---

COMPETITION TRIBUNAL

REPUBLIC OF SOUTH AFRICA

Case Numbers: 39/LM/Jul01 and

47LM/Aug01

In the large mergers between

Massmart Holdings Ltd

and

Jumbo Cash and Carry (Pty) Ltd

and

Massmart Holdings Ltd

and

Picardi Liquors (Pty) Ltd - Sip ‘n Save division

Reasons for the Competition Tribunal’s Decision – NON-CONFIDENTIAL VERSION

Approval/Prohibit

1. The Competition Tribunal issued a Merger Clearance Certificate on 21

September 2001 approving the acquisition by Massmart Holdings Ltd of Jumbo Cash and Carry (Pty) Ltd and of Sip ‘n Save, a division of Picardi Liquors (Pty) Ltd, from Rebhold Ltd. The acquisition was approved without conditions. The reasons for our decision are set out below.

The transaction

2. The transactions will result in Massmart acquiring control of Jumbo Cash & Carry and its subsidiaries Browns and Weirs from Rebhold, and of Sip ‘n Save, a division of Picardi Liquors, also controlled by Rebhold.

3. Rebhold owns 70% of Jumbo with the remaining 30% held by Tiger. Sip ‘n Save is a division of Picardi, which, in turn, is wholly owned by Rebhold. Rebhold informs us that it is disposing of these businesses because it wants to focus on its core business, namely the provision of business support and facilities management services. The target companies, Jumbo and Sip ‘n Save, are involved in the retail and wholesale trade and, as such, do not form part of Rebhold’s core business focus.

4. The core business of Massmart, on the other hand, is precisely in the retail and wholesale trade. Massmart believes that this transaction will allow it to diversify its range of retail and wholesale activities and to deepen its involvement with particular consumer segments and product groups. Massmart avers that it has the skills, expertise, experience and commitment required to invest the capital and other resources to further develop the target businesses.

5. The parties argue that the acquisition of Jumbo and Sip ‘n Save constitute a single transaction. They point out that both the target firms are controlled by Rebhold and that the completion of the Sip ‘n Save transaction is conditional upon the successful completion of the Jumbo transaction. The Commission, on the other hand, insists that they be treated as separate transactions. We are however persuaded by the parties’ argument and find that we are here dealing with a single transaction.

The parties

6. Massmart comprises the following wholesale and retail chains:

• Massdiscounters, a chain of discount stores trading as Game and Dions which retail a wide range of general merchandise.

• Makro, a chain of large warehouse club outlets, involved in the retail and wholesale distribution of food, liquor and general merchandise.

• Shield Buying & Distribution (Pty) Ltd (Shield), a buying association procuring products on behalf of 241 independently owned wholesaler businesses and 254 independently owned retail businesses.¹ Shield’s wholesale members collectively constitute 70% of Shield’s sales.

• CCW, a peri-urban and rural chain of 18 “cash and carry” or wholesale² warehouses distributing basic food, liquor and groceries to retailers trading in a low income retail market. CCW co-owns some stores with its managers who own up to 48% of the shareholding in their respective stores.

7. Rebhold’s involvement in the retail and wholesale trades comprises:

• Jumbo, a wholesale distributor of cosmetics, toiletries, and hair care products for the lower to middle income urban consumers. In the recent past Jumbo has widened its focus to include food, grocery products, hardware and cigarettes. Jumbo operates in the urban areas of Gauteng, KwaZulu Natal, the Northern Province and Free State.

• Brown’s and Weir’s, wholesale distributors of basic grocery products and a limited range of general merchandise. Brown’s and Weir’s are active in the rural areas of KwaZulu Natal and the Eastern Cape respectively. There are 22 Brown’s and 22 Weir’s stores, 22 of which (11 Browns and 11 Weirs stores) will be acquired by Massmart.

• Sip ‘n Save, a division of Rebhold’s wholly owned subsidiary, Picardi Liquor (Pty) Ltd, comprises 3 stores involved in the wholesale and retail liquor trade. It sells mainly beer and mass-market wine to low income consumers in the Port Elizabeth area.

