EC, July 22, 1992, No M.190

THE COMMISSION OF THE EUROPEAN COMMUNITIES,

Having regard to the Treaty establishing the European Economic Community, Having regard to Council Regulation (EEC) No 4064-89 of 21 December 1989 on the control of concentrations between undertakings (1), and in particular Article 8 (2) thereof, Having regard to the Commission decision of 25 March 1992 to initiate proceedings in this case, Having given the undertakings concerned the opportunity to make known their views on the objections raised by the Commission, Having regard to the opinion of the Advisory Committee on Concentrations (2), Whereas:

I. BACKGROUND

(1) On 25 February 1992, Nestlé SA notified a public bid for 100 % of the shares of Source Perrier SA which was launched by Demilac, a jointly controlled subsidiary of Nestlé and Banque Indosuez. Nestlé has an option to purchase the shareholding of Banque Indosuez in Demilac, which it has announced it will take up. On 30 January 1992, Nestlé concluded an agreement with BSN following which the Volvic source of Perrier will be sold to BSN if Nestlé acquires control over Perrier.

(2) By decision dated 17 March 1992, the Commission continued the suspension of the concentration pursuant to Article 7 (2) of Council Regulation No 4064-89 (the Merger Regulation). Nestlé has acquired the majority of the shares in Perrier but is refrained from exercising the voting rights attached to these shares pursuant to Article 7 (3) of the Merger Regulation.

(3) By decision dated 25 March 1992, the Commission declared that the proposed concentration raised serious doubts as to its compatibility with the common market. The Commission therefore initiated proceedings in this case pursuant to Article 6 (1) (c) of the Merger Regulation.

II. CONCENTRATION AND COMMUNITY DIMENSION

(4) Nestlé has notified a public bid by which it intends to acquire control over Source Perrier SA and its subsidiaries (hereinafter referred to as 'Perrier'). The operation would thus lead to a concentration within the meaning of Article 3 of the Merger Regulation. The undertakings concerned by the concentration, i.e. Perrier and Demilac, including the turnover of Nestlé, meet all the thresholds of Article 1 (2) of the Merger Regulation.

III. THE UNDERTAKINGS

NESTLÉ

(5) Nestlé is a publicly held Swiss company which is active in many sectors of nutrition. The Nestlé group manufactures and sells food products throughout the world. Nestlé has a consolidated worldwide turnover of more than ECU 28 000 million.

PERRIER

(6) Perrier is a French company which is mainly active in the manufacture and distribution of bottled waters. Perrier has also some activities in the cheese market. Perrier has a consolidated worldwide turnover of over ECU 2 000 million.

IV. AFFECTED MARKETS

(7) The proposed concentration between Nestlé and Perrier affects primarily the business of bottling water originating from a natural spring or source ('source water'). Source waters may be labelled as 'mineral water' provided they fulfil certain legal requirements regarding their composition and their quality, and provided they obtain an authorization from the competent authorities to that effect. Source waters which are not mineral waters will hereafter be referred to as 'spring waters'. The production and marketing of spring waters are also subject to an authorization from the competent authorities.

Nestlé's bottling activities in the source water sector are mainly located in France with the well-known brands Vittel and Hépar (turnover in 1991 : FF 1 564 million, with sales of 916 million litres) and in Germany (turnover in 1991 : DM 196 million with sales of 468 million litres). Perrier's bottling activities in the source water sector are mainly located in France with the wellknown brands Contrex, Volvic, Perrier, Saint-Yorre and Vichy (turnover in 1991 : FF 4,014 million with sales of 1,885 million litres). In France, Nestlé is almost exclusively active in the still mineral water segment while Perrier is present in both the still and sparkling mineral water segments. Both companies export water from their sources in France to other countries. The third major supplier on the French source water market is BSN. Sales of Nestlé, Perrier and BSN in the European Community are set out in Annex 1 (3).

RELEVANT PRODUCT MARKET

A. Distinction between bottled source waters and soft drinks

(8) Nestlé has submitted in its notification that there is no separate market for bottled source water, and that the relevant market to asses the proposed concentration should be that of non-alcoholic refreshment beverages, including both bottled source water and soft drinks. This market definition is argued on the basis that all beverages have as basic function to quench the consumer's thirst.

(9) The Commission considers that a limited substitutability in terms of functionality alone is not sufficient to establish substitutability in competition terms. In the present case, if the only criteria to establish substitutability was to be quenching thirst, many products of very different nature which fulfil that function would have to be considered as belonging to the same market (tea, milk, beer, certain fruits, etc.). Several factors, however, indicate the existence of a distinct market for bottled source waters, where operators are able to act with a significant independece of the actions of companies selling soft drinks, in particular in the area of pricing. These factors are examined below.

(a) Demand considerations

(i) Motivation of final consumers to purchase bottled water originating from a natural source

(10) Bottled source water, in particular in France, is bought and regularly consumed because of its image as a natural product (water originating from a natural source) and its association with purity, cleanliness, absence of contamination and, in general, health and a healthy style of life.

Several consumer surveys confirm this view. An example is provided by the survey carried out by the Institut français de démoscopie in October 1991 for Evian. It shows that the main reasons for which consumers drink still bottled water originating from a natural source relate to its positive action on the human body, in particular because of :

- elimination,

- provision of mineral salts,

and

- positive effect on health.

Other studies, where other types of questions were proposed to the comsumers, tend to confirm this view. The economic advisers of Nestlé themselves (see Lexecon Report) have submitted that: 'The clear evidence of consumer surveys in France is that many people drink bottled water mainly because they trust it to be clean, pure and free of contamination'. Furthermore, the market research study carried out by Sofres at the request of Nestlé's economic advisers has provided the following result: '... more importantly, the survey also asked people why they consumed bottled water in preference to ordinary tap water; and the most popular reason given was its purity, and this was followed by its lack of a bad taste and only on third place the presence of minerals' (Commission's emphasis).

In addition, the publicity campaigns of the three suppliers Nestlé, Perrier and BSN are all directed towards promoting these types of elements in the mind of consumers.

(ii) Differences in composition of the products, their taste and their intended use

(11) Soft drinks are normally manufactured with tap water, and contain additions of flavour and sugar and therefore have a different taste from source waters. Soft drinks are drunk in smaller quantities, not only to quench thirst, but also to satisfy a particular taste pleasure, often in a social context.

(12) Source waters are bought regularly by the final comsumer for daily use in large quantities to fulfil a basic alimentary need, whereas soft drinks are consumed more occasionally. The markets in terms of volumes demanded are very different; the levels of per capita consumption in France in 1990 were the following: bottled water 104,8 litres, carbonates, 29,9, still drinks 9,6, fruit juices 8,6 litres (Source Canadean report, Table 2.1.1). This significant difference in volumes shows that waters and soft drinks are demanded to fulfil largely different needs (low demand-side substitutability); it also shows that soft drinks do not constitute a sufficient substitution alternative for bottled source water.

(iii) Low responsiveness to price changes

(13) Three main factors indicate that it cannot be reasonably expected that an appreciable non-transitory increase in the price of source waters compared with that of soft drinks would lead to a significant shift of demand from source waters to soft drinks for reasons of price only.

1. Substantial price difference in absolute terms between bottled water and soft drinks

A comparison of manufacturer's ex-works price lists in 1991 shows the following price ranges (see Annexes 2 to 7(4)):

- still spring waters (PVC bottles of 1,5 litres): between FF 0,85 and 1,4,

- still mineral waters (national brands in PVC bottles of 1,5 litres): between FF 2,49 and 2,56,

- sparkling mineral waters (national brands): between FF 2,99 and 3,47 (PVC bottles of 1,25 litres), FF 2,98 (PVC bottle of 1 litre) and FF 3,65 (glass bottle of 1 litre, non-returnable),

- soft drinks: FF 6,1 for Coca-Cola (PET bottle of 1,5 litres), between FF 6,1 and 9,4 for tonics (PET bottles of 1,5 litres), between FF 6,1 and 8,76 for still orange flavours (PET bottles of 1,5 litres).

Taking as a reference national still mineral waters, which account for the bulk of consumption in France (over 70 %), soft drinks are, at the manufacturer level, in the range of two to three times more expensive than source waters.

Retail prices (prices for the end comsumer) reflect this same gap. The Canadean Report, Nielsen and Secodip statistics provided by suppliers of source waters and soft drinks and information provided by the main food distributors all confirm this (see Annex 12(5) ).

Nestlé has denied the existence of such gap. Its representatives have submitted that: About 50 % of all soft drinks sold within France are colas. The normal consumer price per litre of Coca-Cola (non-returnable 1,5 litres PVC bottle) is, FF 3,73, whereas the price per litre of Perrier (non-returnable glass bottle) is FF 4,10 and of Vittelloise flavour is even FF 5,23 (non returnable 1,5 litres PET).

Nestlé took these prices from the Canadean Report which indeed quotes a price for Coca-Cola of FF 3,73, applied in March 1991 by the retailer Casino. Nestlé's representatives have not given any explanation of why this price is the normal consumer price of Coca-Cola. The same Canadean Report also quotes much higher prices per litre of Coca-Cola (for instance, prices applied by Euromarché in the same period were FF 4,23 and even FF 4,83 for Coca-Cola Light).

In addition, whereas it seems plausible to select Coca-Cola (among the cheapest soft drinks) because it accounts for a large part of soft drinks consumption, Perrier or Vittelloise flavour are not indicative with regard to bottled source water. The brand Perrier, with 204 million litres sold in 1991 represents some 3,8 % of the bottled source water in France. All flavoured waters added represent less than 1 %. In addition, the packaging and pricing of these two products, which are positioned by the companies on the fringe between source water and soft drinks, is the exception rather than the norm in the bottled water market. Perrier and Vittelloise flavour cannot be considered either in terms of volume or marketing as representative of the source water market.

Moreover, Nestlé's price comparison adjusts for differences in volume in the respective bottles of Coca-Cola and Perrier, but it ignores the differences in the materials of those bottles. Perrier is sold in glass, which is more expensive than the plastic (PET) package of the Coca-Cola chosen to carry out the comparison. However, it is obvious that the price per litre of the same Coca-Cola sold in a glass bottle would be higher. In this respect, a comparison of prices as seen by consumers on the shelves in a retail store might therefore be more appropriate than adjusted retail prices. In this case, the price differences become even larger (see Annex 12).

In view of the above considerations, the Commission concludes that manufacturers' and retail prices of soft drinks, as a rule, are much higher (between 200 and 300 %) than those of bottled waters in France. A price-ratio between two and three is of such a magnitude that it cannot be reasonably expected that an appreciable, non-transitory increase in the price of source waters, would lead to a significant shift of demand from source waters to soft drinks for reasons of price only.

2. The market of bottled source water in France is characterized by strong marketing and promotional activities, with high advertising budgets for the three main suppliers. As a consequence of the resulting image of natural, pure and healthy product associated with source waters (see recital 10) consumers do not recognize soft drinks as a substitute of bottled source water for daily use at home. The general image directly associated by consumers with branded source water as a consequence of the common axes of the publicity of each brand, and their additional attachment to each particular brand reduce the importance of price as purchasing criteria (small elasticity of demand). Therefore, a small increase in the price of source waters is not likely to induce a reaction of consumers resulting in a significant substitution of bottled source water by soft drinks.

3. Price evolution

Manufacturers' prices of source water and soft drinks have had a very different evolution during the last five years (see recital 16). Suppliers of the national mineral waters have been able to substantially increase their prices in both nominal and real terms, in spite of the decreasing trend of soft drink prices during the same period. Manufacturers in both sectors do not seem to take into account in their pricing policies possible substitution by consumers of source waters by soft drinks. This price evolution seems to indicate that even strong and sustained reductions of soft drink prices in real terms would not force source water suppliers to also reduce their own prices, nor would it affect their ability to increase them.

(iv) Views of retailers

(14) The vast majority of retailers consulted by the Commission regard bottled source waters as a distinct market, where demand is basically led by brand awareness and health concerns. Furthermore, they regard mineral waters, in particular national brands, as products they need to offer to their customers in order to cover the full range of basic alimentary products. Because of the regularity of purchases of bottled water for household consumption and the high brand awareness of consumers, retailers regard the national mineral waters as products they need to offer to their customers, with little or no possibility of substitution by other drinks. Retailers are therefore not likely to substitute source waters by soft drinks in response to changes in prices, since they need to sell both kinds of products.

