category management function to the category captain, BATSA. We do though accept his evidence regarding listing and ranging and the use of rate of sale as the criterion ultimately governing the allocation of space, certainly to the extent that a BATSA-inspired deviation from a rate of sale rule may have resulted in out of stock situations which would not have been tolerated by the retailer.
We also accept Ackerman’s evidence regarding BATSA’s flexible attitude towards compliance. The centrality accorded by Ackerman to the incentive payments, and the role that he himself was willing to play in the process of reporting non-compliance, may be read in one of two ways. It may testify to the importance that Ackerman attached to ensuring his every store’s compliance with the details of the agreements so as to be certain of securing the full package of incentive payments. This is obviously the interpretation favoured by the applicants. Or it may signal that serious non-compliance had to be established before payments were withheld. That, in other words, BATSA would forebear from escalating trivial acts of non-compliance to Ackerman’s level and would adopt a flexible attitude to the agreements including flexibility around space and positioning designed to cater for the specificity of its divergent store demographics, and, as such, inevitably managed by store management.
We doubt that this flexibility would have extended to the exclusion of BATSA from the ‘hot spot’ but it does not suggest a rigid approach to the allocation of space.60 On the whole we accept the version that effectively says that Shoprite Checkers attaches great significance to incentive payments and, while this means that it will do its best to ensure broad compliance, neither will it readily accept non-payment in the event of deviations attributable to the requirement for flexibility – for example, when confronted by the demographic divergence between its stores - rather than to wilful, serious non-compliance.
The situation with respect to Pick n Pay supermarkets is less clear in keeping with the decentralised nature of the Pick n Pay operation. This business model underscored the requirement for flexible application of agreements concluded at the head office. In fact it appears that from 2006 separate agreements were concluded between BATSA and the individual regions of the Pick n Pay supermarkets.
For much of the complaint period, it does not seem that Pick n Pay concluded binding written agreements with BATSA. It’s not clear how much weight to attach to the non-binding character of the agreements because there were clearly understandings that were enforced in relation to the cigarette POS and while these conformed in many respects to the agreements concluded with Shoprite Checkers there were important differences.
In keeping with the approach of Shoprite Checkers, rate of sale is the fundamental criterion in allocating space and ranging and listing remain the preserve of Pick n Pay. Again we take this latter to mean that BATSA’s limited exclusive did not extend over ranging and listing, that is, it could not determine its rivals’ listings – this was the preserve of the retailer – but it did secure an agreement that all of its own products would be listed and that an incentive would be paid for compliance.61
However, in marked contrast with Shoprite Checkers, Pick n Pay utilised its own merchandising furniture, including the costly CDUs, apparently because it feared compromising its independence from the supplier. This having been said, there is no evidence that the provision of free CDUs to Shoprite Checkers ever interfaced with the trade investment agreements. It appears that in Shoprite Checkers the agreements regarding the CDUs were concluded at store level and there is no evidence that withdrawal or threatened withdrawal of the merchandising furniture was ever used to enforce the nationally negotiated trade investment agreements. We will say more about the provision of CDUs below.
As with Shoprite Checkers, Pick n Pay was incentivised to comply with BATSA drivers by the payment of cash incentives. However it appears that while the agreements regarding the total payment by BATSA to Pick n Pay were concluded at national level, implementation, including the application of regional particularities, was negotiated with the regional buyers.
Planogramming in Pick n Pay was attended by even more than the usual uncertainties and disagreements. Hence van Vuuren insists that despite receiving regular compliance payments from BATSA (as well as considerable sums in exchange for Pick n Pay’s EPOS data), the Western Cape region of Pick n Pay did not implement planograms submitted by BATSA. The reason advanced was that he, van Vuuren, had never accepted, never agreed to, any of the planograms proposed by BATSA which he consistently described as mere ‘wish lists’. However, his refusal to accept or comply with the planograms drawn up by BATSA did not, in van Vuuren’s estimation, impact on Pick n Pay’s right to receive the full incentive payments from BATSA which, he insisted, he continued to receive. van Vuuren’s rather eccentric interpretation of the agreement is that BATSA was obliged to furnish him with a planogram acceptable to him and until that was done Pick n Pay was under no obligation to implement an important element of its side of the bargain – to implement planograms agreed with BATSA - without thereby impacting on BATSA’s obligation to make good on its side of the bargain, that is to continue making the full incentive payments that had been agreed upon.
van Vuuren’s interpretation of the terms of his agreement with BATSA may reflect little more than the dynamics between a powerful buyer at South Africa’s largest grocery chain and a supplier. More likely, and for the sake of argument we should assume this to be case, it indicates that, regardless of whether or not its planograms had been formally endorsed by Pick n Pay, the actual extent of conformity between what BATSA desired and what was acceptable to Pick n Pay and therefore implemented was sufficiently close for BATSA to continue making the incentive payments. The much debated case of the northern Gauteng region of Pick n Pay appears to demonstrate that when the CDUs ceased to reflect in any way BATSA’s planograms either with respect to space or positioning, BATSA stopped making payments in respect of that region to the retail chain.
The 2007 agreement with Pick n Pay supermarkets makes reference to a 90:10 space allocation as between BATSA, on the one hand, and the brands of other manufacturers, on the other. van Vuuren insists that this was not applied. It clearly reflects BATSA’s stated aspiration that space allocated in all channels reflects its national market share – indeed this is somewhat higher than its 2007 national market share. Moreover the national share is skewed by BATSA’s particularly large share in independent convenience and the ‘main market’ in particular, and is certainly higher than the grocery share.
We would conclude from this that the organisation of the cigarette POS in Pick n Pay supermarkets generally conformed to BATSA’s planograms and contractual terms. However they were clearly not applied unexceptionally. We have already referred to the Waterfront store where the demographics and demand patterns of the clientele appear to have accounted for a significant departure from BATSA planograms and in which store management accorded Marlboro pride of place both with respect to allocations of space and position. It does not appear that incentive payments due in respect of this store were withheld because of this instance of non-compliance, nor is it clear how many other similar instances of non-compliance occurred. The decentralised nature of the Pick n Pay operation would have facilitated deviations from the national norm and would account for the northern Gauteng rebellion, the most widely discussed instance of non-compliance which we deal with below.
The most plausible conclusion to be drawn from van Vuuren’s evidence is that while the shelf space and positioning allocations used in the cigarette category in his region conformed broadly with BATSA’s planograms, this is principally because, while they did not entirely conform to van Vuuren’s view and thus were never formally agreed, they were nevertheless not thought to conflict fundamentally with Pick n Pay’s own interests. It is difficult to believe that the incentive payments were not factored into Pick n Pay’s interest and that they did not exercise some influence on the organisation of Pick n Pay’s CDUs, but they certainly do not appear to have been the only factor. van Vuuren believed emphatically that he was entitled to the payments regardless of whether or not he implemented BATSA’s planograms and there is clear evidence of significant deviation from BATSA’s planograms when the demographics and demand patterns of a particular store warranted it. This is all of a piece with van Vuuren’s blunt assertion, already cited, that ‘the best form of category management depends on the strategy of the retailer’.
van Vuuren’s approach and his refusal to be bound by BATSA’s planograms also lends weight to his assertion that he would be prepared to consider a planogram presented by any supplier who met his minimum condition for participation in the category management process, namely that it purchases Pick n Pay’s EPOS data. Indeed as indicated above, van Vuuren testified – and we do not understand this testimony to have been challenged - that since JTI and PMI have commenced paying for the EPOS data they have been permitted input into the holy sanctum of planogramming. As with Shoprite Checkers we do not believe that this would have extended to significant compromise on the ‘hot spot’ display – this is, after all, the central aspect of the limited exclusive that BATSA has purchased - but the evidence suggests that it may well improve both the positioning and the space allocated to BATSA’s rivals without resulting in a withholding of the incentive payment. As elaborated below we believe that the northern Gauteng incident bears out this contention.
The reasonable conclusion to draw – both with respect to Shoprite Checkers and Pick n Pay supermarkets - is that JTI’s lack of influence at the POS was in significant part attributable to its reluctance to expend any resources on promotion, even including purchasing the EPOS data, a threshold condition for participating in category management. Indeed even beyond JTI’s refusal to purchase the EPOS data and its effective abdication from participation in the planogramming process, both Ackerman and van Vuuren were clearly struck by what can only be described as JTI’s starkly supine attitude towards trade marketing in the grocery channel. Witnesses from the retail trade persistently remarked upon the sparse attempts at contact made by JTI.62
We hasten to add that JTI’s refusal to expend time and money on trade marketing in the organised grocery channel may well reflect a rational, considered commercial decision on its part. Like van Vuuren and Ackerman we are sceptical of the promotional power of the organised grocery POS, particularly given the channel’s relatively low contribution to cigarette sales and the demographics of its customers. It makes perfect sense for a manufacturer with constrained resources to abandon promotional activity at the grocery channel’s point of sale, particularly as the limited exclusive and the retailers’ listing requirement guaranteed it a POS presence without, we should add, having to spend a cent. Moreover, BATSA’s category captaincy ensured a neat, uncluttered POS and CDU. However, if JTI has indeed decided, on solid commercial grounds, not to expend its trade marketing resources in organised grocery, it behoves us to adopt a sceptical approach to a strategy that seeks to invoke competition law in order to neutralise BATSA’s promotional expenditure, largely by permitting JTI to free ride on its rival’s efforts.
Our reluctance to permit JTI to use legal stratagem to blunt its rival’s efforts at trade marketing is bolstered by the observation that where JTI did elect to expend resources on trade marketing, even in organised retail, it succeeded in wresting promotional opportunities from BATSA. As mentioned above, when JTI eventually decided to purchase the Pick n Pay EPOS data it was given input into planogramming. This does not mean that it would have displaced BATSA from the hot spot or even that it would have been allowed to ascend from the bottom left hand corner of the CDU. In order to do this it – presumably with the support of PMI - would have to have used its opportunity to make input in order to demonstrate to the retailer that the present arrangement regarding space and position redounds to the ultimate detriment of the retailer and its customers, precisely the case that it claims to make here. Their arguments would have undoubtedly entailed the offer of an incentive payment, an important consideration in the advantage that accrues to both retailer and customer from the present arrangement. That is to say, appeals to the ‘principles’ of category management are extremely unlikely to suffice in the real world. However, as noted above, we believe that a combination of factors would enable BATSA’s rivals to make some headway with respect to space and positioning without the retailer risking the withholding of BATSA’s entire incentive payment.
This is, we believe, demonstrated by the northern Gauteng incident. To cut short a story debated at inordinate length, it appears that the Pick n Pay buyer for the northern Gauteng region insisted on placing Camel in the top shelves alongside the other premium brands. BATSA objected. The positioning of its rival was nevertheless improved along the lines desired by the northern Gauteng buyer and the amount of space allocated to BATSA was reduced significantly. It appears that BATSA initially continued making its incentive payments with the exception of the payments in respect of the planogramming driver. However, far from dissuading the buyer from continuing along this path, BATSA’s position, both in term of space and position continued to deteriorate with a concomitant improvement in Camel’s allocations of space and position, including in relation to its privileged place in the hallowed ‘hot spot’. Ultimately BATSA concluded, in the words of its representative, that it was ‘paying for nothing’ and withheld its entire incentive payment.
As far as we are able to glean – and the evidence is not clear – the northern Gauteng buyer’s demands did not arise as a result of any sudden outraged sense of fairness or even commercial epiphany on his part regarding the appropriate location of Camel in the CDU, but rather in consequence of unusually robust trade marketing, including the offer of a competing incentive, on the part of JTI. Indeed JTI effectively bid to become category manager, a bid in which the persuasiveness of its representatives was backed up by the payment of a considerable incentive.
This incident has received the degree of attention that it has because, allege the applicants, it reveals the ‘all or nothing’ character of BATSA’s incentive package, the refusal of BATSA to sever one element of its incentive package from another. We are not confident that it conclusively clarifies BATSA’s position on severability, although we are constrained to observe that it certainly does establish that it is, at least, possible to compete for promotional opportunities in Pick n Pay on a regional basis. That is to say, it establishes that a rival to BATSA did not have to secure promotional privileges or category captaincy across the national Pick n Pay supermarket chain, but rather that it was possible to contend for a particular region. This alone calls into question the ‘all or nothing’ contention advanced by the applicants, at least as it applied to Pick n Pay.
On the question of the severability of one driver from another, we believe that the northern Gauteng incident reveals that, to a degree, there is indeed severability. That is to say, BATSA will accept quite significant deviations from its planogram – especially where the quantum of space allocated to it is concerned – and, at most although not inevitably, withhold its incentive payment in respect of the planogram driver only, while continuing to make good on the other incentive payments. Nor is this surprising. Consistent with our view on the question of countervailing power, we cannot see that BATSA has an interest in adopting an overly aggressive stance vis a vis an important retailer and so it will continue to maintain the relationship, which is defined by the incentive payments, until a particular, albeit blurred, line is crossed when, in the words of its head office representative who dealt with the northern Gauteng saga, Mr. Garthwaite, it will conclude that it is ‘paying for nothing’ and it will stop all of its payments.
This point will inevitably be reached when BATSA’s hold on the ‘hot spot’ is severed. And nor should this be surprising because when this line is crossed BATSA will, in effect, no longer be category manager, it will no longer be in possession of the limited exclusive which permits it to determine the allocations of space and, of particular importance, position, to its rivals. In fact what the Commission conspicuously omits to mention in its heads of argument is that JTI precisely displaced BATSA by assuming the role of category captain and by paying a competitive incentive for this privilege.
This view – namely, that BATSA will forebear from terminating its relationship with a major retailer, manifest in the incentive payments, only with considerable reluctance – applies with even greater strength to the notion that BATSA would lightly withdraw its CDUs.
The applicants make much of the provision by BATSA of free merchandising equipment, most notably the costly macrobat CDUs.63 In particularly, they seek to establish an interface between the provision of the merchandising furniture and the plannogramming rights. They point to clauses in the equipment loan agreements (ELAs) – the agreements that govern the provision of the merchandising equipment - that make reference to plannogramming rights. They argue that the duration of the ELAs – typically five years – effectively extends the duration of, ensures the roll over of, the much shorter trade merchandising agreements because the presence of BATSA furniture, and the prospect that BATSA may, in the event of failing to renew its trade merchandising agreements, terminate the ELA. BATSA, for its part, points out that the ELAs are concluded on a store-by-store basis and are not part of the remit of the trade merchandising teams.
While all of this may be true we are not persuaded by what is effectively the applicants’ core contention in this regard, namely that BATSA would withdraw its CDUs in order to enforce and extend its trade merchandising rights. The provision of free CDUs represents an extremely costly investment on BATSA’s part, particularly in organised retail, where the retailers value achieving a similar look and feel across its branches.64 BATSA will undoubtedly attempt to recoup as much of this investment as possible and so may well suggest a linkage between, on the one hand, continued access to this equipment and, on the other, the enforcement of its limited exclusive, and particularly the planogramming driver. On the other hand it is a very valuable benefit to the recipients and so any prospect of losing the CDUs may be a powerful incentive for compliance. However, on any sensible construction, BATSA’s actual ability to withdraw its CDUs, particularly in organised retail, is extremely limited. By withdrawing CDUs from a small number of non-compliant stores in a chain, the retailers’ efforts to achieve a similar look and feel is immediately undermined. It is unthinkable that BATSA would, in the event of a disagreement with, say, Shoprite Checkers start withdrawing its CDUs. The loss of goodwill that this would entail would severely damage the relationship between BATSA and a significant customer. In this little corner of the world, removal of the CDUs is equivalent to exercising the nuclear option and the likelihood of that is close to zero.
We have no doubt that the provision of free merchandising equipment purchases significant goodwill for BATSA and that BATSA will seek to milk this to the maximum extent possible, in particular by attempting to assert a linkage between the free CDUs and compliance on space and positioning. However, we are of the view that, for the most part, it represents an investment in the category, an investment which only BATSA would make on a significant scale. Certainly, the counterfactual, less adequate CDUs or the retailers paying for the equipment itself, will result in either a less well managed and presented category or significantly less pass-through to end consumers. It is telling that for all that the applicants seek to make of the threat posed by removal of the CDUs, we cannot recall a single example – even from independent retail where BATSA’s ability to contain any possible fall out resulting from a withdrawal of a freely provided CDU is much greater – of BATSA having acted, or even having threatened to act, on this. As noted Pick n Pay has installed its own CDUs – though with support from BATSA – in preference to using free BATSA-owned CDUs. This preference is rooted in Pick n Pay’s desire to remain independent of its suppliers. While this indicates that Pick n Pay believes that free CDUs could be used as leverage by BATSA, we do not believe that, once installed, the threat of withdrawal is credible. Indeed Pick n Pay also helps to establish that compliance with BATSA’s desired POS arrangements did not depend upon the provision of free merchandising equipment.
It appears that the in-store merchandising architecture driver is more concerned with the requirement to remove non-tobacco products from the CDU and the placement of secondary promotional materials at the POS than with the plannogramming of the CDU. The former requirement – to keep the CDU free of non-tobacco products – brings no particular advantage to BATSA. Category promotion is clearly the consideration here.
Where secondary displays are concerned – and this included secondary CDUs – to the extent that we are able to decipher the conflicting evidence, it appears that while BATSA offered an incentive to ensure that its secondary promotional material was displayed, there are numerous examples of BATSA’s rivals placing secondary material at the POS. The merchandising architecture right was not an exclusive right although JTI pointed to certain agreements which appeared to require that only BATSA secondary material be placed at the POS or, at least, that the secondary materials of its rivals not obscure the CDU in any way.65 However Mr. Ackerman of Shoprite Checkers testified that PMI approached him and purchased the right to place a dedicated dispensing unit on the countertop. He testifies further that only after PMI secured this for its Marlboro brand was he approached by JTI.66 Mr. Van Vuuren testified to the secondary promotional materials – including dedicated counter-top CDUs – that had been taken up by PMI, Gallaher and JTI. He too testified that he was only approached by JTI after PMI had secured the placement of its secondary material in the Pick n Pay stores.67
The hearings have tended to focus on the main CDU, with the secondary promotional materials implicitly viewed as a poor relative.68 But this is not necessarily true. Mr. Wittenberg, a BATSA witness, argued persuasively that in the promotion of a new brand, dedicated secondary promotional material is more effective than a facing amongst dozens of brands on the main CDU.69 BATSA argues that in the independent convenience channel which is examined below secondary materials were particularly important because a large number of the CDUs in the stores in this channel were pacrobats which offered fewer promotional opportunities and planogramning than did the macrobats.70
We acknowledge that both grocery chains, but particularly Shoprite Checkers attach considerable significance to the payment of incentives by their suppliers. We recognise that a resonant theme of Ackerman’s evidence and that of van Vuuren is the view that the POS display, and certainly the CDU display, has little influence over their customers’ choice of cigarette brand.71 In other words, they are of the view that they are receiving a significant incentive from BATSA that has no impact on their customers’ choice of grocery outlet because the requirement of in-store variety is met by the terms of the limited exclusive (which required all listed brands and SKUs to be allocated space) combined with an expansive listing policy (which ensured that most of the brands and SKUs of all of the manufacturers were listed). This is the basis on which they are only prepared to offer a limited exclusive.
