EC, January 31, 1994, No M.315

THE COMMISSION OF THE EUROPEAN COMMUNITIES,

Having regard to the Treaty establishing the European 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 20 September 1993 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. THE PARTIES

(1) Dalmine is an Italian undertaking, active in the steel, metals, engineering, electric and mineral sectors. Dalmine's ultimate parent is Ilva SpA, an Italian State-owned company.

(2) MRW is a German undertaking active in the production of pipes and tubes, the production of steel and prematerial for pipes and tubes as well as the processing, finishing and sale of such products. MRW is a subsidiary of Mannesmann AG, a German company active in the steel industry, the manufacture and sale of various industrial products as well as the planning and construction of factory sites.

(3) Valtubes is a French company active in the production and sale of pipes and tubes as well as products made therefrom. Valtubes is a wholly owned subsidiary of the French company Vallourec SA which is active in the manufacturing, processing and selling of steel materials. The French State-owned company Usinor holds a 28 % stake in Vallourec and none of the other shareholders hold a stake in excess of 5 %.

II. THE OPERATION

(4) The parties will establish DMV (Dalmine/MRW/Valtubes) as a holding company in which they will each have an equal interest. DMV will own, manage and control four companies (one in each of the home countries of the parents plus one in the United States of America) running the previous seamless stainless steel tube businesses of the parties.

As part of this operation, the parties intend to close down the Dalmine extrusion press at Costa Volpino, keeping the two presses located in France and Germany which they intend to specialize to realize cost benefits and economies of scale. They will also close down some cold-finishing facilities in the DMV French and German plants.

III. CONCENTRATION

Joint control

(5) Each of the parties will hold 33,33 % of DMV's shares. None of the parties will therefore be able to dispose alone of a majority of the votes in a shareholders' meeting. Any disposal of shares to a third party is prohibited for a period of [...] (3) unless the prior written consent of all parties has been given.

(6) Under the terms of the joint-venture agreement of 28 June 1993, a supervisory board of six members, with two members appointed by each party, shall be established. The supervisory board shall take its decisions unanimously and will deal with all essential business, financial and R& D questions which go beyond daily business matters.

The supervisory board will set up an Executive Committee composed of the respective chief executives of the parties. The Executive Committee shall take its decisions unanimously. It shall select the managing directors who shall be members of the board of directors.

The board of directors will be composed of three managing directors. It will manage DMV and run its day to day activity. The board of directors will be supervised by the supervisory board.

Consequently, DMV will be jointly controlled by its parent companies.

Joint-venture performing on a lasting basis all the functions of an autonomous economic entity

(7) Articles 3 to 6 of the joint-venture agreement provide that each of DMV's European subsidiaries will be endowed with the assets (including all intellectual property and related rights), fixed assets (with the exception of land and buildings which will be leased) and liabilities, which relate to DMV's activities as described above. 30 % of the parties' current production capacity shall be closed down, with the remaining plants each specializing in certain sectors to realize economies of scale.

In addition, Article 14 of the joint-venture agreement states that each party shall maintain in or transfer to DMV all key people (technical, commercial and administrative) necessary to ensure the proper functioning of the joint venture.

(8) To enable the joint venture to acquire autonomy rapidly and to operate successfully on the market, Article 20 of the abovementioned agreement provides that the parties undertake to continue to supply products, material and services linked to the transferred plants. However, the parties have stated that DMV will remain free to determine its own purchasing policy and therefore to make its purchases elsewhere.

(9) As a result of the above, DMV will be endowed with all the necessary means to perform on a lasting basis all the functions of an autonomous economic entity.

Absence of coordination of competitive behaviour

(10) Following the operation, the parties will not remain active in the production of seamless stainless steel tubes. Only Vallourec, through its subsidiary 'Valinox nucléaire', was and will remain active in seamless tubes for sensitive parts of nuclear plants such as steam generators. These tubes require a specialized workshop and are subject to specific technical and regulatory requirements. They should therefore be considered as belonging to a separate product market.

(11) All other nuclear activities, not related to the sensitive parts of the nuclear plants and not requiring specific workshops, were performed by all parties prior to the operation and would be merged in DMV. However, both DMV and Valinox nucléaire could produce a very limited quantity of very special tubes such as heat exchangers and guide tubes. These products are only manufactured in order to meet customers' demands for a full range of products, and the overall tonnages involved amounted to [...] in 1992 (less than 0,1 % of the overall seamless stainless steel tube market).

This particular activity of one of the parent companies in DMV's field of business should therefore be considered as de minimis.

(12) The three parent companies will remain competitors in the field of finished oil country tubular goods (OCTG) tubes. 98 % of the OCTG business relates to carbon steel tubes and the remaining 2 % to seamless stainless steel tubes.

As they will have transferred their manufacturing facilities to DMV, the parties will no longer be able to manufacture the latter tubes. Consequently, to enable a full programme offer, Article 19 of the joint-venture agreement provides that the parent companies will purchase unfinished OCTG stainless tubes from DMV, will finish them and will sell them through their respective specialized networks to the oil companies. The highly sophisticated finishing techniques and high-tech threading machines needed for the finishing of OCTG tubes, both stainless and carbon, make it unlikely that DMV would subsequently enter this market in competition with the parent companies.

The production cost of the unfinished tubes does not represent a major part of the final price for these specific tubes, and it is therefore unlikely that the abovementioned purchasing of unfinished tubes from DMV would give rise to an appreciable risk of coordination of competitive behaviour between the parent companies.

(13) The three mother companies themselves, their parent or ultimate parent companies, or companies with which they have capital links, will also remain important competitors in the market for bar and billets. Bars and billets are the raw material for seamless stainless steel tubes as well as for other inoxydable products. These products are covered by the ECSC Treaty and the effects of the operation on the market for these products are currently being examined according to the rules of the ECSC Treaty.

(14) As described below, the market for seamless stainless steel tubes should be considered as a separate product market. The parties will transfer to DMV their extrusion presses and will consequently no longer be able to manufacture seamless stainless steel tubes. Given the high investment represented by an extrusion press, it is highly unlikely that any of the parent companies would re-enter the market.

(15) The notified operation therefore constitutes a concentration within the meaning of Article 3 of the Merger Regulation.

IV. COMMUNITY DIMENSION

(16) The combined aggregate worldwide turnover of the three undertakings concerned exceeds ECU 5 billion. At a worldwide level and in 1992, Ilva had a turnover of ECU 6,33 billion, Mannesmann of ECU 13,6 billion and Vallourec of ECU 966 million, even without taking into consideration the Usinar group turnover. The aggregate Community-wide turnover of each of the undertakings concerned is more than ECU 250 million. The Community-wide turnover in 1992 was as follows: ECU 5,43 billion for Ilva, ECU 8,9 billion for Mannesmann and ECU 803 million for Vallourec. None of the undertakings concerned achieves more than two-thirds of its turnover in one and the same Member State. The proposed operation therefore has a Community dimension.

V. COMPATIBILITY WITH THE COMMON MARKET

A. Relevant product market

(17) Unlike carbon steel tubes, seamless stainless steel tubes (SST) have a high content of chromium, a low content of carbon and may contain non-ferrous metal such as nickel or titanium. The high content of alloying elements in stainless steel tubes makes them inoxydable whereas carbon steel tubes are not. Such a content of alloying elements also leads to prices 5 to 10 times higher than those of carbon steel tubes. They are therefore only used for applications that require inoxydable material.

(18) The three parent companies have stated that steel can be considered as 'inoxydable' (stainless) when the chromium content is of at least 13 %, while the CEN standard for stainless steel starts at 10,5 % (CEN 10,020).

The parties will remain active in the range of stainless steel tubes with a chromium content below 13 %.

From the demand side, most customers questioned by the Commission have confirmed that below 13 % chromium content, a tube is an alloyed steel grade tube and not a stainless tube.

From the supply side, 13 % constitutes a dividing line. With 13 % or higher chromium content, seamless stainless steel tubes have to be manufactured on an extrusion press. Below 13 % chromium content, seamless carbon steel tubes and alloyed steel grade tubes are manufactured on rolling mills. As a result, from both demand and supply side, stainless steel tubes with a 13 % or higher chromium content and alloyed steel grade tubes with a less than 13 % chromium content must be considered as belonging to separate product markets.

(19) Stainless steel tubes can be either seamless or welded. They are used for a variety of applications such as, on the one hand, in the nuclear, chemical and oil industries but also, on the other hand, in the construction, car manufacturing and food industries. The former type of applications only use seamless tubes while the latter predominantly use welded tubes. 20 years ago, seamless accounted for around 80 % of the overall stainless tube applications and welded for around 20 %. Their respective shares are now reversed, i.e. 20 % for seamless and 80 % for welded, mainly due to the progressive upgrading of the welded technology and a high price difference between welded and seamless tubes. This substitution process is now almost complete and seamless stainless steel tubes are generally only used when it is unacceptable to use welded stainless steel tubes.

