EC, December 17, 2008, No M.5141

THE COMMISSION OF THE EUROPEAN COMMUNITIES,

Having regard to the Treaty establishing the European Community, Having regard to the Agreement on the European Economic Area, and in particular Article 57 thereof, Having regard to Council Regulation (EC) No 139-2004 of 20 January 2004 on the control of concentrations between undertakings (1, and in particular Article 8(1) thereof, Having regard to the Commission's decision of 8 September 2)008 to initiate proceedings in this case, Having regard to the opinion of the Advisory Committee on Concentrations, Having regard to the final report of the Hearing Officer in this case, Whereas:

1 INTRODUCTION

1. On 17 July 2008, the Commission received a notification of a proposed concentration pursuant to Article 4 of Council Regulation (EC) No 139-2004 (the "EC Merger Regulation") by which the undertaking KLM Royal Dutch Airlines NV ("KLM", the Netherlands), ultimately controlled by Air France-KLM Holding SA ("Air France-KLM", France) acquires control, within the meaning of Article 3(1)(b) of the EC Merger Regulation, of the whole of the undertaking Martinair Holland NV ("Martinair", the Netherlands) by way of purchase of shares.

2. After examination of the notification, the Commission has concluded that the proposed transaction falls within the scope of the EC Merger Regulation and would not significantly impede effective competition in the common market or in a substantial part of it and that it is therefore to be declared compatible with the common market and the functioning of the EEA Agreement pursuant to Article 8(1) of the EC Merger Regulation and Article 57 of the EEA Agreement.

2 THE PARTIES

3. KLM is a full-service network carrier with its hub at Amsterdam Schiphol. KLM is part of the Air France-KLM group. Both KLM and Air France have two core activities: air transport of passengers and cargo. They also provide engineering and maintenance services. KLM has a subsidiary, Transavia.com, which is active in the passenger market as a low-cost carrier.

4. Martinair is a Dutch point-to-point carrier with its base at Amsterdam Schiphol. Martinair provides charter and scheduled air transport of passengers and cargo. Martinair's fleet only serves intercontinental destinations. It operates passenger flights to the following eleven long-haul destinations: Vancouver, Toronto, Miami, Havana, Varadero, San José, Punta Cana, Puerto Plata, Cancun, Curacao and Aruba. (2) The cargo flights operate between Europe and the following regions: Africa, the Middle East, the Far East, North America and Latin America.

3 THE CONCENTRATION

5. KLM currently owns 50% of Martinair, sharing the equity 50/50 with Maersk Holding B.V., a Dutch-based wholly-owned subsidiary of A.P. Moller-Maersk A/S ("Maersk"), a Danish company with interests mainly in sea and land transport. The parties submit that Maersk no longer has a strategic interest in the air transport business and wishes to dispose of its stake in Martinair in order to fully concentrate on the sea shipping industry. Pursuant to a Letter of Intent of 7 April 2008, KLM therefore intends to acquire the remaining 50% shareholding of Maersk in Martinair.

6. According to Martinair's Articles of Association, KLM and Maersk share the voting rights equally in Martinair, as well as the rights to appoint the members of Martinair's management board and supervisory board. Each year, the management board must submit a draft budget to the supervisory board which must be approved by the latter. In addition, the management board needs the approval of the supervisory board for a number of important decisions, including any decisions relating to the acquisition and sale of aircrafts and decisions to take out loans. Operational decisions such as decisions relating to the routes on which Martinair operates, pricing and positioning on the markets are all taken within Martinair. Depending on their importance, management board approval may be required.

7. Given the company organisation and corporate governance of Martinair, KLM and Maersk currently jointly control Martinair. The acquisition of Maersk's 50% shareholding in Martinair by KLM will therefore involve a change from joint to sole control. Accordingly, the proposed transaction constitutes a concentration within the meaning of Article 3(1)(b) of the EC Merger Regulation.

4 COMMUNITY DIMENSION

8. The undertakings concerned have a combined aggregate worldwide turnover in excess of EUR 5 000 million (3), and each of them have a Community-wide turnover in excess of EUR 250 million (4). They do not reach two thirds of their aggregate Community-wide turnover within one and the same Member State. Therefore, the notified concentration has a Community dimension within the meaning of Article 1(2) of the EC Merger Regulation.

5 PROCEDURE

9. Based on the results of the first phase market investigation, the Commission concluded that the proposed transaction raised serious doubts as to its compatibility with the common market due to competition concerns identified with respect to air passenger transport between Amsterdam and Aruba and Curacao. On 18 August 2008, the parties submitted a commitment proposal. The proposed commitment consisted in benchmarking the post-merger price evolution of economy class fares on the routes where competition concerns were identified, namely, the routes connecting Amsterdam to Aruba and Curacao, vis-à-vis the price evolution of economy class fares on a basket of comparable routes. However, the market test of the proposed commitments revealed doubts as to the workability and effectiveness of the proposed price control mechanism and showed that the proposed commitment was too complex to implement and monitor. The proposed commitment therefore proved to be not sufficiently clear-cut to eliminate

10. Accordingly, on 8 September 2008, the Commission concluded that the transaction raised serious doubts as to its compatibility with the common market and with the functioning of the EEA Agreement and decided to initiate proceedings under Article 6(1)(c) of the EC Merger Regulation.

11. The parties submitted comments on the Article 6(1)(c) decision on 15 September 2008.

6 CONCEPTUAL FRAMEWORK FOR THE COMPETITIVE ASSESSMENT

6.1. Introduction

12. KLM and Martinair are both active in air cargo transport and long-haul air passenger transport from Amsterdam to various long-haul destinations worldwide. The proposed transaction results in horizontally affected markets with respect to air cargo transport and certain long-haul routes for scheduled passenger air transport.

13. In view of the fact that KLM already has joint control of Martinair, the proposed transaction raises one general question with respect to the relevant counterfactual for the assessment of its competitive impact as regards both cargo and passenger air transport - the extent to which the competitive interaction between KLM and Martinair will actually change as a result of the proposed transaction. In addition, Air France-KLM's membership in the SkyTeam alliance should be taken into account in assessing the competition effects of the proposed transaction.

6.2. KLM's joint control over Martinair

14. The parties submit that, generally speaking, a change from joint to sole control only has a limited impact on the competitive structure of the relevant markets as the acquiring undertaking already exercises some degree of control (5).

15. Indeed, in several previous cases, the Commission highlighted the fact that the acquirer had joint control prior to the relevant transaction as further evidence of the absence of negative competitive effects resulting from such transactions. (6) However, in other more complex cases, the fact that the acquirer already had joint control did not prevent the Commission from carefully analysing the competitive effects of the transaction. In these cases, the extent to which the parties had been competing pre-merger played an important role in the assessment of the effects of a transaction consisting in a change from joint to sole control. (7) Moreover, in the Commission's previous investigation into KLM's attempt to acquire the remaining 50% shareholding in Martinair, which was notified to the Commission in December 1998, the fact that Martinair was jointly controlled by KLM did not prevent the Commission from raising serious doubts as to the compatibility of the transaction with the common market and initiating proceedings pursuant to Article 6(1)(c) of the EC Merger Regulation.(8)

16. In the present case, in order to properly assess the impact of KLM's acquisition of Maersk's 50% shareholding in Martinair, and the effects of the resulting move from joint to sole control by KLM, it is first necessary to analyse the extent to which KLM and Martinair have constrained each other pre-merger. To this end, the Commission must consider the organisation and corporate governance of Martinair as well as the extent and nature of the competitive interaction between the parties on the markets where their activities overlap.

17. In this respect it must be noted that, according to the parties, contrary to what had been suggested by some competitors in the Commission's investigation of the merger between Air France and KLM, there is currently no exchange of budgets, marketing plans or similar sensitive information between Martinair and KLM, nor is there any coordination of networks to avoid overlap of services. (9)

18. Moreover, the parties have confirmed that KLM has no influence on Martinair's decisions on entering a particular route, business cases with respect to individual routes, increasing or decreasing capacity or frequencies on individual routes or allocating aircraft to such routes. All these decisions are taken independently by internal decision-making bodies within Martinair. In fact, according to Martinair's Articles of Association, KLM's rights enabling it to influence Martinair's market conduct are limited to its rights to nominate members of Martinair's Executive Board and Supervisory Board together with Maersk (10) and its veto rights via the Supervisory Board with regard to more far-reaching, strategic and operational choices including, in particular, strategic choices to operate within certain business segments, such as the short-haul or long-haul passenger business or the cargo segment, and decisions to extend or reduce Martinair's fleet. (11)

19. Martinair's internal documents also confirm that Martinair operates to a large extent independently from KLM and, in particular, that it takes decisions relating to its market conduct, including pricing strategies and entry on new routes, independently from and in competition with KLM. (12) Nevertheless, it must be noted that this operational independence is limited. If KLM were to agree with Maersk on the line to take, KLM could in fact decide upon all aspects of Martinair's market conduct, whether they are of a strategic or operational nature.

20. Finally, the Commission's market investigation largely confirmed that, despite the fact that KLM jointly controls Martinair, KLM and Martinair currently compete with each other on the relevant markets and that the proposed concentration is likely to eliminate the existing pre-merger competition. Indeed, the parties' competitors and customers generally perceive KLM and Martinair as competing with regard to both their passenger and their cargo operations. However, as will be set out in more detail in section 8.4.1, several tour operators indicated that Martinair's competitive strength has recently declined with regard to its passenger air transport operations. In addition, to be able to continue operating its passenger flights beyond the next one or two years, Martinair depends on KLM's agreement on the renewal of its fleet, as this is one of the areas where KLM enjoys veto rights. The competitive constraint that Martinair currently exerts on KLM therefore appears to be significantly more limited with respect to the passenger business than with respect to the cargo business.

21. Moreover, it should be pointed out that KLM's joint control over Martinair reduces, at least to some extent, its incentives to compete with Martinair. Indeed, it can be argued that, while setting its fare strategy, despite having the incentive to maximize profits from its own operations (from which it collects 100% of revenues), KLM would internalise the effect that such a strategy has on Martinair's profits, because it is entitled to half of those profits. However, as regards Martinair's market conduct, the effect of KLM's 50% shareholding is mitigated by Maersk as the other controlling shareholder. It is in Maersk's interest that Martinair fully compete with KLM so that it may maximise its revenues.

22. For the purpose of the assessment of the present case therefore, the two airlines can be regarded, in principle, as competitors. The proposed change from joint to sole control would remove Martinair's current operational independence from KLM and eliminate the existing degree of competition between the two airlines as described in this section, in particular as regards the routes where Air France-KLM and Martinair are currently competing. However, the fact that KLM already has control of Martinair pre-merger clearly reduces the likely anti-competitive effects of the merger. The considerations set out above and explained in more detail with respect to the parties' passenger air transport operations in section 8.4.1 will therefore be taken into account in the competitive assessment of the relevant markets.

6.3. Air France-KLM's Membership in SkyTeam

23. Air France and KLM are both members of the SkyTeam alliance, and have close cooperation agreements both for passengers (SkyTeam alliance (13)) and cargo (SkyTeam Cargo alliance (14)). Martinair is currently not part of SkyTeam and according to the parties there are no plans that Martinair joins SkyTeam post-transaction.

24. As regards cargo, KLM has a 26% shareholding in Kenya Airways and a joint venture with Kenya Airways on the Amsterdam-Nairobi route, under which both companies cooperate extensively (15). On the other routes from Europe to Kenya, cooperation is more limited.

25. In addition, as a result of the 2004 merger of Air France and KLM, the 2008 merger of Delta and Northwest and the forecasted four-way joint venture between these four players on transatlantic flights, the Commission has, on a fail-safe basis for the purpose of the present decision, aggregated their market shares with respect to both passenger and cargo businesses. This is, however, without prejudice to any future assessment of the four-way joint venture under Article 81 of the Treaty. In addition, in light of the close cooperation between Kenya Airways and Air France with respect to the Amsterdam-Nairobi route, the Commission has aggregated, on a fail-safe basis, the market shares of these two players with respect to air transport of cargo on the route between Kenya and Europe16. In any event, whether the markets shares of various partners are aggregated or not does not ultimately affect the assessment of the present transaction.

7 AIR TRANSPORT OF CARGO

7.1. Introduction

26. Air cargo can be transported by three different types of air cargo carriers, namely combination airlines, all-cargo airlines and integrators.

27. Combination airlines are typically network carriers (such as Air France-KLM, Lufthansa or United Airlines) operating scheduled passenger and cargo services. They typically carry cargo in the belly-hold and/or on part of the main deck of the passenger aircraft or on dedicated full freighter aircrafts. All-cargo airlines (such as Cargolux, Nippon Cargo Airlines, Air Bridge Cargo or Atlas/Polar Air Cargo) operate full freighter aircrafts on both scheduled and non-scheduled operations. Integrators (such as DHL, UPS, FedEx or TNT) rely on two types of aircrafts depending on the demand on a particular route. They operate their own aircrafts on routes with high traffic flows (mostly within Europe and between Europe and North America or Asia) and rely on other carriers' extended networks for other destinations. The latter carriers are called integrators because, in addition to airport to airport services, they also provide door-to-door services, that is to say, they handle cargo from the point of origin to the airport and from the airport to the final destination, including legal formalities such as customs clearance.

7.2. The relevant product and geographic markets

28. With respect to the relevant market for cargo, the Commission investigated whether, in light of recent market developments and in particular the increase of fuel prices, intercontinental transport of cargo should include not only the transport of cargo by air but also the transport of cargo by ship, and whether the market for air transport of cargo should further be subdivided according to the nature of the goods transported. Moreover, the Commission investigated to what extent the market for air transport of cargo should be divided on the basis of a point of origin/point of destination approach based on continents or countries respectively.

7.2.1. Substitutability with other means of transport

29. In the past, the Commission has found that for intra-European air cargo transport the relevant product market includes road and/or other modes of transport (17). As regards inter-modal competition for international and in particular intercontinental cargo transport, the parties claim that due to ever-increasing fuel costs and relatively low time-sensitivity of most cargo, sea transport exerts considerable competitive pressure on air transport, resulting in a gradual shift from air transport to sea transport.

30. The market investigation widely confirmed that, given the high fuel prices, the increased speed and reliability of ships, improvements in temperature-controlled equipment, technology and processes for ocean freight and enhanced efficiency of ocean connections with rail and truck, sea transport is increasingly becoming an economically attractive option for intercontinental cargo, especially for non-urgent cargo. However, many types of cargo still require transport by air rather than transport by sea due to their time-sensitive character.

31. Irrespective of the above developments which lead to ocean freight increasingly constraining air cargo transport, it is not necessary for the purpose of this decision to decide whether intercontinental cargo, air and sea transport are part of the same product market, as Air France-KLM and Martinair are only active in cargo air transport and the proposed concentration does not significantly impede effective competition in the common market or in a substantial part of it even on the basis of the narrowest product market definition (namely, a separate market for the intercontinental transport of cargo by air).

7.2.2. Distinction on the basis of the category or nature of the goods transported

32. In previous decisions, the Commission has not conclusively decided whether the market for air cargo transport should be subdivided by category or nature of the products transported (18). The parties submit that the air cargo transport market should not be subdivided further according to the nature of the goods transported, as most cargo is transported using standardised forms of packaging (boxes, pallets, containers), as a result of which the same handling is required to load and unload the cargo. In addition, with the exception of very dangerous goods which would require transport by a full-freighter aircraft, the same aircraft is typically used to transport the different types of cargo.

33. In this regard, the market investigation has confirmed (19) that there is no need for a subdivision of the air cargo market by category or nature of the goods transported, as customers generally buy space which they fill with all sorts of different cargo, including perishables. The special requirements of certain products relate mainly to activities ancillary to air cargo transport such as ground handling in warehouses at airports which may, for instance, require a temperature-controlled room. These services are often outsourced. In addition, for efficiency purposes, carriers need to fill the aircraft with different kinds of cargo (in terms of weight, shape etc.) to make best use of the aircraft space. They can adapt their services relatively easily to the specific requirements of certain goods.

34. Nevertheless, for the purpose of the current decision it can be left open whether a further subdivision is necessary, given that the proposed concentration does not significantly impede effective competition in the common market or in a substantial part of it, irrespective of the exact market definition.

7.2.3. Continent-to-continent/continent-to-country approach

7.2.3.1. All types of cargo

35. The Commission has previously held that the strict "point of origin/point of destination" city-pair approach is inappropriate to delineate the relevant market in air transport of cargo (20).

36. In its previous decisions, the Commission has defined the market for air cargo transport as Europe-wide (including other means of transport such as road transport) for intra-European cargo transport (21) and on a continent-to-continent basis for intercontinental transport of cargo, at least when it concerns continents with a sufficiently developed infrastructure allowing for onward connections (22). The Commission has indicated that, for countries without sufficient infrastructure, a further division on a country-by-country basis may be necessary (23).

37. The market investigation indeed confirmed that a continent-to-continent approach is appropriate for Europe and North America. The situation is, however, different for Africa, South America and part of Asia. According to the majority of respondents, those continents have to be subdivided on a country-by-country basis because of complicated customs formalities, insufficient road networks and restricted air feeder capacity.

38. Furthermore, according to the parties' claims which were confirmed by the market investigation, air cargo transport markets are one-directional in nature as the demand for air cargo transport can differ substantially between the ends of a route, a circumstance which often results in a structural overcapacity on certain routes.

39. In view of the above, the air cargo transport market for the purpose of this decision is to be divided on the basis of one-way continent-to-continent pairs with respect to routes connecting Europe and North America, and on a continent-to-country basis for all other routes.

7.2.3.2. Specific types of cargo

40. Given that the present transaction relates to a merger between two parties having their cargo hub based at the same airport, namely Schiphol airport in Amsterdam, the Commission has investigated whether this could have an impact on competition on any routes connecting to Europe in the sense that, post-transaction, the parties would enjoy a significant competitive advantage, at least with respect to certain kinds of cargo, in particular flowers. According to the parties, and in line with their submission on the geographic market definition encompassing the entire European continent as the origin or destination point, there are no goods for the transport of which there would be no alternative airport to Schiphol airport.

41. In the course of the market investigation some respondents indicated that, due to the time-critical nature of the flower market and the existence of the flower auction in Aalsmeer, close to Amsterdam, there is a traditional concentration of the flower business in the vicinity of Schiphol airport which may, due to its geographic location, make it the preferred option for freighters active in the distribution of flowers. However, the market investigation widely confirmed that, due to the well-developed European highway infrastructure, many cargo airports and, in particular, those of Brussels, Luxembourg and Frankfurt are substitutable to Schiphol airport for all kinds of cargo, including flowers. (24) One respondent indicated a range of around 800 km within which airports were substitutable for the reception and the sending of cargo while another respondent put forward a 12-hour truck range. (25) One of the respondents active in the shipment of flowers indicated that it ships flowers from Ecuador to Europe via airports in Luxembourg, Brussels and Madrid.

42. For the purpose of this decision, it can therefore be concluded that there is no ground to adopt a narrower market definition for the transport of perishable cargo transport, since there are numerous alternative airports to Schiphol for perishable cargo, including flowers.

7.2.3.3. Direct vs. indirect services

43. According to previous decisions (26), cargo is by definition less time-sensitive than passengers and thus may be routed with a higher number of stopovers as a result of which any indirect route is substitutable to any direct route. The market investigation in the present case confirmed that, except for extremely time-sensitive perishables and goods sensitive to multiple handling, for which direct flights appear to be more suitable as the total transport time should be as short as possible, indirect flights are generally substitutable to direct flights.

7.2.3.4. Conclusion on the relevant market

44. In light of the above, the relevant markets for the purpose of the present decision are the markets for air transport of general cargo on a uni-directional basis between (i) continents for transport of cargo between Europe and North America and (ii) Europe and specific countries for transport of cargo between Europe and South-American, Greater Caribbean, Far Eastern and African countries. In this respect, it is not necessary to further delineate the relevant market according to the nature of the product transported as the proposed concentration would not significantly impede effective competition in the common market or in a substantial part of it, irrespective of the exact market definition.

7.3. General market conditions

Multitude of competitors

45. According to the parties, despite differences in their business models, all three types of air cargo carriers, that is to say combination airlines, all-cargo airlines and integrators, compete for business with the same customers. Although integrators used to focus on the overnight delivery of mail and smaller parcels, they currently have the ability to carry all kinds of cargo without any weight limitation. While all-cargo airlines have a tendency to carry general cargo and scheduled combination airlines tend to carry more time-sensitive goods, all types of carriers compete on the market for air cargo transport.