We acknowledge that while it is not clear whether or not Ackerman and van Vuuren believe that the management of the category is having any positive impact on overall sales in the category, they are manifestly sceptical of the cigarette companies’ view that the limited exclusive is, through the organisation of the CDU and the POS, impacting on their customers’ choice of cigarette brand. They clearly believe that this choice is overwhelmingly determined by brand loyalty strengthened by the fact that an attendant mediates the interplay between the customer and the shelf display thus vitiating the impact of the display.
We are, however, prepared to accept that, notwithstanding van Vuuren and Ackerman’s views on the extent of brand loyalty, what little evidence there is suggests that, at any given point in time, a small but not insignificant number of smokers are indeed prepared to contemplate a change in their cigarette brand. However this does not undermine the validity of the retailers’ reading of the situation in their chains regarding the lack of impact of the POS display and merchandising on the distribution of sales between brands and manufacturers.
The fact is that the data regarding the proportion of smokers willing to change brands are overwhelmingly concentrated amongst precisely the demographic targeted by the cigarette companies, namely young adult smokers. However these are not well represented amongst the cigarette customers of the large grocery chains. In fact it would be reasonable to hypothesise that they are not well represented in the overall customer base of the large grocery chains.72 The POS in this channel, including the CDU display, is thus unlikely to have little independent impact on promotion. Indeed, in contrast with the hub-spoke relationship hypothesised to apply in the HORECA-forecourt interplay, the grocery points of sale are even unlikely to track the impact of successful promotional activity elsewhere. The data establish clearly that their customer demographic accords closely with the implicit categorisation of those who know them best, namely the retailers themselves. They are brand loyal, older cigarette consumers, well used to shopping in this retail format and, so, certain that their cigarette brand of choice is available, and, presuming that they even bother to cast their eye over the display before making their request of the attendant, unembarrassed at the thought of asking the attendant for their brand should their eye not immediately alight upon it.
We conclude then that even if BATSA has comprehensively planogrammed the CDUs of the grocery chains – and the evidence ranging from the Pick n Pay Waterfront store to the northern Gauteng incident suggests something less than total coverage and a degree of pragmatic flexibility on BATSA’s part – their role in cigarette promotion is relatively inconsequential and so could make little, if any, contribution to foreclosure of promotional opportunities. The grocery channel’s contribution to total cigarette sales is relatively small and its customer base does not match the target demographic of the cigarette manufacturers. Moreover, to the extent that it does contribute to foreclosure, competitive retailers are favoured with significant cash incentives which, it is reasonable to infer, are partly passed through to their customers.
It is little wonder then that JTI – and, it would appear, the other non-BATSA manufacturers – have chosen to devote few resources to outbidding BATSA in the organised grocery channel. The low transactions costs involved in dealing with two relatively centralised head offices may, on a superficial reading, make it a relatively attractive channel to target, but privileged space and positioning allocations do not come cheap from the large grocery chains and the returns, insofar as promotional opportunities are concerned, appear, at best, highly uncertain. This is presumably why JTI has attempted to use this regulatory intervention in order to try and achieve what it is not willing to pursue through the more costly expedient of actively competing for space and position.
Organised convenience
The relationship of the cigarette manufacturers to organised convenience is somewhat different to that of organised grocery. This channel is, in sheer quantitative terms, a significantly more important sales channel for cigarettes, accounting for 10% of total cigarette sales in 2003 and 9,5% in 200673. For this, among other important reasons, JTI appears to have more actively courted promotional opportunities in this channel.
Note however that while cigarette sales from the organised convenience channel represents a significant proportion of national sales, the demographic make up of this channel’s custom, though more favourable than grocery, is also skewed against the target demographic with 32% of adult smokers in this channel under the age of 30 and 17% under 24. 5,7% of all smokers and 4,4% of ASU30s buy their cigarettes most often in this channel.74
The Spar group is the largest participant in the channel – indeed it appears to be the largest seller of cigarettes in the country - and it alone accounts for approximately half of total cigarette sales in the channel or 5% of total national cigarette sales. The other large chains in the organised convenience channel are Pick n Pay Franchise, which, under various brand names, accounted for about 30% of sales in this channel, OK franchise, Seven-Eleven and Friendly.
Organised convenience is a franchise format, a factor that might reasonably be predicted to facilitate access to promotional opportunities. Franchise arrangements give the franchisees varying degrees of autonomy vis a vis the franchisors. Accordingly reaching an agreement with the head office does not provide anything like the prospect of compliance that reaching an agreement with the head office of managed stores does. And as the Pick n Pay northern Gauteng incident shows, powerful regional or store managers ensure that not even that structure can offer absolute certainty of compliance. The franchise relationship, by its very nature, eases the ability of smaller manufacturers to cherry pick, to carve out selected stores and enter into promotional agreements with them.75
It appears that the arrangement between the Spar and BATSA specifically conceded that each franchisee could choose whether or not to adopt the agreed incentive arrangement and it also specifically provided for the severability of the various drivers which formed the basis for the incentive payments.76 Testimony by a senior Pick n Pay executive confirmed that agreements signed by the franchise division of Pick n Pay are not enforceable on individual franchisees.77
Because of the relative autonomy enjoyed by Spar franchisees, approaches to cigarette merchandising within the various Spar stores range over a significant spectrum. For example certain of the stores use BATSA-provided merchandising furniture, while others use their own furniture or furniture provided by other manufacturers. The degree of store-level compliance with the agreement concluded between BATSA and the Spar franchisor varies widely. Hence some Spar stores use the cigarette kiosk as a catch-all for all products which, for one reason or another, they do not wish to sell from the open shelves – razor blades and condoms are items commonly sold from the cigarette kiosk. There also appears to be considerable variation regarding the use of secondary promotional material as well as smaller, dedicated secondary CDUs at the point of sale.
But of critical importance, it appears that while the trade investment agreement is concluded between the franchisor and the BATSA head office, individual stores are not thereby precluded from entering into independent agreements with other suppliers. While the evidence shows that Spar franchisor certainly encourages its franchisees to abide by the agreement and endeavours to monitor and secure compliance, the extent to which the incentive payments in the BATSA-Spar agreement are passed through to the franchisors is not clear and this will surely significantly influence the extent of franchisee compliance. Our previous experience with franchise arrangements would suggest that the pass through of supplier incentives from franchisor to franchisee is an arena of considerable contestation between the parties to such an arrangement.78
It seems that at one stage approximately 10% of Spar’s 700 plus stores had been successfully cherry-picked by JTI suggesting either that the incentive payments are not passed through or that they are not sufficiently attractive to secure anything like comprehensive compliance with the BATSA-Spar agreement. It is reasonable to assume that JTI would have cherry-picked stores on the basis of their attractiveness as sites of cigarette promotion. Indeed it appears that all of the Spar stores secured by JTI were in the urban areas, which are the only localities that JTI identified as significant from a promotional perspective.79
As regards the Pick n Pay franchise the evidence is that approximately 80% of franchisee stores adopted BATSA incentives while 20% did not. Approximately 50% of Pick n Pay franchise outlets made use of BATSA merchandising units while 20% used JTI units and 30% had their own units.80
In addition the severability of the drivers allowed individual franchisees to enter into agreements with BATSA rivals that carved out particular promotional opportunities. BATSA argues that the ability to do this was especially significant with regard to the placement of secondary promotional materials at the points of sale given that a large proportion of the CDUs used in organised convenience were not macrobats but rather pacrobats which provided little of the visibility that the applicants consider to be such a powerful element in the promotional armoury.81
In summary while independent convenience is a more important site of cigarette sales than organised grocery – and this surely enhances its importance as a site of promotion for the obvious reason that more smokers visit its cigarette points of sale – its promotional opportunities are limited by the demographic composition of its clientele which, though slightly more favourable than grocery, is still skewed against the preferred demographic target for cigarette promotion.
But the most important distinction between organised grocery and organised convenience is the franchise model employed used in the latter channel.82 While the Spar franchisor was clearly incentivised to encourage its franchisees to comply with its agreement with BATSA this did not prevent some 10% of its franchisees from entering into agreements with JTI. It is reasonable to surmise that if this number of franchisees were willing to turn their backs on the BATSA agreement with their franchise owner to the extent of entering into alternative agreements, that a great many more would have felt comfortable to deviate in a less significant and dramatic manner, for example, by altering the CDU layout or selling space for secondary displays. The long and short of it is that, given the franchise format that dominates organised convenience, BATSA’s rivals had ample opportunity to outbid BATSA and did so on a significant number of occasions. Indeed an agreement concluded with the franchisor – as opposed to the individual franchisees – could only hold if the manufacturer was willing to pay an incentive large enough to be shared between the franchisor and the franchisee. A cherry-picked agreement with an individual store had only to be large enough to incentivise the individual franchisee and so had only to exceed that portion of the BATSA incentive passed on down from the Spar head office.
We conclude then that the limited exclusive between BATSA and Spar constituted at most a moderate foreclosure of promotional opportunities. The extent of foreclosure achieved by the limited exclusive is reduced by the unfavourable demographic profile of the customer base, by the proven ability of non-BATSA rivals to undermine BATSA’s coverage over the franchised stores and by the limited ability of BATSA and the franchisor to enforce the agreement.
With respect to organised convenience and Spar in particularly we should note that the only retailer called by the applicants was Mr. Denby, a manager at the Hermanus Spar. While Mr. Denby’s witness statement suggested that he was going to testify to BATSA’s aggressive stance in the use of its ownership of the cigarette dispensing units as a means of compelling compliance from the retailers in organised convenience, his oral testimony rather evidences JTI’s passivity with respect to seizing opportunities presented to it.
While we do not want to make too much of the evidence of a single, rather inconsequential retailer, he was the only retailer called by the applicants and his evidence further indicates JTI’s apparent unwillingness to avail itself of opportunities offered. Again we must conclude either that JTI accorded little significance to the promotional opportunities afforded by this channel (and thus it would be hard pressed to simultaneously claim significant foreclosure in this channel) or else it has chosen to fight its battles in the Competition Tribunal rather than on the more testing terrain of the market.
Organised forecourts
These are the retail outlets located on petrol station forecourts. Certain of these are operated by the oil company which owns the site, others are franchised operations. Yet others – the minority it seems - are located on the forecourts of petrol stations which are not owned by the oil stations and are operated by the owner of the station. These latter fall into the category of independent convenience and are not considered here.
The data indicate that this channel is not only important from the perspective of the cigarette companies – it accounts for some 3.6% of national cigarette sales and it boasts of the highest national contribution of premium cigarette sales of any channel, apparently more than double the national contribution83 - but, in addition, cigarette sales represent a significantly greater portion of total forecourt sales than in the case of the other organised channels. The tobacco category contributes 20-30% of the turnover of the organised forecourts channel.84 It is also a significant profit generator with margins of between 19-23%.
Moreover the demographic composition of those who purchase cigarettes from organised forecourts outlets accords with the target demographic of the cigarette manufacturers and the regularity with which these young adults purchase their cigarettes from these outlets significantly exceeds the other channels within organised retail. Forecourts, together with independent convenience, have by far the highest percentage of ASU30 smokers of any of the channels with approximately 52.7% of the smokers below 30 and approximately 24.41% is below 24.85
This may also bear out the hub and spoke theory regarding HORECA venues which, as will be elaborated, are sites that are of little consequence from a sales or distribution perspective but are important sites of promotion. The hub and spoke theory postulates that young adult smokers will try a brand at a HORECA venue and then purchase a packet at a garage forecourt en route to or from the HORECA venue. It may simply establish that, as we have observed above, cigarettes are classic convenience purchases and what more convenient site for a smoker (who also owns a car of course) to pick up her pack of cigarettes than from a garage en route to or from work or entertainment. But be that as it may, the data do establish that many impressionable young smokers purchase their popular and premium brands of cigarettes on a regular basis from garage forecourts and that fact alone renders this channel potentially significant from a promotional perspective. The data clearly establish that a great many of those entering garage forecourts are young adult smokers whose clear purpose is to purchase a pack of cigarettes, regardless of the fact that the retail price is higher in this channel than in other organised retail channels.
It is not surprising then that of all the organised retail channels this has proved to be the site of greatest contestation between the cigarette manufacturers and, inevitably then, between the retailers and the manufacturers. The power dynamics between the manufacturers and the retailers are more finely balanced in this channel. They are important sites of distribution and, relatively speaking, promotion; and they are important revenue earners for the forecourt stores. Thus the manufacturers would be intent on ensuring high levels of visibility and favourable positioning; the retailers would be intent upon ensuring that ‘excessive’ promotion of a particular brand or of a particular manufacturer’s brand portfolio did not undermine the sales volumes of an important source of revenue.
This channel is also largely a franchise channel. Therefore, as with organised convenience, the individual stores enjoy a degree of autonomy. However a unique feature of this channel is that certain of the agreements governing cigarette merchandising are international agreements concluded between the multinational oil and tobacco companies. However as with the relationship between franchisor and franchisee, the local franchisor appears to enjoy a degree of autonomy vis a vis its international head office. For all these reasons, it is extremely difficult to draw a general picture reflecting the state of cigarette merchandising in organised forecourts. What can however be said with confidence is that it is a hard fought channel with each of the leading manufacturers – BATSA, JTI and PMI – robustly contesting promotional space.
While it is not easy to provide a general description of the forecourt agreements, it appears that in 2003 BATSA’s forecourt agreements (then with Shell, Caltex and Engen) covered approximately 800 of the 1250 organised forecourts. The agreements gave BATSA the right, in exchange for the usual incentive payments, to install its CDUs and allocate space and positioning of the cigarette brands, including those of its rivals. It appears however, that agreements concluded between the international oil companies and the international cigarette companies did not always square with the agreements struck at the national level. Hence, although BATSA had concluded a national agreement with Shell along the lines indicated above, it appears that during the currency of this local agreement, an international agreement was concluded between Shell and JTI which entitled the latter to 30% of the cigarette shelf space at Shell branded outlets, a share that exceeded the rate of sale of JTI brands at Shell’s South African forecourts, and, probably, in a great many other countries beside. This excessive (‘excessive’, that is, in relation to the rate of sale) allocation of space conflicted with the interests of local Shell franchisors, probably accounting, along with the incentives offered by BATSA, for the low level of local compliance with the international agreement. In similar fashion BATSA and Caltex concluded a local agreement containing the usual drivers and incentive payments while an agreement was simultaneously struck at the international level that entitled Camel to access to the ‘hot spot’ in the back wall units supplied by BATSA.
The only organised forecourt franchisee who was called to give evidence was Ms. Sally Talbot who owns and operates a Sasol forecourt in the Eastern Cape. While she acknowledged that she had not been involved in the negotiations for the agreements which incorporate the planogram, she averred that she determined the space allocations to the cigarette brands based on rate of sale and that the planogram gave general directions of where cigarettes should be positioned and to which she attempted to adhere.86 With regards to secondary units in store, Ms Talbot said that she allowed Gallaher to place a dispensing unit to promote Sobranie cigarettes on her till for some time. In addition, she referred to an e-mail she received from the Sasol Merchandising team stating that a deal had been struck to allow a secondary dispensing unit for “Camel Naturals”, and that display remained in place for some months. At the time of the hearing in May 2008, Ms Talbot confirmed that she was running another campaign for Camel Naturals and also an awareness campaign for Marlboro involving Ferrari. The evidence of Ms Talbot is that for the past four years approximately one third of the space on the shelf was occupied by non-BATSA brands.87 This seems to be confirmed by the photo of Talbot’s store.88
In about 2003 JTI and BATSA put up bids to provide cigarette dispensing units for some of Engen’s stores that were being branded independent convenience stores to be called “Engen-on-the-Go” stores. JTI outbid BATSA to supply the “Engen-on-the-Go” stores with cigarette dispensing units branded largely with Camel. Although there is a dispute as to how many of Engen’s stores participated in the “Engen-on-the-Go” program (with BATSA saying it was 600 stores and Mr Doyle, for JTI saying it was 88 evidencing the frustrating and ubiquitous extent of factual dispute)89, this opportunity undoubtedly enabled JTI to install Camel branded unitary in all participating stores which gave Camel extensive branding as well as a single extensive illuminated stock-holding area only for JTI brands.90 JTI had to pay R30 000.00 per unit with no further payment required to the stores.91
While BATSA supplied Caltex with the back wall units, BATSA, JTI and PMI paid Caltex a fee that enabled them to participate in the planogramming process. JTI has approached BP with the same offer, which has been accepted.92 In addition, after the expiry of the global agreement BATSA was allocated 85% of the space and undertook to do category management for Shell. In his supplementary witness statement Mr Potgieter stated that JTI was offered 10% of the space in Shell against payment of R600 000.00 but preferred to take 7.5% of the space against payment of R250 000.00.93 Although Doyle testified that the 10% offered by Shell was at rates JTI could not justify on commercial grounds and failed to represent JTI’s rate of sale in Shell which stands at 12%, JTI agreed to a payment to acquire 7.5% of the space.94
The experience of this channel – and similarly of the organised convenience channel – is that when the other cigarette manufacturers, particularly large multinationals like JTI and PMI, devote their resources, including their marketing and merchandising skills and experience, to contesting a channel then they are more than capable of successfully competing for access to promotional resources even against a dominant competitor like BATSA. Given that their lower market share gives them a lesser interest in expending massive resources on promotion they may opt for greater selectivity in where to place their promotional efforts. They have chosen to focus on organised forecourts and HORECA. In organised forecourts they have achieved some considerable success. And, as we shall elaborate below, this is also true of HORECA.
The real source of frustration of those who hold small shares of the South African market is that no matter how much additional space they achieve – additional, that is, to their rate of sale equivalent – or how successful they are in acquiring choice positions, the quantity of space and quality of position lost by BATSA will be limited largely by the retailers’ interest in the application of minimal category management principles in the actual context of BATSA’s market share. That is to say, the retailers will not risk out of stocks thus, in consequence of BATSA’s market share, ensuring BATSA brands considerable space. By the same token, the sheer proportion of the CDU commanded by BATSA’s market share combined with the retailers’ interest in ensuring that its best selling brands are clearly positioned, will ensure that BATSA brands are competitively positioned. Conversely, the grocers, who are clearly less concerned with the cigarette category than are the forecourts will not, despite the supine approach of BATSA’s competitors, grant BATSA more than a limited exclusive, that is regardless of the incentive offered by BATSA, they will ensure that non-BATSA brands are allocated space and that they are displayed, albeit in inferior positions.