The relevant market

8. The relevant product market comprises all of those products and/or services which are regarded as interchangeable or substitutable by the consumer, by reason of the products’ or services’ characteristics, their prices and their intended use.

9. The parties to this transaction are involved in the provision of a distribution service. They effectively serve as the intermediaries between, on the one hand, a vast number of manufacturers of a wide range of products and, on the other, the consumers of those products. Certain of their customers are the final consumers of the product – this describes the retailing activities of the parties. In other instances the customers are themselves retail outlets who purchase the products for on-sale to the final consumer. This latter activity describes the wholesaling activities of the parties.

10. The activities of three of the target firms – Jumbo, Brown’s and Weir’s – are overwhelmingly directed at the wholesale trade. That is, a relatively insignificant portion of their revenue is derived from sales made to final consumers - their customers are predominantly retailers who on-sell the product that they purchase to the final consumers. The fourth target firm, Sip ‘n Save, is involved in both wholesaling and retailing.

11. One of the acquiring firms, Massdiscounters, trading as Dion’s and Game, is exclusively engaged in retailing. A second acquiring firm, Makro, is engaged in both wholesaling and retailing. CCW, a third acquiring firm, is overwhelmingly involved in the wholesale trade – a relatively insignificant portion of its sales is made to final consumers. Shield, the fourth acquiring firm, undertakes bulk purchases on behalf of it members, who are both retailers and wholesalers. The bulk of Shield’s purchases are undertaken on behalf of wholesalers.

Percentage sales by customer type:

	Wholesale “cash & carry”	Direct to general public
Makro (SA)	57%	43%
CCW (SA)	95%	5%
Massdiscounters (SA)	-	100%
Jumbo	95%	5%
Browns stores	95%	5%
Weirs stores	95%	5%
Sip ‘n Save	50%	50%

12. The above table underlines Massmart’s strong involvement in the retail trade: Dion and Games are exclusively in retail and some 43% of Makro’s considerable revenues are generated through direct sales to the public. However, the target firms in the Rebhold stable – Jumbo, Brown’s, Weir’s and Sip ‘n Save - have a very limited exposure to the retail trade. There is, thus no competitive overlap of any consequence in the retailing trade. Accordingly little purpose is served by further examination of the retail market in grocery products (in which Makro is active) or general merchandise (in which Makro, Dions and Game are active). The one exception is Sip ‘n Save, half of whose revenues are generated through retailing liquor. There is some competitive overlap in both the wholesale and retail liquor trades arising from the acquisition of Sip ‘n Save and this will be briefly examined – as we shall show Sip ‘n Save’s geographic focus as well as the specific consumer segment that it serves considerably limits the competitive overlap with the acquiring parties.

13. Accordingly, with the limited exception of liquor retailing, our analysis focuses exclusively on the activity in the wholesale market. The parties have contended for a relevant market that denies the existence of a boundary between wholesale and retail markets. They argue that the large retailers – be they grocery, general merchandise of liquor retailers - effectively set the upper limit on the wholesale prices of those products. It is argued that the limit imposed by competition between the supermarkets and the small retailers effectively constrains the ability of the wholesalers to unilaterally impose price increases on their retail customers – in the event that such action on the part of the wholesalers compromised the price competitiveness of their retail customers, the result would be a decline in the market share of all those retailers dependent upon wholesale distribution and, by extension, a decline in the customer base of the wholesalers. We do not accept that this eliminates the distinction between the wholesale and retail markets. It may, however, through the exercise of a countervailing power, influence the assessment of post-merger market power and the argument will be considered at that stage.³

14. This is not to deny the major impact that the advent of mass, supermarket-type retailing has had on the character and extent of wholesaling. In essence, and at the risk of considerable oversimplification, previously wholesaling was, to all intents and purposes, the only mechanism for the distribution of manufactured products to a myriad of small retailers. However, the advent of the supermarket effectively shortened the link between the manufacturer and the final consumer by eliminating the wholesaling stage – these relatively gigantic retailers purchased in sufficient bulk to permit economies of scale in warehousing and distribution and enabled them to demand discounts similar to, or even greater than, those available to the large wholesalers. The rise of retail franchising operations, with the franchisor generally performing the intermediating function between the manufacturer and the individual franchisees, further excluded the wholesaler.

