CCE, 11 février 2004, n° M.3280

THE COMMISSION OF THE EUROPEAN COMMUNITIES,

Dear Sir/Madam,

Subject: Case No COMP/M.3280 - Air France/KLM Notification of 18.12.2003 pursuant to Article 4 of Council Regulation No 4064-89 [1]

1. On 18/12/2003, the Commission received a notification of a proposed concentration by which Société Air France (" Air France") acquires within the meaning of Article 3 (1) (b) of Council Regulation No 4064-89 as amended ("the Merger Regulation") control of the whole of Koninklijke Luchtvaart Maatschappij N.V. ("KLM") by way of a recommended exchange offer.

2. In the course of the proceedings, the parties to the concentration submitted undertakings designed to eliminate competition concerns identified by the Commission, in accordance with Article 6 (2) of the Merger Regulation. After examination of the notification and in the light of these modifications, the Commission has concluded that the operation falls within the scope of the Merger Regulation and does not raise serious doubts as to its compatibility with the Common Market and the EEA agreement.

I. THE PARTIES

3. Air France is a French-based full-service air carrier with significant international operations. The Air France group has three main activities: passenger airline transport, cargo transport and maintenance services. Air France operates a hub-and-spoke network with its principal hub for international operations at Paris Charles de Gaulle airport and its main domestic hub at Paris-Orly airport. It is also one of the founding members of the SkyTeam alliance, which members are, in addition to, Alitalia, Delta, CSA Czech Airlines, Korean Air and Aeromexico. Air France is currently controlled by the French State, which holds 54.4 % of the voting rights.

4. KLM is a Dutch-based full-service carrier operating worldwide. The KLM group has four main activities: passenger airline transport, cargo transport, maintenance services and the operation of charter and low-cost/low-fare scheduled services by its subsidiary Transavia. KLM operates a hub-and-spoke network with its principal hub at Schiphol airport. KLM has an alliance with Northwest Airlines covering principally operations on North Atlantic routes and related feeder routes. There is currently no majority shareholder in KLM, although the Dutch State currently holds 14 % of the voting rights.

II. THE OPERATION

5. Pursuant to a Framework Agreement signed on 16 October 2003, the notified operation will be implemented by way of a recommended exchange offer by Air France for all issued and listed securities of KLM.

III. CONCENTRATION

6. During an initial period of three years from completion of exchange offer, Air France will hold a 100 % economic interest and 49 % of the voting rights in KLM. After the initial period, Air France will hold 100 % of the economic interest and acquire 100 % of the voting rights in KLM. The Framework Agreement provides that already during the initial period, Air France will inter alia gain veto rights over KLM's business plans and budgets and the right to veto the remuneration and appointment and to dismiss senior KLM management. The operation therefore constitutes a concentration within the meaning of Article 3 (1) (b) of the Merger Regulation.

IV. COMMUNITY DIMENSION

7. The undertakings concerned have a combined aggregate world-wide turnover of more than EUR 5 billion [2] (Air France: EUR 12,687 million and KLM: EUR 6,485 million). Each of Air France and KLM have a Community-wide turnover in excess of EUR 250 million (Air France: EUR [...] million and KLM: EUR [...] million), but they do not achieve more than two-thirds of their aggregate Community-wide turnover within one and the same Member State. The notified operation therefore has a Community dimension.

V. RELEVANT MARKETS

8. In the notification, the parties to the concentration have identified the following relevant markets:

a) Scheduled air transport of passengers

b) Air cargo transport

c) Maintenance, repair and overhaul ("MRO").

A. Scheduled air transport of passengers

(i) Point of origin/point of destination (O&D) pairs vs. network market definition

9. When defining a relevant product market, the Commission first considers demand substitution. The competitive constraint arising from supply side substitutability is normally only considered in the market definition when it has an immediate and effective impact on the relevant product market [3]. In the air transport sector, the Commission has therefore found that the relevant product market for passenger air transport services should be defined on the basis of the "point of origin/point of destination" (O&D) pair approach [4]. According to this approach, every combination of a point of origin and a point of destination should be considered to be a separate market from the customer's viewpoint. [5] To establish whether an O&D pair forms a relevant product market, the Commission considers the different possibilities offered to consumers to travel between these two points.The Commission will not only consider the direct flights between the two airports concerned, but also other transport alternatives to the extent that they are substitutable to these direct flights. These alternatives may include direct flights between other airports which have overlapping catchment areas with the airports concerned at each end (airport substitution), indirect flights between the airports concerned, or other means of transport such as road, train or sea (inter-modal substitution). Whether one of those alternatives is substitutable to the direct route depends on a multiplicity of factors, such as the overall travel time, frequency of services and the price of the different alternatives. This can only be decided on a route-by-route basis. [6]The parties have accepted this approach. However, the parties also maintain that, from the supply side perspective, there are clear indications that airlines increasingly compete on a network-to-network basis.

10. During the market investigation some network competitors have contested the traditional O&D approach and have suggested that network competition should instead be considered. As network carriers operate a hub-and-spoke system and because of the increasing size and scope of airline alliances, such competitors submit that competition occurs on a network basis. In their view, the O&D approach fails to capture the nature and the extent of such competition. By following that approach the Commission would fail to appropriately assess all the competition concerns that the present transaction might raise, particularly in regards of the reinforcement of the SkyTeam alliance.

11. The reference to network competition represents a supply side perspective and follows the business model of network carriers rather than the customer needs. Arguably, the network approach is normally of little relevance to the individual consumer.If confronted with high prices due to a monopoly on a particular O&D pair, a passenger may find little comfort in the fact that airlines compete world-wide in the development of their respective networks. Moreover, network carriers only represent one, if important, part of the industry. The market investigation has shown that low cost point-to-point operators and many regional carriers instead tend to agree with the O&D approach.

12. The Commission therefore maintains that consumers wishing to travel from a point of origin to a specific point of destination will consider the various possibilities to travel to the point of destination. Hence, consumers will take into account the network aspects such as for example frequent flyer programs only to the extent that airlines or alliances serve the O&D pair between which they wish to travel.

13. Some competitors have further suggested that corporate customers are increasingly considering contracting with airlines on the basis of the geographic coverage of their network (including those of their partners). Unlike an individual consumer, such corporate customers therefore would follow rather the network than the O&D approach. Typically, corporate agreements provide for a general discount on prices when flying with an airline/alliance's network. The discount often depends on the customer's overall travel expenditure.

14. The market investigation has shown that most corporate customers deal with more than one company or alliance and that they do not conclude exclusive contracts. The reason seems to be that no single airline (or alliance) could yet offer a complete world-wide coverage. Thus, even if a large number of customers may have, from the corporations' point a view, a preference for a corporate contract their employees travelling needs are still point-to-point. This is consistent also with the Commission's view in the United Airlines/US Airways decision.

15. As most corporate customers conclude such contracts with several airlines/alliances, whenever flying on a specific O&D pair, among other elements, they will consider the prices charged on this particular route and the overall discount. The more advantageous the latter is and the better the carrier's network covers the customer's travel needs, the more likely it is that the corporate customer remains with one carrier/alliance. In that case the customer's choice is determined by network competition and not by competition on individual city pairs. In this regard, the market investigation has shown a mixed result. In their choice of an airline, some corporate customers give more importance to the fare level while others give priority to the overall discounts offered.

16. For these reasons it is concluded that demand substitution justifies the O&D approach in defining the relevant market. However, in the case of corporate customers it is recognised that demand is driven both by network effects as well as O&D considerations.

17. Supply side considerations instead play an important role when considering the competitive constraint which individual carriers may exert on a certain market. The hub-and-spoke system determines the network carriers' decision to operate (or not) a passenger air transport service on a particular O&D pair. Network airlines concentrate traffic into a specific hub and disperse passengers via connection to numerous spokes. This increases the load factor of aircrafts and allows airlines to exploit economy of density. They normally refrain from entering city pairs which are not connected to their respective hubs. At the same time the concentration on their hubs reinforces their position at these airports which often makes entry of competing airlines more difficult.

18. The competition between networks is also reflected in the fact that the Commission considers that for long-haul flights, generally speaking, indirect services actually compete with direct flights. Network effects are taken into account when considering the competitive strength exerted by the merging parties on O&D city pairs linked to their individual hubs, and, in general, when considering the barriers to entry stemming from the existence of such networks (see further below).

(ii) Time-sensitive vs non-time sensitive passengers

19. The Commission has found that passengers travelling on unrestricted tickets (time-sensitive passengers with a need for flexibility) may be in a different market from passengers with restricted tickets (non-time sensitive), which are more interested in the price than the frequency and accept longer journey times [7].Some competitors have suggested that this difference no longer reflect the requirements of passengers and the commercial policy of the airlines. However, customers and travel agents have confirmed the importance for certain travellers (mainly business passengers) to the possibility to modify their tickets and return as soon as possible to their point of origin, preferably within the day if a short-haul route (this requirement does not, obviously, exclude that these customers are also -more and more, as argued by many respondents - price-sensitive). Time-sensitive passengers have therefore different requirements than non-sensitive ones and normally will only choose to fly airlines offering a high number of frequencies in a given O&D pair and the possibility to use unrestricted tickets. Therefore, the Commission considers the existence of two distinct markets for time-sensitive and non time-sensitive passengers when assessing the proposed merger's impact on competition on the different routes.

(iii) Substitutability of indirect flight with direct flights on short-haul routes and long haul routes

20. In the past, the Commission has considered that on short-haul city pairs indirect flights do not provide a significant competitive constraint on direct services. However, this has to be examined on a case-by-case basis. In exceptional circumstances, indirect flights could exert a certain competitive constraint over direct flights. This is the case of some of the routes affected by the present transaction where a significant number of time-sensitive passengers even prefer the indirect service over the direct one. This is due to the inadequacy of the direct flight to meet the specific requirements of time sensitive passengers, namely a high number of frequencies allowing the completion of a roundtrip in a day.

21. In the United/US Airways decision, the Commission found that on long-haul flights, such as transatlantic services, "indirect flights are at a lower disadvantage than on short- and medium-haul services because intermediate stops have a lower relative impact on total elapsed time as the total trip duration increases". The Commission concluded that indirect flights may constitute a competitive alternative to non-stop services if they are marketed as connecting flights on the city pair on computer reservation systems, are operated on a daily basis and cause only a limited extension of the trip (150 minutes waiting time). In recent transatlantic alliance decisions [8], the Commission followed a similar approach and concluded that indirect flights, under certain conditions, appear to exert a sufficient competitive constraint on non-stop long haul services. The degree to which an indirect flight is substitutable should be considered on a route-by-route basis.

22. During the course of the market investigation some corporate customers indicated that only services with a connection time of a maximum of 120 minutes fully constitute competitive alternatives to direct non-stop flights. However, corporate customers are only one customer group and, according to the market investigation, probably the most time-sensitive one. Other customer groups can accept longer connection times. Moreover, other elements brought by the market investigation mitigate this suggestion since, on long haul routes, the dividing line between time sensitive passengers and non time sensitive passengers has become less clear. As was reported by several parties (including the merging parties themselves) an increasing number of time-sensitive passengers appear to have become more price sensitive. In particular an increasing number of undertakings require their employees to take the most economical flights, including indirect flights, irrespective of total flight duration or longer connecting times. It is therefore still considered that the availability of flights with a maximum connection time of 150 minutes exert sufficient competitive constraint on long-haul direct flights.

23. In light of the above, it is concluded that, on the long haul routes considered in this investigation, competitive indirect flights can be generally seen as a suitable alternative to non-stop services. Those competitive indirect flights are therefore included in the markets for the provision of scheduled air services for passengers on transatlantic city-pairs concerned.

(iv) Substitutability of airports in Paris, Milan, New-York

24. Airport substitutability has to be looked at both from the demand and the supply side. From the demand side, passengers who begin or end their journey in a catchment area of two or more airports can choose from and to which airport they wish to fly. This increases competition if the choice between airports also implies a wider choice between different airlines, to the extent that they are independent from each other (i.e. that they are not alliance partners or franchisees). To the extent that such areas exist, the share of customers on a certain route who actually live in an overlapping catchment area will have to be established. Only if the share of such passengers is sufficiently high, a carrier would take them into account when setting prices.

25. There is a correlation between the extra time a traveller is willing to spend by travelling to an airport which is farther away and the overall travelling time [9]. For flights within Europe, it can be assumed that the radius of an individual airport's catchment area is small, given the overall short travelling time, while for long-haul flights the catchment area is larger.

26. On the supply side, i.e. from the perspective of airlines providing the transport service, the substitutability of airports depends on their passengers' needs, but also on the kind of service they wish to provide, on the particular ground services provided by the airport and on the airline's existing activities. A network carrier which provides significant connecting and feeder services has different needs at the airport than a low-cost carrier providing mainly point-to-point services. Moreover, airlines have a strong preference in bundling all their activities at one airport due to fixed overhead and operation costs at the airport.

Paris airports

27. Paris is served by two main airports: Charles de Gaulle (CDG) and Orly (ORY), both of which are "fully coordinated", i.e. slot co-ordination is necessary since demand for slots exceeds supply at least at certain times during the day. The parties have argued that from the demand side, both airports are substitutable. They would be equally convenient for point-to-point traffic as they are located at approximately the same distance from the centre of Paris. While both airports could be used for destinations within the EU and to third countries, the parties point out that in practice most network carriers have focused their activities on CDG airport while airlines operating at Orly mainly focus on domestic traffic.

28. The market investigation has shown that many corporate customers do not consider both airports to be substitutable, since Orly offers fewer connections. On the supply side most network carriers consider the two airports to be substitutable. However, this is considered differently in particular by low-cost airlines. They emphasise the importance to have the choice of the airport with a view to minimise their unit costs [10].

29. Therefore, from the demand side for point-to-point traffic it is considered that both airports are substitutable as they are located in the same catchment area and have comparable access facilities. As regards transfer traffic, there exists a preference for CDG due to the number of connections offered. From the supply side, the substitutability of the airports depends on the individual carrier's needs. For some airlines they are not substitutable depending on the markets they serve (European or long-haul traffic) and due to different airport costs.

30. As the parties have not argued that Beauvais-Tille (BVA) [11] is a substitute for CDG / Orly for intra-European flights or that Brussels-Zaventem is a substitute for long-haul flights, this is not further investigated.

Milan airports

31. Milan is served by two main airports, Malpensa (MXP) and Linate (LIN) both of which are "fully co-ordinated". In addition, there exists a regional airport at Bergamo. The market investigation has shown that as it is closer to the city of Milan [12], time-sensitive passengers with a final destination in Milan generally prefer Linate. This, however, situation may be different for passengers travelling long-haul and in particular for passengers with connecting flights. Bergamo is generally considered an option only for non-time sensitive passengers.

32. From the point of view of low cost airlines providing point-to-point services, some carriers prefer Bergamo due to the cheaper handling fares and slots while others have a strong preference for Linate, due to its proximity to the city centre. Also in this case a large majority of airlines argued that the entrant should be allowed to choose the airport at Milan, as they consider Linate and Malpensa are not substitutable for every service.

33. Therefore, from the demand side for point-to-point traffic, taking into account both time-sensitive and non-time sensitive passengers, it is considered that Milan-Linate and Milan-Malpensa are substitutable. As regards transfer traffic, there exists a preference for Malpensa due to the number of connections offered. From the supply side, the substitutability of the airports depends on the individual carrier's needs. For some airlines they are not substitutable depending on the markets they serve (i.e. mainly local time-sensitive passengers or mainly transfer passengers) and due to different airport costs.

John Fitzgerald Kennedy (JFK) and Newark in New-York

34. The market investigation has shown that since JF Kennedy and Newark airports are at similar distance from downtown Manhattan and are both served by major international airlines flying non-stop to European cities, these airports are considered to be substitutable.

(v) Substitutability of alternative means of transport on short-haul routes

35. The parties, based on previous cases [13], claim that high speed train services may be substitutable for direct short haul flights, depending on travelling time, the frequency of the high speed train service and the price of the different alternatives. This can only be assessed for each individual city pair (see section VI).

B. Air cargo transport

36. The Commission has in previous decisions departed from the "point of origin / point of destination" (O&D) pair approach to delineate the relevant market in air cargo transport cases [14]. However, the Commission has found in previous cases that a wider market could be defined as, unlike passengers, cargo may be routed with a higher number of stop-overs and hence any indirect route is substitutable to any direct route. According to these decisions.

(1) For intra-European cargo transport, the relevant market could be defined as a European-wide and would include road transport and, to a limited extent, train. This was based on the finding that in the great majority, if not all, of intra-European air cargo prices are constrained by alternative modes of transport, notably road transport [15].