Indeed, as BATSA has pointed out – and this does not seem to have been contested – the organised grocery and convenience channel certainly, and very probably, the organised forecourt channel, appear to go beyond the application of strict category management considerations in their approach to the listing of small brands and SKU’s. It has never been suggested to us that the retailers in any of the channels – with the obvious exception of the highly space-constrained vending machine operators – have refused to list and place a new brand or even SKU on the CDU. The applicants have sought to make something of the provision contained in many of the BATSA agreements that require the retailers in question to stock all BATSA brands and SKUs. In truth, though, this clause seems superfluous because, from most accounts, it appears that the retailers, of their own volition, insist on listing and displaying all available brands and SKUs, regardless of the manufacturer from whom they emanate. The inevitable upshot of this is that it is BATSA’s allocation of space, and particularly that of its larger brands, that is compromised because only it has a space allocation large enough to be reduced without altogether removing a brand or SKU from the CDU.
Organised Retail – a note on the ‘all or nothing’ argument, asymmetric incentives and free riding
We have already encountered the ‘all or nothing’ argument in our discussion of the organised grocery segment of this channel. It is JTI’s ‘nightmare scenario’ as articulated by Mr. Doyle, essentially the fear that because the retailer faces the prospect of losing all of its incentive payments in consequence of non-compliance with a single driver or in a single corner of a vast retail empire, a company with a much smaller market share would, in order to purchase the category captaincy, be compelled to replace all of BATSA’s incentive payments across the entire breadth of say Pick n Pay or Shoprite Checkers. This would necessitate the expenditure of considerable resources spread over a sales base significantly smaller than that of BATSA. From a foreclosure perspective this nightmare scenario only looms as a serious threat in circumstances where BATSA is able to conclude an agreement at a central level that is then imposed upon a large number of retailing outlets, where the collective ‘all’, so to speak, is considerably larger than the ‘nothing’ and so difficult for a smaller competitor to meet or to justify meeting. It is, in other words, an argument particularly applicable to the organised retail sector and so is convenient to raise now, although our reasons for questioning its validity do, in some measure, draw upon an analysis of independent convenience and HORECA, channels which are dealt with below.
In our discussion of the ‘all or nothing’ argument we concluded that, even in the grocery segment – the segment of organised retail in which BATSA’s coverage is most comprehensive - reasonably resourced competitors are able to gain incremental access to promotional opportunities. We will not revisit these access points, save to say that our examination of the other organised retail channels – convenience and forecourts – and, as is elaborated below, independent convenience, has revealed additional points of access to promotional opportunities at the retail points of sale that do not require substituting BATSA’s entire incentive package. In short the evidence shows that there is indeed a spectrum between all and nothing, and BATSA’s smaller rivals in the South African market, the most significant of which are, of course, subsidiaries of enormous multinationals, move along the spectrum and are by no means comprehensively excluded from promotional opportunities at the retail points of sale by the operation of an ‘all or nothing’ paradigm.
The fact that BATSA’s ‘small’ rivals are subsidiaries of multinational giants is mentioned advisedly. These are firms with experience of operating in a vast array of markets, many of them ‘dark’ to one degree or another. Needless to say they are possessed of huge financial resources and, certainly where those rival manufacturers referred to most frequently in these hearings are concerned, the second applicant, JTI, and the American multinational, PMI, are also possessed of immensely powerful international brands like Camel, Winston and Marlboro. In these circumstances entry is possible, difficult but possible, with immeasurably the greatest barrier to be found in the regulations prohibiting above the line advertising and sponsorship, regulations whose clear objective, and largely realised, was to eliminate cigarette promotion.
Mr. Pretorius, for the Commission, when discussing the failure of the Independent Tobacco Company, the company founded by Mr. Piet Botha, and its major brand, Matrix, asserts that it ‘cannot be possible that you must have hundreds of millions of dollars to be able to get some market share’.95 While ‘hundreds of millions’ is an exaggeration, it can and does require a very significant investment, particularly when one enters, as did Mr. Botha, with a super-premium product, at a price level appropriate to that elevated status, in a dark market, pitted against brands like Peter Stuyvesant and Dunhill, not to mention Camel and Marlboro. In order to sustain entry under these conditions, the entrant should indeed be prepared to spend very significant sums (not least on brand building and promotion) and to wait patiently for market share growth.96 After all, the applicants themselves acknowledge that the impact of promotional strategies is only discernible in the long term. Entry, particularly in dark market conditions, may then well be beyond the likes of Mr. Botha’s company, but it is not beyond the likes of a JTI and PMI determined to make serious inroads into this market, a determination which has not been demonstrated by JTI. We may indeed have encountered our first asymmetric incentive: BATSA, it seems, is more strongly incentivised to maintain its position in this market – downhill is the only way that its market share, and particularly Stuyvesant’s market share, can realistically go – than are PMI and JTI to conquer what will, inevitably, remain a modest share, at least for the foreseeable future.97
So in summary, for the likes of JTI and PMI, we do not see a stark all or nothing choice in accessing promotional opportunities. We see a spectrum and BATSA’s rivals are active, if by no means uncontested, agents in deciding where best to locate themselves on that spectrum. JTI certainly chose its place on the spectrum – we know little about PMI in this regard – by its unwillingness to put the requisite resources behind its promotional effort, even to the extent of refusing to purchase the EPOS data from the two grocery chains despite knowing that without meeting that threshold condition there was little or no prospect of influencing the retailers’ merchandising approach. In those few pockets of organised grocery in which JTI did invest in brand building it achieved some success, and in forecourts and organised convenience, and, as we shall see, HORECA, it achieved greater success in achieving improved allocations of space and position and the placement of secondary promotional material, including dedicated, secondary CDUs.
The related asymmetric incentive and free-riding issues, already alluded to above, strike us as arguments more deserving of serious consideration than the ‘all or nothing’ argument. It is argued that a major difficulty confronting the manufacturers with a smaller market share and few brands and SKUs and arising directly from BATSA’s unusually large market share and its vast array of brands and SKUs, is that even if one of the smaller market participants were to expend the resources necessary to capture and retain the category captaincy – and so acquire the limited exclusive that is the private return arising from category captaincy – the sheer size of BATSA’s market share and of its portfolio of brands will still guarantee it a major slice of the space available and place in the hot spot. While, as the northern Gauteng and certain of the forecourt examples show, it may well be an outcome significantly less favourable than that which BATSA will acquire through exercising the category captaincy itself, no retailer could permit a change in BATSA’s allocation of space and position that left it without a very significant allocation of space and a position, albeit shared, in the hot spot. To limit BATSA too severely would inevitably mean out of stocks and the delisting of BATSA brands. This, as we have seen, would be contrary to the interests of the retailers and their customers.
On the other hand BATSA is able, by exercising its limited exclusive, to somewhat reduce the space allocated to rival brands – again limited by listing requirements and the need to avoid out of stocks of rival brands – and it is, in particular, able to exclude them from the hot spot. Co-operating with BATSA not only results in the payment of a handsome incentive to the retailer, but also meets the latter’s requirement for variety and stock control. In other words the advantage that will accrue from category captaincy to a small player is disproportionately smaller than that which will accrue to a large player, or, otherwise expressed, the free ride enjoyed by the larger player as a result of the smaller player exercising the category captaincy role is disproportionately greater than that which accrues to the smaller player when the larger player is category captain. However, we are constrained to observe that if the applicants accept the arguments of their own marketing experts then they will place a high premium on being able to enter the hot spot and on being able to dilute the ‘big brand’ effect that BATSA is said to be able to achieve through the limited exclusive that it has purchased. If the hot spot is all that it is cooked up to be, then the likes of JTI should be willing to expend significant resources to get there, even if this does leave BATSA with a very significant remaining presence as a result of its superior market share.
This is an advantage that accrues to BATSA because of its market share in the South African market and its brand portfolio. It does not derive from the structure of the incentive payment and so is clearly distinguishable from South African Airways. In that instance the dominant airline structured its incentive to ensure that the marginal sale of an SAA ticket ensured the travel agent the payment of an incentive that was calculated back over the full extent of all its South African Airways ticket sale. In other words the agent was actively discouraged by the structure of the incentive payment from selling tickets on any other airline. One may envisage a similarly structured incentive payment in respect of cigarette sales. Indeed the evidence shows that some form of loyalty or market share incentive was actively contemplated by BATSA and, indeed, inserted into certain of its agreements. However, the applicants have made nothing of this and the evidence suggests that, acting on legal advice, these loyalty based incentives were never activated. Indeed had these loyalty incentives been activated the outcome of this case may have been different.
The incentive payments which we are currently examining do not reward the retailer for selling more BATSA products or penalise the retailer for selling more non-BATSA cigarettes. They simply constitute investment in promotion, which is undoubtedly costly and facilitated by the bountiful resources of a large firm, but no different in principle from the enhanced ability of a large company to sponsor high profile sporting events or produce and flight costly television or cinema advertisements. Nor is the limited exclusive a unique arrangement – it would be no different if, absent the regulations, a large cigarette company were to enter into an arrangement with the country’s key publishers and the cinema duopoly that gave it access to whatever is the equivalent of the ‘hot spot’ in the largest selling magazines or on the country’s cinema screens. The question to answer is not whether the large company is uniquely capable of affording the purchase of these promotional privileges; but rather whether it has used its resources to foreclose promotional opportunities without generating any countervailing pro-competitive gains.
We have examined the evidence of foreclosure from promotional opportunities in the various retail formats and HORECA and have concluded that there is promotional capacity available to a well resourced rival to BATSA provided that it was indeed willing to commit the requisite resources. In our view, were BATSA’s rivals to take full advantage of these and other marketing opportunities discussed below, they may well, particularly given the strength of their international brands, mount a more robust and successful challenge to BATSA. However, given BATSA’s head start and, most particularly, the regulatory restrictions, it is difficult to envisage BATSA being displaced from its dominant position in the market, and, hence, its dominant position at the POS and in the CDUs. However, the larger its competitors grow, the more difficult it will be for BATSA to resist the expansion of the limitations imposed on their exclusive.
BATSA’s Trade Investment Agreements – Independent Convenience
The omnibus term ‘independent convenience’ refers to all of those retail outlets that are not part of the organised channels; effectively they are not part of chains but are rather typically owner-managed retail outlets that sell cigarettes. These may range from upmarket delicatessens in the leafy suburbs of the metropolitan areas to farm stores on country roads to bottle stores in small towns, to spaza shops in the historically black townships. This category naturally incorporates a vast number of stores. We are told that there are some 140 000 outlets in this category.
The channel accounts for the significant majority of cigarette sales – 65,4% of turnover from all cigarette sales – although the degree of variation between the sales volumes of the wide range of individual outlets in this category is predictably vast. It is noteworthy that despite JTI’s insistence that this channel does not attract consumers of premium brands, sales of Camel from independent convenience accounts for the significant majority of the brand’s total sales. For instance, in 2007 the independent convenience channel contributed 39.35% in JTI’s total sales by volume compared to 4.6% in shebeens, 27.21% in organised forecourts, and 28.24% in grocery.98 Camel accounted for more than three quarters of JTI’s sales.99 This channel also caters for a significant proportion of young adult smokers as indicated in each of the subsets of this channel: independent forecourts have 52% ASU30 customers and 24% are ASU24; independent retailers have 42% ASU30 customers and 24% are ASU24; Spazas have 42% ASU customers and 24% are ASU24; and Bottle stores have 48% ASU30 customers and 36% are ASU24. 34% of smokers buy their cigarettes most often at the independent retail stores (including 35% ASU30 smokers).
However despite the undeniable significance of this channel from a distribution perspective, the applicants insist that it offers limited promotional opportunities. They argue that it is impossible for the trade investment staff to conclude agreements with more than a small proportion of the vast number of outlets within this category. And they argue that it would be extremely difficult to monitor compliance if trade investment agreements were concluded with more than a handful of independent convenience stores, lacking, as they do, the co-ordinating centre of organised retail. In part because of the absence of that co-ordinating centre they insist that the vast majority of outlets in this category, including specifically the cigarette points of sale, are badly organised and untidy and so inappropriate sites for promotional activity. In general they argue that the stores in this category and the customer base that they serve do not generate the sort of associations, the imagery that supports brand promotion.
These arguments are not entirely persuasive. BATSA points out that there are particular brands – Courtleigh Gold Band for example – that have made particularly significant inroads in the so-called ‘main market’ and thus are largely sold through township-based independent convenience outlets which include shebeens and spaza shops. Why would JTI not direct its promotional activities at having Camel compete with this brand, particularly as Camel already enjoys a well-established presence in this market?100 The fact – as opposed to the atmosphere created by the applicant’s counsel – is that far from Camel being the preserve of wealthy white consumers and purchased at places where wealthy white consumers purchase, the lion’s share of its sales take place in independent convenience and to participants in the main market, or, less euphemistically, to black consumers. This is consistent with JTI strategic material which identifies independent convenience to be a ‘focus channel’ for Camel promotional activity.101
However, while in argument the applicants reject the prospect of successful promotion in independent convenience, they do acknowledge that there are exceptional stores in this category that are suitable sites for promotional activity. The impression gained is that these are stores in upmarket metropolitan areas with significant cigarette sales. Having then reduced the promotional opportunities in these channels to a small proportion of the total number of outlets we find that they coincide, for the most part, with those stores that participate in one or other of the BATSA programmes – dubbed the Sprint and Vecta programmes – directed at the independent retail channel. And so again, the applicants, having narrowed the promotional opportunities to the small number of stores deemed suitable for promotional activity, allege that BATSA has appropriated the significant majority of available promotional opportunities in the channel.
The Vecta and Sprint Programs are BATSA’s trade and merchandising schemes directed at the independent convenience channel.102 Not all stores in this channel qualified for participation in the Vecta or Sprint programs. The qualifying criteria were based on cigarette volumes sold, geographical location, outlet type and consumer profile. Vecta was aimed at superior quality outlets while Sprint was directed at lower quality outlets that did not qualify for Vecta.
The Vecta programme was launched in 2002 and it targets the independent convenience stores that sell higher volumes of cigarettes in priority urban areas and metropolitan areas. It particularly targets stores in Cape Town, Durban, Johannesburg, Pretoria and Bloemfontein.103 Mr Potgieter, a BATSA witness, averred that BATSA focused on metropolitan stores because of the cost entailed in taking value added services, the incentive to participate in Vecta, to stores from outlying areas and that it would be difficult for the retailers to come from those outlying areas to attend the sessions in the centres.104 While BATSA classified the independent convenience stores into Superior Quality (“SQ”), High Quality (“HQ”), and Normal Quality (“NQ”), Vecta seems to have been more targeted at the SQ stores and HQ stores which had the desired volumes and right profile of consumers. There has been a dispute concerning the number of participating stores in the Vecta programme, BATSA argues that approximately 2000 stores were participating in Vecta in 2007.105 JTI, on the other hand, concluded after analysing additional data from BATSA, that the number of stores participating in Vecta may well be less than a thousand.106 JTI further argued that the universe of independent convenience stores appropriate for trade marketing is decidedly small.
BATSA designed specific drivers for the Vecta programme which include availability, visibility, merchandising, performance and pricing.107
Availability was assigned a maximum of 200 points, of which 150 points are available for ensuring that nominated BATSA brands are available for sale at the point-of-sale at all times and that no out of stock situations are identified when the representative visits the store. A further 50 points are available for compliance with BATSA’s product merchandising guidelines.
Visibility is also assigned a maximum of 200 points. This allocation is directed at rewarding retailers for using BATSA merchandising units as the primary visible cigarette merchandising unit in the store. A retailer would receive a maximum of 25 points if the unit had been kept clear of non-tobacco products. 50 points would be allocated to a retailer who followed BATSA’s product merchandising guidelines by grouping its drive brands together. The 75 points for “product visibility” can only be earned if all variants of the key BATSA brands like Dunhill, Peter Stuyvesant and Rothmans are displayed in a face-on position on a visible shelf, while all other cigarette brand variants, including those of BATSA’s competitors, are displayed in an end-on position.
Compliance with the pricing driver attracts a maximum of 250 points which were earned on a sliding scale depending on the proximity of the retailer’s selling price to the BATSA recommended retail price. BATSA argued that the objective of incentivising retailers to drop prices of BATSA products is to grow the category margins and improve the business both for the retailer and for BATSA.108
Performance has a maximum of 150 points that can be allocated to a retailer. In this category BATSA incentivises retailers to sell BATSA products. 80 points are available for “market share performance” and are awarded to those retailers for which BATSA brands accounted for at least 93% of total cigarette sales. 60 points are awarded to retailers if 8% or more of their total monthly BATSA cigarette sales are comprised of BATSA premium brands, and a further 60 points if 84% or more of their total monthly cigarette sales is comprised of BATSA popular brands. Again the loyalty and market share incentives do not seem to have been enforced. It is unclear whether BATSA re-assigned the points available in respect of these unenforced drivers or whether it simply made the payments despite non-compliance.
Retailers can obtain a further 200 points for participating in and implementing cycle activities initiated by BATSA. This usually takes place when BATSA wishes to place secondary material as part of promoting the launching of new product or innovation.
Under the Vecta scheme, BATSA offers participating retailers points which can be redeemed for a range of rewards. For example Vecta members are entitled to attend an annual conference for retailers, an annual get away convention for the top points earners, a quarterly business bulletin providing information regarding trends in retail and business advice, and preferential business partnerships with Vecta business partners like ADT Security.109
JTI has argued that the Vecta program’s rewards are not “objectively justifiable” as they do not relate to true efficiency gains on the part of the supplier but are directly aimed at ensuring loyalty to BATSA on the part of the retailer and driving the sale of BATSA brands, rather than generating efficiencies in BATSA’s business. Arguably, this is made manifest particularly in the “visibility”, “merchandising”, and “performance” drivers. The Vecta Guiding Principles document states that:
“Retail Relationship Marketing entails a marketing programme designed to establish long term loyalty with identified retailers where member retailers are offered a range of business benefits, achievable through driving BATSA’s business at the point of sale”.110
BATSA on the other hand argues that JTI’s desire to gain access to what it says are desirable stores, is a tacit acknowledgement of the effectiveness of the Vecta programme in “uncluttering” the environment in independent convenience stores thus ensuring that they have become desirable sites of promoton. In other words, BATSA argues that it is precisely the Vecta programme that has created an environment conducive to the promotion and distribution of cigarettes upon which JTI now wishes to free-ride.
For stores that do not meet the criteria for participation in Vecta, BATSA introduced the Sprint program which was launched shortly after Vecta and has between 200 and 300 stores participating in the programme at any given time.111 Sprint required compliance with the availability and visibility obligations. As discussed under Vecta, the visibility driver requires that certain BATSA brands be displayed in a face-on position on a visible shelf. The retailer is required to follow the BATSA product merchandising guidelines and has to make the BATSA units the primary visible cigarette merchandising unit in store.
Under the Sprint programme retailers are offered cartons of BATSA cigarettes and entry into a monthly cash prize draw in reward for compliance with the BATSA guidelines.112 Retailers would normally participate in the Sprint programme for a period of 3 months and have a month’s break from participation. We are asked to prohibit the Vecta and Sprint programmes on the grounds that they are exclusionary, effectively that they foreclose access by BATSA’s rivals to promotional opportunities in independent convenience.