With respect to the product market definition, the following points should be taken into account:

- there is only a limited technical overlap between seamless and welded stainless steel tubes due to technical production limitations in terms of diameters and wall thickness,

- within the overlapping area, there is a high price difference, the average price for seamless tubes being twice the average price of welded tubes; the manufacturing techniques are very different as seamless steel tube production requires a much higher initial investment than welded tube production; the evolution of prices is different for seamless tubes than that of welded ones; international standards used in oil, chemicals and pharmaceuticals (e.g. the US ASTM standard) impose the use of seamless tubes; as stated above, the main fields of use for the seamless and welded tubes are different.

Customers and end users questioned by the Commission confirmed these points and stated that they could normally not substitute a welded tube for a seamless one.

Consequently, the market for seamless stainless steel tubes should be considered as a separate relevant product market.

(20) This relevant product market is composed of two distinct product segments, i.e. hot-finished and cold-finished tubes. All tubes are extruded and hot-finished. They can then be sold as such or be cold-finished. Cold-finishing is typically done to manufacture tubes below a certain diameter, or to improve the tube characteristics in order to meet special tolerance standards.

In the first case, there is no size overlap between hot and cold products whereas an overlap does exist for the latter. Therefore, together cold and hot tubes make up the full range of sizes.

Both hot- and cold-finished products are bought by the same customers as part of packages; they are both produced by the same manufacturers (although in some countries, specialized companies called redrawers carry out part of the cold-finishing using extruded hot hollow tubes or cold tubes). The price for cold-finished tubes is substantially higher than that for hot-finished tubes, and the costs and manufacturing technologies as well as the competitors' capacity loading is different for hot and cold products. Finally, cold-finishing is more labour intensive than hot-finishing.

The competitiveness of a supplier in the market depends, to a significant degree, on its ability to offer a full range of products including both hot- and cold-finished tubes.

B. Geographical reference market (a) The geographical market for seamless stainless steel tube products (hot and cold) is western Europe (Community and EFTA countries)

(21) There is no tariff between the Community and EFTA areas. There is intensive trade in seamless stainless steel tubes between the different countries which belong to these two areas. A large part of Community consumption comes from EFTA countries. Transportation costs are very low within the Community and EFTA (hereafter referred to as western Europe).

In addition, there is a high degree of mutual interpenetration from State to State within western Europe. For example, the proportion of imports from other Community and EFTA countries into countries where a national producer exists is high (around 60 % in France and Germany, and around 30 % in Italy). To date, the leader within western Europe as a whole is an EFTA company, Sandvik from Sweden, even though each supplier holds a strong position in its own domestic market (i.e. MRW in Germany, Valinox in France, Dalmine in Italy and Tubacex in Spain).

Finally, the structure of supply and demand is similar from State to State. In particular, distribution is widely achieved through a number of stockholders in each country.

(b) The geographical market is not wider than the Community and EFTA

(22) At a worldwide level, the SST business is characterized by three main areas of consumption and two main areas of production. Each of these areas is characterized by different patterns of supply and demand.

According to estimates provided by the parties and by other producers, western Europe is the biggest market (consumption of [between 50 and 60] KT) followed by North America (30 KT) and Japan (25 KT). As regards production, two main areas should be distinguished: western Europe and Japan. In western Europe as well as in Japan, production largely exceeds consumption, resulting in a high level of overcapacities. Japanese exports to western Europe have not exceeded 11 % of total Community consumption and western European exports to Japan have been negligible. Producers from both areas make substantial exports to North America. Sandvik and Sumitomo recently began to jointly operate an extrusion press in the United States of America. Nonetheless, North American production will still only cover a small proportion of total US consumption.

(23) There is a low rate of mutual interpenetration between the different areas. Imports are stable at a level of approximately 10 % in western Europe (mainly due to the Japanese producers, with North American exports being insignificant) and are said to be negligible in Japan. By contrast, US imports are high (60 to 80 % coming mainly from Europe and Japan). Part of these imports seem to be re-exported, mainly to South America.

The transport costs faced by producers located outside Europe would represent on average 2 or 3 % of total costs.

Most of the imports to western Europe have to face a 10 % import duty. This is true in particular with regard to the Japanese producers. However, within the framework of GATT the reduction of import duties on steel products, including seamless stainless steel tubes, is presently being negotiated. The text under negotiation will provide for a linear reduction of import duties over a 10 year period, starting from 1 January 1995.

(24) The markets are also characterized by different price levels. The producers and stockists of seamless stainless steel tubes who replied to the Commission's inquiries have described the prices as being up to 35 % higher in Japan (on average approximately 20 %) and on average 5 % higher in the United States of America than in western Europe.

(25) The customers are different from area to area. Only a few large stockists are active worldwide where they compete for international bids which represent, according to the parties, no more than 5 % of the overall SST business.

Even if the present price level in the western European market was to increase by a small but significant percentage, e.g. 5 %, it could not be expected that such an increase in price would bring about imports by the larger stockists from Japan, as the difference in price level between Japan and western Europe, taking into account the 10 % import duty and transport costs, would still be substantial.

If prices in western Europe were to increase, for example by 5 %, reimports of SST exported by western European or Japanese producers to North America would also not be possible given the higher price level in North America, the 10 % import duty and transport costs. Reimports would only become attractive with a stronger price increase in western Europe.

(26) Furthermore, the structure of distribution in the three areas is different. In the United States, all the seamless stainless steel tubes are distributed only through independent stockists of very large size and no direct sales take place. By contrast, in western Europe, even the largest stockist accounts individually for less than 1 % of the market.

(27) Three other areas can be distinguished: eastern Europe, India and Asia. The latter two regions are promising and fast growing markets and extrusion presses are located in these areas but their production is low.

Imports from Asia and India to western Europe are negligible and are not expected to increase significantly in the foreseeable future given the high underlying rate of growth of domestic demand. In addition, such imports are subjected to duties which still constitute an obstacle for an increase in sales within the western European market.

With regard to eastern Europe, trade flows are very low, and production and quality standards are still significantly lower than the western European ones. The producers located in the CIS are subject to import duties of 10 % when importing SST to western Europe. This does not apply to producers of other eastern European countries, where duties are gradually being phased out by 1996. The Commission has negotiated tariff quotas for steel products (including the SST) with the Czech and Slovak Republics. Duty-free ceilings exist for the trade with Romania, Russia and Ukraine. The parties mentioned the existence of producers in the Czech Republic (Chomutov), Romania, Ukraine and Russia (see below paragraph 121).

(28) Despite the abovementioned factors, the parties assessed the relevant geographical market as world-wide. Their assessment relied on four main elements as indicated below; these elements are price correlations, international trade flows, the existence of international suppliers as well as international stockists and customers.

(i) Price correlations

(29) According to the parties, correlations between prices charged in western Europe and the United States - ranging from coefficient 0,5 to 0,7 - are close to correlations between prices charged in the various western European countries. This situation would 'imply that producers are unable to price discriminate in different parts of the world'.

(30) In analysing this result, it is important to first analyse to what extent these price correlations are influenced by other factors which influence price series in different markets.

The parties' correlations are based on a comparison between the prices charged in western Europe and in the United States by the four Japanese producers for products manufactured in Japan. It is clear that this basis is more likely to be influenced by factors common to the producers rather than, for example, a correlation based on domestic prices in the United States, western Europe and Japan.

However, even on the basis of these figures, the parties' conclusion that the alleged high correlations would demonstrate that the Japanese implemented similar pricing policies in western Europe and in the United States is to be questioned.

Based on the parties' data, the price difference between Japanese exports to western Europe and to the United States between 1987 and 1992 is of 3,2 % in 1987, of 18,7 % in 1988, of 8 % in 1989, of 9,5 % in 1990, of 16,4 % in 1991 and of 18,4 % in 1992. These figures do not confirm the parties' conclusion that 'the Japanese cannot change prices in one of the markets without correspondingly changing the other'.

(31) A second series of data concerning two typical stainless steel products of the notifying parties led to totally different price correlations. For example, Dalmine's correlation for one of these products between the United States and western European prices is extremely high (0,98, i. e. almost identical price movements). The parties draw the conclusion that Dalmine is unable to discriminate on the basis of price. For the same product, Vallourec's price correlation is 0,47 (i. e. half that of Dalmine). The correlation for Mannesmann's prices was not shown. Such differences between the two correlations are inconsistent with the concept of a world market.

(32) Finally, while absence of price correlation between two geographical areas is a strong indicator of different geographical markets, the existence of price correlations does not necessarily indicate a Single Market in the absence of other elements such as mutual interpenetration or similar structures of supply and demand in the different areas.

(ii) International trade flows

(33) The parties relied on the Elzinga-Hogarty test according to which, above 10 % of exports or imports with respect to production and consumption respectively, an area has to be considered as belonging to a wider market. However, this test is widely regarded as questionable in particular because it does not give any indication with respect to mutual interpenetrations between the different areas considered. Only if there is mutual interpenetration can purchasers arbitrage price differences by buying in other areas. In the present case, even though the 10 % export threshold is met, interpenetration is very low.

(iii) International producers, stockists and customers

(34) Finally, the parties infer the existence of a world market from the activities of international producers, stockists and customers.

(35) As far as producers are concerned, it is true that the Japanese and western European producers are both active in western Europe and North America. However, the western European producers are not active on the Japanese market and no US producer is currently active on the western European market.

