Anticipated acquisition by Asda Stores Limited of Netto Foodstores Limited ME/4551/10 The OFT's decision on reference under section 33(1) given on 23 September 2010. Full text of the decision published on 20 October 2010. Please note that the square brackets indicate figures or text which have been deleted or replaced in ranges at the request of the parties or third parties for reasons of commercial confidentiality. PARTIES 1. Asda Stores Limited (Asda) is one of the UK's largest supermarket chains and is a wholly owned subsidiary of the US retailer Wal-Mart Stores Inc. Asda operates around 350 stores in the UK, almost all of which are large supermarkets (that is, having more than 1,400 sq m of sales space). Although groceries are its core business Asda also sells (among other products and services) clothing (through the 'George' range), general household items, financial products (for example, insurance products and credit cards), travel agency services and in-store pharmacy services. 2. Netto Foodstores Limited (Netto), a wholly owned subsidiary of Dansk Supermarked A/S which in turn is 68 per cent owned by the Danish company AP Møller-Maersk A/S. Netto operates 194 stores in the UK. It is a limited assortment discounter (LAD) offering fewer product lines than similarly sized mainstream supermarkets but at generally low prices.1 Unlike Asda, Netto does not offer non- grocery products to a significant extent. 1 Other LADs operating in the UK include Aldi and Lidl. 1 THE TRANSACTION 3. On 27 May 2010 Asda entered into a sale and purchase agreement with Dansk Supermarked A/S to acquire the entire issued share capital of Netto. 4. Once the deal is completed Asda intends to convert the Netto stores to trade under the Asda fascia by around [ ]. Asda submitted that once the conversion process is complete the number of product lines sold in the former Netto stores that will be retained by Asda will increase from around 1,650 today to 5,000–6,000 in the smaller stores and over 10,000 in the larger stores. 5. After extensive pre-notification discussions the parties provided the relevant information to the Office of Fair Trading (OFT) on 12 July 2010. The OFT's administrative deadline in this case is 22 September (extended by agreement with the parties). JURISDICTION 6. As a result of the proposed merger the enterprises Asda and Netto will cease to be distinct. 7. For the year ending 31 December 2009 Netto had a UK turnover of around £745 million. Therefore, the turnover test in section 23(1)(b) of the Enterprise Act 2002 (the Act) is met. 8. Consequently, the OFT considers that arrangements are in progress or contemplation which, if carried into effect, will result in the creation of a relevant merger situation.2 THE COUNTERFACTUAL 9. In order to assess whether the proposed merger gives rise to a realistic prospect of a substantial lessening of competition (SLC), 2 The turnovers of the merger parties are such that the merger would have a European Union dimension but since both parties earn more than two-thirds of their European Union turnover in the UK the case falls within the jurisdiction of the UK (Article 1(2) of the EC Merger Regulation, Council Regulation (EC) No 139/2004). 2 the OFT needs to make an assessment of what would be the case without the merger (the counterfactual). In most instances this is simply the pre-merger situation.3 10. In this case, at the national level the OFT has not received any evidence that the pre-merger situation is not the appropriate counterfactual. As such, the OFT has used the pre-merger situation as the counterfactual in this case. 11. At the local level, the parties have submitted some evidence that the pre-merger situation is not the appropriate counterfactual in some areas (for example, because of the imminent entry of rivals). These have been incorporated in the OFT's analysis where relevant to the outcome of the OFT's decision (and all such areas are discussed below). MARKET DEFINITION Product market 12. Based on the Competition Commission's (CC's) groceries market report the parties submitted that Asda operates mostly in a market of large supermarket stores, or one-stop stores (greater than 1,000 to 2,000 sq m) constrained by other national grocery multiples (Tesco, Sainsbury's, Morrisons, Co-operative Group (CGL), Marks & Spencer and Waitrose) operating one-stop stores.4 13. Netto, on the other hand, operates in the LADs segment. All of its stores are greater than 280 sq m (below which a store is classified as a convenience store) but below 1,400 sq m.5 The parties 3 'Merger Assessment Guidelines', A joint publication of the Competition Commission and Office of Fair Trading, September 2010, paragraph 4.3.5. 4 Competition Commission, 'The supply of groceries in the UK market investigation', 30 April 2008, paragraphs 4.135 ('CC groceries report'). 