46. The market investigation confirmed that the cargo market is generally competitive in the sense that it is characterised by a multitude of players competing for business. The market investigation widely confirmed the parties' view indicating that all three different types of air cargo carriers offer the same services to the same customers and many of them are involved in air cargo transport in multiple ways. Besides, the boundaries between the different types of air cargo carriers are becoming increasingly blurred. Some respondents to the market investigation indicated that competition will become even tougher post-transaction. They expect prices to decrease as a result of the transaction.

47. Furthermore, although it has been left open whether air and sea transport are part of the same product market for intercontinental cargo, the market investigation clearly showed that sea transport exerts a competitive constraint on air transport. The parties therefore face competition not only from other air cargo carriers, but also, to some extent, from sea freight operators.

Low barriers to entry and expansion

48. The barriers to entry on the relevant markets are relatively low. As for regulatory barriers, the parties stress that although air cargo transport services require slots, as opposed to passenger services, they do not require slots for a precise schedule. On the contrary, cargo operations can easily take place at night. In addition, while Schiphol airport is congested during peak hours, the other airports concerned by the transaction are not (or are only to a very limited extent) congested.

49. The market investigation also confirmed that cargo transport is 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. (27) In the course of the market investigation, customers expressed their readiness to switch to a different supplier should the parties decide to increase their prices.

50. As far as barriers to expansion are concerned, the market investigation has not identified any. Most competitors have indicated that they have sufficient flexibility to increase or decrease the number of weekly freighter operations and to adapt their freight capacity. (28)

Countervailing buyer power

51. With the exception of large corporate customers, air carriers rarely deal with end customers (or "shippers", namely, the customers that have the goods shipped). Rather, they provide services to shippers through freight forwarders.

52. The market investigation confirmed the parties' claim that cargo customers, especially the major freight forwarders (such as DHL, Panalpina, DB Schencker, and Expeditors), are able to exert a certain degree of countervailing buyer power vis-à-vis air cargo carriers, given their size and importance. (29) Over the last twenty years, freight forwarders have gone through a process of concentration resulting in a handful of worldwide players with large bargaining power. According to the parties, the freight forwarders' pressure on the market is two-fold:

(i) as customers, freight forwarders have gradually consolidated in the past, thereby gaining buyer power on these markets, while shippers have increasingly asked for "turn key" solutions.

(ii) in a strategic move, freight forwarders have integrated upstream, thus becoming competitors. Freight forwarders such as DHL or TNT acquired their own dedicated cargo aircraft fleet, thus putting pressure on "dry" air cargo service providers such as the parties.

53. Air cargo customers' buyer power constitutes an additional factor showing that the proposed transaction does not significantly impede effective competition in the common market or in a substantial part of it with respect to the markets for air cargo transport.

7.4. Competitive assessment

54. According to the parties' claims confirmed by internal documents, the acquisition of Martinair's cargo business is the strategic key driver of the transaction for KLM (30).

55. At the outset, it should be noted that KLM and Martinair are not close competitors in the supply of air cargo services as they pursue different business models in this respect. Martinair mainly provides air cargo services with 11 dedicated full freighter aircrafts and has a belly capacity on passenger flights representing less than 5% of its total cargo capacity. KLM, on the other hand, relies mainly on capacity in the belly-hold of its passenger aircrafts and has only 3 full freighter aircrafts based in Schiphol (Air France has 10 full freighter aircrafts based at its cargo hub at Roissy Charles de Gaulle). Therefore KLM's cargo business is, to a considerable extent, driven by its passenger business, which determines both available cargo capacity and destinations. (31)

56. Taking into account Air France's market share, the operation will result in the creation of the second largest air cargo operator worldwide (32). Routes between Europe and North America and Europe and several African, Caribbean and Far Eastern countries are affected by the present transaction. In the following recitals the effects of the proposed transaction will be assessed individually with regard to each of the affected routes.

7.4.1. Europe/North America

57. As regards international air cargo transport between Europe and North-America, the parties' combined market shares will remain below 25% and the parties will continue to face strong competition from several other major combination airlines, such as Lufthansa, British Airways and integrators such as FedEx and UPS.

<table>

58. Moreover, as was indicated above, barriers to entry and expansion are not high: there is a certain degree of countervailing buyer power; and KLM and Martinair are not close competitors as they pursue different business models with regard to cargo.

59. Finally, as set out in section 6.2, the fact that KLM jointly controls Martinair pre-transaction is likely to reduce the anti-competitive effects of the proposed concentration.

60. From all these elements, it must be concluded that the proposed concentration would not significantly impede effective competition in the common market or in a substantial part of it with respect to the markets for air cargo transport from Europe to North America and from North America to Europe.

7.4.2. Europe/South American countries

61. With respect to air cargo transport between Europe and South America, the transaction will lead to high market shares and a significant increment of market shares on the following routes as follows: Europe to Colombia [50-60]*% (Air France-KLM [10-20]*%, Martinair [30-40]*%), Colombia to Europe [60-70]*% (Air France-KLM [10-20]*%, Martinair [50-60]*%), Peru to Europe [40-50]*% (KLM [20-30]*%, Martinair [10-20]*%), Europe to Peru [40-50]*% (KLM [40-50]*%, Martinair [5-10]*%), and to a lesser extent on the routes from Ecuador to Europe [30-40]*% (KLM [5-10]*%,

Martinair [20-30]*%) and Europe to Ecuador [30-40]*% (KLM [20-30]*%, Martinair [10-20]*%).

<table>

62. The parties claim that their market shares, as reported through IATA, are exaggerated since they not take into account significant indirect services to and from Europe provided by competing North American carriers via U.S. airports (Miami and Houston for example). According to the parties, IATA data is also misleading as it does not include non-scheduled flights.

63. The market investigation (35) confirmed that IATA data needs to be adjusted for South American countries, as the market coverage of IATA is quite low in that continent as only a few airlines and/or forwarders participate in IATA's Cargo Account Settlement System which is the basis for detailed statistics. The market investigation also confirmed that IATA data does not cover non-scheduled flights. One respondent estimated that approximately 20 % of cargo coming from Ecuador to Europe is carried via airports in the U.S. and thus counted as cargo transported from North America to Europe by IATA, although it actually comes from Ecuador. Estimates by other respondents of the percentage of cargo from Latin America to Europe that is not covered by IATA figures reach up to 40%. The market investigation also confirms that the IATA data are the only source of information. This makes it impossible to adjust the market shares given in section 7.4.2 in any meaningful way. Taking into account that the parties are among the airlines which do give detailed data to IATA, it is concluded from the above elements that the market shares given in section 7.4.2 are likely to overstate the parties' position on the relevant markets.

64. On routes where the parties will have high market shares post-transaction, they will continue to face several other competitors. On the route from Peru to Europe and from Europe to Peru, the parties' main competitor is Iberia with a market share of [20-30]*% and [20-30]*% respectively. On the route from Ecuador to Europe the parties' main competitor was FemkeFly with a market share of [20-30]*% (36). On the route from Europe to Ecuador, Iberia has a market share of [10-20]*%. On the route from Europe to Colombia, the parties face competition from Cargolux ([0-5]*%) and Iberia ([10-20]*%). On the route from Colombia to Europe, Iberia's market share is [5-10]* % and Cargolux' market share is [5-10]*%.

65. The market investigation indicates that barriers to entry on the routes from Europe to South American countries and from South American countries to Europe are relatively low. (37) Although there are traffic right restrictions for flights to Colombia, Peru and Ecuador, the parties claim that they do not constitute a serious obstacle since the respective governments adopt a liberal attitude to (new) charter operations. A respondent to the market investigation pointed out that traffic right restrictions do not apply to carriers with their home base in those countries.

66. Several new players have recently entered various routes connecting Europe to South America, in particular Cargo B based in Brussels, which flies inter alia to Ecuador and Colombia. Lan Chile has started to operate cargo flights from Amsterdam to Ecuador.

FemkeFly, an airline from Panama, was also operating on the routes between Ecuador and Europe. (38)

67. Some of the parties' competitors expressly indicated in the market investigation that they have extra capacity to take up additional demand should the parties decide to increase their prices. Besides, most markets are characterised by overcapacity. In addition, according to the parties, the market is expected to grow by 4%-6% over the next 20 years. In light of the above, any price increase by the merged entity would necessarily result in a loss of business.

68. Moreover, there is a certain degree of countervailing buyer power and KLM and Martinair are not close competitors as they pursue a different business model with regard to cargo.

69. Finally, as set out in section 6.2, the fact that KLM jointly controls Martinair pre-transaction is likely to reduce the anti-competitive effects of the proposed concentration.

70. From all these elements, it must be concluded that the proposed concentration would not significantly impede effective competition in the common market or a substantial part of it with respect to the markets for air cargo transport on the routes between Europe and South American countries.

No competition concerns even in case of a separate market for perishables

71. Despite the fact that, on the routes connecting Colombia, Peru and Ecuador to Europe, perishables make up the largest share of traffic, the competitive assessment would not change, even if a separate market for perishables would need to be defined. As previously noted in recital 41 in the context of the market definition, the market investigation confirmed that all airports in Europe are substitutable for all types of cargo. Even for perishables, there are many alternatives to Schiphol airport. Some respondents to the market investigation indicated that they regard Schiphol as the preferred airport for the transport of perishables given the traditional concentration of the flower business near this airport, but the majority of them admitted that Schiphol is not irreplaceable. (39) Therefore, the fact that both parties have their cargo hub at Schiphol airport would not alter the conclusion that the proposed merger does not significantly impede effective competition in the common market or a substantial part of it, even if there were a separate market for the transport of perishables on these routes.

7.4.3. Europe/Greater Caribbean countries

72. As concerns the routes connecting Europe to some Greater Caribbean countries, the parties market shares will be high on the routes from Europe to Mexico where the parties' combined market shares is [50-60]*% (Air France-KLM [30-40]*%, Martinair [10-20]*), Mexico to Europe with a combined market share of [50-60]*% (Air France-KLM [40-50]*%, Martinair [5-10]*%) and Europe to Cuba where the parties reach a [40-50]*% market share (Air France-KLM [40-50]*%, Martinair [5-10]*%).

<table>

73. For the same reasons as set out in recital 63 with respect to South America, IATA data is likely to overstate the parties' position on the Greater Caribbean markets. The market investigation confirmed that IATA coverage is low in Greater Caribbean countries and that IATA data does not include indirect flights and non-scheduled flights. (40)

74. On the above mentioned routes, competitors are in a position to substantially constrain the competitive position of the parties. On the route from Europe to Cuba and from Cuba to Europe, the parties face competition from Iberia ([30-40]* %) and Virgin Atlantic ([10-20]*%). On the route from Europe to Mexico, Cargolux has a market share of [20-30]* % and Iberia has a market share of [10-20]* %. On the route from Mexico to Europe, the parties' competitors include Cargolux ([10-20]* %), Lufthansa ([5-10]* %) and Iberia ([5-10]* %).

75. Although the combined market shares of the parties will be high for routes connecting Europe to the Netherlands Antilles, the increment brought about by the present transaction is rather limited ([0-5]*%). It should be pointed out that there are no significant cargo flows from the Caribbean to Europe as there is hardly any industry with airport potential on the islands. The Caribbean is a major tourist attraction which is mainly serviced by passenger aircrafts with limited belly capacity, as most capacity is used for passengers' luggage. KLM indicates that a large part of KLM cargo traffic is actually from outside the Caribbean (such as flowers from Colombia and Ecuador that are brought to Curacao and Aruba and then brought to Amsterdam on KLM flights). (41)

76. Although the table indicates that the parties' combined market share would also be substantial for the routes connecting Europe to Surinam, Martinair has recently terminated its air cargo transport operations on this route.

77. For the same reasons as set out in recitals 48-50, barriers to entry on the routes from Europe to the Greater Caribbean countries and from the Greater Caribbean countries to Europe are relatively low. New entries during the last five years have taken place, such as the entry of Lufthansa on the route to/from Mexico.

78. Some of the parties' competitors indicated that they have extra capacity to take up additional demand.

79. Moreover, as was indicated above, there is a certain degree of countervailing buyer power, and KLM and Martinair are not close competitors as they pursue different business models with regard to cargo.

80. Finally, as set out in section 6.2, the fact that KLM jointly controls Martinair pre-transaction is likely to reduce the anti-competitive effects of the proposed concentration.

81. From all these elements, it must be concluded that the proposed concentration would not significantly impede effective competition in the common market or in a substantial part of it with respect to the markets for air cargo transport on the routes between Europe and countries in the Greater Caribbean.

No competition concerns even in the case of a separate market for perishables

82. Perishables only account for a very small percentage of total traffic from Europe to Greater Caribbean countries (less than 5%). In contrast, cargo transported from Greater Caribbean countries to Europe mainly consists of perishable goods. Exceptions are Mexico and Cuba which play hardly any role in the perishable markets. As stated above, for the transport of perishables from Greater Caribbean countries to Europe, there are many alternative airports to Schiphol airport. Therefore, the fact that both parties have their cargo hub at Schiphol airport would not change the conclusion that the proposed merger does not significantly impede effective competition in the common market or in a substantial part of it with respect to the markets for air cargo transport on the routes between Europe and Greater Caribbean countries, irrespective of the precise market definition.

7.4.4. Europe/African countries

83. With respect to routes connecting Europe to African countries, the parties' market shares will be high on the routes from Europe to Kenya where the parties' combined market share is [50-60]*% (Air France-KLM/Kenya Airways [30-40]*%, Martinair [10-20]*%), Kenya to Europe with a combined market share of [30-40]*% (Air France-KLM/Kenya Airways [20-30]*%, Martinair [10-20]*%) and Europe to Sudan where the parties reach a combined market share of [60-70]*% (KLM [10-20]*%, Martinair [40-50]*%).

<table>

84. As for South American and Greater Caribbean countries, the parties claim that IATA data does not give a realistic picture of the market situation in African countries since major carriers like Avient and Africa West do not report their data. The parties also point out that Emirates Airlines and Etihad carry cargo from Africa to Europe via Dubai and Abu Dhabi, which is counted as Middle Eastern cargo, even though it originates in Africa.

85. The market investigation confirmed that, for African countries, IATA data is likely to overstate the merged entity's market shares and needs to be adjusted since it does not account for cargo transported via indirect routes nor for cargo transported by non-scheduled airlines. One respondent indicated that for transport from South Africa, IATA data only covers 50% of cargo effectively transported from this country. (42)

86. Post-transaction, the parties will face strong competitors. On the route from Kenya to Europe, the parties' competitors include, among others, Cargolux ([20-30]*%) and Lufthansa ([10-20]*%). On the route from Europe to Kenya, the parties' main competitor is British Airways ([10-20]*%). For the route from Europe to Sudan, the parties face strong competition from Lufthansa which, according to the parties, has a market share of [50-60]*% on direct flights to Sudan. The carriers from the Gulf (namely Emirates, Etihad and Qatar Airways) also exert a competitive constraint in that they transport cargo from Europe to Sudan via Dubai, Abu Dhabi and Doha.

87. For the same reasons as set out in recitals 48-50, barriers to entry on the routes from Europe to African countries and from African countries to Europe are relatively low. One respondent to the market investigation even claimed that, for cargo carriers operating to and from Africa, there are no barriers to entry at all.

88. The market investigation indicated that there have been several new entries in the last five years. For the assessment of this case, the entry of SN Brussels Airlines and Cargolux on the route Europe to/from Kenya is especially important.

89. Some of the parties' competitors also indicated that they have extra capacity to take up additional demand.

90. Moreover, as was indicated in recitals 51-53, there is a certain degree of countervailing buyer power, and KLM and Martinair are not close competitors as they pursue different business models with regard to cargo.

91. Finally, as set out in section 6.2, the fact that KLM jointly controls Martinair pre-transaction is likely to reduce the anti-competitive effects of the proposed concentration.

92. From all these elements, it must be concluded that the proposed concentration would not significantly impede effective competition in the common market or in a substantial part of it with respect to the markets for air cargo transport on the routes between Europe and African countries.

No competition concerns even in the case of a separate market for perishables

93. The parties' activities overlap with regard to the transport of air cargo between Europe and South Africa, Europe and Kenya, Europe and Uganda as well as Europe and Sudan. With regard to a potential separate market for perishables, the parties' overlaps on the routes between Europe and Uganda and Europe and Sudan will not significantly impede effective competition in the common market or in a substantial part as Uganda and Sudan do not export perishables to any significant extent. Sudan's main export products are oil and cotton, whereas Uganda's main export products are coffee and tea. Since Martinair does not ship perishables from South Africa to Europe, even in the case of a further subdivision of the cargo market into a separate market for perishables, the proposed concentration will not significantly impede effective competition in the common market or in a substantial part of it with respect to the route from South Africa to Europe, either. Regarding the transport of perishables from Kenya to Europe, the parties face competitive constraints given that there are many alternative airports to Schiphol airport.

7.4.5. Europe/Far Eastern countries

94. With respect to international air cargo transport between Europe and Far Eastern countries, the parties' combined market shares will reach a maximum of 25% and the parties will continue to face strong competition from several other carriers, such as Lufthansa, Cargolux, Cathay Pacific and Air China.

<table>

95. Furthermore, there have been numerous new entrants on the routes Europe to/from countries in the Far East. Such entrants include Jade Air, Yangtze River Express, Great Wall, Etihad and Qatar Airways. Other carriers, like Emirates, have substantially increased their presence in these markets in recent years.

96. Moreover, as was indicated in recitals 51-53, there is a certain degree of countervailing buyer power, and KLM and Martinair are not close competitors as they pursue different business models with regard to cargo.

97. Finally, as set out in section 6.2, the fact that KLM controls Martinair pre-transaction is likely to reduce the anti-competitive effects of the proposed concentration.

98. From all these elements, it must be concluded that the proposed concentration would not significantly impede effective competition in the common market or in a substantial part of it with respect to the routes connecting Europe with the various countries in the Far East.

7.4.6. Conclusion

99. In light of the above and, in particular, of the fact that the parties will continue to face significant competition from other air cargo carriers; that barriers to entry and expansion in these markets are relatively low; that large freight forwarders have buyer power; that KLM and Martinair are not close competitors by virtue of their different business models; and that KLM currently controls Martinair, any post-merger price increase by the parties is likely to be unprofitable. In addition, the market investigation did not bring to light any serious concerns with regard to the impact of the proposed transaction on any of the affected cargo markets. For all these reasons, it must be concluded that the proposed concentration would not significantly impede effective competition in the common market or in a substantial part of it with regard to air transport of cargo.

8 AIR TRANSPORT OF PASSENGERS

8.1. Introduction

100. Martinair is essentially a point-to-point carrier which is active in air transport services to leisure passengers. It performs both scheduled and full charter operations; the latter have been discontinued as of October 2008. Following its restructuring in 2007, it is active only on a number of long-haul routes which are essentially holiday destinations.

101. Accordingly, KLM's and Martinair's services overlap only with respect to a number of long-haul routes out of Amsterdam. In winter 2008/2009, both Air France-KLM and Martinair will offer air transport services to passengers on the following eight routes: Amsterdam-Curacao; Amsterdam-Aruba; Amsterdam-Vancouver; Amsterdam-Toronto; Amsterdam-Punta Cana; Amsterdam-Miami; Amsterdam-Cancun; and Amsterdam-Havana (43) (referred to together as "the overlapping routes" or "the affected routes").

102. Martinair operates direct flights to all of these destinations. KLM operates direct flights to Aruba, Curacao, Vancouver and Toronto and offers indirect flights through Air France's Charles de Gaulle/Roissy hub to Punta Cana, Miami, Cancun and Havana.

8.2. The investigation and the survey conducted at Schiphol airport

103. In order to assess the effects of the proposed transaction, apart from using the traditional investigative tools (namely, questionnaires to market participants, follow up teleconferences, and the like), in its second phase investigation, the Commission commissioned a customer survey which was conducted at Schiphol airport in Amsterdam. The aim of the survey was to clarify the relevant product and geographic markets through analysis of the behaviour (in particular their price and destination sensitiveness) of Dutch travellers flying to the destinations affected by the present transaction. The survey was conducted at Schiphol airport by an independent contractor (selected through a tender procedure) between 26 September and 5 October 2008.

