The Competitive Assessment of Mergers in Retail Grocery Markets: A Basket Case?

This paper shows that competition authorities in Europe have adopted very different approaches to assess the competitive effect of horizontal mergers on retail grocery markets. Specifically, we highlight substantial differences between the UK CMA, the German Cartel Office, and the French Competition Authority. All these authorities have considered the effect of mergers between grocery retailers on local markets where the merging parties overlap, however, the method and scope of the competitive assessment they have developed vary significantly. The CMA has demonstrated willingness to apply a mechanistic approach, recently using an upward pricing pressure test in their investigations. This may lead to less discretion when compared with the approach taken by the German Cartel Office or the French Competition Authority, which appears more qualitative in nature despite heavy reliance on the degree of post-merger market concentration. Further, the CMA approach applies a very low threshold for intervention, while in contrast the Adlc has been significantly more lenient, in instances finding no competition concerns in local markets where the merged entity would have at least 50 percent market share.

By Benoît Durand¹

 

I. INTRODUCTION

One major motivation behind retail grocery chains looking to merge is to achieve synergies, notably those associated with scale. Indeed, bigger size can generate significant cost advantages. In theory, such size advantage can be achieved organically, but constraints on growth, in particular planning authorization, often direct retail chains to finding mergers as the most practical way to generate cost reductions.

Cost advantages that stem from bigger scale span several dimensions. For instance, by grouping their purchases, two grocery retailers can obtain better deals from suppliers, notably through volume discounts. Such cost reductions can be passed on to consumers. Mergers between retail chains may also generate other types of savings, including around logistics and product assortment. These synergies can be significant and strengthen the competitive ability of merging parties. Indicative of this is the 2016 merger between Ahold and Delhaize, which generated €750 million in cost savings over three years, exceeding expected synergies.²

However, horizontal mergers can also give rise to more concentrated retail grocery markets. Hence, competition authorities may have concerns that these transactions could confer significant market power to the merged entity. This in turn may manifest in higher prices for the supply of groceries or, more globally, may worsen the retail offering.³ This worsened retail offering may involve additions to price, diminished quality, a reduced range of products, and a narrowing of services delivered to customers.⁴

In Europe, mergers in the retail grocery sector are reviewed by national competition authorities, as these transactions typically concern only one country. Recently, authorities blocked or threatened to block some of the largest mergers between grocery retailers. In 2019, the UK Competition Market Authority (“CMA”) prohibited the merger of Sainsbury’s and Asda.⁵ In 2016, Auchan and Système U abandoned their plan to merge after the Autorité de la concurrence (“Adlc”) opened an in-depth investigation.⁶ And in 2015, the German Bundeskartellamt (“BKartA”) published its decision to block the acquisition of hundreds of Kaiser’s Tengelmann stores by EDEKA.⁷

These decisions show that national competition authorities can intervene to preserve competition on retail grocery markets. However, behind these decisions, there are major differences in approaches to evaluate the risk these mergers pose to retail competition. Moreover, and crucially, there are significant variations in the threshold applied for intervention across jurisdictions, such that the same merger could have a very different fate depending on which authority reviews the transaction.

Some authorities, in particular the CMA, have developed a rather mechanistic approach to assessing the competitive effect of retail grocery mergers. A major motivation behind this is that, otherwise, some transactions may require the authority to conduct an in-depth investigation of hundreds of markets. By way of example, the CMA identified 537 local areas in which the merger between Sainsbury’s and Asda may give rise to a substantial lessening of competition.⁸ In this case the CMA heavily relied on the outcome of an upward pricing pressure test (GUPPI) to determine whether the merger would give rise to a substantial lessening of competition in local markets.⁹ While there are good reasons to deploy a systematic, mechanistic approach when the number of grocery stores involved is very large, this algorithmic method fails to account for dimensions of competition that cannot be captured by quantitative tools. For instance, competition in retail grocery markets includes factors such as quality of service, which may involve investment to refit and refurbish local stores or spending on training for staff to improve the level of service. These factors are not taken into account in upward pricing pressure tests.