(b) Supply considerations

(i) Conditions of production and marketing

(15) Production and marketing are subject to different constraints in the soft drink market and in the market of bottled source water in particular in France:

1. Regulatory requirements

the production of spring and mineral water is subject to authorisation (for mineral waters this can take between two and five years, see Goldman Sachs report, p. 16). Spring and mineral water has to be bottled at the source. Location of bottling plants is therefore subject to a major legal constraint which manufacturers of soft drinks do not have. Producers of soft drinks can locate their bottling plants according to an adequate balance between economies of scale and transport costs, a rationale water bottlers are not able to follow.

2. Manufacturers of soft drinks may and often do grant trade mark licences to independent bottlers who produce the soft drink in their own premises and sell the product under the original brand. According to industry sources, the soft drink industry has traditionally operated on the basis of national bottling/distribution franchises, although more and more the major companies tend to acquire controlling stakes in independent bottlers. In contrast, water from a given spring may only be bottled at the source. In addition, mineral waters (which represent over 75 % of the total French water market by volume), can only be marketed under the trade mark linked to their repective source, ie. mineral water extracted from a source has, by law, to be marketed under the trade mark registered for that source, and only water originating from that source can be marketed under that trade mark.

3. Different actors in each market

In France, no major soft drink producer (Cadbury Schweppes, Coca-Cola, Pernod Ricard) has been able to develop a presence in the market of bottled water, in spite of the past growth of demand and increasing manufacturers' prices for bottled source water. Similarly, the mineral water manufacturers do not have a major activity in the soft drink market (for instance, Perrier sold to Schweppes its Oasis still soft drinks in 1990, within an explicit strategy of focusing its business on the water market).

(ii) Manufacturers' pricing

(16) Pricing in the soft drink and the bottled source water market seems to follow a different logic, which further confirms underlying differences in the competitive constraints suppliers in each market face.

A comparison of manufacturers' list prices (ex-works, VAT and discounts excluded) for comparable packagings (PVC or PET plastic bottles of 1,5 litres) in the last five years permits certain conclusions to be made:

- prices of soft drinks in real terms, as deflated by the monthly French consumer price index, have shown a declining trend since 1987. The reduction in real prices of soft drinks contrasts with the evolution of real prices for mineral waters. All national mineral water brands have been able to substantially increase their real prices in the same period and gain at the same time sales in volume in the face of an expanding demand (see Annexes 2 to 7);

- the differences in pricing policies are further illustrated by an anlaysis of real price correlation. The coefficient of correlation of real prices among the different brands of waters ranges between a minimum of 0,85 (Badoit and Vittelloise) and 1 (Hépar and Vittel). Real price correlation among soft drinks marketed by different companies is often positive and relatively high (see for instance correlation coefficients between Coca-Cola, Indian Tonic (Schweppes) and Banga (Pernod Ricard)). However, correlation between each soft drink and bottled water is in most cases negative, or when positive, very low. Companies seem to acknowledge therefore in their pricing policies that the bottled water market and the soft drink market are subject to different competitive constraints.

It is particularly illustrative in this respect that BSN's Orange Passion is not significantly correlated with BSN's water brands, although it is the same company deciding on the pricing of both products.

(17) An examination of discounts granted on list prices by manufacturers does not alter the validity of these conclusions for the following reasons:

- list prices constitute, in particular in the source water market and in the soft drink market, an adequate indication of the pricing policies of the companies. The absolute level of discounts negotiated with each particular client (between [...] % (6) and [...] % in the source water market and between [...] % and [...] % in the soft drink market according to supplier and client) and their past evolution are not of such magnitude or variability that they would invalidate the representative value of list prices as indicators of the pricing policy of each company,

- a large portion of discounts cannot be considered as such in the sense that they cannot be passed on to final consumers, but rather as part of the publicity and promotional activities of the supplier or as purchase of a service from the distributor. Suppliers of mineral water include under the term discount budgets of varying amounts, linked for instance to the space allocated in retail stores for the suppliers products, to mailings to end consumers, promotional contests, etc.,

- from the point of view of the distributor, discounts are to a large extent compensated by transport costs, which are borne by the distributor himself. Transport costs are not the same for all distributors throughout France. However, since they have been estimated by Nestlé at roughly [...] % of the ex-works price for a distance of 300 km (page 16 of the notification), list prices constitute a fair approximation of the actual prices paid by distributors.

(iii) Supply side substitutability

(18) The conjunction of the production and marketing constraints described under recital 15 make it impossible for soft drink or beer producers to switch their installed capacity from production of soft drinks to production of source water, be it spring or mineral water. Even if soft drink producers currently had excess capacity, the legal requirement to bottle spring and mineral water at the source makes it impossible for them to use any possible spare capacity already installed to bottle source water.

Nestlé has suggested that bottlers of soft drinks and beer could use their bottling plants to produce purified tap water and enter the source water market with such a product. No evidence of the existence of soft drink or beer producers with excess capacity has been produced to support the submission.

Nestlé has further suggested that purified tap water is marketed in certain countries (United States of America, Denmark, Greece) and is starting to make inroads in Germany, where it accounts for 1,5 % of total consumption of bottled water. According to Nestlé, the fact that purified tap water is sold in certain countries constitutes sufficient evidence that it could be marketed in France, although it is not sold there at present.

Nestlé's submission rests on the implicit assumption that purified tap water would be considered by consumers as a substitute for source water. This assumption is not supported by any evidence. On the contrary, the characteristics of demand in the relevant geographic market indicate that this is not the case. Moreover, it is striking that in spite of:

- strong demand growth and the considerable rate of expansion of market volumes until 1990,

- constand price increases (after discounts) in the French market of bottled source water, in both nominal and real terms,

- the decreasing trend of list prices of the leading brands of soft drinks in real terms (in the case of Coca-Cola, even in nominal terms), which is further aggravated when discounts are taken into account,

and

- the technical and industrial ability of soft-drink or beer producers to bottle purified tap water,

no soft-drink or beer manufacturer has ever entered the source water market in France. Past evidence strongly suggests, therefore, that supply-side substitution has not been considered as commercially viable by the companies concerned.

In addition, the soft drink producers consulted by the Commission have confirmed that they do not currently envisage producing purified tap water for the French market, nor did they suggest it either to be commercially feasible or rational. In this respect, it has to be noted that Coca-Cola has fulfilled its needs for canned water for its vending machines in France with supplies of mineral water from Spa. It has not decided to use its own purified tap water (Bonaqa) sold in Germany, or to use its available installations in France to produce purified tap water.

For these reasons, considerations of supply-side substitutability cannot lead in the present case to a different view of the relevant product market.

(c) Conclusion

(19) It can be concluded that the relevant product market is that of bottled source water

B. Distinction between still source waters and sparkling and flavoured waters

(20) The assessment of the proposed merger would not be materially different whether sparkling and flavoured source waters are excluded form the relevnt product market or not. Several demand factors indicate that sparkling and flavoured waters could be excluded from the relevant product market. There are a number of differences between still source waters and sparkling and flavoured waters, in terms of physical characteristics, taste, intended use, volumes consumed and price levels. In certain exceptional cases, the companies concerned position sparkling and flavoured waters closer to soft drinks in terms of packaging, marketing and price. The sparkling water Perrier (3,8 % of the market) and certain flavoured waters (less than 1 %) constitute examples of such exceptions.

However, it is not possible to exclude competitive interactions between sparkling flavoured and still source waters. Most companies marketing still source waters also market sparkling source waters, and to a lesser extent, flavoured source waters. A bottler of still source water can easily switch production to a sparkling or flavoured source water, at least from a technical point of view.

For these reasons, it might be difficult to justify a radical exclusion of sparkling and flavoured source water from the relevant product market. However, still sparkling and flavoured waters certainly constitute different categories of products, or segments, within the overall market of bottled source water. In the present case the assessment of the proposed merger would in any event not materially differ even if sparkling and flavoured waters were excluded.

RELEVANT GEOGRAPHIC MARKET

(21) In view of the structural elements set out below, and after examination of the competitive environment in the source water markets in the Community, the Commission has concluded that the relevant geographic market within which the power of the new entity has to be assessed is France. Several factors indicate that the parties to the concentration are and will be able to determine their competitive behaviour in France without suffering significant competitive constraints from outside France.

(22) Nestlé's representatives have argued that the level of exports to other areas should be taken into account in the market definition. According to Nestlé, price discrimination between France and the areas where it sells abroad is not possible, at least in relation to Belgium and certain German Laender. Therefore, its pricing in France would be restrained by the conditions prevailing in these other markets where Nestlé is present, although conditions of competition in these other markets are substantially different.

On this basis, and after an analysis of French exports of bottled source water to Belgium and Germany, Nestlé has submitted that the relevant geographic market should be extended to include Belgium and certain German Laender. It has argued that if excessive prices were to be applied in the French market, parallel imports into France would develop. The mere threat of parallel imports would jeopardize any dominant position in the French market according to Nestlé.

However, in view of the different competitive environment prevailing in each Member State, the practical impossibility of the development of parallel imports and the absence of Community competitors capable of overcoming the strong barriers to entry into the French market, the Commission concludes that the relevant geographic market is France.

A. Competitive environment in the Community

(23) The competitive environment in the bottled source water market is very heterogeneous throughout the Community. Demand presents different characteristics in each Member State, trade flows are negligible in terms of consumption and supply is highly fragmented in most Member States.

(24) Demand of a very different nature in the various Member States. In countries such as the UK, Ireland, Netherlands, Denmark and Portugal, demand for bottled source water is extremely low or non-existent (very low consumption per capita) and the market is in a stage of take-off.

In other Member States demand is basically for sparkling waters only, which account in Germany for 75 % of the market, in Italy for over 60 %, in Belgium for 34 %. In these Member States, per capita consumption of bottled water is close to the consumption of carbonated soft drinks, which shows to what extent patterns of consumption are different.

(25) Trade flows in the Community are of minor significance. Water is a low-value/high-volume product which in general cannot bear transport costs over long distances. The high impact of transport consts tends to regionalize the bulk of the market. Nestlé itself has estimated the impact of transport costs at 10 % for 300 km for the most expensive still waters. For glass, transport costs are over twice as much.

Imports into the countries surrounding France remain negligible in terms of consumption: they represent around 5 % in Germany, less than 1 % in Italy and Spain. Basically these are imports of the French mineral waters, who appeal to the higher segment of demand, willing to pay substantial premiums for a reputed brand offering the extra benefit of an international flair. Only in Belgium do imports attain a certain significance, because of the low degree of concentration of the supply and the relatively small size of the market. Cnsumption in Belgium represents around 10 % of French prodction in volume terms.

(26) Supply is highly fragmented in Germany, Italy, Spain and to a lesser extent Belgium, which are the only countries in the Community where the market development would allow the presence of a major player.

In Germany none of the leading companies (Ueberkinger group, Gerolsteiner, Blaue Quellen (Nestlé), Apollinaris) accounts for a market share exceeding 10 % each, even when all their respective sources and brands are added. Only a handful of brands have national status: Apollinaris (less than 6 % of the market) and Fachinger (representing 7 % of the market) and also Gerolsteiner and Hirschquelle might be considered as national brands.

In Italy springs are spread throughout the country and there are over 200 companies marketing bottled source water. The leading brand Ferrarelle (owned by BSN) represents just 10 % of the Italian market. The situation is similar in Spain, where the leading brand Font Vella is also owned by BSN and accounts for 18 % of the market. Apart from Grupo Vichy Catalan (12 % of the market) dozens of other small companies with total annual sales below 100 million litres per year supply the Spanish market (source: Alimarket).

In Belgium the two main producers Spadel and Chaudfontaine enjoy higher market shares basically because of the smaller size of the market. Spadel's total sales of bottled source water in Belgium amount to 256,3 million litres (27 % of the market) and Chaudfontaine's sales to 95 million litres (10 % of the market). None of them has any significant export activity to France, in spite of the advantageous location of their sources.

(27) The fragmentation of these other markets and the absence of major companies comparable to the French suppliers is explained by a number of factors. In Germany, almost all source water is sold in (usually returnable) glass bottles, which increases considerably the impact of transport costs. In Italy, glass accounts for almost half of the total market and the leading brands are normally supplied in glass.