104. The Commission's first phase market investigation had already provided sufficiently clear indications that the proposed concentration would not significantly impede effective competition with respect to the parties' overlapping passenger air transport activities on the routes connecting Amsterdam to Vancouver, Toronto and Miami. Therefore, these three destinations were not covered by the customer survey. By contrast, for the routes connecting Curacao and Aruba to Amsterdam, the Commission identified serious doubts during the first phase investigation and, with respect to the routes connecting Amsterdam or Amsterdam, Düsseldorf and Brussels to Punta Cana, Cancun and Havana, the Commission identified a need to further investigate the effects of the proposed concentration during its second phase market investigation

105. Accordingly, passengers to six destinations in the Caribbean region, namely Aruba, Curacao, Punta Cana, Cancun, Havana and Paramaribo using services of four carriers, namely KLM, Martinair, ArkeFly and Surinam Airways, were surveyed, which resulted in a total number of 1 005 responses. Two flights per carrier per route were surveyed. To the extent possible, in order to increase the representative value of the sample, one weekend flight and one week day flight were surveyed with respect to each carrier. The resulting sample is large enough to allow statistical inference. However, it should be noted that with respect to a certain type of questions, designed to address only a sub-sample of the respondents, and at a certain degree of disaggregation, the sample may be too small to achieve statistically significant results with standard parametric analysis. (44)

106. Bearing in mind the caveat identified above with respect to some questions, overall the sample appears to be representative of the population of passengers flying from Schiphol airport to the Caribbean area. Indeed, most of the respondents are residents of the Netherlands and travel for leisure purposes. The sample is also representative of both travellers purchasing unbundled flights and travellers purchasing package holidays, the two groups accounting each for approximately half of the sample.

107. As already indicated, one of the main aims of the survey was to clarify the relevant product and geographic market with respect to the overlapping routes. It aimed at identifying the relevant airport catchment area at the point of origin, in particular whether the airports in Brussels and Düsseldorf belong to the same catchment area as Amsterdam airport, as well as the substitutability between various long-haul holiday destinations in the eyes of Dutch leisure travellers. (45)

8.3. The relevant product and geographic market

108. The parties submit that, for the purpose of the assessment of the present transaction, the relevant market should be defined as the market for air transport of Dutch leisure travellers departing from Amsterdam, Brussels and Düsseldorf to long-haul (holiday) destinations, comprising direct and indirect flights as well as flights sold separately and flights sold as part of package holidays (46). According to the parties, there is only one market for the sale of seats to end customers and the wholesale supply of seats to tour operators ("TOs") (the latter to be bundled in a package holiday). Alternatively, the parties propose to define a market for air transport of Dutch leisure travellers to the Greater Caribbean region from the airports in Amsterdam, Brussels and Düsseldorf.

109. It was also suggested by the parties in the present case that unbundled seats offered by charter airlines or tour operators constrain the parties, in particular Martinair, and should thus be included in the relevant market. For the purpose of the present decision, a charter airline is defined as any passenger airline that operates non-scheduled flights (exclusively or not).

110. The following recitals analyse (i) the extent to which the alleged constraint from charter airlines and tour operators is relevant to the assessment of this case in the context of the definition of the relevant market and the competitive assessment (see recitals 111-115); (ii) the elements of the Commission's investigation in the present case on the basis of which it must be concluded that the sale of seats to end customers and the wholesale supply of seats to TOs constitute separate markets (section 8.3.1). Subsequently, the exact market definitions with respect to the sale of seats to end customers (section 8.3.2) and the wholesale supply of seats to TOs (section 8.3.3) will be discussed.

Competitive constraint from charter airlines and tour operators

111. As regards the competitive constraint on scheduled operations from charter airlines/tour operators selling "seats only" (unbundled seats) to end customers, the Commission left this issue open in its Ryanair/Aer Lingus decision (47).

112. With respect to sales of seats by tour operators, the market investigation showed that TOs sell, to different extents, seats that they purchase from carriers, mainly KLM and Martinair, as unbundled airline tickets. However, these sales are mainly performed by TOs in their role as travel agents in which case they act in the name and on behalf of the carrier against a standard commission. With respect to these seats, TOs do not have the ability to set the price (48). In addition, to a very limited extent TOs also sell seats where they can determine their own margin. These sales, however, concern only the resale of outstanding guaranteed seats purchased in advance by large TOs that they did not manage to sell as part of a package (49). In the course of the market investigation, several TOs indicated that they normally try to avoid selling large numbers of seats as "seats only". Instead, whenever possible, they still try to include left-over seats in a package and sell the package for a last-minute reduced price. Given the limited number of seats which are sold by tour operators as "seats only" and the fact that they might actually have to sell such seats at a loss, these sales cannot therefore exert much competitive pressure on direct sales by the airlines themselves.

113. With respect to the sale of seats by charter/non-scheduled airlines, the only airline falling under this category flying to long-haul destinations from Schiphol airport is ArkeFly, belonging to the TUI group. ArkeFly has non-scheduled operations to all of its long-haul destinations except Aruba and Curacao where it operates a scheduled service. Most of its capacity (even with respect to scheduled flights) is, however, captive, allocated to TUI's downstream tour operating activities (50).

114. Accordingly, even if the Commission were to admit that, in principle, unbundled seats sold by tour operators or charter airlines exert competitive pressure on seats sold by airlines themselves as suggested by the parties (51), this is of little relevance for the case at hand as, with the exception of a limited amount of seats sold by ArkeFly (and TUI group TOs) on those affected routes where ArkeFly does not operate scheduled services (essentially only the Amsterdam-Punta Cana and Amsterdam-Cancun routes) (52), all seats sold by Dutch tour operators are seats purchased from the parties. Even if, in light of the market definition with respect to the point of origin adopted in the present decision, seats sold by airlines on long-haul flights out of Brussels and Düsseldorf are taken into account, the distinction between scheduled and charter airlines has become blurred over recent years as even full-leisure airlines, such as ArkeFly, LTU and Jetairfly operate scheduled services to some destinations nowadays.

115. In light of the facts of the case at hand, in particular the limited sale of seats by non-scheduled/charter airlines, it is not necessary to conclude on this issue, as in any event, the proposed concentration would not impede effective competition in the common market or a substantial part of it.

8.3.1. Wholesale supply of seats to tour operators v. sales to end customers

116. In its previous decisions, the Commission regarded the supply of airline seats to tour operators (that is, the "wholesale market") as a market distinct from the market for supply of scheduled air transport services to end customers (53). In the Ryanair/Aer Lingus decision, the Commission concluded that the wholesale supply of seats to tour operators is not in the same market as scheduled air transport services for end customers as, from a demand-side perspective, market conditions on the market for sales of seats to tour operators differ significantly from those on the market for sales to individuals. In this respect, the Commission noted that "the market for sales of seats to tour operators is a market which is upstream to the market for sales of seats to individuals. Accordingly, the competitive conditions in this market are manifestly different since tour operators have different requirements than individual customers (for example, buying of large seat packages, negotiation of rebates, taking into account of customers' needs in terms of flight times etc.)". (54)

117. With respect to the case at hand, both the first phase and the second phase market investigation as well as the results of the customer survey broadly confirmed the Commission's findings in the Ryanair/Aer Lingus decision which distinguishes separate markets for the supply of passenger air transport services to end customers and the wholesale supply of airline seats to tour operators (TOs) (to be bundled in a holiday package). This is due to the fact that, from the demand side perspective, according to the results of the customer survey, end customers do not consider package holidays and independent holidays as substitutable, contrary to the parties' claim that, from the customers' perspective, unbundled flights and package holidays are part of the same market, and that there is therefore one market comprising both the sale of seats to end customers and the sale of seats to TOs.

118. According to the parties, the majority of the parties' passengers on the affected routes are on a holiday trip and therefore, for a significant part of these customers, the fact of booking a package holiday with a tour operator is a substitute to booking a flight and accommodation separately, if the package turns out to be cheaper overall. Conversely, there would be an indirect constraint from end customers who increasingly buy components of their holidays separately, rather than buying a holiday package. Therefore, according to the parties, package holidays exert competitive pressure on unbundled flights and vice versa.

119. The Commission's first phase market investigation indicated that according to tour operators, package holidays and unbundled tickets could be, at least to some extent, substitutable for typical leisure passengers in the case of a price increase of flights with respect to the overlapping routes. However, some tour operators also indicated that package holiday customers tend to stick to packages because of the care-free nature of these type of holidays as compared to independent holidays. The market investigations also revealed that travellers visiting friends and relatives (so-called "VFR" travellers) (55) would generally not consider switching between a ticket only and a package holiday.

120. In addition, the customer survey conducted in the second phase investigation showed that the final consumer actually does not consider unbundled flights and package holidays as substitutable and that, therefore, unbundled flights and package holidays do not belong to the same product market. Indeed, only 7.4% of the customers who replied to the relevant question and who had bought an unbundled flight would consider buying a package (see Graph 1 in this recital). This is because the majority of customers purchasing unbundled flights cite the independence and the flexibility that unbundled flights offer as the reason for their choice. Similarly, a relatively small proportion of customers purchasing packages (17%) indicated that they would switch to unbundled flights. This appears to be mainly due to the comfort and the safety guaranteed by package holidays organised by tour operators. Thus, for the vast majority of passengers in both groups of customers, the decision to buy either a package tour or an unbundled flight depends mainly on the specific characteristics of the product rather than on pure price considerations.

Graph 1: Composition of the sample of unbundled flight buyers by whether they considered buying a package tour.

121. For the reasons stated in this section, in line with the existing precedents, it must therefore be concluded that there are two separate markets for (1) the supply of passenger air transport services to end customers, and (2) the wholesale supply of airline seats to tour operators.

8.3.2. The supply of passenger air transport services to end customers

8.3.2.1. Point of origin/point of destination ("O&D") pairs traditional approach

122. In the present case, both KLM and Martinair operate flights out of Schiphol airport to various long-haul, mainly leisure, destinations.

123. In the past, the Commission has defined the relevant market for passenger air transport services on the basis of the O&D city pair approach (56) because from the end customers' viewpoint (wishing to go from point A to point B), only the combination of the same point of origin and the same point of destination are substitutable.

124. In this respect, KLM claims that as the parties with respect to the affected routes transport mainly leisure passengers to holiday destinations, the O&D approach is inappropriate in this case as it fails to take into account the competitive pressure from (i) package holidays offered by tour operators in the Dutch market, (ii) flights aimed at leisure passengers departing from neighbouring airports (Düsseldorf and Brussels) and (iii) long-haul flights aimed at leisure passengers from Amsterdam to other holiday destinations than those where the parties overlap.

125. The issue of the alleged competitive constraint from package holidays was discussed in recitals 119-125. As discussed in these recitals, the market investigation confirmed the existence of a separate market for the wholesale supply of airline seats to tour operators. With respect to the second argument, namely the inclusion of the Brussels and Düsseldorf airports in the point of origin, the Commission investigated the relevant airport catchment area with respect to Dutch leisure passengers travelling to long-haul holiday destinations ("point of origin substitutability") (see section 8.3.2.2.). The Commission has also investigated the third argument raised by the parties, namely the alleged pressure exercised by alternative holiday destinations ("point of destination substitutability") and their possible inclusion in the relevant market (section 8.3.2.3).

8.3.2.2. Point of origin substitutability

126. The parties take the view that, with respect to the overlapping destinations, they face competition from flights from Düsseldorf and Brussels to the extent that there are alternative flights. According to KLM, these airports are close to the Dutch border and for a large group of Dutch passengers, it will take less time to travel there than to Amsterdam. Moreover, the parties submitted internal studies of the impact of the introduction of the "vliegtaks" (a Dutch government tax collected on environmental grounds, amounting to up to EUR 45 for flights to long-haul destinations departing from the Netherlands) suggesting that its introduction has resulted in an increase of bookings by Dutch passengers of flights out of Belgian and German airports.

127. Previous Commission decisions have recognised that flights from or to airports which have overlapping catchment areas can be regarded as substitutes (57). Such airport substitution has often been accepted for several airports located in the same city, such as the Milan-Linate and Milan-Malpensa airports or the JFK and Newark airports in New York. (58) Moreover, in recent cases, the Commission has considered distances and travelling times of 100 km/1 hour as a benchmark for an airport catchment area. The Commission in its Ryanair/Aer Lingus decision noted, however, that the airport catchment area may be broader in the case of routes with a high percentage of leisure passengers due to the fact that leisure passengers are usually more price sensitive and less time constrained and are thus willing to travel further distances to board a flight if the overall price is lower than for flights from their local airports. The Commission stated that leisure passengers "are more inclined to trade a longer total journey time in exchange for a lower total journey cost [...] and it is reasonable to assume that airport substitutability from the standpoint of passengers is wider rather than narrower". This finding (in a case concerning short-haul routes) may be even more relevant in the present case concerning long-haul routes leisure destinations.

128. With respect to Brussels airport, the neighbouring Dutch regions are Zeeland, Noord-Brabant and Limburg. These regions have a combined population of around 3.9 million, which represents around 24% of the total population in the Netherlands. The average distance from the centre of these regions to Brussels airport is 102 km for Zeeland (Terneuzen), 115 km for Noord-Brabant (Tilburg) and 147 km for Limburg (Roermond) compared to the average distance of these regions to Schiphol airport amounting to 222km, 127 km and 189 km for Zeeland, Noord-Brabant and Limburg respectively. Also, travel time by car from these three regions to Brussels airport is shorter and the costs of car transport are lower than to Amsterdam airport. Travel time by train to Brussels airport is shorter or equal for Zeeland and Limburg compared to Amsterdam airport and travel time by train is slightly longer for Noord-Brabant (approximately 45 minutes). At the same time, travel costs by train from the three regions to Brussels airport are lower or equal to travel costs by train to Amsterdam airport. Accordingly, a large part of these regions is actually closer to Brussels airport than to the airport of Amsterdam and thus Dutch passengers living in these regions can be expected to be at least equally inclined to travel out of Brussels airport than to travel out of Schiphol.

129. Concerning the airport in Düsseldorf, the situation is analogous to that of Brussels. The Dutch neighbouring regions to Germany from which Dutch passengers would potentially travel to Düsseldorf for the purpose of boarding a flight include Limburg, Noord-Brabant and Gelderland having a combined population of around 5.5 million accounting for approximately 34% of the total population in the Netherlands. The average distance from the centre of these regions to Düsseldorf airport is 60 km for Limburg (Roermond), 153 km for Noord-Brabant (Tilburg) and 130 km for Gelderland (Rheden) compared to 189 km, 127 km and 115 km from Limburg, Noord-Brabant and Gelderland respectively to Amsterdam. Travel time by car from Noord-Brabant and Gelderland to Düsseldorf airport only slightly exceeds travel time by car from these two regions to Amsterdam airport, while from Limburg it is significantly shorter. Equally, costs of car transport are only slightly higher from Noord-Brabant and Gelderland to Düsseldorf airport than to Amsterdam airport, while from Limburg they are significantly lower. Travel time by train from Noord-Brabant and Gelderland to Düsseldorf airport is longer than to Amsterdam (around 70 minutes in case of Noord-Brabant and 20 minutes in case of Gelderland) and travel costs are higher, with the difference in travel costs however currently not exceeding EUR 15. Travel time by train from Limburg to Düsseldorf airport is shorter than to Amsterdam airport and the corresponding travel costs are lower. Accordingly, there are large parts of these regions that are actually closer to Düsseldorf than to Amsterdam. Dutch passengers living in these regions can therefore be expected to be at least equally inclined to travel out of Düsseldorf airport than to travel out of Schiphol.

130. The respondents to the Commission's market investigation considered that mainly Dutch customers from regions neighbouring the airports of Brussels and Düsseldorf do or would travel out of these airports. Should the prices of flights out of Amsterdam increase, the catchment area of Brussels and Düsseldorf airports would expand even further. More specifically, tour operators and travel agents indicated that a proportion of between 5% and 25% of Dutch leisure passengers (on top of those Dutch customers who already travel out of these airports) would switch their point of departure to Brussels or Düsseldorf if prices for flights to long-haul holiday destinations from Amsterdam were to increase by 5-10% and alternative flights were available from Brussels or Düsseldorf.

131. The customer survey confirms the findings of the market investigation. Indeed, the majority of respondents, 49% of the whole sample, replied that they would switch to either Brussels airport or Düsseldorf airport, or both, while only 36% of the whole sample replied that they would not switch to either Brussels airport or Düsseldorf airport where the price of their trip increased by 5-10%. (59) In addition, these figures refer to Dutch leisure travellers currently flying out of Amsterdam airport and thus do not take into account those customers who could have potentially flown out of Amsterdam but eventually chose to fly from another airport. Accordingly, the elasticity at the origin which is indicated by these figures is likely to be lower than the actual elasticity of demand. Overall, one would therefore expect that air transport services offered by competing carriers out of Brussels or Düsseldorf exert a competitive constraint on the parties.

132. It must therefore be concluded, for the purpose of the present decision, that the airports in Brussels and Düsseldorf are substitutable with Schiphol for numerous leisure passengers travelling out of Amsterdam to the long-haul "sun and beach" holiday destinations covered by this market investigation. As will be discussed in section 8.4.3.3 with respect to the routes to North America, namely Miami, Vancouver and Toronto, the proposed concentration would not significantly impede effective competition irrespective of whether the airports of Brussels and Düsseldorf are included in the relevant market. With respect to these routes it can therefore be left open whether the relevant market at the point of origin is limited to Amsterdam or whether it includes Brussels and Düsseldorf.

8.3.2.3. Point of destination substitutability

133. Among eight destinations affected by the present transaction, two main groups can be identified. The first group contains three North-American destinations, namely Miami, Vancouver and Toronto, which are served by numerous network carriers and have diverse demand composition including a fair portion of business travellers. The second group includes five destinations in the greater Caribbean region, namely Cancun, Punta Cana, Havana, Aruba and Curacao. These are mainly holiday destinations where a vast majority of customers travel for leisure purposes and which are served mainly by leisure airlines.

134. The parties claim that, with respect to holiday destinations affected by the present transaction, the O&D approach is inappropriate as leisure passengers are more price-sensitive than destination-sensitive and thus would be inclined to fly to another holiday destination in the event the tariffs for flights to such other holiday destinations were more attractive. In view of this fact, the parties argue that the relevant market should include flights offered to Dutch leisure travellers to a bundle of destinations including in principle all long-haul holiday destinations, or alternatively, all long-haul "sun and beach" destinations or at least the bundle destinations in the greater Caribbean region.

135. Respondents to the first phase market investigation considered there to be a certain degree of substitutability between various holiday destinations but that the degree of substitutability varies from one destination to another, depending on the customer mix. While in the eyes of an average Dutch holidaymaker, typical long-haul "sun and beach" destinations are to a certain degree substitutable, demand for some of the destinations is more inelastic given both the specificity of the destination and the customer base.

136. Indeed, the first phase market investigation indicated that demand for flights between the Netherlands and the Netherlands Antilles, namely Aruba and Curacao, is so specific that even if there were a price increase on the part of the merged entity on these routes post-transaction, only a very small proportion of customers would substitute them for other (long-haul) holiday destinations. This lower price sensitiveness seemed to be due to the historic, cultural, and linguistic ties to the Netherlands, favourable climate (hurricane free area) and high popularity of these two destinations amongst Dutch holidaymakers. In fact, a significant proportion of the demand on these two routes (approximately 25-30%) according to the Schiphol survey (60) submitted by the parties) is constituted by so-called "VFR" passengers, that is to say, passengers visiting family and relatives whose main reasons for travel are family ties and who are therefore destination-sensitive. Approximately 5% of the demand is represented by destination-sensitive and price-insensitive business passengers going to the islands for specific business reasons. (61) The largest group, 65-70% of customers, is represented by leisure passengers. Within this last group, there is a small subgroup of Dutch travellers having their second home on these islands and thus not willing to switch for another destination. Accordingly, a significant proportion of the demand to these two destinations seemed to be inelastic.

137. To dispute the findings of the Commission's first phase market investigation, the parties provided figures from the Tourist Boards of Aruba and Curacao which have collected information on the basis of declarations made by travellers on the landing cards, and according to which a significant proportion of all Dutch holidaymakers visiting Aruba and Curacao (75%) are first time visitors. From these figures, the parties inferred that the majority of travellers to these islands are not going there for a specific purpose (VFR or second home), but are rather represented by typical holidaymakers who made a one time choice when deciding on their "sun and beach" holiday. However, the parties did not provide any comparable figures for other holiday destinations in the same geographic area which could have demonstrated that demand of Dutch holidaymakers flying to Aruba and Curacao does not differ significantly from demand to the other destinations (62). These submissions were therefore insufficient for the Commission to be able to dismiss the findings of the market investigation that demand for flights to Aruba and Curacao is actually less elastic than demand for other holiday destinations. More importantly, these figures did not constitute sufficient evidence to demonstrate that a sufficient proportion of Dutch leisure passengers would actually switch to other holiday destinations in the case of a price increase for flights to Aruba and Curacao. In light of these findings, the Commission had therefore considered in its first phase investigation that, at least for flights to Aruba and to Curacao, the proposed transaction had to be assessed on the basis of the traditional O&D approach.

138. The results of the second phase market investigation indicate that in any case for typical "sun and beach" destinations in the Caribbean, such as Cancun or Punta Cana, other long-haul "sun and beach" destinations in the same region exert some competitive pressure. This is also true, albeit to a lesser extent, for destinations in the Dutch Antilles, namely Aruba and Curacao, as well as for Havana, three destinations which appear to be specific and thus display a limited substitutability with other destinations.