In sharp contrast, the BKartA claims to rely on a broad approach in its review of EDEKA’s acquisition of around 450 Kaiser’s Tengelmann stores, even though the transaction involved hundreds of stores. The BKartA expressly dismissed the “mathematical” approach, claiming to conduct a thorough assessment that accounts for several important factors, notably those related to market structure.¹⁰

Across jurisdictions, however, competition authorities do not consider post-merger local market entry or expansion by rival retailers as likely. This is despite recognition that competitors’ reactions could maintain prices at the competitive level. In large part, the authorities can attribute this simplification to the fact that opening a new supermarket requires authorization from local authorities, which even if granted, might take significant time. In some local markets, however, such a possibility might exist, but to assess whether entry is likely and timely would require a detailed analysis in each market. While authorization for a new supermarket may not be forthcoming, it may be easier for a grocery store to open if a merger has reduced local competition. This new entry could significantly boost competition, and potentially apply downward pressure on prices. Furthermore, competitors may respond to a price increase by repositioning their product offering such that greater competitive pressure is applied on the merged entity. Such a move may also reduce prices.

 

II. DEFINING LOCAL MARKETS

All competition authorities consider that retail grocery markets are local, sometimes involving very small territories, which means that when large retailers notify a transaction, the regulatory review may involve an in-depth investigation of hundreds of markets. Authorities, however, have defined local markets differently. The CMA and the Adlc consider the catchment area of the merging parties’ supermarkets as the relevant market, while the BKartA defines geographic markets around administrative boundaries.

A. Local v. Nationwide Competition

To assess the competitive effect of retail grocery mergers, competition authorities define local markets despite transactions involving nationwide retail grocery chains. The main reason why competition authorities focus on the competitive effect of grocery mergers in local areas is simply because consumer demand is mostly from the area in question.¹¹ Price sensitive customers are highly unlikely to travel long distances to find another grocery store if prices are raised at the supermarket they usually visit. These customers will typically consider local alternatives.

There is plenty of evidence that this is the case. For instance, the BKartA and the CMA report that the entry or exit of grocery stores has an impact on the revenue of local rivals.¹² Further, according to the CMA, its exit-store survey conducted in the Sainsbury’s/Asda case shows that most customers of a given supermarket would not consider shopping in stores that are more than a 15-minute drive-time from the store in question.

Competition authorities have also noted that many aspects of competition are essentially local. That is, stores can flex local retail offerings to attract customers away from competing stores in proximity. For example, the range of available products at a store or the level of service offered may depend on the local competitive conditions. These parameters can be flexed by grocery retailers to improve their local offer in areas where competition is more intense. Even if part of a national retail chain, local stores may have a degree of control over their pricing, with the possibility of applying discounts to the price of some key items. Additionally, a retail grocery chain can decide to refurbish individual local stores to improve the layout and create new amenities, which will strengthen the local competitive position of these stores.

Some aspects of competition, however, appear to be set over a much broader geographic scope than local areas. In some cases, key parameters of competition are decided centrally and applied uniformly across all stores of a retail chain. In the UK, for example, the Competition Commission had determined that prices of major grocery retailers were largely uniform. That is, the price of goods is the same across all stores of a major retailer. On that basis, it has been argued that the effect of a merger will not be felt locally but across all stores of the merging parties, whether the stores overlap or not.¹³ In that context, it might be more relevant to define broader geographic markets, potentially nationally, as large brand chains compete nationally.

Competition authorities have resisted this argument. In large part this is because the main source of competitive constraint is essentially local, even if it is the case that prices are set nationally. One way to see this is to consider a simple hypothetical example. Consider that two retail chains, A and B, seek to merge, they set price uniformly, and their national market share is respectively 12 percent and 7 percent. If the geographic market is defined at the national level, in principle such a transaction would be unlikely to give rise to competition concerns. However, upon a closer look, the parties’ stores overlap in a significant number of concentrated local areas in the west of the country, and in fact, in many local markets, this merger would lead to local monopoly. Further, if the two grocery retailers do not have many other stores elsewhere in competition with other retailers, then post-merger prices can be expected to increase to the monopoly level. An assessment at the national level would fail to see this.¹⁴