15. Many of the old names in South African wholesaling simply disappeared. However, there arose in their stead a new breed of wholesaler dedicated to serving the still considerable slice of the retailing trade that was not subsumed by the new supermarkets and franchise operations. However, by contrast with their predecessors, these new wholesalers offered an extremely pared down service, a level of service geared towards enabling the remaining small retailers to achieve price-competitiveness with their efficient large-scale counterparts. Credit was limited as were other services such as transport. The small retailer would typically go in his or her own transport to the massive warehouse-type facilities of the new wholesaler, trawl the aisles and choose the required stock, pay cash, and depart with the wares – hence the term ‘cash and carry’. There is, of course, an element of caricature in this description. In fact it appears that many of the ‘cash and carry’ outlets do offer limited credit and some do offer, at a price, transport and other distribution services.⁴ However, the point remains: the new breed of wholesaler responded to the new retail environment by offering a service to small retailers that gave them the opportunity to remain price competitive in the process eliminating convenient but costly services such as telephone ordering and delivery. Certain of the parties to this transaction – who will be referred to as ‘wholesalers’ - belong to the new breed of ‘cash and carry’ wholesaler.

16. The manufacturers whose products are distributed through the mechanism of wholesale by the parties include the producers of grocery products, liquor and general merchandise.

17. General merchandise, as the name implies, encompasses a disparate array of products including office supplies, DIY equipment, hi-tech products, household appliances and even certain categories of clothing. It appears that a relatively small proportion of trade in these products is conducted through wholesale channels, that is, for the most part, retailers of general merchandise tend to source their product directly from the manufacturers. Hence, it is to be expected that Dion’s and Game, who trade overwhelmingly (to the extent of some 91% of their sales) in general merchandise, are not involved in the wholesale trade at all. 36% of Makro’s sales are in the area of general merchandise (with 51% and 13% in grocery items and liquor respectively) with 57% of its sales directed at wholesalers and 43% purchased by end consumers. It appears that the lion’s share of Makro’s retail activity is in general merchandise and liquor, with grocery products making up the bulk of its wholesale revenues.

18. Grocery products encompass food, cigarettes, health and beauty products and non-edible consumables such as detergents and house care products. In contrast with general merchandise there is a considerable wholesale trade in grocery products. Certainly the competitive overlap between the parties to this transaction occurs in the grocery products wholesale market. Jumbo’s, Brown’s, Weir’s, CCW and Makro are principally involved in the wholesale distribution of grocery items. It is however important to keep in mind that although Jumbo does sell groceries it has established itself as a leading player in a particular segment of the grocery market namely the distribution of cosmetics, toiletries and hair care products. Makro, as noted above, does earn significant revenues from its retailing activities but this is largely accounted for by its sales of general merchandise

19. Note that while a single store may and, indeed, in this particular case, does distribute grocery products, general merchandise and liquor, it is also not uncommon, and also occurs in this instance, for stores to specialize in the distribution of one or other of these broad product categories. We are satisfied that there is not meaningful substitutability between groceries, general merchandise and liquor, and, accordingly that they belong in separate relevant markets no matter that they are frequently traded under the same roof.

20. By the same token it may be argued that one grocery item is not substitutable for another and that within the overarching grocery products market the specific product categories – say detergents versus canned food products – should be treated as separate relevant wholesale grocery markets. However, it appears that mass grocery products wholesalers do not specialize in the distribution of a small number or limited range of grocery or general merchandise items - successful mass wholesaling appears to demand that a full line of items is stocked. The range of items will, to be sure, vary, principally, it appears, in relation to the customer segment upon which the wholesaler store or chain is focused, but, within that parameter, each will carry several thousand different lines. We are, in short, dealing with full-line wholesaling of grocery products.⁵ A distributor specializing in the distribution of a small number of select grocery product brands is not part of the same market as a full line grocery wholesaler whose product offerings are intended to satisfy the full range of requirements of a typical retailer.⁶