139. According to the customer survey, 34% and 25% of customers travelling to Punta Cana and Cancun respectively would switch to other holiday destinations if there were a price increase. This proportion is much lower for the other three destinations, amounting to only 17%, 14% and 13% of the sample of customers travelling to Aruba, Curacao and Havana, respectively. (63) The customer survey therefore indicated that (assuming that the market conditions on the other destinations remain unchanged) the respondents' price elasticity suggested by these figures is relatively high. To take an example, for the route Amsterdam - Havana, - allegedly the least elastic destination - the price elasticity of demand of respondents would amount to approximately 1.7. (64) An increase in price would thus result in a considerable reduction in total revenue and possibly, in profits, given the low incremental operating cost per passenger which characterizes air transport (65).

140. With respect to Miami, Toronto and Vancouver, the first phase market investigation seemed to indicate that the traditional O&D approach is appropriate. As no competition problems arise irrespective of the market definition, the Commission has not investigated the scope of the relevant market with respect to any of these three destinations further in the customer survey.

141. As the results of the market investigation with respect to the scope of the point of destination are inconclusive, the effects of the present transaction will be assessed on the basis of the narrowest possible relevant market, that is, destination by destination. In any event, it is not necessary for the purpose of the present decision to conclude on the level of destination substitutability in the eyes of Dutch holidaymakers, as regardless of the result, the proposed concentration would not impede effective competition in the common market or a substantial part of it.

8.3.2.4. Substitutability of direct and indirect flights

142. With respect to long-haul flights the Commission has found in the past (66) that indirect flights constitute a competitive alternative to non-stop services under certain conditions, in particular, when they are marketed as connecting flights on the relevant O&D pair in the computer reservation systems, they are operated on a daily basis and cause only a limited extension of the trip (maximum 150 minutes connecting time). Long-haul flights have been defined as flights with flight durations in excess of three hours (67). The notifying party agrees with the Commission's past practice.

143. The market investigation in the present case broadly confirmed the Commission's practice in past cases with respect to the substitutability of indirect flights to non-stop services under certain conditions. This is even more the case for leisure passengers, given that leisure passengers are generally less time-sensitive than other passengers, notably business travellers.

144. For the purpose of the present decision, the Commission therefore takes the view that indirect flights with a connecting time of a maximum of 150 minutes are part of the same market as direct flights.

8.3.2.5. Time-sensitive v. non time-sensitive passengers

145. The Commission has traditionally found that passengers travelling on unrestricted tickets, so-called time-sensitive passengers which are primarily business customers requiring flexibility, are in a different market than passengers travelling on restricted tickets, the so-called non time-sensitive passengers who attach more importance to the price than to the schedule or the direct/indirect itinerary of the flight. (68)

146. Taking account of the fact that, according to the parties, the vast majority of Martinair's passengers are non-time sensitive passengers (between 96% and 97%) and that Martinair does not currently offer business class seats, the present decision will only focus on the effects of the transaction on non-time sensitive passengers who are the ones most affected by the present transaction.

8.3.2.6. Conclusion on the market definition with respect to the supply of air transport services to end customers

147. For the purpose of the present decision, in line with Commission precedents, the Commission has assessed the effects of the proposed transaction primarily on the market for the supply of air transport services to end customers on the basis of the traditional O&D approach, whereby, with respect to the routes to Aruba, Curacao, Punta Cana, Cancun and Havana, the point of origin comprises flights out of Amsterdam, Brussels and Düsseldorf. Given that the results of the market investigation with respect to substitutability of the airport of origin concerning the flights to Miami, Toronto and Vancouver were inconclusive, the effects of the present transaction with respect to these markets are assessed against the narrowest possible alternative, that is, comprising only flights out of Amsterdam to each of these three cities respectively. In its assessment, the Commission included indirect flights as competitive alternatives to direct services to the extent that they are substitutable.

148. Moreover, the Commission has also analysed the results of the proposed transaction on the basis of the two wider markets proposed by the parties: (i) the market for the supply of air transport services to end customers between the airports of Amsterdam, Düsseldorf and Brussels and all long-haul destinations; and (ii) the market for the supply of air transport services to end customers between the airports of Amsterdam, Düsseldorf and Brussels and all long-haul destinations in the Greater Caribbean.

8.3.3. The wholesale supply of airline seats to tour operators

149. As already mentioned, due to the indirect pressure from the downstream market where end customers travelling to leisure destinations have a choice between unbundled flights and package holidays, the parties consider that it is not appropriate to define a separate market for the carriage of passengers on behalf of tour operators. However, should this market exist, the parties consider that the market should be defined as the market for the sale of seats to tour operators addressing the Dutch market for all long-haul destinations (or at least all long-haul "sun and beach" destinations) (69). This would also include seats supplied to tour operators addressing Dutch leisure passengers departing from Düsseldorf and Brussels.

150. As discussed in recitals 99 to 106, the market investigation confirmed the existence of a separate market for the wholesale supply of airline seats to tour operators. In addition, respondents to the Commission's market investigation have confirmed that, in the Netherlands, with respect to long-haul destinations, scheduled airlines are the main source of supply of airline seats to tour operators. ArkeFly, the only non-scheduled airline serving the Dutch market, sells almost all of its capacity internally to tour operators belonging to the TUI group.

8.3.3.1. Overall long-haul market v. distinction on a route-by-route basis

151. The parties, on the basis of past Commission decisions, consider that if there is a market for the supply of seats to tour operators, this market should comprise the supply of airline seats to all long-haul destinations without further sub-division of the market by destination.

152. The Commission has, in previous cases, considered that the relevant market for the wholesale supply of seats to tour operators (which is upstream to the markets for the supply of foreign package holidays that are still largely national in character) is national in scope (the demand for air transport services included in holiday packages arises mainly from customers resident in the country of origin) without the need for any further geographic subdivision. (70)

153. The market investigation revealed that large tour operators, in particular, negotiate their deals with airlines on a route-by-route basis (to allow the airlines to be able to accommodate the demand with respect to each route and forecast the capacity needs). These large tour operators often benefit from allotments, namely, pre-reserved amounts of seats expiring 4-6 weeks before the departure date at which moment the TOs must decide whether they are going to purchase the seats or release them (return them to the airline). In addition, with respect to Martinair, some large TOs, currently only Thomas Cook, purchase guaranteed seats at the beginning of the season for which they negotiate volume and a flat rate but which have to be paid in advance so that TOs support the risk of not selling them (71). All other TOs purchase seats from airlines, although in different conditions than end customers, only at the moment when their customers express interest in purchasing a package to a specific destination. These seats are purchased on a first-come-first-served basis. Accordingly, except for the purchase of guaranteed seats, with respect to which tour operators may have the ability to arbitrate between various destinations (and thus direct their downstream demand towards destinations including cheaper flight), the level of substitutability between different holiday destinations is largely driven by the one prevailing on the market for the sale of seats to end customers.

154. This was confirmed by the results of the market investigation. All of the tour operators who replied indicated that they would continue to offer the affected destinations in case of a 5-10% price increase. They would simply pass on the price increase to their customers, who, of course, in turn may decide to travel to another destination, or not travel at all. Similarly, as in the case of the market for the sale of seats to end customers, tour operators indicated that Aruba and Curacao, in particular, may be "must have" destinations due to their high popularity among the Dutch population. As a result, some tour operators considered that a 5-10% price increase would not hurt the demand to these destinations.

155. The perspective of the end customer is somewhat different. The customer survey indicates that customers purchasing packages tend to be more flexible than customers purchasing an unbundled flight. At an aggregate level, 27% of the respondents who bought a package tour asserted that they would switch to another destination in response to a 5-10% price increase. Similarly, only 33% of the whole sample of respondents who bought a package tour indicated that they would not switch to either the Brussels airport or the Düsseldorf airport if the price of their trip was increased by 5-10%.

156. In any event, it is not necessary for the purpose of the present decision to conclude on whether the market for wholesale supply of seats to tour operators should comprise all long-haul destinations (or alternatively all long-haul "sun and beach" destinations) or should be further subdivided on a route-by-route basis as regardless of the answer, the proposed concentration would not impede effective competition in the common market or a substantial part of it.

8.3.3.2. Airport substitutability

157. The notifying party considers that if there is a market for the supply of seats to tour operators, this market should comprise the supply of airline seats to all long-haul destinations from Amsterdam, Düsseldorf and Brussels.

158. The market investigation confirmed that TOs already offer packages with a departure flight out of Brussels and Düsseldorf. In addition, tour operators indicated that in case of price increases for flights out of Amsterdam, at least for part of their requirements, they would switch the point of departure for their package holidays to long-haul destinations to neighbouring airports, in particular Brussels and Düsseldorf. Nevertheless, Dutch tour operators also indicated that they would still have to meet a high percentage of the demand by offering package tours departing from Amsterdam since only customers living in the border regions would be willing to travel to Düsseldorf and Brussels.

159. However, as tour operators have a tendency to pass any price increase on to the end customers who, as indicated in recitals 180-183, are price-sensitive and clearly consider airports in Brussels and Düsseldorf as alternative points of departure for their long-haul holiday trips, it can be concluded that, for the purpose of the present decision, the airports in Amsterdam, Brussels and Düsseldorf are substitutable and thus part of the same relevant market for the wholesale supply of seats to Dutch tour operators.

8.3.3.3. Conclusion on the market definition with respect to the wholesale supply of airline seats to tour operators

160. Given the inconclusive results of the market investigation with respect to the exact scope of the relevant market, the fact that most tour operators do not have the ability to arbitrate between various destinations and the indications that the market should be subdivided by destination, the Commission has assessed the effects of the proposed transaction on the market for the wholesale supply of seats to tour operators both on a route-by-route basis and on the basis of an overall market for the wholesale supply of seats to Dutch tour operators for all long-haul holiday destinations, whereby in both cases the point of origin comprises flights out of Amsterdam, Brussels and Düsseldorf.

8.4. Competitive Assessment

161. As regards passenger air transport, the main impact of the proposed concentration is on the transport of Dutch leisure passengers from Amsterdam to various holiday destinations. In accordance with the market definitions considered in section 8.3, the effects of the proposed concentration on competition must be assessed separately for (i) the supply of air transport services to end customers (section 8.4.3), and (ii) the wholesale supply of airline seats to tour operators (section 8.4.4).

162. In assessing the effects of the proposed concentration on these two markets, account must be taken of general features of long-haul air transport of Dutch leisure travellers (section 8.4.2) as well as of the relevant counterfactual for the assessment of the present transaction with respect to the air transport of passengers. To this end, the extent to which Martinair currently constrains KLM in its conduct on the market for the supply of air transport services to end-consumers and the market for the wholesale supply of airline seats to tour operators has to be assessed (section 8.4.1).

8.4.1. The relevant counterfactual for assessing the effects of the proposed transaction relating to the parties' operations in the air transport of passengers

163. The parties claim that, in the absence of the acquisition by KLM of Maersk's 50% shareholding in Martinair, Martinair's long-haul passenger operations are likely to be discontinued in the near future. This is due to the fact that Martinair's passenger business has been loss-making during the last few years and Maersk lacks any incentive to make the necessary investments given that it no longer has any strategic interest in the aviation business. The parties argue that, for these reasons, the competitive pressure that Martinair currently exerts on KLM is already limited and in any case, under the current conditions, it will only further decrease.

164. Indeed, Martinair's internal documents confirm the loss-making nature of Martinair's long-haul passenger business as well as the need for a restructuring of these operations in order to make them sustainable. This is evidenced, among others, by figures from Martinair's management accounts in 2006, 2007 and 2008 (72) as well as the minutes of meetings of Martinair's internal decision-making bodies from 2004-mid 2008 (73), showing that Martinair's long-haul passenger business has reported negative profits since 2005.

165. In a report which Martinair commissioned in 2007 to review its strategy (74), VS Consulting advised Martinair that its long-haul passenger business could become sustainable if its lack of scale would be addressed. (75) As a consequence, Martinair restructured its long-haul passenger operations by reducing the number of destinations served, while increasing the number of flights to the remaining destinations, thus aiming to allow for a more efficient use of crews and aircraft. (76)

166. Despite these restructuring measures, however, the financial situation of Martinair appears to have gotten considerably worse. In fact, whereas the parties estimated Martinair's incremental losses, that is to say, the difference between passenger business revenues and those costs (variable or fixed) which are solely attributable to the passenger business and which could only be avoided if the passenger business were closed down, to amount to [...]* in 2007, they anticipate its incremental losses to reach [...]* by the end of 2008. (77)

167. The parties further explained that, in order to mitigate these losses, Martinair has initiated a number of measures including staff cuts, measures to reposition Martinair's brand, measures to improve its yield management and changes in its policy on fuel surcharges. According to the parties, the cost savings that can be achieved as a result of these measures would amount to approximately [...]*. (78) These initiatives would therefore not be sufficient to overcome the [...]* of incremental losses anticipated for 2008.

168. In addition, the parties have explained that, to continue to operate its passenger air transport business as a stand-alone business beyond the next two years, namely, without the integration of this business in a larger passenger air transport business like that of KLM, Martinair will need to make some large investments, including investments in its fleet and in the development of systems for inventory and check-in. Martinair's financial situation, however, makes it difficult to raise finance for such investment. According to the parties, Martinair has defaulted on loan agreements in recent years because the required level of equity was insufficient. Amendments to its loan agreements allowing a lower level of equity have resulted in higher financing costs. (79) Moreover, Martinair's difficulty in raising finance is also reflected in the lending rates it achieves. These are considerably less attractive than those that KLM normally achieves. (80)

169. The main investment that Martinair would need to make to continue operating its passenger air transport business in the medium and long term concerns the renewal of its fleet. Martinair currently has six aircrafts, of which three are owned and three are leased. Martinair's entire fleet are old (averaging over 80 000 flights hours (81)) and inefficient, incurring high maintenance costs and inefficient fuel usage. One of the aircrafts in Martinair's ownership was taken out of service in September 2008. Two of the three leased aircrafts have accumulated more combined flight hours and flight cycles than any other aircraft in Martinair's fleet and therefore suffer from a high dispunctuality (that is, aircraft grounded for repair). The leases for two of the remaining three aircrafts expire on 15 July 2009 and the lease of the other one expires on 18 June 2011. In addition, the two other aircrafts which are owned by Martinair will also have to be replaced within a short period of time as they have also accumulated considerable combined flight hours and flight cycles. Accordingly, unless the necessary investments are made, Martinair might be left with only one aircraft from mid-2009. This appears unlikely to be sufficient for profitable operations and it will be clearly insufficient to continue operations on all eleven routes currently served by Martinair.

170. During the market investigation, one tour operator explicitly confirmed that, due to technical problems with its old fleet, Martinair flights are often delayed. As a result, the reliability of Martinair's services has decreased and its brand image has been suffering. This appears to have led to a consistent decrease of Martinair's competitive strength with respect to the long-haul air transport of passengers. Furthermore, according to a number of other tour operators, the age of Martinair's passenger fleet is Martinair's most serious problem, and it needs to be addressed for Martinair to be able to operate profitably in the long term. Additionally, several tour operators noted that, in the absence of the proposed concentration, they did not consider Martinair's long-haul passenger operations to be sustainable and they would therefore expect Martinair's passenger business to be closed down in the short term.

171. To restore its competitive strength or even to maintain it at its current, already somewhat weakened level, Martinair would need to invest in either purchasing or leasing new aircrafts. However, as set out in recital 18, this is a strategic decision which is subject to approval by Martinair's supervisory board and with respect to which both KLM and Maersk have a veto right.

172. Maersk has already explicitly stated that it has no strategic interest in its stake in Martinair (82) as it intends to withdraw from the airline business and focus on the shipping business. Accordingly, Maersk has no interest to further invest in a Martinair's loss-making business.

173. More importantly, however, it is also uncertain whether KLM would agree to far-reaching investments in Martinair's fleet as long as it holds only 50% of the shares in Martinair, while co-shareholder Maersk is reluctant to make the necessary investments and KLM has only limited control over Martinair's operations. As previously noted, KLM's influence on Martinair is limited with respect to operational decisions, such as decisions on pricing, capacity or frequency decreases or increases on individual routes or aircraft allocation to such routes. However, the situation seems to be different at this crucial point in time where Martinair has to renew almost its entire fleet within only a few months. Due to its veto right, it can be assumed that KLM will be able to exercise decisive influence over Martinair with respect to its future competitive strength in the passenger air transport market.

174. KLM's position is reinforced by the fact that it appears unlikely that Maersk could sell its stake to a third party instead of KLM. Pursuant to Martinair's Articles of Association, Maersk has to offer its shares to KLM first. (83) Only in the event that no agreement between KLM and Maersk is reached can Maersk negotiate the sale of its shares to a third party. However, even in that case, it appears that it might be very difficult for Maersk to find a buyer for a 50% share in a company which is highly loss-making, in which KLM owns the other 50% and as such, is able to block strategic decisions, such as the renewal of Martinair's fleet. Financial buyers are unlikely to be interested in acquiring Maersk's stake as long as they have to share control with KLM and there are no credible prospects that the necessary investments in Martinair's fleet would actually be made (84). Strategic buyers would likely be even more reluctant to share ownership with KLM, in particular given that Martinair depends on KLM not only for the renewal of its fleet, but also for part of its operations, namely for revenue accounting, reservations and sales systems. (85) Given the loss-making nature of Martinair's passenger business, other airlines would likely be more inclined to start up their own passenger operations from Schiphol rather than sharing ownership in Martinair with their competitor KLM. According to the parties, Maersk has confirmed to them that it has had discussions with other parties about the sale of its share in Martinair, but that none of these other parties were interested in acquiring a share in Martinair as long as KLM held the other 50%.

175. These specific circumstances constitute an important element of the counterfactual for the assessment of the proposed concentration. While the former evidence seems to indicate that the parties are competitors (86), Martinair's specific situation makes it likely that the competitive constraint exerted by Martinair will be eroded in the foreseeable future. From the above it can therefore be concluded that the merger-specific effects of the proposed concentration with respect to the parties' passenger air transport activities are likely to be limited. (87)

8.4.2. General market conditions

176. When analysing transactions between airlines, numerous variables must be taken into account. The load factor (that is, the capacity utilization of the aircraft during a given period) and the average yield per passenger determine the profitability of airline operations. Accordingly, the number of passengers carried and the revenue per seat sold are key to growth in the airline industry.

177. On the basis of their operational costs and expected revenue, airlines can calculate their break-even or critical minimal load factor which determines the minimum scale of profitable operations on a particular route. Given high-fixed costs in the airline industry, the marginal cost of an extra passenger on board is relatively low and therefore the more passengers on the plane, the higher the profitability of the operation of a particular flight. While, generally speaking, a high load factor is needed for profitable operations, it should be noted that although the load factor is one of the indicators of profitability, other parameters such as the availability of seats in different booking classes, the number and configuration of seats on the flight or the type of aircraft, may influence its importance in the overall profitability calculations.

178. Depending on the business model and the composition of the customer mix of each carrier, the high load factor may have different importance. With respect to the case at hand, KLM and Martinair indeed rely on different business models and the composition of their customer base differs substantially. While KLM is a full-service network carrier, relying (at least to some extent) on high yield business passengers, Martinair is a full-leisure point to point carrier.

179. Given the higher average yield per passenger KLM achieves (due to its business class component) as compared to Martinair, KLM's dependence on a high load factor is less crucial than that of Martinair, which sells virtually all of its capacity to leisure travellers. (88) The market investigation in the present case generally confirmed the need for a high load factor to operate profitably and indicated that the loss of a small number of passengers may have a negative impact on the profitability of the service. (89) The need for a high load factor makes that airlines, in particular the ones relying solely on leisure traffic, are extremely sensitive to a small decline in the number of passengers given that even such a small decline can make the difference between a profitable and an unprofitable service. As a consequence, a small number of marginal passengers is capable of exercising sufficient price pressure to prevent price increases. (90) The need for a particularly high load factor of a leisure carrier can be corroborated with the fact that Martinair's current average load factor amounts approximately to 87.5% and it still makes significant losses.

8.4.2.1. Specificities of air transport of leisure travellers

180. The routes concerned by the present transaction are mostly routes connecting Amsterdam, Düsseldorf and Brussels to holiday destinations. Except for Miami, Vancouver and Toronto, all other routes, namely Aruba, Curacao, Punta Cana, Cancun and Havana, are "sun and beach" destinations. Within this group, Aruba and Curacao can be distinguished from the other three destinations by a relatively high proportion of ethnic traffic (91) (around 20%) and the existence of cultural ties to the Netherlands. In any case, the five "sun and beach" destinations have in common a low proportion of business traffic in the overall composition of demand. (92) Accordingly, these leisure destinations account for low yield traffic.