Furthermore, distributors of bottled water in Italy, Germany and Spain are relatively fragmented. Unlike in France, nationwide food retailers are much less important in these Member States, which further hinders the possiblities of suppliers of bottled water in these Member States to develop a large national home base allowing them to envisage expansion in other markets. At present, exports from either Italy, Spain or Germany represent less than 1 % of their respective production. Everything indicates that the struggle of companies in these Member States to overcome their regional limits and achieve at least a national dimension will still continue in the foreseeable future.

B. French exports and the threat of parallel imports

(28) Bottled source water is a relatively inexpensive and bulky merchandise with a high impact of transport costs. It is not likely, even if margins would be high, that a parallel importer could move throughout the various Member States the large volumes required to earn significant revenue and profits. Moreover, it has to be noted that, as a rule, exports of the French mineral waters are carried out through subsidiaries of the main suppliers. For instance, all export activity of Vittel to Belgium and Germany is carried out through Vittel's subsidiaries in those countries (Société Vittel Import and Vittel Mineralwasser GmbH respectively). This arrangement allows for a certain degree of control of the water exported to these two countries.

Prices in Belgium and Germany (prices delivered to customers) are significantly higher than the ex-works prices in France. Nestlé has quoted FF 3,36 per bottle of Vittel (1,5 l PVC) in Belgium and FF 2,82 for the same bottle in Germany for April 1991(7). These prices have to be compared with an ex-works price in France of FF 2,39 at that date. The result of this comparison is that the price at which any hypothetical parallel importer could buy Vittel in Belgium or Germany is currently much higher than list prices charged by Vittel in France (40 and 18 % respectively). In addition to paying these higher prices, the hypothetical reexporter would incur the costs associated with:

- transport back to France,

- unpackaging the pallets, relabelling each bottle and repackaging them again,

- its own margin.

Discounts and rebates granted at the year end on the basis of volumes and sales progression and range discounts also allow the control of parallel imports into France. An independent parallel importer would not be in a position to benefit from these discounts. By their very nature, this sort of discounts enables different commercial conditions to be applied to individual customers, even within France.

In Germany, the packaging of source water is different from that in France. The typical packaging in the German source water market is glass bottles of 75 cl, which are not used in France. 89 % of all sales of bottled water in Germany are packaged in returnable glass, and the bulk of French exports to Germany is packaged in glass. Vittel's exports to Germany consist basically of glass bottles of 1,25 l and 1 l, which together represent [...] of their exports. Transport costs for glass are substantially higher than for plastic (according to Nestlé's estimates as much as 200 %). In addition, French retailers have confirmed that distribution of still source water in France in glass bottles would face several problems (consumers are used to purchase still source water in plastic bottles and distributors would incur higher costs in handling glass instead of plastic).

For these reasons, the Commission considers that the conditions prevailing in the source water market show that the parties' exports and the threat of parallel imports do not and will not constrain the parties' competitive behaviour in the relevant geographic market.

C. Barriers to entry isolating the French market

(29) In addition to the different competitive conditions prevailing in France on the one hand and the rest of the Community on the other hand, there are strong barriers to entry into the French market of bottled source water. These barriers (see recitals 30 to 34) in themselves clearly indicate that France constitutes a separate geographical market, where French suppliers are able to impose appreciable non-transitory price increases without suffering any external competitive constraint.

(30) Absence of imports

In spite of the absence of tariff and legal barriers to trade, imports of bottled source water into France are negligible (between 1 and 2 % of total French consumption) and show no increasing trent. This is easily explained by the market structure prevailing in other Member States. Imports from remote areas or imports of glass bottles are practically excluded for reasons of price. It is important to remember that mineral and spring waters have to be bottled at the source and therefore can be produced only at one specific location. To the extent that there might be foreign sources located near France their transport costs might be absorbed. However, even if that were to be the case, the absence of imports is also explained by several other factors.

(31) Distribution logistics

French mineral water producers generate large enough volumes to transport water by train in complete wagons. Their logistics and that of most of their main customers are adapted to this competitive type of transport. Even companies located near France (such as Spa) would suffer a transport cost disadvantage to export to France, since they would have to either transport by truck or develop the necessary logistics and generate sufficient volume to take advantage of transport by train.

(32) Access to distribution

The French market is a mature market in terms of the number of brands and range of products. According to industry sources, access to the French retail market with a new source water brand additional to the well-established national brands is difficult; only brands with an estabished name can reasonably expect to survive in the medium to long term, rendering entry into the market a high-risk strategy. In addition, water shelf space in retail stores is necessarily limited, and it might in certain cases prove difficult to replace an existing brand.

(33) Advertising (sunk) costs

The French water market is characteried by the predominance of brands. The combined advertising budget of the three national French suppliers - Nestlé, Perrier and BSN - amounted to over FF 680 million in 1991. Heavy advertising by the three established suppliers has existed for several years. The establishment of a new brand would require heavy investment and could take a long time. In addition, the multiplicity of existing brands makes the establishment of a new one mor difficult (access to the media at the appropiate moment) and involves a high level of risk, in particular in view of the national image attached to brands belonging to Nestlé, Perrier and BNS.

(34) Additional barriers to entry

The high degree of concentration in the French source water market, with three companies holding 82 % of the market share by value, constitutes an additional barrier to entry and increases the risks associated with new entry (see recital 98).

The apparent failure of past attempts to enter the French market, together with the absence of imports confirms the actual economic impact of the barriers mentioned above. According to Perrier, there were 15 attempts at entry during the last five years. Only four are still present in the French market: Ferrarelle, which belongs to BSN, San Pellegrino, where Perrier holds a 20 % stake and has three of the nine members of the board of management, and Apollinaris and San Benedetto, which remain insignificant in terms of sales. According to food distributors' sources, while certain brands might enjoy an initial acceptance by the French consumer, consumers in general return to the well-known national brands in the short or medium term. The fact that the Belgian company Spadel directs its export effort to the UK rather than to France, in spite of the geographical situation of its source, further shows to the difficulties of penetrating the French market.

V. COMPATIBILITY WITH THE COMMON MARKET

(35) For the reasons set out above, the assessment of the compatibility of the proposed merger is made on the basis of a separate French market for bottled source waters (hereinafter referred to as 'bottled water') which is composed of still waters and sparkling waters. The focus of the assessment will mainly lie on the segment of still waters because this segment represents approximately 84 % of the total bottled-water market in volume terms and because not material change occurs in the segment of sparkling waters where Nestlé itself has only a small acitivity in flavoured sparkling waters.

(36) The smaller local waters (mineral and spring waters) are included in the market to be assessed, although they differ in a number of respects from the national mineral waters. These differences will be set out below and taken into account in assessing whether local waters (mainly spring waters) provide a sufficient competitive constraint on the nationally distributed mineral waters.

(37) The assessment of the proposed merger will also take account of the agreement concluded between Nestlé and BSN on 30 January 1992 by which these two parties have agreed, subject to Nestlé acquiring control over Perrier, to transfer Volvic, one of the major still mineral source waters of Perrier, to BSN. In its assessment of the proposed merger, the Commission must take into account any existing agreement, the implementation of which would have an appreciable impact on the future market structure. Such an agreement, like any other market element, may lead to the conclusion that the take-over of Perrier under those future market conditions cannot be declared compatible with the common market. The assessment must be made for two situations, one where the Volvic agreement would be implemented and one where the parties would decide not to implement the Volvic agreement.

A. Market shares

(38) In its notification of the proposed merger, Nestlé has provided figures of the total market volume and the market shares of the respective market participants based on its own data and data collected by the Chambre syndicale des eaux minérales and the Chambre syndicale des eaux de source. Nestlé provided figures for the years 1986 to 1991. Nestlé also provided figures of the total market value and the market shares by value.

(39) According to these figures, in 1991 the total market value and the total market volume and the respective market shares of the three national water suppliers, ie. Nestlé, Perrier and BSN, and of the local water suppliers ('others') were as follows:

<emplacement tableau>

(40) The market shares in value terms better reflect the real market strength in this market than the market shares in volume because the French water market is composed of two categories of products which are very different in terms of price, ie. the nationally distributed mineral waters and the local waters, which are mainly spring waters. There exists a wide and constantly increasing price gap between these two categories of waters. On average, the manufacturers' ex-works price of local spring waters is about FF 1 per 1,5 l PVC bottle while the ex-works prices of national mineral waters currently range between FF 2,49 and 2,56 for the same quantities. After the deduction of a flat-rate rebate of 10 % the gap between national waters and local spring waters thus amounts to between FF 1,24 and 1,30. Given this considerable price difference between these two categories of waters and the importance of financial resources in the water market for investment in publicity and marketing, the Commission considers that it is more appropriate to take account of the market shares expressed in value than in volume. The market shares in value provided by Nestlé show that the three national suppliers hold a market share of 82,3 % of the total French bottled water market and that the local suppliers have a market share of only 17,7 %.

(41) In its written and oral submissons following the statement pursuant to Article 18 of the Merger Regulation, Nestlé changed the volume figures it had originally presented in the notification which were based on data from the two relevant trade associations in favour of figures taken from the Canadean Report 1990. According to that report the total market volume would not be 5,250 million but 5,900 million litres, the difference resulting mainly from the view that local suppliers would have a larger share of the market, representing not 25,8 but 35,5 % for still waters. BSN which was also heard by the Commission presented estimates of two panels (Intercor and Secodip) which would suggest a higher market volume than that presented by Nestlé in its notification.

(42) Nestlé has not changed its figures on the total market value, ie. FF 9,334 million. BSN has suggested a total market value of FF 9,400 million. These figures are remarkably close and thus confirm the value figures provided by Nestlé on notification. Therefore, the discussion about the market volume is less important since in any event the market shares by value would remain at the level of 82 % for the three national suppliers and 17 to 18 % for the local water suppliers.

(43) As regards the market volume, the figures suggested by the Canadean Report 1990 and the Intercor/Secodip panels are only estimations which cannot be given more credit than the figures originally presented by Nestlé which are mainly based on data received from the relevant trade associations. The Canadean Report and the Intercor/Secodip panels include in the category of others local mineral and spring waters belonging to Nestlé and Perrier. In 1991, these waters represented a volume of 221 million litres. In addition, unlike the figures of the trade associations, these reports and panels are only estimations based on samples of part of the sales of retailers or purchases of consumers. Finally, they contradict themselves because they present three entirely different estimations of the total market volume, ie. Intercor 5 390 million, Canadean 5 900 million and Secodip 6 690 million litres.

(44) In 1991 the strong demand growth of 1989 and 1990 (mainly due to exceptionally hot weather conditions) was interrupted and slowed down to below 1 %. Nestlé's figures provided on notification take account of this interruption in demand growth and provide a total market volume for 1991 of 5 250 million litres. In Annex 2, the Commission sets out the sales volumes obtained through its enquiries and from the data of the relevant trade associations. This calculation leads to a total market volume of 5 249 million litres (still waters 4 419 million litres - sparkling waters 830 million litres). These figures are remarkably close to the figures provided by Nestlé in its notification.

The figures provided by the Chambre syndicale des eaux minérales and the Chambre syndicale des eaux de source show that in 1991 there was a decline in sales of local mineral waters compared to 1990 and only an increase of 3,3 % in sales for local spring waters. This confirms the general trend of the market in 1991 which only experienced a small increase in total volume compared to 1990.

(45) It can be concluded from the above that both the value and volume figures provided by Nestlé in its notification are reliable figures which can be used for the assessment of the proposed merger. These figures are also to a large extent confirmed by the share of local waters in the total water sales (still and sparkling) of a number of major retailers and wholesalers. A weighted average of those sales shows that the local water suppliers represent in volume only approximately 22 % of total water sales. By value, this share falls to approximately 16 to 17 %.

(46) From 1986 to 1991, the market shares for the total French water market by value and volume evolved according to Nestléy's own figures as follows:

<emplacement tableau>

The above table shows that the market shares of the three national suppliers remained relatively stable over the period considered. The average variation in market shares for all three suppliers amounted to 2,1 % by value and 2,4 % by volume. This compares with a variation of the market shares of all others of 2,3 % by value and 4,7 % by volume. In absolute terms, the three national suppliers have constantly increased their sales in terms of value (average increase of 9,4 %) and in terms of volume (average increase of 4,6 %). It is only in 1991 (strong reduction in demand growth) that the three national suppliers lost a volume of 49 million litres (see recital 68) while still increasing their sales in value by 4,9 %. In the same year, the local mineral water suppliers also lost volume. Only the local spring water suppliers could increase their volume by 3,3 %.