We should say at the outset that the applicants are so disparaging of promotional opportunities in this channel that, on their own argument, it would be difficult to make a finding of significant foreclosure on the basis of exclusion from promotional opportunities in this channel. In particular they argue that promotional opportunities are beyond the capacity of manufacturers with low market shares and limited resources, particularly compliance-monitoring resources. Certainly if JTI’s inactivity in this channel is a measure of its inability to promote its brands there then the upshot of voiding the Sprint and Vecta programmes would appear to be a brand promotional vacuum and a category promotional vacuum in the independent convenience channel, a particularly barren level playing field. That is to say, even if we were to limit BATSA’s promotional activities in this channel by voiding the Sprint and Vecta agreements, the applicants, on their own argument, would not be able to engage in promotion themselves – promotion is, on their own argument, limited by the character of the channel rather than by BATSA’s exclusionary activities.
However even if we were to accept the applicants’ winnowing down of any promotional prospects to the stores that participated in the Sprint and Vecta programmes, we would still resist acceding to their request to prohibit these programmes. Firstly, as just indicated, there is no evidence that these programmes could remotely account for significant foreclosure. Secondly, the incentives offered in order to secure compliance are relatively insignificant and certainly not beyond the resources of another cigarette manufacturer, none of whom appear to have attempted to compete for whatever limited promotional opportunities this channel offers. Thirdly, the rewards offered for compliance with the specified drivers appear, in the main, to constitute what may be best described as ‘freebies’ for small businesses and removing them at the behest of a public authority like the Commission or a large multinational like JTI appears particularly churlish. However to the extent that they do amount to anything, the incentives would fall into the category of technical assistance for small business. If anything these would appear to be manifestly pro-competitive.113
We must though emphasise that this channel is not only the largest retail distributor of cigarettes but it is the largest seller of Camel. The fact that next to no effort has been made by JTI indicates to us that either it is indeed impossible to actively promote a brand through this channel (in which case a claim of significant foreclosure cannot be sustained) or that JTI has adopted a thoroughly supine stance with respect to promotion in the channel – not unlike the stance that it adopted in organised grocery – and has preferred to rely on regulatory intervention in order to eliminate as much competitive activity at the point of sale as possible rather than engage in the costly and arduous business of robust competition.
Organised Retail and Independent Convenience – summary of conclusions
We are not persuaded that the POS in organised retail and independent convenience is, in general, an important source of cigarette promotion, although clearly some channels, for example, garage forecourts, are more effective than others. As we have already elaborated at length, we are of the view that the fact that smokers know that the vast majority of cigarette retail outlets stock all brands and SKUs, the fact of brand loyalty, the fact that cigarette purchasers do not have to search the shelves for their product but simply ask an attendant, and, in respect of grocery and organised convenience certainly, the fact that the target demographic at whom promotion is aimed does not purchase cigarettes in certain of these channels, all lead us to conclude that the promotional significance of the cigarette points of sale is significantly exaggerated.
The applicants have attempted to persuade us of the significance of POS promotional activities through, firstly, the assertion of general marketing theories that are, we believe, of limited applicability to the peculiar circumstances of cigarette retailing and, for the most part, serve to negate rather than promote competition. Secondly, the applicants have sought to persuade us to infer that BATSA’s willingness to devote considerable resources to POS promotion in the retail channels establishes that BATSA itself believes this to be an important promotional site.
We have stated our view on category management principles at length and will not repeat these here. We should though note that it has been suggested that marketing advantage accrues to BATSA from the bill board or ‘big brand’ effect that is created as a result of blanketing the CDU ‘hot spot’ with a single brand. This is lent credence by evidence suggesting that BATSA has made some effort to split the displays of other brands. There may be some validity in the big brand effect argument. It is impossible to evaluate – even qualitatively – its impact or to remedy it without imposing a highly invasive set of regulations on the retail POS.114 Moreover, as noted above, if the ‘big brand’ effect is so significant, then the gains to be derived from a smaller, though well resourced competitor like PMI buying its way into the hot spot and thus reducing the big brand effect are increased.
Regarding the inference that we are asked to draw from BATSA’s willingness to spend significant resources at the POS, our view is that this is clearly a market of overriding significance for BATSA and that it has elected to devote significant resources, even to err on the side of overspending, in order to defend it is unremarkable. Certainly the assessment of experienced retailers like van Vuuren of Pick n Pay and Ackerman of Shoprite Checkers is that the cigarette companies, and BATSA in particular, are overspending on POS promotion, that is to say that the payments made for their POS privileges are not justified by the extent to which organised grocery customers are influenced in their brand choice. As we shall see, in the case of HORECA, which we do believe to be an effective promotional site, a witness with considerable experience in HORECA management stated quite bluntly that he believed the cigarette companies to be overpaying for its HORECA promotions.115
Secondly, as noted above, we believe that the scale of payments for securing promotional privileges represents the extent of the retailers’ countervailing power vis a vis the manufacturers.
Thirdly, while BATSA has undoubtedly spent considerable sums on POS promotion, it has, as we shall show, also spent considerable sums on direct marketing, of which HORECA is a particular case. Our interpretation of the evidence is that, absent the ability to exercise social and peer group pressure through above the line advertising and sponsorship, direct marketing, including, but not limited to, the hosting of events at HORECA venues, is likely to be the most effective form of promotion in the cigarette category. With the important exception of HORECA it is not alleged, and nor could it be, that BATSA has somehow contrived to foreclose direct marketing opportunities.
Fourthly, given BATSA’s market share it does retain a clear incentive to promote the category. As has been frequently pointed out BATSA’s huge market share presupposes that its share of growth in the category is similarly significant. Indeed it is this feature of the market that makes BATSA the obvious category manager – it cannot afford not to promote the category. Accordingly it should be acknowledged that a significant part of its trade marketing spend is rationally directed at promotional activity from which the entire category, including BATSA’s rivals, benefits. For example, the provision of free merchandising furniture and the incentive payments designed to keep the CDU free of non-tobacco products certainly have a significant category promotion aspect. The conduct which is in question – the limited exclusive with respect to space, position and secondary display - represents BATSA’s attempts to secure a private advantage from a POS into which it has invested without the ability to appropriate a private return, at least without the ability to ensure that part of the return is not appropriated by its rivals.
Accordingly, the suggestion that BATSA’s trade investment spend infers its calculus of the importance of the POS as a site of brand promotion must, if it is to be at all useful, be based on that part of its trade investment spend that is directed at promoting its own brands, and should not include that part of its trade investment spend that is directed at supporting the category. It is, of course, extremely difficult to separate these. However there is, in our view, no doubt that a very significant part of the BATSA’s trade investment spend promotes the category, from which BATSA’s rival’s benefit, at least to the extent of their share of the category.
BATSA’S Trade Investment Agreements – HORECA
There are two promotional opportunities afforded by HORECA venues. The first is at the cigarette distribution sites in these venues. These venues increasingly distribute cigarettes through vending machines placed in the premises by agreement with one or other of the vending machine operators (VMOs). It is alleged that BATSA incentivises both the VMOs and the venue owners in order to secure privileged space allocations and positioning in, and branding opportunities on, the vending machines.
Secondly, the cigarette companies host promotional events at which direct marketing and other methods that brand the venue are employed in order to encourage trial of the brand being promoted and, more generally, to associate the venue and the type of clientele it attracts with a particular cigarette brand. JTI and the Commission allege that BATSA is entering into exclusive arrangements with venue owners, the coverage of which effectively forecloses its rivals from promotional events at HORECA venues.
While the applicants have sought to link these mechanisms of promotion – that is, the VMO display and the promotional events – they can clearly be distinguished for the purpose of assessing the extent of foreclosure, if any, arising from BATSA’s alleged conduct. We will explore the linkages between the vending machines and the promotional events where appropriate.
Vending Machines
It appears that while there are still many instances where cigarettes are sold from behind the bar or at the till, vending machines are increasingly the preferred mechanism for selling cigarettes in upmarket and midmarket HORECA venues, that is those venues which, as we shall see, are deemed appropriate for promotional purposes. Cigarette vending machines are not owned by HORECA outlets but by one of a number of vending machine operators who own, operate, install and maintain vending machines. The preference of HORECA venue owners for vending machines over other forms of distribution relates to the threat of pilferage and the desire to avoid having to trade in a product that is not core to their businesses, particularly as it represents a negligible portion of the revenues of any single venue. Indeed, as already indicated, sales of cigarettes in HORECA venues represent 15% of total cigarette sales of which vending machines account for just 0.7%.116
The South African cigarette vending market is highly concentrated with five VMOs accounting for over 93% of cigarette sales by volume and 83% of cigarette vending machines.117 Mr Ari Zelezniak of Vendomatic and “Magic Vending”, the largest of the VMOs, gave evidence before the Tribunal. He testified that the VMOs run their businesses as independent operators and not as agents of cigarette manufacturers and wholesalers, that is, they are retailers in their own right. As with any retailer they derive their profits from the sale of cigarettes from the margins that they charge over the wholesale purchase price. They charge higher prices than those charged in any of the other channels. The VMOs contract with the venue owners to put their machines in the premises of the venue owners and pay them a commission on each packet of cigarettes sold.118 However, as we shall show, the lion’s share of the revenue earned by the VMOs is derived from the incentives paid, in exchange for space, facings and advertising materials in and on the vending machines.
When considering the impact of vending machines from a promotional perspective it is important to bear in mind that HORECA venues, unlike the other retail channels and particularly the convenience channel, are not visited for the purpose of purchasing cigarettes. The overwhelming majority of the clientele of HORECA venues will have purchased their cigarettes before entering the venue. Indeed, if the applicants are to be believed, should a HORECA visitor require cigarettes for a long night out there is a strong possibility that she will have purchased her cigarettes from a garage forecourt on the way to the venue. This is certainly a clear option for all who have reason to doubt that their chosen brand will be available at the space constrained vending machines in the HORECA venue.
It is important to appreciate that the vending machines are highly space constrained relative to the retail macrobats and, certainly, relative to the number of brands and SKUs available in the cigarette market. While the VMOs have a wide range of machines available, many of the machines available and installed in HORECA venues only have 5 columns and 8 facings, with the largest machines having 12 columns and 24 facings. In 2001 Zelezniak had less than 3000 machines and approximately 900 were five column machines, between 400-500 eight column machines of one variety and 400 of another variety still having eight columns, and approximately 400 twelve column machines. Therefore in HORECA, unlike in the other retail channels, there is inevitably a listing issue, that is to say a large number, indeed the overwhelming majority, of the brands and SKUs that find their way onto the retail CDUs, do not appear in the vending machines at all. Inevitably then this limited space is robustly contested. However we are of the view that the primary contestation is over limited distributional space and this does not impute promotional significance to vending machine display.
Not only is it common cause that the vast majority of smokers in a HORECA venue bring their cigarettes with them, it is also so that despite the presence of a vending machine, a HORECA client may request cigarettes from a waiter who would then get the cigarettes from the vending machine or, if the vending machine did not carry his chosen brand, the smoker would then go to a nearby café or forecourt to purchase the required brand or request a waiter to go out and make the desired purchase. Slightly ridiculous as they are to recall, much was made of these points which appear even more trivial in their recounting than they did in the hearings.
But, trivial or not, BATSA insisted, at some considerable length, that it was perfectly reasonable to expect that either the smoker or the waiter at a HORECA venue would be willing to leave the venue for the purpose of purchasing a brand that was not available in the vending machine, backing its learned arguments with evidence purporting to demonstrate that at the HORECA strips most frequently cited in evidence – Camps Bay and Long Street in Cape Town, Melrose Arch, Melville and Rosebank in Johannesburg – alternative sources of cigarettes were readily available. The applicants – or, rather, their legal representatives - poured scorn (at equal length) on the notion that any client would leave a HORECA venue, insisting that the temptation to try a new brand was the most likely outcome of the disappointed smoker’s frustration at not finding his chosen brand in the vending machine.
However what this proliferation of conjecture bolstered by anecdote does reinforce is the inescapable fact that the vending machines in HORECA venues are not frequently visited. The data demonstrate that few cigarettes are sold, it is common cause that most HORECA clientele would enter the venue clutching their cigarettes, if cigarettes are bought in the venue it is wholly possible that it is the waiter who effects the purchase and, while we are regrettably unable to make a definitive finding on whether or not the client would be more willing to leave the venue for a few minutes rather than try a new brand, there are clearly other purchasing options available. Given all this we may confidently conclude that the display on the vending machine does not amount to a significant source of promotion for the simple reason that not many people have much reason to view it. This is bolstered by the evidence suggesting that the vending machines are generally not prominently located in the HORECA venues. Certainly in those HORECA venues where there are vending machines but no complementary promotional activity – and this would describe the overwhelming majority of HORECA venues – it is our firm view that the vending machines play an extremely limited role, if any, in promotion. They serve purely a distribution function and it is the constrained size of the vending machines that create the limits on the range of cigarette distribution in the HORECA channel.
As might be expected given the scarcity of distributional space, the cigarette manufacturers have entered into a variety of agreements with the VMOs and, at times, with the venue owners themselves. BATSA stands accused of providing various incentives in exchange for the VMOs carrying their brands in the vending machines and, as with the other retail channels where listing is not an issue, for using these incentives to secure preferential allocations of space and position in the vending machines as well as the right to place its branding material on the vending machine.
It is particularly difficult to understand precisely the various contractual arrangements that govern the relationships between the VMOs and the various manufacturers, the VMOs and the venue owners, and the cigarette manufacturers and the venue owners and the interfaces between the various agreements. What does seem reasonably clear is that, as with the retail CDUs, even machines that were contracted by the VMOs to BATSA, were not necessarily the exclusive preserve of BATSA and that, for the most part, provision was made for non-BATSA brands and/or for those brands that the venue owner required to be placed in the vending machine (the so-called ‘owner’s choice’ column). BATSA concluded agreements with HORECA venue owners in terms of which BATSA would get preferential access for availability and distribution of BATSA products as well as consumer related contact. Certain of these agreements clearly gave BATSA exclusivity in the vending machines.
Mr. Zelezniak demonstrated his commercial acumen, and, as the largest player in the preferred retail channel in HORECA, his countervailing power vis a vis the various manufacturers, by entering into multiple agreements with the three manufacturers. By 2004, BATSA had contracted approximately 6000 machines with VMOs.119 Zelezniak confirmed that in 2004 BATSA contracted 2571 of his machines out of a universe of 3126 with JTI only getting 29 machines plannogrammed according to its requirements.120 Later however he averred that a third of his machines were not subject to any agreed planograms.121 However, in 2001 JTI had entered into an agreement with Zelezniak for 20% of the column space in his vending machines.122 Ultimately the JTI agreements appear to have covered some 2500 of Zelezniak’s machines. Therefore over the same period Zelezniak concluded agreements with BATSA in terms of which BATSA paid a large amount of cash and an allocation of free cigarettes in exchange for a limited exclusive over certain of his machines and JTI entered into an agreement guaranteeing them minimum column space in a great many of these machines. PMI also appears to have entered into agreements with Zelezniak. In summary, at one time after 2001 Zelezniak’s received some R1,8 million each month from BATSA and in the same period he garnered some R1 million each month from his agreement with JTI.123
Note that it appears that JTI’s agreement was focused on the number of columns, as opposed to facings, that it received. This suggests a primary concern – on JTI’s part at least - to avoid out of stock situations, and confirms our view that the vending machines were viewed primarily as distribution channels rather than promotional sites.
This already murky water is further muddied in those venues with which one or other of the manufacturers have entered into an exclusive promotional agreement. These are examined below. Suffice to say that where a manufacturer had entered into a costly agreement that effectively sought to superimpose a cigarette brand on one or other desirable HORECA venue, the agreement appears, predictably, to have extended to that particular brand’s domination of the vending machine. However even in these cases, it does not appear that the vending machine was necessarily exclusively reserved for the brand or manufacturer in question. In the case of one such venue, Ignite, which PMI had secured, at considerable cost, as a Marlboro promotional venue, it seems that 80% of the vending machine was reserved for Marlboro.
Much was made of evidence showing that it is in this channel in which Camel’s share appears to have declined most significantly, from 23% in 2000 to 16% in 2004, with BATSA posting an increase in its share of sales from 77% to 83% in the same period. JTI’s economist attributes this decline to BATSA’s agreements with the “most desirable venues”.124 This cannot be true. The agreements covered a miniscule proportion of the HORECA venues and so the promotional agreements cannot account for a decline of sales in the HORECA channel of this scale. This interpretation was challenged with BATSA arguing that Camel’s was losing significant share to Marlboro rather than to the BATSA brands. While it was conceded that Marlboro was indeed eating into Camel’s market share, Mr. Baker suggested that this was because of the limited amount of space available to non-BATSA brands – particularly in the 5 slot machines – where BATSA’s agreements effectively ensured that either Camel or Marlboro, but not both, were available in the vending machines. However this argument is weakened by other evidence that showed that while Dunhill and Peter Stuyvesant were losing market share to Marlboro due to switching, it was Carmel that was losing the largest share to Marlboro among white smokers between 18 and 44.125
Even if we were to conclude that BATSA had acquired some or other significant promotional advantage from its agreements with the VMOs, the impact is completely overshadowed by the major competitive gains attested to by Mr. Zelezniak, gains, moreover from which the entire category benefits handsomely. Zelezniak’s unchallenged testimony is that fully two thirds of his revenue is attributable to incentive payments from cigarette manufacturers with the remaining one third accounted for by cigarette sales.126 He testified further – and the bald figures support this averment - that the incentive payments are the basis for the significant expansion of the vending machine format of distribution in HORECA.
In other words the growth of the distribution mechanism, strongly favoured, and for good reason, by the HORECA owners is accounted for by the incentive payments made predominantly, if by no means, exclusively, by BATSA. In the absence of vending machines there is no doubt that the cigarette category in those HORECA venues that rely upon vending machine distribution would be relatively poorly managed. The insignificant quantity of sales in these venues would not justify expenditure on costly CDUs and so, absent CDUs or vending machines, cigarettes would be sold, and probably not displayed at all, from behind the bar or at the till.127 Even if we were to accept the claims made by the applicants regarding the promotional gains said to flow to BATSA as a result of the incentive payments, when these are placed alongside the pro-competitive gains that are manifest in the very existence of the cigarette vending machine industry, there is nothing further to be said about this aspect of the case against BATSA. It has no merit whatsoever.
We note too that the vast majority of cigarettes sold in HORECA are not sold through vending machines or at upmarket venues but through shebeens in the historically black townships. We are constrained to observe again that JTI may garner better returns from promoting Camel in these venues – of which we heard near to nothing - and in the ‘main market’ in general, than from fighting legal battles over promotional opportunities in channels that, in the best of circumstances, appear to offer little capacity for promotion.