181. Generally speaking, demand of leisure travellers tends to be unstable given that customers are very sensitive to demand shocks, as strictly speaking, they do not need to travel for leisure. This was largely confirmed during the market investigation and the customer survey, both showing that Dutch leisure travellers tend to be sensitive to demand shocks such as price increases.

182. In order to support the claim that leisure travellers are very price sensitive, the parties have provided some estimates of airfare elasticity as well as results of a natural experiment provided by the introduction of the Dutch ticket tax at Schiphol. (93) With respect to price elasticity of air travel, the parties have reviewed some papers demonstrating that the price elasticity of leisure travellers is far in excess of 1. (94) In addition, from the effects of the introduction of the Dutch environmental tax at Schiphol airport estimated by the Netherlands Bureau for Economic Policy Analysis, the parties infer a price elasticity of 1.7. This evidence indicates that there is usually a significant customer response to changes in airfares.

183. These studies provide some interesting indications. However, they cannot be taken as definitive: (95) Given the high heterogeneity of travellers, the extent to which the results reported by the quoted sources (which tend to focus on passengers travelling to other destinations and, in particular, passengers travelling from/to Australia) can be applied to Dutch leisure travellers' price sensitivity is rather limited.

8.4.2.2. Closeness/Asymmetry of competition between KLM and Martinair

184. There are some clear differences between KLM and Martinair in terms of their market positioning and business models. Martinair's customer base predominantly consists of leisure passengers. (96) Martinair has no business class and very few time-sensitive passengers. As a point-to-point carrier focusing on leisure travellers, Martinair has a business model that is fundamentally different from the business model of Air France-KLM. Air France-KLM is a network carrier with a different passenger mix in which leisure passengers are generally in a minority.

185. The customer survey has shown a considerable degree of asymmetry between the parties. More specifically, the survey checked directly whether customers of the different airlines would have considered travelling with a different airline if the price they paid had been increased by 5-10%. (97) According to the responses received, most of Martinair's passengers travelling to Aruba and Curacao consider KLM as a valid alternative. With respect to Aruba, 55% of Martinair's passengers considered KLM as an alternative, 17% did not consider any other airline and 15% considered ArkeFly. Similarly, for Curacao, 51% of Martinair passengers considered KLM as an alternative, 21% did not consider any other airline and 17% considered ArkeFly. On the other hand, most of Martinair passengers flying to Havana and Cancun did not consider any carrier as alternative (43% and 53% respectively).

186. On the other hand, a much smaller percentage of KLM passengers seem to consider Martinair as a possible alternative in all of the surveyed routes. Indeed, the majority of KLM's passengers indicated not to have considered any other airline, irrespective of the route. Moreover, the percentage of KLM passengers travelling to Curacao who considered ArkeFly as an alternative carrier (27%) appears to be higher than those who considered Martinair as an alternative carrier (23%).

187. On the basis of these results, it must be concluded that KLM exerts significantly more competitive pressure on Martinair than vice versa.

188. The differences between KLM and Martinair in terms of business models also translate into the wholesale market for the sale of seats to tour operators in that Martinair is much more a "tour operator airline" than KLM. KLM, on average, sells not more than [10-20]* % of its capacity on flights to holiday destinations to tour operators. Martinair, on the other hand, sells approximately [50-60]*% of its capacity to tour operators and on some routes even more than [90-100]*%. As Graph 2 below shows, the composition of the sample of respondents to the customer survey confirms such difference between KLM and Martinair. Indeed, as it can be noted, the majority of respondents flying with Martinair bought a package tour (62%) while the vast majority of KLM customers bought an unbundled flight (82%).

Graph 2: Composition of the sample of respondents by carrier and by type of travel

189. Therefore, despite the fact that both KLM and Martinair have their hub or base at Amsterdam, the two airlines cannot be considered as close substitutes in every respect. In addition, as indicated by the market investigation and the customer survey, not only Martinair, but also ArkeFly, tends to have a higher percentage of package tour customers than KLM. It thus appears that, in terms of market positioning, Martinair is closer to ArkeFly than to KLM.

8.4.2.3. Barriers to entry and expansion

190. When considering airline markets, it has to be borne in mind that starting up or expanding an airline business is subject to a number of regulatory conditions which vary depending on the country of operation and may render entry on the market difficult. Barriers to starting or expanding airline operations are usually threefold; consisting in (1) regulatory barriers related to traffic rights, (2) slot restrictions that may exist at a particular airport and (3) high sunk costs associated with setting up operations.

Traffic rights

191. As concerns traffic rights, it should be noted that international aviation is regulated by a web of bilateral air services agreements ("ASAs") concluded between States or at the level of regional organisations. The ASAs regulate the conditions under which airline carriers can operate services between the two signatories concerned. The ASAs may specify routes that may be served, the number of carriers as well as their nationality that can be authorised to operate the routes, the capacity in terms of weekly/monthly seats, the number of frequencies, the size of aircraft as well as prices.

192. The resulting regulatory framework is diverse. It ranges from restrictive agreements whereby only one carrier from each country can be designated to operate the route (single designation) with all other operational parameters determined by the ASA to comprehensive agreements, such as "Open skies agreements" which reduce many of the regulatory barriers. Open skies agreements provide for unlimited designation and market access, unlimited frequencies etc. One example of a comprehensive agreement is the EU-US Air Transport Agreement (98).

193. One common aspect of all these agreements is that, irrespective of whether they provide for single/multiple or unlimited designation, the authorisation to operate on the relevant routes is usually reserved to carriers owned and controlled by nationals of the signatory countries, which constitutes a barrier to entry. In November 2002, the European Court of Justice found that, if an ASA between a Member State and a third country permits designation only of companies owned and controlled by nationals of the signatory EU Member State, such designation is discriminatory and is in breach of EC law. (99) As a result, every Member State is required to grant equal market access for routes to destinations outside the Community to any EU carrier with an establishment on its territory (the so-called 'nationality clause' must be replaced by the 'Community designation' clause). ASAs between Member States and third countries must be amended to ensure that traffic is operated on a sound legal basis. So far, 777 existing agreements have been regularised either by Member States and third countries, or by a agreements negotiated directly by the Community (38 'horizontal' agreements have been negotiated).

194. When assessing the market structure and the possibility of market entry for long-haul flights, the Commission took into consideration the existence of specific regulatory barriers following from this web of ASA, including the recognition or not of the Community designation, in order to determine to what extent other carriers, be it European or third countries, may benefit from the necessary traffic rights and operation flexibility to compete effectively with the parties. The relevant ASAs and, in particular, the extent to which they may restrict entry of expansion on the affected routes in the present case will be described in more detail in section 8.4.3.

Slot restrictions

195. Another barrier air carriers often face is the availability of slots at congested airports. There are three levels of congestion whereby airports are classified as non-coordinated airports (airports with capacities which are adequate to meet the demand), schedules facilitated airports (where there is potential for congestion at some periods but it is resolved by voluntary coordination between air carriers facilitated by a schedule facilitator) and coordinated airports (where demand exceeds capacity and landing and taking off slots must be allocated to air carriers by an independent coordinator on the basis of transparent rules). Accordingly, for carriers to operate a service out of a congested airport, they must have permission to use the infrastructure given by the slot coordinator. Depending on the service (short-haul/long-haul) different slots are convenient. While short-haul services require early morning and late night slots allowing for at least two round trips during the day, long-haul services tend to be operated from afternoon slots to allow for morning arrivals at the destination.

196. With respect to the case at hand, the parties claim that none of the airports concerned by the services affected by the present transaction are congested with respect to long-haul services to leisure destinations, which are by nature flexible as to the time of departure. The Commission's market investigation generally confirmed the parties' submission, except in the case of Düsseldorf airport, which seems to be the most congested of all three alternative departure airports (100). While Schiphol is congested during peak hours, airlines serving non-time sensitive leisure passengers are flexible to schedule their flights outside of peak hours. With respect to the availability of slots at the destination airport, none of the airports seems to be congested as concerns long-haul flights (101). Accordingly, with respect to this specific case of long-haul leisure air transport of passengers, it seems that slot restrictions are no significant barrier to entry.

Minimum scale for profitable long-haul passenger air transport operations

197. Furthermore, establishing airline operations involves high sunk costs associated mainly with purchasing or leasing aircrafts and hiring the crew. In addition, establishing a base or a hub at an airport normally requires minimal scale of operations to make a base/hub profitable. With respect to the case at hand, and taking into account the specificity of long-haul leisure markets, the market investigation indicated that an airline would need, in addition to a good cost structure, a minimum of 3-4 aircrafts allowing for a good frequency to 4-5 long-haul destinations to operate profitably. In addition, when starting operations on individual routes, airlines have to be able and ready to cope with the financial risk and possible retaliatory measures of the incumbent. Accordingly, for a new player which is not active in long-haul passenger air transport at all, or for an existing player wishing to start operating on the affected routes which does not have a long-haul base or hub either at the origin airports of Amsterdam, Düsseldorf and Brussels or at the destination airports served by both parties ( Vancouver, Toronto, Miami, Punta Cana, Cancun, Havana, Curacao or Aruba), the minimum scale required for profitable operations described in this recital constitutes another significant barrier to entry or expansion.

8.4.3. The supply of air transport services to end customers

198. With regard to the supply of air transport services to end customers, the main impact of the proposed concentration is on the transport of Dutch leisure passengers from Amsterdam to various holiday destinations.

199. In line with the market definitions discussed in section 8.3, the effects of the proposed concentration will be assessed on the basis of all possible alternative markets, namely:

(i) on the basis of the hypothetical wider markets for the supply of air transport services to Dutch leisure passengers to all long-haul destinations, both departing from Amsterdam only and from Amsterdam, Brussels and Düsseldorf,

(ii) on the basis of the hypothetical wider markets for the supply of air transport services to Dutch leisure passengers to destinations in the Greater Caribbean region, both departing from Amsterdam only and from Amsterdam, Brussels and Düsseldorf , and

(iii) at the level of each of the O&D pairs, on which the parties' activities overlap, namely:

(a) Amsterdam-Vancouver ("AMS-YVR") and Amsterdam/Düsseldorf/ Brussels - Vancouver ("AMS/DUS/BRU-YVR"),

(b) Amsterdam-Toronto ("AMS-YYZ") and Amsterdam/Düsseldorf/Brussels - Toronto ("AMS/DUS/BRU-YYZ"),

(c) Amsterdam-Miami ("AMS-MIA") and Amsterdam/Düsseldorf/Brussels - Miami ("AMS/DUS/BRU-MIA"),

(d) Amsterdam/Düsseldorf/Brussels - Havana ("AMS/DUS/BRU-HAV"),

(e) Amsterdam/Düsseldorf/Brussels - Punta Cana ("AMS/DUS/BRU-PUJ"),

(f) Amsterdam-Cancun ("AMS/DUS/BRU-CUN"),

(g) Amsterdam-Curacao ("AMS-CUR") (102) and

(h) Amsterdam-Aruba ("AMS-AUA"). (103)

8.4.3.1. Air transport of Dutch leisure passengers departing from Amsterdam, Brussels and Düsseldorf to all long-haul destinations

200. On the basis of the widest possible market definition suggested by the parties, namely a market for air transport of Dutch leisure travellers departing from Amsterdam, Brussels and Düsseldorf to all long-haul destinations, the parties estimate their combined market share to amount to [40-50]*% (KLM: [30-40]*%, Martinair: [10-20]*%). (104) If this market were to be confined to Dutch leisure passengers departing from Schiphol airport only, the parties' combined market share would be slightly higher, namely [40-50]*% (KLM: [30-40]*% and Martinair: [10-20]*%). (105)

201. Martinair and KLM are the only carriers based in Amsterdam which are not vertically integrated with a tour operator and thus it is only natural that their market shares are relatively high. However, these relatively high market shares cannot, as such, be interpreted as an indication of market power in this widest possible hypothetical market proposed by the parties.

202. Firstly, as explained in more detail in recitals 184-189, there are clear differences in the parties' business models and market positioning as a result of which there is a significant asymmetry in the way the parties constrain each other pre-transaction. This limits any possible anti-competitive effects of the proposed concentration from the outset.

203. Secondly, on the basis of this widest possible hypothetical market proposed by the parties, customers can still choose to substitute flights offered by the parties to a large number of other long-haul destinations worldwide which are not served by the parties or to flights to the same destination which are served by competing carriers. These alternatives still account for around [50-60]*% of the market. Indeed, there are numerous competitors active in this hypothetical market to which customers could switch. In addition to several airlines operating flights from Amsterdam, Düsseldorf and Brussels to the Greater Caribbean Area (see recital 209), there is a large number of other airlines serving destinations in Africa, the Middle East, North and South America and Asia both from Schiphol and from the neighbouring airports of Düsseldorf and Brussels. Examples of airlines carrying leisure passengers to long-haul destinations from Düsseldorf and Brussels include Air Berlin, Emirates (both operating from Düsseldorf) and Etihad (operating out of Brussels).

204. Given the specificities of long-haul passenger air transport, it also appears that many of these competitors would be able to quickly increase capacity, at least on those routes on which they are already present. As has been noted in recitals 195-196, long-haul passenger air transport services do not necessarily require slots during peak hours and, outside peak hours, slots at the airports of Amsterdam, Düsseldorf and Brussels are still available. Regulatory restrictions under the applicable bilateral air services agreements may limit the maximum number of frequencies that can be operated on certain routes, but this would be the case only with respect to some destinations included in the relevant market and sufficient other destinations would remain where such regulatory restrictions would not prevent competitors of KLM and Martinair from increasing capacity or launching new services.

205. Moreover, as set out in section 6.2, the fact that KLM jointly controls Martinair pre-transaction is likely to reduce the anti-competitive effects of the proposed concentration. Finally, as discussed in section 8.4.1, the competitive pressure that Martinair currently exercises on KLM in passenger air transport in general is decreasing due to its financial situation and the need for the timely renewal of its fleet, which is ultimately controlled by KLM. Therefore the effects of the proposed concentration with respect to the parties' passenger air transport activities will in any event be limited.

206. Considering these circumstances and in particular taking account of the large number of alternatives available in this hypothetical market for air transport of Dutch leisure travellers departing from Amsterdam, Brussels and Düsseldorf to all long-haul destinations, the significant price-sensitivity of leisure travellers in general and the limited barriers to expansion for existing competitors already present on routes connecting Amsterdam, Düsseldorf and Brussels to various long-haul destinations worldwide, it must be concluded that the proposed concentration would not significantly impede effective competition in the common market or a substantial part of it insofar as this wide hypothetical market is concerned.

8.4.3.2. Air transport of Dutch leisure passengers departing from Amsterdam, Brussels and Düsseldorf to the Greater Caribbean region

207. On the basis of the alternative market definition suggested by the parties, namely a market for air transport of Dutch leisure travellers departing from Amsterdam, Brussels and Düsseldorf to the Greater Caribbean region, comprising Jamaica, Cuba, Haiti, the Dominican Republic, Surinam, the Bahamas, Antigua, the Netherlands Antilles and other Caribbean islands, the parties' combined market share would be approximately [60-70]*% (KLM: [30-40]*%, Martinair: [20-30]*%). (106) If this market were to be confined to Dutch leisure passengers departing from Schiphol airport, the parties' combined market share would be slightly higher, namely [60-70]*% (KLM: [30-40]*% and Martinair: [30-40]*%). (107)

208. Again, these relatively high market shares cannot as such be interpreted as an indication of market power in this alternative market proposed by the parties. As set out in more detail in recitals 184-189, there are clear differences in the parties' business models and market positioning as a result of which there is a significant asymmetry in the way the parties constrain each other pre-transaction. This limits any possible anti-competitive effects of the proposed concentration already at the outset.

209. Moreover, similar to what has been explained in relation to the wider market comprising all long-haul destinations, also on the basis of this alternative market comprising all destinations in the Greater Caribbean region, customers can still choose to substitute flights offered by the parties to a large number of other long-haul destinations in this region which are served by competing carriers or to flights to the same destination which are served by competing carriers. These alternatives account for around [30-40]*% of the market. Competitors include, in particular, ArkeFly (operating flights between Amsterdam and the Netherlands Antilles, the Dominican Republic, Jamaica, Mexico and Cuba), British Airways and Virgin Atlantic (operating flights between Amsterdam and, amongst others, Barbados, Dominica, Grenada, St. Lucia and Tobago (108)), LTU (operating flights between Düsseldorf and the Dominican Republic, Jamaica and Cuba), Condor (operating flights from Düsseldorf and/ or Frankfurt to Antigua, Barbados, the Dominican Republic, Grenada, Jamaica and Tobago) and Jetairfly (operating flights between Brussels and the Dominican Republic and Jamaica).

210. In light of the specificities of long-haul passenger air transport in terms of slots constraints and regulatory restrictions set out in recitals 191-196, it also appears that many of these competitors would be able to increase capacity, at least on those routes on which they are already present.

211. Finally, the effects of the proposed concentration with respect to the parties' passenger air transport activities will in any event be limited for the reasons set out in sections 6.1 and section 8.4.1. More specifically, as set out in section 6.2, the fact that KLM jointly controls Martinair pre-transaction is likely to reduce the anti-competitive effects of the proposed transaction. Moreover, these effects of the proposed transaction will be reduced even further because the competitive pressure that Martinair can currently exercise on KLM in passenger air transport in general is decreasing due to its financial situation and the need for the timely renewal of its fleet and the latter is ultimately controlled by KLM.

212. Considering these circumstances and taking into account the large number of alternatives for customers in this hypothetical market for air transport of Dutch leisure travellers departing from Amsterdam, Brussels and Düsseldorf to long-haul destinations in the Caribbean region, the significant price-sensitivity of leisure travellers in general and the limited barriers to expansion for existing competitors already present on the competing routes, it must be concluded that the proposed concentration would not significantly impede effective competition in the common market or a substantial part of it insofar as this alternative market is concerned.

8.4.3.3. Routes connecting Amsterdam, Brussels and Düsseldorf to North American cities (Vancouver, Toronto and Miami)

8.4.3.3.1. Vancouver

213. KLM currently serves Vancouver with a direct service 7 times a week during the IATA summer season and intends to operate 5 or 6 direct flights to Vancouver per week during the IATA 2008 winter season. Martinair does not operate flights to Vancouver in the winter season. During the summer season, it operates 3 direct flights to Vancouver per week (2 flights per week with effect from 1 September 2008).

214. The table below shows the market shares of the parties and their competitors on the Amsterdam-Vancouver route on a yearly basis for the year 2007/2008 (109) as well as the market shares of the parties and their competitors for the hypothetical wider market comprising flights from Amsterdam, Düsseldorf and Brussels to Vancouver on a yearly basis for the year 2006/2007. (110)

<table>

215. Although the parties have a significant combined market share of more than [70-80]*% on this route, this market share will decrease since Martinair has reduced its frequencies to 2 flights per week. In addition, as set out in more detail in recitals 184-189, there are clear differences in the parties' business models and market positioning as a result of which there is a significant asymmetry in the way the parties constrain each other pre-transaction. This limits any possible anti-competitive effects of the proposed concentration at the outset.

216. Moreover, there are numerous credible alternatives available to end customers wishing to travel from the Netherlands to Vancouver. In particular, Air Transat operates two direct flights per week (115) between Amsterdam and Vancouver. Furthermore there are several indirect alternatives, namely the flights of Air Canada, United Airlines, British Airways and Lufthansa, all of which operate between Amsterdam and Vancouver with a stopover duration of less than 150 minutes. (116)

217. In addition, there are several competing flights to these destinations from the airports of Brussels and Düsseldorf which constitute a viable alternative for a share of Dutch passengers, namely at least those living in the border regions, and thus exert at least some degree of competitive pressure on flights departing from Amsterdam. In particular, LTU operates two direct flights per week between Düsseldorf and Vancouver. Apart from that, British Airways offers 13 indirect flights per week from Düsseldorf and 13 indirect flights per week from Brussels. Also Air Canada operates 14 indirect flights per week from each of Düsseldorf and Brussels. Finally, Lufthansa offers 7 indirect flights per week from Düsseldorf. If all of these competing flights are included in the relevant market, the parties' combined market share is reduced to less than [60-70]*%.

218. As regards regulatory barriers to entry and expansion, the bilateral air services agreement concluded between Canada and the Netherlands provides for the designation of two authorised airlines by each country. KLM and Martinair are such designated airlines for the Netherlands while Air Canada is the only designated airline for Canada at this stage. The agreement does therefore currently not allow for the designation of another airline by the Netherlands. However, the bilateral air services agreement does not contain any limitations on charter flights. Accordingly, Air Transat currently operates charter services on the route AMS-YVR. Given that Air Transat is the parties' largest competitor on this route, regulatory barriers do not appear to be a major constraint for Air Transat.