21. We will confine our analysis of the competitive impact of this transaction to those markets in which both the acquiring and target firms are active, in which, in other words, there is competitive overlap. This refers principally to the wholesale distribution of grocery products, although we will briefly examine the wholesale and retail markets for liquor. There is, as already elaborated, a limited wholesale market in general merchandise and there is no competitive overlap between the acquiring firms considerable involvement in the retail side of this trade and the activities of the target firms given their near exclusive involvement in grocery wholesaling.

22. Note that the market is further limited by its geographic boundaries. Where the retail liquor market is concerned the geographic market is very narrow, confined, at its widest, to Port Elizabeth. Hence the geographic markets for retail liquor would be very narrow, certainly not extending beyond the city or town in which the retail outlets are active. In the case of the Sip ‘n Save transaction then the widest retail geographical market would be Port Elizabeth and, given that its trade is directed at low income consumers this may be narrowed to specific sections of the city.

23. On the other hand, geographic markets for wholesale products are clearly larger than that of their retail counterparts. The Commission has concluded that retailers are prepared and able to purchase product from wholesalers within a radius of some 200 kilometers of their stores. This is borne out by the data supplied by the parties although it appears that most retailers have access to wholesale facilities in closer proximity. Nevertheless retailers are clearly prepared to travel some considerable distance to access the desired products and, this fact, combined with the assertion that, in these circumstances, a ‘chain of substitution’ will further widen the geographic parameters of the market, leads us to accept the contention that the geographic markets for the wholesaling of grocery products are local.⁷

24. In a previous matter, the Tribunal held that, even if the ability to physically substitute an alternative source of supply was geographically bounded, in order to sustain a claim for a sub-national definition of the geographic market, it still had to be demonstrated that national chains allowed prices and other competitive conditions to be set within these geographic bounds rather than at a national level.⁸ This has significant implication for the place of independent stores operating in limited geographical areas – if the national chains pricing and competitive strategies are not influenced by the competitive behaviour of local or regional stores then the latter are clearly not within the relevant market. Conversely expressed, this would mean that the relevant market was the market for national chains of wholesale groceries or, as in the case referred, national chains of furniture retailers. In the present case, however, we are persuaded that store managers play a significant role in determining prices, certainly in the stores of the acquiring party.⁹ This is consistent with a finding that the relevant geographic market for the wholesale trade in grocery products is indeed local.

25. This finding further limits the areas of competitive overlap as illustrated in the following table:

Provinces	Grocery Wholesale (X)		Liquor Wholesale/retail (xx)
	Acquiring Firm	Target firm	Acquiring Firm	Target firms

Gauteng	X	X	Xx
Kwazulu Natal	X	X	Xx	X
Free State		X
Western Cape	X		Xx
Eastern Cape	X	X	Xx	Xx
Mpumalanga	X		Xx
Northern Province	X	X	Xx

26. Thus the provinces in which both the acquiring and target firms are present with regard to grocery wholesale are Gauteng (GP), Kwazulu-Natal (KZN), Northern Province (NP) and Eastern Cape (EC). With regard to liquor wholesale and retail there are two geographic markets in which they overlap namely Kwazulu-Natal and Eastern Cape.

27. We will, therefore, focus our analysis on the wholesale grocery market in GP, KZN, NP and EC and on the liquor wholesale and retail market in KZN and EC since these are the relevant markets in which competition will be affected by the merger. As noted it is our view that the geographical markets for wholesaling are somewhat narrower than the province but data limitations oblige us to use the provinces as a proxy.

The Impact on Competition

28. In terms of Section 12A(1) of the Act we are enjoined to determine whether or not the merger is likely to substantially prevent or lessen competition in the relevant market. In terms of Section 12A(2) we are, in making this determination, required to assess the strength of competition in the relevant market, and the probability that the firms in the market after the merger will behave competitively or co-operatively. Section 12A(2) provides a non-exhaustive list of factors that, if relevant, we are required to consider in making our determination.