We should add that the applicants’ case is further weakened by the vastly exaggerated claims made by certain of their witnesses regarding the unavailability of Camel in certain important outlets, indeed in entire entertainment precincts, for example Melrose Arch, Rosebank, and Melville. These averments made by Mr. Makola were persuasively countered by retailers from those precincts who gave evidence. Mr Doffay who managed BJ restaurant/club in Melrose Arch for the past 3 years confirmed that all vending machines in the area offered brands other than BATSA brands, and those brands include Camel.128 This contradicted Makola who had testified to the unavailability of Camel. Mr Chiki, who has managed Nino’s restaurant in Rosebank for the past 8 years testified that Nino’s has always sold Camel and Marlboro (after its re-entry into SA), and has never had an agreement with any cigarette company.129 Concerning Melville, Mr Boshielo, the manager of Xai Xai restaurant testified that the outlet has always stocked Camel130 and that it hosts Camel parties131 since 2004/5.132 Mr Makola had averred that JTI was ousted from the News Cafe chain as a result of BATSA’s strategy,133 an allegation that was refuted by Mr Van der Westhuizen of News Cafe who stated that they stock Camel and Marlboro in their stores as BATSA,s agreement does not require total exclusivity.134
Promotional Events
We are persuaded that the branding by the cigarette manufacturers of selected HORECA venues and the holding of promotional events at these venues are important promotional mechanisms. In other words, we are persuaded that peer suggestion and pressure is an effective form of promotion. Absent above the line advertising and sponsorship, the most efficient, if by no means the only, mechanism through which peer pressure can be exerted in an organised and systematic manner is through associating cigarette brands with desirable entertainment venues.135 Predictably then the right to superimpose the cigarette brands on these venues, principally through the staging of promotional events, is hotly contested terrain, and, as we shall show, by no means terrain on which BATSA is unchallenged.
There are of course thousands of HORECA venues. However it appears that only a select number of these are deemed suitable sites for the hosting of promotional events. BATSA employs a detailed classification in order to rate the suitability, from a promotional perspective, of the thousands of HORECA venues. Key criteria are the number of ASU30s who frequent the venue, the location of the venue and particularly the existence of a HORECA ‘hub’ or cluster of suitable HORECA venues – for example Melville and Melrose Arch in Johannesburg and Long Street and Camps Bay in Cape Town - and, critically, the number of ‘opinion formers’ whom the venues count amongst their clientele. It is evident that cigarette brand promotion is directed at associating a brand with particular individual and social characteristics and lifestyles – rich and glamorous, suave and worldly, rugged and individualistic, feminine. Accordingly, it appears that representative HORECA venues are sought out for promotional purposes and so an upmarket sports bar may be used to promote a particular portfolio of brands while an upmarket night club or cocktail lounge may be a more suitable venue for the promotion of other brands.
Utilising these criteria BATSA classified the Horeca outlets into normal quality (“NQ”) outlets, high quality (“HQ”) outlets and superior quality (“SQ”) outlets. Adopting a similar classification, the applicants allege that BATSA has entered into exclusive agreements with the most attractive venues, thus foreclosing access by their rivals to an important promotional channel.
Again the approach taken by the applicants renders them vulnerable to the argument that they have defined the HORECA universe of promotional opportunities by reference to the category of venues at which BATSA has chosen to direct its campaign for hosting promotional events thus ensuring that foreclosure is a self fulfilling prophecy. But let us, at least for the sake of argument, put this to one side. The evidence is that even if we accept the narrow universe of opportunities identified by the applicants and by Mr. Botha, the owner of the failed Matrix brand, as suitable for promotional events, it is clear that BATSA, on the one hand, and JTI and PMI, on the other, have successfully contested for promotional opportunities within the universe of suitable HORECA venues. All the cigarette majors have been prepared to pay significant sums for exclusivity in these venues and while BATSA has won many of these contests, it has periodically lost significant battles. Moreover, the duration of the contracts and, indeed, of the very life span of many of the most attractive venues, is sufficiently short for the battle for most of the desired venues to be periodically joined and re-joined.
For example, JTI and PMI competed vigorously with BATSA for HORECA promotional venues in the Camps Bay hub. JTI won a bidding war with BATSA for an exclusive arrangement with Caprice. In 2006 JTI paid a fee of R300 000 for its exclusive arrangement with Caprice; in 2007 this was raised to R450 000.00.136
For its part, PMI managed to secure Ignite, a premier venue in Camps Bay, at the beginning of 2006. It thus became a Marlboro branded venue. In 2007 PMI increased its offer by 20% in the face of BATSA competition.137 In addition to the payment of a significant fee. Mr Parker, the operational director of Ignite, testified that PMI had contributed to renovations of the club. In addition to securing the exclusive right to use the venue for Marlboro promotional events, PMI was entitled to set up a Marlboro kiosk outside the VIP entrance, to exclusive distribution of Marlboro from behind the bar counter and to an 80% share of the vending machine space for Marlboro. Parker stated that Marlboro concluded similar Horeca contracts with 2 large night clubs in Cape Town, namely Voodoo Lounge and Chrome – the latter was won from BATSA - as well as Patapata, a well known club in Gugulethu. Galaxy, another large Cape Town club, was also lost by BATSA but this time to JTI. Parker, who was intimately familiar with the Camps Bay hub, identified 16 venues in the locale, of which only 3 were contracted to cigarette manufacturers.
Note that, in contrast to the trade investment agreements in respect of the retail POS which we have characterised as limited exclusives, the significant agreements regarding the staging of promotional events at the most attractive HORECA venues are properly characterised as exclusive. That is for the duration of the agreement, only the cigarette company that has entered into the agreement is permitted to ‘brand’ the venue.138 As already noted, the exclusivity did not necessarily extend to vending machines, although the cigarette company that won the promotional contract does seem to have also secured highly preferential space and positioning in the vending machines. PMI, for example, achieved 80% coverage or space for Marlboro in vending machines in its exclusive outlets like Ignite. While exclusivity clearly enhances the likelihood of significant foreclosure, from a marketing perspective it makes intuitive sense that, particularly in respect of those venues that are most attractive from a brand promotional perspective, these agreements should be exclusive given that the promotional strategy consists not merely in the hosting of a series of discrete events, but rather in the embedding of an association between the venue, on the one hand, and the cigarette brand, on the other. In other words the point is to brand the venue as a Marlboro venue of a Dunhill venue.
In summary, we do not accept the applicants’ contention that BATSA’s rivals have been foreclosed from access to promotional opportunities in what have been designated suitable HORECA venues, even if one accepts, and we do not, the BATSA SQ venues as representing the totality of suitable venues. We do not believe that the vending machines are particularly attractive promotional channels and even if we were to accept this, it is clear that BATSA’s rivals have successfully contended for promotional opportunities in this channel. Moreover, the evidence clearly shows that were it not for the incentive payments made by the cigarette companies to the VMOs, this means of distribution and limited promotion would not exist at anything like its present scale of operation with the alternatives – behind the bar counter sales – infinitely inferior from both a distributional and promotional perspective.
We do, on the other hand, believe that the branding by the cigarette manufacturers of desirable HORECA venues and the staging of promotional events at these venues are important mechanisms of promotion. Here BATSA’s rivals appear to have successfully contested for key HORECA venues. Moreover it appears that even if a narrow universe of suitable venues is accepted, there are many attractive venues that remain available for promotional use by the cigarette companies, included venues that cater predominantly to the ‘main market’. Note too that in the branding of selected HORECA venues the relatively small market share of JTI is not confronted by the relatively large market share of BATSA, but rather the market share of a single brand (like Camel) is confronted by the market share of another single brand (usually Marlboro or Dunhill).
We are, in general, struck by the exaggerated claims made by the JTI witness, Mr. Makola, regarding the exclusion of BATSA rivals from key HORECA hubs.
Alternative marketing mechanisms
We note finally in our discussion of foreclosure that this case has been cast as one solely concerning promotional opportunities at the retail POS and the holding of promotional events at selected HORECA venues. We are indeed told that these represent the only opportunities for boosting the sale of cigarettes and the market shares of the participating manufacturers. Where the retail POS is concerned, the evidence and argument has focused entirely upon the CDU display and, to a far lesser extent, on the use of secondary promotional materials. We have determined that the limited exclusive enjoyed by BATSA at many important retail points of sale does not constitute significant foreclosure. Nor have the exclusive agreements concluded with HORECA venues precluded their rivals from bidding for these or other venues.
We should, however, record briefly, that promotion, cast in the terms that the applicants in this matter have chosen, is not the only mechanism for boosting sales and market share. These alternative mechanisms barely rate a mention by the applicants, fixated as they are on POS and HORECA promotional opportunities. They are dealt with in somewhat greater detail by BATSA. We simply flag them as possible alternative marketing mechanisms and, as such, factors that partly account for movements in market share. In particular we refer to price, to changes in the extrinsic features of the brand, particularly packaging, and to direct marketing.
As regards price competition, there is one particular episode immediately prior to the introduction of the dark market regulations in which JTI engaged in a major effort to iron out anomalies in the pricing of Camel. It appears that, overall, this involved a significant reduction in the Camel price that, once the exercise had been completed, aligned the price of Camel with that of Dunhill, a BATSA drive brand and major direct competitor of Camel. This much is conceded by Mr. Oscar Makola, a leading JTI witness.139 This price reduction resulted in a significant spike in sales of Camel and in its market share which then reached a plateau once the relative prices of Camel and Dunhill had been brought into alignment. This correlation is, it seems, conceded by Mr. Baker, JTI’s economic expert.140 We are satisfied that this clearly establishes the impact of price, and particularly the price of a given brand relative to that of a directly competing brand, on demand.
We must however add an important caveat which is reflected in Makola’s evidence, namely that this price realignment was complemented by an intensive above-the-line campaign informing consumers of this change in the relative price of Camel. Given the regulatory proscription on above-the-line communication it is difficult to imagine a shift in relative prices impacting on demand to quite the same extent as it did in the period in question.
We should also note that the applicants insist that the positioning of Camel and Marlboro - both premium brands – alongside the VFM brands is likely to, indeed is intended to, confuse consumers by suggesting that both are relatively expensive brands, when in fact their prices compare favourably with their premium peers. This may be so although we observe that, as far as we are able to tell, prices are not prominently displayed on the CDU.
Our conclusion here is that while the regulations undoubtedly create obstacles to engaging in price competition, it should not be beyond the ken or the resources of a major cigarette manufacturer to devise means of communicating to consumers that it is engaging in price competition. If JTI used its opportunities for secondary promotion – many of which, as we have pointed out, are woefully underutilised – as well as direct marketing opportunities to communicate price competition, then we have no doubt that this would get through to consumers. Given the highly positive experience that it has already had of price competition, we would have expected a JTI determined to increase market share to take another stab at this. Of course if JTI prefers to use BATSA pricing as an umbrella for its own pricing, the better to share in the monopoly rents that it alleges are accruing to BATSA, then it will not engage in price competition. But then it should not expect competition law to rescue it from the form of competition – promotional competition at the POS – that is its rival’s preferred terrain of competition. In other words, BATSA engages in robust POS promotional competition. JTI finds it difficult to compete on BATSA’s terrain. Its response is to come before the competition authority in order to prevent BATSA from competing on its chosen terrain, in the process eschewing alternative forms of competition, such as price competition.
Much the same may be said about another acknowledged form of competition, namely changes in the product and particularly in the packaging and other forms of presentation, often referred to as the re-launching of a brand. Again it seems that JTI largely eschewed this form of competition. As to the product itself, it was not until 2005 that Camel introduced a new SKU – Camel Super Lights – and it never introduced a menthol variant. Similarly the Camel pack was not redesigned until 2005.141 And again it seems that on the one occasion that it did launch a re-designed pack that this was reflected in a spike in demand.142 As with price competition, it should be acknowledged that the impact of a brand re-launch will be limited by the dark market regulations as well as by a combination of Camel’s relatively small market share and BATSA’s purchase of a limited exclusive which conspire to deny it the ‘billboard’ or ‘big brand’ effect that a blanketing of the hot spot on the CDUs would achieve. However again we must insist that by using the access that it does have at various points of sale, secondary materials, HORECA promotional events and direct marketing it should prove well within the ability of a well-resourced, massively experienced cigarette manufacturer with a widely recognised international brand to communicate a brand re-launch.
Then there is direct marketing. This involves an array of mechanisms ranging from cell phone messaging to inviting prospective customers to HORECA venues (as distinct, that is, from major promotional events which seek to align the venue with the cigarette brand).143 This form of marketing tended to be dismissed out of hand by the applicants.144 However, on reflection we do not believe that it should be quite so easily dismissed. We do know that BATSA spent heavily on this form of marketing.145 While we do not infer anything about the importance of direct marketing simply from the fact of BATSA’s considerable expenditure, the likely force of direct marketing does square with the significance that is assigned, by all parties and their various experts, to peer pressure and imaging as a mechanism of brand promotion.
Scammel Katz – whose expertise is in POS promotion – and other of the applicants’ witnesses expressed the view that this form of marketing would reach too few smokers or potential smokers to prove effective from a promotional point of view. However HORECA promotional events also reach relatively few individuals. The point is not the quantity of people reached but rather their quality as brand role models. Once the universe of smokers at whom direct marketing is aimed is reduced to these opinion leaders, the ability to reach relatively large numbers is concomitantly enhanced.146 Moreover, as BATSA has pointed out a miniscule proportion of cigarette smokers engage with vending machines and yet much is made of this entirely passive interaction between, so to speak, man and machine.147 This is to be contrasted with the direct marketing method of addressing a customer or potential customer directly utilising all the various persuasive attributes of a ‘brand ambassador’. We think it more probable that direct marketing is likely to be a significantly more powerful form of brand promotion than a few facings on a vending machine languishing outside of the toilets in an ill lit HORECA venue.
A further note on ‘harm’
The applicants have, for the overwhelming part, sought to rely on proof of significant foreclosure in order to infer anti-trust harm. As elaborated at length we do not believe that significant foreclosure has been established. It is our view that were manufacturers of the size and with the brand portfolios of JTI and PMI to make a determined effort to penetrate this market, then there is sufficient access to promotional opportunities – in organised retail, in independent convenience and in HORECA – to reduce, through competition on the merits, the foreclosing effects of BATSA’s limited exclusive in organised retail and in vending machines, its trade investment programmes in independent convenience, and its exclusive arrangements with selected HORECA venues. Our conclusions are partly based on evidence that shows that where BATSA’s rivals have been prepared to expend effort and resources, they have accessed promotional opportunities. In addition, there is evidence to the effect that POS promotion is not the only mechanism for cigarette marketing and, again, it has been shown that when JTI has engaged in price and quality (brand re-launches) competition, then it has made inroads. All this, of course, aside from our view that the practice of category management, when viewed as the purchase of a limited exclusive, while clearly a contravention of idealised category management principles, is not, on its own, a convincing theory of anti-trust harm.
Some rather token efforts to identify direct consumer harm have been made by both JTI and the Commission.148 It has been suggested that had BATSA not, in violation of category management principles, competed over POS promotional opportunities, then it would have been obliged to compete on price. The Commission has suggested that the range of consumer choice has been compromised by BATSA’s conduct. The latter is, as a result of universal distribution, manifestly untrue.149 As for price competition, JTI has indeed established that dropping the price of Camel relative to the BATSA premium brands boosted demand. BATSA has elected to focus on alternative forms of competition – competition for retail shelf space and positioning. As we have indicated this is both a legitimate form of competition and one, whose benefits are almost certainly experienced by retail customers.
The applicants have also sought to rely on what they believe to be Camel’s indifferent performance as evidence of the foreclosing effect of BATSA’s conduct. We should note at the outset that BATSA’s rivals have, in the period under question, maintained or increased their market share. While JTI’s market share has remained somewhat stagnant at approximately 4.7% between 2001 and 2005,150 BATSA experienced a decline in volumes of 24% while JTI experienced a 19% decline in volumes. When the market grew between 2005 and 2007, JTI managed to increase its volumes by approximately 49% and BATSA also experienced an increase in volumes of approximately 18%, albeit from a much higher base than JTI. In effect while the general market between 2001 and 2007 declined by 4.9%, BATSA volumes declined by 10.6% and JTI volumes increased by 20.8%. 151
Nor, indeed, has it been established that Camel’s performance has been disappointing. The contention that Camel has performed poorly overall is based upon its own vastly exaggerated expectations of its performance and particularly on the projections of market share contained in the evidence of Mr. Makola. He took the view that Camel was capable of achieving a 12% market share in South Africa. The evidence revealed that no-where in the world had Camel achieved a national market share as high as 12%.152 It was also pointed out that all premium brand sales accounted for 12% of the entire South African cigarette market.153 Indeed Makola conceded that a 3% market share is significant for a premium brand.154
However, even if there had been a decline in the market shares of BATSA’s rivals this would be difficult ground on which to rest an abuse of dominance claim. At the risk of stating the obvious, in the absence of a coherent theory of harm and evidence of foreclosure, the inference to be drawn is that JTI’s disappointing performance derives from the superior product offering of BATSA or its superior competitive strategy, or, simply, the advantage of incumbency and a long established dominant brand. However, in order to dispel any residual doubt on this score we will examine the claim that Camel’s performance has been impaired by BATSA’s conduct, although we emphasise that poor performance by Camel would, on its own, not constitute evidence of exclusionary conduct.
The applicants point to market share trends as evidence of the impact of BATSA’s conduct on Camel’s performance. They point to trends that purport to demonstrate a sudden flattening out of market share performance by Camel after a period of significant improvement and assert a correlation between the introduction of BATSA’s trade investment programmes and this deteriorating performance. Between 1997 and 2000, Camel’s share in the market went up markedly from approximately 2.5% and then flattened out starting from around 2000 to 2006 at approximately 4.7%.155 However, the correlation that the applicants seek to draw with BATSA’s trade investment programme is not persuasive.
Firstly, as already noted, the sharply improved performance in Camel sales and market share immediately prior to the introduction of the regulations that brought about the dark market appears to be a response to a significant reduction in the price of Camel particularly vis a vis the price of Dunhill. Once the prices of these two competing brands had been re-aligned Camel’s share flattened out. Note too that the applicants observe a spike in demand for Camel (which coincides with the drop in price) and then attribute to BATSA’s conduct the flattening out of demand for Camel which coincides with, actually slightly precedes, the introduction of BATSA’s trade investment promotion, despite having accepted that the impact of promotional activities on the demand for a cigarette brand can only be measured over the long term.156
Secondly, as already intimated, it is extremely difficult to separate the impact of the dark market regulations from the impact of BATSA’s trade investment programme.157 However, it cannot be disputed that the most significant foreclosing conduct is to be found in the regulations that imposed dark market conditions. As already elaborated the applicants insist that because the regulations have left only a small window – essentially the POS and HORECA promotional events – available for promotional activities, any conduct that further reduces promotional opportunities must be judged in a particularly serious light, its starting point must be the market already constrained by purposeful regulatory intervention. Whatever, the merits of this argument, and, as already indicated, we do not believe that it has merit, it does not assist us in correctly identifying the cause of the travails of BATSA’s competitors.