219. Furthermore, on the basis of a proposal of the European Commission (117), the Council authorised the Commission on 2 October 2007 to open negotiations with Canada on a comprehensive aviation agreement. The aims of the negotiations, which started officially on 27 and 28 November 2007 in Brussels, are to establish an open aviation area between the EU and Canada, to replace the nationality clause by the Community designation clause, to remove additional market restrictions and to achieve a high level of regulatory convergence.

220. As regards other barriers to entry and expansion, including the availability of slots at Schiphol airport, reference is made to recitals 191-197 which describe the situation for flights between Schiphol airport and all long-haul destinations in general.

221. Finally, the market investigation did not provide any indication that, as a result of the proposed transaction, the parties would be in a position to raise prices, or reduce capacity or service levels on the route AMS-YVR and the Commission has not received any substantial complaints in this respect.

222. It follows that, after the implementation of the proposed concentration, the parties will continue to face sufficient competition from established airlines such as Air Canada, Air Transat, United Airlines, Lufthansa and British Airways which operate flights to Vancouver from Amsterdam as well as Brussels and Düsseldorf, while regulatory barriers to entry on the route AMS-YVR appear to be limited.

223. Moreover, as set out in section 6.2, the fact that KLM jointly controls Martinair pre-transaction is likely to reduce the anti-competitive effects of the proposed concentration. Finally, as discussed in section 8.4.1, the competitive pressure that Martinair currently exercises on KLM in passenger air transport in general is decreasing due to its financial situation and the need for the timely renewal of its fleet, which is ultimately controlled by KLM. Therefore the effects of the proposed concentration with respect to the parties' passenger air transport activities will in any event be limited.

Conclusion on the assessment with respect to the route AMS-YVR or AMS/DUS/BRU-YVR

224. In light of these circumstances, it can be concluded that the proposed concentration will not significantly impede effective competition in the common market or a substantial part of it with respect to this route.

8.4.3.3.2. Toronto

225. KLM operates 14 direct flights a week to Toronto. Martinair does not operate flights to Toronto in the winter season. During the summer season, it operates 4 direct flights to Toronto per week (2 flights per week with effect from 1 September 2008).

226. The table below shows the market shares of the parties and their competitors on the route Amsterdam-Toronto on a yearly basis for the year 2007/2008 as well as the market shares of the parties and their competitors for the hypothetical wider market comprising flights from Amsterdam, Düsseldorf and Brussels to Toronto on a yearly basis for the year 2006/2007.

<table>

227. Also on this route, the parties have a significant combined market share of more than [70-80]*%. However, this market share will go down since Martinair has reduced its frequencies to 2 flights per week. Furthermore, as explained in more detail in recitals 184-189, there are clear differences in the parties' business models and market positioning as a result of which there is a significant asymmetry in the way the parties constrain each other pre-transaction. This limits any possible anti-competitive effects of the proposed concentration at the outset.

228. Moreover, there are numerous credible alternatives available to passengers wishing to travel from the Netherlands to Toronto. Air Transat, which operates one direct flight per week, is again the second largest competitor on this route with a market of [10-20]*%, whereas indirect flights are operated, in particular, by of Air Canada (daily flights), United Airlines, British Airways and Lufthansa. As for Vancouver, all of these airlines offering indirect services operate between Amsterdam and Toronto with stopover durations of less than 150 min.

229. In addition, there are several competing flights to these destinations from the airports of Brussels and Düsseldorf. In particular, Lufthansa operates 6 direct flights per week between Düsseldorf and Toronto whereas Jet Airways serves Toronto with direct flights from Brussels 7 times per week. Apart from that, Air Canada operates 28 indirect flights per week from each of Düsseldorf and Brussels via Heathrow and 14 indirect flights from Düsseldorf to Toronto via Frankfurt, whereas British Airways operates 17 indirect flights per week from each of Brussels and Düsseldorf. Finally, Lufthansa offers another 13 indirect flights per week from Düsseldorf. If all of these competing flights are included in the relevant market, the parties' combined market share is reduced to less than 60%.

230. Regulatory barriers to entry and expansion and slot restrictions which may limit entry and expansion with regard to the route AMS-YYZ appear similarly limited as those described in recitals 218-219 for AMS-YVR. As regards other barriers to entry and expansion, including the availability of slots at Schiphol airport, reference is made to recitals 191-197 which describe the situation for flights between Schiphol airport and all long-haul destinations in general.

231. Finally, the market investigation did not provide any indication that, as a result of the proposed transaction, the parties would be in a position to raise prices, reduce capacity or service levels on the route AMS-YYZ and the Commission has not received any substantial complaints in this respect.

232. It therefore appears that, after the implementation of the proposed concentration, the parties will continue to face sufficient competition from established airlines such as Air Canada, Air Transat, United Airlines, Lufthansa and British Airways which operate competing flights to Toronto from Amsterdam as well as Brussels and Düsseldorf, while barriers to entry on the route AMS-YYZ again seem to be more limited than in case of other routes.

233. Moreover, as set out in section 6.2, the fact that KLM jointly controls Martinair pre-transaction is likely to reduce the anti-competitive effects of the proposed concentration. Finally, as discussed in section 8.4.1, the competitive pressure that Martinair currently exercises on KLM in passenger air transport in general is decreasing due to its financial situation and the need for the timely renewal of its fleet, which is ultimately controlled by KLM. Therefore the effects of the proposed concentration with respect to the parties' passenger air transport activities will in any event be limited.

Conclusion on the assessment with respect to the route AMS-YYZ or AMS/DUS/BRU-YYZ

234. In light of these circumstances, it can be concluded that the proposed concentration would not significantly impede effective competition in the common market of a substantial part of it with respect to this route.

8.4.3.3.3. Miami

235. The table below shows the market shares of the parties and their competitors on the route Amsterdam-Miami on a yearly basis for the year 2007/2008 as well as the market shares of the parties and their competitors for the hypothetical wider market comprising flights from Amsterdam, Düsseldorf and Brussels to Miami on a yearly basis for the year 2006/2007. 62

<table>

236. Also on this route, the parties have a significant combined market share of more than [70-80]*%. However, KLM, Air France, Delta/Northwest only offer indirect flights to Miami. Furthermore, as explained in more detail in recitals 184-189, there are clear differences in the parties' business models and market positioning as a result of which there is a significant asymmetry in the way the parties constrain each other pre-transaction. This limits any possible anti-competitive effects of the proposed concentration already at the outset.

237. Moreover, as for the routes AMS-YVR and AMS-YYZ, there are numerous credible alternatives available to passengers wishing to travel from the Netherlands to Miami. These alternatives include indirect services of British Airways and the American carriers United Airlines, Continental Airlines and US Airways. Again all competitors of the parties listed in the table in recital 235 operate indirect services between Amsterdam and Miami with stopover durations of less than 150 min. (120)

238. Furthermore, there are numerous competing flights to these destinations from the airports of Brussels and Düsseldorf. In particular, the German airline LTU operates 5 direct flights per week between Düsseldorf and Miami. Apart from that, American Airlines operates 14 indirect flights per week from Düsseldorf, British Airways serves Miami with 14 indirect flights per week from each of Brussels and Düsseldorf, and Lufthansa operates another 7 indirect flights per week from Düsseldorf. As for Vancouver and Toronto, the parties' combined market share would be reduced to less than 60%, if all of these competing flights were to be included in the relevant market.

239. As a result of the EU-U.S. comprehensive Air Transport Agreement which took effect on 30 March 2008, there are no meaningful regulatory barriers to entry or expansion on the route AMS-MIA. As a result of the same agreement on the same is valid for the routes DUS-MIA and BRU-MIA either. As regards other barriers to entry and expansion, including the availability of slots at Schiphol airport, reference is made to recitals 191-197 which describe the situation for flights between Schiphol airport and all long-haul destinations in general.

240. Finally, the market investigation did not provide any indication that, as a result of the proposed transaction, the parties would be in a position to raise prices, reduce capacity or service levels on the route AMS-MIA and the Commission has not received any substantial complaints in this respect.

241. It therefore appears that, after the implementation of the proposed concentration, the parties will continue to face sufficient competition from established airlines such as British Airways, United Airlines, Continental Airlines and US Airways which operate competing flights to Miami from Amsterdam as well as LTU, American Airlines and Lufthansa which serve Miami from Brussels or Düsseldorf or both. Moreover, barriers to entry on the route AMS-MIA seem to be even more limited than in the case of the AMS-YVR and AMS-YYZ routes.

242. Moreover, as set out in section 6.2, the fact that KLM jointly controls Martinair pre-transaction is likely to reduce the anti-competitive effects of the proposed concentration. Finally, as discussed in section 8.4.1, the competitive pressure that Martinair currently exercises on KLM in passenger air transport in general is decreasing due to its financial situation and the need for the timely renewal of its fleet, which is ultimately controlled by KLM. Therefore the effects of the proposed concentration with respect to the parties' passenger air transport activities will in any event be limited.

Conclusion on the assessment with respect to the AMS-MIA route or AMS/DUS/BRU-MIA route

243. In light of these circumstances, it can be concluded that the proposed concentration would not significantly impede effective competition in the common market or a substantial part of it with respect to this route.

8.4.3.4. Routes connecting Amsterdam, Brussels and Düsseldorf to holiday destinations in the Caribbean region and Mexico (Punta Cana, Cancun and Havana)

8.4.3.4.1. Punta Cana

244. The table below shows the market shares of the parties and their competitors on the Amsterdam/Düsseldorf/Brussels-Punta Cana route on a yearly basis for the year 2006/2007. 64

<table>

245. The parties' combined market share on this route amounts to [70-80]*%. Martinair operates 2 direct flights per week between Amsterdam and Punta Cana. Its pre-transaction market share on the basis of MIDT data is [60-70]*%. Based on MIDT data, Air France is only the fourth largest player on this route. Accordingly, the market share increment resulting from the proposed transaction is limited (compared to other routes on which the parties' activities overlap), amounting to only [0-5]*%. In addition, as explained in more detail in recitals 184-189, there are clear differences in the parties' business models and market positioning as a result of which there is a significant asymmetry in the way the parties constrain each other pre-transaction. This limits any possible anti-competitive effects of the proposed concentration already at the outset.

246. Moreover, Air France operates only indirect services on the AMS/DUS/BRU-PUJ route, whereas there are three other airlines which operate direct services on this route, namely the German airline LTU, ArkeFly and Jetairfly, a Belgian airline, also ultimately owned by the TUI group. LTU offers 2 direct services per week from Düsseldorf and its market share is about four times higher than that of Air France. Similarly, ArkeFly serves Punta Cana from Amsterdam with 2 direct services per week and Jetairfly operates two weekly flights to Punta Cana from Brussels. Finally, Condor, a German airline owned by the tour operator Thomas Cook, operates 5 indirect flights per week from each of Düsseldorf and Brussels (via Frankfurt).

247. It should be noted that Table 9, which shows the market shares of the parties and their competitors on the AMS/DUS/BRU-PUJ route, is based on MIDT data. As already set out in footnote 111, MIDT data usually does not include "direct" sales, that is sales made by the airlines themselves, and sales of seats on non-scheduled flights. Whereas these flaws in the data appear to apply to all airlines in more or less the same way, MIDT data seems to underestimate the market positions of two types of air carriers: (i) low cost carriers which generally tend to sell a higher proportion of seats directly through their website instead of through travel agents and (ii) airlines which are vertically integrated with a tour operator with respect to seats that are included in package holidays of the respective tour operator). Therefore, with respect to the AMS/DUS/BRU-PUJ route, MIDT does not report market shares for either ArkeFly (122) or Jetairfly. As a consequence, the size of the overall market is necessarily larger and therefore table 9 is likely to overstate the market shares of all airlines reported in the table, including those of the parties. (123)

Barriers to entry and expansion - AMS-PUJ

248. As regards regulatory barriers to entry, the bilateral air services agreement between the Netherlands and the Dominican Republic provides for the designation of only one airline by each country. Currently, the designated carrier for the routes between the Netherlands and the Dominican Republic is Martinair. Entry on the AMS-PUJ route is thus restricted by the applicable air services agreement and any competitor to Martinair can at this stage operate only non-scheduled services on the AMS-PUJ route. However, it should be noted that ArkeFly, although not being a designated carrier under the bilateral air services agreements, operates two direct flights from Amsterdam to Punta Cana as charter operations. The bilateral air services agreement concluded between the Netherlands and the Dominican Republic contains a nationality clause which has not yet been replaced by a Community designation clause. This constitutes a regulatory barrier for non-Dutch EU-based airlines wishing to receive the traffic rights for flights between the Netherlands and the Dominican Republic. According to the parties, however, the Dutch government intends to send a letter to the government of the Dominican Republic shortly in which it will request the Dominican Republic to regularise the bilateral air services agreement so as to include a Community designation clause.

249. In this regard the Commission's market investigation has revealed that, under the current Dutch regulation, more specifically the Decree Inclusive Tours (Besluit IT-Reizen) (124), the possibilities for non-scheduled airlines to sell "seats only", namely, seats sold directly to end customers, are limited. In principle, according to one of the market participants, the interpretation given to the Decree Inclusive Tours by the Dutch government is that for long-haul holiday destinations outside of Europe, non-scheduled airlines are only allowed to sell seats integrated in package holidays. Sales of "seats only" are permitted only with respect to flights on routes to destinations outside of Europe, on which no scheduled services are already being performed, unless the Dutch Minister considers it desirable from an aviation political view to allow sales of "seats only" by non-scheduled carriers.

250. However, according to this market participant, the Dutch government has not been enforcing this prohibition recently so that, in practice, ArkeFly can currently sell part of its capacity as "seats only" for flights to various destinations outside of Europe via its website, including Punta Cana. Although the situation of lack of enforcement may change in the future, this would not necessarily mean that ArkeFly would no longer be an option for customers that wish to purchase an unbundled seat. In fact, two market participants have pointed out to the Commission that a similar restriction on the sale of "seats only" by non-scheduled airlines with respect to flights to the Caribbean applies in Belgium. Nevertheless, it appears to be possible for airlines operating non-scheduled flights from Brussels to destinations in the Caribbean to sell "seats only" via their website, for example by bundling the sales of the respective flight tickets with one night hotel accommodation. (125)

Barriers to entry and expansion -DUS-PUJ and BRU-PUJ

251. As regards the DUS-PUJ and BRU-PUJ routes, the applicable bilateral air services agreements are less restrictive than in the case of the AMS-PUJ route. The bilateral air services agreement between Belgium and the Dominican Republic and between Germany and the Dominican Republic respectively both provide for multiple designations, that is to say each country can designate several airlines. However, the bilateral air services agreement between Belgium and the Dominican Republic limits the number of frequencies to a maximum of 14 per week for each contracting party. There are no restrictions as regards the number of frequencies under the bilateral air services agreement between Germany and the Dominican Republic.

252. As previously noted in recital 250, there are also regulatory restrictions on the sale of "seats only" by non-scheduled airlines in Belgium which are similar to those in the Netherlands. However, these restrictions do not appear to prevent non-scheduled airlines from selling "seats only" on flights between Brussels and Punta Cana.

253. As regards slot restrictions at the Schiphol, Düsseldorf and Brussels airports, reference is made to recitals 195-196 which describe the situation for flights between these three airports and all long-haul destinations in general.

254. On the whole, barriers to entry and expansion appear to be lower with respect to flights to Punta Cana from Düsseldorf and Brussels as compared to flights to this destination from Amsterdam.

High price elasticity of demand is likely to make post-transaction price increases unprofitable

255. Finally, the Commission's market investigation, as well as the customer survey conducted at Schiphol airport, have clearly indicated that, in the case of a price increase on the part of the merged entity on the AMS/DUS/BRU-PUJ route , a significant proportion of customers would substitute Punta Cana by other (long-haul) holiday destinations (126). This flexibility with regard to the destination constrains the merged entity to such an extent that, post-transaction, it would likely not be able to implement a significant non transitory price increase on this route on a lasting basis. Indeed, according to the results of the customer survey, 34% of the sample of customers who were travelling to Punta Cana indicated that they would switch to other destinations in case prices for flights to Punta Cana increased by 5-10% (127). At the same time, 13% of customers would refrain from travelling in response to a price increase. Thus, only 53% of customers would have kept travelling to their destination if the price was 5-10% more expensive With regard to Martinair's customers only, the proportion of customers who would keep going to the same destination in response to a price increase of 5-10% is somewhat higher amounting to 65%. (128)

256. Indeed, a vast majority of the interviewed customers travelling to Punta Cana, (91%) were leisure passengers.

257. Accordingly, the majority of tour operators who responded to the Commission's market investigation indicated that, in the case of a 5-10% price increase on the AMS-PUJ route, a proportion of the Dutch leisure travellers would substitute Punta Cana, a typical "sun and beach" holiday destination by other "sun and beach" holiday destinations in the Caribbean or elsewhere, and another proportionwould stop travelling altogether.

258. In addition, the parties argue that the number of passengers who need to switch from a flight to make a price increase unprofitable is very low. According to the parties, this is due to the fact that the incremental cost of serving an additional passenger on a flight is almost negligible, so that the additional revenues of the marginal passengers are almost entirely profitable. Indeed, the large majority of competitors who responded to the market investigation confirmed that airlines are generally very sensitive to even a small decline in the number of passengers because this can have a significant impact on the profitability of their services. Given the high load factor required for profitable passenger air transport operations to long-haul leisure destinations (see recitals 176-179), this appears to be particularly true insofar as passenger air transport operations to this type of destinations are concerned.

259. Finally, the parties argue that they are not able to price-discriminate between passengers that are sensitive to the destination and those that are not and that consequently, the group of leisure travellers who are sensitive to a specific destination is protected against price increases by the threat that other passengers who are not sensitive to a particular destination would switch to another destination in response to a price increase. Several competitors have confirmed that, when selling flight tickets to end customers via travel agents, through the airline's website or through other distribution channels, it is indeed not possible for an airline to distinguish those customers who are destination-sensitive (for example, VFR travellers) from those who are not and to charge higher prices to the former as compared to the latter. It therefore appears that, as long as a significant proportion of customers will switch to other destinations or will stop travelling at all in the case of a price increase, not only these customers, but also destination-sensitive customers would be protected from a price increase.

260. In light of all these circumstances, the fact that the parties cannot price-discriminate between different groups of customers, and, in particular, in view of the high load factor required for profitable passenger air transport operations to long-haul holiday destinations (129) and the significant proportion of passengers who would switch away in case of a significant non transitory price increase on the AMS/DUS/BRU-PUJ route, it appears unlikely that any such price increase on the part of the merged entity would be profitable.

261. Moreover, as set out in section 6.2, the fact that KLM jointly controls Martinair pre-transaction is likely to reduce the anti-competitive effects of the proposed concentration. Finally, as discussed in section 8.4.1, the competitive pressure that Martinair currently exercises on KLM in passenger air transport in general is decreasing due to its financial situation and the need for the timely renewal of its fleet, which is ultimately controlled by KLM. Therefore the effects of the proposed concentration with respect to the parties' passenger air transport activities will in any event be limited.

Conclusion on the assessment with respect to the AMS/DUS/BRU-PUJ route

262. Accordingly, it must be concluded that, with respect to the supply of passenger air transport services to end customers on the AMS/DUS/BRU-PUJ route, the proposed concentration would not significantly impede effective competition in the common market or a substantial part of it.

8.4.3.4.2. Cancun

263. The table below shows the market shares of the parties and their competitors on the Amsterdam/Düsseldorf/Brussels-Cancun route on a yearly basis for the calendar year 2007.

<table>

264. On this route, the parties achieve a combined market share of [80-90]*%. Martinair operates 2 direct flights per week between Amsterdam and Cancun. Its pre-transaction market share on the basis of MIDT data is [70-80]*%. KLM and its partners Delta Airlines (131) and Northwest Airlines (132) operate indirect services between Amsterdam and Cancun and, based on MIDT, they are the second largest player on this route with a market share of [10-20]*%. However, this combined market share overstates the combined pre-merger market share of KLM and Delta Airlines given that, with respect to flights between Amsterdam and Cancun, KLM and Delta Airlines do not offer competitive stopover durations. The stopover durations on one or both legs of their indirect flights clearly exceed the 150 min. maximum stopover duration for such flights to be part of the same market as direct flights (see recitals 142-144). More specifically, the stopover duration with regard to KLM's indirect flights between Amsterdam and Cancun is 240 min. on the leg AMS-CUN and 160 min. on the leg CUN-AMS and also Delta Airlines serves Cancun from Amsterdam with a stopover duration of 240 min. on the leg AMS-CUN (on the leg CUN-AMS, the stopover duration is 135 min.).