Furthermore, an analysis of the evolution of prices on the western European market as compared with the evolution of the market share of Japanese firms on the same market leads to the conclusion that the level of Japanese imports is not strongly affected by the level of prices. Imports into western Europe do not appear to react to changes in demand in western Europe, as they have been fairly constant in volume since 1991. They would therefore not appear to justify a wider definition of the reference market. One explanation for this is the fact that Japanese producers have concentrated their exports to western Europe on very small diameter high value-added tubes.

(36) On the demand side, competition only takes place at worldwide level for big international tenders launched by the largest end users. According to the parties, these international tenders represent only 5 % of the overall business concerned. The parties also indicated that the western European market should be divided as follows, one third being directly sold to end users (of which 5 % through big international tenders) and the rest to the stockists.

The parties also point to the activities of large international stockists active on a worldwide level. There is a very small number of large stockists (four to six). One (TAD) is a group stockist of Dalmine, another [. . ] is connected to one of the parties through a minority capital link, while a third one [. . .] is connected to the Avesta-Sandvik group.

These large stockists compete through international tenders for the supply of SST products and associated services, including inter alia stainless steel tubes, to international oil, chemical or pharmaceutical companies. However, this activity is unrelated to their traditional stockholding and distribution activity in western Europe (no stock is held in anticipation of these large tenders). For example, [name], presented by the parties as the most internationalized stockist, has in fact only sourced from western European producers within the last three years. Total purchases of [name] have in any event amounted to less than 1 % of the total western European demand for SST in 1992.

(c) Conclusion on geographic market

(37) There is a high penetration of the US market by European and Japanese producers, Japanese producers also have a permanent significant but relatively small presence on the European market. Consumption at world level is characterized by low mutual interpenetration, different price levels between the main markets of the United States, Europe and Japan, different mixes of customers and distribution structures, and a different balance between domestic production and imports. In addition, western Europe is protected by a 10 % duty which is applicable to imports from a range of third countries including Japan and the CIS countries. On the other hand, the Japanese market is characterized by high barriers to entry, whereas there is a 6 % import duty in the US market. The geographic reference market should therefore be considered to be western European.

C. Assessment of compatibility with the common market (a) General features of the market

(i) General overview

(38) The market for hot and cold-finished seamless stainless steel tubes is to a large extent a commodity market which is in a mature phase. Substitution between seamless and welded tubes is now said to be virtually complete.

Two structural changes took place in the SST sector within the last years. The first occured in the mid 1980's, when the production of thick carbon steel seamless tubes was shifted from extrusion presses to more cost-efficient rolling mills. The second occurred in the early 1990's, when the demand from the former eastern European countries (a traditional outlet for west European manufacturers) collapsed.

Following a drastic decrease over the last 10 years, demand is now expected to be stagnant in the long run in western Europe (i. e. 0 to 1 % growth in general) and the elasticity of global demand to prices is low. Furthermore, the demand in the market for seamless tubes is very sensitive to overall business cycles in the economy. Market conditions have worsened during the early 1990's in parallel to the worldwide economic slowdown.

This overall situation as well as the drop in prices for raw materials (in particular nickel) led to decreasing prices within the last years. The current price level in the western European market appears to be close, in nominal terms, to the 1988 price level.

(ii) Overcapacities

Theoretical overcapacity

(39) The structural changes described above have resulted in a significant overcapacity in western Europe.

In the production of seamless stainless steel tubes, the maximum theoretical output of a plant is in principle decided by two factors: the capacity of the extrusion press (including adjacent facilities for piercing, cutting, heat-treatment, etc.) which determined the maximum possible production of hot-finished products, and the capacity of the cold-finished facilities (pilgering and drawing facilities). The maximum capacity is influenced by the product range produced. However, for the purpose of assessing capacities, this influence can be disregarded as all producers on the market produce a range of sizes and grades which is sufficiently homongeneous.

(40) All seamless stainless steel tubes have to go through the hot-finishing process before they can be sold as cold-finished products. This means that the total output of hot-finishing products equals the sum of hot-finished tubes that are used as the prematerial for the cold-finished process plus hot-finished tubes that are sold as such. For all seamless stainless steel tube producers, the maximum output of the higher value-added cold finished tubes is determined by the available facilities in cold-finished. When those facilities are used to a maximum, the producer has no other opinion than to sell the remaining production as hot-finished products.

(41) On the basis of these definitions, the post-merger production capacities of hot- and cold-finished tubes within western Europe are estimated at approximately 125 kilotons (KT). The current rate of utilization of the production for hot- and cold-finished products is approximately 50 %.

Employment and output

(42) Within these maximum theoretical capacities, the effective capacity is determined by the current level of employment, i. e. number of working days per year and shifts per day. Based on current employment and production in 1992, the average use of capacity of the producers in the western European market, in the post-merger situation, is estimated at approximately 79 % in hot-finishing and 87 % in cold-finishing. A certain margin therefore already exists at current employment levels for producers to increase output. Producers also have a certain degree of flexibility to increase production through increases in overtime and temporary employment, etc., although beyond a certain level, output could only be increased by significant increases in employment (e.g. extra shifts).

(43) However, in the context of a market with a highly concentrated supply, a long-term stagnant demand and a low price elasticity of demand, it would be unreasonable for a producer to increase its current level of employment to any substantial extent, thus bringing its effective capacity closer to its maximum theoretical capacity. For this reason, none of the western European producers of seamless stainless steel tubes have indicated a willingness to do so. The current levels of employment are therefore an important, though not necessarily determinant factor in the competitive environment of the industry for the foreseeable future.

(iii) Market structure

(44) For the reasons stated in paragraph 21, the relevant geographic market for assessing the impact of the proposed concentration is western Europe. However, according to the parties, the shares of the main players at world level would be (volume-based estimates): Sandvik 23 %, Sumitomo 23 %, DMV 16 %, other Japanese producers 15 %, Tubacex 10 %, SBER 5 %, Sammi Steel 3 % and others 5 %.

Suppliers and customers in the geographic reference market

(45) Six suppliers are curently located in the geographic reference market: the Sandvik group, Dalmine, Mannesmann, Vallourec, Schoeller-Bleckmann (SBER) and Tubacex. After the merger only four competitors would remain. The main other suppliers in the market are the Japanese firms which traditionally account together for around 10 % of the market, although this figure has been much lower in recent years. There are virtually no other suppliers in the market.

(46) There are two categories of customers: end users and stockists. The end users purchase either small quantities from the traditional stockholders or bigger quantities ex-mill, i. e. directly from the producers. Some large international end users launch international tenders for large first equipment bids. A few large stockists (four or six in Europe) compete regularly in this small end of the market (i. e. 5 % of the market) where they hold no stock but offer packaged services either in association or in competition with the seamless stainless steel tube producers. In addition, large stockists sell a small proportion of their total sales to other smaller stockists.

Market shares

(47) According to the parties, in 1992, the apparent consumption in the Community amounted to [between ECU 250 to 300] million representing [between 50 and 60] Kilotons (KT).

In order to take into account the different product mixes of the various suppliers to better reflect the competitive strength of each competitor, the Commission based its assessment on market shares calculated in value. However, even if market shares in volume are used, the assessment would not be materially different.

Prior to the operation, the parties achieved the following value-based market shares: Dalmine 14 % (ECU [. . .] million), MRW 13 % (ECU [. . .] million) and Valinox 9 % (ECU [. . .] million). Following the operation, the overall value-based market share of DMV would amount to 36 %.

According to the parties, the strongest producer and competitor would be Sandvik which would achieve a 29 % value-based market share composed of 11 % through its UK subsidiary, Sterling Tubes, and 18 % through imports to the Community from Sandvik's plant in Sweden. The remaining two western European competitors, Schoeller-Bleckmann (SBER) from Austria and Tubacex from Spain would achieve respectively 11 and 13 % value-based market shares. Two small German competitors (P& P and TPS) together account for 3 % (value-based) of the overall Community market. Imports from outside western Europe are indicated as accounting for 8 %, mainly Japanese firms.

(48) The investigation carried out by the Commission broadly confirmed the parties' estimates, and also confirmed that the respective overall market shares would not be significantly different if the EFTA territories were taken into account. It also showed that the Community sales in tons as well as the shares of the Japanese producers were overestimated, and that the shares of the main competitor of the merging parties have been somewhat underestimated. The combined market share of DMV and Sandvik after the merger would amount to approximately 70 % in value of the Community market (DMV approximately 36 % and Sandvik approximately 33 %).

(49) A breakdown, Member State by Member State, shows that Germany with a total apparent consumption of close to ECU 100 million (ECU [. . .] million) is by far the largest outlet and is almost twice as large as the second largest outlet, Italy ECU [...] and [...] million. France and the United Kingdom account respectively for ECU [...] and [...] million. Spain and the Netherlands account respectively for ECU [...] and [...] million. Belgium/Luxembourg and Denmark account for ECU [...] and [...] million. Finally, the Republic of Ireland, Portugal and Greece account respectively for (ECU [...], [...] and [...] million).