It is however reasonable to hypothesise that in a market in which above the line advertising and sponsorship was overwhelmingly the most significant mode of promotion, the abrupt proscription of these activities will impact on growth in the overall market as well as in relative market shares. Certainly it would appear reasonable to expect a freezing in relative market shares as a result of the proscription of the major form of promotion. Several of the witnesses – including JTI’s economist – accepted this proposition.158
Conversely, Mr. Scammell Katz, JTI’s marketing expert, admits that:
‘It is very difficult to prove evidentially a direct link between merchandising strategies in retail outlets or forced trial in horeca venues, and long term switching and consumption decisions by smokers. This is because of the long time over which people switch, their willingness to use a repertoire of brands and the marginal effect each retail and horeca experience may be expected to have’.159
How then are we able to identify the causative factor in the flattening out of Camel’s market share which coincided with the introduction of both the dark market regulations and BATSA’s trade investment programme? Of course, the fact that above the line promotional activity terminated abruptly while the trade investment programme rolled out gradually with, all agree, the impact of the latter only discernible in the long term, would lend weight to the view that it is the former that is responsible for the flattening out of Camel’s share. Having flattened out, it appears eminently reasonable to assume that a further change in relative shares would, in the absence of above the line promotional activity, occur only very slowly. Again we emphasise that it is common cause that the impact of POS promotional activity is only discernible over the long term.
This argument would apply with even greater force to new entrants though clearly those, like Marlboro, a very powerful international brand with the advantages of major sponsorships that permit of a continuing degree of above the line promotion and backed by the deep pockets and experience of a major tobacco multinational, would more easily overcome the strictures of the dark market in any given national market than would an entirely new brand that did not enjoy significant corporate backing. Accordingly, as already indicated, the failure of Mr. Botha’s Matrix brand comes as no surprise to us whatsoever. No amount of POS promotion or HORECA events would have rescued this brand.160
Accordingly even if we were inclined to use Camel’s performance as a proxy for competitive harm, indeed were we inclined to conclude that JTI could reasonably have expected a better performance from Camel, there is no satisfactory method for ascribing causation. There is, on the other hand, good reason for looking for the culprit, if culprit there be, in that set of public interventions that had as its avowed purpose the elimination of cigarette promotion and that was bound to do its work more rapidly and effectively than the private interventions of BATSA, interventions that relied upon the conclusion of a range of agreements with retailers and HORECA owners, that required the allocation of corporate resources, and that met with the opposition of their competitors, albeit less resolute opposition than might have been expected.
There is indeed another likely source of Camel’s travails and that is Marlboro, a large, albeit unseen and unheard, presence in this hearing. While we do not intend entering the Virginia versus American blend debate, these are clearly direct competitors for what appears to be a very large market – at least if Marlboro and Camel’s international sales are anything to go by – in those attracted to the macho image associated with these brands.161 Indeed it is clear that some of JTI’s most robust competitive actions were in response to Marlboro’s entry.162
In summary, not only can we not identify consumer harm or find significant foreclosure arising from BATSA’s promotional activities, we cannot even ascribe harm to competitors from the allegedly anti-competitive conduct. Firstly, the impact on the market and the performance, absolute and relative, of market participants arising from BATSA’s promotional activities is dwarfed by the impact of the regulations proscribing above-the-line promotion. Secondly, we are not persuaded that Camel has substantially underperformed what should have been reasonably expected, particularly given the regulatory environment.
The legal tests
We have determined that to sustain an allegation that Section 8(c) or (d) of the Act has been contravened it must be established that the conduct in question has given to rise to anti-competitive effects. We have further determined that these effects may be reflected in direct consumer harm or that such harm may be inferred from significant foreclosure. For its part Section 5 of the Act – the prohibition of restrictive vertical practices which JTI contends BATSA has contravened - explicitly requires a showing that the complained of agreement between vertically related parties has had the effect of substantially preventing or lessening competition.
After an exhaustive examination of the evidence, and contrary to the preconceptions with which many may approach an abuse of dominance allegation against a firm with a near monopoly market share, we do not believe that the applicants have discharged the obligation to show harm to competition. In particular we have closely examined the allegation that promotional opportunities have been significantly foreclosed. We do not believe that this has been shown to be so. Not only are there promotional opportunities available even within the narrow universe of POS promotional opportunities to which the applicants have sought to confine this enquiry, particularly to firms with the resources, including major international brands, and the experience of JTI and PMI, but there are also alternative mechanisms and sites of marketing and promotion that have been studiously ignored by the applicants.
It is certainly true that in certain of the retail channels investigated BATSA has purchased the right to determine, within limits, the space and positioning allocated to its rival’s brands. However, we do not believe this conduct, essentially the sale by the retailer of shelf space, to be anti-competitive. We certainly reject the view that this promotional facility should be placed beyond the terrain of competition. Indeed, as we have shown, it is an arrangement which not only enables the contracting retailer and manufacturer to jointly maximise their profits, but, in addition, there is every reason to expect that in a competitive retail market a significant portion of these payments would be passed through to the customers. We believe this to be a considerably more persuasive theoretical standpoint than that which relies on adherence to the principles of category management – which is predicated on the neutralisation of competition for shelf space – and the vague consumer benefits that are postulated to result from this. Ultimately the impact on competition of any form of exclusive arrangement must be measured by the extent of foreclosure that results from the agreement. As noted, we have exhaustively examined the evidence and we find that there is no significant foreclosure and hence no anti-competitive effect.
The matter may end there because each of the alleged contraventions rely on a showing of anti-competitive effects. However we will briefly comment on the conduct and show that, for the most part, it does not comport with that proscribed by the relevant statutory provisions.
With respect to the allegation that BATSA’s conduct contravenes Section 8(c), we observe that in the absence of evidence of significant foreclosure the impugned conduct cannot be said to be exclusionary, that is, it cannot be shown that the conduct ‘impedes or prevents a firm entering into, or expanding within, a market.’ The fact that Camel (and indeed Marlboro) has actually expanded, or, at worst, maintained, its market shares in the relevant period is noteworthy, although not dispositive. We are cognisant that the counterfactual may, had they not been the victim of exclusionary conduct, greater expansion. What does dispose of the Section 8(c) complaint is that in the absence evidence of significant foreclosure, the allegation of exclusionary conduct cannot be sustained.
As for the allegation that BATSA has contravened Section 8(d)(i) – that is, that BATSA has required or induced a supplier or customer not to deal with a competitor - the applicants have set themselves an exceptionally high hurdle. Their difficulty does not turn on whether BATSA’s relationship is indeed with a ‘supplier or customer’. Despite the intermediation of the wholesalers, we are satisfied that the retailers are in fact customers of the cigarette manufacturers, regardless of what the black letter of the law may have to say.163 And if this does not suffice to bring them within the ambit of the statutory provision, then we believe that it can equally be shown that the retailers, including the VMOs, and the owners of the HORECA venues, are suppliers of promotional opportunities, opportunities which they have conditionally sold to BATSA. We observe that by the same token one of the key elements for a Section 5 allegation is established, that is the owners of the promotional opportunities are indeed in a vertical relationship to those contending for these opportunities.
The applicants difficulty in establishing a contravention of Section 8(d)(i) is then not to be found in some or other technical legal impediment but rather lies in the extremely thin line between anti-competitive inducement, on the one hand, and the very essence of competition, on the other. In Mittal the panel observed:
‘After all, on the face of it, the practice of competition consists precisely in inducement. While we can envisage – as clearly does the Act – a species of anti-competitive inducement, the facts of this case are insufficiently clear to arrive at so far-reaching a conclusion.’164
In Senwes the panel reasoned that
‘It would seem that Section 8(d)(i) requires that the exclusionary Act complained of constitutes a process of enticing or persuading a customer or supplier not to deal with a competitor. Absence (sic) the features of persuasion or enticement to either a specific customer or supplier or a class of them, the requirements of this subsection would not be met.’165
In this matter, BATSA paid those who controlled supermarket and vending space and position a fee in order to achieve preferential space and position. BATSA paid the owners of selected HORECA venues a fee in order to permit it to associate the venue with a BATSA brand. These facilities were available on the market and, as we have outlined, there was nothing to prevent its rivals bidding for this space – indeed on many occasions they did precisely so, in the process successfully acquiring certain promotional facilities for their own use. BATSA, to be sure, possessed formidable financial muscle and capacity which enabled them to pay the rates demanded by these suppliers of valuable retail real estate. However, by paying the purchase price, BATSA is no more inducing Pick n Pay to refuse to deal with other cigarette manufacturers, than would Pick n Pay itself be guilty of inducement of its rivals’ customers were it to charge a lower price for its Corn Flakes. It is only those who hold that retail shelf space should somehow be allocated by a mechanism other than competition, who could seriously argue that what is effectively nothing other than paying a higher price to a supplier of this space is tantamount to ‘inducement’ within the meaning of Section 8(d)(i).
As intimated, a payment structured around loyalty or market share, a payment that effectively rewards the recipient of the payment for not selling the product of a rival to a customer is likely to fall foul of Section 8(d)(i) and, in South African Airways, has already done so. But to view the act of offering a higher price for an input (promotional resources or any other) or a lower price for an output as an illegal inducement of, respectively, a supplier or a customer is to penalise rather than promote competition.
JTI’s counsel urges us to view inducement in the following way:
‘So if one wants to understand what inducing is, it is offering something, either gratuitously or with money that is paid to back up that offer so as to gain some kind of advantage which can be understood in the context of the language of a refusal to deal.’166
But this takes us nowhere. If an advertiser purchased the back cover of an important magazine, it would naturally be effectively precluding the magazine owner from concluding a deal for the same space with a rival advertiser. This is a refusal to deal, a restraint of trade, only to the extent that there is an element of restraint implicit in any contract, be it with supplier or customer. This is why it is so important to privilege content over form in the assessment of exclusive dealing, and this is achieved precisely by examining the extent of foreclosure that results from the arrangement in question.
A glance at the authoritative ‘Antitrust Law Developments’ produced by the American Bar Association will confirm the primacy accorded by US jurisprudence to evidence of foreclosure in deciding an exclusive dealing allegation. In the context of the case presently before us we note particularly the observation that
‘Practices that effectively operate as partial exclusive dealing arrangements – partial requirements contracts, minimum purchase requirements, and sales quotas – generally have been upheld since they do not totally foreclose competing sellers from selling to the buyers on whom the partial exclusive dealing requirements have been imposed, and may not even be exclusive dealing.’167
JTI, in pressing its inducement claim, has sought to rely on the judgement of Europe’s Court of First Instance in Van den Bergh Foods. This judgment does frequently use the language of inducement and this in the context of a refusal to deal case and so it is perhaps understandable that the applicants would have sought support in the reasoning of the Court of First Instance.168
The facts of van den Bergh are quite distinct from the matter before us. Critically, it did impose exclusivity – as opposed to a limited exclusive – on the outlets using van den Bergh’s fridges. It was, in other words, an exclusive in distribution. In its summary of the matter the European Commission notes that
The fact that a manufacturing industry in a dominant position on a market ties de facto 40% of the outlets in the relevant market by distribution agreements containing an exclusivity clause which in reality creates outlet exclusivity constitutes an abuse of dominance within the meaning of Article 86 of the Treaty (now Article 82 EC). The exclusivity clause has the effect of preventing the retailers concerned from selling other brands of the same product or of reducing the opportunity for them to do so, even though there is a demand for such brands, and of preventing competing gains from gaining access to the relevant market.169
It is this effective outlet exclusivity which the Court of First Instance judged to be abusive. It is a world away from the situation in the South African cigarette market.
Moreover, the Act provides that in respect of each of the alleged contraventions, the conduct, if found to be anti-competitive, may be countervailed by proof of technological, efficiency or other pro-competitive gains arising from the conduct in question. In the case of contraventions of Section 5 and 8(d)(i), the onus is on the parties to the agreement (Section 5) or the perpetrator of the unlawful unilateral conduct to prove the countervailing pro-competitive gains, while in the case of a Section 8(c), the onus is on the applicant to establish that the anti-competitive effect is not countervailed by pro-competitive gains.
Given that the allegations of anti-competitive conduct have not been sustained, we do not have to examine the efficiency gains. However we observe that there are considerable pro-competitive gains. We do not include the joint profit maximising outcome of the purchase of the limited exclusive and the likely pass through of the incentive payments to the retailers’ customers as a countervailing pro-competitive gain. This underpins our theoretical conclusion that the complained of conduct – the purchase of the limited exclusive – is not necessarily anti-competitive.
The pro-competitive gains are rather to be found in the positive impact on the category of the free provision by BATSA of the CDUs, the maintenance of an orderly cigarette POS, the very existence of the vending machine channel which, as we have noted, is predicated on the incentives paid by the cigarette companies, most notably, BATSA, and the significant improvement – which is common cause – in the stocking situation, notably the reduction in out of stock situations.170 Each of these is category promoting. These are, of course, not acts of charity by BATSA whose massive market share ensures that what’s good for the category, is good for BATSA. Because they all involve considerable expenditure by BATSA, it is not at all surprising that it would attempt to appropriate as much private gain from these measures as possible, as with its attempts, a bluff we believe, to link the continued provision of CDUs to compliance with their trade merchandising drivers. Indeed, as previously noted, so great is BATSA’s interest in the health of the category (if not in the health of its customers), we suspect that JTI calculates that BATSA would continue with its category promoting activities even if we effectively prevented it from appropriating any private gain at all thus ensuring the free ride that JTI seeks. We however, think that JTI should compete for its market share, rather than have us order the elimination of critical platforms of competition.
For all of these reasons we have dismissed the applications of the Competition Commission and JTI.
ORDER
The Tribunal orders as follows:
The complaint referral of the Commission and the complaint referral of JTI, consolidated under case number 05/CR/Feb05, are dismissed;
JTI is to pay 50% of the costs of BATSA in these proceedings including the costs of three counsel and the qualifying costs of BATSA’s expert witness, Dr. Theron;
BATSA is to pay JTI’s costs for the exception application, which costs were reserved on 11 January 2006, including the costs of two counsel; and
No costs order is made with respect to any other interlocutory applications.
________________ 25 June 2009
D Lewis DATE
Tribunal Member
N Manoim and M Mokuena concur in the judgment of D Lewis
Tribunal Researcher : R Kariga
For the Commission : Advocate W Pretorius, instructed by the State
Attorney
For the intervenor : Advocate D Unterhalter (SC) assisted by J
Wilson and A Gotz, instructed by Webber Wentzel Bowens Attorneys
For the respondent : Advocate Cilliers (SC) Assisted by D Turner, and
F Snyckers, instructed by Edward Nathan Sonnenbergs Attorneys
1 As explained below JTI has also been recognized as an intervener in the complaint referred by the Commission. For convenience they are at all times referred to as an applicant.
2 Record page 684. JT International was created in 1999 when its holding company, Japan Tobacco Inc, purchased RJRI, the international operations of RJ Reynolds. JTI is headquartered in Geneva, Switzerland.
3 See Econex Report (1) paragraph 30 (Expert Witness Bundle, page 170).
4 Rothmans is a result of the merger of the tobacco interests of Rembrandt (now known as Remgro Limited) and Compagnie Financiere Richemont AG in 1995. Pleadings Bundle page 112.
5 See JTI Heads page 6.
6 T2882-3; Pleadings pp40-41
7 Pleadings page 35.
8 T7050. The Commission’s counsel describes the issue of positioning within the cigarette dispensing unit ‘as the aspect (that) is probably the most important in this matter….the pillar (of) this entire exclusionary conduct of BATSA.’
9 The term HORECA – which essentially refers to entertainment venues - is explained below.
10 See RBB (1) paragraph 2.4 (Expert Witness Bundle, page 36); T6.
11 RBB (1), paragraphs 2.8-2.11 (Expert Witness Bundle, page 37).
12 It appears that in 2005 JTI changed Camel’s positioning from premium to popular, a ‘downward’ move that nevertheless coincided with changes in packaging and the introduction of new brand extensions or SKUs. However at present Camel appears to be marketed in the premium category (see witness statement of Oscar Makola, p692, and RBB report at Table 2, p7) with no apparent accompanying changes in the brand’s intrinsic or extrinsic characteristics.
13 See T6770 and Econex (1), paragraph 175 (Expert Witness Bundle, pages 158.10-158.11).
14 Competition Board Annual Report page 12.
16 Pleadings page 111
17 T7168
18 See JTI HOA pars 22-49.
19 BATSA’s counsel puts it thus: ‘The evidence is that after the merger, at the time of the merger Rothmans International were (sic) strong in distribution, weak in marketing, trade marketing BAT were (sic) stronger in the latter, so they changed their strategy and said we have got to build up relationships we mustn’t just feed them with distributing build up relationships with the retailers and build up relationships with a customer that is what the strategy was to market our own product better than we have done in the past, but they put some devious strategy to them.’ (T7168) It seems that Mr. Botha, the owner of the new entrant, and ultimately unsuccessful, Matrix brand, entertained the hope that the dark market restrictions may assist a new entrant by eliminating the advantages of a large budget in what had hitherto been a promotional strategy in which the various rivals focused costly above the line advertising campaigns. (see Commission’s HOA para 1.1). The restrictions on advertising achieved precisely what Mr. Botha hoped for in the realm of above the line advertising, but it did not end the competition in the cigarette market or the contribution of promotional campaigns to that competitive battle. It simply shifted the terrain from above the line advertising to other, also costly, promotional campaigns, a necessity clearly recognised by BAT and transmitted to their recently acquired South African subsidiary.
20 JTI HOA para 7.5
21 JTI HOA para 13.1 our emphasis
22Conwood Company, L.P. v. United States Tobacco Co., [2002] USCA6 235; 290 F.3d 768 (6th Cir. 2002) page 16.
23 Elyria-Lorain Broad. Co. v. Lorain Journal Co., [1966] USCA6 84; 358 F.2d 790, 793 (6th Cir. 1966). See also Conwood Supra.
24 JTI HOA para 3.3
25 Commission HOA para 3.4.2
26 The Commission’s marketing expert, Dr. Cuthbertson trod carefully around this issue. See T4307 in which he holds that while there is no single model, it is nevertheless standard practice for a manufacturer to pay to be ‘part of the process’ of category management Cuthbertson explains that this payment would entitle the payee to ‘see any planogram that was produced and they would have some say in what they thought about how that planogram was produced and whether it was good or useful.’ In this case this would appear to take the form of payment for the EPOS data which, as we will discuss below, was the threshold payment required by the large organized retail chains and which JTI does not appear to have agreed to pay until relatively recently.