Given the fact that demand growth is largely due to environmental reasons leading people to replace tap water by the cheapest bottled water which are spring waters (tap-water replacement segment), the above variations in market shares are relatively small. They do not indicate a fundamental change in the market positions of the three national suppliers amongst themselves or with regard to local water suppliers.

(47) Based on Nestlé's figures provided on notification, the post merger market shares in the still water segment with and without the sale of Volvic to BSN would be as follows:

<emplacement tableau>

(48) It can be concluded from the above that the French bottled water market is already a highly concentrated market before the merger and that the merger would further increase this concentratin because only two suppliers would hold 82 % of the market by value and nearly 75 % by volume with over 90 % of all still mineral waters.

B. Capacities and portfolio of sources

(49) Nestlé has submitted that nearly all major sources of Nestlé, Perrier and BSN are working close to their maximum capacity and consequently cannot maintain their market share in a strongly growing market.

(50) Nestlé itself, however, recognizes that the growth rate of the French bottled water market is likely to slow down and indeed did not reach 1 % in 1991. Nestlé speaks of a realistic growth rate of 5 % over the next years.

(51) In the still water segment at the centre of discussion in this case, Nestlé has the following free capacities in mineral water (ie. output capacity of the source per year):

1. Vittel

- annual capacity: [...] million litres,

- free capacity: [...] million litres,

2. Hépar:

- annual capacity: [...] million litres,

- free capacity: [...] million litres,

3. Abatilles:

- annual capacity: [...] million litres,

- free capacity: [...] million litres.

Nestlé also owns the spring water Pierval which has a capacity of [...] million litres and a free capacity of [...] million litres.

(52) Perrier owns two major still mineral water sources with considerable free capacities, i.e. Contrex and Volvic. The exact figures are known to the Commission but cannot be disclosed for reasons of trade secrecy. The free capacity of both sources combined exceeds the total still-water market volume. Perrier owns in addition other still mineral water sources, in particular Thonon which has an estimated capacity of [...] million litres and a free capacity of approximatively [...] million litres (Nestlé's estimation).

Perrier further owns a number of important sparkling mineral waters: Perrier, Saint-Yorre, Vichy and several smaller sources. The first two sources have a considerable overall and free capacity.

Finally, Perrier owns a large number of local spring water sources, in particular the sources Castel, Saint-Roch, Les Ormes, Carola, Saint-Lambert, Montegut, Regina and Sergentale.

(53) BSN has one major still mineral water source, ie. Evian, which has an estimated free capacity of approximately [...] million litres and one major sparkling mineral water source, i.e. Badoit which has reached its maximum capacity. BSN is presently developing a new sparkling mineral water source called Salvetat.

(54) After the merger and the sale of Volvic to BSN, the two suppliers would have a considerable numer of sources, the overall free capacity of which would by far exceed the total water market volume (5 250 million litres) and each one of these two suppliers would have at least one major still mineral water source with huge free capacities compared to the overall market volume and all other local water suppliers. They would thus be in a position to respond to an increase in demand without any capacity limitation, even if one were to assume a constant growth rate of demand of 5 % per year. Although it is not possible to mix mineral waters from different sources, Nestlé and BNS would considerably increase their respective portfolios of major and reputable sources and brands with the result that they can better adapt their marketing strategies to demand increases.

(55) None of the local water suppliers has water sources in number or in size similar to those of Nestlé and BSN after the merger. The capacities of most of the local sources are below 200 million litres and generally much lower than that. Even if different spring waters can be sold under one brand (e.g. a retail brand) they must still be bottled at the source with the result that, given their low price, they cannot be transported over longer distances 200 to 250 km). This means that if a big retail chain intends to develop a retail brand with several spring waters, it can only do so with waters coming from different sources and thus having different tastes and characteristics. These types of waters are at the lowest end of local spring waters and thus do not strongly compete with the national mineral waters which are known for their mineral characteristics, their stability in composition and their identical geographic origin. This fact is confirmed by most retailers (see Annex 15 (8)).

(56) It can be concluded that after the merger and the sale of Volvic to BSN, both Nestlé and BSN would have very considerable capacity reserves and thus be able to respond to an increase in demand. The acquisition of Perrier and its subsequent division would considerably increase the overall and free capacities of both Nestlé and BSN.

C. Prices

(57) Even without the merger a narrow oligopoly of three suppliers exists between whom price competition is considerably weakened and for whom the degree of market transparency is very high.

(58) Price gap between local spring waters and national mineral waters

The ex-works prices (before rebates and VAT) of Nestlé, Perrier and BSN for their five major still mineral waters range currently between FF 2,49 and 2,56 per 1,5 l PVC bottle: Contrex FF 2,56, Hépar FF 2,55, Evian FF 2,51, Vittel FF 2,49 and Volvic FF 2,49. After the deduction of a flat-rate rebate of [...], these prices exceed the average ex-works prices of local spring waters by FF 1,24 to 1,30 per 1,5 l bottle (see recital 40).

(59) Price increases and price parallelism

The ex-works prices (before rebates and VAT) of the five major still mineral waters of the three national suppliers have constantly increased in a parallel way since at least 1987 (see Annexes 5 and 6). The average annual increase of prices between 1987 and 1992 was: Contrex 5,10 % (nominal) - 1,69 % (real); Volvic 4,83 % (nominal) - 2,10 % (real); Vittel 5,17 % (nominal) - 1,41 % (real); Hépar 5,70 % (nominal) - 1,60 % (real); Evian 5,82 % (nominal) - 2,29 % (real). Whoever first increased its prices was always followed by the other two suppliers. There was no price decrease during the whole period considered. The price leader seems always to have been Perrier which has traditionally maintained the highest price level for most of its products. The rebates granted by the three suppliers were according to the Commission's information, in the range of [...] to [...] %.

(60) Elasticity of demand

With regard to local waters (mainly spring waters), the cross-price elasticity of demand of the national still mineral waters is relatively low. The three national suppliers have created an extremely high consumer fidelity through long standing and strong publicity campaigns for national mineral waters. This is confirmed by the fact that retailers, who only reflect the demand of consumers, do not consider price as the first criterion for the choice of national mineral waters (see Annex 14(9) ) and by the fact that they consider themselves dependent on the major national mineral water brands (see Annex 15). This low cross-price elasticity of demand with regard to local waters is clearly borne out by the fact that since at least 1986 the three national suppliers have maintained and constantly increased their absolute sales volumes (which represent over 70 % of the total water market) with only a small loss in 1991 when total demand growth slowed down considerably. There thus exists a low cross-price elasticity of demand for mineral waters which facilitates price increases or at least the maintenance of high prices without losing significant sales volumes to local water suppliers.

(61) Production cost-price margin

The production cost-price margin of the three national suppliers for a 1,5 l PVC national still mineral water bottle is very high: it varies between approximately FF [...] and [...]. The cost basis used includes variable and fixed production costs and advertising costs. From this very high production cost-price margin and from the important price difference with local spring waters which are priced at a competitive level it can be concluded that prices for national mineral waters are probably already at a very high, supra-competitive price level.

(62) Market transparency

The three national suppliers publish their list prices with the basic quantity rebates. Since they all supply the same customers, there is also a considerable feedback from these customers. In addition, the three suppliers provide the Chambre syndicale des eaux minérales with their monthly sales volumes and each one receives the monthly sales quantities broken down by brand of the other suppliers. In a narrow oligopoly such a practice further increases the market transparency and permits each supplier to follow the evolution of the market positions of the others.

D. Cost structures

(63) Nestlé recognizes that the major brands of the three national suppliers have similar cost structures (see Lexecon Report, p. 45). The manufacturing process consists basically of bottling. The need to bottle at source and the resulting multiplicity of bottling plants reduces the scope for economies of scale. As a result neither Nestlé nor BSN has a major cost advantage which could give either one of them an incentive for aggressive competitive action vis-à-vis the other. It must also be noted that the three national suppliers all apply a system of ex-works prices, transport costs being borne directly by the customers (retailers/wholesalers).

E. Competition from local water suppliers

(64) Nestlé, supported by BSN, has submitted that local waters (mainly spring waters) constitute a competitive constraint for the three national suppliers. In its written submissions to the Statement pursuant to Article 18, Nestlé claims that local suppliers represent 34 % (still waters: 35,5 %) of the French water market and that in certain hypermarkets their share would have reached 50 %.

(65) According to Nestlé's own figures, the market share of others by value amounted in 1991 to only 17,7 %. Given the considerable price gap between the average ex-works prices of local spring waters and national mineral waters, it is the market share by value which best reflects the market strength of local waters in the total water market (see recital 40). It is not disputed by Nestlé that in value terms, the sales of others are generally below 20 % of the total market.

(66) As regards the sales of others by volume, their market share in 1991 amounted according to Nestlé's figures to 24 % for the total water market and 25,8 % for the still water segment. These market shares are confirmed by the share of sales held by local waters in the total water sales of a representative sample of retailers and wholesalers for which the weighted average sales of local waters represents approximately 22 % of total water sales.

(67) The evolution of the market share of others in the total water market was as follows for the last five years; by value: 16 % in 1986, 17,7 % in 1991 (increase 1,7 %); by volume: 20,3 % in 1986, 24 % in 1991 (increase 3,7 %) (see recital 46).

The evolution in absolute figures of sales values and sales volumes (total waters and still waters) of the three national suppliers and the local suppliers (others) is set out in the following table:

<emplacement tableau>

(68) The above table shows that following two years of very strong growth mainly due to exceptionally hot weather conditions, market growth slowed down considerably from 8,5 % in 1990 to 0,9 % in 1991. In spite of this strong decline in demand growth, the market value increased in 1991 by FF 534 million, of which the three national suppliers took FF 364 million (68 %) and the others FF 170 million (32 %). During both the last five and three years, the average increase in value of the three national suppliers always exceeded by more than three times that of the local suppliers. They have constantly increased their turnover, even in 1991 when they lost 49 million litres in volume.

(69) The above table further shows that in volume terms the sales growth of not only the three national suppliers but also of the local suppliers declined in the years 1990 and 1991. In spite of a stronger decline of the three national suppliers in 1991, the average increase in sales of the three national suppliers during the last five and the last three years always exceeded that of the local suppliers.

(70) The volume losses of the three national suppliers in 1991 were small: 47 million litres for still waters and two million litres for sparkling waters. This represents only 1 % in the still water segment where the three suppliers together hold nearly 75 % of the market volume. Given the strong reduction in demand growth in 1991 following two exceptional years of growth, this loss is not the sign of a significant shift in demand in favour of the local spring waters. Also, the loss of sales of the national suppliers in 1991 is not necessarily or only due to competitive action of local spring waters because with a price increase of 7,11 % in 1991, Evian still gained a volume of 20 million litres. Volvic and Hépar, which also increased their prices, equally gained in volume, respectively 35 and 16 million litres. Only Contrex and Vittel lost in volume. Furthermore, it is not denied that a certain degree of competition exists between the national mineral waters and local spring waters. However, the above changes are not sufficient to show that local spring waters constitute a real competitive threat to the market positons of the three national suppliers. Even lively competition is not incompatible with market power as long as such competition does not compel the undertakings holding the market power to lower their prices (see the Court's judgment in Hoffmann-La-Roche, paragraphs 70 and 71, 1979 [ECR] 461).

(71) It must always be kept in mind that the prices of national mineral waters are already at a very high level. In spite of this situation, the three national suppliers have succeeded in regularly increasing their prices and further increased their prices in 1992 (Contrex 3,64 %, Volvic 4,18 %, Vittel 4,18 %, Hépar 5,80 % and Evian 4,14 %). Thus, the three suppliers themselves consider that it is still profitable to increase their prices and that the pressure from local waters is not such as to prevent them either from maintaining the present high prices or even from increasing them. This ability constantly to increase prices, even in real terms, is a strong proof that competition from local water suppliers has not been sufficient in the past and is not likely after the merger to significantly contrain the market power of the three national suppliers.