Equally, however, if prices are set uniformly, a narrow focus on the local analysis of mergers may overstate the effects of the transaction, failing to account for the competitive constraint exerted by rival grocery retailers. One way to see this is to use another simple example. Consider that two grocery retailers, C and D, seek to merge, and they respectively have 500 and 400 stores, but they overlap only in three local markets. The price set by chain C in all of its supermarkets is the result of the competitive constraint that its 500 stores face locally. In particular, if the 500 stores face intense local competition, then C’s price will be set at a very low level. In this context, a merger that relaxes competition for only three stores is unlikely to affect nationwide prices significantly (assuming that the retail grocery chains continue to set price uniformly across all stores).

B. Defining the Boundaries of Local Grocery Markets

While competition authorities consider that markets are local, each has developed its own approach to define the boundaries of individual geographic markets. Importantly, none of these approaches are strictly in line with the hypothetical monopolist test, which is the method to define antitrust markets. In most cases, the geographic market is defined around a center point, and the boundaries are found using a radius from that center, either measured as a straight line in kilometers/meters or as drive-time, which looks at the car journey time of customers. As we explain below, however, not all competition authorities define the center point in the same way.

The German BKartA defines the geographic market around a regional center (“regionales Oberzentrum”), using either a 20 km radius or 20-minute drive-time.¹⁵ Every postal code area that falls within this range is assigned to that relevant market. The BKartA thus splits Germany into many geographic markets – 345 to be exact.¹⁶ That said, in the case of large cities over 500,000 inhabitants, the BKartA defines markets according to smaller territories, following the boundaries of boroughs (“Stadtbezirke”).¹⁷

This approach is imperfect in several ways. First, it fails to account for the possibility that stores in differently defined geographic markets can compete. For instance, in large cities the BKartA follows the boundaries of boroughs yet stores in separate administrative territories can compete. Indeed, two stores in such close proximity that they are obviously competing for the same customers could still be separated into different geographic markets according to the rule applied by the BKartA. Equally, two supermarkets that are far apart from one another, for example a distance over 35-minute drive-time, may not compete, or not compete as closely as nearby stores, yet they would belong to the same geographic market.

Other competition authorities, such as the UK CMA and the French Adlc, have defined local markets using each supermarket as the center point. Specifically, each local market corresponds to the catchment area of the merging parties’ stores, which means that the number of geographic markets will be very large if the grocery retailers have many supermarkets. They do not provide justification for this approach, which differs significantly from the one adopted by the BKartA, but we can conjecture that it might be more pragmatic to consider the competitive constraint exerted on a given grocery store by examining the number and identity of competing stores within its catchment area.

The size of the catchment area varies according to the size of the grocery store in question.¹⁸ Large grocery stores have wide catchment areas, as people may travel greater distances to do their weekly shopping. Small grocery stores have small catchment areas. Competition authorities make such a distinction. For example, in the UK, the CMA considers that large stores are constrained by other large stores within 10 (15) minute drive-time in urban (rural) areas.¹⁹ The rule is different for smaller stores. Medium stores are constrained by other medium stores within 5-minute drive-time and by large stores within 10-minute drive-time in urban area.²⁰ (The catchment area is broader in rural areas).

As noted, the size of the catchment area can vary depending on whether stores are in densely populated areas or in rural areas, which is clear from the rules set out by the CMA. Competition authorities, however, can go beyond the simple arbitrary split between rural and urban areas. The French Adlc considers reducing the boundaries of the catchment area depending on population density, which leads to a series of rules. The starting position is that the catchment area of a hypermarket is 30-minute drive-time and that of a supermarket is 15-minute drive-time.²¹ However, for the 10 most populated cities in France, the rules become 20- and 10-minute drive-time respectively.²² In the Paris region, the catchment area of grocery stores is smaller. In the close suburbs of Paris (known as la petite couronne), the authority uses 10-minute drive-time as the starting position.²³ In Paris, customers are expected to walk to do their grocery shopping as a very small proportion of residents own a car and the size of grocery stores is typically small. For Parisian stores greater than 400 square meters, the catchment area is defined by the territory within 500 meters of the store, and for smaller stores, the size of the catchment is based on a radius of 300 meters.²⁴