265. In addition, as set out in more detail in recitals 184-189, there are clear differences in the parties' business models and market positioning as a result of which there is a significant asymmetry in the way the parties constrain each other pre-transaction. This limits any possible anti-competitive effects of the proposed concentration already at the outset.

266. Moreover, post-transaction there will be several alternatives left for Dutch leisure travellers wishing to travel to Cancun and some of these alternatives might be much more attractive to Dutch leisure travellers than the indirect services offered by KLM and Delta Airlines. This is likely to be the case, in particular, with respect to the competing direct services offered on this route by ArkeFly and LTU. ArkeFly serves Cancun from Amsterdam with two direct flights per week and LTU offers 1 direct service per week from Düsseldorf. In addition, Condor (belonging to the tour operator Thomas Cook) operates 3 indirect flights per week from Düsseldorf, Iberia operates 7 indirect flights per week from each of Düsseldorf and Brussels and several American airlines offer a limited number of indirect services from Brussels or Düsseldorf, but hold only minor market shares.

267. For the reasons set out in footnotes 114 and 122 and recital 247, it should be noted that also table 10, which shows the market shares of the parties and their competitors on the AMS/DUS/BRU-CUN route, is likely to overstate the market shares of all other airlines listed in the table, including those of the parties, because this table is based on MIDT data and it does not properly report market shares for ArkeFly which does not operate scheduled services on these routes.

Barriers to entry and expansion - AMS-CUN

268. As regards regulatory barriers to entry, the bilateral air services agreement between the Netherlands and Mexico provides for the designation of only one airline by each country. The designated carrier by the Netherlands is KLM. No other airline from the Netherlands can thus carry out scheduled services between the Netherlands and Mexico. However, it should be noted that ArkeFly, although not being a designated carrier under the bilateral air services agreements, operates two direct flights from Amsterdam to Cancun as non-scheduled services. Also Martinair provides semi-scheduled passenger services between Amsterdam and Cancun on the basis of a charter permit. These examples show that, although the bilateral air services agreement provides only for single designation by each country, this does not necessarily prevent other airlines from successfully operating flights on the AMS-CUN route. In addition, it should be noted that, on 23 May 2008, the European Commission has initialled an agreement with Mexico, which, once concluded, removes the barrier initially created by nationality restrictions in bilateral agreements between Member States and Mexico.

269. The restrictions under the current Dutch regulation with respect to the sale of "seats only" by non-scheduled airlines described in paragraphs 249-250 also apply with respect to the sale of seats on ArkeFly's flights between Amsterdam and Cancun. However, as set out in recital 250, these restrictions are currently not being enforced so that, in practice, ArkeFly can currently sell "seats only" on the AMS-CUN route. Apart from that, even if these restrictions were to be enforced, they could be overcome by offering the flight together in a bundle with accommodation for only one night. In any case, similar restrictions applying in Belgium do not appear to prevent airlines operating non-scheduled flights between Brussels and Cancun from selling "seats only" on such flights. (133)

Barriers to entry and expansion - DUS-CUN and BRU-CUN

270. As regards regulatory barriers to entry for the BRU-CUN and DUS-CUN routes, the bilateral air services agreement between Belgium and Mexico provides for dual designation. It limits the number of frequencies to four per week, which can, however, be increased through consent of the aeronautical authorities should market conditions justify such an increase. The bilateral air services agreement between Germany and Mexico provides for single designation plus the designation of a second carrier by Germany for routes from Germany to Cancun. It provides for limitations in the number of frequencies, which are, however, set at a level that appears relatively generous as compared to limitations under other bilateral air services agreements (10 weekly frequencies from Germany to Cancun).

271. As already noted in paragraphs 250 and 269, in Belgium, there are regulatory restrictions on the sale of "seats only" by non-scheduled airlines which are similar to those existing in the Netherlands, but these restrictions do not appear to prevent non-scheduled airlines from selling "seats only" on flights between Brussels and Cancun.

272. As regards other barriers to entry, including the availability of slots at the Schiphol, Düsseldorf and Brussels airports, reference is made to recitals 191-197 which describe the situation for flights between these three airports and all long-haul destinations in general.

High price elasticity of demand is likely to make post-transaction price increases unprofitable

273. Finally, the Commission's market investigation as well as the customer survey conducted at Schiphol airport, have clearly indicated that, also in case of price increases on the part of the merged entity on the AMS/DUS/BRU-CUN routes, a significant proportion of customers would substitute Cancun by other (long-haul) holiday destinations thereby constraining the merged entity to such an extent that, post-transaction, it would not be able to implement on a lasting basis a significant non transitory price increase on the AMS/DUS/BRU-CUN route.

274. Also for Cancun the results of the customer survey suggest that, in case of such a price increase, a significant proportion of Dutch leisure travellers flying from Amsterdam to Cancun could switch to other long-haul "sun and beach" holiday destinations in the same or in other geographic regions. Indeed, according to the results of the survey, 25% of the sample of customers who were travelling to Cancun indicated that they would switch to other destinations in case prices for flights to Cancun increased by 5-10%. At the same time, 8% of them would refrain from travelling in response to a price increase. Only 67% of customers would have kept travelling to their destination if the price was 5-10% more expensive. The proportion of customers who would keep going to the same destination in response to a price increase of 5-10% gets higher, 79%, if we focus on Martinair passengers only. (134) As in the case of Punta Cana, these results reflect the composition of the sample in terms of trip's purpose. Indeed, almost the entire sample of passengers travelling to Cancun, 95%, was travelling for leisure purposes.

275. Similarly, also most tour operators who responded to the Commission's market investigation indicated that, in case of a 5-10% price increase on the AMS-CUN route, part of the Dutch leisure travellers would substitute Cancun, a typical "sun and beach" holiday destination by other "sun & beach" holiday destinations in the Caribbean or in other geographic regions and that another part would stop travelling altogether.

276. Moreover, as set out in paragraph 179, the large majority of competitors who responded to the market investigation confirmed that airlines are generally very sensitive to only a small decline in the number of passengers because this can have a significant impact on the profitability of their services, in particular in case of passenger air transport operations to long-haul leisure destinations.

277. Finally, as set out in recital 259, the parties cannot price-discriminate between passengers that are sensitive to the destination and those that are not. Therefore, as long as a significant proportion of customers will switch to other destinations or will stop travelling at all in case of a price increase, the group of leisure travellers who are sensitive to a specific destination is protected against price increases by the threat that other passengers who are not sensitive to a particular destination will switch away in response to such a price increase.

278. In light of all these circumstances, the fact that the parties cannot price-discriminate between different groups of customers and, in particular, in view of the high load factor required for profitable passenger air transport operations to long-haul holiday destinations (135) and the significant proportion of passengers who would switch away in case of a significant non transitory price increase on the AMS/DUS/BRU-CUN route, it appears unlikely that any such price increase on the part of the merged entity would be profitable.

279. Moreover, as set out in section 6.2, the fact that KLM jointly controls Martinair pre-transaction is likely to reduce the anti-competitive effects of the proposed concentration. Finally, as discussed in section 8.4.1, the competitive pressure that Martinair currently exercises on KLM in passenger air transport in general is decreasing due to its financial situation and the need for the timely renewal of its fleet, which is ultimately controlled by KLM. Therefore the effects of the proposed concentration with respect to the parties' passenger air transport activities will in any event be limited.

Conclusion on the assessment with respect to the AMS/DUS/BRU-CUN route

280. Consequently, it must be concluded that, also with respect to the supply of passenger air transport services to end customers on the AMS/DUS/BRU-CUN route, the proposed concentration would not significantly impede effective competition in the common market or a substantial part of it.

8.4.3.4.3. Havana

281. The table below shows the market shares of the parties and their competitors on the route Amsterdam/Düsseldorf/Brussels-Havana on a yearly basis for the year 2006/2007.

<table>

282. The parties' combined market share on this route amounts to [80-90]*%. Martinair operates 3 direct flights per week between Amsterdam and Havana. It is the only airline which operates direct flights from any of the three airports of Amsterdam, Düsseldorf and Brussels to Havana. Although Air France has the second largest market share on this route, it operates only indirect services on the AMS/DUS/BRU-HAV route. Moreover, Iberia operates 7 indirect flights per week via Madrid from each of the airports of Schiphol, Düsseldorf and Brussels and has a market share which is only slightly lower than that of Air France. Finally, also Virgin Atlantic offers indirect services via London Gatwick on this route (136), but only has a negligible market share of [0-5]*%. (137) In addition, as explained in more detail in recitals 184-189, there are clear differences in the parties' business models and market positioning as a result of which there is a significant asymmetry in the way the parties constrain each other pre-transaction. This limits any possible anti-competitive effects of the proposed concentration already at the outset.

Barriers to entry and expansion - AMS-HAV

283. As regards regulatory barriers to entry, the bilateral air services agreement between the Netherlands and Cuba provides for the designation of two airlines by each country. Currently, the only carrier appointed as a designated carrier by the Netherlands for the routes between the Netherlands and Cuba is Martinair. The maximum number of frequencies that can be operated by the designated airline(s) of each country is limited to seven per week for scheduled flights (route I) and to six per week for non-scheduled leisure services (route II). At present five frequencies per week are operated by Martinair for scheduled services and one frequency per week for non-scheduled services. The bilateral air services agreement concluded between the Netherlands and Cuba contains a nationality clause which has not yet been replaced by a Community designation clause. This constitutes a regulatory barrier for non-Dutch EU-based airlines wishing to receive the traffic rights for flights between the Netherlands and Cuba. According to the parties, however, the Dutch government has sent a letter to the government of Cuba in August 2008, in which it has requested Cuba to regularise the bilateral air services agreement so as to include a Community designation clause. Based on various oral contacts, the government of the Netherlands expects a positive response to this request.

284. Entry on the AMS-HAV route is thus restricted by the applicable air services agreement mainly through limitations on the number of frequencies. A second designated carrier can only operate flights in as far as the total maximum number of frequencies provided for by the bilateral air services agreement is respected. Taking into account the flights currently operated by Martinair, this would allow a second designated carrier to operate two flights per week for route I and five flights per week for route II.

285. The restrictions under the current Dutch regulation with respect to the sale of "seats only" by non-scheduled airlines described in recitals 249-250 also apply with respect to the sale of seats on non-scheduled flights between Amsterdam and Havana. However, for the reasons set out in recital 250, these restrictions would not necessarily prevent ArkeFly or any other airline from entering the AMS-HAV route by starting non-scheduled operations on this route.

286. Indeed, in response to the Commission's market investigation ArkeFly indicated that it would be able to accommodate extra demand on the AMS-HAV route by starting (charter) operations on this route if the merged entity were to increase prices as a result of which customers were wishing to switch to ArkeFly's services. In this respect, ArkeFly has indicated that it currently uses part of the capacity of its long-haul fleet (Boeing 767) for short term and ad-hoc rental to third parties (ACMI). (138) ArkeFly explained that this is not its core business and that it could therefore use this capacity at short notice to extend the frequency on existing routes or to start operating flights on routes where it is currently not present. As a result several extra rotations per week could be added within a short period of time without the need for an expansion of the fleet in the short term.

287. Moreover, ArkeFly also noted that, in the mid term (six months to one year and longer), it could increase its capacity even further by putting into operation new and/or leased aircrafts, provided that demand is sufficient to ensure profitable operations on the respective routes. In this respect, also several other competitors confirmed that there is a market for the leasing of aircrafts which makes it possible to lease additional aircrafts within a time frame of several months up to one year. In addition, ArkeFly explained that, in the mid term, new and/or leased aircrafts could also come from other companies belonging to the TUI group. (139)

288. It therefore appears that ArkeFly is likely to constrain the parties post-transaction on the AMS/DUS/BRU-HAV routes to the extent that there is a realistic threat that ArkeFly might enter this route in case post-transaction price increases on the part of the merged entity make such entry attractive for ArkeFly. In combination with the price elasticity of demand for this route, which will be discussed in recitals 292-294, this threat of entry would likely deter the parties from raising prices on this route post-merger.

Barriers to entry and expansion - DUS-HAV and BRU-HAV

289. As regards regulatory barriers to entry for the BRU-HAV and DUS-HAV routes, the applicable bilateral air services agreements are less restrictive. The bilateral air services agreements between Belgium and Cuba and between Germany and Cuba respectively both provide for multiple designations, that is to say, each country can designate several airlines. Moreover, the number of frequencies that may be operated by the respective designated airlines is significantly higher under these two agreements than under the bilateral air services agreement between the Netherlands and Cuba.

290. Again, insofar as there are regulatory restrictions on the sale of "seats only" by non-scheduled airlines in Belgium which are similar to those existing in the Netherlands (see recitals 250 and 271, such restrictions do not appear to prevent non-scheduled airlines from selling "seats only" on flights between Brussels and various long-haul destinations outside of Europe.

291. As regards other barriers to entry, including the availability of slots at the Schiphol, Düsseldorf and Brussels airports, reference is made to recitals 191-197 which describe the situation for flights between these three airports and all long-haul destinations in general.

Price elasticity of demand is likely to make post-transaction price increases unprofitable

292. Finally, the Commission's market investigation as well as the customer survey conducted at Schiphol airport, suggest that, in case of price increases on the part of the merged entity on the AMS/DUS/BRU-HAV routes, a significant proportion of customers would substitute Havana by other (long-haul) holiday destinations or would not have travelled at all (the proportion of customers that would be lost by the merging parties as a consequence of a price increase is lower than in the case of Punta Cana and Cancun, though).

293. More specifically, while most tour operators and travel agents who responded to the market investigation were sceptical whether a significant proportion of customers would switch away from Havana to other destinations, the results of the customer survey indicate that still 14% of the of customers travelling to Havana would switch to other destinations in case the price of their trip would be increased by 5-10%. An additional 7% of them indicated that, in case of such a price increase, they would not travel at all. Accordingly, about 20% of the sample of customers travelling to Havana would be lost by the merging parties in case of 5-10% price increase.

294. It should be noted, however, that the indications which can be drawn from the customer survey on the Amsterdam - Havana route are rather limited. That is due to the small number of responses which were collected on this route (only 45 passengers travelling to Havana expressed a definite answer on the question relative to substitutability at destination). This is well reflected in the size of the 95% confidence interval for the loss of customers due to a price increase in the Amsterdam - Havana route: the confidence interval is, indeed, 8% - 32%. (140)

295. To provide evidence that any price increase by the parties on the Amsterdam-Havana route would be unprofitable, the parties submitted a study on demand elasticity. (141) The study provides an estimate of the 'critical loss', that is, the minimum loss of sales which will make a price increase unprofitable. According to the results presented in the study, the critical loss on the Amsterdam-Havana route for the merging parties would be [5-10]*% in 2007. This would mean that a price increase would become unprofitable if it leads to a loss of at least [5-10]*% of their passengers. As the parties consider that any price increase would be necessarily followed by a higher percentage loss of passengers, they claim that they have no incentive to increase prices on the Amsterdam-Havana route post-transaction.

296. The critical loss analysis can be a useful tool to determine the parties' scope for price increases post-transaction. However, in this particular case for at least three reasons, the reliability of the result is, doubtful. Firstly, the study claims that the gross margins on passengers are very high. According to the critical loss analysis, whenever gross margins are high, the critical loss is small, that is,it is sufficient that a small number of customers would stop buying the good in order to make unprofitable a 5% price increase. However, high gross margin can also indicate small actual loss. Indeed, if it is assumed that pre-merger the firms are maximizing their profits, they would set the price (hence their gross margins) to be equal to the inverse of the elasticity of demand they face. It turns out that high margins signal low elasticity of demand. Thus, in order to present a coherent story, the study would need to explain this contradiction (that is: if the actual loss is low a price increase could be profitable in spite of a low critical loss). For instance, the study should address the issue of why the demand is very inelastic when the price is decreased but very elastic when the price is increased (this is the case if there is a 'kink' in the underlying demand curve). Otherwise, the study should address the cost structure of KLM and Martinair and explain whether it would render any price decrease unprofitable (this could be related to the considerable amount of fixed costs which have to be sustained in order to expand capacity in the airline sector). However, the study fails to provide any coherent explanation of these inconsistencies.

297. The second reason concerns the fact that the analysis is carried out on a per flight basis with a short term perspective. In this way, the analysis fails to account for changes in the frequencies and in the flight schedule which could be implemented by KLM and Martinair in order to achieve additional cost savings. In other words, if KLM-Martinair were to raise the price by 5%, a certain amount of passengers will be lost. That would have a negative impact on sales but also a negative impact on total costs. In the short term the negative impact on cost is limited to the elimination of the (relatively low) marginal cost due to each of the lost passengers. However, in the medium term, the negative impact on costs can be relatively high if the reduction in the number of passengers is sufficient to make it convenient to stop a flight. If that is taken into account, a price increase can be more profitable (as the operating costs for a flight are rather high). By not taking this effect into account, the critical loss analysis could yield biased results.

298. Finally, the third reason relies on how the incremental costs per passengers are estimated. Indeed, the study assumes that only 9% of variable costs of a flight are related to the number of passengers. However, some costs which are presumably related to the number of passengers (for example "fuel", "catering handling", "credit card fees", "handling fees" etc.) are excluded from such estimation and this would likely impact the size of the estimated critical loss.

299. For these reasons, the estimate of the critical loss provided by the parties can only offer a rough indication of the amount of loss of customers which would render a price increase unprofitable. In this context, however, it should be noted that the average loss of respondents due to a price increase as indicated by the customer survey is about 4 times the critical loss calculated by the parties' study.

300. In addition, as set out in recital 179, the market investigation confirmed that airlines are generally very sensitive to only a small decline in the number of passengers because this can make the difference between profitable and unprofitable services, in particular insofar as passenger air transport operations to long-haul leisure destinations are concerned.

301. Finally, as explained in recital 259, there is no possibility for the parties to clearly distinguish between passengers that are sensitive to the destination and those that are not and to charge higher prices to the former as compared to the latter. Therefore, as long as a significant proportion of customers will switch to other destinations or will stop travelling at all in case of a price increase, the group of leisure travellers who are sensitive to a specific destination is protected against price increases by the threat that other passengers who are not sensitive to a particular destination will switch away in response to such a price increase.

302. In t light of these circumstances, despite the fact that the price elasticity of demand is lower in case of Havana than in case of more typical "sun and beach" holiday destinations, it appears more likely than not that any significant non transitory price increase on the part of the merged entity would be unprofitable.

303. Moreover, as set out in section 6.2, the fact that KLM jointly controls Martinair pre-transaction is likely to reduce the anti-competitive effects of the proposed concentration. Finally, as discussed in section 8.4.1, the competitive pressure that Martinair currently exercises on KLM in passenger air transport in general is decreasing due to its financial situation and the need for the timely renewal of its fleet, which is ultimately controlled by KLM. Therefore the effects of the proposed concentration with respect to the parties' passenger air transport activities will in any event be limited.

Conclusion on the assessment with respect to the AMS/DUS/BRU-HAV route

304. Consequently, it must be concluded that, with respect to the supply of passenger air transport services to end customers on the AMS/DUS/BRU-HAV route, the proposed concentration would not significantly impede effective competition in the common market or a substantial part of it.

8.4.3.5. Routes connecting Amsterdam to the Netherlands Antilles (Amsterdam-Curacao and Amsterdam-Aruba)

305. In line with the Commission's findings with regard to the relevant market in recitals 122-148, for the purpose of the present decision, the effects of the proposed concentration with regard to flights connecting Amsterdam to Aruba and Curacao, the proposed transaction needs to be primarily assessed on the basis of the traditional O&D approach.

306. In this respect, it must be recalled that (i) there are no flights to Aruba or Curacao from either Düsseldorf or Brussels and (ii) the market investigation suggested that, given historic, cultural, and linguistic ties to the Netherlands (Aruba and Curacao are former Dutch colonies), a relatively large proportion of Dutch customers intending to fly to Aruba and Curacao is destination-sensitive (and thus less sensitive to price) and therefore less willing to consider switching to alternative "sun and beach" destinations in the event of a price increase.

8.4.3.5.1. Market positions on the Amsterdam-Curacao and Amsterdam Aruba routes

Curacao

307. The table below shows the market shares of the parties and their competitors on the Amsterdam-Curacao route on a yearly basis for the year 2007/2008.

<table>

308. KLM operates 7 weekly flights to Curacao, whereas Martinair serves this destination 5 times a week during the winter season and 4 times a week during the summer season. The parties' combined market share on this route is relatively high reaching [80-90]*% in 2007/2008. In addition, both parties operate direct flights on this route and the market share increment resulting from the proposed concentration is considerable, not only when comparing it to the increments on other routes connecting Amsterdam to the Greater Caribbean region. The only competitor on this route, ArkeFly, operates 6 weekly flights to Curacao and reaches a market share of [10-20]*% despite the fact that it uses most of its capacity captively for inclusion in package holidays offered by TUI or other tour operators belonging to the same corporate group as TUI and ArkeFly.