(72) There exist a number of structural disadvantages for local spring water suppliers which limit their competitive pressure on the national suppliers.

The suppliers of spring waters are small companies which are dispersed numerically and geographically. None of the local suppliers achieves 10 % of the total market or of the still water market segment. In 1991, the family Papillaud and the group CGES each reached a sales volume representing a market share in the region of 8 to 9 % by volume and 3 to 4 % by value. Most of the other local water suppliers have market shares below 5 % by volume and much smaller by value.

(73) The local suppliers are all independent companies which may not be aggregated to determine whether they would be able to constrain the market power of the two remaining national suppliers. Their individual financial strength compared with that of Nestlé and BSN is extremely weak. Therefore, the combined market share of others of 17,7 % value and 25,8 % by volume in fact overstates the competitive strength of local water suppliers. Even if a few local waters are owned or supported by big retailers, this activity does not generate for these retailers a sufficient profit for them to invest important resources for the development of such sources. By contrast, Nestlé and BSN have a considerable financial strength which they are able and likely to use to develop their water activity which represents an important and very profitable business for them.

(74) Even if retail brands can combine several local spring waters under one umbrella, these waters are not a sufficient alternative and counterweight to the national mineral waters according to the retailers themselves (see Annex 15). This strategy of combining several spring waters under one brand can only to a limite extent improve the position of the local waters but cannot change the fact that there is a long-standing and well-established demand for the national mineral waters (over 70 % of total demand). In addition, spring waters from different sources sold under retail brand are generally at the lowest end of the local still waters. These waters are not really sold to compete with, but rather to complement the major national brands. They are the less likely to be accepted as a reasonable substitute by those consumers who regularly drink national mineral waters. At best, the retail brands will be successful in that segment of the market where consumers are merely looking for a replacement of tap water by the cheapest drink or household water (the tap-water replacement segment).

According to Nestlé's declarations at the oral hearing, retail brands represent at most 7 to 8 % by volume (much less by value) of the total still water segment. Since the retailers have these waters bottled by local water suppliers, their volume is already mostly included in the market share of others. It is only exceptional that a retailer owns a source and bottles itself the water.

Local mineral waters cannot by law be marketed under a retail brand or under several different brand names. According to Nestlé's figures, local still mineral waters represent only 6,5 % of all still mineral waters against 93,5 % in the hands of the three national suppliers.

(75) Given their low sales prices, the majority of local waters can only be marketed economically within a radius of 200 to 250 km. There is not a sufficient margin for the producer to absorb transport costs over longer distances. In addition, the majority of local water sources have capacities of below (mostly far below) 200 million litres.

As a consequence, local water producers do not have sufficient volume and financial return to justify significant important invenstment in brand promotion, in national publicity or in a national distribution organization. Their waters have no strong brand image and are only sold regionally.

In contrast, the three national suppliers own sources with considerable capacities. They have invested for years in the image of their brands and in national distribution systems. They have crated a high fidelity of consumers who expect to find these national brands in all retail stores. The three national suppliers spent over FF 400 million in 1991 in advertising for the five major brands, ie. Contrex, Volvic, Vittel, Hépar and Evian, and in total over FF 680 million. After the merger, the financial resources of Perrier will be further increased by the combination with Nestlé's massive financial resources. No local spring water supplier could possibly try to compete on equal terms with the ramaining two suppliers as far as investments in advertising or nation-wide distribution systems are concerned. In its letter of 12 March 1992, Nestlé recognised that 'for smaller sources it is difficult to become a significant player in the French water market'.

(76) It must be concluded from the above that local water suppliers would not be able, at least in the short term, to significantly constrain the market power of the two remaining national water suppliers. Nestlé and BSN have clear competitive advantages compared with the local suppliers (much bigger sources, much better known brands, much better distribution and stocking systems) and have important financial resources to continue to invest heavily in national publicity and promotion campaigns. This strategy is liable to increase consumer fidelety to their well-known brands and thereby to reduce the cross-price elasticity of demand of their waters in particular vis-à-vis local spring waters. In addition, after the merger Nestlé would itself own a considerable number of local spring water sources (see recitals 51 and 52). It would thus have a full range of sources comprising mineral and spring waters which it could use in its sales strategy on the market.

F. Buying Power

(77) Nestlé, supported by BSN, has submitted that the three national suppliers are constrained by the buying power of the big retailers. Nestlé's 10 largest customers represent 67,3 % of its total turnover. Nestlé and BSN have submitted that four major retail groups, ie. Intermarché, Leclerc, Carrefour and Promodes, would represent 50 % of the total sales volume of the three national suppliers. In addition, some of the retailers sell local waters under own retail brands so that they are becoming competitors of the national water suppliers. The buying power would also be demonstrated by the fact that major brands have been delisted and that the average rebate granted by Vittel to its retailers has increased by 50 % during the years 1988 to 1992 (see Lexecon Report, p. 33).

(78) As regards the concentration of buyers, it is true that the 10 largest buyers represent an important part, around 70 %, of the turnover of all three national suppliers and, thus, have a certain purchasing power. However, concentration of these buyers is much less important than the concentration of the supply side, in particular after the merger. While 82 % by value and 75 % by volume (for still mineral waters over 90 %) of the supply side would be in the hands of only two suppliers which can easily engage in anticompetitive parallel behaviour (see recital 108 and following), the demand side is composed of a much greater number of independent companies, none of which exceeds 15 % of the total water turnover of all three/two suppliers. For Nestlé, the individual share of the ten largest buyers varies between 1,9 % and 11,6 % of Nestlé's total turnover. Out of ten buyers only four reach 10 to 11 %, the remaining six vary between 1,9 and 5,5 %. These purchasers do not all have an equal buying power based on the volume purchased. This is confirmed by the fact that the rebates granted to these retailers are not all identical but adapted to each individual buyer. All this shows that in front of a highly concentrated supply side, there is a demand side which is composed of a number of buyers which are not equally strong and which cannot be aggregated to conclude that they may constrain the market power of the three, and after the merger, only two national suppliers.

In the enforcement of the competition rules, the Commission must also pay attention to the protection of the weaker buyers. Even if some buyers might have a certain buying power, in the absence of sufficient competition pressure on the market, it cannot be excluded that Nestlé and BSN apply different conditions of sale to the various buyers.

(79) The buying power of retailers and wholesalers is also significantly limited by the fact that they cannot afford not to sell for a longer period of time the well-known brands of mineral waters supplied by Nestlé, Perrier and BSN because of the risk of losing customers. Over 70 % of total French water consumption and 90 % of all mineral water consumption would be concentraded on the two remaining national suppliers.

(80) In other cases decided by the Commission where buying power was one of the elements creating a counterweight, the products involved were generally intermediary products or products where long-term contracts or cooperation agreements for development of the products were involved which can create a more balanced seller-buyer relationship, provided the buyers are sufficiently concentrated. This is not the case for bottled waters which are widely used consumer products where the retailers only reflect the demand of consumer.

(81) Contrary to the statements made in the Lexecon Report (p. 48) on brand loyalty, it cannot be denied that the national supliers have created a high consumer fidelity for their respective brands through long-standing and heavy publicity campains which make their products to a large extent 'produits prévendus'. This is generally recognized by the professionals in the industry and by the retailers themselves who only reflect the demand of consumers.

(82) Thus, the Chambre syndicale des eaux minérales writes in its letter of 18 February 1992:

'...on trouve d'une part les grandes sources d'eaux minérales qui sont distribuées a l'échelon national et pratiquent une politique de notoriété de marque appuyée sur une forte publicité et autorisant des prix relativement élevés; d'autre part, des sources régionales, d'eaux de source ou d'eaux minérales, dont le dynamisme de commerialisation repose essentiellement sur des prix très bas.

Dans le premier groupe, qui fournit plus de 70 % des ventes intérieures, le marché est caractérisé par une forte identité de chaque source indépendamment de la Société ou du Groupe qui controle son exploitation et un attachement du consommateur habituel a ses marques préférées que l'on ne rencontre pas au même degré dans les autres pays...'

The Commission questioned a representative sample of retailers and wholesalers about the order of priority of their purchase criteria. A majority mentioned as first criteria the demand of consumers and the reputation of the brand (see Annex 14). Brand reputation and brand loyalty is also mentioned as a important feature of national mineral waters in a number of other replies (see Annex 15).

(83) After the merger and the sale of Volvic to BSN, Nestlé and BSN would own a bigger portfolio of well-known brands than each one of these two suppliers presently owns. They would eliminate competition from all major still and sparkling water brands of Perrier, in particular Perrier, Saint-Yorre, Contrex, Volvic, and thereby reduce the choice of retailers between several currently independent brands. This would inevitably increase the dependency of retailers and wohlesalers on the two remaining supliers. Not only would their choice be reduced from three to two suppliers but the tying of these brands operated by each supplier (rebates are made dependent on the sale of the whole range of products) would cover a wider range of reputable brands and thus further increase the dependency of retailers on the two remaining suppliers.

(84) A majority of the retailers and wholsesalers referred to in recital 82 have stated that they would become even more dependent on the national suppliers than up till now and that local spring waters are not a sufficient alternative for the national mineral waters. These statements are set out in full in Annex 15. The retailers and wholsesalers concerned (both major and smaller companies) express serious fears about their future margin of negotiation vis-a-vis Nestlé and BSN and consider that they have to sell the brands of these two suppliers. They also consider in general that local waters are more a complementary product than a real substitute for drinkers of national mineral waters.

(85) The weigthted average of the share of the sales of Nestlé, Perrier and BSN in the total water sales (still and sparkling 1991) of a representative sample of retailers and wholesalers amounts to approximately 78 % by volume and even higher by value. This shows the considerable weight of the brands of the two remaining suppliers in the total water sales of a number of major retailers and wholesalers.

(86) As far as delisting of brands is concerned, Nestlé admits that it is unlikely that retailers would delist all the brands of a national supplier at the same time. This confirms the overall dependency of retailers on the national suppliers.

It is not denied that even only partial delisting can have a certain disciplinary effect on the supplier and that to that extent at least the biggest retailers have a certain buying power. However, this buying power is not the same for all retailers and is limited the larger the portfolio of brands in the hands of one supplier because each delisting immediately affects the annual full range rebates on all products (still, sparkling, flavoured, etc.). Therefore, the merger and the sale of Volvic to BSN would increase the portfolio of the two remaining suppliers and would make delisting in future more difficult than it is at the present time. In addition, delisting always provokes a certain switch of customers to other shops which sell their preferred brand. The exact percentage of such switch in each case is unpredictable and thus constitutes a factor of risk for the retailer (see Annex 15, retailer L) which limits its bargaining power.

(87) Nestlé has also stated that the fact that bottled water represents only an estimated 1 % of the sale of a retailer, whereas those same sales represent a much higher percentage of the turnover of the supplier, creates an imbalance in favour of the retailers. This does not reflect accurately the bargaining position of the retailers. According to the information provided by certain retailers, it is not excluded that a delisting of a national brand can result in a loss of one out of 10 end-consumers, who, because of their attachment to the brand, would switch retail outlet in order to find the particular brand. Therefore, a delisting of one single major brand can have a much higher impact on the retailer's sales than the 1 % figure would suggest, since the loss of each client implies a loss of all the client's purchases in the retail outlet concerned.

(88) Nestlé has finally submitted that buying power is demonstrated by the significant increase of rebates between 1988 and 1992. This argument cannot be accepted. The significant increases in list prices in both nominal and real terms have more than compensated for the increase in discounts. The effect is a net increase of the prices paid by retailers. Taking Vittel itself, the largest discount it granted to a retailer in 1987 was [...]% (calculated on the basis of all discounts in French francs granted during the year, including commercial cooperation in relation to the turnover in French francs achieved with that client).

In 1992, the largest discount granted by Vittel to a retailer represented [...] % of the turnover achieved with that customer. Certainly, the discount has nearly doubled in the last five years, as Nestlé has submitted. However, the list price of Vittel (1,5 l PVC) has increased from FF 1,92 (most favourable list price) in January 1987 to FF 2,49 in April 1992. Assuming that the total amount of discounts could be considered as such (including commercial cooperation), the price after discount has increased by [...] %. All other major brands have been able to increase their net prices even more (see Annex 13(10)).