Grocery Products

The level and trends of concentration, and history of collusion in the relevant market

29. The competitors in the wholesale grocery products trade in each relevant province, that is in each relevant geographic market, include both wholesale chains (such as Metcash, Sentra/Mega Save and Rainbow Cash & Carry) and various independent wholesalers. The Commission avers that, on the basis of interviews conducted with competitors, that there is robust competition in each in each geographic market and that price is the overwhelming basis for competition.

30. Neither the parties, nor the Competition Commission could provide reliable market share figures for the independent wholesalers or the main chain wholesalers. The Commission presented us provincial market shares although it appears that these shares are calculated on the basis of estimates of the provincial turnovers of the wholesale chains active in the respective provinces. In other words it does not include the sales figures of the various independent wholesalers.¹⁰

Geographic market	Massmart: Pre-merger %	Target: Pre-merger %	Post-merger %
Gauteng	10.40	6.15	16.56
KZN	9.49	3.35	12.84
EC	8.88	23.81	32.69
NP	9.21	10.50	19.71

31. However based on the available market share information we conclude that there will be active competition in each geographic market post-merger. The market share data is corroborated by the number of competitors competing in each province. In Gauteng there are 18 Metcash stores, 1 Rainbow store and more than 60 Independent stores as opposed to the 8 of the merged entity. In KZN there are 23 Metcash stores, 19 Megasave/Sentra stores and more than 50 Independent wholesalers as opposed to the 17 of the merged entity. In the EC, where the market share figures provide the most serious grounds for concern, our fears are somewhat allayed by the fact that there are 24 Metcash stores, 11 Megasave /Sentra stores, I Rainbow and over 60 Independents as opposed to the 18 stores of the merged entity. In the NP there are 3 Metcash and 5 Independents competing with the 2 stores of the merged entity.¹¹ Furthermore we are reassured by evidence presented by the Commission suggesting that many of the independents are long-established, large and, in certain instances referred to, growing businesses.

32. That having been said, the transaction clearly results in the absorption of a formidable competitor – at least insofar as the sale of Jumbo is concerned – into the ranks of a robust and substantial player in the same market.¹² Our concern here is somewhat ameliorated by Jumbo’s focus on a particular market niche, the health care and beauty products market. Indeed some consideration was given to defining this segment as a separate relevant market. Although this was ultimately rejected and this product niche was included in the grocery products market, the fact that Jumbo’s focus is so distinctive does, unquestionably, ameliorate the impact of the transaction on competition.¹³ This is, to some extent, borne out by Massmart’s intention to treat Jumbo and the product market segment in which it is active as an additional division of the Massmart group.¹⁴

33. Does the transaction increase the likelihood of collusion in the grocery products wholesale market? Obviously, a particular danger here is of collusion between the relevant members of the Metcash and Massmart groups. We have, however, no reason to believe that collusion is likely to occur post-merger. There is no obvious history of collusion – quite the contrary, there is evidence of robust competition. Moreover, although Metcash and Massmart will, between them, command an important share of the relevant market, competition from established independent wholesalers and from the various buying groups will make it difficult to collude successfully.¹⁵

The ease of entry into the market

34. It has been suggested that entry into this market is relatively easy. Suppliers, we are informed, are generally willing to extend credit to new entrants.¹⁶ Moreover, while the experience of firms within both Rebhold and Massmart is that a wholesaler must be prepared to stock a substantial range of the grocery products line items, it is not necessary to extend beyond grocery products, and, as Jumbo’s experience demonstrates, it is even possible to flourish by focusing on a broad product niche within the grocery products relevant market. Stores are basic in their design – they are essentially warehouses – and there is no commercial requirement that they be located in the high rent parts of the towns in which they are based. Wholesalers attract little brand loyalty from customers whose overriding concern is with price – hence there is little of the cost that retailers, for example, generally have to incur in advertising and in building up store or chain brand reputation.