This approach also appears to be imperfect. In theory, retail grocery stores compete if their catchment areas overlap. In practice, however, authorities consider that stores compete only if they are in each other’s catchment area. This approach, while practical, is necessarily flawed for the following reasons:

• First, it may fail to identify some of the merging parties’ stores that compete. Supermarkets with overlaps in their catchment areas but not located in each other’s catchment area may compete for customers located in these overlaps. This means that grocery stores outside of the catchment area defined by the authority can exert a significant competitive constraint on the center store. A possible solution to this would be to adjust isochrones according to population center, though this would require detailed local information on population density.Second, it may lead to an overly narrow scope of analysis by misrepresenting competition conditions in local markets. If overlaps between catchment areas are not accounted for, the resulting analysis may be flawed. By failing to consider the competitive constraint exerted by stores outside the catchment, this analysis may be biased towards finding a lessening of competition.

 

III. ASSESSING THE COMPETITIVE EFFECTS OF MERGERS ON GROCERY RETAIL MARKETS

Depending on the extent of overlaps, and because geographic markets are defined narrowly, mergers between grocery retail chains can affect many markets. To address this large number problem, competition authorities typically deploy a two-step approach. First, they apply a filter to identify local markets which require further investigation; then, they assess the competitive effect of the merger in each of these markets.

In some mergers, the authority may have to investigate hundreds of markets even after the application of a filter. Interestingly, competition authorities have developed very different approaches when confronted with this issue. In Sainsbury’s/Asda, the CMA relied almost solely on a mechanistic approach based on upward pricing pressure tests, which is calculated for all overlapping stores. In sharp contrast, the French Adlc and the BKartA’s assessment is based largely on traditional concentration measures, such as market shares and the number of competitors. In Casino/Monoprix, the French Adlc relied on the Parties’ market shares and the number of competitors’ stores in each local market, while in EDEKA/Kaiser’s Tengelmann the German BKartA considered the degree of market concentration in large cities, not focusing on any particular narrow geographic market.

A. Screening Local Grocery Markets

Because the number of local retail markets that are potentially affected by a horizontal merger between two large grocery retail chains can be very large, competition authorities apply a filter based on concentration rules as a first step to screen out markets. This is done to ensure that they investigate the likely effect of a merger only in sufficiently concentrated markets. However, the concentration rules they apply vary substantially.

Until the Sainsbury’s/Asda merger, the CMA (and the OFT before it) consistently used a simple fascia count. That is, if a merger resulted in a reduction of grocery chain fascia from four to three or fewer in the center store’s catchment area, then the store in question was flagged for further inquiry; this occurred irrespective of the size of the merging parties in the affected market.²⁵

In EDEKA/Kaiser’s Tengelmann, the BKartA applied a simple market share threshold. It considers that the merger could restrict competition in every geographic market in which the parties’ combined market share exceeds 25 percent in the overall retail food market (“Gesamtmarkt”), which includes full range grocery retailers as well as discounters.²⁶

The French Adlc adopted perhaps the most convoluted set of filters to screen out markets where it is unlikely that the merger could result in competition concerns. This approach, which is set out in Casino/Monoprix, rests on a combination of rules based on market share and HHI thresholds. Market shares are calculated using the size of stores, or if possible store area dedicated to the supply of groceries. First, the Adlc applies a simple rule based on the dominance criterion. Every local market in which the parties’ combined share exceeds 50 percent is investigated further.²⁷ Moreover, the Adlc considers that markets where the combined market share of the merging parties is between 45 and 50 percent and the following conditions apply also merit further inquiry:

• markets in which the HHI level is between 1000 and 2000, and the delta HHI (the increment in the HHI level caused by the merger) would be at least 250; or
• markets in which the HHI level is greater than 2000, and the delta HHI would be at least 150; or
• markets where the market share increment is at least 10 percent.²⁸