Therefore, even in a situation where three main suppliers are present on the market, retailers have not been able to counterbalance the ability of Nestlé, Perrier and BSN to increase their prices considerably during the last five years. After the merger, the buying power of the retailers is likey to be even more reduced.

(89) It must be concluded from the above that the buying power of retailers and wholesalers would not be sufficient to constrain significantly the market power of the two remaining national water suppliers after the merger. The merger would increase the portfolio of major well-known brands in the hands of Nestlé and BSN which must be on the shelves of each big retail store. The merger would also reduce the choice of the retailers from three to two sources of supply and would thus increase their dependency on the two major suppliers on the market. Given their increased dependency on the major brands, it would be more difficult for retailers to play one producer against the other. Also, the local spring waters do not represent a realistic alternative, at least not in the short term.

G. Potential Competition

(90) Nestlé has submitted that in view of the strong market growth for bottled water almost all major beverage and food manufacturers can be considered as potential competitors which would constrain the market power of Nestlé and BSN after the merger. Nestlé also refers to the large number of available sources in France and to a number of new market entries both by French local suppliers and by foreign suppliers. Nestlé mentions in particular, imports into France from Spa, Apollinaris and some Italian producers.

(91) To address the question of potential competition it needs to be examined whether there exists competitively meaningful and effective entry that could and would be likely to take place so that such entry would be capable of constraining the market power of the two remaining national suppliers. The question is not whether new local water suppliers or foreign firms can merely enter by producing and selling bottled water but whether they are likely to enter and whether they would enter on a volume and price basis which would quickly and effectively constrain a price increase or prevent the maintenance of a supracompetitive price. The entry would have to occur within a time period short enough to deter the company(ies) concerned from exploiting their market power.

(92) The Commission is of the opinion that there exist significant barriers and risks to enter the French bottled water market. These risks exist equally for new local water suppliers and new entrants from other neighbouring geographic markets. The reasons for difficult entry are set out hereinafter.

(93) Although the French water market has been growing very rapidly until 1990, this growth was halted in 1991 when the annual growth rate slowed down to below 1 %. France already has one of the highest per capita consumptions of water in the Community and it is therefore expected that the French water market will in future grow more slowly than the rest of the Community. This is admitted by Nestlé (see Lexecon Report, p. 47 and Mr Sedemund's letter of 12 March 1992, p. 16). The strong decline in demand growth in 1991 is probably not permanent but indicates that there is a certain limitation on demand growth in France.

(94) The French water market is a mature market in terms of number of brands and range of products. The market is already filled with a number of well-known and well-established brands which fully satisfy the retail stores and fill their limited shelf/stocking place. In the still water segment there are five major brands (Evian, Contrex, Vittel, Volvic, Hépar) established in virtually all distribution outlets. Entry with a sixth brand on a national or large scale basis would be very difficult. In addition, new brands with little or no reputation would not only have to be sold at much lower prices (e.g. the local waters) but would also have to offer retailers much larger discounts to induce a retailer to list a new brand which provides much less volume than any of the established well-known major brands.

(95) Access to French distribution is further rendered difficult by the fact that the established national suppliers grant to retailers and wholesalers annual rebates linking their respective products (rabais de quantité, rabais de gamme, rabais de progression, accords de coopération, etc.). These rebates and cooperation agreements go beyond pure quantity rebates because they pursue the objective of binding the buyers to the full range of products of each particular supplier. This type of rebate strengthens the position of the established suppliers and raises the barriers to entry for newcomers. Any newcomer would have to offer comparatively much greater rebates given his lower volume and lower range of products.

(96) One of the key problems of access to the French water market is the reputation of the established brands of the three national suppliers which grants Nestlé, Perrier and BSN a long-run advantage over any local supplier or newcomer.

The three national suppliers have invested for many years and continue to invest considerable sums in publicity and promotion for their respective brands (see recital 75). In 1991, Nestlé spent FF [...] million for the promotion of its brand Vittel. This represented around 10 % of its turnover achieved with the sale of that brand. In the same year, BSN spent also around 10 % of its turnover on the promotion of its brand Evian. The same was true for Perrier and its Volvic brand. High and long-standing spending in advertising has created a low elasticity of demand and thus a significant barrier to entry vis-à-vis local spring waters and vis-à-vis potential newcomers by constantly increasing consumer fidelity to the established brands. After the merger, Nestlé and BSN are capable and likely to continue heavy advertising of their respective brands because they have the required financial resources to do so and because this strategy contributes to deterring potential entrants and maintaining the price gap and reduced elasticity of demand towards the local water suppliers.

(97) Advertising and promotion are sunk costs which are not recoverable in the case of failure in the market. For newcomers the establishment of a new brand is not only very costly and time consuming (consumer acceptance can take several years), but is also extremely risky because in case of failure (non-acceptance of the brand by the consumer) all the investment is lost. As was explained in recital 96, the establishment of a new brand in the mature French market would be extremely risky. In the bottled water market, advertising is required to sell significant amounts of the product. The advertising costs are costs which constitute a considerable disadvantage for newcomers compared with the established firms because newcomers cannot recuperate these costs on high current sales volumes.

(98) High concentration in itself is a barrier to entry because it increases the likelihood and the efficiency of single or concerted reaction by the established firms against newcomers with a view to defending the acquired market positions and profitability. The joint action by Nestlé and BSN to oppose the public bid made by Ifint (the Agnelli Group) on Exor (Perrier) was a clear sign of a fierce reaction against the acquisition of Perrier by a firm which could have disturbed the existing market structure.

(99) All these difficulties were confirmed to the Commission by virtually all interested companies questioned about the ease or difficulty of access to the French water market (see extracts made available to Nestlé by Commission letter of 14 May 1992). One very illustrating reply on this question is set out in Annex 16(11).

(100) Contrary to the assertion made by Nestlé, all the companies contacted by the Commission confirmed that they had no concrete plans to enter the French water market. The answers received show that if there are potential entrants they consider entry too risky, particularly in case the merger Nestlé/Perrier were to take place. In any case, entry ex novo would require the building of a plant, the application for an authorization to exploit a source, the establishment of distribution links and the establishment of a brand, all of which can take a considerable period of time for a newcomer.

(101) Nestlé and BSN have drawn the Commission's attention to the fact that there exist a large number of mineral and spring sources in France which can be developed and have pointed towards a number of entries of local suppliers (see in particular Lexecon Report, p. 58). However, none of these sources are of a size in capacity or sales comparable to the major sources of Nestlé, Perrier and BSN. There is no indication that any of these sources would be able to develop quickly to a point that they could significantly restrain the market power of the two remaining national suppliers. In its letter of 12 March 1992, p. 14, Nestlé admitted that for smaller producers the investment in marketing and advertising makes it difficult to become a significant player in the French water market. It is not denied however that local suppliers can develop and have, but this development is limited to regional areas and to a particular segment of the market which does not and is not likely in the short term to significantly restrain the market power of the national mineral waters. This is best proved by the fact that the three national suppliers were able to constantly increase their prices both in nominal and in real terms despite the increase of sales of local spring waters.

(102) The agreement concluded between Nestlé and BSN by which Nestlé has agreed to sell the Volvic source to BSN for a price exceeding FF [...] million could not be explained if it was in fact very easy to enter the French water market by developing new springs and new brands. This agreement shows that a significant entry or expansion in that market can be made only through the acquisition of a major source. With the acquisition of Perrier owning the biggest capacities on the market and the agreement to sell Volvic to BSN, all possibilities for acquisition of major water sources and brands in France by a newcomer would be foreclosed by the two suppliers remaining in that market.

(103) Nestlé has further referred to some imports by Spa, Apollinaris and some Italian producers.

Firstly, it is noted that so far no foreign beverage or food company of a size similar to Nestlé, Perrier and BSN has acquired a French water source with a capacity equivalent to that of the major sources owned by the three established suppliers.

Secondly, according to Perrier, several foreign companies have tried to import and sell on the French market during the last five years. Only a few are still present but none has achieved any significant volume of sales and none has been able to establish a brand name which could in the short term significantly compete with the wide range of brands owned by the three established supppliers.

(104) The sales of Spa in France (mainly canned still water sold through Coca-Cola vending machines) are negligible and have not succeeded in increasing since their introduction into France. The same is true for the sales of Apollinaris and the Italian producers San Pellegrino (20 % owned by Perrier), Ferrarelle (belonging to BSN), Levissima and San Benedetto. The imports of these latter companies are not in the still water segment but in the sparkling water segment and are principally directed towards the CHR (cafés, hotels, restaurants) sector. They are handicapped by a higher cost of transport in particular because they are mainly sold in glass bottles. As regards the German and Italian waters, these are sold at a higher price than French sparkling waters and thus cannot be considered as a price-constraining factor for French waters.

(105) Nestlé has also argued that there is a strong production substitutability between soft drinks and bottled water and that this should at least be taken into account for the assessment of the question of dominance. However, this production substitutability is only possible as far as purified treated table water is concerned, which can easily be produced by soft drink producers. This type of water is not marketed by anyone in France. It would not be a very good way to enter into competition with the existing French mineral and spring waters and it would involve high marketing costs and very high risk/high sunk costs. There is no existing or foreseeable demand for purified treated water in France (see recital 18).

(106) Finally, Nestlé has argued that end consumer prices of numerous soft drinks, in particular Coca-Cola drinks, are in the same range as end-consumer prices of bottled water and that Coca-Cola has confirmed to the Commission that its prices are targeted at bottled water. There would thus exist a certain degree of demand substitutability between soft drinks and waters which should be taken into account for the question of dominance. This argument cannot be accepted. The end consumer prices for colas and still mineral waters (which are the products at the centre of discussion in this case) are clearly different (see recital 13). In its letter of 12 March 1992, p. 5, Nestlé recognised that still mineral water products appear to be a cheaper alternative for the consumer. Basically, still water is roughly half the price of colas (see also Goldman Sachs Report of 9 September 1991, p. 8).

(107) It must be concluded from all the above that there is no effective price-constraining potential competition which could quickly and significantly constrain the market power of the two suppliers remaining on the French water market. The risk associated with entering that market would be increased by the proposed merger since newcomers would have to face a strengthened power and since all the major brands of Perrier would be foreclosed for acquisition by potential entrants.

H. Impact on the maintenance or development of effective competition

(a) In the case of implementation of the sale of Volvic to BSN: duopolistic dominance

(108) Given the high market shares and capacities of Nestlé and BSN after the merger and the sale of Volvic to BSN, the insufficient competitive counterweight from local mineral and spring waters, the increased dependency of retailers and wholesalers on the portfolio of brands of Nestlé and BSN, the absence of effective price-constraining potential competition from newcomers and the other characteristics outlined hereafter, the Commission must conclude that the proposed merger between Nestlé and Perrier would create a duopolistic dominant position which would significantly impede effective competition on the French bottled water market. This market constitutes a substantial part of the common market.

(109) Nestlé, supported by BSN, has submitted that Article 2 (3) of the Merger Regulation does not apply to oligopolistic dominance and, if it did, that the Commission has not proved the absence of significant competition between the established national suppliers and the absence of actual or potential competition from outside the oligopoly.

(i) Application of Article 2 (3) of the Merger Regulation to oligopolies

(110) Article 2 (3) of the Merger Regulation stipulates: A concentration which creates or strengthens a dominant position as a result of which effective competition would be significantly impeded in the common market or in a substantial part of it shall be declared incompatible with the common market.

The question is whether this provision covers only a market situation where effective competition is significantly impeded by one firm which alone has the power to behave to an appreciable extent independently of its competitors, customers and consumers, or whether this provision also covers market situations where effective competition is significantly impeded by more than one firm which together have the power to behave to an appreciable extent independently of the remaining competitors, of customers and ultimately of consumers.

(111) Nestlé has not denied the fact that from an economic point of view, both single firm dominance and oligopolistic dominance can significantly impede effective competition under certain market structure conditions (see Mr Sedemund's submissions of 19 May 1992, p. 44 and following; Lexecon Report, pp. 61 to 62).

(112) The Commission considers that the distinction between single firm dominance and oligopolistic dominance cannot be decisive for the application or non-application of the Merger Regulation because both situations may significantly impede effective competition under certain market structure conditions. This is in particular the case if there is already before the merger weakened competition between the oligopolists which is likely to be further weakened by a significant increase in concentration and if there is no sufficient price-constraining competition from actual or potential competition coming from outside the oligopoly.