5 The CC used the threshold of 1,400 sq m for much of its analysis (CC groceries report, paragraph 13). In Tesco plc and the Co-operative Group (CWS) Limited: a report on the acquisition of the Co-operative Group (CWS) Limited's store at Uxbridge Road, Slough by Tesco plc (November 2007) the CC also used 1,400 sq m as the threshold. 3 submitted that Netto is not in the same product market as Asda but nevertheless the OFT should be mindful that after the proposed merger the Netto stores will eventually trade under the Asda fascia. 14. In previous decisions the OFT has examined supermarkets according to their size.6 One-stop stores were classified as those with a net sales area of 1,400 sq m or above. These stores were found to form their own product market. Mid-size stores were classified as those with a net sales area of less than 1,400 sq m but above 280 sq m. These stores were found to be constrained by one-stop stores and other mid-size stores. 15. Moreover, an aspect of product market definition in previous grocery investigations has been which store fascia to include within the relevant product market. From the perspective of the large grocery retailers (defined by the CC as comprising Asda, CGL, Marks & Spencer, Morrison, Sainsbury's, Tesco and Waitrose) the effective competitor store fascia, has included the large grocery retailers, regional grocers and symbol groups. Importantly for this case, LADs, frozen food retailers and specialist retailers have not previously been included in the product market from the perspective of large grocery multiples (even where their stores are above 1,400 sq m). 16. The OFT has not investigated any case involving the acquisition of a LAD and therefore has no case precedent on the extent to which there would be a loss of competition between a larger grocery retailer and a LAD. 17. Third party competitors were mixed in their responses on this point. Some considered that LADs should be included in the same relevant product as the large grocery multiples, especially in the mid-size sector. 6 See, for example, Completed acquisition by Wm Morrison Supermarkets plc of 30 stores from Co-operative Group Limited, Case ME/4132/09, OFT decision of 10 July 2009 ('Morrisons/30 CGL stores'); and Anticipated acquisition by Co-operative Group Limited of Somerfield Limited, Case ME/3777/08, OFT decision of 20 October 2008 ('CGL/Somerfield'). 4 18. Others pointed out that the competitive constraint between Asda and Netto may be asymmetric in that Asda may place a strong constraint on Netto but in turn Netto only places a weak constraint on Asda. 19. There is some evidence to support this proposition. For example, the CC found in its groceries report that: • with the exception of the impact of Lidl's entry on the revenues of Sainsbury's stores, the entry of a LAD does not impact on the revenues of the large grocery retailers,7 and • the limited number of products offered by the LADs means that they are not close substitutes for similarly sized stores operated by the large multiples.8 20. Evidence submitted by the parties in this case show that the simple average of the (expenditure weighted) diversion ratio from surveyed Netto shoppers to Asda was around [ ] per cent whereas from Asda to Netto it was around [ ] per cent.9 This is suggestive that Netto does not constrain Asda to any material extent but Asda does constrain Netto. On this basis, the OFT has included Asda, and the other large grocery multiples, as competitors to Netto for the purpose of assessing this merger. 21. Evidence indicating that the OFT should include the other LADs in the same product market as Netto comes from the CC's groceries report, which concluded that LADs stores are competitively constrained by other LADs stores. A key rationale for this finding was their similar product offering by number of Stock Keeping Units (SKUs, a measure of the number of products).10 7 CC groceries report, paragraph 4.71. 8 CC groceries report, paragraph 4.80. 9 These figures are based on stores which failed the stage 1 fascia counting filtering test. The OFT is not aware of any reason why including the stores which passed the stage 1 test would not suggest an asymmetric constraint from Asda to Netto. 10 CC groceries report, paragraph 4.85. 