(2) For intercontinental routes, where at least one of the O&D pairs is located outside Europe, the corresponding catchment area broadly correspond to continents, at least within those continents where local infrastructure is adequate to allow for onward connections. This would be the case of the North Atlantic routes.

37. The parties argue that this market should not be further divided by category or nature of the products transported. In this regard, most of the replies to the market investigation agree with the parties that the air cargo transport is basically the same for all type of products and a further segmentation of the market would not be necessary. The special requirements of certain products relates mainly to activities ancillary to the cargo transport such as ground handling. Nevertheless, it appears that a limited number of products would require a specific type of carrier. For instance, certain oversized cargo and dangerous goods could not be transported by carriers transporting passengers, due to physical or legal constraints. Such services are therefore more likely provided by carriers operating full freight aircrafts. On the other hand, extremely time-sensitive products (such as perishables goods) would require a high number of frequencies and a better inter-connectivity, which normally only combination airlines (i.e. airlines combining the services of passengers and cargo) could guarantee.

38. However, it is not necessary to further delineate the relevant market as in all alternative market definitions considered, effective competition would not be significantly impeded in the EEA or any substantial part of that area.

C. Maintenance, repair and overhaul ("MRO")

39. In its decision in KLM/Alitalia, the Commission found that the MRO sector could be further divided into numerous product markets for the maintenance of each type of aircraft and equipment. The parties consider that the Commission's approach corresponds broadly with commercial reality and, in their view, the MRO market should be divided into four broad segments: (i) line maintenance; (ii) heavy maintenance; (iii) engine maintenance and (iv) component maintenance. The parties however do not agree with the Commission view that line and heavy maintenance could be further divided in A, B, C and D-checks [16]. They submit that this nomenclature has become old-fashioned and the boundaries between different types of checks have become blurred. Some competitors share these views; however, most of the replies to the market inquiry still agrees with the Commission and consider this sub-division acceptable.

40. The Commission found in the above decision that the relevant geographic market for MRO services is worldwide. The parties agree with the exception of line maintenance. For this segment, they argue that it is not necessary to take the aircraft out of service, which is the case with other type of maintenance services. Further, the service is carried out at the airports of origin or destination, or at an aircraft's operational base. The parties are therefore of the view that this segment should be considered local, confined to the airport at which such services are rendered. The market investigation has broadly confirmed this geographic market definition, although some replies suggested that the geographic market could be narrowed in the light of intercontinental transport costs. These replies were of the view that the heavy maintenance market (C- and D-checks) for narrow-body aircrafts (e.g. Boeing 737) and component maintenance should be considered regional rather than world-wide in scope.

41. The definition of the relevant markets can be left open as the operation does not raise competition concerns irrespective of the market definition considered as concluded below.

VI. COMPETITIVE ASSESSMENT

A. Scheduled air transport of passengers

42. The parties have identified (a) eight short-haul O&D city pairs between France and The Netherlands; (b) nine short-haul O&D city pairs between The Netherlands and Italy; and (c) 73 O&D long-haul city pairs on which both parties (and/or their partners) operate scheduled direct or indirect services.

Alliances and partnerships of Air France and KLM

The SkyTeam alliance and the KLM/NW Joint Venture

43. Air France is part of the Skyteam alliance, as indicated above. The framework of this alliance is set by the "Global Alliance Agreement" and the "Co-ordination Agreement", which provides the legal framework for further integration of their activities as a single operational entity on certain routes. These agreements are currently being investigated by the Commission [17]. Under this contractual framework, the Skyteam members have entered into separate bilateral co-operation agreements. On 18 January 2002, Air France, Delta, Alitalia and CSA Czech Airlines have been granted antitrust immunity by the US Department of Transportation (DoT) to fully co-operate on the transatlantic routes.

44. Air France has further entered in a joint venture with CSA Czech Airlines for the Paris-Prague route.

45. KLM has, on its side, concluded an "Enhanced Alliance Implementation Agreement" with Northwest to form the KLM/NW Joint Venture. This agreement establishes a transatlantic joint venture pursuant to which the two partners coordinate scheduling, marketing, sales and prices on all their transatlantic operations and on routes beyond their respective hubs. This agreement benefits from antitrust immunity by the DoT. After investigating this agreement under article 85 EC the Commission decided to close the proceeding without imposing remedies. [18]

46. KLM has entered into a joint venture with Kenya Airways on the Amsterdam-Nairobi route.

Co-operation of Air France, Alitalia and KLM on the Italian-Dutch routes

47. Air France and Alitalia concluded a bilateral co-operation agreement in the frame of Skyteamwhich the Commission is currently investigating [19]. The agreement between those two airlines focus on the Italy-France routes, although it provides the legal basis for a extensive use of free-flow code-sharing and co-ordination of schedules and policies on prices, on a world-wide basis. In its preliminary assessment, the Commission has only raised serious competition concerns on the French-Italian bundle of routes, since the parties' operations on other intra-European routes, though restricting competition on some of these, were considered unlikely to be significant enough to eliminate competition. The proposed transaction, however, changes the basis for the assessment of the Commission as regards the Italy-The Netherlands routes. In effect, by acquiring control over KLM, Air France will extend its footprint to the Dutch market. The proposed transaction will thus provide a change in incentives for the two parties to co-ordinate their behaviour also on these routes, which will have as an effect the elimination of competition between Alitalia and the merged entity on these routes.

48. The intention of Air France and Alitalia to co-ordinate their activities beyond the French-Italian bundle of routes is confirmed by the conclusion of i) a bilateral agreement between KLM and Alitalia for the Dutch-Italian routes and ii) a multilateral agreement between Air France, Alitalia and KLM, with the purpose of further integrating their activities, including the possibility for Alitalia to merge with the parties.The parties submit that the potential spill-over effect of the co-operation between Air France and Alitalia to the Dutch/Italian routes is too hypothetical to be considered under the present procedure. Accordingly they consider these agreements should be assessed under the scope of Article 81 of the EC Treaty,

49. The Commission sustains that the competition concerns related to the Italian-Dutch bundle of routes derive directly from the proposed transaction and should be assessed under the present proceeding.This is without prejudice of the possible assessment of such agreements under an Article 81 proceedings in which due account will be taken of the undertakings by the parties in this procedure.

Cooperation between KLM, Northwest and Continental

50. Some third parties argue that the Commission should not, when assessing the impact of the merger on the transatlantic routes, consider Continental as a competitor of the parties. This is based mainly on three elements: (i) the long history of cooperation between Continental and Northwest; (ii) the existence of a cooperation marketing alliance with KLM and a strategic alliance between Continental and Northwest, and (iii) Northwest's limited partnership in Continental giving the former the possibility to block business combinations and similar change of control transactions involving Continental and a third major air carrier. Further, third parties consider that the incentives for Continental to compete with the merged entity and its partners post-merger will be diminished even further, as Continental has already declared its intention to join the Skyteam alliance.

51. The parties argue that the existing cooperation between Continental, Northwest and KLM only provides for "non-immunised" forms of commercial cooperation and, therefore, the Commission should consider Continental as a true competitor of Air France and KLM. In addition, although it is true that Continental intends to join Skyteam during the current year, this cannot be anticipated as a fact and therefore the Commission should disregard this intention for the assessment of the above mentioned merger.

52. The parties, Continental and Northwest explained to the Commission that KLM/Continental, are parties to a code share agreement (October 2001) affecting journeys connecting via Amsterdam-Houston and Amsterdam-NYC (hub-to-hub routes) to markets behind the others' hub. This means that Continental can place its code on KLM flights behind Amsterdam to 68 cities in Europe, Africa and the Middle East; and KLM can place its code on Continental flights behind Houston and Newark to 30 cities in South America and 44 in North America; which allows them to extend their respective networks to destinations which they do not serve.

53. They further explain that since KLM, Northwest and Continental, today do not enjoy anti-trust immunity in the US, their cooperation will have to remain limited to code sharing and other joint commercial activities which will not have a negative impact on competition. Today, they do not operate under a joint venture arrangement, do not share revenues and expense, and do not coordinate scheduling and pricing. The behind and beyond code sharing arrangement, while allowing the carriers to expand their respective networks, also enhances competition on the relevant city pairs by allowing airlines to offer services on more routes. The incentive to compete between the marketing and the operator carriers results from the fact that the carriers get paid only when a customer flies on their individual flights (except for a nominal selling commission the marketing carrier receives to cover its selling costs). Therefore, the carriers continue to have an incentive to market and sell its own services over the other's services.

54. In June 2003, KLM and Continental agreed to deepen their cooperation on the behind and beyond routes by agreeing to extend their code share in a limited manner to the hub-to-hub routes. The objective was twofold: for Continental and KLM to improve connectivity and build itineraries for KLM to and from Latin America and for Continental to/from Europe, Middle East and Africa; [...]. This enhanced code share agreement has not been implemented [...]. In any event, KLM and Continental are not permitted to place its respective designator code on local hub-to-hub operations of the other carrier.

55. [...].

56. Consequently, the Commission considers that - from the above facts - the existing degree of commercial cooperation between KLM, Continental and Northwest allows it to conclude that these companies compete today and that they will maintain an incentive to compete on the transatlantic routes. Further, internal documents of KLM show that KLM and Continental, despite their cooperation arrangements, do compete on price, schedules and capacity on the transatlantic routes.

57. In addition to the existing cooperation between KLM/Northwest and Continental, some third parties have argued that Northwest holds some shareholding in Continental. The clarifications that were provided on this issue show Northwest owns no shares of Common Stock of Continental. It only holds one share of "Series B preferred stock" containing some limited voting rights intended to enable Northwest to block an acquisition of Continental by another major airline. However, Northwest has no direct or indirect membership or participation on the Board of Directors or other management committees of Continental. With this respect, the carriers have entered in a standstill agreement in November 2000 whereby Northwest is prevented from owning any shares of any class of capital stock of Continental which would allow Northwest to vote for the election of Continental's Directors; to affect or influence the Board of Directors or the control of management; or to seek itself any representation in the Board of Directors.

58. By way of background, this arrangement between Northwest and Continental was agreed in order to settle a law suit the US Department of Justice had filed against Northwest to force divestiture of Northwest stockholding out of concern Continental would not maintain its competitive independence.

59. It has also been argued that Continental has expressed its intention to join Skyteam and that it is already behaving in relation to corporate customers in the US as if it were already a member of Skyteam. The Commission's investigation has provided elements that allow it to conclude that Continental's admittance to Skyteam is not a fact and therefore it would not be appropriate to take it into consideration in the present investigation. If and when Continental enters in bilateral agreements with Air France and Alitalia in the framework of Skyteam, the relevant competition authorities will then have the possibility to review its competitive impact under the appropriate proceeding. Moreover, it cannot even be anticipated what will be the degree and intensity of cooperation these carriers will agree upon.

60. Further, the Commission queried whether Continental, Northwest and Delta have jointly participated in association with other Skyteam members or KLM/Northwest in launched bids for corporate contracts which covered the transatlantic routes. Continental provided a list of joint corporate contracts they have with Northwest in the framework of their US alliance. Some of these contracts also extend to KLM transatlantic flights; but since they do not benefit from anti-trust immunity in the US they build these contracts and the discounts awarded to the clients on their respective published fares (they cannot agree on a corporate fare). Continental has also explained that under their tripartite alliance with Delta and Northwest, they have the possibility, if requested, to offer joint corporate contracts with these carriers, but they are under a number of carve outs imposed by both the US Department of Transportation and the Department of Justice. They remain independent competitors and can only jointly market their services to a limited extent. [...].

61. It follows from above that the elements provided for during the investigation and the clarifications provided to the Commission by the parties, Continental and Northwest have removed possible serious doubts whether Continental is to be perceived today as an effective competitor in particular of KLM. The Commission further considers that the elements obtained in the present investigation do not show that KLM/Northwest and Continental's present cooperation provides either for reduction of actual competition between them, or for elimination of incentives for Continental to compete post-merger with the merged entity.

Conclusion

62. As a result of the above mentioned agreements, the competitive effects of the proposed transaction will not only consist in the elimination of competition between Air France and KLM, but also between their respective partners, Delta, Alitalia, CSA Czech Airlines, Northwest and Kenya Airways, on the routes falling under the scope of these agreements or directly affected by them as a consequence of the merger. Consequently, in line with the United/US Airways decision, the Commission has aggregated the market shares of each of the parties to those of their partners on the transatlantic routes, the Italian-Dutch bundle of routes and other routes such as Paris-Prague and Amsterdam-Nairobi.

63. For the reasons stated above Continental is considered a competitor of the parties on the transatlantic routes.

Scheduled air transport of passengers within the EEA

64. The merger eliminates competition between the parties world-wide. Within the EEA, it is considered that this affects potentially all city pairs between The Netherlands and France, as these are the parties' respective home countries. As regards markets between France or The Netherlands and any other EEA member states (i.e. Paris - Frankfurt) or between two other EEA member states (i.e. Vienna -London), there only exists routes with direct-indirect or indirect-indirect overlap between Air France and KLM.

65. As pointed out above, as there exist a bundle of bilateral and multilateral agreements between Air France, KLM and Alitalia and since both Air France and Alitalia are part of the SkyTeam alliance, the Commission considers that the merger eliminates the incentive for Air France /KLM and Alitalia to compete on city pairs between The Netherlands and Italy.

66. In what follows therefore every city pair between The Netherlands and France and between The Netherlands and Italy on which Air France, KLM and/or Alitalia operate a scheduled direct service is considered.

67. Further, the data used is MIDT for the twelve month period ending in March 2003 (the investigation period). Further, total traffic or overall traffic indicates the total number of passengers flying on that the relevant city pair including connecting passengers while O&D passengers means passengers flying point to point.

Franco-Dutch short haul routes

Overview

68. Between The Netherlands and France there exist two direct overlap routes, i.e. Paris-Amsterdam and Lyon-Amsterdam. These two city pairs are also the most important ones in terms of passengers. On both city pairs there are serious doubts that the transaction, as initially notified, might create a dominant position. Apart from this, there exist five non-overlap routes where one party provides direct and the other one indirect service. On three of them, there are similarly serious doubts that the proposed merger, as initially notified, might create or strengthen a dominant position.

Paris - Amsterdam

69. With more than [1] million air passengers a year (including transfer), Paris - Amsterdam is one of the busiest city pairs within the EU. In addition, the high speed train Thalys carries several hundred thousand passengers between Paris Gare du Nord and Amsterdam Centraal. On this city pair, which connects the respective hubs of Air France and KLM, the parties compete directly, each operating seven daily frequencies. For air travel, the parties have estimated their combined market share post-merger for O&D passengers to amount to [>90 %] for time sensitive passengers and [>90 %] for non-time sensitive passengers. The proposed merger will therefore combine the only two operators of direct flights on this market.

70. Apart from air transport, Thalys, a joint-venture between the French and Belgian national railways SNCF and SNCB, operates 6 daily frequencies between Paris and Amsterdam with a travelling time of 4h09 (one direction). This compares to a flight travelling time of about 3 hours [20] (city-centre to city-centre). As of 2007, it is expected that the new infrastructure of the High Speed Line between Brussels and Amsterdam will enable Thalys to reduce its overall travel time between Amsterdam and Paris to about 3 hours.

71. The investigation has shown that for the time being, Thalys can be considered an effective competitor for non-time sensitive passengers. If one takes into account both air and rail travel, on the basis of data provided by Thalys and the Parties for 2002, the combined market share of Air France /KLM for non-time sensitive O&D passengers is below [45-55 %] . The situation appears to be different for time-sensitive passengers. With regard to this passenger group, Thalys is handicapped due to a travelling time which is almost one hour longer in each direction and due to a smaller number of frequencies (6 daily frequencies against 14 operated by the parties).Even if one considered that all first class Thalys passengers are time sensitive, the Parties still would have a very high market share for O&D time-sensitive passengers. Most likely, however, a significant part of Thalys first class passengers are not very time sensitive as they can accept more than two hours additional travelling time (both directions) and less frequencies. For the time being, Thalys does not allow for a typical business return trip within the same day.

72. Also a price comparison shows that the competitive constraint of Thalys is limited to non-time sensitive passengers. While the fares of the airlines are similar to those of Thalys for economy class, non-restricted tickets of Air France and KLM are more than twice as expensive as Thalys' business class tickets. This conclusion is also supported by the market investigation, in particular by corporate customers and travel agencies, which do not consider Thalys to be a competitor for time-sensitive customers at this stage.However, this may change with the possible introduction by 2007 of a new high-speed link between Brussels and Amsterdam, which will result in a reduction in the total journey time of one hour, and if Thalys adds additional frequencies.

