Market definition can be an important analytical step in competition cases, but it only takes us so far. Once markets have been defined, competition authorities and regulators need to assess the competitive interactions between firms in order to understand the impact, for example, of a merger between two firms in the same market. A previous Agenda article focused on market definition in the context of local markets in France. We now turn our attention to competitive assessments by the French Competition Authority in similar local market cases.
Once product and geographical markets have been defined and areas of overlap between the merging parties identified, the first step in the Authority’s competitive assessment is to identify the firms that are exerting a competitive constraint on the merging parties, and their ‘type’ (e.g. supermarkets, specialised shops, independent outlets). The second step is to compute the market shares of the parties and their competitors within each relevant market. Based on market shares and the number of local stores owned by each of the competitors, the Authority uses screening techniques to discard all areas that are unlikely to raise concerns. Finally, with the potentially most problematic areas identified, the Authority undertakes a more detailed competitive assessment, based on local market characteristics, to determine whether divestments are required to alleviate competition concerns.
Finding ‘credible’ competitors
In recent cases concerning non-food retailing mergers in France, the Competition Authority was presented with a number of competitor types that potentially exert a competitive constraint on the merging parties’ behaviour. These typically included large supermarkets and hypermarkets with non-food sales, specialised outlets with a national network (e.g. large consumer goods outlets or professional plumbing product outlets), and more or less specialised outlets that form part of local or regional networks or are even independently owned.
An important element in the Authority’s competitive assessments has always been the role played by independently owned or dedicated specialised outlets. In several cases, the Authority has questioned the ability of independent outlets to exert a strong competitive constraint. This is because these outlets are perceived as being unable to secure favourable deals with their suppliers. As such, they are seen as unlikely to be able to price as low as stores that are part of a national (or regional) network.
In Saturn, a case involving white, brown and grey goods, the Authority considered multi-specialised outlets (‘grandes surfaces spécialisées’, GSS) such as Boulanger or Darty, and hypermarkets and supermarkets (‘grandes surfaces alimentaires’, GSA), to be the relevant competitors to the parties. Kitchen specialists were not considered to exert a strong constraint.
The Brossette and Toutelectric cases involved construction materials sold to builders and installers (plumbing and electrical products, respectively). Here, the Authority drew a distinction between national or multi-regional networks (i.e. those present in at least two of the 21 French regions), independent outlets that are part of purchasing groups, independent outlets that are not part of purchasing groups, and outlets that are owned by their customers (‘coopératives’). National and multi-regional networks were considered strong competitors to the parties. Independent outlets that are part of purchasing groups were also considered ‘credible competitors’ (‘concurrents crédibles’), while independent outlets not part of a purchasing group were considered to exert a limited competitive constraint. ‘Coopératives’ were not considered part of the competitive landscape because, in most cases, only customers that were members of a ‘coopérative’ could buy from it (i.e. demand substitution towards ‘coopératives’ was limited to customers who were already buying there, and customers who were willing to pay the fee to access these stores).
In Coverpro, the Authority made an interesting statement in terms of case law about some of the competitors present locally. Some independent outlets selling construction materials in France had formed a group, not only as a joint-purchasing agreement (which was the case in Brossette), but also in order to trade under the same brand name with the same catalogues, undertake common national advertising campaigns, and act as co-owners of the group. Their business model was therefore close to a franchising model and went beyond a mere joint-purchasing agreement. The Authority argued that, under these circumstances, although market shares are provided at the group level, competitive interactions still occur between individual stores belonging to such groups. As such, the Authority considered that market power needs to be assessed on a store-by-store basis. This consideration does appear to have had some bearing on its Decision to impose divestments, in particular in the area around Cormeilles in Normandy. In this case, the Authority considered that the competitive landscape was ‘very fragmented’ (‘concurrence fortement atomisée’), although there were only six distinct competing groups in the market before the transaction (of which two were groups of independent stores).
Nevertheless, it is not obvious that individual outlets within such groups would retain enough commercial freedom to actively compete against each other. In addition, there is a risk that customers will assume that stores with the same brand sell the same products at the same price, and therefore do not consider these stores to be competitors, thereby muting the competitive constraint that the stores actually impose on each other.
In previous cases involving franchisees, the Authority looked into whether the stores kept an independent purchasing policy as well as independent pricing and branding. In Mr. Bricolage/Passerelle, which involved a merger of DIY stores, the Authority distinguished between two types of stores:
stores that are members of the purchasing group only (and therefore have no minimum purchasing requirements and do not carry any of the group’s brands). These were considered as separate entities to the merging parties;
stores that carry the group’s brands, and are bound by minimum purchase clauses and other constraints on their strategy that prevent them from behaving independently. These stores were therefore considered to contribute to the group’s market power and market share.
Using this type of filter, the ‘groups of independents’ identified in Coverpro might be considered to constitute whole entities for the purpose of assessing market power, although this is not how the Authority approached the issue.
In conclusion, recent cases have led the Authority to take a view on which alternative outlets exert an effective competitive constraint on the parties. The Authority has been sceptical and reluctant to take into account highly specialised competitor stores and independently owned rivals. The competitive constraint exerted by alternative business models such as franchised outlets, customer-owned ‘coopératives’ or very strong purchasing entities is not easy to evaluate, and the Authority has not always taken a consistent approach.