(50) After the merger, the parties and Sandvik would hold relatively balanced market shares in the overall western European market (see paragraph 48) as well as in the largest geographical outlet (Germany, where there would be a 6 to 8 % market share difference).

Sandvik would be the leader in northern Europe, namely Benelux (around 50 % value-based market share), the United Kingdom and the Republic of Ireland (around 50 % value-based market share in the United Kingdom and Ireland taken as a whole) and Denmark (above 75 % value-based market share). In Sweden and Norway, Sandvik's market share is equally very high.

DMV would be the leader in southern Europe, namely France (above 50 % value-based market share), Italy (above 50 % value-based market share), and Greece (above 50 % value-based market share).

Tubacex would be the leader in its home market (i. e. Spanish market).

(b) Assessment of dominance

(i) Sole dominance

(51) It results from the above figures that although DMV would hold a value-based market share of approximately 36 % following the merger, it cannot be considered that the proposed operation would lead to a single dominant position of DMV as a result of which effective competition would be significantly impeded in the common market. This is true in view of the strength of the second largest competitor, the Swedish firm Sandvik, which would hold a very similar market share to that of DMV.

(ii) Duopolistic dominance

(52) The Commission considers that the proposed joint venture does not lead to the creation of a joint dominant position of DMV and Sandvik significantly impeding competition in the common market.

(53) The merger would bring about a radical change in the market leading to close market shares and a high combined market share between two firms.

While a merger can be the vehicle for reducing structural overcapacities in a market and mitigating the effects of a recession, it is important that the higher level of concentration which results does not lead to the creation of a position of joint dominance which could harm effective competition on a lasting basis.

Such oligopolistic dominance would exist if a small group of companies were in a position to behave to an appreciable extent independently of the other actual and potential competitors, the customers and ultimately the consumers.

(54) It is true that the post-merger western European SST market structure, would increase the probability of collective dominance. Nevertheless the overall assessment of the Commission is that, on the basis of available fact and evidence as described below, collective dominance within the meaning of Article 2 of the Merger Regulation will not be created by the post-concentration market structure.

A. Structural market conditions creating a strong incentive to engage in anti-competitive parallel behaviour

(55) Although Sandvik and DMV do not have any structural links between them, an incentive for them to enter into anti-competitive parallel behaviour can be based on several factors. The existence of a level of market concentration has to be sufficient to allow anti-competitive parallel behaviour. Moreover, this incentive, i. e. the profits arising out of such behaviour, increases both with the level of concentration and the symmetry of market shares.

After the merger, Sandvik will have a 33 % market share and DMV 36 %. Given the resultant high level of concentration and the close similarity of the market shares of the two principal producers, there is prima facie a strong incentive for these producers to act together to sell the same volumes at higher prices, rather than to compete with each other on price in order to obtain higher market shares.

The incentive to engage in anti-competitive behaviour is further strengthened because demand is stagnant and price-inelastic. The overall demand price-elasticity is low on the SST market due to the fact that SST have no substitutes, and, moreover, since the cost for SST in any industry forms only a small part of total costs. Therefore, even a significant increase in prices is unlikely to result in a large drop in demand.

Furthermore, effectively only four European players would remain active in the seamless stainless steel tube market and two of them (Sandvik and DMV) would hold two-thirds of the overall Community/EFTA consumption. The remaining two European competitors, i. e. Tubacex and SBER, would only have a market share of respectively 13 and 11 %.

In such a situation, both Sandvik and DMV would have an incentive to enhance their position by restricting output and increasing prices. This is true in particular as any increase in output by one of them would have a direct and significant impact on market price, and hence on their respective revenue.

(56) Subject to the further factors to be examined below, it follows that the structural conditions would exist for the two main players, DMV and the Sandvik group, to have a strong incentive to enter into anti-competitive parallel behaviour leading to a significant increase of market prices.

B. Structural differences between DMV and Sandvik

(57) The incentive to engage in anti-competitive parallel behaviour might be less strong, if there were significant structural differences between the two market leaders that resulted in different possibilities and/or incentives to react to changes in market conditions.

(58) According to the parties, such significant structural differences exist between DMV and Sandvik in terms of corporate size and activities, as well as capacities and costs. The parties' statements are examined in detail below. The conclusion from the Commission's investigation is however that, given all the other elements of the market, the existing structural differences between DMV and Sandvik are not sufficient as to offset their incentives to engage in anti-competitive parallel behaviour.

C. Market share and corporate size of DMV and Sandvik

(59) As shown above, following the operation, the respective market positions of both DMV and Sandvik would be relatively symmetric in western Europe and also in Germany, the most important national market. DMV has a stronger market position in the south, whereas Sandvik is stronger in the north. But this fact is not likely to disturb the overall balanced positions of interest on the western European market. None of these two companies can afford to engage in a price war against the other one in its strongest markets without the risk of retaliation and a substantial loss of revenue for both companies. The situation would be different if either DMV or Sandvik had a higher market share in a clearly bigger consumption area compared to that of the other company. This however is not the case here.

According to the parties, there are a number of factors that put Sandvik ahead of DMV, namely its larger volume of stainless products, its global strategy (including cold-finishing production facilities in North America), the newly started extrusion press in a US joint venture with the Japanese producer Sumitomo, the probable acquisition of the Czech company Chomutov and Sandvik's worldwide market share in SST.

(60) A comparison between the total business volumes of, on the one hand, the whole Sandvik group and, on the other hand, DMV excluding its parent companies would not be appropriate. The parent companies of DMV are all part of large groups and will, according to the notification, maintain business activities in several fields related to stainless steel production that are distinct from the production of SST, but nevertheless related to it. These are however included in the activities of Sandvik, e. g. welded stainless tubes, OCTG tubes and specialized nuclear equipment. If the respective business volumes of DMV and Sandvik are compared in relation to the relevant market i. e. the production of SST within western Europe, the corporate sizes are indeed similar, which is reflected by their respective market shares in western Europe.

(61) In addition, even if measured at the level of total world-wide seamless stainless steel tube sales in 1992, the two groups are not significantly different. Sandvik sells slightly more volumes than DMV, what is mainly attributable to Sandvik's production in Canada and the United States. However, Sandvik's production in North America is solely intended for sale in that market. It would not be economically reasonable for Sandvik to ship any part of their North American production to western Europe, where it would be hit by import duties, especially as Sandvik has sufficient production facilities and spare SST capacity in western Europe. This is confirmed by the parties' own figures, according to which the total exports of SST from North America to western Europe (by Sandvik and all other North American producers) in 1992 was less than 0,5 KT.

D. Production capacity of DMV and Sandvik

(62) According to the parties' estimates, the capacity situation of DMV and Sandvik is asymmetric, since Sandvik's worldwide as well as European capacities are twice as large as those of DMV. This question requires careful analysis of the various levels at which capacities can be measured: in particular the theoretical throughput of the extrusion presses and the theoretical capacity for hot- and cold-finished tubes.

Maximum theoretical throughput of extrusion presses

(63) At the highest level, capacities can be calculated at the level of extruded hollows, as significant sunk costs are associated only with the extrusion press. According to the parties, there is a readily available market for second-hand finishing equipment. In principle, this would mean that a producer could react flexibly to an increase in demand and would only be constrained significantly by its extrusion capacity. Nevertheless, given the large amount of overcapacity in extrusion in Europe, it is unlikely that this capacity will be the determinant limit on expansion of production. It is more likely that given the underlying market conditions in the seamless stainless steel tube sector, which are generally unfavourable even in the medium term, any significant investment even in hot- or cold-finishing would require careful analysis of the expected revenus from increased production.

(64) However, even if the capacities of DMV and Sanvik are compared at the level of maximum theoretical capacity of extruded hollows, DMV will have excess capacity corresponding to approximately one third of the overall western European demand for hot- and cold-finished products, whereas Sandvik will have excess capacity in western Europe corresponding approximately to half of the overall western European demand for hot- and cold-finished products. At such levels of spare capacity, even if there are differences in absolute terms, neither Sandvik nor DMV would have a significant incentive to increase its output since they know that the increase would be at the respective disadvantage of the other and that the other one will always have sufficient capacity to respond.

Maximum theoretical capacity in hot- and cold-finishing

(65) DMV's maximum theoretical capacity, at European and world level, is estimated to be [...] KT in hot-finishing. For cold-finishing, the parties have indicated that DMV's maximum theoretical capacity will be [...] KT (including [ . . .] KT in the United States). Sandvik's worldwide capacities, of which most are situated in Western Europe, are above DMV's figures.

Even at the level of maximum capacity in hot- and cold-finishing, DMV will have excess capacity sufficient to cover between one-third and half of the overall western European demand for hot- finished products, and roughly one-fourth of the demand for cold-finished products. Sandvik will for its part have excess capacity in western Europe sufficient to cover about two-thirds of the overall western European demand for hot-finished products, and one-fourth of the demand for cold-finished products.