For the city of Paris, which is viewed as different from the rest of the country, notably because most residents do not own cars and tend to spend less on each shopping trip, the French Adlc established a different set of rules to screen out markets. The Adlc decided that in markets where the merging parties’ combined share is below 60 percent, where there are at least three competing fascia (with stores greater than 120 square meters) and at least half of the stores are in the hands of rivals, the merger would not raise significant competition concerns.²⁹

B. Assessing the Competitive Effect of Retail Grocery Mergers

In principle, competition authorities investigate the competitive effect of a merger only in local retail markets where the merging parties overlap and the increase in market concentration is deemed sufficiently high, although the CMA, and to a lesser extent the BKartA, have considered the nation-wide effect of transactions. The approach that authorities have developed to predict the competitive effect of these mergers on numerous markets varies significantly, however.

The BKartA and the French Adlc relied mostly on market shares, though the threshold used by the Adlc appears significantly more lenient than that adopted by the BKartA.

For the EDEKA/Kaiser’s Tengelmann merger, the BKartA first considered closeness of competition between the merging parties, but also their competition with rival grocery retailers in the wider food retail market, not focusing on any particular local market.³⁰ For each retailer the BKartA examined several characteristics, such as the size of stores, the range of goods, as well as the proportion of private labels in retailers’ profit margins. In addition, the BKartA developed several data analyses, looking in particular at how retailers overlap in the supply of particular brands of goods, similarities in the baskets of their respective customers, as well as the impact of entry/exit on rival grocery retailers.³¹ The BKartA concluded that the wider food retail market could be segmented between full-range grocery retailers and discounters, and within full-range retailers, the market can be further segmented. In particular, the BKartA found that a group of three major retailers, EDEKA, Kaiser’s Tengelmann, and Rewe, are very close competitors. This led the BKartA to conclude that the merged entity would face competition from only one very close competitor: Rewe.³²

Second, using a series of concentration measures, the BKartA determined whether the merger would result in a significant impediment to effective competition in local markets. The German authority did not examine each local market in any detail, but rather favored a broader overarching view. The BKartA took into account the market share of the merging parties in the overall food retail market, but also their market position in a narrower segment comprising full-range retailers.³³ For example, in Berlin, the BKartA found that in most districts of the city, the merging parties and Rewe have a combined market share of at least 50 percent in the overall food retail market.³⁴ Further, the authority found that in the narrower full-range retail segment, EDEKA, Kaiser’s Tengelmann, and Rewe would together control about two-thirds of the market.³⁵ The BKartA concluded that the proposed transaction would lead to the loss of competitive pressure from Kaiser’s Tengelmann as well as giving rise to dangerously high market concentration, especially given the reduction of three to two competitors in the full-range grocery retail segment. ³⁶

Concluding their analysis, the BKartA found that the merger between EDEKA and Kaiser’s Tengelmann restricted competition in a series of markets, and in particular in large cities such as Berlin, Munich, and Düsseldorf.³⁷

When conducting in depth investigation of markets flagged by its market concentration filter in Casino/Monoprix, the French Adlc mostly considered features of the local market structure in determining its overall findings regarding competitive effects. For instance, in two markets, Arles and St Raphael, despite the merged entity having a market share above 50 percent and a market share increment greater than 5 percent, the Adlc ruled out competition concerns on the basis that there were five competing fascia in the catchment area, which had many competing stores.³⁸ In Ajaccio, even though the merged entity would have a market share above 45 percent and it would face only two competing fascia, the Adlc found that the merger would not stifle competition. The reasons invoked for this decision are that two-thirds of stores are owned by rivals and one of the competing fascia, Carrefour, has a market share of 40 percent.³⁹

In Paris, the Adlc applied a less interventionist filter than in the rest of the country. In spite of this, it had to investigate 47 local areas in greater depth for the Casino/Monoprix transaction.⁴⁰ The competitive assessment essentially boiled down to determining the extent to which hypermarkets (stores which dedicate at least 2,500 square meters for groceries) owned by rival grocery retailers across locations just beyond and inside Paris exert a competitive constraint on the merging parties’ stores in Paris. Using the results of various analyses, the Adlc made downward adjustments to the merging parties’ market share in some local areas to account for this particular constraint.⁴¹