73. The market investigation has furthermore confirmed that other network carriers are not interested in operating on this city pair. As regards competition from low-cost and regional carriers, there exist a number of entry barriers.Competitors have argued that the parties' dominance at their respective hubs in Paris and Amsterdam make it particularly unattractive to enter this route. Airports in Paris and Amsterdam are slot constrained which, in particular in Paris, makes it very difficult for competitors to obtain sufficient slots to operate at a minimum efficient scale. Given the parties' large scale of operations at their respective hubs, in comparison to their competitors they benefit from economies of scale at both airports for instance regarding the overhead and operation costs, the need for a fixed level of standby crew cover and the increased leverage to negotiate pricing with third party service providers such as engineering, ground handling services and airport facilities etc. As a result, the merged entity would have a very strong position on this hub-to-hub route.

74. With regard to high yield (business) customers, the merged entity would be in a much stronger position than a potential entrant due to its high number of frequencies, its well-known brand, its FFP programme and the world-wide network within the SkyTeam. The latter is in particular attractive to corporate customers. Unlike a regional or low-cost airline which can only compete for O&D traffic, the merged entity benefits from economies of density due to the high volume of connecting passengers on this city pair [21]. Finally, as an additional hurdle on this route, any entrant would face competition from Thalys with regard to non-time sensitive passengers.

75. [...].

76. It is therefore concluded that due to the elimination of one direct competitor (KLM) and a number of market entry barriers linked to the parties' strong position on the market, there is a risk that without appropriate remedies allowing for effective entry, the proposed merger would create a dominant position on this city pair for time-sensitive customers.

Amsterdam/Lyon

77. Also Amsterdam-Lyon is a direct overlap city pair on which Air France's subsidiary Régional and KLM operated three daily frequencies between Lyon St. Exupéry and Amsterdam Schiphol in summer 2003. With [between 175,000 and 225,000] passengers a year [65,000-70,000] O&D it is a medium sized European city pair on which the parties' combined market share amounted to [>90 %] and [>90 %] for time-sensitive and non-time sensitive passengers, respectively. The parties have not suggested that other transport modes can be considered as a substitute transport service. On this city pairs, there are market entry barriers similar to those mentioned for Amsterdam-Paris.

78. Therefore, without appropriate remedies allowing for effective entry, there is the risk that the merger creates a dominant position on this city pair for time-sensitive and non-time sensitive customers.

Other routes between France and The Netherlands

79. Besides the Paris/Amsterdam and Lyon/Amsterdam city pairs, there are five other "regional" city pairs on the Franco-Dutch bundle which are affected by the proposed transaction. On the city pair Nice - Amsterdam, in summer 2003 KLM and its subsidiary Basiq Air operated 3 and 2 daily frequencies, respectively. They compete with easyJet which operates two daily frequencies and has a significant market share. Operating only an indirect service, Air France's market share was below 1 % in 2002. Given that Air France has no hub operations at Nice and that KLM as well as two low cost carriers each offer already daily frequencies on this city pair, it is considered that Air France does not have a real commercial possibility of entry. The merger will therefore not affect substantially competition on that market.

80. As an exception to the general presumption that indirect services do not provide a competitive constraint to direct ones for short-haul flights, it is concluded that on two regional city pairs a direct / indirect overlap exists between KLM's subsidiary Basiq Air and Air France. On the city pairs Bordeaux - Amsterdam and Marseille - Amsterdam, in 2002 Air France's indirect service (via Paris) had a significant market share with a [20-30 %] and [25-35 %], respectively. While Basiq Air's direct service offers a shorter travelling time, Air France's indirect service offers many more frequencies as it is operated via its main hub. As a one day return trip is only possible with Air France, its indirect service becomes particularly attractive to time-sensitive passengers. By way of contrast, KLM's (Basiq Air) direct service does not offer a comparable number of frequencies and it is addressed mainly to non-time sensitive passengers [22]. With a combined market share of 97 % on these routes in 2002, the Merged Entity will provide almost the only air transport service on these city pairs.

81. Among the five regional routes between France and The Netherlands, with [between 125,000-175,000] passengers [45,000-55,000 O&D], Toulouse - Amsterdam is the second most important one. Offering 3 daily direct services, in 2002 KLM's market share amounted to [>90 %] and Air France's indirect service to [< 5 %].It is considered that Air France provided some competitive constraint on KLM, as it could have either entered this route with a direct service or by expanding its indirect services via its French hubs. The parties have put forward that Air France had no interest doing so, as this would imply feeding a competitor's hub in Amsterdam. However, given that Air France competes on other routes which are connected to the other competitor's hub, this argument was not accepted. [23]

82. The merger therefore creates the risk of establishing or strengthening a dominant position on these three city pairs without appropriate remedies allowing for effective entry.

83. With [less than 8,000] true O&D passengers, the route Clermont-Ferrand - Amsterdam is very thin. Air France Régional operates this route mainly for the purpose of connecting passengers in Clemont-Ferrand. Given the very small O&D traffic and that in 2002 less than 10 % of transfer passengers took connecting flights in Amsterdam, the parties' argument was accepted that KLM could not make a business case for offering a service on this city pair. In the case of Eindhoven - Paris, Air Exel /KLM operated three daily frequencies in summer 2003. Air France does not offer neither a direct nor an indirect service. The parties' argument was accepted that there would be no business case for Air France to enter this route, given the limited size of local traffic [25,000-30,000] and given that they could not expect to attract many connecting passengers. The latter would travel mainly from Amsterdam - Schiphol or Düsseldorf. It was therefore concluded that the merger will not affect substantially competition on that market.

84. If follows from the above that the notified transaction raises serious doubts on the following routes: Paris-Amsterdam; Amsterdam-Lyon; Bordeaux-Amsterdam; Marseille-Amsterdam; and Toulouse-Amsterdam.

Italian-Dutch short haul routes

Overview

85. Between The Netherlands and Italy there exist two direct overlap routes, i.e. Milan-Amsterdam and Rome-Amsterdam. These two city pairs are also the most important ones in terms of overall passengers. On both city pairs the Commission considers that the transaction risks creating a dominant position. Apart from this, there exist seven non-overlap routes where one party provides direct and the other one indirect service. On two of them, i.e. Amsterdam-Bologna and Amsterdam-Venice, the Commission considers that there is a risk that the merger creates or strengthens a dominant position.

Overlap hub-to-hub routes with direct operations: Rome-Amsterdam and Milan-Amsterdam

- Amsterdam/Milan

86. KLM and Alitalia are the main competitors on the Amsterdam/Milan city pair, which accounts for [between 675,000 and 725,000] [300,000-350,000 O&D] passengers. This is therefore a very busy city pair in the EU which connects KLM's and Alitalia's respective hubs. On the basis of MIDT data provided by the parties for 2002, Alitalia's market share amounts to [25-35 %], KLM [45-55 %] and KLM / Basiq Air for [15-25 %] of total traffic.This does not take into account Virgin Express, which has started operating one daily flight between Amsterdam and Milan Linate in the last season. However, the competitive constraint exerted by the latter is limited as the expansion of services is difficult due to slot constraints at the Milan airports.

87. In particular as this is a hub-to-hub route, potential entrants face significant entry barriers as described above for the Paris - Amsterdam city pair. In addition, there exists almost no indirect traffic on this route. It is therefore concluded that, without appropriate remedies allowing for effective entry, there is the risk that the merger creates a dominant position on this city pair for time-sensitive and for non-time sensitive customers.

- Amsterdam/Rome

88. Amsterdam/Rome is the second direct overlap hub-to-hub route between The Netherlands and Italy. Also this is an important route in terms of passengers, with in total [500,000-550,000] passengers [250,000-300,000 O&D] in 2002. Both Alitalia and KLM operate on this city pair with an overall market share of [35-45 %] and [45-55 %], respectively [24].As this is based on MIDT data, it does not include Virgin Express which started operating one frequency per day in June 2002. The parties estimate Virgin Express to account for about [20-30 %] of overall traffic. Given the limited number of frequencies of Virgin Express, presumably a majority of that traffic is from non-time sensitive passengers. Also on this city pair, there exist numerous entry barriers as described above. It has also been argued by 3rd parties that the merger would re-enforce Alitalia's position on this market which could squeeze out competitors.

89. It is therefore concluded that, without appropriate remedies allowing for effective entry, there is the risk that the merger creates a dominant position on this city pair for time-sensitive and, depending on Virgin Express' future position in the market, for non-time sensitive customers.

- Other routes between The Netherlands and Italy

90. Among the non-overlap routes, Amsterdam - Venice and Amsterdam - Bologna are the most important ones in overall traffic with [175,000-225,000] and [150,000-200,000] passengers, respectively in 2002. Both are comparable also with regard to O&D traffic, which on each city pair amounted to about [85,000-95,000] the same year. In both cases KLM operates the direct service with a market share of roughly [80-90 %] while Alitalia's indirect service accounts for [< 5 %]. In addition, Air France's indirect service on this city pair is at about [< 5 %].Lufthansa's indirect service accounts for [5-10 %]. Alitalia and Air France' indirect services therefore provided marginal competitive constraint on KLM. Considering overall traffic as well as O&D traffic, these city pairs are sufficiently large to cater for two operators.

91. Alitalia has pointed out that it cannot be considered a potential competitor on these city pairs. Alitalia's network is focused on its hubs in Milan and Rome and services which are not connected to these platforms are incompatible with Alitalia's business strategy. However, in contrast to these assertions, before it entered into an alliance agreement with Air France, Alitalia competed with the former on city pairs which were not connected to its hubs. This includes city pairs from Bologna, Naples and Venice to Paris [25]. In addition, Alitalia's subsidiary Alitalia Express, which operates on city pairs like Rome-Düsseldorf and Milan-Geneva would be equipped to enter such regional routes from Italy to Amsterdam.

92. Given the overall size of these markets and since both city pairs on one end are linked to Alitalia's domestic market, Alitalia is therefore considered to be the most likely potential competitor providing direct services on this route. The competitive constraint from this potential entry has been eliminated as a result from the merger for both non-time sensitive and time-sensitive passengers.On these city pairs slot constraints exist in Amsterdam. Other entry barriers, as discussed above, also apply on these markets. It is therefore concluded that, without appropriate remedies allowing for effective entry, there is the risk that the merger creates or strengthens a dominant position on these city pairs.

93. On the regional city pairs Amsterdam-Pisa, Amsterdam-Naples and Amsterdam-Catania Alitalia provides an indirect service while since summer 2003 KLM's low cost subsidiaries Basiq Air and Transavia offer a direct service which is addressed to tourist customers. The parties have emphasised that on these city pairs Basiq Air / Transavia has entered only recently and that it has developed a new market. As its service is addressed to tourists (i.e. seasonal traffic), it would be outside defined market for scheduled air services. As a result and due to the little brand awareness of Basiq Air in Italy it would not compete with Alitalia's scheduled flights. Instead, Basiq Air/ Transavia would compete with charter services on these markets. This is confirmed by Basiq Air / Transavia's schedules. During the winter season, Basiq Air operates 2 to 3 weekly frequencies (Monday-Friday) between Amsterdam and Naples and, depending on the month, between 0 and 2 frequencies on the city pair Pisa - Amsterdam. Transavia operates only in the summer between Amsterdam and Catania.

94. On the city pair Amsterdam - Turin, in 2002 there were in total 120,000 passengers. The market is served by KLM, offering the only direct flights, which accounts for a market share of [>90 %], and some indirect services offered by Lufthansa [< 5 %] and Air France [< 5 %]. Alitalia does not offer even indirect services on this city pair. Given that the entire transfer takes place in Amsterdam and that the O&D market is relatively small [35,000-45,000], the parties' position (confirmed by Alitalia) was accepted that it would not be a viable business strategy for Alitalia to start operating on this city pair. Finally, on the city pair Amsterdam - Florence with a total traffic of [75,000-125,000] passengers KLM does not operate itself but offers services in code-share with Meridiana. Alitalia's indirect service accounts for [< 5 %]. On this market Meridiana is the main carrier.

95. It is therefore concluded that the merger does not affect substantially competition on the city pairs from Amsterdam to Pisa, Turin, Naples, Catania and Florence.

Long haul routes

Regulatory restrictions on long haul services

96. International aviation is regulated by a web of Bilateral airservice agreements (ASAs), which determine the scope and degree of freedom to provide air services between specific city-pairs (the "freedoms of the air" provided for in the Chicago Convention. [26]When assessing the market structure for long haul flights, the Commission took into consideration the existence of specific regulatory barriers flowing from this web of ASAs in order to determine to what extent other carriers, be it European or third country carriers, may benefit from the necessary traffic rights and operation flexibility to effectively compete with the parties and, therefore, exert sufficient competitive constraints on the parties.

97. In addition, during the market investigation, competitors argued that both in France and The Netherlands, the national authorities imposed certain restrictions which limited their ability to effectively compete with the parties. The concerns raised by competitors during the market investigation relate in particular to (i) 6th freedom price leadership restrictions; and (ii) 5th freedom traffic rights for third country carriers.

(i) 6th freedom price leadership restrictions:

98. A sixth freedom is defined as carriage between two other countries via the carrier's home country (e.g. British Airways flying from Amsterdam to New-York via London);

99. Price leadership restriction on services operating 6th freedom are normally applied to prevent competitors providing indirect services on long haul destinations to third countries, via their European or third country hubs, undercutting the price of the direct services provided by the national designated carriers.

100. With the liberalization of air transport in Europe, "price leadership" restrictions were removed for such flights wholly operated within the EEA. However, 6th freedom price leadership restrictions still exist for flights between some European countries and third countries, flowing from national legislation. Some countries have abolished it on the basis of reciprocity, but the market investigation has shown that such restrictions exist and are sometimes applied in France and the Netherlands. More specifically, two competitors have argued that the price leadership restrictions prevent them from offering competitive indirect services out of France at lower published fares than what is charged by Air France for direct long-haul flights to/from France. This naturally removes effective competition from indirect flights on such routes.

101. In the past, and likewise in this proceeding, the Commission has concluded that indirect flights constitute competitive alternatives (within certain limits) to direct flights between a given long-haul city-pair and therefore they belong to the same relevant market. Absent regulatory constraints, the Commission would therefore normally not raise serious doubts on inter continental routes where the parties enjoy substantial market shares but strong competitors such as Lufthansa, British Airways, Iberia and others offer significant indirect competitive services.

102. For this conclusion to hold in the present proceeding, and for the Commission not to raise serious doubts on all routes where the main competitive constraint for the parties flows from indirect service offered by other carriers, the French and the Dutch authorities have agreed by way of declarations to remove 6th freedom price restrictions on all these routes. [27]

(ii) 5th freedom traffic rights for third country carriers:

103. 5th freedom is carriage between two foreign countries on a route with origin/destination in its home country (e.g. Singapore Airlines operating Amsterdam-NYC with origin in Singapore). This freedom is granted either through the ASAs or on an ad hoc basis. Therefore, on the intercontinental city-pairs where the Commission considers that competition will likely be affected in a meaningful way and accordingly has raised serious doubts, the French and the Dutch authorities have agreed by way of declarations to grant 5th freedom traffic rights and not to impose other restrictions as to price and capacity to facilitate new entry on each of these routes.

104. The declarations referred to in paragraphs 102 and 103 are annexed to this decision and form an integral part of the Commission's assessment

Overview

105. On long haul routes, the Commission differentiates the routes by the type of overlap between the parties and/or its partners: direct - direct overlap; direct - indirect overlap; and indirect - indirect overlap.

106. In the case of direct - direct overlap, any route was regarded as affected upon the sole existence of such overlap. However, there is only one market where both Air France and KLM and/or its partners offer both direct services and that is Amsterdam - New-York. For direct - indirect overlap and indirect - indirect overlap routes all O&D city pairs which are economically significant and where the proposed merger has an appreciable effect on the relevant long-haul city pair have been taken into account [28]. On this basis the parties notified 73 long-haul routes [29].

107. After a thorough analysis the Commission has concluded that out of these 73 long-haul routes, serious competition concerns would occur on 5 routes.

108. The data used was MIDT for the twelve period ended in March 2003 (the investigation period).

Overlap routes where both parties have direct operations:

- New-York-Amsterdam

109. With almost [375,000-425,000] passengers, this is a large route. KLM/NW and Delta (Air France's immunised partner) offer both direct services on this route, having respectively [25-35 %], [10-20 %] and [10-20 %] market shares. The proposed transaction will therefore lead to the combination of the three biggest non-stop competitors on this route which will collectively account for [55-65 %] of all O&D traffic between these two cities. By combining the flights between KLM/Northwest and Delta, this will allow the merger entity and its Skyteam partner Delta to offer a higher number of frequencies than before(27 as from the IATA summer season) [30].