It is worth mentioning that both DMV's and Sandvik's maximum capacity reserves for hot-finished tubes individually exceeded the total sales of hot-finished products in 1992 by the remaining western European competitors, i. e. Tubacex and SBER. For cold-finished tubes both DMV and Sandvik each have maximum reserve capacities exceeding half of the total sales in 1992 of Tubacex and SBER. DMV and Sandvik would therefore be in a position enabling them to successfully retaliate if Tubacex and/or SBER increased their output in the western European market.

Capacity at current levels of employment

(66) If the respective capacities of DMV and Sandvik are compared at the level of manned capacity, the capacity situations are as follows: at the extrusion/unfinished products level, DMV will have a western European and worldwide capacity of [ . . .] KT, whereas Sandvik will have a higher capacity in western Europe and even more if their newly started joint venture in the United Kingdom is included. The differences are smaller in hot-finishing, where DMV will have [...] KT capacity in western Europe and worldwide. DMV is therefore below Sandvik's manned capacities in hot-finishing, measured both in western Europe and at world level. In cold-finishing, DMV will have [ . . .] KT capacity in western Europe, and a further some [...] KT in the United States. Sandvik currently has less manned capacity in cold-finishing in western Europe than DMV, but is likely to acquire Chomutov's cold-finishing facilities. (In addition, Sandvik also has a small volume of cold-finishing in the United States).

Conclusion on capacities

(67) While the respective capacity situations of Sandvik and DMV would be different, each producer has substantial capacity reserves on whatever definition to respond in a significant way to aggressive sales policies of any other producer on the market. In view of the stagnant demand and low-price elasticity of demand for SST, the fact that the two firms dominating the western European market do not have the same reserve capacities in absolute terms would therefore not weaken the incentive for Sandvik and DMV to engage in anti-competitive parallel behaviour on the western European market.

E. Cost situation of DMV and Sandvik

(68) According to the parties, the respective cost situation of DMV and Sandvik are different. The main difference is that Sandvik is vertically integrated (up- and downstream), while DMV is not. The parties indicate that the two companies would therefore have different cost structures, and hence different incentives to increase or decrease production in relation to market changes.

However, for one of the two producers to have an interest in breaking out of any parallel behaviour, there would have to be significant cost differences for one of them to be able to expect higher profits from competing. The higher the market shares of the two producers are, the higher the expected profits from their tacit collusion would be, and thus the higher the cost differences would have to be to offset the incentive to enter into anti-competitive parallel behaviour.

(69) The Commission made an attempt to assess the situations of DMV and Sandvik with reference to average costs. However, it appeared that factors like differences in accountancy standards, product mixes and methods of allocation of costs inherently limited such an exercise. The Commission therefore focused on the stated structural differences between DMV and Sandvik in order to examine the extent to which these differences are likely to influence significantly the marginal cost and indeed the future behaviour of the two market leaders.

It appeared upon close examination that these structural differences apply to only part of Sandvik's and DMV's activities. Indeed, Sandvik's UK subsidiary Sterling Tubes buys part of its prematerials from outside the Sandvik group, and distributes its tubes independently through independent stockists. Only a limited part of Sandvik's steel output in fact goes to SST production. Regarding the DMV companies, they purchase a significant portion of their raw material within their respective groups, and distribute [...] of their tubes through group stockists.

The structural differences between the two groups will also influence the amount of their fixed costs and, consequently, their marginal costs. In general however, production of SST as well as production of stainless steel through electric steel production technology involves a relatively low proportion of fixed costs. In fact, most of the costs involved in the production of stainless steel are variable costs (scrap, electricity) in contrast to those incurred in the production of carbon steel. Finally, marginal costs are not likely to decline steeply enough to constitute a strong incentive to break out of tacit anti-competitive parallel behaviour; this is all the more true when demand is price-inelastic and stagnant, and the two producers both have enough spare capacity to flood the market.

Vertical upstream integration

(70) The degree of vertical upstream integration is a factor which may influence the cost situation and the degree of fixed costs of a firm. According to the parties, this should be true with respect to Sandvik, given that this company has its own steel shop that produces the prematerial (bars and billets) for its SST.

This was confirmed by Tubacex which stated that part of its favourable cost position was due to the fact that its wholly owned steel shop is fully dedicated to the needs of its tube shop.

However, Sandvik's steel shop division is not dedicated to the needs of its SST division, the steel shop supplies numerous customers inside and outside the Sandvik's group and it is run as an independent profit centre, as is the tubes division. This is reflected in particular in the fact that the purchasing policy of Sandvik's UK subsidiary Sterling Tubes, with the exception of certain steel grades which have to be supplied from Sweden, is not restricted to internal sources and indeed varies according to market conditions and currency fluctuations. Sandvik is planning to adopt the same policy in terms of prematerial as for Sterling Tubes for the new extension press in the United States.

(71) According to the parties, DMV will buy its prematerial on the open market, from steel producers. However, the DMV ultimate parents, or companies with which they have capital links, are also important players in the upstream market (see above paragraph 13). Furthermore, DMV will benefit from greater purchasing power than the three companies had separately prior to the merger (although the parties have not foreseen any change in costs for prematerial in their feasibility study).

There is no reason to believe that, under normal competitive market conditions on the market for prematerial, non-integrated producers like DMV would suffer a cost disadvantage compared to vertically integrated firms as far as raw material is concerned. This is true in particular in the steel industry as the problem of overcapacities is also present at the level of billets and bars production.

This is further confirmed by the fact that a close examination of Sandvik's internal transfer price for bars and billets belonging to the most commonly used steel grades shows that the impact of vertical upstream integration on the cost of Sandvik's tube business is very low.

(72) According to the parties, the fact that Sandvik is integrated upstream should mean that it has a higher degree of fixed costs than DMV, which would buy its prematerial on the market, thus the total cost for prematerial would be variable for DMV. In addition, the parties consider that Sandvik will have higher fixed costs due to the fact that it has a central storage in Sweden and that it operates four extrusion presses for western Europe, as opposed to DMV which will only operate two extrusion presses. The parties conclude that Sandvik, particularly in a poor market situation and in view of its higher degree of fixed costs, will continue to book orders when DMV is below its variable (prematerial) costs in order to have the steel mill and its extrusion presses as fully loaded as possible.

The extrusion presses of Sandvik on the western European market, as well as those of DMV, are largely written off, and should therefore not generate a significantly higher degree of fixed costs for Sandvik than for DMV. Given that the Sandvik presses are specialized in producing different ranges of sizes, Sandvik will have less costs than DMV for changing the container that decides the size of the extruded hollow. Moreover, Sandvik did not state a different average fixed costs ratio than the other producers who answered the Commission's investigation.

Finally, the substantial revenue effect of anti-competitive parallel behaviour on prices between two companies achieving significant shares is unlikely to be compensated by the cost effect of better loading of Sandvik's steel shop, which is only partly dependent on tube manufacturing.

(73) There is no indication in Sandvik's previous performance on the SST market that would suggest that it would be willing to sell at low prices in order to increase its volume, be it for higher fixed costs or other reasons. On the contrary, Sandvik is widely recognized in the SST industry as a traditionally high-priced producer. In addition, it can be deduced from figures provided by the parties that Sandvik's market share in western Europe has been decreasing constantly between 1988 and 1992. In so far as market prices also decreased during this period, the fall in volumes sold should be interpreted as meaning that Sandvik did not follow an aggressive pricing behaviour resulting from a cost advantage. (During this same period, the market share of the notifying parties remained fairly stable).

In conclusion, there is no evidence under current and foreseeable market conditions that being vertically integrated upstream is either an advantage or a disadvantage. Nor are there any indications that Sandvik, due to its upstream integration, would have an incentive to adopt a more competitive policy thereby disturbing the possibilities of engaging in anti-competitive parallel behaviour with DMV.

Vertical downstream integration

(74) The degree of downstream vertical integration is a further factor which may influence the cost situations of DMV and Sandvik given that the latter has a network of subsidiaries in different countries that act as its own stockists, while the former does not. However, Sandvik's UK subsidiary Sterling Tubes sells mainly through independent stockists.

Within the respective groups of the DMV parent companies, or companies with which they have capital links, there are several companies which are also important players in the downstream market (namely TAD, Starval, Mannesmann Handel, Stappert Spezial Stahl and Thyssen). Furthermore, DMV will be less sensitive to purchasing power on the downstream market than the three companies had been separately prior to the merger.

(75) According to the parties, the fact that Sandvik is integrated downstream should mean that it has a higher degree of fixed costs than DMV. However, the Commission's investigation has not been able to obtain evidence that downstream vertical integration is either an advantage or a disadvantage. Nor are there any indications that Sandvik, due to its downstream integration, would have an incentive to adopt a more aggressive price policy (4) thereby disturbing the possibilities of engaging in anti-competitive parallel behaviour with DMV.

Labour costs

(76) According to the parties, Sandvik will have an advantage over DMV in terms of labour costs, as it could shift more of its production to the United Kingdom and in particular to the Czech Republic following the acquisition of Chomutov. However, the labour costs in the countries where DMV will be active also differ substantially. If a comparison is made between the figures of labour costs with those of Sandvik in different countries (all figures provided by the parties), the results are similar. If the parties are able to reach projected savings from the planned laying-off of labour following the merger, it would probably put them slightly ahead of Sandvik even after the acquisition of Chomutov.