Unlike the German BKartA and the French Adlc, which rely mostly on concentration measures (i.e. market shares and number of competitors), in Sainsbury’s/Asda the CMA relied exclusively on the results of an upward pricing pressure test, namely the GUPPI, that is calculated for each market in which the merging parties overlap. The CMA found that the merger would give rise to a significant lessening of competition in markets where the GUPPI is above 2.75 percent when merger-specific efficiencies are accounted for, and above 1.5 percent without efficiencies. These thresholds are particularly low, and lead to an adverse finding in many overlapping markets.⁴²

The GUPPI is based on two inputs, the diversion ratio between the merging parties and their profit margin. The higher the parties’ diversion ratio and the higher their profit margin, the greater the GUPPI will be. In past cases, diversion ratios were estimated using an exit-store survey. That is, customers shopping on the day of the survey at the center store were asked where else they would shop if the store in question were no longer in operation. On the basis of the surveyed customers’ response, it was possible to estimate the proportion of customers who could consider switching to the other merging party.

In Sainsbury’s/Asda, the parties overlapped in many markets, therefore conducting a customer survey in all these markets would have been not only costly but also time consuming.⁴³ The CMA developed an alternative approach, called the weighted share of shops (WSS), which assigns a weight to each store in the local market based on the brand, its size and its distance from the center store.⁴⁴ All else equal, if a store is large, closer to the parties’ center store, or if it belongs to a brand seen as exerting a greater competitive constraint, it shall receive a higher weighting. For each local area considered, with one of the merging parties’ stores at the center, the weights add up to 100 percent. These proportions were then used to compute the relevant diversion ratios between the merging parties.

In sharp contrast to the approach followed by the German BKartA and the French Adlc, the CMA did not rely on traditional concentration measures, such as the parties’ market share and the number of competitors, as these may not adequately reflect the extent of the competitive constraint exerted by rival grocery retailers in each local market. Furthermore, through diversion ratios, the CMA took into account the competitive constraint exerted by discounters, such as Aldi and Lidl, which are not always included in the relevant antitrust market, as well as what it termed “out-of-market” constraints, which include bargain stores, specialist stores but also supermarkets located outside the immediate local areas (i.e. more than 10- or 15-minute drive-time from the center store).⁴⁵

Unlike market shares or other concentration measures, this method takes into account the fact that rival local supermarkets apply a varying degree of competitive pressure on the center supermarket. In the Asda/Sainsbury’s case, this manifested in greater weights for large Tesco stores than medium Waitrose stores at a given drive-time distance, for example.

The CMA found that the Sainsbury’s/Asda merger would result in a significant lessening of competition in 537 local markets centered on the Parties’ supermarkets.⁴⁶ Importantly, this represents about 40 percent and 50 percent of Sainsbury’s and Asda’s stores, respectively.

 

IV. CONCLUSION

As we have shown in this article, major competition authorities in Europe have not applied a consistent approach to predict the competitive effect of mergers on retail grocery markets. Crucially, the threshold for intervention varies substantially across jurisdictions. That said, they share an important, common feature: the result of the competitive assessment is based on a static analysis. This is because in each jurisdiction, authorities have found that there are significant barriers to expansion and entry in the grocery sector, therefore finding reactions from competitors appear quite unlikely. This implies that for mergers that lead to more concentrated local markets, unless the merging parties can convince the authorities that significant synergies would offset any predicted anti-competitive effects, the merger will be cleared only if the parties agree to divest stores to reduce or eliminate the overlap in the local markets at issue. The extent of the divestment package depends, however, on the threshold for intervention, which can be particularly low in some jurisdictions.

---

1 Partner at RBB Economics. The author would like to thank Lachland Lindsay and Johanna Lunghini for excellent research assistance. The views expressed in this article are the author’s own and do not reflect the view of RBB Economics.