110. Continental and Singapore airlines operate direct services on this route with [10-20 %] market share each. However, given that it had only limited 5th freedom rights on this route [31], Singapore Airlines has decided to withdraw its services as from the summer of 2004. Therefore Continental will be the sole competitor offering a direct service with 7 frequencies a week, compared to the 27 frequencies by the merged entity and partner.Indirect services only account for a limited number of bookings (the next largest competitor, British Airways has a market share of about [< 5 %] of total O&D bookings).

111. It follows that, in order to effectively compete with the merged entity and Delta on Amsterdam-NYC, operators would have to commence new direct or indirect services with a significant number of frequencies. However, in order to offer such services they would have to obtain the necessary assets and traffic rights, and in particular the relevant take-off and landing slots at the airports concerned. The results of the Commission investigation suggest that, although this should not be a problem at the NYC airports, this would prove difficult at Amsterdam airport as this airport has congestion problems.

112. In light of the above, there are serious risks that the operation would create a dominant position on the Amsterdam- NYC city pair.

Overlap routes where only one party has direct operations and the other party operates an indirect service and is a potential direct competitor:

- Paris-Detroit

113. This route registered [over 100,000] O&D passengers in the investigation period. Today only Northwest offers direct services on this route with [80-90 %] market share. Air France, together with its partner Delta, and KLM operate indirect services via, respectively, Cincinnati, NYC and Amsterdam. Air France and Delta account for [< 5 %] market share, being the largest indirect service provider on this route(followed by Lufthansa and British Airways with [< 5 %] each). The merger leads, therefore to elimination of the most important actual competitor.

114. In addition, Paris is a hub of Air France and Detroit a hub of Northwest. Air France has never operated the route, but according to the parties this is a route with enough traffic to support two direct competitors [32]. Air France has admitted it is a potential direct competitor of Northwest and therefore its mere potential entry on the route created a constraint which will be removed by the present transaction.

115. It follows that, in order to effectively compete with the merged entity and Northwest on this route, operators would have to commence new direct or indirect services with a significant number of frequencies. Being a hub to hub route, this will contribute to make entry or expansion more difficult for competitors, and in particular could preclude the entry of alternative non-stop operators who do not have the same connecting facilities at each end of the route. In addition, in order to offer such services they would have to obtain the necessary assets and traffic rights, and in particular the relevant take-off and landing slots at the airports concerned. The results of the Commission investigation suggest that, although this should not be a problem at the Detroit airport, this would prove difficult at both Paris airports, which have congestion problems.

116. In light of the above, there are serious risks that the operation would create a dominant position on the Paris-Detroit city pair.

- Amsterdam-Atlanta

117. This route accounted for [over 80,000] O&D passengers in the relevant period. Only Delta operates a direct service, while KLM/Northwest, Air France and Alitalia operate indirect services via Detroit, Houston, Paris and Milan, respectively. Delta accounts for [75-85 %] of total market shares while the combined market shares for the parties and their partners post-merger would be [>90 %]. Other carriers operating indirect services only account for [5-10 %] of the O&D traffic,with British Airways being the most important with [< 5 %]. The proposed transaction will have the effect of eliminating its most important competitor, KLM/NW with [5-10 %] market share.

118. In addition, Atlanta is a hub for Delta and Amsterdam for KLM. KLM operated this route until the end of IATA summer 2001 when it decided to withdraw due to financial losses. For this reason, the parties consider that the prospect of re-entry of KLM on the route is low and therefore it is very unlikely that Delta considers KLM as a serious or realistic potential (re)entrant.This is because the route would be too thin to support two head-to-head network carriers flying direct services on a daily basis. Even if there is connecting traffic, this is far more limited at Amsterdam, where only one third of the passengers transferred to other services. Further, KLM has little connection possibilities in Atlanta and therefore would not be able to attract US connecting traffic.

119. The Commission considers that KLM is very likely to exert a potential competitive constraint on Delta on this route because of the following: the routes links Atlanta (an important US destination in terms of traffic) to its hub. The withdrawal from the route in 2002 (probably enhanced by decrease in traffic after the events of 9/11) does not prevent KLM from re-entering the route once the general economic context improves and specifically when demand for air transport services on this market increases. The route is sufficiently thick to support two direct hub operators because of the connecting possibilities these enjoy at their hubs.In addition, KLM carried alone more than [60,000] O&D passenger in IATA summer 2000 (one year before it abandoned the route), which indicates that KLM's direct service satisfied consumer demand. Further, KLM has sufficient connecting possibilities in Atlanta because its partner Northwest operates 44 services out of Atlanta to other US destinations. [33]

120. It follows that, in order to effectively compete with the merged entity and Delta on this route, operators would have to commence new direct or indirect services with a significant number of frequencies. Being a hub to hub route, this will contribute to make entry or expansion more difficult for competitors, and in particular could preclude the entry of alternative non-stop operators who do not have the same connecting facilities at each end of the route. In addition, in order to offer such services they would have to obtain the necessary assets and traffic rights, and in particular the relevant take-off and landing slots at the airports concerned. The results of the Commission investigation suggest that, although this should not be a problem at the Atlanta airports, this would prove difficult as Amsterdam airport has congestion problems.

121. In light of the above, there are serious risks that the operation would create a dominant position on the Amsterdam-Atlanta city pair.

Overlap routes where only one party has direct operations and the other party operates and indirect service

- Amsterdam-Lagos and Paris-Lagos

122. These routes accounted for respectively [< 45,000] and [>35,000] O&D passengers in the relevant period. On Amsterdam-Lagos, KLM is the sole provider of direct services and accounts for [80-90 %] of the O&D market, being Air France the largest competitor offering indirect services with [< 5 %] market share. On Paris-Lagos, Air France is the sole provider of direct services and accounts for [80-90 %] of O&D bookings, followed by KLM with [5-10 %] of bookings via Amsterdam.In addition, the market investigation suggests that other indirect service providers will not be able to increase capacity or frequencies due to the restrictions flowing from their ASA with Nigeria.

123. The parties argue that these are relatively thin routes and the investigation period was quite exceptional since during the year of 2002 British Airways, a major operator on the London-Lagos route, had temporarily withdrawn its services. Therefore, during the year of 2002, passengers who used to fly with British Airways were directed to use Air France and KLM's services and connect at Paris or Amsterdam. British Airways has re-commenced services on the London-Lagos route in 2003 and according to the parties they would soon recover an important market share also of the Paris-Lagos and Amsterdam-Lagos routes. Therefore, after the merger, BA would exert a competitive constraint on the merged entity. In addition, a new Nigerian carrier would start operating these routes in the near future.

124. The Commission investigation suggests that to this date, even after having re-entered the route, British Airways has a very limited market share [34] and therefore, until otherwise demonstrated, the present operation will reinforce the already considerable market power the parties have on these two routes. In addition, the perspective of a new Nigerian direct entrant on these routes is not a fact the Commission can take into account since it is a future and merely hypothetical fact.

125. It follows from above that there are serious risks that the operation would strengthen the dominant position Air France and KLM already enjoy today on, respectively, Paris-Lagos and on Amsterdam-Lagos city pairs.

The other direct-indirect overlap routes

126. The operation will also create overlaps on other routes where one of the parties (or one of its partners) provides direct services, while the other party provides competitive indirect services. On 7 routes, the parties' market shares after the transaction are below [45-55 %] and the proposed transaction will not materially alter the competitive conditions on these routes. These are Athens-NYC; Paris-Montreal; Madrid-San Francisco; Paris-Mexico; Paris-Damascus; Amsterdam-Tokyo and Paris-Jakarta. On 18 other routes, while the parties market shares are between [45-55 %] and [75-85 %], the investigation has shown that there is sufficient competitive constraint exerted by either other carriers providing a direct service [Barcelona-NYC; Amsterdam-Los Angeles; Amsterdam-San Francisco; Marcelona-Quyaquil; Barcelona-Quito; Amsterdam-Casablanca; Paris-Amman and Paris-Osaka] or providing indirect services [Venice-NYC; Milan-Miami; Amsterdam-Boston; Milan-Boston; Paris-Caracas; Amsterdam-Johannesburg; Amsterdam-Manila; Paris-Manila; Amsterdam-Shanghai; Amsterdam-Osaka]. On all these routes, it is considered that either the direct service operated by one competitor and/or the other indirect services operated exert a sufficient competitive constraint on the parties. It is generally true that in all these city pairs, there is an ample number of possible indirect alternatives connecting at a variety of hub airports. Furthermore, certain third parties have indicated they have extra capacity on the long haul routes concerned. In two of these routes, while the parties and its partners will have a combined market share slightly over [75-85 %] [Venice-NYC and Amsterdam-Shanghai], the investigation has shown that Lufthansa offers an indirect service which accounts for [10-20 %] of the O&D bookings and that Lufthansa alone is able to exert a competitive constraint on the parties.

127. It follows from above that the elimination of one or even two indirect alternatives stemming from the transaction will not materially affect the conditions of competition on the city-pairs concerned. The remaining indirect services either individually or collectively will most likely exert significant competitive constraint on the merged entity and passengers will still have a wide range of alternative travel if they choose not to fly with Air France/KL.

Overlap routes without direct operations [indirect-indirect overlap]

128. The proposed operation leads also to a number of overlaps on routes where the parties and/or its partners only operate indirect services. There are two sets of indirect-indirect overlap routes: the routes where none of the parties or its partners offers a direct service but a third airline does; and routes where no carrier offer a direct service. There is no indication that in any of these markets the transaction will create or strengthen a dominant position because of the presence of other competitors is considerable. The only routes where the parties will have a substantial part of total O&D bookings [55-65 %] and the operation will result in a significant addition of bookings [5-10 %] are Lyon-Montreal; and Paris-Nomea. However, it appears that despite these market shares, the merged entity will remain subject to sufficient competitive constraints. Other competitors account for market shares around [10-20 %] and since these are city-pairs served via an intermediate airport, passengers will remain having an ample choice of connecting services, within and outside Europe.

129. In the light of the above, it is concluded the proposed transaction would not lead to the creation or strengthening of dominant position on the indirect-indirect overlap routes.

Network effects

130. As pointed out in paragraph 14, corporate customers generally conclude non-exclusive contracts with airlines or airline alliances. Their choice of a carrier depends in particular on whether its network complies with their travelling needs. Most corporate customers are likely to deal with the airline which has a hub that can be easily reached from the headquarter or its operational facilities. Thus, a corporate customer located in Paris would often have a corporate deal with Air France and SkyTeam, as they would be able serve the most important city pairs for this customer. Only corporate customers which are widely spread geographically would significantly benefit from using different networks. In their case the choice of a particular airline for a particular flight would depend more on the price on an individual O&D city pair and the overall conditions (discounts) offered by the contract. It is acknowledged that such corporate customers to some extent may be affected by the merger's impact on network competition.

131. In this respect, some competitors have argued that the merger between Air France and KLM effectively reduces the number of world-wide alliances from four (SkyTeam, Star, the KLM/NW Joint Venture and OneWorld) to three, as the KLM/NW Joint Venture and SkyTeam alliance merge. In their view, this would have a serious impact on network competition, as in fact only SkyTeam and Star are immunised alliances.

132. However, the market investigation has not confirmed that corporate customers are generally concerned with the merger's impact on alliance competition. As regards long-haul travel, many emphasise the benefits which arise as the concentration will open new routes to them. Where corporate customers raise concerns about reduced competition leading to higher prices and/or less schedules, this is linked mainly to intra-European city pairs which are of particular importance to them. To the extent that corporate customers propose remedies, they are addressed to the need to generate competition on individual city pairs. Thus, generally speaking, while they have concerns regarding particular city pairs, corporate customers do not consider to be negatively affected by the merger as far as network competition is concerned.

133. With regard to non-corporate costumers, it is even less evident that a merger between the KLM/NW Joint Venture and the SkyTeam alliance as such would have an appreciable effect on competition. The effect on transatlantic flights (direct and indirect), has already been addressed in the above O&D analysis. Because airline alliances tend to serve most O&Ds with lower traffic densities via their hub-and-spoke system, the present transaction would lead to the elimination of one alliance on these indirect routes. It could be argued therefore that the reduced number of alliances raises competitive concerns on routes where passengers connect at both ends (e.g. Kansas City - US hub - EU hub - Warsaw). However, the Commission considers that it is unlikely that the present transaction creates competition concerns on these double connection routes for a host of reasons. First, airlines via interlining and code-share agreements can provide air service for low densities traffic between two points. Although these agreements are not immunised from antitrust scrutiny, these rival air services exert a competitive constraint on immunised alliance service on double connection routes. As a result, passengers travelling on the vast majority of O&D requiring a double connection will still have a number of options even after the merger between Air France and KLM [35]. Second, in some cases competition may appear to be more limited because not all alliances operate a feeder service between the point of origin or the point of destination and their respective hub. However, there is little evidence that competition concerns would arise after the merger. Indeed, potential competition may exert a significant competitive constraint on the existing air service. In double connect O&D, market entry would be relatively easy. A competing alliance that is not yet present may have to introduce only one feeder flight to its own hub to enter a double connect O&D. Alternatively, when several airlines already operate passenger air service between all points on a double connect O&D, some of the airlines would simply have to agree on code sharing or interlining to propose a new air service to passengers.

134. Finally, there may be thin double connecting routes where the traffic is so small that competing alliances would refrain from introducing such a feeder service. Arguably, for these low traffic double connecting routes the price set by an airline / alliance is not primarily influenced by the competitive situation on that particular market (Kansas City - Warsaw). The price level on such a route would depend much more on the competitive situation on the transatlantic hub-to-hub flight and on the capacity situation. If, due to the competitive situation on the transatlantic flight, the airline / alliance face a low load-factor, given the very low marginal cost in this industry, it would have an incentive to reduce prices for such feeder traffic.

135. In light of the above arguments it can be concluded that the network effects of the present transaction do not raise serious doubts.

B. Air cargo transport

136. The Commission's assessment of the proposed operation's impact on air cargo transport takes into account, for the same reasons provided for in the passenger market, the existing co-operation agreements between the parties with Delta and Northwest for the North Atlantic routes and Kenya Airways (Kencargo) for Kenya. In the light of the extensive co-operation agreements between Air France and Alitalia [...], the market shares of the Alitalia has also been considered on a failsafe basis, without prejudice of the specific scrutiny of these agreements by the Commission under other proceedings.

137. The parties submit that Martinair, despite being jointly controlled by KLM, is commercially independent from KLM and has not concluded any co-operation agreement withy any of the parties. In the KLM/Martinair case [36], the Commission considered that Martinair maintained substantial operational and commercial independence from KLM. The market investigation has generally confirmed this view, except for some competitors, who suggest that both companies coordinate their network to avoid overlap. The Commission investigations indicate that KLM and Martinair however appear to compete on an important number of routes. Notwithstanding this, if Martinair's market shares have to be added to those of KLM, the assessment of the transaction would not be substantially altered.

138. As regards intra-European routes, the parties have estimated their combined market share of less than 1 % if alternative means of transport are considered, and around [0-10 %] in volume of total air cargo transport in Europe.

139. On intercontinental routes, the figures provided by the parties show that the merged entity will become one of the main players on the routes between Europe and Africa, with around [20-30 %] ([30-40 %] with Martinair) and between Europe and Latin America-the Caribbean, with around [30-40 %] ([45-55 %] with Martinair) of the respective markets. On the North Atlantic, Europe-Near/Middle East and Europe-Far East routes, the parties' combined market share would be around 15 % or less [37].

140. In line with the proposed market definition, the African, Asian and Latin America routes have been further subdivided on a regional or country by country basis. Out of the sixteen routes in which the parties activities overlap (nineteen if Martinair is considered) [38], the figures provided by the parties shows that their combined market share is only significant [39] as regards routes to/from Cameroon [40-50 %], Nigeria [45-55 %], St Martin [50-60 %] and Mexico ([40-50 %] or [65-75 %] with Martinair). Kenya should be added to this list if Martinair's position is taken on board, where the merged entity would hold [40-50 %] of the market.

141. The increment for Cameroon is insignificant, where KLM has a market share of about 1 %. Further, the merger entity will face the competition of other important players such as Cameroon Airlines and SN Brussels Airlines. It is therefore concluded that the proposed transaction does not raise serious doubts on this route.

142. On the other four routes, the increment after the merger is more significant: [10-20 %] for Nigeria, [5-10 %] for Kenya, [5-10 %] for St Martin and [15-25 %] for Mexico ([20-30 %] if Martinair is considered).

143. Nigeria is a relatively thin route (e.g. outbound traffic of less than 2,500 tons a year). Important carriers such as British Airways and Lufthansa are present in this market. In addition, the merger entity will compete with a large number of all-cargo carriers such as Hydro Air, DAS Air, Panalpina/ASB Air, KM Air, KABO Air, Fresh Air (charters hold around [25-35 %] of the market) and integrators such as DHL.