(77) As labour represents on average around one fifth of total costs for production of seamless stainless steel tubes according to the data provided by the parties and their western European competitors, such a close labour cost situation could not in itself trigger overall cost differences sufficiently significant to create an incentive for either company to break out of the anti-competitive parallel behaviour.

Cost savings through innovation

(78) A possible means for a producer to gain an advantage over its competitors would be through innovation. However, the characteristics of the industry are such that innovation has only a relatively minor role to play, given that 80 to 90 % of the SST are commodity products subject to widely accepted international specifications. The amount of resources allocated to R& D in this industry is rather low, and neither of the two market leaders would be able to gain any significant competitive advantage even if resources currently invested in R& D were to be increased.

Conclusion on costs

(79) To conclude, the groups' structures do not indicate that one of them would enjoy a sufficiently significant structural cost advantage to offset against the substantial gains that would result from engaging in anti-competitive parallel behaviour.

F. Market transparency

(80) In order to assess whether anti-competitive parallel behaviour would be feasible between the DMV and Sandvik groups, the Commission has examined the degree of transparency in the market for seamless stainless steel tubes.

The result of this examination is that the degree of transparency would be sufficient for DMV and Sandvik to find out rapidly if the other one was not following the tacitly agreed pricing behaviour. The transparency of the SST market basically results from the following factors: the concentrated supply structure, the commodity character of the products, the low rate of innovation, and the fact that all the producers have structured their price list around a single standard product. In addition, it is important to note that the demand side of the market is widely dispersed and that customers tend to make frequent purchases concerning relatively small volumes. DMV and Sandvik would, either directly or indirectly (through their group stockists), meet the same customers on all segments of the market. The feedback from customers on prices would therefore be both rapid and highly reliable. These factors are examined in detail below.

Price lists

(81) The results of the inquiry launched by the Commission show that no price lists are released to customers for international equipment bids, where prices are quoted on an individual basis. However, regarding the distribution to stockists within western Europe, all European competitors release price lists to their customers.

In particular, Vallourec, Mannesmann, Dalmine, Sterling Tubes (which belongs to the Sandvik group) and SBER periodically publish a price list for the whole western European market. Tubacex only published its price list for the first time in 1992, and Sandvik publishes price lists in France, Italy, Germany and Sweden.

(82) Close examination of these price lists reveals that a distinction can be made between two categories of price lists corresponding to two groups of competitors depending on the distribution channel used.

On the one hand, most producers (namely Dalmine, Vallourec, Mannesmann, Sterling Tubes and Tubacex) distribute their products to a large extent through independent stockists. These producers launch a production run for each order, they hold no stocks and are therefore unable to react very rapidly or to deliver small quantities.

On the other hand, two producers (namely Sandvik and SBER) have established central storage and a distribution system (Sandvik through subsidiaries and SBER through exclusive dealers). For these reasons, these competitors are able to provide small quantities very rapidly; they achieve a lower proportion of their turnover with independent stockists than those competitors falling in the first group as described above.

(83) It results from the above differences that the competitors who belong to the first group publish ex-mill price lists targeted at the independent stockists whereas the competitors who belong to the second group publish ex-stock price lists indicating higher prices and none of the extras and rebates typically associated with ex-mill deliveries.

However, all price lists (including to a lesser degree those of Sandvik and SBER) are based on a common principle (described below). Moreover, the price lists of all competitors within each of the two categories are highly similar.

(84) A comparison of Sandvik and SBER's ex-stock price lists shows that both price lists indicate the levels of prices corresponding to the same two groups of steel grade. Neither indicates extras or rebates.

(85) The ex-mill price lists have a more comprehensive structure. All price lists include the different tubes (according to their size, diameter and thickness) and indicate their prices for the cheapest most commonly sold stainless steel grade (i.e. the ASTM TP304 steel grade). The TP304 price is used as a pricing reference, and prices for other grades of steel are quoted as a percentage of that of TP304. Each customer can then calculate the price for each specific grade by multiplying the TP304 price for a specific product by the given premium which reflects the cost difference with the basic steel grade. In addition, most competitors indicate in their price lists extra price factors referring to the same length, tolerances and certification. Three out of five lists indicate rebates for similar quantities. As a result, the structure of the different price lists for the first category of competitors is very similar.

(86) The high degree of similarity among the different competitors' price lists is further confirmed by a comparison of the scales used for some of the most important extras and rebates mentioned in these price lists, such as extras for specified steel grades, length, certification, tolerances or quantity rebates.

All European competitors use identical or almost identical scales which indicates a high transparency of the price structure in the market for SST. This is true with regard to grade scales, but also with regard to tolerance scales, to length-cutting scales and to quantity rebates.

(87) The market transparency is further demonstrated by the similar publishing dates of the various competitors' price lists during the abovementioned period (from end of 1988 to beginning of 1992). The analysis of these publishing dates demonstrates that all competitors publish price lists within the same month, with the exception of early 1992 when the first list was published two months before the last one.

(88) The parties have indicated that these similarities were due to the fact that all competitors in the market have decided to structure their price list according to the Mannesmann price list. The parties' main competitors and customers have confirmed this point to the Commission.

In particular, one competitor stated that during the 1980's all the competitors changed their price list and adopted the structure of the Mannesmann price list. This change of structure led to such similar price lists that some customers and one competitor referred to it as 'the European price list' in their answers to the Commission.

It was further indicated that the name of 'European price list' itself does not refer to a single price list but refers to the fact that, given the high similarity in the structure of the different price lists, it is common practice in the seamless stainless steel tubes business to quote prices in a given Member State according to the price list of the market leader. This behaviour of competitors regarding pricing further strengthens the effects of price lists on market transparency.

'Secret rebates'

(89) The parties have communicated to the Commission estimates of the overall average 'secret rebates' granted during the abovementioned period as well as estimates of the rebates granted to the main stockists. A comparison of the average rebates might lead to a wrong conclusion with regard to similarity of pricing policy because the level of these rebates is closely linked to the structure of the sales portfolio of the given supplier. Therefore, even though the average rebates granted by the three parties are similar, this can be considered as no more than an indication that pricing policies are similar.

(90) A comparison of the rebates granted to the largest stockists is consequently more meaningful.

The average rebates granted to the largest stockists are very high and increased each year between 1988 and 1992. Although a high rate of rebates does not facilitate market transparency, it should be noted that, in the present case, the rebates granted by all competitors are similar, and that the difference between the lowest rate (Mannesmann) and the highest rate (Vallourec) has always been constant except for a brief period between 1989 and 1990. In particular, it should be mentioned that even though no price list was published during the 16-months period between the 1990 and 1992 price lists, the average 'secret rebate' rate granted to the largest stockists for the overall period was of [...] for Dalmine, [...] for Mannesmann and [...] (5) for Vallourec.

(91) In order to assess whether such similarity indicates a degree of price transparency sufficient to allow a mutual monitoring of parallel behaviour, the Commission took into account the following factors which should have led to differences amongst the 'secret rebate' rates of the parties.

Given the abovementioned price behaviour in the market (i.e. each competitor determines its rebate in a given Member State on the basis of the home-company price list), the abovementioned average level of rebates is expressed as a percentage of prices (in the price lists) which are not used as a reference when selling abroad. For this reason, this average rate is influenced by the relative level of price in each country, the diverging evolution of currencies from country to country and by the structure of each company's sales portfolio geographic-wise.

In addition, it must also be taken into account that the average rebates provided by the parties are related to the period of time starting from the publication of one price list to next publication. The average rate is therefore also influenced by the structure of each company's sales portfolio time-wise (i.e. a given large order will not influence the average in the same way depending on whether it occurs at the beginning or the end of the period).

Taking into account all these elements, the Commission comes to the conclusion that the difference in the average rebate rates granted to the largest stockists does not contradict the existence of a high degree of transparency for prices.

(92) In a commodity market in which a few suppliers have commercial contacts with numerous customers, the feedback from customers on prices is very rapid and highly reliable, and there is little scope for price discrimination. This is true in particular, as the stockists have greatly reduced their stocks in the last years. They now generally purchase smaller quantities more frequently than in the past, in principle with orders once or twice a month. This provides the suppliers with a huge number of requests for price quotes per month (Vallourec referred to some [several hundreds] requests a month). Such a high frequency of contacts with numerous customers greatly increases market transparency. In addition, the parties would have the possibility to further verify the reliability of such information through their group stockists.

Subsequent to the merger, the number of suppliers would be reduced; hence DMV and the Sandvik group companies will be confronted with more frequent and thus more accurate information on the prices of competitors, what they would be even more likely than in the past to be able to identify.

(93) This is further confirmed by the competitors' answers to the Commission's requests for information.

None of the competitors who answered the Commission's requests for information have confirmed the parties' opinion that the SST market is not price-transparent; in fact, all confirmed that the SST market was characterized by a high degree of price transparency.

For example, one competitor stated that its firm 'is in constant contact with all categories of clients and thereby receives feedback from them on changed price lists (within hours) and discounts, i.e. actually applied prices (less than one week)'. The same competitor declared that a 'normal' secret rebate rate on the market would be [...] - which is very close to the average rates indicated by the parties (see above). Another competitor indicated to the Commission that 'everybody knows the average actual price level of his competitors through contacts with both competitors and clients'.