27 Olympia Equip. Leasing Co. v Western Union Tel. Co. 797 F.2d 370.375.378 (7th Cir. 1986)
28 Benjamin Klein and Joshua D. Wright – Antitrust Analysis of Category Management: Conwood v United States Tobacco (p2). The contention of the Commission’s economist Dr. Jenkins that ‘the restrictions that are inherent in the legislation in terms of what is allowed at the point of sale for advertising, it is that which, it is my understanding that BATSA sought exclusivity over’ is misleading. The rights, which are limited, and therefore not exclusive, operate at two interrelated levels: first, the right to manage the category or organize the POS is limited; second, the exercise of that right, the manner in which the POS may be organized, is subject to limitation. BATSA purchased the right to organize the point of sale, to manage the category, even though it is not clear that the various agreements ever record the right in ‘exclusive’ terms. It is likely that it is not recorded as such precisely because this exclusive was limited, first, by the retailers’ retention of ultimate control over their shelves; second, by the requirement, certainly from the large grocers, that each supplier that purchased EPOS data would be construed a ‘partner’ and that this would presumably give them certain rights – let us say the right to be consulted by BATSA or by the retailers themselves – over the management of the category, consisting of the right to allocate space and position at the POS. However, this was effectively not tested because the other manufacturers, including JTI, did not buy the EPOS data and so did not meet the threshold conditions, imposed by the retailers, for ‘partnership’ and thus participating in limiting the ‘exclusive’. However, once BATSA had acquired the category management right – with or without the requirement to consult its rivals – the exercise of the right was restrained at least to the extent that it could not extend to eliminating a brand or SKU listed by the retailers from the POS and particularly the CDU itself. In other words, as we clarify below, BATSA effectively had the right to determine the space and positioning of its rivals brands in the CDU and their secondary displays at the POS but this was limited because it did not allow them to remove their rivals from the POS. The purchase of the EPOS data secured for BATSA the right to organize the POS which, given the failure of other manufacturers to purchase the EPOS data, was a de facto exclusive right; and the incentive payments paid against compliance with the drivers were directed at enabling BATSA to achieve as privileged a position on the POS as possible subject to the limitation that prevented them from removing their rival’s brands from the POS altogether. So the exclusive was limited at two levels: first, the right to organize the POS was limited by the retailers retention of ultimate control and the ability of the rivals to acquire ‘partnership’ status; second, the right to exclusively promote BATSA products was limited by the retailers requirement that their customers’ demand for variety in the available cigarette offerings be met. We are principally concerned with the substantive organization of the POS because it is this that goes to the question of foreclosure.
29 This is confirmed time and again by the retailers who appeared before the Tribunal. Note the following exchange between the panel and the witness from Shoprite Checkers:
‘CHAIRPERSON: But so do you think there’s a difference between tomato sauce and cigarettes in that sense, that in the case of tomato sauce people stand in front of the tomato sauce shelves and make up their mind about what tomato sauce they want to buy, but whereas cigarettes are concerned people do not do that, they walk up to the kiosk and ask what they want?
MR ACKERMANN: There would be browsing categories and there would be destination categories. The one thing that all 3 cigarette suppliers agreed with me, was that the biggest thing that we had to ensure is in-stock and that is what Shoprite have been focussing over the last years, because all 3 representatives from all 3 supplier or all 3 companies have said to us that if you’ve got the stock on the shelf, the consumers will ask for it.’ (T4531)
30 Witness statement of Siemon Scammell-Katz at par 6.4.7, Expert Witness Statements page 13.
31 This is why the issue of whether space allocations are determined by national shares or regional shares or channel shares or individual store shares is so clearly beside the point. This is discussed ad nauseam in JTI’s HOA especially at Paras 7.83-7.125. The nettle to be grasped is that the category captain has purchased the right to organise the display of its own brands and that of its rivals and it is understood that the consequential allocation of space will privilege the brands of the category captain. However the extent of this privilege is limited by the retailers’ requirement for variety and for avoiding out of stocks thus the space allocation will approximate the rate of sale in the individual store because if it didn’t then out of stock situations would be commonplace. The question of what measure of rate of sale is right in ‘principle’ is a red herring. What is ‘right’ from the retailer’s perspective is that cigarette brands and SKUs should be available in sufficient variety and sufficient quantity to satisfy the needs of the customers, nothing more and nothing less. It is common cause that the position with regard to out of stocks improved for all brands during the BATSA category captaincy regime thus proving that BATSA did not overstep the boundaries of its limited exclusive as far as space allocation is concerned. Note that we understand that there is potentially a difference, though small, between stocking levels and facings, that is, for purposes of limiting visibility and hence promotion the number of facings may be limited while the stocking level of the CDU may exceed this to ensure that out-of-stock situations are limited
32 We will in fact argue that BATSA has indeed invested heavily in the category. The principle indicator of this is its investment in merchandising architecture. BATSA has, to be sure, attempted to extract the greatest possible private return from this investment largely by using the free provision of superior and costly merchandising architecture to reinforce its limited exclusive over space and positioning. However, it is not at all clear that the private return is justified by the size of the private investment.
33 It is for this reason that in a market characterized by a small number of large participants the application of pure category management principles is likely to provide a fruitful platform for collusion.
34 T3752
35 JTI and the Commission argue that they are precluded from bidding for the role of category captain by, inter alia, being obliged to bid for entire channels or at least entire chains of supermarkets or convenience stores. This they attribute to BATSA’s conduct, which conduct, they allege, is driven by the fact that only BATSA possesses the scale necessary to absorb the considerable costs of ‘buying’ an entire channel or chain. We will return to this later.
36 This is confirmed in the Commission’s HOA which sums up certain of the evidence of the witness from Shoprite Checkers as follows: ‘Mr. Ackerman stated that in the absence of BATSA’s requirements, he believes that the principle applicable to a merchandising cigarettes should be a flow of premium, popular and value-for-money products for all suppliers’. (CC HOA para 9.2.4.4). Effectively Ackerman confirms Klein and Wright’s thesis that with no incentive payment the retailer would adopt competitively neutral merchandising but because of universal distribution, the specific character of cigarette retailing, brand loyalty and the incentive payments, Ackerman is prepared to accede to BATSA’s preferred merchandising practices as long as they did not give rise to out of stock situations. (CC HOA para 9,2,4.4ff)
37 Klein and Wright pp7-8 our emphasis
38 Klein and Wright p47 our emphasis
39 The retailers’ interest in a limited exclusive is well articulated in the following extract from the cross examination of Mr. Ackerman, a witness from Shoprite Checkers: “Coca-Cola coolers, Coca-Cola coolers, if they put Coca-Cola coolers in a store, they insist that only Coca-Cola products go in the Coca-Cola coolers, plain and simple, they do not pay for door space, they do not pay for electricity, they give you the refrigeration equipment and only their products go into the Coca-Cola coolers. If we are in a position now where we have CO2 problems, then Coca-Cola put cans, they put water, they put whatever else they want to in the Coca-Cola coolers. So that’s the merchandise that they pay for. When it impacts on a specific category and that’s the only merchandise according to you where you are going that’s allowable in that specific category, then I have to provide for the merchandising for my customers somewhere on that merchandising unit for stocking the opposition’s brands, because I’m not prepared not to have the opposition’s brands there.” T4523-4 our emphasis
40 Klein and Wright p22. We will show that BATSA adhered – without the exigency of a specific contractual obligation – to a minimum allocation of space to its rivals. With respect to position it was significantly less accommodating – ‘bottom left’ was a mantra reluctantly compromised – but there is evidence of some compromise. For example, we learnt of the markedly different allocations of position (and space) as between the (relatively low income) Mitchell’s Plain Pick n Pay store and the (very high income) Waterfront store with Marlboro accorded significantly better space and positioning in the latter. (T4920 and T4982-4986).
41 Mr. Unterhalter (for JTI): ‘I’m dealing with the planograms and I’m saying how can it ever be that somehow there is a welfare enhancement that comes about because of an imposition of a planogram where BATSA pays for space and visibility? We just can’t see the link’. T6979 The Commission attempted to make something of the fact that in her response to this proposition, Dr. Theron, BATSA’s economist, did not identify the efficiency gain is neither here nor there (CC’s HOA para 13.10). The efficiency gain is clearly in the pass through to the retailers’ customers and the joint profit maximization that underpins the limited exclusive.
42 RJ Reynolds Tobacco Company v Philip Morris Incorporated 199 F.Supp.2d 362, 2002-1 at p8 par 19
43 This very elementary proposition was, as one might expect, supported by the JTI’s expert economist, Mr. Baker, who notes that ‘..there are strict limits on how tobacco products may be promoted. Access to the few remaining means of brand promotion is therefore vital. Since the remaining means of cigarette brand promotion are scarce, we would expect manufacturers to compete to obtain access to this key marketing input so as to be able to promote their brands.’ (RBB report cited JTI’s HOA para 5.27, our emphasis). JTI’s counsel also recognizes the ‘scarcity’ of promotional opportunities – indeed this is what the applicants case is predicated on – but appears to conclude that because , unlike the days when above the line advertising was permitted, promotional opportunities are scarce, or have become more scarce, they should be allocated according to category management principles rather than through the competitive process, which, at the risk of a trite observation, is the process competition law and economics favour for allocating scarce resources.
44 T6555
45 JTI HOA para 1.17. ‘The low market shares of rival brands and the high level of payments that BATSA makes means that retail outlets and HORECA venues are well compensated for any loss of rival brand sales caused by the restriction on their availability or, because of limited visibility in retail outlets, their perceived availability.’
46 JTI HOA para 9.91 – 9.95
47 Competition Commission v South African Airways (Pty) Ltd [2005] 2 CPLR 303 (CT) at para 131 and 132 where the Tribunal stated that “harm to structure suffices to show an infringement of the Act, rather than requiring direct evidence of harm to the consumer welfare”.
48 Mr. Baker, JTI’s economist, conceded that in assessing the impact of the POS on choice one would make, in the words of BATSA;s counsel, ‘huge allowances for the fact that well I’m dealing with a product that has an enormous amount of brand loyalty and I’m dealing with a product that people asked for whether they see it or not.’
49 More specifically, JTI claims that it does not consider the low income market a credible market at which to direct the promotion of premium and popular brands. BATSA disagrees. Note the following exchange at T5980-1 between JTI’s counsel and Mr. Wittenberg, a BATSA executive:
“ADV UNTERHALTER: And I’m assuming that particularly when we think about premium brands and so on, they are often directed towards the better off people that you marked out as being around R10 000 a month. They are the class who are less price resistant.
MR WITTENBERG: No, no, no, no, it’s actually interestingly in South Africa the other way around. The poorest part of the market, which are the black consumers, they actually have the highest premium brand incidence and high levels of your loyalty. They would rather smoke less. They smoke, a consumer, only 6,7,8 cigarettes per day but they want Peter Stuyvesant, versus in the white market there is the highest incidence of value-for-money brands. And as you’ve seen in the switching analysis that you’ve taken me to earlier, in the white consumer segment there is the highest down-trading”.
50 BATSA characterizes the applicants’ approach as ‘a wholly transparent attempt, first to construct a ‘promotions market’, and then progressively to discard various parts of that ‘market’ until a ‘market’ was yielded that was sufficiently small and focused around BATSA’s activities to create the impression of foreclosure.’ BATSA HOA para 8.36
51 Of course the association is meant to appeal far beyond those who actually attend these salubrious venues. Just as the vast majority of Stuyvesant smokers of yore will never have gone to St. Moritz, or the vast mast majority of Camel smokers will never have crashed fearlessly through the Amazon, it is the association with those people and those places that is thought to appeal, possibly most especially to those who will never visit there. So too with glamorous and expensive HORECA venues.
52 The Commission argues that the retailer should decide which SKU to promote (CC HOA Para 7.2.3). We cannot understand how the retailer would be able to make this decision or why it would want to. If the manufacturer’s better informed reading of the cigarette market suggests a long term increase in demand for an international brand like Dunhill and a long term decline in demand for a mature national brand like Stuyvesant then surely the retailer would want to defer to this judgement with the residual right to disagree if he concludes that the manufacturer has read the market incorrectly. For example, a retailer may well have concluded that Lucky Strike’s prospects were, on the basis of persistently excessive stock holdings, not as rosy as suggested by the quantum of space allocated it by BATSA and the retailer would then retain the right to instruct BATSA to accord it less prominence on the CDU. See also Commission’s HOA Para 7.2.8 where BATSA’s strategy is outlined and its pro-competitive character clearly revealed: ‘In this respect BATSA tends to take space that might be allocated to its most dominant brand (generally Peter Stuyvesant at around 40%), but which is not necessary to support its rate of sale, and allocate it to BATSA brands that it wishes to promote but which do not have the rate of sale to justify increased facings. For example, BATSA may increase the amount of Dunhill in the hotspot if JTI’s competing brand, Camel, is making headway in a particular store.’ (our emphasis) How is competition conceivably promoted by requiring a manufacturer to adopt a merchandising approach which is ‘not necessary’ to promote the sale of a particular brand, in preference to a strategy that enables it to pit one of its other brands against a competitor brand or even against one or more of its own brands? There is a long discussion of this issue in JTI’s heads Paras 7.83-7.92 which confirms – persuasively – that were category management principles to be strictly applied, the allocation of space on the planogram would be by brand, or, better, SKU and would be entirely backward looking, that is, allocation would be determined by what the brand had achieved in terms of market share in the past, rather than by a strategy to grow a new brand or revive an old brand. These paragraphs reveal sharply that the appeal to category management principles is an attempt to neutralize the point of sale and the retailers’ shelves as sites of competitive engagement.
53 Not least because international brands can still take advantage of above the line advertising. Marlboro’s sponsorship of Formula 1 motor racing is a case in point. The organizers of Formula 1 may be obliged to hold their events in countries that have not prohibited above the line advertising but the sponsors are still able to take advantage of international television coverage. This would not be a commercially viable promotions strategy for a national brand to pursue but it makes sense for a global brand.
54 Exhibit Commission File A p73
55 Exhibit Commission File C page 147
56 Exhibit Commission File A p269
57 Note that at least where Shoprite is concerned certain of the agreements included drivers related to market share and volume growth. However these drivers do not appear to have been pursued and were ultimately removed altogether. In any event nothing was made of them by the applicants. Note too that where planogramming is discussed, the grocers both insist that ranging and listing is a retailer prerogative and yet they are also incentivized to list all BATSA offerings. This presumably means that the retailers do not accept constraints imposed by the category captain on the listing of BATSA’s rivals’ products and that the listing driver only provides an incentive to list all of BATSA’s products.
58 BATSA HOA para 89.4.1
59 Each of the Shoprite agreements describe the right acquired by BATSA as that ‘to determine the physical display of the product positioning of the cigarette category in each outlet, subject to the particular customer profile of each outlet and after consultation with the company’. BATSA HOA Para 89.2.1. The applicants claim that this proviso was never implemented. A feature of this case is the extent to which it is claimed, with little uncontested evidence either way, that clear contractual commitments between large and powerful firms are simply ignored and meaningless.
60 This is in fact conceded by BATSA’s counsel presentation of the evidence of Mr. Potgieter, a BATSA witness: ‘After expiry of the yearly contract, BATSA would not react to incremental enhancements of the display of rival products, unless it involved a significant loss of space or the allocation of the ‘hot spot’ and the space adjacent to it to a rival manufacturer. Only in the latter case would BATSA withdraw its incentive offer in respect of planogramming – until it reached a stage where the rival was, in terms of space (compared to ROS) and position so dominant (on) the CDU layout that BATSA was ‘paying for nothing’. BATSA HOA para 101.2.4 our emphasis. This is reinforced by the single example provided to us of a BATSA complaint of non-compliance escalated to Ackerman. The complaint concerned precisely an incident where a Shoprite Checkers store placed Marlboro in the ‘hot spot’, an aberration from the agreed incentive drivers which Ackerman immediately corrected.
61 Exhibit 30 p86
62 Witness Statement Bundle p461-2 and T2471-2; Witness Statement Bundle p1246 and T4884, 4957 and 4951; Ackerman Witness Statement par 5.4.
63 A macrobat is an open back wall unit containing cigarettes in open shelves. A pacrobat is a gravity dispenser or overheard unit which is usually on top of the till and contains cigarettes which will largely be out of sight of the customer. The cashier, and not the customer, has sole access to the macrobat and the pacrobat. (T2092-3)
64 See JTI HOA para 7.64.1 in which the evidence estimating the cost of CDU provision is estimated. Mr. Ackerman estimated that by 2000 BATSA had invested R21 million in furniture in Shoprite Checkers alone, an amount more than five times larger than the total annual incentives paid to Shoprite Checkers in terms of the BATSA agreements.
65 T7315
66 Cited in closing argument T7268-9
67 Van Vuuren Witness Statement para 47 (Witness Bundle p1238)
68 Siemon Scammel-Katz Witness statement para 6 (Expert Witness Statement p11-13). Scammel-Katz states that: “When studying visual activity throughout the shopping process, fixations begin at the centre of the fixture. The centre of the cigarette fixture is therefore the strongest location in terms of attracting fixations. However, the location of the signpost brand (generally the best-selling brand) can bias this as smokers will often search for the signpost brand, so as to locate the cigarette category regardless of whether or not it is their preferred brand. Once the centre or the signpost brand has been fixated upon, visual activity moves to either side as the smoker seeks his or her preferred brand. If the fixture is at least relatively well merchandised, a diamond shaped hotspot appears. Respondents tend to fixate primarily on their own brand, but other brands will also be fixated upon, particularly those in the diamond of visual activity”. (At Par 6.4.4). We note that the necessity to ‘search for the signpost brand’ in order to ‘locate the cigarette category’ does not apply to the cigarette category because one does not have to search through the shelves to locate the cigarette brands. As any cigarette purchaser will know they are either conveniently located at the till or in a kiosk immediately proximate to the till. So while, if Scammel-Katz is to be believed, a shopper looking for breakfast cereals may fixate on finding a box of Kellog’s cornflakes to locate the breakfast cereal category, this will not apply to cigarettes.
69 In asserting the importance of secondary displays, Wittenberg argues: ‘In the dark market, it is not good enough to merely put the brand on the shelf and put traditionally acceptable promotional material to encourage trial. You have to do something new and innovative that attracts the consumers’ attention’. (cited BATSA HOA para 70.5). By contrast the Commission’s counsel asserts: ‘We say that as a result of the agreements of BATSA, as a result of a problem with space, as a result of the ineffectiveness of these secondary displays, these residual point-of-sale marketing opportunities are irrelevant really and they should be discarded as a real strategy for competitors to overcome the limitations placed on them by BATSA’ reinforcing his argument that the case is essentially concerned with the position assigned in the CDU. (CC HOA T7065) We find Wittenberg’s argument to be more persuasive.
70 While effectively downplaying the importance of secondary displays, JTI and the Commission have nevertheless made much of the alleged unilateral removal by BATSA of the secondary promotional displays of their rivals. We do not intend dealing with this. Little evidence has been advanced in support of the allegation. Certainly, nothing approximating a pattern of conduct of this sort has been established. The lack of evidence notwithstanding, we have little doubt that this occurred on occasion. However, while not condoning it, it speaks more to the existence of extremely – one may say, in this instance, excessively - robust competition than to impeachable anti-competitive practice. In Conwood the defendant referred to this as ‘insignificant tortuous behaviour’ that had little impact, if any on the extent of foreclosure, an assessment with which we are strongly inclined to concur.