144. The market shares of Air France in Kenya ([5-10 %]) are outweighed by those of DAS Air ([10-20 %]) or British Airways and Lufthansa (each holds around [5-10 %] of the market).

145. Saint Martin, a Caribbean island of 37 square miles partly Dutch and partly French, is a particularly thin route (e.g. outbound traffic of [< 150] tons a year). On such a small route, the merged entity still will face competition from Corsair.

146. The market shares in Latin American estimated by the parties do not take into consideration significant indirect services to and from Europe provided by competing North-American scheduled and non scheduled carriers via US airports (e.g. Miami, Huntsville and Dallas). This regards St. Martin and particularly Mexico. Further, the parties and some competitors argue that Mexico (or part of it) could be regarded as part of the North American market, given that Mexico is part of NAFTA and that some cargo to/out of Mexico is flown to/from the United States, while the leg to/from Mexico is transported by road. There are important players present in this market, such as Cargolux (with a market share of [10-20 %]), Iberia (around [5-10 %]), Atlas/Polar, Aeromexico, Lufthansa, British Airways, DHL and LTU, operating from airports such as Mexico City, Cancun, Guadalajara and Monterrey.

147. On the four routes, the market investigation shows that competitors (combined airlines, all cargo carriers and, to a lesser extent, integrators) are in a position to substantially constrain the competitive position of the parties. These competitors offer a sufficient number of frequencies and/or operate full freight aircrafts enabling them to efficiently compete with the merged entity in all possible segments of the market. Further, the market investigation has indicated that airports in Nigeria, Kenya, St. Martin or Mexico are not slot congested. Some of these competitors have expressly indicated during the market investigation that they dispose of extra capacity to take up additional demand should the parties decide to increase their prices, particularly as regards Nigeria, Kenya and Mexico.

148. It also appears that customers of air cargo transport (freight forwarders, integrators and consolidators) are able to exert a considerable countervailing buying power vis-à-vis air cargo carriers, given their size and importance on certain routes. Important freight forwarders such as Schenker, Panalpina and Küehne & Nagel are particularly present in Latin America. Cargo transport seems to be less affected by national or brand loyalty than passenger transport and contracts are normally concluded for a short duration (six months to a year) and with no exclusivity clause. In the course of the market investigation, customers have declared their readiness to switch to a different supplier should the parties decide to increase their prices.

149. It is therefore concluded that the proposed concentration does not create or strengthen a dominant position on those routes as a result of which effective competition would be significantly impeded in the EEA or any substantial part of that area.

C. Maintenance, repair and overhaul

150. The merged entity would have a worldwide market share in the overall merchant MRO business of less than 5 %. Even if one consider the narrowest possible market definition the impact of the proposed transaction is not significant on any market.

151. In the line maintenance segment, the parties' activities do not overlap if A-checks are B-checks are considered separately. If A and B-checks are taken together, their combined market share for overall line maintenance do not exceed 15 % in the three airports where the parties overlap: Rome-Fiumicino, Barcelona and Paris-Charles de Gaulle. In the heavy maintenance segment, the parties' activities only overlap in the provision of C-checks in the EU for the Boeing 737, with less than [5-10 %] of the market (less than [0-5 %] world-wide), and the Boeing 747, with less than [10-20 %] (less than [10 %] world-wide). Important competitors on these markets are Singapore Technologies Aerospace, FLS Aerospace, TIMCO, Goodrich, Ameco, Taeco, SR Technics, Japan Airlines, Lufthansa Technick (LHT) and Iberia.

152. The parties' activities also overlap for world-wide engine maintenance services for General Electric CF6-80 and CF6-50 engines and world-wide component maintenance services for Boeing 737, 747 and 767. For all alternative market definitions discussed above their market share would not exceed [5-10 %], with the exception for the General Electric CF6-50 engine, where the market share would be less than [20-30 %]. However, the proposed transaction only leads to a non-significant increment of less than 1 % from Air France. The main players on these markets are GE Engine Services (with a market share of [25-35 %]), Snecma Engine Services, LHT and MTU for engine maintenance and Honeywell, LHT, Goodrich, Hamilton Sundstrand and Rockwell Collins for component maintenance.

153. As a consequence, the proposed operation does not raise serious concerns in any MRO market considered as regard its compatibility with the Common Market.

VII. UNDERTAKINGS SUBMITTED BY THE PARTIES

154. In order to remove the serious doubts raised by the proposed concentration the parties have submitted Undertakings pursuant to Article 6(2) Merger Regulation, which are intended to remove the competitive concerns identified above. The Undertakings are attached to this decision and form an integral part thereof.

155. Furthermore, with a view to assuaging the Commission and third parties concerns that the effectiveness of the Undertakings may be impaired by certain regulatory measures which may hinder new entry on a number of long-haul routes, the competent authorities of France and The Netherlands have submitted declarations whereby they give reassurances that such regulatory barriers will not be triggered on those long-haul routes. The Commission has taken note of these declarations, which form an integral part of its assessment.

156. All competitive concerns identified in the course of the investigation are addressed by the package of Undertakings. Furthermore, comments by third parties during the market test have been taken into account and, as a result, the Undertakings initially offered by the parties have been significantly improved.

Summary of the Undertakings

157. The following is a concise summary of the Undertakings, highlighting the main features and the main thrust of the package of the Undertakings and it does not constitute a legal interpretation or an exhaustive list thereof.

158. The parties undertake to make slots available on the city-pairs identified in the investigation, both for short and long hauls. Furthermore, the parties have offered additional Undertakings, such as frequency freeze, and to enter into agreements with new entrants regarding interlining and pro-rate, block space, frequent flyers programs and intermodal transport so as to make the "package" more effective and new entry more attractive (see further below).

Undertaking concerning slots

a) Short-Haul/European routes: The Parties commit to make available, without any compensation and in accordance to the procedure specified in the Undertakings, a number of slots at Amsterdam and/or Paris and/or Lyon and/or Milan and/or Rome, and to allow one (or more) new entrants to operate new or additional non-stop scheduled daily passenger air services in the mentioned European city pairs.This will allow new entrants to operate on the following routes: Paris/ Amsterdam; Lyon/ Amsterdam; Marseille/Amsterdam; Toulouse/Amsterdam; Bordeaux/ Amsterdam; Milan/Amsterdam; Rome/Amsterdam; Venice/Amsterdam and Bologna/ Amsterdam.

As regards the Paris-Amsterdam route, the Parties have undertaken to make available six frequencies per day. In case no third party has applied for and has been awarded any of the slots within a certain period, new entrants shall be granted "grandfather" rights, once it has operated the service between Paris and Amsterdam for six IATA seasons. Slots shall be allocated preferably in block to one single new entrant.

b) Long-Haul/Intercontinental routes: Slots will also be made available at the Amsterdam and Paris airports for the long haul routes identified in the decision, where competition concerns arise. The purpose is to allow a new entrant to provide on each of such routes one non-stop or indirect (with a connecting time of not more than 150 minutes) scheduled passenger air service operated on a daily basis (or in any case not less than 6 times a week).

c) Conditions: The slots released by the Parties shall be within 90 minutes for long haul city pairs and within 30 minutes for intra-European city-pairs of the time requested by the new entrant (see further Section 1.3 of the Undertakings). Reflecting the need for effective entry, at Paris the entrant can choose whether to take slots in CDG or in Orly. Given the very limited slot portfolio held by Air France and KLM in Milan-Linate, the parties only offer slots at Malpensa.

d) The duration of the Undertakings relating to slots is unlimited.Slots may be claimed, in accordance with the procedure laid down in section 3 of the Undertakings, by a new entrant with no limitation in time. The parties will nonetheless be able to invoke, in exceptional circumstances, the Review Clause (paragraph 14) pursuant to which the Commission may decide to waive, modify or substitute any one or more of the Undertakings.

e) If, following such a review, the Commission concludes that the Merged Entity's obligation to release slots on a particular city pair is extinguished, the new entrant can continue using the slots it has already received beforehand. Once it stops using the slots on the particular city pair, these slots have to be surrendered to the slot co-ordinator.

Other measures

a) Frequency freeze: The Parties commit themselves not to add frequencies on Paris-Amsterdam or Lyon-Amsterdam or, with a marginal degree of flexibility, Amsterdam-New-York for a period of six IATA seasons.

b) Interlining agreements: The Parties agree to enter, at the request of a new entrant, into an interline agreement concerning any of the city pairs identified in the decision;

c) Special prorate agreements: The Parties undertake to enter, at the request of a potential new entrant, into a special prorate agreement for traffic with a true origin and destination in either France and/or the Netherlands, provided part of the journey involves the Paris-Amsterdam city pair. The Agreement shall be at the same conditions as what applies to the parties' alliance partners;

d) Frequent flyer programme (FFP): The Parties will allow a new entrant, on request, to be hosted in the Parties' FFP for the identified city pairs at the same conditions as the parties' alliance partners;

e) Inter-modal services: The Parties shall enter, at the request of a railway or other surface transport company operating between France and The Netherlands and/or between Italy and The Netherlands, into an inter-modal agreement.

f) Blocked space agreements (BSA): The Parties shall enter, at the request of a potential new entrant, into a BSA for traffic with a true origin and destination, on the one hand, in The Netherlands and, on the other, at Marseille, Toulouse or Bordeaux, provided part of the journey involves the Paris-Amsterdam city pair. This would allow an operator on the Paris-Amsterdam city pair to feed passengers from other cities in France. Passengers from these cities would obtain an alternative service to the indirect services already offered by Air France on these markets. More generally, the Parties commit to enter, into a BSA pertaining to the identified European city pair(s) if the number of non-stop frequencies offered by such potential new entrant on that city pair is lower than the number operated by the Parties.

g) Condition pertaining to fares: Whenever the Merged Entity reduces a published fare on the city pair Paris-Amsterdam, it agrees to apply an equivalent reduction on the corresponding fare on the city pair Lyon-Amsterdam, as long as there is no competitive air service on the latter route.

Assessment of the Undertakings

159. The Undertakings submitted by the Parties constitute a complex and comprehensive package which built on past practice of the Commission as regards competitive concerns arising in the civil aviation sector.

160. The conditions of slot release have been simplified and clarified. In particular, the parties have agreed to reduce the "bracket periods" from 45 minutes to 30 minutes for short-haul routes and from 120 minutes to 90 minutes for long-haul routes. These conditions have been considered by new entrants as crucial to facilitate entry and represent an important improvement.

161. While the analysis for the purpose of the present case has been based on the O&D market definition adopted in past cases, the Commission has also taken into account concerns raised by competitors in particular with regard to hub dominance. This is reflected in undertakings (a) with regard to the possibility that entrants may obtain grandfather rights for slots obtained operating on the city Paris-Amsterdam city pair and (b) by the rule that, once entrants stop operating on a particular city pair, the related slots do not eventually go back to the Merged Entity but to the slot co-ordinator. Such "structural" measures are meant to increase the attractiveness of market entry and will have an effect on competition wider than on the specific routes identified in this decision.

162. The parties have submitted Undertakings, which are of unlimited duration save for specific exceptions. This is more constraining than in previous cases, in which slots were released by the parties only for a limited duration in time, and in practice characterises the slot-related Undertakings as "structural" in so far as the parties have committed to "divest" the use of such slots indefinitely.

163. With regard to long-haul markets, the Statements by the French ad the Dutch Governments regarding the "fifth" and "sixth" freedoms are an important element for the Commission to conclude that effective competition is possible on these markets. 6th freedom price restrictions in fact may render competition between indirect and direct services less effective. This could affect a large number of markets where competitive constraints on the operator offering a direct service stems from such indirect competition. The French and Dutch governments have declared that they will not restrict such price competition on notified long-haul routes which include a direct service to or from France or The Netherlands. 5th freedom traffic rights are important to allow competitors to operate one stop flights via Amsterdam or Paris to a third country (i.e. Singapore-Amsterdam-New-York). As, from the perspective of the European consumer, such services could provide a direct connection (i.e. Amsterdam - New-York), and given the very strong preference of time-sensitive passengers for direct flights the possibility of effective 5th freedom competition can be a significant competitive constraint on certain long-haul city pairs.

164. Furthermore, traditional industry measures (interlining, prorate, frequent-flyer and blocked-space agreements) figure as well prominently in the package of Undertakings and will help new entrants to obtain a minimum number of passengers necessary to start operations on particular city pairs.

165. On the market Paris-Amsterdam, competition from Thalys could be strengthened by the possibility to conclude inter-modal agreements. This, for instance, could allow Thalys to offer time-sensitive customers to combine a one way railway trip with a return flight. Inter-modal travelling could be offered at attractive conditions. For instance, many promotional air fares require a minimum stay before the return trip can be undertaken. As these restrictions are waived for the inter-modal service, travellers could benefit from the promotional air fare while choosing any return trip offered by the railway undertaking.

166. The purpose of the condition pertaining to fares is twofold. As potential entrants have raised concerns that the Merged Entity, given its overall size and strength in the market, will apply predatory pricing, the condition makes such behaviour considerably more costly on the city pair Paris-Amsterdam (since all price reductions have to be passed on also to the monopoly route Lyon-Amsterdam). Second, as long as there is no entry on the city pair Lyon-Amsterdam, the condition pertaining to fares would allow consumers on this city pair to benefit from price reductions which result from competition on the city pair Paris-Amsterdam.

167. Finally, the package of Undertakings foresees the establishment of a monitoring system under the supervision of the Monitoring Trustee and a "fast-track" arbitration procedure. These features should contribute to ensuring the effective implementation of the Undertakings.

Conclusions on the Undertakings

168. The Commission considers that the Undertakings submitted are sufficient to eliminate the serious doubts as to the compatibility of the transaction with the Common Market. In particular the Commission considers that the Undertakings are suitable to solve the competition concerns in the markets identified during the investigation by allowing and facilitating third parties' entry.

169. The Undertakings provided for in paragraph 1, 2, 3.3, 3.4, 4 to 10 and 13.1 constitute conditions of this decision, as only through full compliance therewith can the necessary changes on the relevant markets be achieved. The remaining Undertakings constitute obligations as they concern the implementing steps, which are necessary to achieve the sought changes.

VIII. CONCLUSION

170. The Commission has concluded that the Undertakings submitted by the parties are sufficient to address the competition concerns raised by the concentration. Accordingly, subject to the full compliance with the Undertakings submitted by the notifying parties, the Commission has decided not to oppose the notified operation and to declare it compatible with the Common Market and with the EEA Agreement.This decision is adopted in application of Article 6(1) (b) and Article 6(2) of Council Regulation (EEC) No 4064-89.

COMMITMENTS PACKAGE

CASE COMP/M.3280-AIR FRANCE/KLM

Pursuant to Article 6(2) of Council Regulation (EEC) 4064-89 as amended (the "Merger Regulation"), Société Air France ("AF"), in agreement with Koninklijke Luchtvaart Maatschappij N.V. ("KLM"), submits the commitments specified below (the "Commitments") in order to enable the European Commission (the "Commission") to declare the proposed combination of AF and KLM (the "Concentration") compatible with the Common Market by its decision pursuant to Article 6 (1) (b) of the Merger Regulation (the "Decision").

These Commitments shall take effect upon receipt of the Commission's Decision declaring the Concentration compatible with the Common Market pursuant to Article 6(1) (b) of the Merger Regulation and will be binding on AF, its subsidiaries, successors and assigns.

These Commitments are offered exclusively in the context of the notified concentration between AF and KLM and are without prejudice to the position of AF and/or its alliance partners (Aeromexico, Alitalia, CSA, Delta Air Lines, Korean Air) in other cases currently being examined by the European Commission under any of Regulation 3975-87, Regulation 17-62, Articles 84 and 85 EC Treaty.

This text shall be interpreted in the light of the Decision to which the Commitments are attached as conditions and obligations, and in the general framework of Community law, in particular in the light of the Merger Regulation, and by reference to the Commission Notice on remedies.

0 DEFINITIONS

ACA

the Alliance Co-operation Agreement entered into by AF and Alitalia Linee Aeree Italiane S.p.A. on 27 July 2001, together with its related agreements, notified to the Commission, updated from time to time and currently being examined pursuant to Regulation 3975-87 in Case COMP/38284

Competitive Air Service

* a non-stop scheduled passenger air service which is operated on a daily basis (or in any case not less than six (6) times a week) on one or more of the Identified Long-Haul City Pairs;

* an intra-European feeder passenger air service from Paris or Amsterdam (the point of origin or destination on one of the Identified Long-Haul City Pairs) to the Prospective New Entrant's European hub which connects with a passenger air service which is operated by the Prospective New Entrant on a daily basis (or in any case not less than six (6) times a week) from that hub to New-York, Detroit, Atlanta or Lagos (the point of origin or destination on one of the Identified Long-Haul City Pairs) provided the connecting time at the Prospective New Entrant's hub is not more than 150 minutes; or

* a non-stop scheduled daily passenger air service which is operated on one or more of the Identified European City Pairs.