The overwhelming majority of the competitors and large customers who answered the Commission's inquiry indicated that the reaction of the other competitors to a reduction of price by one supplier would occur within a very short period of time (at most three weeks). In particular, most of the competitors of the parties stated that they would react within one to two weeks following the initial price modification.

Conclusion on market transparency

(94) In light of the above, and given that the market for seamless stainless steel tubes is a mature market with a low rate of innovation, where the products sold are to a large extent (80 to 90 %) commodity products belonging to only six different grades of steel, the price lists' similarities as described earlier and the close publishing dates allow for a very rapid comparison among the various competitors. This pricing behaviour, as well as the fact that the main suppliers have commercial contacts on a regular basis with the same customers, further strengthens this transparency and enables each competitor to have a good knowledge of the other competitors' prices.

(95) The degree of transparency would be sufficient for each of the two principal producers to find out rapidly if the other one was not following the tacitly agreed pricing behaviour by decreasing prices so as to put greater volumes onto the market. It is important to note in this context that stockists tend to make frequent purchases during the year, and that transactions usually concern relatively small volumes. In this sense, the functioning of the market offers the necessary conditions for competitors to monitor the behaviour of other competitors, and does not offer the possibility of making important 'one-off' gains through the conclusion of high-volume long-term contracts.

G. Actual and potential competition

(96) The Commission must also evaluate, as a second step, the extent to which the two leaders' behaviour would be constrained by other competitors and/or by the customers.

Actual competition

(97) Two of the remaining competitors are located in the geographic reference market (namely Tubacex and Schoeller-Bleckmann) while the others are Japanese competitors.

Western European producers

(98) With respect to competitors located within western Europe, the fact that western European demand for SST is stagnant, while structural overcapacities prevail, means that none of the remaining producers primarily active in western Europe can realistically hope to increase their sales volume without exposing themselves to likely joint retaliation by the two leaders. In fact, any increase of the sales volume by one firm would necessarily be at the expense of the sales of the main competitors.

Further, DMV and Sandvik would each have sufficient reserves to punish any attempts by the smaller competitors, to gain market shares.

(99) Tubacex is a Spanish company which only recently focused (since 1987) its activity on the SST market. Tubacex's activities are almost exclusively concentrated on the SST market and the company does not, unlike DMV and Sandvik, belong to a larger group of companies. According to the parties, Tubacex holds around 13 % of the market (the Commission's investigation in fact reveals a slightly lower share). Tubacex is the leader in its home market, the Iberian peninsula (a relatively low volume area), whereas it has some difficulties in imposing its presence in the main western European areas of consumption (i.e. Germany, France, Italy and the United Kingdom) due to the presence of domestic producers. The Commission has no indication whether Tubacex will be able or not to maintain its strong position in Spain in the new market structure.

Tubacex is the only supplier in the market which uses a specific production method allowing quicker extrusion of the billets and the use of square billets instead of round billets. Such a process enables Tubacex's press to achieve larger theoretical extrusion capacity than competitors' presses. In addition, this partly explains why Tubacex achieves lower costs when using square billets. On the other hand, this generates certain quality problems, in particular regarding the hot-finished tubes' eccentricity. Even though from a technical point of view this only creates significant problems for specific applications (higher quality products), given the high average quality standard of the market, such problems damage Tubacex's position as a quality supplier and have led it to compete mainly at the lower end of the market. This is already confirmed by Tubacex's weak position in the hot-finished tube business and by the fact that it does not currently offer the full range of steel grades. This was further confirmed by the declarations of both Sandvik and some stockists to the Commission.

Due to its specific manufacturing process, Tubacex has to offer cold-finished tubes in ranges where its competitors are offering hot-finished material. In such ranges Tubacex does not enjoy any cost advantage. Its current use of actual and theoretical capacity in cold-finishing is unusually high. To significantly increase its capacity in cold-finishing, Tubacex would therefore not only have to employ more labour, but would also have to make further investments in facilities for cold-finished production.

(100) Tubacex has just overcome a very difficult financial situation and has concluded an agreement with its suppliers with a view to carrying out important internal restructuring. Although Tubacex focused on the SST business only since 1987, it represents an important competitor, having played until now the role of the 'disturbing competitor' in the western European market. Tubacex has been selling at generally lower prices than the other producers in order to penetrate the market and thus acquire its current market share. Recently, as prices have fallen, Tubacex's prices have come closer to the average market prices.

(101) However, the investigation of the Commission showed a change in Tubacex's strategy since 1992. As of this date, Tubacex started publishing price lists. The already published price list (published mid-1992) has a similar structure to those published by the other competitors in the market. In the meantime, Tubacex concentrated on certain export markets outside of western Europe which are growing faster, where domestic producers are weak and hence risks are lower. This is reflected in the fact that Tubacex's market share at world level is relatively high, if compared with its western European market share.

(102) The parties acknowledged in their written answer to the Commission's statement of objections that 'Tubacex has on purpose designed its manufacturing route with an initial cost-saving on raw material and hot-finishing partly counterbalanced by an additional cold-pilgering step which solves the so-called eccentricity problems'. However, they argue that 'Tubacex has been able to set up additional cold-pilgering facilities by buying second-hand pilger machines which are used in the final manufacturing step of tubes' and such tubes are sold as 'hot-finished tubes'.

(103) The parties consider that Tubacex would have the possibility to overcome its eccentricity problems by buying additional second-hand pilger machines. Tubacex would therefore be able to use part of its spare capacity in the field of hot finishing to increase its market share should DMV and Sandvik increase their prices. In the context of long-term stagnant demand and low-price elasticity of demand, it is however doubtful whether an increase of output on the western European market would be economically rational behaviour for Tubacex. Tubacex has a relatively low degree of fixed costs. In addition, Tubacex already covers its total costs at the current level of production and prices, and is therefore not bound to increase its output.

Furthermore, Tubacex would have very little possibility of pursuing such a strategy without making the abovementioned new investments in pilger machines. Such a decision is normally only taken when a long-term increase in demand can be expected, and a supplier has reasonable prospects of recovering its investments through a durable increase in prices and thus profits.

In this respect, Tubacex would have little economic incentive to pursue such a competitive strategy, as its possible gains would have to be balanced against the risk of a price war. This risk is particularly high when the supply side is dominated by two players achieving an overall market share of 70 %, and where the demand side is characterized by a long-term stagnating demand and low-price elasticity. Thus the further decrease of revenues which would be the inevitable result of such action can be expected to be particularly damaging for a company that has recently gone through serious financial difficulties.

(104) Tubacex has stated that it would follow any price increase imposed on the market by DMV and Sandvik. Such a statement is not only consistent with the reasoning developed above, but it is also in line with Tubacex's policy of the last years. Tubacex's prices moved closer and closer to the other competitors' prices when Tubacex began publishing its price lists. In the meantime, Tubacex's high rate of exports out of the western European market shows that this firm targeted its activity on other markets outside western Europe where it was possible to use its cost advantage without the danger of a counter-reaction from other western European competitors.

Finally, whereas market operators, including the two leaders, usually sell two-thirds of their production to stockists and one-third to end-users, Tubacex focused its sales differently and it achieves [. . .] of its business as direct sales to end users To achieve a significant increase in its market share, Tubacex would have to be more active with regard to stockists where price transparency is higher and would be further strengthened following the proposed operation. Further it would, to a higher extent, have to produce tubes in the sizes and steel grades where it does not enjoy the same cost advantage.

(105) Schoeller-Bleckmann, (SBER), is an Austrian State-owned company which, according to the parties, holds around 11 % of the European market. SBER is in a difficult financial situation and the fact that it is a publicly-owned entity cannot be considered to guarantee that the Austrian government will be prepared to cover any level of losses that the company might have.

(106) During the last five years, SBER has gone through a restructuring process which has led to a reduction of around 40 % in the number of employees. This is reflected by the fact that its use of actual capacity is already very high for both hot- and cold-finishing. It would therefore only have a limited possibility of increasing its production to constrain the principal producers' behaviour.

(107) In addition, SBER has one of the least favourable cost positions with respect to the other western European competitors. It also has a relatively high fixed cost ratio and is registering heavy losses. The parties have stated that such a situation should in principle lead SBER to maximizing its output in order to contribute as much as possible to covering its fixed costs. However, SBER's break-even point is, at current price levels, quite close to its maximum theoretical capacity. It could be argued that SBER might increase its prices, but still try to undercut a price increase imposed by DMV and Sandvik, in order to gain in market share. However, in a market where the global demand is price-inelastic and expected to be stagnant, if SBER started to increase production, according to the parties, this would then quickly and severely depress prices.

In addition, in such a scenario, the other competitors and in particular the two leaders might retaliate and further depress prices, which SBER could not support given its weak financial situation as well as its low capacity reserves compared to those of the two leading firms. Such a price decrease could even raise SBER's break-even point up to an unrealistic level and possibly even above its maximum theoretical capacity.