5 22. In its groceries report the CC did not conclude that the LADs were constrained by frozen food retailers such as Iceland. The OFT has, however, reviewed other evidence indicating that Iceland should be included in the same product market as Netto. This includes: • analysis based on Kantor Worldpanel data show that around 55 per cent of Netto shoppers also shop at Iceland (making it with Aldi the most popular alternate destination for Netto shoppers) and relative to UK shoppers as a whole are significantly more likely to choose Iceland as their alternate shopping destination11 • the OFT finding in 200812 that Iceland had a comparable number of SKUs to Somerfield (across various product lines) which suggests that Iceland can compete with Netto, at least to some degree, on product range, and • internal Netto customer research finding that [ ].13 23. On this basis, the OFT has taken a broad view of the relevant competitor set in this case, and has included Aldi, Lidl, Iceland, the large grocery multiples, regional multiples and symbol groups in the counting of the relevant fascia during the stage 1 tests (described below and in annexe A) when examining the proposed merger from the point of view of Netto shoppers. It should be noted that evidence on diversion ratios, from customer survey results, are used later in this decision for the purpose of competitive assessment. These provide a clearer indication of the extent of competitive constraint between individual fascia in specific local areas. 11 Based on data from Kantar Worldpanel for the year to 18 May 2010. When all UK shoppers are indexed to 100, the data show that Netto shoppers choosing Iceland as an alternate destination are indexed to 141 which makes Iceland second only to Aldi (index = 155) in the ranking of Netto shoppers' alternative destination. 12 CGL/Somerfield, paragraph 17. 13 [ ]. 6 Geographic market 24. Previous CC and OFT reports into supermarkets have found that there are both national and local aspects to competition.14 The parties did not disagree with this. 25. Based on previous CC and OFT reports, the OFT's proposed candidate geographic market in this case for local level competition is a 10-minute drive time isochrone for one-stop stores in urban areas and a 15-minute drive time isochrone in rural areas.15 For mid-size stores the candidate geographic market is a five-minute drive time isochrone for urban locations and a 10-minute drive time isochrone for rural locations. 26. For Netto, the parties submitted that [ ]. 27. In some past cases the OFT has used the catchment area capturing 80 per cent of a firm's customers or sales as a good rule of thumb in determining a candidate geographic market.16 However, the current case involves a multi-step analytical process which uses a conservative fascia-based filtering exercise within local areas before another exercise is employed which uses customer survey responses. Therefore, the OFT has adopted a cautious approach in the first step of the filtering exercise and allowed the customer survey responses to dismiss competition concerns in some or all of the remaining areas.17 14 For example, Morrisons/30 CGL stores; CGL/Somerfield; Anticipated acquisition by Tesco plc of 45 outlets from Adminstore Ltd, OFT decision of 5 March 2004 (paragraph 11); Competition Commission, 'Safeway plc and Asda Group Limited (owned by Wal- Mart Stores Inc); Wm Morrison Supermarkets plc; J Sainsbury plc; and Tesco plc: a report on the mergers in contemplation', September 2003 (paragraph 2.65). 15 For example, the CC groceries report, CGL/Somerfield and Morrisons/30 CGL stores. 16 For example, Completed acquisition by Aggregate Industries Limited of Atlantic Aggregates Limited and Stone Haul Limited, Case ME/3978/08 (aggregates); and Anticipated merger between Co-operative Group (CWS) Limited and United Co- operatives Limited, OFT decision of 23 July 2007 (funeral services). 17 This is a conservative approach because of the asymmetric constraint. [ ]. 7 28. [ ], the OFT has used a five minute drive time isochrone in the first stage of the filtering exercise with respect to the Netto stores (apart from the one Netto store in a rural area in which the relevant drive time used is 10 minutes). This mirrors the approach taken for mid-size grocery multiple stores. HORIZONTAL ISSUES Competition at the national level 29. The OFT does not consider that competition concerns arise as a result of the proposed merger at the national level. Netto's share of grocery retailing at the national level is less than one per cent and other competitors, such as Tesco, Sainsbury's, Morrisons and Co- operative Group, will remain after the proposed merger to constrain the merged entity at the national level. This reasoning – that the proposed merger does not remove an important constraint at the national level – applies for both unilateral effects concerns and coordinated effects concerns. (Nor does the OFT consider that the proposed merger will create or strengthen coordination at the local level. The OFT has no evidence of pre-existing coordination at the local level nor does the OFT believe that the removal of Netto will make local level coordination more likely.) 30. Vertical effects concerns do not arise in this case. 31. The OFT does not believe that the proposed merger will lead to issues of anti-competitive buyer power in respect of the merged firm's suppliers since the increment to national share of supply arising from the proposed merger is very small. 