71 Wittenberg makes the important point that ‘where a consumer only sees the brand displayed in-store, it may improve or maintain his awareness (of the brand) but it does not tell him anything about the brand or who smokes it’ (cited BATSA HOA para 70). Given the significance assigned to imagery and peer association in attracting smokers to a new brand, this argument seriously questions the effectiveness of the CDU as a promotional tool because there is no imagery or peer group association conveyed through the CDU display.
72 Ironically, Mr. Baker, the JTI economist, argued that organized retail was a particularly important site for the promotion of premium cigarette brands because of its ‘superior demographics’ relative to independent convenience. However aside from the fact that the demographics of organized retail diverge considerably between their various stores and so it is not possible to apply the ‘superior demographic’ argument to organized retail generally, the data show that the ‘superior demographic’ from the point of view of cigarette promotion – that is, young adult smokers – is to be found in independent convenience rather than in organized retail. (RBB 1, par 8.14, p74; RBB 2, par. 26, p9) Moreover, to the extent that ‘superior demographic’ is meant to refer to the disposable income of the demographic slice so described, it appears that the implicit underlying assumption that high income groups smoke premium brands while low income groups smoke VFM brands is also open to question. Indeed surely the whole point of associating a ‘hip and happening’ HORECA venue with a particular brand is precisely to say to those who cannot afford to spend their evenings in Camps Bay wearing expensive clothes and sipping overpriced cocktails, that they too can share in the magic of the venue and the lifestyles of those frequenting it by buying Marlboro or Camel or Dunhill.
73 Exhibit Commission File B page 158 at page 163. See also Oxera Report, Table 6.2, page 27. While the number of stores has changed over time, the franchise chains which contributed to the channel were Spar (with 760 stores), Pick ‘n Pay franchise (with 285 stores), OK franchise, including various brand names (with 381 stores), Seven Eleven (with 170 stores in South Africa), and Friendly stores
74 Exhibit Commission File B p164.
75 The autonomy extended to Spar franchisees is confirmed in the testimony of Mr. Denby, the manager of a Spar store, and the only retailer called to testify by the applicants. It is difficult to understand the basis for the Commission’s claim that the autonomy enjoyed by the franchisees does ‘not detract from the exclusionary nature of BATSA’s conduct in this channel.’ (T700)
76 T219 and T3125-6.
77 Stephen Van Vuuren Witness Statement page 4.
78 Simelane & Others NO v Seven Eleven Corporation SA (Pty) Ltd & Another [2001-2002] CPLR 13 (SCA)
79 BATSA HOA para 150.1. We note that Mr. Baker, the JTI economist, dismissed these as ‘relatively marginal activities around a core of BATSA controlled outlets’. This is not a credible assessment of either JTI’s achievements in this channel (particularly when weighted by the localities of the JTI affiliated stores) or of the possibilities that the franchise format offered to manufacturers willing to compete for promotional opportunities.
80 Witness Statements Bundle p274. NB The reference is silent on 80% of the stores having taken up BATSA incentives.
81 BATSA HOA para 138.4
82 Note that the grocery stores also use the franchise model but to a significantly lesser extent than the convenience stores. Our views on the impact of autonomy would apply with equal force to franchised grocery stores.
83 Commission File C, page 71; BATSA Organised Channel Plan 2003, Commission File C, page 5; Commission File C, page 69; T252, T486-487; Exhibit 121 pages 242-46 at 244. The channel is referred to as “top end convenience with high premium appeal”
84 Exhibit Commission File C p147
85 Exhibit Commission File C page 151 and Exhibit Commission File D page 151.
86 T5424-5425
87 T5441-5442.
88 See Exhibit 105 on pages 75-76.
89 T2642-3.
90 Exhibit 29, T2088.
91 T2091.
92 T2052.
93 Potgieter Supplementary Witness Statement, paragraphs 8.2 and 13.7.2.
94 Witness Statements Bundle page 668.14-668.15; T2003.
95 T7091
96 In fact Mr. Botha’s company lost some R17 million and, although it expected to achieve a market share of !% after 3-4 years it closed down after 2 years. See CC HOA paras 11.3.5-6
97 JTI argues that BATSA’s stronger incentives derives from the fact that it is defending monopoly rents while its rivals are only fighting for competitive returns (JTI HOA para 12.36). Despite the voluminous record we have not learnt a great deal about the pricing of cigarettes, though we note JTI’s contention that had JTI not engaged in trade investment programmes to maintain its market share, it would have been obliged to compete on price. But the trade investment programme is precisely an element of competition, competition for space, and at considerable cost to the manufacturer at least part of which is, as we have noted, passed down to consumers, thus reducing whatever rent has accrued to BATSA in consequence of the dominant market share that its trade investment programme may have defended.
98 Econex Supplementary Report, Figure 23, page 51.
99 Camel’s share in the total market was 3.6% in 2007 (Exhibit 171), and JTI’s market share in independent convenience in 2007 was approximately 3.7% (Econex Supplementary Report Figure 13 p31)
100 BATSA’s counsel make the point that there are clearly large niche markets in which particular brands are successful – Courtleigh in the ‘main market’ and Rothmans in the Western Cape are offered as examples. Clearly then there are different promotional strategies that apply to different markets and there seems, on the face of it, no reason why a powerful brand like Camel should not have identified a particular market and focused on penetrating that niche. Given its inroads into the ‘main market’ this would appear to have been one obvious possibility. Instead, pleading the impossibility of promoting in the ‘main market’ and suggesting, quite incorrectly, that premium brands like Camel do not sell in that market segment, it abdicated.
101 Exhibit 31, p156, discussed with Makola at T2602.
102 The independent convenience channel is made up of four subtypes namely, independent retail stores, independent forecourts, Spaza shops and bottle stores.
103 T6389.
104 T6389.
105 BATSA Answering Affidavit, Pleadings page 356; and Potgieter Supplementary Witness Statement, Witness Statement page 1411.
106 JTI Heads of Arguments, pages 167-168.
107 Vecta Guiding Principles, pages 538-540.
108 Witttenburg supplementary statement para 116; witness bundle page 1320; and Exhibit 139. Wittenberg said that: “The most memorable conversations with retailers are those where the retailer had taken a leap of faith, encouraged by the BATSA program and reduced its price levels from the traditionally high convenience levels to the recommended resale price. Their invariable response was that the overall margins in the category grew and, importantly, footfall into their business as a whole”.
109 Pleadings page 537
110 Vecta Guiding Principles, page 16, Annex RW6 to BATSA’s Answering Affidavit (Pleadings Bundle, page 550).
111 Martin Potgieter’s First Witness Statement (Witness Statements Bundle, page 1018).
112 Sprint Business Drivers, BATSA’s Answering Affidavit, Pleadings Bundle page 558.
113 Potgieter First Witness Statement par 45.1, and 47.6 (Witness Statements Bundle p1007, and 1014); T6203-6207; Exhibit 139.
114 Mr. Botha asserted that ‘one of the key issues of displaying products is trying to make your product look bigger than it is. So you’re always looking for opportunities to get mass displays and bigger displays to make the product look big…’ cited CC HOA para 7.3.4. If the CDUs are to be relied upon to ‘make the product look big’ then, in a dark market, on the application of the strictest category management principles (where brand shares rather than manufacturers shares are used to allocate space), Peter Stuyvesant’s ‘bigness’ would dwarf all other brands. To the extent that this ‘big brand’ effect is indeed an important means of brand promotion, the application of category management principles would then exacerbate the effect that we believe is, in any event, caused by dark market conditions, namely that of freezing market shares and acting as an absolute barrier to new entry. Certainly by appealing to category management principles Mr. Botha could not have expected a CDU display to make his brand ‘look big’ for the simple reason that its market share was miniscule. BATSA was able to make a small brand like Lucky Strike ‘look big’ by purchasing a limited exclusive and, on the basis of manufacturer’s share, by sacrificing space that would have, following category management principles, been allocated to one of its big brands, inevitably Stuyvesant. We should add – as already noted – that Lucky Strike seemed incapable of growing its market share despite the marketing and trade investment attention that has been lavished on it.
115 These ‘over payments’ have been incorrectly characterized by JTI’s counsel as an instance of raising rivals costs. We will deal with this in our discussion of HORECA.
116 Commission File D p158-185. Shebeens accounted for over 14% of the cigarette sales in this channel.
117 BATSA Vending Account Plan 2002, Commission File D p132-134
118 T1300
119 T1576
120 T1343-1347. See T1347 where Zelezniak states that JTI did not require 20% of all his machines but 20% of the total column space and he achieved that by splitting and moving machines.
121 T1409
122T1619-1620. The Agreement defined the “JTI specified portion” as “20% of the overall available column space area of the vending machines …” (Ari Zelezniak Witness Bundle File 1 p62)
123 BATSA HOA par 246.6.
124 RBB 1 par 7.37
125 Exhibit 45.3 which is a sample of white smokers between 18-44 years. See also Oxera Table 4.1, p12. See also T3255.
126 T1544, T1683
127 Note T5516 where the manager of Nino’s, a prominent Rosebank venue, testified that there was not a vending machine on the premises and that cigarettes were kept in a locked cupboard.
128 T5481-5483
129 T5523
130 T5496 and T5497.
131 T5499
132 T5502
133 T2565-2574
134 T5333-5338
135 ‘Desirable’, as we shall elaborate, appears on the cigarette manufacturers’ reckoning to exclude ‘main market’ venues, Again we point out that these are venues frequented by large numbers of Camel smokers, and, as noted above, significantly more cigarettes are consumed at these venues than in the HORECA universe in dispute in this matter. We repeat that we cannot understand why Camel has insisted so strongly that the ‘main market’ is not a Camel market and that promotional resourced directed at these markets would not be well spent.
136 Exhibit 31 p36 and 37.
137 T5232
138 It appears that these agreements characteristically identify a particular brand to be promoted at the venue. That is to say BATSA would identify a venue suitable for the promotion of Dunhill and, by dint of marketing considerations if not necessarily legal agreement, it would not alternate between promoting Dunhill at this venue and, say, Stuyvesant or Rothmans. The venue would, from BATSA’s perspective, become a Dunhill venue.
139 T2548Bff
140T3687-3698, where Exhibit 58 and 50 were discussed.
141 Wittenburg Supplementary Witness Statement par 130 (Witness Statements Bundle p1327) Mr, Wittenberg, a BATSA witness, also cautions that sight should not be lost of the importance of maintaining the loyalty of existing committed consumers of a particular brand and that, for these customers, pack redesigns and the introduction of new SKUs is particularly important. (T5920-5921) In Wittenberg’s words ‘loyalty management is the business that we’re in’. Cited in BATSA HOA para 64.4.2
142 T2556B and T2557B. See also Econex 2, p11-22. .
143 The CC’s counsel suggests that direct marketing is limited by the exclusive HORECA agreements. We think that this misconstrues a form of marketing that essentially invites a group of opinion leaders to a drink at a venue at which the guests are encouraged to sample the host’s product. This has nothing whatsoever to do with the HORECA agreements which after all do not extend to not permitting rival products from being consumed on the premises..
144 See CC HOA para 11.2.11 and particularly lengthy extract from cross examination of Mr. Scammell-Katz at para 11.2.11.3
145 Wittenburg Supplementary Statement p1331 paras 138-140 and Exhibit 12 pp4 and 5. Here the estimate is 50% of promotional spend but this seems to include the cost of pack redesigns.
146 Mr. Scammel-Katz argues: ‘The thing with one-to-one merchandising and one-to-one sampling is that I don't know how many million people there are that live in a country. I don't know how many million people smoke. I would imagine we are into tens of millions of people. To go and talk to tens of millions of people is going to cost you an awful lot of money. So whatever you do, you are only going to talk to a very small number of people whereas a display in a store that people go into every single one tells a message. It tells a message’.(cited CC HOA para 11.2.11.3) We believe this misses the point entirely and is at odds with the critical and universally accepted significance of peer leadership in cigarette brand promotion. Mr. Botha is cited approvingly by the Commission when he explains that in dark market conditions ‘you’ve nearly got to address them one by one, but its super efficient in the sense that it’s eye-to-eye contact and you have a much greater opportunity to make a full impact, positive impact on a consumer’.( Cited in CC’s HOA para 3.5.6.) For the same reason we believe the assertion that ‘a display in a store that people go into every single one tells a message’ is highly debatable. Note Wittenberg’s view, cited above, that argues that ‘where a consumer only sees the brand displayed in-store, it may improve or maintain his awareness (of the brand) but it does not tell him anything about the brand or who smokes it’ (cited BATSA HOA para 70). Our conclusion from the evidence is that in cigarette brand promotion it is the message ‘about the brand’ and ‘who smokes it’ that is critical.
147 BATSA HOA para 13.8.
148 CC HOA para 11.4
149 The Commission argues that choice is particularly limited in vending machines. (CC HOA para 11.4.2) However where most vending machines have 5 facings in a market where there are well over 100 SKUs available, then clearly the choice restriction has more to do with the size of the vending machine than with any exclusionary conduct. In any event given the miniscule volume of cigarettes purchased through vending machines, this restriction of choice would not appear to be of much significance.
150 T3685-3686.
151 Econex 2 figure 1, discussed at Econex 2, p6-7 and discussed with Dr Theron at T6759 and T6965-6966
152 T7187
153 T7177. In JTI’s HOA (paras 2.3.1 and 2.3.2) super premium is said to account for 1.1% of cigarette sales and premium for 8,5%, a combined total of 9,6% which Makola expected Camel, on its own, to exceed by a relatively considerable margin. Note too that JTI’s premium offering, Camel, which with 13 SKUs represented just under half of all JTI SKU’s on offer, enjoyed, according to JTI, a 41% share of the premium category. However JTI had hardly any presence in the popular category where it seems to have 3 brands and approximately 10 SKUs. Although no-one has contended for separate markets based on the premium-popular-VFM segmentation, and nor do we do so here, it would appear that JTI’s problem is not so much Camel’s poor performance but its failure to penetrate the popular segment of the market. JTI has one large international brand in the popular segment, namely Winston, which has only managed to achieve a 0,4% market share in South Africa. On this performance, on any definition of category management, Winston commands neither space nor position. Indeed on strict category management principles it may not even have received a listing. Interestingly we heard next to nothing about Winston in the hearings which, from JTI’s perspective, appeared to focus exclusively on Camel which, as we have noted, is argued to have performed well below reasonable expectations. However, on these figures JTI’s problems may stem not from Camel’s underperformance which has not only performed extremely well in the premium segment but, on share of the total cigarette market, is in the same league as Benson and Hedges, Courtleigh, Dunhill and Craven A (4 out of BATSA’s 6 largest brands), and only substantially outgunned by Peter Stuyvesant with approximately 40% and Rothmans (with its very large Western Cape niche) at about 11%. Rather JTI appears to suffer from too small a brand portfolio and, particularly, poor representation in the popular segment, although, we repeat, this problem, and Winston’s inability to get its head above the parapets, was barely mentioned in the hearings. (See JTI HOA paras 2.6-2.13). At every turn it appears that JTI has under-invested in the South African market, preferring to take its chance with the competition authorities.
154 T2605.
155 Exhibit 54; and RBB Supplementary Report Figure 1, p23
156 Acknowledged by Baker at T3703. See also RBB 2, par 41, p13.
157 Note Mr. Baker’s argument cited at para 11.3.11.4 of the Commissions HOA. Baker states: ‘What I see, the effect I see is Camel growing briskly, the onset of a dark market and the conclusion of this suite of arrangements and that brand which had previously been growing rather briskly then flat lines. That is consistent with it then living off of its installed base of regular users acquired in the late 1990s, but struggling to augment that with new recruits to the brand. That’s the effect that I see in the Camel data, accepting that, you know, there are some difficulties herein because of the lags and so on.’ (our emphasis). A more plausible account would have acknowledged massive difficulties with the lags given that there it is common cause that POS promotional activities only manifest in the long run, while, it is to be reasonably expected, the abrupt termination of above-the-line advertising, especially for a brand like Camel that appeared to have relied upon a massive above-the-line campaign, would manifest immediately particularly insofar as winning new customers is concerned.
158 T3705-6 . JTI’s counsel holds a view somewhat different to that of its economics expert. In JTI’s HOA at para 13.8 it is stated that ‘this flattening coincided with the onset of the so-called dark market, but more particularly with the introduction of BATSA’s trade investment agreements and strategies in organized retail and HORECA venues’. (our emphasis). We cannot understand the term ‘more particularly’ which is not qualified except by an extraordinary footnote to the same paragraph which states ‘that there is no cogent reason why regulatory restriction would have impacted on JTI’s growth’ and supports this by reference to a statement by Mr. Botha (already cited) to the effect that the coming of the dark market would level the playing field by reducing the possibility of large advertising spends thus relatively privileging the ‘small guy’. JTI is not a ‘small guy’ in this sense at all. All agree that it had spent massively on above-the-line advertising and thus there is every reason to believe that its proscription would have a significant effect on its growth unless it, like BATSA, was prepared to compete on new terrain.
159 Para 8.1.1 on Scammel-Katz’s witness statement
160 Not only would the dark market make new entry, and the particularly the entry of new brands, more difficult, it seems reasonable to hypothesise that the limitations on above the line advertising, the health warnings and the restrictions on where tobacco can be consumed are likely to make a particularly powerful impression on those who have never smoked before, a large part of the demographic at whom promotion is aimed.
161 Exhibit 45.3 which is a sample of white smokers between 18-44 years. See also Oxera Table 4.1, p12. See also T p3255.
162 Van Vuuren Witness Statement par 47 (Witness Bundle 1238).
163 It is for this reason that we have decided to award costs against BATSA in its exception application in which it alleged that the manufacturers were not in a vertical relationship with the retailers.
164 Harmony Gold Mining Company Ltd and Durban Roodepoort Deep Ltd vs Mittal Steel South Africa Ltd and Macsteel International Holdings BV Tribunal Case No 13/CR/Feb04 Para 204
165 Competition Commission of South Africa vs Senwes Ltd Tribunal Case No. 110/CR/Dec06 Para 162
166 T7339
167 Antitrust Law Developments (third) p176
168 See for example para 22- ‘The Commission states ‘HB abuses is dominant position in the relevant market…in that it induces retailers …who do not have a freezer cabinet for the storage of impulse ice-cream either procured be themselves or provided by another ice-cream supplier than HB to enter into freezer-cabinet agreements subject to a condition of exclusivity’ and that ‘the inducement takes the form of an offer to supply the freezer cabinets to retailers, and to maintain them, at no direct charge to the retailer’’. See also para 151 ‘HB has induced them to enter into agreements containing an exclusivity clause, and this constitutes an abuse’. Also para 159
169 [ 2003] ECR II-4653 Case T-65/98 para 7 our emphasis
170 Indeed the evidence suggests that Camel’s stocking situation has shown the most significant improvement as a result of BATSA’s management of the category. (Exhibit 59; Baker T3747-3750; and Theron T6757-6758)