Conversely, in the event of a price increase, SBER would better cover its fixed costs, without taking any risk, if it followed the increase.

(108) This is confirmed by all the statements made by SBER with respect both to its cost situation as well as to its marketing strategy. In particular, SBER confirmed in a written statement that it would not have any other solution post-merger but to follow any price increase imposed on the market by DMV and Sandvik.

Japanese producers

(109) The Japanese seamless stainless steel tube industry is the second largest on a worldwide level (just below the western European industry). There are five major Japanese suppliers (Sumitomo, Kobe, Sanyo, Nippon Steel, and NKK). These producers enjoy an aggregate production capacity which is around one-third of the overall worldwide production capacity and their actual production in 1992 amounted to approximately 51 KT. They currently have a common share of the western European market which is below 10 %. However, it should be noted that the Japanese have had a stable although stronger presence in western Europe over the past (their overall joint-market share amounted to around 10 %).

(110) One of these suppliers, namely Sumitomo, is one of the two market leaders at worldwide level (its worldwide market share is around 20 %). Sumitomo is also the Japanese producer with the strongest presence on the western European market. Sumitomo, as well as all the other Japanese producers, is widely recognised as a high-quality producer and is able to supply the full range of seamless stainless steel tubes. In addition, Sumitomo has special strengths in the production of value-added, small-sized tubes and therefore competes, like Sandvik, in the higher end of the product market.

(111) The Japenese producers were affected by the same structural changes in the field of seamless stainless steel tubes as their western European competitors, namely by the substitution process from extrusion presses to rolling mills and by the collapse of the market in the eastern bloc (see above paragraph 38). They therefore suffer from a low rate of utilization of their capacities.

(112) As already mentioned, the Japanese competitors have a long-term established, significant presence in the western European market. However, none of the Japanese producers is primarily active in Europe. If DMV and Sandvik were to jointly increase the prices in this market by a small but significant amount, the Japanese competitors would therefore be well placed to use their large spare capacity to increase their sales in western Europe immediately, while, unlike Schoeller and Tubacex, they would not have to fear retaliation from the two market leaders. This is all the more true as import duties are to be progressively reduced as of 1 January 1995.

(113) Although the Japanese producers did not significantly enter the western European market in the early 1990's when prices were higher but demand was heavily decreasing, it should be expected that, in view of the future long-term stagnant demand, they would adopt different behaviour.

(114) Any price increase carried out in common by DMV and Sandvik is thus likely to provoke a significant entry onto the western European market of Japanese suppliers. Given the overall demand price-inelasticity, such entry would even with limited volumes quickly depress prices.

(115) It follows that the Japanese producers are likely to act as a sufficient constraint to prevent the creation of a joint dominant position of DMV and Sandvik in the western European market for seamless stainless steel tubes.

Potential competition

(116) The parties have mentioned as an additional constraining factor the high likelihood of a rapid entry of eastern European suppliers in the western European market.

(117) The investigation showed that SST producers are presently located in eastern European countries (Czech Republic, Romania, Ukraine and Russia). Valcovny Trub Chomutov is the best known of these companies. According to a competitor and as confirmed by the parties, Chomutov recently sold its extrusion press to China and is now active only in cold-finishing. Sandvik is currently negotiating the acquisition of the Czech company and a definite decision is expected shortly. It is highly likely that soon, Chomutov will no longer be an independent firm, but will belong to the Sandvik group. Chomutov can therefore not be taken into account in the assessment of potential competition from eastern Europe.

(118) So far, there is only a very limited presence on the western European market by a few eastern European producers. Most western European stockists therefore did not feel able to give an opinion on the precise time frame within which the eastern European industry would become competitive in the field of SST.

(119) Nevertheless, two stockists indicated that eastern European firms would be able to offer competitive quality within two years. Several other stockists, end users and competitors estimated that eastern European firms would not be competitive before two or more years.

(120) The eastern European suppliers are widely rcognized as holding modern production facilities, supplied in particular by the Mannesmann group within the last 10 years. The aggregated production capacity, at the level of extruded hollows, of the Russian, Ukrainian and Romanian producers amounts to 125 KT.

However, at the present time, they suffer from poorly organized and poorly managed production, lack of commercial skills and logistics. This explains their current low degree of penetration in the western European market.

(121) No structural reasons would prevent these producers from rapidly obtaining the necessary quality and certification (it takes six to eight months to obtain such certificates). This is further confirmed by the fact that most eastern producers rapidly obtained certificates with respect to high-quality carbon steel tubes. Moreover, two eastern producers already hold western European certificates in the SST area. One of these is the Czech producer Chomutov. The other one is a Russian producer, Pervouralski Novotrubnij, which has recently received a Tuev certificate for heat exchanger tubes.

(122) The parties have suggested that the market penetration by the eastern European producers would be accelerated as partnerships are concluded between such producers and western competitors. The Commission's investigation shows that at least two western companies actively seek for such partnerships at present.

(123) The Community has concluded agreements with the eastern European countries which govern among others the exchange of seamless stainless steel tubes. These agreements provide for duty-free ceilings and tariff quotas above which duties apply. The tariffs charged on imports from the Czech Republic and Romania are gradually being reduced and will be abolished in 1996.

(124) It follows from the above that the eastern European competitors hold the necessary means to be competitive rapidly on the western European market. Given the cost advantage that the eastern European producers enjoy, (in particular due to low labour costs), it is therefore very likely that the eastern European producers will have a more significant impact on the western European market in the near future.

H. Customers

(125) The structure of demand for SST is far less concentrated than the structure of supply. However, the Commission has to assess the countervailing power of customers differently depending on whether they are stockists or end users. Based on the information available to the Commission, about a third of the overall stainless steel tubes is distributed directly by the producers to the end users. The remainder is distributed through stockists, who are either independent or vertically integrated with the producers.

(126) None of the stockists individually have a market share of more than 1 % of total sales in western Europe; they will therefore not constitute a counterweight to highly concentrated supply. At the same time, the other typical clients of the tube manufacturers, the end users, do not enjoy any significant countervailing power, as they confirmed in their answers to the Commission, due to the fact that they purchase only occasionally and only relatively small quantities of tubes.

(c) General conclusion

(127) The proposed operation will increase the degree of concentration in an already highly concentrated market. DMV and Sandvik will each hold a high and relatively balanced market share in western Europe as well as in the most important country (i. e. Germany). The joint market share held by DMV and Sandvik will amount to approximately 70 %, with none of the existing competitors holding more than 13 %.

(128) Both companies would hold large capacity reserves and, although Sandvik is more integrated in prematerial production, stocking and distribution, it would not have sufficiently dissimilar ex-mill costs following the merger to offset against the possible gains to be made from engaging in anti-competitive parallel behaviour. This assessment is not modified in its substance by the probable take-over of Chomutov by Sandvik.

(129) The limited number of suppliers on the market and the degree of price transparency is sufficient to allow Sandvik and DMV to rapidly monitor conditions offered on the market and to engage in parallel behaviour. Moreover, given the general features of the market (commodity market, low rate of innovation, etc.), such market transparency would not allow fierce competition between the two companies, which would in fact lower the sales margins of both producers without improving their sales volumes. Their mutual interdependence thus creates a strong common interest and incentive to maximize profits by engaging in anti-competitive parallel behaviour.

(130) Neither Tubacex nor Schoeller-Bleckmann would have any real possiblity of, nor an interest in, significantly constraining such market behaviour by increasing their production and offering lower prices with a view to gaining market shares. Both companies are primarily active in the western European market and are in a difficult financial situation. They could therefore not withstand retaliation from the two market leaders.

(131) The same does not apply to the Japanese competitors. On the basis of the Commission's findings and analysis, any small but significant price increase is likely to provoke further entry by Japanese competitors. Such entry would be sufficient to constrain the behaviour of the two market leaders.

(132) Moreover, it is likely that eastern European competitors currently not active or only marginally active on the western European market will significantly enter this market in the near future, either alone or through partnership with western European firms.

(133) Consequently, there is an ongoing threat undermining the creation and stability of any possible anti-competitive parallel behaviour.

VI. OVERALL ASSESSMENT (134) For the reasons outlined above, the Commission has come to the conclusion that the proposed operation does not create or strengthen 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,

HAS ADOPTED THIS DECISION:

Article 1

The proposed concentrated between Mannesmann, Vallourec and Ilva is declared compatible with the common market.

Article 2

This Decision is addressed to:

Dalmine SpA, Piazza Caduti 6 luglio 1944, n. 1, I 24044 Dalmine (BG);

Mannesmannroehren-Werke AG, Mannesmannufer 3, D 40213 Duesseldorf

and

Valtubes SA, 130, rue de Silly, F 92100 Boulogne Bilancourt.

(1) OJ No L 395, 30. 12. 1989, p. 1. (Corrigendum: OJ No L 257, 21. 9. 1990, p. 13).

(2) OJ No C 111, 21. 4. 1994, p. 6.

(3) 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. All [...] replace, therefore, figures or information deleted for business secret reasons.

(4) For the likely effects of Sandvik's possible higher degree of fixed costs, see paragraph 73.

(5) Within five percentage points for all three parties.