32. Finally, following evidence submitted by the parties on their holdings of land banks, and taking into account the recent Order on controlled land,18 the OFT does not consider that the proposed merger would result in competition concerns arising through land banking. 18 The Groceries Market Investigation (Controlled Land) Order 2010. 8 33. Therefore, any competition concerns which arise in this case will do so solely through possible unilateral effects at the local level. Unilateral effects at the local level 34. The proposed merger removes Netto as a supermarket operator independent from Asda. As discussed above, the available evidence suggests that Asda is a stronger constraint on Netto than vice versa. As such, the primary theory of harm in this case is that the merger removes, at a local level, the competitive constraint from Asda on Netto, which will in turn enable the merged firm to raise prices at the Netto site or worsen some non-price factor of competition (such as by reducing quality of service, reducing the quality or range of goods offered, or reducing investment levels). 35. In carrying out its investigation, the OFT did not rule out the possibility that there may be some local areas in which an important constraint imposed by Netto on Asda is lost. The CC, in its groceries report, found that LAD stores did not, in general, constrain stores operated by large grocery retailers (but that mid- size and one-stop stores operated by large grocery retailers did constrain the LADs stores).19 However, notwithstanding the generic value of this evidence, the OFT adopted a cautious approach and examined whether there were, at a local level, any areas in which Netto imposes a material constraint on the Asda store, such that its removal through merger would result in a realistic prospect of the Asda store worsening its competitive offering. In practice, however, the diversion ratios from Asda to Netto in every surveyed local area were sufficiently low to dismiss this theory of harm. 36. In order to identify possible competition concerns at the local level the parties adopted a variant of the two-stage methodology that has been employed in other recent supermarket cases. The methodology for this case is detailed in annexe A. However, the key components of the exercise were as follows: 19 CC groceries report, paragraphs 4.81 and 4.84. 9 • Stage 1 filter: A desktop fascia counting exercise was carried out in order to remove from the analysis any local area where, despite both Asda and Netto being present, concerns are highly unlikely to arise. Specifically, areas were removed if there were at least three other relevant competitors present within the relevant isochrones. Isochrones were also re-centred on census output areas in order to ensure that this stage 1 filtering exercise led to a cautious and over-inclusive identification of all possible problem areas. • Stage 2a: Customer surveys were conducted at the Netto stores and used to calculate expenditure-weighted diversion ratios from Netto to Asda. These were then combined with variable profit margins to estimate an illustrative price rise (IPR) for each area, based on a symmetric version of the IPR formula (stage 2a).20 Given the expectation that diversion ratios from Asda to Netto would be lower than those from Netto to Asda, this approach was expected to generate over-estimates of the IPR in each local area. A number of areas were ruled out from further analysis at this stage, on the basis that the (over-stated) estimate of IPR was below the threshold value of five per cent.21 • Stage 2b: For the remaining local areas, customer surveys were then conducted at the Asda stores. The diversion ratios from Asda to Netto, resulting from this survey, were combined with 20 The illustrative price rise is a measure which the CC used in Somerfield/Morrison and which the OFT has subsequently used in some of its supermarket cases (for example, CGL/Somerfield and Morrisons/30 CGL stores). It incorporates the diversion ratio from a store (which provides some indication of the closeness of competition between the merger parties and so whether a store has the ability to raise its prices) and the variable profit margin of a store (which provides some indication of price sensitivity of demand of the store's customers and whether a store therefore has the incentive to raise its prices). Therefore, the illustrative price increase is probative of unilateral effects theory of harm. Importantly, the IPR is not the same thing as merger simulation and therefore it does not try to predict post merger prices nor does it suggest that the OFT is willing to tolerate post merger price increases of up to five per cent. 21 The threshold of five per cent is based on recent grocery merger cases adopting the IPR methodology (for example, Somerfield/Morrisons; CGL/Somerfield). 10