(113) Article 3 (f) of the EEC Treaty provides for the institution of a system ensuring that competition in the common market is not distorted. One of the principal goals of the Treaty is thus the maintenance of effective competition. The restriction of effective competition which is prohibited if it is the result of a dominant position held by one firm cannot become permissible if it is the result of more than one firm. If, for instance, as a result of a merger, two or three undertakings acquire market power and are likely to apply excessive prices this would constitute an exercise of a collective market power which the Merger Regulation is intended to prevent by the maintenance of a competitive market structure. The dominant position is only the means by which effective competition can be impeded. Whether this impediment occurs through single firm power or collective power cannot be decisive for the application or non-application of Article 2 (3) of the Merger Regulation.

(114) In the absence of explicit exclusion of oligopolistic dominance by Article 2 (3) it cannot be assumed that the legislator intended to permit the impediment of effective competition by two or more undertakings holding the power to behave together to an appreciable extent independently on the market. This would create a loophole in the fundamental Treaty objective of maintaining effective competition at all times in order not to jeopardize the proper functionning of the common market. If, in order to avoid the application of the Merger Regulation, it sufficed to divide the dominant power between for instance two companies in order to escape the prohibition of Article 2 (3), then, in contradiction to the basic principles of the common market, effective competition could be significantly impeded. In such a hypothesis the objective of Article 3 (f) of the EEC Treaty could be overturned.

(115) Seen in the light of these legal and economic considerations, Article 2 (3) must be interpreted as covering both single firm and oligopolistic dominance. It is also significant to note that all other major antitrust systems with a merger control system apply or can apply their rules to both single firm and oligopolistic dominance, e.g. the American system, the French law (Article 38 of the Law of 1 December 1986); German law (22 GWB) and UK law (Fair Trading Act, Section 6 (2)). In most of these systems, it is an established practice to control mergers raising a problem of oligopolistic dominance. It cannot be the case that following the adoption of the Merger Regulation mergers which previously were subject to such control would now be subject only to single firm dominance control. The Merger Regulation would not only have transferred the national merger control powers to the Community but those Member States which had a system with oligopolistic dominance control would at the same time have abandoned such control altogether without any substitute for it at Community level. In the absence of any express provision to that effect, such a cession of control cannot be assumed.

(116) The argument of BSN that the Commission would be violating the principle of legal certainty because this case would be the first case to apply the oligopolistic dominance concept cannot be accepted. As explained above, the correct interpretation of the Merger Regulation leads to the conclusion that Article 2 (3) has always covered dominance which significantly impedes effective competition independently of whether such situation is the result of one or more than one firm. Furthermore, if the argument of BSN was right, it would mean that the Commission could never develop any of its administrative case law. Although BSN's argument could under other circumstances have some relevance in particular where a posteriori control takes place and involves interference in acquired rights, the merger control system is an a priori control system which by definition does not allow the implementation of mergers without prior authorization by the Commission.

(ii) Creation of a duopolistic dominant position

(117) Although Nestlé has stated that the Commission has not proved absence of significant competition between the suppliers, it has not brought forward any arguments to rebut the indicators mentioned by the Commission in its statement pursuant to Article 18 to the effect that, even before the merger, price competition between these suppliers was greatly weakened in particular the high degree of price parallelism over a long period of time, the very high production cost-price margin, and the large gap between ex-works prices of national mineral waters and local spring waters.

(118) The Commission has not asserted that there existed an oligopolistic dominance between Nestlé, Perrier and BSN even before the merger. The facts and market structures show, however, that the French bottled water market is already a highly concentrated market where price competition is considerably weakened. The maintenance or development of whatever competition there remains on that market therefore requires particular protection. Any structural operation restricting even more the scope for competition in such a situation has to be judged severely. The combination of the market structure arising from the merger and of certain additional factors (see recitals 119 to 120) lead to the conclusion that the proposed merger would have the effect of creating a duopolistic dominant position allowing Nestlé and BSN to jointly maximize profits by avoiding competition among themselves and acting to a large extent independently of their customers and competitors.

(119) After the merger, the degree of concentration would be extremely high with Nestlé and BSN holding over 82 % of the total French water market by value and 75 % by volume (nearly 95 % of all still mineral waters). There would only be two national still and sparkling mineral water suppliers on the French market. Both suppliers would operate in the most profitable segment of bottled water and have a strong interest and incentive to jointly defend their position.

After the merger, there would be no competitor in the Community approaching the size, the range of well-known brands and the geographic spread of either Nestlé or BSN. The remaining main company not linked to Nestlé or BSN would be Spa in Belgium, with total sales of 256,3 million litres of bottled water in Belgium and a turnover of Bfrs 8,05 billion in 1989 (ECU 185 million) which is several times smaller than that of Nestlé or BSN. The leading German bottlers are also much smaller. Apollinaris sales in Germany amounted in 1990 to approximately 345 million litres. Gerolsteiner Brunnen (427 million litres) and Mineralbrunnen UEberkingen (490 million litres) are also well below Nestlé or BSN in size.

(120) The reduction from three to two suppliers (duopoly) is not a mere cosmetic change in the market structure. The concentration would lead to the elimination of a major operator who has the biggest capacity reserves and sales volumes in the market. Perrier sources and brands would be divided between the two remaining suppliers. In addition, the reduction from three to only two national suppliers would make anticompetitive parallel behaviour leading to collective abuses much easier.

(121) The mineral water suppliers in France have developped instruments of transparency facilitating a tacit coordination of pricing policies

Retail prices of bottled water are transparent. The packaging and product size of the main products are the same in this market. Retail prices can easily be checked, and statistics are regularly supplied by different organizations.

In addition, Perrier, Vittel and BSN publish all tariff prices which are readily comparable (ex-works, taxes excluded, per bottle and for different means of transport). These basic prices constitute a reference on which pricing policies can be tacitly coordinated. An analysis of tariff price evolution through time (in real terms, i.e. effect of general inflation excluded) shows a high degree of price parallelism (see Annexes 5 to 8(12) ).

(122) Companies have developed instruments allowing to control and monitor each other's behaviour

Homogeneity and transparency of tariffs allow the mutual monitoring of pricing policies of competitors. In addition, Perrier, BSN and Vittel have implemented a mechanism of regular exchange of information on quantities sold each month, broken down by major brands. This allows each one of them to follow permanently, on a monthly basis, the behaviour and the sales evolution of each other. Moreover it would also permit immediate detection of any deviation by any single major brand of the expected performance.

After the merger, these instruments may become less indispensable given the reduction in suppliers on the market and the resulting increased transparency of the market itself. It also needs to be remembered that these suppliers carry out the bulk of their sales with the same customers.

(123) The two players remaining in the market are similar in size and nature

After the merger, there would remain two national suppliers on the market which would have similar capacities and similar market shares (symmetric duopoly): in the still water segment their market shares by volume would be Nestlé [...] %, BSN [...] % for all still waters and Nestlé [...] %, BSN [...] % for still mineral waters. By value, their market shares would be Nestlé [...] %, BSN [...] % for all still waters.

Given this equally important stake in the market and their high sales volumes, any aggressive competitive action by one would have a direct and significant impact on the activity of the other supplier and most certainly provoke strong reactions with the result that such actions could considerably harm both suppliers in their profitability without improving their sales volumes. Their reciprocal dependency thus creates a strong common interest and incentive to maximize profits by engaging in anti-competitive parallel behaviour. This situation of common interests is further reinforced by the fact that Nestlé and BSN are similar in size and nature, are both active in the wider food industry and already cooperate in some sectors of that industry.

(124) Demand is relatively price-inelastic; prices might be increased without fear of offsetting losses in volume

Because of the motivation of consumers to buy and consume mineral water on a daily basis, there are no real substitutes for mineral water. The overall image of branded water and brand loyalty accentuate the price-inelasticity of demand.

Increases in real prices have therefore a proportionately smaller impact on quantities demanded; a price increase would result in increased total revenue and profits. The incentive and possibility to increase prices jointly in this market seems to have already been recognized by the companies in the past (see the Annexes(13) showing price evolution in the last five years, in particular real prices). The proposed concentration facilitates and reinforces the likelihood of such a strategy being tacitly adopted by Nestlé and BSN. Both companies share a long common experience in the bottled water market in France.

(125) Costs are similar. No company seems to enjoy a significant cost advantage

It could be argued that companies facing different cost conditions could have a very different view on the prices they would like to prevail in the market whereby tacit coordination of pricing policies, without express and binding agreements among the companies, could become extremely difficult. Significant differences in costs can reasonably be considered an element that would hinder the implementation of tacit parallel behaviour.

This element is not present in this market. The raw material (source water) is free. The manufacturing process consists basically in bottling, which is a well-established technology where it would be difficult to gain a major technological advantage. The need to bottle at the source, and the resulting multiplicity of bottling plants reduces the scope for economies of scale. Finally, Nestlé has confirmed that the costs of the three major suppliers are not very different and the information provided to the Commission in this respect confirms this view.

(126) Technology is mature. R & D play no major role

The basic manufacturing process of the parties consists in bottling, which is a well-established technology. Nestlé has not given any indication of expected major changes in this respect. In addition, the scope for product innovation in the market of water is reduced. The water industry cannot be considered as an industry based on research, where new technological developments might quickly erode acquired positions. Research and development activities in this market are in fact of minor importance. The total expenditure of Nestlé in R & D with respect to waters (1,5 % of the annual turnover achieved in the sector) is a clear indication in this respect. This figure is lower than the average expenditure in the food sectors (see p. 33 of the notification). It can also be compared with the 15 to 25 % of turnover devoted by typically innovative industries to R & D.

(127) The notified operation in itself weakens the likelihood of the development of competition among Nestlé and BSN

Nestlé and BSN have together reacted strongly to the attempt by an external actor (Ifint) to acquire Perrier. Since BSN and Vittel have smaller spare capacities than Perrier, Perrier under ownership of a third party coming from outside the water market could have constituted an element of uncertainty and disruption in the market. The joint reaction of Nestlé and BSN to Ifint's bid can be viewed in this sense as a joint entry deterrence action.

(128) The subsequent agreement between Nestlé and BSN to share Perrier (sale of Volvic) further weakens the likelihood of development of competition between Nestlé and BSN. The competition between the two suppliers would further be reduced by the fact that BSN would hold two brands (Evian and Volvic) with a low mineralization degree (below 500 milligrams per litre) while Nestlé would hold three brands (Contrex, Vittel and Hépar) with a high mineralization degree (above 500 milligrams per litre). This split of brands would increase the product differentiation between the two suppliers with the result that competition would be further weakened and buyers would become even more dependent on each of these two suppliers. Finally, end consumers will suffer the consequences of the absence of competition in this market in terms of higher prices.

(129) Fringe firms (local spring waters) or retailers do not constitute a sufficicent competitive constraint

Local spring and mineral waters are too small and dispersed to constitute a significant alternative to the national waters. As examined in recitals 64 ff, none of these companies constitutes a sufficient price-constraining competitive force, which would be able to undermine the possibilities offered by the structure of the affected markets to BSN and Nestlé to jointly maximise their profits by avoiding competition among themselves. The same is true as regards the alleged buying power of retailers (see recitals 77 ff). None of these forces in the market has been able in the past to prevent the national mineral water suppliers from constantly increasing in parallel their prices in real terms.

(130) Barriers to entry are high

The conditions of entry and the effect of potential competition on the French water market have been examined above. It cannot reasonably be expected that actual entry or the threat of likely entry would jeopardize the ability of Nestlé and BSN to profitably increase their prices jointly in the future.

In addition, the merger in itself increases even further the associated risks and difficulties faced by newcomers.

(iii) Conclusion

(131) It must be concluded from all the above that the market structure resulting from the merger between Nestlé and Perrier (followed by a sale of the Volvic source to BSN) would create a duopolistic dominant position on the French bottled water market which would significantly impede effective competition and would be very likely to cause a considerable harm to consumers. Given the value of the total French water market (over FF 9 000 million per year) and the volume of water bought by French consumers (among the highest consumption rate per capita in the Community), the power to increase the prices of bottled water by only a small, but significant percentage (for instance 5 %) would lead to considerable losses for consumers.