2 See https://www.grocerydive.com/news/grocery–ahold-delhaize-profits-surge-68-with-merger-savings/534814/ and https://www.aholddelhaize.com/media/8807/ahold-delhaize_q4fy18_summary-report.pdf.

3 Some competition authorities have also been concerned that mergers between two large grocery retailers would confer the merged entity significant buyer power, at the expense of suppliers.

4 The CMA refers to these price and non-price factors as “PQRS,” for price, quality, range and service.

5 See https://www.gov.uk/government/news/cma-blocks-merger-between-sainsburys-and-asda

6 See https://www.autoritedelaconcurrence.fr/en/communiques-de-presse/24-may-2016-retailing-sector

7 See https://www.bundeskartellamt.de/SharedDocs/Entscheidung/EN/Fallberichte/Fusionskontrolle/2015/B2-96-14.html?nn=3591568. The German ministry of economic affairs, however, vetoed the decision, which de facto allowed the merger to go through.  (See https://www.bundeskartellamt.de/SharedDocs/Entscheidung/EN/Fallberichte/Kartellverbot/2017/B2-31-17.html?nn=3600108.)  However, that ministerial authorization was appealed before the Higher Regional Court of Düsseldorf, which declared the ministerial decision to be unlawful. (See https://www.vbb.com/insights/competition/merger-control/edekakaisers-tengelmann-german-merger-review-saga-comes-to-an-end.

8 See paragraph 3b, https://assets.publishing.service.gov.uk/media/5cc1434ee5274a467a8dd482/Executive_summary.pdf.

9 See https://www.rbbecon.com/publications/brief-60-sainsburysasda-and-the-cmas-guppi-decision-rule-on-the-money-or-basket-case/.

10 See paragraphs 239 and 409 of the decision (BKartA, B2-96/14, EDEKA/Kaiser’s Tengelmann, 31.03.2015).

11 Andreas Mundt, President of the Bundeskartellamt, said in reference to the EDEKA/Kaiser’s Tengelmann case: “In this case the local market situation is the deciding factor. The argument that Kaiser’s Tengelmann has a relatively low share of the national market misses the point. No one travels across Germany to go shopping, not even across a large city.” https://www.bundeskartellamt.de/SharedDocs/Meldung/EN/Pressemitteilungen/2015/01_04_2015_Edeka_Untersagung.html;jsessionid=581104F2EED97BC053C8762F522F79E9.1_cid371?nn=3591568.

12 See the CMA decision on Sainsbury’s/Asda and the BKartA decision on EDEKA/Kaiser’s Tengelmann.

13 As indicated in the UK Competition Commission Final Report on the Groceries Market Inquiry at paragraph 4.98, “Tesco submitted that the prevalence of nationally set, and largely uniform, aspects of the retail offer for grocery stores indicates a national rather than a local geographic market.” The Competition Commission also noted that most retailers set their prices uniformly, or mostly uniformly, across their store network. See Competition Commission, The supply of groceries in the UK market investigation, April 30, 2008.

14 Paul Dobson and Michael Waterson, “Chain-store pricing across local markets” Journal of Economics and Management Strategy (Spring 2005) 14, 1, 93-119.

15 See paragraphs 217 and 236 of the decision (BKartA, B2-96/14, EDEKA/Kaiser’s Tengelmann, 31.03.2015).

16 See paragraph 209 of the decision (BKartA, B2-96/14, EDEKA/Kaiser’s Tengelmann, 31.03.2015).

17 See paragraph 215 of the decision (BKartA, B2-96/14, EDEKA/Kaiser’s Tengelmann, 31.03.2015).

18 In the 2008 groceries market investigation, the UK Competition Commission recalls that the catchment area is the area from which a store draws most of its customers. For example, the Commission shows that a majority of customers (more than 80 percent) shop at large stores within a 15-minute drive-time. See Final Report at paragraph 4.103.

19 Grocery stores that are greater than 1,400 square meters are named large stores by the CMA. The CMA considers that these stores are suitable for one-stop shopping.

20 The CMA defines medium stores are those which have a size between 280 and 1,400 square meters. Stores under 280 square meters fall in the convenience store category.