Consummation of the Concentration

the date, following the end of the offer period, on which shares in KLM are exchanged for shares in AF

Effective Date

The date of the Decision, unless otherwise specified

Identified Long-Haul City Pairs

Amsterdam-New-York, Paris-Detroit, Amsterdam-Atlanta, Paris-Lagos, Amsterdam-Lagos

Identified European City Pairs

Paris-Amsterdam, Lyon-Amsterdam, Marseille-Amsterdam, Toulouse-Amsterdam, Bordeaux-Amsterdam, Amsterdam-Milan, Amsterdam-Rome, Amsterdam-Venice, Amsterdam-Bologna

Indemnified Party

has the meaning given to it in Paragraph 11.2.6

Intermodal Partner

has the meaning given to it in Paragraph 7.1

Merged Entity

Air France and KLM and companies and/or affiliated businesses controlled by the entities after the Consummation of the Concentration

Monitoring Trustee

means an individual or institution, independent from AF, KLM or the Merged Entity, who is approved by the Commission and appointed by AF and who has the duty to monitor the Merged Entity's compliance with the conditions and obligations attached to the Decision as more fully described in Clause 11

Morning Peak Time

the period between 08h00 and 10h59 (Paris/Amsterdam time)

New Air Service Provider

* any airline that commences a new non-stop service on an Identified European City Pair or that increases the number of frequencies it operates on an Identified European City Pair; or

* any airline (or airlines that are each members of Star Alliance or oneworld) that individually or collectively by codeshare provide(s) a new or additional Competitive Air Service.

Prospective New Entrant

any airline (or airlines that are each members of Star Alliance or oneworld) able to offer a Competitive Air Service individually or collectively by codeshare and needing a slot or slots to be made available by the Merged Entity in accordance with the Commitment(s) to operate a Competitive Air Service

Requesting Party

has the meaning given to it in Paragraph 12.2

Standard Slot Allocation Procedure

has the meaning given to it in Paragraph 1.3.1

1 SLOTS

1.1 SLOTS FOR IDENTIFIED LONG-HAUL CITY PAIRS

1.1.1 The Merged Entity undertakes to make slot(s) available at Amsterdam or Paris to allow a Prospective New Entrant to operate one (1) new or additional Competitive Air Service on each of the Identified Long-Haul City Pairs.

1.1.2 The Merged Entity will make slots available to the Prospective New Entrant selected in accordance with Clause 3.

1.2 SLOTS FOR IDENTIFIED EUROPEAN CITY PAIRS

1.2.1 The Merged Entity undertakes to make slot(s) available at Amsterdam and/or Paris and/or Lyon and/or Milan and/or Rome to allow one or more Prospective New Entrants to operate a new or additional Competitive Air Service on the Identified European City Pairs.

The Merged Entity shall be obliged to make available to Prospective New Entrants the number of slots needed to support in aggregate:

* for flights between Paris and Amsterdam: up to six (6) frequencies per day;

* for flights between Lyon and Amsterdam: up to three (3) frequencies per day;

* for flights between Marseille and Amsterdam: up to two (2) frequencies per day;

* for flights between Toulouse and Amsterdam: up to two (2) frequencies per day;

* for flights between Bordeaux and Amsterdam: up to two (2) frequencies per day;

* for flights between Milan and Amsterdam: up to four (4) frequencies per day;

* for flights between Rome and Amsterdam: up to three (3) frequencies per day;

* for flights between Venice and Amsterdam: up to two (2) frequencies per day;

* for flights between Bologna and Amsterdam: up to two (2) frequencies per day.

The Merged Entity's obligation to make slots available to Prospective New Entrants for flights between (i) Milan and Amsterdam, (ii) Rome and Amsterdam, (iii) Venice and Amsterdam and (iv) Bologna and Amsterdam shall however lapse in the event that either AF or Alitalia terminates the ACA and AF confirms in writing to the Commission that their co-operation has terminated on the France-Italy bundle.

1.2.2 In the event that no Prospective New Entrant has begun to operate a Competitive Air Service on the Paris-Amsterdam city pair [...], any Prospective New Entrant will be authorised to use any slots released for this city pair thereafter at its discretion on any other city pair once it has operated between Paris and Amsterdam for six full IATA seasons.

1.2.3 Subject to the Commission's confirmation of the competitiveness of the service in accordance with this Paragraph , all frequencies operated by airlines independent of and unconnected to the Merged Entity on the Identified European City Pairs shall be counted against the number of slots to be released by the Merged Entity under Paragraph.

On the basis that Virgin is already operating one (1) frequency between Amsterdam and Rome and one (1) frequency between Amsterdam and Milan, the Merged Entity shall not be required to make more slots available to Virgin on these city pairs than are necessary to enable Virgin to operate three (3) additional frequencies between Amsterdam and Milan and/or two (2) additional frequencies between Amsterdam and Rome. As soon as Virgin starts to operate two (2) or more frequencies between Amsterdam, on the one hand, and Milan or Rome, on the other hand, one of which is operated in the Morning Peak Time, the number of frequencies operated by Virgin shall be counted against the number of slots to be released by the Merged Entity under Paragraph 1.2.1

The Merged Entity will inform the Monitoring Trustee and the Commission of the announced commencement by a carrier of a new or additional Competitive Air Service on an Identified European City Pair that does not use slots made available by the Merged Entity as soon as possible following the announcement of that service. The Commission will confirm to the Merged Entity, in accordance with Clause 3, whether or not such new or additional Competitive Air Service is sufficient for the relevant carrier to qualify as a competitor with respect to that Identified European City Pair once such service is commenced.

1.3 CONDITIONS PERTAINING TO SLOTS

1.3.1 At least three (3) weeks prior to the IATA scheduling conference for the traffic season in which the Prospective New Entrant intends to commence a new or additional Competitive Air Service, the Prospective New Entrant shall notify the Merged Entity of its intention to request slots pursuant to the Commitments. A Prospective New Entrant shall be eligible to receive slots pursuant to these Commitments only if it can demonstrate that all reasonable efforts to obtain slots for the Identified European or Long-Haul City Pair(s) through the normal workings of the slot allocation procedure before the beginning of the relevant IATA traffic season (the "Standard Slot Allocation Procedure") have failed.

To this end, the Prospective New Entrant shall apply for these slots at the forthcoming IATA slot conference through the Standard Slot Allocation Procedure and maintain an "open book" policy for the airports concerned during the entire period between the notification to the Merged Entity of its intention to apply for slots in order to operate services on an Identified European or Long-Haul City Pair and the end of the respective IATA scheduling period, including the final allocation of slots by the coordinator following the Slot Return Date.

The Prospective New Entrant will be deemed not to have exhausted all reasonable efforts to obtain slots if :

(i) slots were obtained through the Standard Slot Allocation Procedure within ninety (90) minutes of the times requested, but such slots have not been accepted by the Prospective New Entrant in the case of an Identified Long-Haul City Pair; and/or

(ii) slots were obtained through the Standard Slot Allocation Procedure within thirty (30) minutes of the times requested, but such slots have not been accepted by the Prospective New Entrant in the case of an Identified European City Pair; and/or

(iii) slots were obtained through the Standard Slot Allocation Procedure more than ninety (90) minutes from the times requested and the Prospective New Entrant did not give the Merged Entity the opportunity to exchange those slots for slots within ninety (90) minutes of the times requested in the case of an Identified Long-Haul City Pair; and/or

(iv) slots were obtained through the Standard Slot Allocation Procedure more than thirty (30) minutes from the times requested and the Prospective New Entrant did not give the Merged Entity the opportunity to exchange those slots for slots within thirty (30) minutes of the times requested in the case of an Identified European City Pair.

The slots released by the Merged Entity shall be within ninety (90) minutes in the case of an Identified Long-Haul City Pair and within thirty (30) minutes in the case of an Identified European City Pair of the time requested by the Prospective New Entrant, if the Merged Entity has slots available within this time-window. In the event that the Merged Entity does not have slots available within this time-window, it shall offer to release the slots closest in time to the Prospective New Entrant's request.

1.3.2 Requests for slots to the slot coordinator and to the Merged Entity shall be renewed by the Prospective New Entrant for each subsequent IATA scheduling season.

If the New Air Service Provider has obtained slots from the Merged Entity pursuant to these Commitments for a particular IATA season and requests for the following IATA season some or all of the slots at the same times as the slots released by the Merged Entity, the Merged Entity shall make slots available to the New Air Service Provider at the same time as in the previous IATA season, unless the Merged Entity does not hold a slot at that time, in which case it shall release a slot as close as possible in time to the slot granted in the preceding IATA season, provided the New Air Service Provider has complied with the conditions and procedure described above.

1.3.3 Where one or more slots have been made available by the Merged Entity, the New Air Service Provider shall remain under a continuing obligation to make all reasonable efforts to obtain suitable slots at Amsterdam and/or Paris and/or Lyon and/or Milan and/or Rome through the normal workings of the slot allocation procedures enabling it to provide a Competitive Air Service.

1.3.4 For the purposes of these Commitments, it is accepted that a single daily connecting service departing from Amsterdam and/or Paris can constitute part of a new or additional Competitive Air Service on more than one Identified European or Long-Haul City Pair.

1.3.5 Where a Prospective New Entrant which has obtained slots pursuant to these Commitments decides (i) not to commence services on an Identified European or Long-Haul City Pair, (ii) to operate a lower number of frequencies or decides to cease operating on an Identified European or Long-Haul City Pair (so-called misused or under-used slots) or (iii) not to use the slots on the Identified European or Long-Haul City Pair (so-called improperly used slots), it shall inform the Merged Entity in writing and return such misused or unused slots to it immediately. In these circumstances, the Merged Entity shall make the returned slots available to any other Prospective New Entrant.

For the purposes of this Paragraph, a Prospective New Entrant will be deemed to have ceased operating on an Identified European or Long-Haul City Pair where it has not used at least 80 % of its slots during the scheduling season for which they had been allocated for the city pair in question, unless this non-use of the slots is justified on one of the grounds referred to in Article 10 (5) of Regulation (EEC) 95-93 or in any other regulation that amends or supersedes it. Should the Prospective New Entrant be considered to have ceased operating the Identified European or Long-Haul City Pair pursuant to this Paragraph, the Merged Entity may refuse to surrender slots to the said Prospective New Entrant for the next IATA season on this Identified European or Long-Haul City Pair.

1.3.6 Should the New Air Service Provider notify the Merged Entity too late in a scheduling season for it to use the returned slots pursuant to Article 10 (3) of Regulation (EEC) 95/93, either with immediate effect or after the deadline provided for in Article 10(4) of that Regulation and before the effective start of the scheduling season, the Merged Entity shall be entitled to require the Prospective New Entrant to transfer to it a comparable slot as compensation in the event the slot is lost. If, for any reason, the Prospective New Entrant is unable to transfer to the Merged Entity a comparable slot, the Merged Entity may justify the non-use of the surrendered slot on the basis of Article 10 (5) of Regulation (EEC) 95/93 in order to recover and retain the unused slot.

1.3.7 To ensure that the slots released by the Merged Entity are used in a manner consistent with these conditions, a mechanism shall be agreed by the Merged Entity and the Prospective New Entrant, and approved by the Commission, that will allow the Monitoring Trustee to monitor how the slots are being used.

1.3.8 Slots made available by the Merged Entity under these Commitments shall be offered without any compensation.

1.3.9 On the basis that Paris CDG and ORY airports are substitutable for intra-European passenger air services, Prospective New Entrant(s) may request slots at either CDG or ORY.

1.3.10 The Merged Entity's portfolio of slots at LIN being very limited, any slots to be made available at Milan airports pursuant to these Commitments may be released from either LIN or MXP at the Merged Entity's discretion.

1.3.11 Slots may be requested by a Prospective New Entrant from the Merged Entity at any time in the day, except that in the light of environmental regulations in force at Amsterdam Schiphol airport, the Merged Entity shall not be obliged to release slots at this airport between 11pm and 7am local time for flights on any Identified European City Pair.

1.3.12 To ensure the viability of the Merged Entity's hub-and-spoke network at Amsterdam Schiphol airport, the Merged Entity shall not be obliged to release more than nine (9) pairs of slots in any one hour period in the Morning Peak Time at Amsterdam Schiphol airport. In the event that a request is made for the release of more than nine (9) pairs of slots in any one hour period under this Paragraph, the Merged Entity shall offer the Prospective New Entrant the next closest slot to the time requested in accordance with Paragraph 1.3.7 to above.

2 DURATION OF THE COMMITMENTS

2.1 To ensure that a Prospective New Entrant on any Identified European or Long-Haul City Pair has guaranteed access to a structural remedy, save where otherwise indicated, these Commitments are unlimited in duration and may be invoked at any time by a Prospective New Entrant, subject to review pursuant to Clause 4.

2.2 Without prejudice to Paragraph 1.2.2, in the event that the Commission has concluded pursuant to Clause that the Merged Entity's obligation to release slots on a particular city pair is extinguished, thereafter any slots actually surrendered by the Merged Entity for that particular city pair that are misused or underused (within the meaning of Regulation 95/-3) or improperly used (that is, not used on the Identified European or Long-Haul City Pair for which they were released) by the New Air Service Provider shall be surrendered to the coordinator and not to the Merged Entity

3 SELECTION PROCEDURE, ROLE OF THE MONITORING TRUSTEE AND APPROVAL OF THE COMMISSION

3.1 A Prospective New Entrant wishing to obtain slots from the Merged Entity pursuant to these Commitments shall notify the Merged Entity of its intention to apply for these slots at the forthcoming IATA slot conference within the time period specified in Paragraph 1.3.1. A copy of this notification shall be sent immediately by the Merged Entity to the Monitoring Trustee and the Commission.

3.2 Should a Prospective New Entrant be unable to obtain slots through the Standard Slot Allocation Procedure at the IATA slot conference for the traffic season in which services are intended to commence, it shall apply to the Merged Entity for slot releases within one (1) week following the end of that slot conference.

The application shall take into account the slots obtained at the slot conference within ninety (90) minutes of the times requested in the case of an Identified Long-Haul City Pair and give the Merged Entity the opportunity to exchange slots obtained more than ninety (90) minutes from the times requested, for slots of the Merged Entity within ninety (90) minutes of the times requested pursuant to Paragraph .

The application shall, similarly, take into account the slots obtained at the slot conference within thirty (30) minutes of the times requested in the case of an Identified European City Pair and give the Merged Entity the opportunity to exchange slots obtained more than thirty (30) minutes from the times requested for slots of the Merged Entity within thirty (30) minutes of the times requested pursuant to Paragraph 1.3.1.

A copy of the application shall be sent by the Prospective New Entrant to the Monitoring Trustee and to the Commission at the same time as it is sent to the Merged Entity.

Within two (2) weeks following the end of the IATA slot conference for the traffic season in which services are intended to commence, the Merged Entity shall identify the exact time of the slots it intends to make available to the Prospective New Entrant.

Within four (4) weeks following the end of the IATA slot conference for the traffic season in which services are intended to commence, the Merged Entity shall submit to the Monitoring Trustee and to the Commission a proposal for the selection of the Prospective New Entrant on the Identified European or Long-Haul City Pair together with an indication as to the probable time(s) at which slots will be released.

3.3 When making a proposal for the selection of the Prospective New Entrant on any particular Identified Long-Haul City Pair, the Merged Entity shall accord a preference to:

(a) a Prospective New Entrant that has indicated that it intends to operate a non-stop scheduled passenger air service; and then

(b) a Prospective New Entrant that has indicated that it intends to operate an intra-European feeder passenger air service from Paris or Amsterdam to that carrier's European hub to connect with a new non-stop scheduled service from that hub operated by the Prospective New Entrant to New-York, Detroit, Atlanta or Lagos, as the case may be; and then

(c) a Prospective New Entrant that has indicated that it intends to operate an intra-European feeder passenger air service from Paris or Amsterdam to that carrier's European hub to connect with an existing non-stop scheduled service from that hub operated by the Prospective New Entrant to New-York, Detroit, Atlanta or Lagos, as the case may be.

3.4 When making a proposal for the selection of the Prospective New Entrant on the Paris-Amsterdam Identified European City Pair, the Merged Entity shall accord a preference to the Prospective New Entrant that has indicated that it intends to operate the highest number of frequencies per day.