(b) Without the sale of Volvic to BSN: single firm dominance

(132) If the Volvic agreement was not implemented, the proposed merger between Nestlé and Perrier would create a dominant position for the new entity by affording it the power to behave alone to an appreciable extent independently of its competitors, its customers and ultimately its consumers.

(133) The factors leading the Commission to this conclusion are as follows:

- the high market share in the French water market (approximately 60 % by value and 53 % by volume) which is more than twice the market share of the next biggest competitor,

- the combined overall and free capacities of the new entity both for still and sparkling waters: these capacities represent a multiple of those of the next biggest competitor and exceed by far the volume of the total water market,

- the portfolio of mineral water sources owned by Nestlé/Perrier: eight still mineral water sources of which three major sources (Contrex, Volvic and Vittel) and seven sparkling mineral water sources of which two major sources (Perrier and Saint-Yorre); this would create a major advantage vis-à-vis its competitors,

- the new entity would also hold an important number of spring water sources,

- at national level there would be one other competitor, BSN, with only one major still water source (Evian) and one major sparkling mineral water source (Badoit); BSN has a reserve of still mineral water which only leaves it a limited scope of action to compete with Nestlé/Perrier in the medium and long term. In the sparkling mineral water segment, its source Badoit has reached full capacity and can thus no more respond to an increase in demand,

- the competition from local spring waters would be insufficient to constrain the scope of action of Nestlé/Perrier,

- retailers and wholesalers would become even more dependent on the well-known brands of the new entity and could thus not effectively constrain the market power of Nestlé/Perrier,

- there is no sign of a price-constraining potential competition which could quickly and effectively enter the French water market.

(134) If in spite of its limited free capacity of Evian, BSN nonetheless maintained some scope of action to compete against Nestlé/Perrier, it cannot be expected that BSN would effectively compete against Nestlé/Perrier since both suppliers would have a strong common interest and incentive to jointly maximize profits.

In the still mineral water segment, BSN has only one major brand, i.e. Evian, with which it supplies both the French market and export markets. It is very unlikely that BSN would decide to use its available capacity to take away any significant volume from Nestlé/Perrier because it will want to participate in the increase in total demand and also maintain and if possible increase its exports. It can therefore be anticipated that BSN could not or would not effectively compete against Nestlé/Perrier.

I. Conclusion

(135) For all the reasons outlined above, the Commission concludes that with or without the sale of Volvic to BSN the proposed merger would create a dominant position as a result of which effective competition would be significantly impeded in a substantial part of the common market within the meaning of Article 2 (3) of Council Regulation No 4064-89.

VI. COMMITMENTS PROPOSED BY NESTLÉ

(136) Nestlé has offered to modify the original concentration plan as notified by entering into the following commitments:

While maintaining fully its position adopted in its written and oral replies to the Statement of the Commission pursuant to Article 18 of Regulation (EEC) No 4064-89 and dated 4 May 1992, Nestlé hereby declares that it is prepared to modify the effect of the operation notified to the Commission on 25 February 1992 on the basis of the following undertaking:

In order to meet the requirements of the Commission to facilitate the entry of a viable competitor with adequate resources in the bottled water market or the increase in the capacity of an existing competitor so that in either case such competitor could effectively compete on the French bottled water market with Nestlé and BSN, Nestlé has undertaken that it will make available for sale both brand names and sufficient capacity of water for bottling to such competitor as will permit that competitor to have not less than 3 000 million litres of water capacity per annum.

The brand names and sources to be divested are the following:

SpringCapacity (million litres)

Vichy[...]

Thonon[...]

Pierval[...]

Saint-Yorre[...]

A number of other water sources 1 195

The above capacity figures represent the best knowledge of Nestlé. The capacity figures are approximate but Nestlé undertakes to have these figures confirmed by an expert opinion.

Nestlé will be restrained from directly or indirectly providing any data that is less than one year old on its sales volumes to any trade association or any other entity which would provide that information to other competitors in the market in a form enabling other competitors to directly or by deduction identify the exact sales volumes of Nestlé. Nestlé agrees not to provide such data for as long as the present narrow oligopolistic market structure persists in the French bottled water market. Nestlé may at any time petition the Commission to show that market conditions and structure in the French bottled water market have changed sufficiently to allow a modification of this undertaking.

Nestlé agrees to hold separate from its own operations all assets and interests acquired from Perrier, until it has completed the sale of all bottled water operations described above to a single entity which is approved by the Commission. The buyer must be an entity that has no structural, financial or personal relationships with Nestlé, Perrier or BSN or any of the parent entities, or any of the subsidiaries of, or any officers, directors, employees, or agents of Nestlé, Perrier or BSN, which relationships are of such a nature that it would be unlikely that the buyer would compete effectively with Nestlé and BSN.

During the aforementioned period of separate holding, pending the sale of the assets to be divested, Nestlé also agrees that there will be no structural changes made in Perrier's business operations, such as the sale of a trademark or brand, the sale of a business unit, the closing of a manufacturing unit, or any disposal of similar major assets (including leasing, assigning or pledging of such assets) without prior Commission approval. Such approval will be deemed to be given if the Commission does not communicate in writing to Nestlé its opposition within 15 working days of the receipt of the notification of the intended structural change.

Until the divestiture required by this Decision has been accomplished, Nestlé:

(a) shall take all reasonable steps to assure that the assets of the Perrier group as a whole are maintained as distinct and saleable assets,

and

(b) shall preserve the documents, books and records relating to Perrier.

During the aforesaid period of separate holding, Nestlé undertakes not to permit the Perrier management to transfer any business secrets, know-how, commercial information or any other industrial information or property rights of a confidential or proprietary nature that it obtains from Perrier to any other commercial entity within the Nestlé group, nor to use any such information within the Nestlé group other than for the purposes of selling the assets which are the subjects of this commitment.

Nestlé agrees to sell the assets concerned by [...]. Nestlé shall be deemed to have complied with this obligation if, by [...], it has entered into a binding contract for the sale of the divestiture assets to a purchaser approved by the Commission, provided that such sale is completed within a time limit agreed to by the Commission.

Nestlé shall take all reasonable steps to accomplish the abovedescribed divestiture on or before [...]. The Commission may, in its sole discretion, upon Nestlé's request and upon showing good cause at any time on or before [...], extend said period for an additional three months.

Nestlé acknowledges that the approval of the purchaser by the Commission is of the essence for the acceptance of its undertaking by the Commission. The establishment of an effective competitor vis-a-vis Nestlé and BSN depends on the strength of the purchaser to develop the sources and brands which will be sold to it. The purchaser must in particular have:

- sufficient financial resources to develop a nation-wide distribution organization and to adequately promote the acquired brands;

and

- sufficient expertise in the field of branded beverage or food products.

These are the two main criteria. The Commission will withhold approval only if it can show that the proposed purchaser will not be a credible competitor or is unlikely to compete effectively with Nestlé and BSN, despite meeting these two criteria.

Nestlé acknowledges that it will be a condition of the Decision that Nestlé will not sell Volvic to BSN unless and until such time as it has fully completed the sale of the bottled water assets to be divested, it being understood that the sale of such assets is a prerequisite to a finding of compatibility with the common market.

Nestlé is enjoined and restrained from re-acquiring, directly or indirectly, any of the sources or brands which it divests pursuant to this undertaking, for a period of 10 years from the date of this Decision, without the prior written approval of the Commission. In this context, the Commission will take into consideration any changes in the market structure of the French bottled water market.

Without prejudice to its obligations under the Merger Regulation in respect of Community-dimension mergers, Nestlé also undertakes to inform the Commission in writing of any acquisition of a bottled water entity in France with a market share of sales exceeding five percent (5 %), for a period of five years from the date of the Commission decision.

(137) The Commission is satisfied that Nestlé's offer contains a range of mineral and spring water sources which would put in the buyer's hands the stated amount of approximately 3 000 million litres of capacity.

Nestlé has undertaken to sell the above capacity and brand portofolio of mineral and spring waters which should create effective competition to counteract the duopolistic dominance which would otherwise result from the proposed concentration and the sale of Volvic to BSN.

Provided that the buyer of the proposed bundle of brands and sources has the ability to develop quickly the assets which Nestlé will sell, the capacities offered would be, in the Commission's opinion, sufficient to create a new competitive entity in the French bottled water market which is likely to constrain the market power of Nestlé and BSN.

The new entity will have at its disposal a capacity of the French bottled water market which amounts to approximately 20 % of the total capacity previously held by the three established national suppliers and which is sufficient to cover more than half of the present total market demand. Although the present sales of these brands are still very low, their potential for development is high given the capacities and the required characteristics that the buyer must possess.

The sources to be sold include sources of still mineral water, still spring water, and sparkling water. They also contain a number of already well-established brands, or brands which are about to become nationally distributed brands in each of these water segments.

The buyer of this package will acquire a foothold in major retail stores, access to which is necessary to enable any new entrant to promote other lesser-known brands and to introduce new brands.

This, however, means that the buyer must have sufficient expertise in developing food brands and has sufficient financial resources to invest in the publicity required to support these brands.

Provided this is the case, it can be expected that the new entity will be albe to offer an alternative to consumers at prices below the high-priced national mineral waters. This type of market positioning could counteract the current pricing policies of the three national suppliers.

Therefore, subject to the fulfilment of the conditions and obligations set out in Nestlé's commitment, the Commission has concluded that the sale of a capacity of 3 000 million litres and of the agreed portfolio of brands to a strong buyer acceptable to the Commission would be sufficient to create a viable competitor able effectively to compete with Nestlé and BSN and thus to prevent the collective dominant position which would otherwise be created by the proposed concentration and the subsequent sale of Volvic to BSN.

(138) If the sale of the assets to be divested has not taken place by the end of the time period set out in Nestlé's commitment, or if any of the other obligations accepted by Nestlé are breached, then the Commission reserves the right, pursuant to

Article 8

(5), to revoke its decision to accept the modifications proposed here and to require that Nestlé divest all assets and interests in Perrier and thereby that Nestlé and Perrier be fully separated in order to restore conditions of effective competition, as provided by Article 8 (4).

These actions will be taken without prejudice to the Commission's right to impose fines pursuant to Article 14 (2),

HAS ADOPTED THE FOLLOWING DECISION:

Article 1

Subject to the full compliance with all conditions and obligations contained in Nestle's commitment vis-à-vis the Commission as set out in recital 136 of this Decision, the concentration notified by Nestlé SA on 25 February 1992 relating to the acquisition of control over Source Perrier SA is declared compatible with the common market.

Article 2

This Decision is addressed to:

Nestlé SA,

55, avenue Nestlé,

CH-1800 Vevey.

(1) OJ No L 395, 30.12.1989, p. 1; corrected version: OJ No L257, 21.9.1990, p. 13.

(2) OJ No C 319, 5.12.1992, p. 3.

(3) Annex not published pursuant to the provisions of Article 17 (2) of Regulation (EEC) No 4064-89 concerning non-disclosure of business secrets.

(4) Annex not published pursuant to the provisions of Article 17 (2) of Regulation (EEC) No 4064-89 concerning non-disclosure of business secrets.

(5) Annex not published pursuant to the provisions of Article 17 (2) of Regulation (EEC) No 4064-89 concerning non-disclosure of business secrets.

(6) In the published version of the Decision, some information has hereinafter been omitted, pursuant to the provisions of Article 17 (2) of Regulation (EEC) No 4064-89 concerning non-disclosure of business secrets.

(7) Prices reported in national currencies (Belgian francs and German Marks) were converted in French francs with the average exchange rate in 1991.

(8) Annex not published pursuant to the provisions of Article 17 (2) of Regulation (EEC) No 4064-89 concerning non-disclosure of business secrets.

(9) Annex not published pursuant to the provisions of Article 17 (2) of Regulation (EEC) No 4064-89 concerning non-disclosure of business secrets.

(10) Annex not published pursuant to the provisions of Article 17 (2) of Regulation (EEC) No 4064-89 concerning non-disclosure of business secrets.

(11) Annex not published pursuant to the provisions of Article 17 (2) of Regulation (EEC) No 4064-89 concerning non-disclosure of business secrets.

(12) Annex not published pursuant to the provisions of Article 17 (2) of Regulation (EEC) No 4064-89 concerning non-disclosure of business secrets.

(13) Annex not published pursuant to the provisions of Article 17 (2) of Regulation (EEC) No 4064-89 concerning non-disclosure of business secrets.