Screening for non-problematic areas
In Saturn, which involved stores selling brown, white and grey goods, market shares were calculated using the amount of space dedicated to each of these types of goods in the stores present in the catchment area of the store under investigation. This approach is similar to that followed in Mr. Bricolage/Passerelle and in all food retailing cases in France, but contrasts with that in Brossette, Toutelectric and Coverpro, where market shares were calculated on the basis of store turnover. The Authority used a combination of criteria based on the parties’ market share, the increase in market share, and the number of (credible) local competitors, in order to eliminate local areas that clearly have no anticompetitive effect.
When screening the 34 overlapping areas identified in Saturn, the Authority first looked at the 18 cases where the parties’ combined market share was lower than 40%. On the basis of a brief overview, it concluded that no anticompetitive effects would result from the merger (particularly because a significant number of alternative stores remained in the market).
In Brossette, having initially identified 196 areas of overlap, the Authority used two consecutive screens to reduce the number of cases for investigation:
73 overlap areas were discarded because at least two national or multiregional competitors were also present in the area. In one case, however, the Authority decided to conduct a more extensive analysis even though three national or multiregional competitors were present, because the parties had high market shares;
70 of the remaining overlap areas were discarded because the parties had a combined market share of less than 35%.
After a short competitive assessment by the Authority, 15 further areas were discarded based on their small increment in market share, the strength of the existing one or two national competitors, and the number and sizes of independent stores. The Authority therefore looked closely at 39 areas in total.
In Toutelectric, of the 37 outlets owned by Toutelectric (the target firm), 33 generated an overlap with the activities of Rexel (the acquirer). On the basis of figures provided by third parties, the Authority calculated the market shares of sales made by all stores located in the catchment areas. To discard any non-problematic cases, it then applied two filters based on market shares:
three areas were discarded because the parties had a market share of less than 30%;
in nine other areas, the parties had a market share below 40% and the increment was below 20% (5–10% in six areas). The Authority considered these cases to be non-problematic.
The Authority therefore looked closely at 21 areas in total.
Competitive assessment of problematic areas
In Saturn, a detailed analysis was undertaken in the remaining 16 areas (14 distinct cities) where the parties’ combined market share (of the sales area) exceeded 40%. This analysis took into account the increase in share contributed by the smallest party; the closeness of competition between the parties (in terms of location and range of products offered); the number, type and size of competing stores; and market dynamics (i.e. how likely competitors were to enter the area). In all areas where concerns were raised as a result of the initial screening, a market test was conducted to assess in more detail the local competitive constraints exerted by competitors, and customers’ actual choices. In seven of the 16 areas, the Authority required the store to be divested.
In Brossette, the Authority looked in detail into the competitive situation around 39 stores owned by Point.P (the party being acquired). In six areas, the transaction led to a merger resulting in a monopoly, and a divestment was immediately required in order to remove the overlap. Further divestments were systematically required in duopoly or quasi-duopoly situations—which was the case in nine other areas.
Seven areas related to stores within the Parisian urban area, where the competitive situation is slightly different from other areas as a result of a dense transport network with heavy traffic. As a consequence of multiple overlapping areas within this area, a set of divestments was designed to simultaneously address all the local competitive issues identified by the Authority. Although not all overlaps were removed, the divestments were designed to restore competition to pre-transaction levels. The 17 remaining areas were all subject to a detailed local assessment which led to four additional divestments.
Before the Toutelectric Decision, Rexel (the acquiring firm) was the largest national network of outlets specialising in electrical equipment, tools and devices, with 30–40% of the market nationally and 380 stores. Toutelectric (the target firm) was small in comparison, with 0–5% of the (national) market and 37 stores, mostly in the South West of France. Two other national networks held 30–40% of the market (approximately 500 outlets) and 0–5% of the market (approximately 300 outlets) respectively. A number of independent outlets were also present, some of which were part of purchasing groups.
Despite conducting a case-by-case-assessment, the Authority in general authorised transactions that were basically 4-to-3 or even 3-to-2 mergers in all local areas, sometimes with very high market shares. In fact, the Authority explained that, in this market, calculating market shares in this way might overestimate market power. In particular, this is because customers are sophisticated and able to force outlets to compete with one another (e.g. through phone orders and deliveries), and because barriers to entry are low.
Although the Toutelectric and Brossette Decisions were taken within a month of each other, the difference in their approaches is striking. In Brossette, the Authority appeared reluctant to accept ease of entry, low barriers to entry and customers’ ability to play outlets off against each other; whereas, in Toutelectric, it fully endorsed such arguments, despite a relatively high degree of existing overall concentration at a national and local level.
Beyond the geographical definition
Once local markets are defined, the French Competition Authority’s attention turns to calculating market shares and, in particular, to identifying credible competitors to the parties’ businesses. Recent Decisions also refer to ways of screening out non-problematic local areas in order to focus detailed analysis on the (potentially) most problematic ones. These screens are based on market share estimates for sales (or store sales areas) of all stores within the relevant market, and sometimes on the number of competitors present locally.
The Authority undertakes careful analysis of cases that have been identified through the screening process as being potentially problematic. However, recent Decisions suggest that the Authority has made inconsistent use of, and reference to, factors that affect all local markets. In particular, it has considered entry barriers to be high or low in cases where such barriers do not appear to differ materially. The ability of customers to play outlets off against each other was also accepted in some cases but not in others, although the types of customer involved were likely to have been similar.