3.5 The Commission shall consider the advice of the Monitoring Trustee (as provided for in Paragraph ) and decide whether or not to approve the Merged Entity's proposal pursuant to the following criteria:

* the Prospective New Entrant is independent of and unconnected to the Merged Entity and;

* the Prospective New Entrant is a viable existing or potential competitor, with the ability, resources and commitment to operate the Identified European or Long-Haul City Pair in the long term as a viable and active competitive force.

3.6 To assist the Monitoring Trustee in the preparation of his advice and/or the Commission in taking its decision, the Monitoring Trustee and/or the Commission may request the Prospective New Entrant to provide to the Monitoring Trustee and/or the Commission a detailed business plan. This plan shall contain a general presentation of the company including its history, its legal status, the list and a description of its shareholders and the two most recent yearly audited financial reports. The detailed business plan shall provide information on the plans that the company has in terms of development of its network, fleet etc, and detailed information on its plans for the route on which it wants to operate. The latter should specify in detail planned operations on the route over a period of three (3) years (size of aircraft, number of frequencies operated, planned time-schedule of the flights) and expected financial results (expected traffic, revenues, profits). The Monitoring Trustee and/or the Commission may also request a copy of all co-operation agreements the Prospective New Entrant may have with other airlines. Business secrets and confidential information will be kept confidential by the Commission and the Monitoring Trustee and will not become accessible to other undertakings or to the public.

3.7 If the Commission does not oppose the Merged Entity's proposal within six (6) weeks following the end of the IATA slot conference, this proposal will be deemed accepted.

3.8 In case the Commission does not approve the proposal submitted by the Merged Entity, if other carriers have applied to the Merged Entity for slots, the Merged Entity shall communicate to the Monitoring Trustee and the Commission without delay the names of other carrier(s) to be selected.

3.9 Within one (1) week after the approval by the Commission of the selection of the Prospective New Entrant on the Identified European or Long-Haul City Pair, the Merged Entity shall submit its written proposal for slot releases to this Prospective New Entrant.

4 FREQUENCY FREEZE

4.1 The Merged Entity shall not add frequencies on Paris-Amsterdam or Lyon-Amsterdam, as the case may be, for a period starting when a New Air Service Provider begins operations on the relevant city pair. This frequency freeze shall last six full consecutive IATA seasons.

4.2 The Merged Entity shall not add frequencies beyond a total of fourteen (14) frequencies per week on Amsterdam-New-York J F Kennedy airport, nor shall it add frequencies on Amsterdam-New-York Newark airport, for a period of six full consecutive IATA seasons starting when a New Air Service Provider begins operating a non-stop service on that city pair.

5 INTERLINING AGREEMENTS

5.1 At the request of a New Air Service Provider, the Merged Entity shall enter into an interline agreement concerning any Identified European or Long-Haul City Pair operated by the New Air Service Provider.

5.2 Any such interline agreement shall be subject to the following restrictions:

* it shall apply to the first class, business class and leisure travel categories only;

* it shall provide for interlining on the basis of the Merged Entity's published one-way fares when a one-way ticket is issued or half of the Merged Entity's published round-trip fares when a round-trip ticket is issued;

* it shall be limited to true origin and destination traffic on the Identified City Pair operated by the New Air Service Provider;

* it shall be subject to the MITA rules and/or normal commercial conditions;

* it shall include the possibility for the New Air Service Provider, or travel agents, to offer a return trip comprising services provided one-way by the Merged Entity and one-way by the New Air Service Provider.

5.3 Subject to seat availability in the relevant fare category, the Merged Entity shall carry a passenger holding a coupon issued by a New Air Service Provider for travel on an Identified European or Long-Haul City Pair. However, to avoid abuse, the Merged Entity may require that the New Air Service Provider or the passenger, where appropriate, pay the (positive) difference between the fare charged by the Merged Entity and the fare charged by the New Air Service Provider. In cases where the New Air Service Provider's fare is lower than the value of the coupon issued by it, the Merged Entity may endorse its coupon only up to the value of the fare charged by the New Air Service Provider. A New Air Service Provider shall enjoy the same protection in cases where the Merged Entity's fare is lower than the value of the coupon issued by it.

5.4 All interline agreements entered into pursuant to this Clause 5 for a particular Identified European or Long-Haul City Pair shall lapse automatically in the event that the New Air Service Provider ceases to operate that city pair.

6 FREQUENT FLYER PROGRAMME

6.1 If a New Air Service Provider does not participate in the Merged Entity's frequent flyer programme ("FFP") or does not have its own comparable FFP, the Merged Entity shall allow it, on request, to be hosted in its FFP for the Identified European or Long-Haul City Pairs operated by the New Air Service Provider. The agreement with the New Air Service Provider shall be on terms such that the New Air Service Provider shall have equal treatment with the Merged Entity's alliance partners on the same Identified European or Long-Haul City Pair.

6.2 Any agreement relating to a particular Identified European or Long-Haul City Pair and entered into pursuant to this Clause 6 shall lapse automatically in the event that the New Air Service Provider ceases to operate that city pair.

7 COMMITMENT TO FACILITATE INTERMODAL SERVICES

7.1 At the request of a railway or other surface transport company operating between France and The Netherlands and/or between Italy and The Netherlands (an "Intermodal Partner"), the Merged Entity shall enter into an intermodal agreement whereby they provide passenger air transport on their services on any of the Identified European City Pairs as part of an itinerary that includes surface transportation by the Intermodal Partner.

7.2 Any intermodal agreement entered into pursuant to this Clause 7 shall be based on the MITA principles (including the Intermodal Interline Traffic Agreement - Passenger and IATA Recommended Practice 1780e) and normal commercial conditions. The Merged Entity shall accept full pro-rating according to the terms applied by MITA members, including on routes where only rail services are provided. No restrictions shall apply to fare combinations between carriers that are IATA intermodal MITA members and the most restrictive conditions rule shall apply only for the applicable segment and its carrier. The Merged Entity and the Intermodal Partner may waive minimum stay requirements on any fare and any city pair they operate. Such decisions are respected and published reciprocally. Where the Intermodal Partner requires notification of a sector mileage, a location identifier or an add-on fare, the Merged Entity shall make such a request to IATA under normal IATA procedures.

7.3 At the request of a potential Intermodal Partner, the Merged Entity shall make efforts in good faith to reach an agreement on conditions comparable to those granted to other Intermodal Partners, provided that the necessary requirements are met especially with regard to safety, quality of service, insurance coverage and liability limits. The conditions of such an agreement shall override the general obligations arising pursuant to this Clause 7.

8 SPECIAL PRORATE AGREEMENTS AND BLOCKED SPACE AGREEMENTS ON SEGMENTS BEYOND PARIS OR AMSTERDAM

8.1 At the request of a Prospective New Entrant, the Merged Entity shall enter into a special prorate agreement with it for traffic with a true origin and destination in either France and/or the Netherlands, provided part of the journey involves the Paris-Amsterdam city pair. The conditions for a special prorate agreement shall be on terms such that the New Air Service Provider shall have equal treatment with the Merged Entity's alliance partners on the same Identified European City Pair.

8.2 At the request of a Prospective New Entrant, the Merged Entity shall enter into a blocked space agreement with it for traffic with a true origin and destination, on the one hand, in The Netherlands and, on the other, at Marseille, Toulouse or Bordeaux, provided part of the journey involves the Paris-Amsterdam city pair. The blocked space agreement shall be based on a fixed number of seats (hard block basis) and apply for at least one entire IATA traffic season. The number of seats covered by such an agreement shall be a maximum of 15 % of the seats offered on a particular frequency and in any event not more than 30 seats on a particular aircraft. The conditions for any such blocked space agreement shall be comparable to those entered into with third non-alliance/other alliance carriers in connection with the Identified European City Pair in question. The Prospective New Entrant shall carry the full commercial risk attaching to the seats covered by the blocked space agreement.

9 BLOCKED SPACE AGREEMENT ON AN IDENTIFIED EUROPEAN CITY PAIR

9.1 At the request of a Prospective New Entrant, the Merged Entity shall enter into a blocked space agreement pertaining to the Identified European City Pair(s) operated by the Prospective New Entrant if the number of non-stop frequencies offered by the Prospective New Entrant on that city pair is lower than the number operated by the Merged Entity.

9.2 Any blocked space agreement shall be based on a fixed number of seats (hard block basis) and apply for at least one entire IATA traffic season. The number of seats covered by such an agreement shall be a maximum of 15 % of the seats offered on a particular frequency and in any event not more than 30 seats on a particular aircraft. The conditions for any such blocked space agreement shall be comparable to those entered into with third non-alliance/other alliance carriers in connection with the Identified European City Pair in question. The Prospective New Entrant shall carry the full commercial risk attaching to the seats covered by the blocked space agreement.

10 CONDITIONS PERTAINING TO FARES

10.1 Each and every time the Merged Entity reduces a published fare on Paris-Amsterdam, the Merged Entity agrees to apply an equivalent reduction (in percent) on the corresponding fare on Lyon-Amsterdam. This commitment applies only as long as the reduction on Paris-Amsterdam remains effective and as long as there is no Competitive Air Service on Lyon-Amsterdam. For purposes of this commitment, a published fare shall include applicable IATA fares, carrier fares that are distributed to CRSs via the public tariff data base of ATPCO (Airline Tariff's Publishing Corporation) and fares marketed on the Internet where such fares are available to the general public.

11 MONITORING TRUSTEE

11.1 APPOINTMENT OF MONITORING TRUSTEE

11.1.1 A Monitoring Trustee shall be appointed in accordance with the procedure described in Paragraph 1.1.2. The Monitoring Trustee must be familiar with the airline industry and have the experience, competence and independence necessary for this appointment. The Monitoring Trustee will have had no direct or indirect employment, consultancy or other relationship with AF or KLM during the past two years and will have no such relationship with the Merged Entity for the three years following the completion of its mandate.

11.1.2 Within two months of the Effective Date and in any event before Consummation of the Concentration, AF, in agreement with KLM, shall submit a list or one or more persons whom AF proposes to appoint as the Monitoring Trustee to the Commission for approval. The proposal shall contain sufficient information for the Commission to verify that the proposed Monitoring Trustee fulfils the requirements set out in Paragraph and shall include:

(i) the full terms of the proposed mandate, which shall include all provisions necessary to enable the Monitoring Trustee to fulfil its duties under these Commitments;

(ii) the outline of a work plan which describes how the Monitoring Trustee intends to carry out its assigned tasks.

The Commission shall have the discretion to approve or reject the proposed Monitoring Trustee and to approve the proposed mandate subject to any modifications it deems necessary for the Monitoring Trustee to fulfil its obligations. If only one name is approved, AF, in agreement with KLM, shall appoint or cause to be appointed the individual or institution concerned as Monitoring Trustee. If more than one name is approved, AF, in agreement with KLM, shall be free to choose the Trustee to be appointed from among the names approved. The Monitoring Trustee shall be appointed within one week of the Commission's approval, in accordance with the mandate approved by the Commission.

If all the proposed Monitoring Trustees are rejected, AF, in agreement with KLM, shall submit the names of at least two more individuals or institutions within one week of being informed of the rejection, in accordance with the requirements and the procedure set out in Paragraph 11.1.1.

If all further proposed Monitoring Trustees are rejected by the Commission, the Commission shall nominate a Monitoring Trustee, whom AF shall appoint, or cause to be appointed, in accordance with a trustee mandate approved by the Commission.

11.2 MONITORING TRUSTEE'S MANDATE

11.2.1 The Monitoring Trustee's mandate shall include, in particular, the following responsibilities:

(i) to monitor the satisfactory discharge by AF or, as the case may be, the Merged Entity of the obligations entered into in these Commitments in so far as they fall within the scope of the Monitoring Trustee's mandate;

(ii) to propose to the Merged Entity such measures as the monitoring Trustee considers necessary to ensure the Merged Entity's compliance with the conditions and obligations attached to the Decision;

(iii) to advise and make a written recommendation to the Commission as to the suitability of any Prospective New Entrant submitted for approval to the Commission under Clause 3;

(iv) to provide written reports to the Commission on the progress of the discharge of its mandate, identifying any respects in which the Monitoring Trustee has been unable to discharge its mandate; and

(v) at any time, to provide to the Commission, at its request, a written or oral report on matters falling with the Monitoring Trustee's mandate.

11.2.2 The Merged Entity shall receive simultaneously a non-confidential version of any written recommendation made by the Monitoring Trustee to the Commission (as provided for in Paragraph 11.2.1 (iii)

11.2.3 The reports provided for in Paragraph 11.2.1 (iv) shall be prepared in English. They shall be sent by the Monitoring Trustee to the Commission within ten (10) working days from the end of every IATA season following the Monitoring Trustee's appointment or at such other time(s) as the Commission may specify, and shall cover developments in the immediately preceding IATA season period. The Merged Entity shall receive simultaneously a non-confidential copy of each Monitoring Trustee report.

11.2.4 The Merged Entity shall provide the Monitoring Trustee with such assistance and information, including copies of all relevant documents, as the Monitoring Trustee may reasonably require in carrying out its mandate and shall pay reasonable remuneration for its services.

11.2.5 The Monitoring Trustee shall have full and complete access to any of the Merged Entity's books, records, documents, management or other personnel facilities, sites, technical information necessary for fulfilling its duties.

11.2.6 The Merged Entity shall indemnify the Monitoring Trustee (and, where appropriate, its employees and agents) (each an "Indemnified Party") and hold each Indemnified Party harmless, and hereby agrees that an Indemnified Party shall have no liability to the Merged Entity for any liabilities arising out of the performance of the Monitoring Trustee's duties under the Commitments, except to the extent that such liabilities result from the wilful default, recklessness, gross negligence or bad faith of the Monitoring Trustee (or, where appropriate, its employees and agents).

11.2.7 At the expense of the Merged Entity, the Monitoring Trustee may appoint advisors, subject to the Merged Entity's approval, if the Monitoring Trustee considers the appointment of such advisors necessary for the performance of its duties under the mandate, provided that any fees incurred are reasonable.

11.3 TERMINATION OF MANDATE

11.3.1 If the Trustee ceases to perform its functions under the Commitments or for any other good cause, including the exposure of the Monitoring Trustee to a conflict of interest:

(i) the Commission may, after hearing the Monitoring Trustee, require the Merged Entity, to replace the Monitoring Trustee; or

(ii) the Merged Entity, with the prior approval of the Commission, may replace the Monitoring Trustee.

11.3.2 If the Trustee is removed the Monitoring Trustee may be required to continue in its function until a new Monitoring Trustee is in place to whom the Monitoring Trustee has effected a full hand over of all relevant information. The new Monitoring Trustee shall be appointed in accordance with the procedure referred to Paragraph 11.1.2.

11.3.3 Aside from being removed in accordance with Paragraph 11.3.1, the Monitoring Trustee shall cease to act as Monitoring Trustee only after the Commission has discharged it from its duties. However, the Commission may at any time require the reappointment of the Monitoring Trustee if it subsequently appears that the Commitments might not have been fully and properly implemented.

12 FAST TRACK DISPUTE RESOLUTION

12.1 In the event that a Prospective New Entrant, a New Air Services Provider or an Intermodal Partner has reason to believe that the Merged Entity is failing to comply with the requirements of the Commitments vis-à-vis that party, the fast track dispute resolution procedure described in this Clause will apply.

12.2 Any Prospective New Entrant, New Air Services Provider or Intermodal Partner who wishes to avail itself of the fast track dispute resolution procedure (a "Requesting Party") must notify the Merged Entity in writing setting out in detail the reasons leading that party to believe that the Merged Entity is failing to comply with the requirements of the Commitments (the "Notice"). The Requesting Party and the Merged Entity will use their best efforts to resolve all differences of opinion and to settle all disputes that may arise through co-operation and consultation within a reasonable period of time not to exceed fifteen (15) business days after receipt of the Notice.

12.3 Should the Requesting Party and the Merged Entity fail to resolve their differences of opinion through cooperation and consultation as provided for in Paragraph 12.2, the Requesting Party shall nominate an arbitrator.

12.4 The Merged Entity shall, within two weeks of receiving notification in writing from a Requesting Party of the appointment of the Requesting Party's arbitrator, nominate its arbitrator and provide to the Requesting Party in writing detailed reasons for its challenged conduct.

12.5 The arbitrators nominated by the Merged Entity and the Requesting Party shall, within one week from the nomination of the former, agree to appoint a third arbitrator. If the arbitrators nominated by the Merged Entity and the Requesting Party cannot agree on the nomination of a third arbitrator, they shall ask the President of the International Chamber of Commerce ("ICC") to appoint the third arbitrator.

12.6 The arbitrators shall be instructed to establish an arbitration tribunal and to make a preliminary ruling on the contested issues within one month of the appointment of the third arbitrator. The preliminary ruling shall be applicable immediately and until the final decision is issued. The final decision shall be taken by the arbitrators within six (6) months of the appointment of the third arbitrator.