21 See paragraph 96 of the decision (Autorité de la concurrence, 13-DCC-90, Casino/Monoprix, 11.07.2013).

22 See paragraph 102 of the decision (Autorité de la concurrence, 13-DCC-90, Casino/Monoprix, 11.07.2013).

23 See paragraph 112 of the decision (Autorité de la concurrence, 13-DCC-90, Casino/Monoprix, 11.07.2013).

24 See paragraph 105 of the decision (Autorité de la concurrence, 13-DCC-90, Casino/Monoprix, 11.07.2013).

25 See CMA, Retail mergers commentary, April 10, 2017, CMA62 at paragraph 3.35. The CMA has used the four-to-three fascia count threshold in mergers in the grocery sector. This was first used in Safeway (2003).

26 See paragraph 410 of the decision (BKartA, B2-96/14, EDEKA/Kaiser’s Tengelmann, 31.03.2015).

27 See paragraph 189 of the decision (Autorité de la concurrence, 13-DCC-90, Casino/Monoprix, 11.07.2013).

28 See paragraph 198 of the decision (Autorité de la concurrence, 13-DCC-90, Casino/Monoprix, 11.07.2013).

29 See paragraph 274 of the decision (Autorité de la concurrence, 13-DCC-90, Casino/Monoprix, 11.07.2013).

30 See paragraph 331 of the decision (BKartA, B2-96/14, EDEKA/Kaiser’s Tengelmann, 31.03.2015).

31 See paragraph 259 of the decision (BKartA, B2-96/14, EDEKA/Kaiser’s Tengelmann, 31.03.2015).

32 See paragraph 325 of the decision (BKartA, B2-96/14, EDEKA/Kaiser’s Tengelmann, 31.03.2015).

33 See paragraph 414 of the decision (BKartA, B2-96/14, EDEKA/Kaiser’s Tengelmann, 31.03.2015).

34 See paragraph 440 of the decision (BKartA, B2-96/14, EDEKA/Kaiser’s Tengelmann, 31.03.2015).

35 See paragraph 457 of the decision (BKartA, B2-96/14, EDEKA/Kaiser’s Tengelmann, 31.03.2015).

36 See paragraph 468 of the decision (BKartA, B2-96/14, EDEKA/Kaiser’s Tengelmann, 31.03.2015).

37 See paragraphs 2-5 of the decision (BKartA, B2-96/14, EDEKA/Kaiser’s Tengelmann, 31.03.2015).

38 See paragraphs 191-192 of the decision (Autorité de la concurrence, 13-DCC-90, Casino/Monoprix, 11.07.2013).

39 See paragraph 200 of the decision (Autorité de la concurrence, 13-DCC-90, Casino/Monoprix, 11.07.2013).

40 See paragraph 275 of the decision (Autorité de la concurrence, 13-DCC-90, Casino/Monoprix, 11.07.2013).

41 See paragraph 353 of the decision (Autorité de la concurrence, 13-DCC-90, Casino/Monoprix, 11.07.2013).

42 See RBB Brief 60, October 2019 for a more detailed discussion of the particularly low GUPPI thresholds adopted by the CMA in this case. (https://www.rbbecon.com/downloads/2019/10/RBB_Brief-60.pdf). It is worth noting that in its commentary on retail merger, the CMA appears to suggest that it find concerns when GUPPI is above 5 percent. See CMA, Retail merger commentary, April 10, 2017, at 5.14.

43 The Parties operate over 1,000 supermarkets across the UK. See paragraph 4.1.

44 The CMA conducted an exit survey that covered 100 stores, 50 Sainsbury’s and 50 Asda supermarkets. The results of the survey together with the finding of an empirical analysis on the impact of entry and exit has been used by the CMA to inform the weight in relation to brand, store size and distance from the center store.

45 Out-of-market constraints, such as non-supermarket stores, brands other than the supermarket chains analyzed, online retailers, and supermarkets outside of the catchment area, were allocated a 25 percent weighting in the estimation of diversion ratios.

46 See paragraph 8.90.