35. The Commission holds that supply-side substitution is likely in the face of anti-competitive behaviour by the wholesalers. That is, the Commission argues that in the event of an exercise of market power on the part of the wholesalers it would be relatively easy for the well-established and richly resourced large retailers to enter wholesale distribution. We do not accept this argument – as Mr. Lamberti of Massmart points out this would essentially entail the retail supermarket chains supporting their own competitors.¹⁷

The degree of countervailing power in the market

36. As indicated above, the parties argue that the coincidence in a single retail market of large supermarkets outside of the wholesale chain of distribution with small retailers dependent upon a wholesale distribution network constrains the ability of the wholesalers to increase the price of their service to their customers – it would be a self-destructive act, one that would sacrifice market share to precisely that element of the retail market, the supermarkets, that has removed itself from the wholesale chain of distribution. In other words, the wholesalers and the small retailers have a mutual interest in the latter maintaining market share against the supermarkets thus providing a peculiarly strong incentive for the wholesalers to raise efficiencies rather than prices.

37. While, as indicated above, we were not willing to accept that this eliminated the distinction between the wholesale and retail markets it is, on the face of it, a factor countervailing the ability of the wholesalers to increase the price of the service that they provide to their retailer customers. In order to make a firm finding on the strength of this countervailing power the degree to which supermarkets and small retailers are substitutable forms of distribution (that is, participants in the same relevant market) would have to be established. Small retailers are able to charge a premium over the prices offered by supermarkets and still retain their custom – they are able to do this because they offer convenience, a feature for which consumers are clearly willing to pay a price.

The ‘failing firm’ defence

(THIS SECTION CONTAINS CONFIDENTIAL INFORMATION)

38. Accordingly the ‘failing firm’ argument supports our decision to approve this transaction.

Liquor Products

43. There is a degree of overlap between the activities of the target and acquiring firms in the retail and wholesale liquor trades. We have accordingly examined the impact that the transaction would have on competition in these markets. As briefly elaborated below we have concluded that there is unlikely to be any discernible impact on competition in either the retail or wholesale liquor markets

44. Makro is the only wholesale chain that has a significant presence in the liquor market, with 13% of its total sales representing liquor sales. The target wholesale chains only sell liquor in KZN and the EC where its market shares are 0.58% and 2.95% respectively. Post-merger the market shares are as follows:

Gauteng 5.39%

KZN 6.88%

EC 4.95%

NP -

45. Sip ‘n Save, the retail arm of the target firm, distributes beer, low quality wine and liquor to the lower income consumer through 3 stores in Port Elizabeth. The major wine brand that is sold by Sip ‘n Save is “Namaqua” which represents 30% of sales generated, with the remainder beer and spirits.

46. The following table sets out the retail market shares:

Market players in Port Elizabeth	Number of large stores	Market shares based on number of large stores
Picardi	5	7.24%
Big Daddies	10	14.49%
Makro/Sip ‘n Save post- merger	4	5.79%
Other	69	72.46%

47. The smaller liquor retail outlets have not been taken into consideration in calculating the market shares and would accordingly reduce the market shares should they be included.

Conclusion

48. In our analysis of the transaction’s impact upon competition we have considered two markets, these are the grocery products wholesale market in selected geographical areas and the liquor retail and wholesale trade in selected geographical areas.

49. We are persuaded that the transaction will not substantially diminish competition in the grocery products wholesale market. While market shares in certain of the geographical areas in question are prima facie cause for concern we are persuaded that the merged entity will not only face robust competition from other national chains but, given evidence of local determination of prices and competitive strategies, will face competition from well-established independents. The suppliers have a positive interest in facilitating the competitiveness of independents who are, moreover, assisted, by the growth of bulk buying groups. Entry into this market is, we conclude, relatively easy. Moreover, the influence of large retailers on the prices of small retailers will act as a countervailing influence upon the ability of the wholesalers to exercise market power over their customers.

50. Market shares in liquor retail and wholesale markets indicate clearly that there is no prima facie cause for concern.

10 October 2001

D.H. Lewis Date

Concurring: F. Fourie, P.E. Maponya

---