12.7 In their preliminary ruling and their final decision, the arbitrators shall also decide the action, if any, to be taken by the Merged Entity in order to ensure compliance with the Commitments vis-à-vis the Requesting Party.

12.8 Any of the arbitrators will be entitled to request any relevant information from the Merged Entity or the Requesting Party in order to enable the arbitrators to reach a decision.

12.9 The burden of proof in any dispute under this fast track dispute resolution procedure shall be borne as follows: i) the Requesting Party must produce evidence of a prima facie case, and ii) if the Requesting Party produces evidence of a prima facie case, the arbitrator must find in favour of the Requesting Party unless the Merged Entity can produce evidence to the contrary.

12.10 The arbitrators shall be instructed not to disclose confidential information and to apply the standards attributable to confidential information and business secrets by European Community competition law.

12.11 The arbitration shall be in English. The arbitration award shall, in addition to dealing with the merits of the claim, impose the fees and costs of the prevailing party upon the party that is unsuccessful.

12.12 In the event of disagreement between the parties to the arbitration regarding the interpretation of the Commitments, the arbitrators shall inform the Commission and may seek the Commission's interpretation of the Commitments before finding in favour of any party to the arbitration. The Commission may, at any time, issue a submission during the arbitration procedure.

12.13 Nothing in the arbitration procedure shall affect the powers of the Commission to take decisions in relation to the Commitments in accordance with its powers under the Merger Regulation and the EC Treaty.

13 GENERAL PROVISIONS

13.1 The Commitments shall take effect on the Effective Date; provided, however, that the Merged Entity shall not be required to provide any slots until one (1) month after the Consummation of the Concentration, except in the case of the Amsterdam-Lagos and Paris-Lagos city pairs where the Merged Entity shall not be required to provide any slots to a carrier whose service falls within the category described in Paragraph (c) until the beginning of IATA Summer 2006 season. If the Concentration is abandoned, abrogated, not approved or disapproved by a relevant Government Authority, or otherwise terminated, then these Commitments shall automatically cease to apply.

13.2 Upon request of the Merged Entity, any deadline provided for in the Commitments can be extended by the Commission if there are circumstances justifying such an extension. The Merged Entity's request shall be properly reasoned and shall specify the recent circumstances that in the Merged Entity's opinion justify an extension.

13.3 If the approval of the Concentration by another governmental authority is made subject to requirements that are potentially inconsistent with these Commitments, AF may request a review and adjustment of these Commitments in order to avoid such inconsistencies.

13.4 The Merged Entity shall use its best efforts to provide the Commission with such information as the Commission may require in connection with these Commitments within ten (10) working days from receipt of the Commission's reasoned request.

14 REVIEW CLAUSE

The Commission may, in response to a request from the Merged Entity justified by exceptional circumstances or a radical change in market conditions, such as the operation of a Competitive Air Service on a particular Identified European or Long-Haul City Pair, waive, modify, or substitute any one or more of the undertakings in these Commitments.

At the request of the Merged Entity, all the Commitments submitted herein may be reviewed, waived or modified by the Commission based on long-term market evolution.

The Commission will consider such a request as a matter of urgency in the case of the frequency freeze provided for in Clause 4, when the Merged Entity can demonstrate that, in the case of exceptional events, additional flights are required on a short term basis.

10 February 2004

______________________________

Jean-Cyril Spinetta,

duly authorised for and on behalf of Société Air France

in agreement with

______________________________

Leo Van Wijk,

duly authorised for and on behalf of Koninklijke Luchtvaart Maatschappij N.V.

Affaire n°COMP/M.3280-Air France/KLM 9 février 2004

DECLARATION DES AUTORITES AERONAUTIQUES FRANCAISES :

Afin qu'elle puisse donner son accord au projet de concentration, qui lui a été notifié le 18 décembre 2003, entre la société Air France (Air France) et Koninklijke Luchtvaart Maatschappij N.V. (KLM), en Phase Un, la Commission a demandé aux autorités aéronautiques françaises de préciser par une déclaration publique sa position sur les questions d'ordre réglementaire suivantes.

1) Droits de cinquième liberté :

Les autorités aéronautiques françaises approuveront les demandes de droits de trafic déposées auprès d'elles par une ou plusieurs compagnies qui souhaiteraient exploiter en cinquième liberté les lignes Paris-Detroit ou Paris-Lagos, dans la limite au total d'un service quotidien sur chacune de ces deux lignes. En outre, les autorités aéronautiques françaises n'opposeront pas de restrictions en matière de prix, capacités ou autre, de nature à empêcher les compagnies d'exercer une concurrence effective sur ces deux lignes.

Pour les demandes émanant des compagnies non communautaires, les autorités aéronautiques françaises s'attacheront à ce que l'octroi de droits de trafic en cinquième liberté s'inscrive dans le cadre de politique communautaire en matière de relations avec les pays tiers.

2) Initiatives tarifaires en sixième liberté :

Nonobstant leurs prérogatives en matière d'initiatives tarifaires prises par des transporteurs aériens sur le marché français pour des services extracommunautaires, les autorités aéronautiques françaises ne désapprouveront pas les tarifs de transport de passagers en sixième liberté sur les lignes suivantes :

a) lignes sur lesquelles la Commission a émis des doutes sérieux quant à la compatibilité de l'opération deconcentration avec les règles du Marché commun :

Pans-Detroit

Paris-Lagos

b) sur lesquelles aucune autre compagnie ne propose une offre concurrentielle en service direct: Paris - Seattle

Lyon Montréal

Paris-Orlando

Lyon-New-York

Nice-Montréal

Paris-Caracas

Paris-Lima

Paris-Le Cap

Paris-Manille

Paris-Djakarta

Paris-Nouméa

S'agissant de la ligne Paris-Nouméa, les autorités aéronautiques françaises se réservent cependant la possibilité de désapprouver les tarifs en sixième liberté qui seraient de nature à porter atteinte à la continuité territoriale entre la France métropolitaine et la Nouvelle-Calédonie.

Les autorités françaises ont bien noté, à l'occasion de leurs entretiens avec la Commission, que celle-ci dans les affaires futures qui lui seraient notifiées dans le cadre des règlements (CEE) n°4064-89 du 21 décembre 1989 et (CE) n° 139-2004 du 20 janvier 2004, ou qu'elle examinerait dans le cadre des articles 81 et 85 du traité ou du règlement (CE) n° 1-2003 du 16 décembre 2002, adoptera la même approche afin de faire bénéficier les transporteurs aériens français d'un traitement équivalent sur d'autres marchés communautaires en matière de tarifs de sixième liberté.

La liste des lignes mentionnées aux points 1) et 2) de la présente déclaration est susceptible d'être réexaminée si la Commission constate un changement significatif dans leur situation concurrentielle.

Les dispositions de la présente déclaration entrent en vigueur à la date de la mise en œuvre de la concentration entre les deux entreprises.

Declaration by the Dutch civil aviation authorities to the European Commission

In order to allow the European Commission to declare, in view of the commitments proposed by Air France and KLM, the proposed concentration between Air France and KLM compatible with the Common Market, the Dutch civil aviation authorities declare the following:

1. The Dutch civil aviation authorities will not use their powers to prohibit, disapprove or in any way hinder sixth freedom price leadership on the following city pairs:

" Amsterdam-New-York;

" Amsterdam-Atlanta;

" Amsterdam-Lagos;

" Amsterdam-Los Angeles;

" Amsterdam-SanFrancisco;

" Amsterdam-Boston;

" Amsterdam-Orlando:

" Amsterdam-Johannesburg;

" Amsterdam-Casablanca;

" Amsterdam-Manila;

" Amsterdam-Shanghai;

" Amsterdam-Osaka.

The Dutch civil aviation authorities have taken note of the fact that the Commission will take the same approach when assessing future transcontinental alliances or merger cases in the air transport sector in order to ensure equivalent treatment for Dutch air carriers with respect to sixth freedom price leadership in the other European (EEA) markets,

2. The Dutch civil aviation authorities are willing to grant fifth freedom traffic rights to one or more additional entrants, to allow such entrant(s) to operate in total one competitive daily direct passenger service on each of the city pairs Amsterdam-New-York, Amsterdam-Atlanta and Amsterdam-Lagos. In addition the Dutch civil aviation authorities undertake not to impose restrictions as to price and capacity or other matters which would impede the entrant(s) to compete effectively on this route:

The Hauge, 10 February 2004

Director-General of Civil Aviation Prof! Dr. F.L. Bussink

[1] OJ L 395, 30.12.1989 p. 1; corrigendum OJ L 257 of 21.9.1990, p. 13; Regulation as last amended by Regulation (EC) No 1310-97 (OJ L 180, 9. 7. 1997, p. 1, corrigendum OJ L 40, 13.2.1998, p. 17).

[2] Turnover calculated in accordance with Article 5(1) of the Merger Regulation and the Commission Notice on the calculation of turnover (OJ C 66, 2.3.1998, p. 25). To the extent that figures include turnover for the period before 1.1.1999, they are calculated on the basis of average ECU exchange rates and translated into EUR on a one-for-one basis.

[3] See Commission Notice on the definition of the relevant market for the purposes of Community Competition law, OJ C 372, 09.12.1997, p. 5-13.

[4] See also judgements of the Court of Justice in Case 66-86 Ahmed Saeed Flugreisen, of 11.04.1989,[1989], ECR 803 and of the Court of First Instance in Case 2-93 Air France/Commission (TAT), of 19.05.1994, [1994], ECR 323.

[5] See in particular Case 66-86 Ahmed Saeed Flugreisen, of 11.04.1989, [1989] ECR 803 and Commission Decision of 05.07.2002, in Case No COMP/37.730 - AuA/Lufthansa, OJ L 242, 10.09. 2002, p. 25.

[6] See also Commission Decisions of 11.08.1999, in Case No COMP/JV.19 - KLM/Alitalia, OJ C 96, 05.04.2000, p. 5 and of 12.01.2001 in Case No COMP/M.2041 - United Airlines/US Airways.

[7] See Commission Decision of 11.08.1999, in Case No. COMP/JV.19 - KLM/Alitalia, OJ C 96, 05.04.2000, p. 5, point 21.

[8] Commission Decision of 30.10.2002, in Cases No COMP/36.201, 36.076 and 36.078 - United/Lufthansa/SAS, OJ C 264, 30.10.2002, p. 5-9 and Commission Decision of 30.10.2002, in Case No COMP/36.111 - KLM/NW, OJ C 264, 30.10.2002

[9] Commission Decision of 05.07.2002, in Case No COMP/37.730 - AuA/Lufthansa, OJ L 242, 10.09. 2002, p. 25.

[10] For instance one 3rd party has shifted all its operations from Orly to CDG. Another operates from both airports but has expressed a strong interest in concentrating its services to Orly.

[11] BVA is about 50 km away from Paris and mainly used by low-cost carriers like Ryanair and Volare.

[12] Linate is a city airport with a travelling time to the city centre of about 15 minutes.

[13] Commission Decisions of 10.03.2003, in Case No COMP/38.477 - BA/SN Brussels Airlines, not published and of 05.07.2002, in Case No COMP/37/730 - AuA / Lufthansa, OJ L 242, 10.09. 2002, p. 25.

[14] See Case No. COMP/M.1328 - KLM/Martinair. See also Commission Decision of 16.01.1996, in Case IV/35.545 - Lufthansa/SAS, OJ L 54, 05.03.1996, p. 28-42.

[15] Commission Decision of 11.08.1999, in Case No COMP/JV.19 - KLM/Alitalia. See also Commission Decision of 10.05.1999, in Case No. IV/M.1506 - Singapore Airlines/Rolls-Royce, OJ C 176, 22.06.1999, p. 11.

[16] A and B-checks are minor checks generally performed at monthly intervals at an aircraft's operational base. A-checks last less than one day while B-checks could take 1 to 2 days approximately. C-checks constitute an intermediate inspection in which one checks the interior of the aircraft and specific structural areas in more detail, generally performed every 18 months. C-checks take 1-2 weeks to perform. Finally, D-checks comprise a thorough inspection, testing and overhaul of an aircraft structure, the systems and the cabins. Such checks are performed every 5-8 years and take 3 weeks to 2 months to be completed.

[17] Case No COMP/37.984 - Air France, Alitalia, Delta [Skyteam Airline Alliance].

[18] Case No COMP/36/111 - KLM/Northwest.

[19] Case No COMP/38.284 - Air France/Alitalia.

[20] Assuming a 40 minute train trip from Paris Gare du Nord to CDG airport, 45 minutes for check-in, a 75 minute flight and 25 minutes travelling time between Amsterdam Centraal and Schiphol Airport by train.

[21] About [55-65 %] of the parties' passengers on this route take connecting flights either in Amsterdam or in Paris.

[22] For instance, in February - March 2004 Basiq Air offers one daily service in each direction between Amsterdam and Marseille in the early afternoon. Air France offers 6 to 7 daily flights in both directions via Paris and Lyon. In the case of Bordeaux - Amsterdam, Basiq Air offers one daily flight in the morning in both directions, while Air France offers four to six daily flights in both directions (weekdays).

[23] For instance, on Lyon-Amsterdam, Air France competed with KLM in spite of the fact that [>90 %] of all transfer passengers have connecting flights in Amsterdam. On Toulouse-Madrid, Air France competes with Iberia and on Lyon-Milan with Alitalia. The latter two city pairs are comparable to Toulouse-Amsterdam in terms of O&D traffic.

[24] In addition, AF has an indirect service, which amounts to about [< 5 %] of total traffic.

[25] As a result of the close co-operation with Air France in particular on the French-Italian market, in summer 2004 Alitalia will stop operating on these city pairs.

[26] Convention of International Civil Aviation signed at Chicago of 7 December 1944.

legislation in issue is compatible with Community law.

[27] The Commission's acceptance of these declarations in no way implies that the national

[28] Long-haul city pairs with less than 30,000 passengers a year are not considered to be economically significant (see case United/US Air). The parties proposed that an appreciable effect for indirect service should be considered when the indirect service is above 3 %. This has also been broadly confirmed by the market investigation.

[29] In general below 30,000 passengers no airline offers a direct service on long-haul routes. On a fail safe basis the Commission has also assessed data submitted by the parties to the US DOJ for 13 transatlantic O&D with an annual number of passengers between 15 000 and 30 000. For six of these routes, the merging parties and their partners have a combined market share in excess of 50 %. It is important to reiterate that none of these routes has a direct service. In four of the six routes (BLQ-NYC, STR-DTT, VCE-BOS and STR-DTT), the current market leader is either Lufthansa or British Airways. In the other two O&Ds (PAR-MSP and OSL-HOU), the merging parties face rivals that currently hold a significant market position. Routes below 15 000 passengers are either one or double connection flights. On these small routes with indirect-indirect overlap and with very volatile market shares it is unlikely that SkyTeam would be dominant.

[30] KLM/Northwest - 13 to JFK and 7 to Newark, Liberty. Delta with 7 to JFK.

[31] Singapore was not allowed to operate more than 4 weekly frequencies.

[32] As a rule of thumb the parties consider that a long haul route has enough passengers to support competing daily direct services if it connects two hubs (i.e. if there are connecting opportunities at both ends of the city-pair) and has traffic of at least 75 000 O&D passengers per annum.

[33] Northwest system timetable, dates 06/01/04 - 11/02/04.

[34] About [< 5 %] in the first half of 2003, and [< 5 %] and [< 5 %] on Amsterdam-Lagos and Paris-Lagos respectively in 2001.

[35] For instance, in the case of Kansas City to Warsaw flights offered by the alliances compete with additional flights offered on the basis of code-sharing between carriers which do not belong to the same alliance. The following options are available on this route: United-LOT, Continental-LOT, USAir-LOT, United-KLM, USAir-KLM, United-Lufthansa, Delta-Air France, Midwest-LOT. Thus, apart from the flights offered by SkyTeam, STAR and Wings (United-LOT, United-Lufthansa, Delta-Air France), there are flights offered by non-aligned airlines with different alliance members (Continental-LOT, USAir-LOT, USAir-KLM, Midwest-LOT) as well as flights offered by members of different alliances (United-KLM).

[36] Case No. COMP/M.1506 - KLM/Martinair.

[37] Market shares calculated in tonnes (source IATA).

[38] Routes to/from Nigeria, South Africa, Egypt, Kenya, Cameroon, Morocco, Brazil, Venezuela, St. Martin, Mexico, China, Hong Kong, South China Sea region, Japan, Thailand and South Korea. The additional overlaps should Martinair be considered are Uganda, Chile and Argentina.

[39] These estimates are based on AEA data, IATA data, Schiphol Airport reports and the estimates of the parties corresponding to the period April 2002 to March 2003.