309. As there are no alternative flights to Curacao from Brussels and Düsseldorf, the parties' market shares are the same for the wider market defined in section 8.3.2 comprising flights to Curacao from Amsterdam, Düsseldorf and Brussels.

Aruba

310. The situation is similar on the AMS-AUA route. The table below shows the market shares of the parties and their competitors on the Amsterdam-Aruba route on a yearly basis for the year 2007/2008.

<table>

311. As for Curacao, these market shares are the same for the wider market defined in section 8.3.2, comprising flights to Aruba from Amsterdam, Düsseldorf and Brussels, as there are no flights to Aruba from Düsseldorf or Brussels.

312. The parties' combined market share on this route will amount to [80-90]*%. KLM operates 4 direct flights per week on this route, whereas Martinair serves this destination with direct flights twice a week during the summer season and 3 times a week during the winter season. ArkeFly, which operates 3 direct flights per week between Amsterdam and Aruba, has a market share of [10-20]*% on this route, despite the fact that, also for this route, ArkeFly uses most of its capacity captively for inclusion in package holidays.

313. Despite the fact that some American carriers offer indirect services via one of their U.S. hubs, there seem to be few, if any, indirect flights that would pass the substitutability test, that is to say, that would extend the travelling time by a maximum of 150 minutes. In addition, such alternative indirect flights offered by US airlines such as United Airlines appear to be much more expensive than the average Martinair ticket. Therefore also for Aruba, ArkeFly will be the only remaining competitor post-transaction.

8.4.3.5.2. Barriers to entry

Curacao

314. As regards regulatory barriers to entry for the AMS-CUR route, operations on this route fall within the application of the Protocol on Air Transport concluded between the Netherlands and the Netherlands Antilles. This protocol provides for multiple designations of carriers and does not contain any restrictions on the number of frequencies that can be operated. Consequently, both ArkeFly and Martinair could start operating on this route as designated carriers (in addition to the existing designated carrier KLM) in 2005 and 2006 respectively. According to the parties, the Dutch government has confirmed that the Community designation clause is currently applicable within the regime of the Protocol on air transport concluded between the Netherlands and the Netherlands Antilles on the basis of the internal constitutional framework of the Kingdom of the Netherlands and that the Netherlands would be willing to affirm this towards the Netherlands Antilles, should that be necessary.

315. As regards regulatory barriers to entry for the DUS-CUR routes, Germany and the Netherlands Antilles have recently concluded a new bilateral air services agreement. This agreement provides for multiple designations. The number of frequencies is subject to the approval of the aeronautical authorities. So far, no German or Dutch Antillean airlines have made use of the traffic rights granted under this agreement.

316. As regards regulatory barriers to entry for the BRU-CUR routes, Belgium and the Netherlands Antilles have also recently concluded a bilateral air services agreement. This agreement provides for multiple designations and does not contain any restrictions on the number of frequencies that can be operated. Again no Belgian or Dutch Antillean airlines have made use of the traffic rights granted under this agreement so far.

317. According to several carriers based at other airports which responded to the market investigation, in particular carriers based in Brussels and Düsseldorf, many of them indicated that, whilst the Netherlands Antilles are especially attractive destinations for Dutch passengers, non-Dutch speaking passengers are not particularly keen on travelling to these islands. Accordingly, outside of the Netherlands there would not necessarily be sufficient demand to economically justify entry into these markets.

Aruba

318. As regards regulatory barriers to entry for the AMS-AUA route, operations between the Netherlands and Aruba are governed by the Protocol on Air Transport. This protocol provides for multiple designations and does not contain any restrictions on the number of frequencies that can be operated. There is therefore no major regulatory barrier. Moreover, according to the parties, the Dutch government has confirmed that the Community designation clause is currently applicable within the regime of the Protocol on Air Transport concluded between the Netherlands and Aruba on the basis of the internal constitutional framework of the Kingdom of the Netherlands and that the Netherlands would be willing to affirm this towards Aruba, should that be necessary.

319. As regards regulatory barriers to entry for the BRU-AUA route, the applicable bilateral air services agreement between Belgium and Aruba provides for single designation. It does not contain any restrictions on the number of frequencies that can be operated. So far, no Belgian or Aruban airlines have made use of the traffic rights granted under this agreement. There is no existing bilateral air services agreement between Aruba and Germany. However, even if a bilateral air services agreement was to be concluded between Germany and Aruba in the near future, for the reasons set out in recital 317, carriers based in Germany would be rather unlikely to enter the DUS-AUA route.

No restrictions for ArkeFly to sell "seats only" on flights between AMS and CUR/AUA

320. The restrictions under the current Dutch regulation with respect to the sale of "seats only" by non-scheduled airlines described in recitals 249-250 also apply with respect to the sale of seats on non-scheduled flights connecting Amsterdam to Aruba and Curacao. However, considering that the applicable bilateral air services agreements provide for multiple designations and do not contain any limitations with respect to the number of frequencies that can be operated on these routes, it may be easier and possibly also more attractive for airlines to operate scheduled instead of non-scheduled services on the AMS-CUR and AMS-AUA routes. The described restrictions on the sale of "seats only", however, do not apply to scheduled operations. Even ArkeFly, which operates non-scheduled flights to all other long-haul or short-haul destinations it serves, is a designated carrier and operates scheduled services on these two routes. In any case, for the reasons set out in recital 250, the restrictions on the sale of "seats only" would not necessarily prevent airlines from entering the AMS-CUR and AMS-AUA routes by starting non-scheduled operations on this route.

Other barriers to entry and expansion

321. As regards other barriers to entry, including the availability of slots at the Schiphol, Düsseldorf and Brussels airports, reference is made to recitals 191-197 which describe the situation for flights between these three airports and all long-haul destinations in general.

8.4.3.5.3. Asymmetry/closeness of competition

322. As already noted in recitals 184-189, there are clear differences between KLM and Martinair in terms of business model and market positioning. Also with respect to Aruba and Curacao, the customer survey has shown a considerable degree of asymmetry between the parties. End customers seem to have a preference for KLM: 23% of KLM passengers travelling to Aruba asserted that, before buying their flight, they considered Martinair as potential alternative to KLM (as compared to 18% who considered ArkeFly); on the other hand, 48% of KLM's customers replied that they did not consider any other airline than KLM. By way of contrast, in reply to the same question, only 18% of Martinair customers did not consider any other airline before purchasing their trip, while 55% considered KLM (as compared to 15% who considered ArkeFly).

323. Similarly, 47% of KLM passengers travelling to Curacao asserted that they did not consider any other airline before buying their trip, while only 24% and 27% considered Martinair and ArkeFly respectively. On the other hand, 51% of Martinair passengers to Curacao considered KLM as potential alternative, 17% considered ArkeFly and 21% did not consider any other airline. (142)

324. This shows that KLM exerts significantly more competitive pressure on Martinair than vice versa. In addition, as indicated by the market investigation and the customer survey, Martinair and ArkeFly both tend to have a higher percentage of package tour customers than KLM. It thus appears that, in terms of market positioning, Martinair is closer to ArkeFly than to KLM.

8.4.3.5.4. Competitive constraints exercised by ArkeFly

325. Moreover, the merged entity would continue to be constrained by ArkeFly on the routes connecting Amsterdam to Curacao and Aruba. ArkeFly achieves a market share of around [10-20]*% with respect to both destinations. It operates scheduled flights on the AMS-AUA and AMS-CUR routes (and therefore the restrictions on the sale of "seats only" under the Dutch Decree Inclusive Tours described in more detail in recitals 249-250 do not apply to ArkeFly for these two routes).

326. Although ArkeFly uses its current capacity to a large extent captively by selling seats on its flights mainly to tour operators belonging to the TUI group for inclusion in package holidays, on the AMS-CUR and AMS-AUA routes [a significant proportion]* of ArkeFly's sales are sales to end customers. This is significantly more than in case of the other routes on which the parties' passenger air transport operations overlap and on which ArkeFly is also active. (143)

327. Moreover, in response to the Commission's market investigation, ArkeFly indicated that it would be able to accommodate extra demand on the AMS-AUA and AMS-CUR routes, for example by adding additional frequencies, if the merged entity were to increase prices as a result of which customers were wishing to switch to ArkeFly's services. As previously explained in more detail (see recital 286), ArkeFly has indicated that it currently uses part of the capacity of its long-haul fleet for short term and ad-hoc rental to third parties (ACMI). As this is not its core business, ArkeFly noted that it could use this capacity at short notice to extend the frequency on existing routes or to start operating flights on routes where it is currently not active. As a result, several extra rotations per week could be added within a short period of time without the need for an expansion of the fleet in the short term.

328. For a carrier already established at Schiphol airport and operating on the two routes, barriers to expansion of capacity are relatively low as it can benefit from economies of scale with respect to its operations. In fact, ArkeFly has already increased its frequencies to Curacao from 5 to 6 per week on 1 June 2008 and from 6 to 7 per week on 1 November 2008.

329. In addition, ArkeFly noted that, in the mid term (six months to one year and longer), it could increase its capacity even further by putting into operation new and/or leased aircrafts, noting that such new and/or leased aircrafts could also come from other companies belonging to the TUI group (see also recital 287).

330. The market investigation has not provided any indication that ArkeFly would have an incentive to follow KLM's price increases rather than to undercut KLM's price in order to gain market share. Indeed, ArkeFly recently invested in additional capacity on the routes to the Netherlands Antilles (on the route AMS-CUR it added two flights per week within the last six months) which indicates commitment to act as a competitive constraint in case of price increases by the merged entity.

331. Therefore, ArkeFly is likely to constrain the parties post-transaction on the AMS-CUR and AMS-AUA routes as it appears to have the ability and the incentive to do so. In combination with the price elasticity of demand for this route, which will be discussed in recitals 332-337, this threat would likely deter the parties from raising prices on this route post-merger.

8.4.3.5.5. Price elasticity of demand is likely to make post-transaction price increases unprofitable

332. Finally, similar to the other Caribbean destinations served by the parties, the Commission's market investigation as well as the customer survey conducted at Schiphol airport have indicated that, in case of price increases on the part of the merged entity on the AMS-CUR and AMS-AUA routes, a significant proportion of customers would substitute these two destinations by other (long-haul) holiday destinations or would not have travelled at all. It should be noted, though, that the proportion of customers that would be likely lost in response to a price increase is lower than in the case of Punta Cana and Cancun (in fact, the results of the customer survey for Curacao and Aruba are comparable to those for Havana).

333. More specifically, while most tour operators and travel agents who responded to the market investigation were sceptical whether a significant proportion of customers would switch away from Aruba and Curacao to other destinations, the results of the customer survey suggest that a significant number of customers might indeed do so.

334. With respect to Aruba, 15% of KLM and Martinair customers asserted that they would switch to an alternative destination in response to a 5-10% price increase (few of them - 2% for Martinair and 5% for KLM - asserted that they would not travel at all). Moreover, given that the number of responses is relatively higher with respect to the Amsterdam - Havana route, the 95% confidence interval for the total loss of customers for the merging parties is smaller: indeed, the customer survey indicates that between 12% and 26% of Martinair and KLM customers would be lost by the merging parties as a response to a price increase.

335. With respect to Curacao, 10% of KLM and Martinair customers who expressed a definite answer asserted that they would switch to an alternative destination in response to a price increase while 10% of them for Martinair and 6% of them for KLM asserted that they would not travel at all. As in the case of Aruba, the relatively higher number of responses determines a bigger 95% confidence interval compared to the Amsterdam - Havana route: indeed, the customer survey indicates that between 12% and 25% of Martinair and KLM customers would be lost by the merging parties as a response to a price increase.

336. Moreover, the customer survey indicated that 15% of Martinair's and 18% of KLM's respondents also considered ArkeFly when booking their flight to Aruba. For Curacao, the relevant percentages are even higher. More specifically, 17% of Martinair's customers and 27% of KLM's customers also considered ArkeFly when booking their flight to Curacao. Out of all customers who considered ArkeFly as an alternative to KLM or Martinair, around 70% indicated that they would indeed switch to ArkeFly in case the merged entity raised prices by 5-10% post-merger. (144)

337. If taken together with the loss of customers due to destination substitutability, it is possible to identify the proportion of customers who would likely absorb a hypothetical price increase by the merging parties. With respect to Aruba, 68% of respondents who bought an unbundled flight ticket asserted that they did not consider ArkeFly before purchasing their trip and that they would keep going to Aruba in case of an unilateral 5-10% price increase. The proportion of respondents who could then be lost by the parties would amount to approximately 32%. This proportion gets even higher in the case of Curacao, where 56% of respondents who bought an unbundled flight ticket asserted that they did not consider ArkeFly before purchasing their trip and that they would keep going to Aruba in case of an unilateral 5-10% price increase. The proportion of respondents who could then be lost by the parties due to the price increase would then amount to approximately 44%.

338. As in the case of Havana, additional information on demand elasticity and profitability of a price increase were provided by the parties with the study submitted to the Commission. (145) The study again provides an estimate of the 'critical loss' for the routes to Aruba and Curacao. According to the results presented in the study, the critical loss on these routes for the merging parties would be [5-10]*% in 2007 as for the Amsterdam - Havana route.

339. These results are however subject to the same critiques reported at recitals 295-299, and a limited weight should thus be attached to them. However, also in the case of Aruba and Curacao, it should be pointed out that the average loss of respondents due to a price increase as indicated by the customer survey is about 3-4 times the critical loss indicated by the parties' study.

340. In addition, as set out in recital 179, the market investigation confirmed that airlines are generally very sensitive to only a small decline in the number of passengers because this can make the difference between profitable and unprofitable services, in particular insofar as passenger air transport operations to long-haul leisure destinations are concerned.

341. Finally, as set out in recital 259, when selling flight tickets to end customers, the parties are unable to distinguish between passengers that are destination-sensitive (in particular VFR travellers) and passengers who are not; consequently the parties would not be able to charge higher prices to those customers who would continue travelling to Aruba and Curacao even in the case of a price increase as compared to those who would switch away. Therefore, as long as a significant proportion of customers could switch to other airlines or alternative destinations or could stop travelling at all in the case of a price increase, the group of travellers who are inelastic to possible alternative destinations is protected against price increases by the threat that other, more elastic, passengers would switch away in response to such a price increase, making that pricing policy unprofitable to the merging parties.

342. In light of these circumstances, despite the fact that the price elasticity of demand is lower in case of Curacao and Aruba than in case of more typical "sun and beach" holiday destinations, it appears more likely than not that any significant non transitory price increase on the part of the merged entity would be unprofitable.

343. Moreover, as set out in section 6.2, the fact that KLM jointly controls Martinair pre-transaction is likely to reduce the anti-competitive effects of the proposed concentration. Finally, as discussed in section 8.4.1, the competitive pressure that Martinair currently exercises on KLM in passenger air transport in general is decreasing due to its financial situation and the need for the timely renewal of its fleet, which is ultimately controlled by KLM. Therefore the effects of the proposed concentration with respect to the parties' passenger air transport activities will in any event be limited.

344. For all these reasons, it must be concluded that it is unlikely that the proposed concentration would result in non-coordinated effects on the AMS-CUR and AMS-AUA routes.

8.4.3.5.6. Coordinated effects

345. It is also unlikely that the proposed concentration could result in coordinated effects in the sense that KLM would be in a position to increase prices for flights on the AMS-CUR and AMS-AUA routes and ArkeFly would follow such price increases.

346. Firstly, the risk that the proposed concentration could lead to coordinated effects appears already minimal in view of the asymmetry between KLM and ArkeFly in terms of market shares. Indeed, the merged entity's combined market share of [80-90]*% on each of the two routes concerned will be more than five times higher than that of ArkeFly (which will have the remaining [10-20]*%). Considering that ArkeFly has stated that it can accommodate substantial extra demand in case the merged entity raises prices post-transaction within a short time-frame of only a few months, any collusive equilibrium would be highly unstable. That is because, in view of its relatively low market share, ArkeFly has a high incentive to undercut KLM to capture part of its demand and expand its market share.

347. Secondly, given the pre-transaction market structure, the likelihood of tacit collusion actually increases only marginally as a result of the merger. Prior to the implementation of the proposed concentration, there are three players serving Aruba and Curacao with flights from Amsterdam. Two of these three players are related by way of a 50% shareholding of the one in the other and therefore, pre-merger, already have an incentive not to compete fiercely (particularly KLM who is entitled to 50% of Martinair's profits). As a consequence, the full integration of Martinair into the Air France-KLM group and its disappearance from the market is unlikely to increase the likelihood of collusion to any significant extent. The impact of the proposed concentration therefore appears to be limited also in this respect.

348. Finally, another reason why possible post-transaction collusion between KLM and ArkeFly is unlikely to occur is the relative instability of the demand in the case at hand. Martinair's long-haul passenger business focuses on leisure travellers and generally, demand of leisure travellers is rather unstable. Indeed, this type of customers is much more likely to cause shocks on the demand than, for example, business travellers, as the former do not strictly have to travel (at least not on that particular route). As a result, it might be difficult for KLM and ArkeFly to identify, on the one hand, change in prices which are due to deviation (and which would need to be punished to sustain a collusive equilibrium) and, on the other, changes driven by demand.

349. In any event, it must be recalled that, as explained in more detail in recitals 184-189, there are clear differences in the parties' business models and market positioning. Whereas Martinair is, to a large extent, a tour operator airline, KLM sells a much lower proportion of its total capacity on flights to holiday destinations to tour operators. Moreover, the competitive pressure that Martinair currently exercises on KLM in passenger air transport in general is decreasing due to its financial situation and the need for the timely renewal of its fleet, which is ultimately controlled by KLM. These differences with respect to business models and market positioning as well as the specific situation of Martinair pre-transaction limit any possible anti-competitive coordinated effects of the proposed concentration already at the outset.

350. In view of all these factors, it appears unlikely that the proposed transaction increases the likelihood that, post-transaction, the new entity and the remaining player ArkeFly would coordinate prices on the AMS-CUR and AMS-AUA routes.

8.4.3.5.7. Conclusion on assessment of the effects of the proposed transaction on the AMS-AUA and AMS-CUR routes

351. Consequently, it must be concluded that, also with respect to the supply of passenger air transport services to end customers on the routes AMS-CUR and AMS-AUA, the proposed concentration would not significantly impede effective competition in the common market or a substantial part of it.

8.4.3.6. Conclusion on the assessment of the transaction with respect to the supply of air transport services to end customers

352. As regards the supply of air transport services to end customers, it must therefore be concluded that the proposed concentration would not significantly impede effective competition in the common market or a substantial part of it with respect to the hypothetical wider markets for air transport of Dutch leisure travellers departing from Amsterdam, or from Amsterdam, Brussels and Düsseldorf to destinations in the Greater Caribbean or to all long-haul holiday destinations and the markets for the supply of air transport services to end customers on the routes connecting either Amsterdam or Amsterdam, Düsseldorf and Brussels to Vancouver, Toronto and Miami, on the routes connecting Amsterdam, Düsseldorf and Brussels to Punta Cana, Cancun and Havana and on the routes connecting Amsterdam to Curacao and Aruba.

8.4.4. The wholesale supply of airline seats to tour operators

353. With respect to all routes affected by the present transaction, both parties are also active on the market for the wholesale supply of seats to tour operators. As the results of the market investigation with respect to the scope of the relevant market at the point of destination are inconclusive, the effects of the proposed concentration will be assessed on the basis of both alternative market definitions, that is to say (i) on the basis of a market for the wholesale supply of seats on all long-haul flights to tour operators serving the Dutch market and (ii) on the narrowest possible alternative markets, that is, on a route-by-route basis where the point of origin with respect to the routes to long-haul "sun and beach" holiday destinations comprises airports in Amsterdam, Brussels and Düsseldorf.

8.4.4.1. The supply of airline seats to tour operators for flights from Amsterdam, Brussels and Düsseldorf to all long-haul destinations

354. The parties were not able to provide estimates on the total market size, their own and their competitors' position on the market for the supply of seats to tour operators for flights out of Amsterdam, Brussels and Düsseldorf to all long-haul destinations. However, based on the 2007 Schiphol survey, the parties provided data with respect to the overall market for the wholesale supply of seats on long-haul flights out of Amsterdam to tour operators addressing the Dutch market, where their combined market share would amount, to approximately [50-60]*% (KLM: [30-40]*%, Martinair: [10-20]*%). (146)

355. As KLM and Martinair are the only two independent airlines with a hub at Schiphol airport, the remainder of the market is likely to be highly dispersed. Although it is probable that ArkeFly accounts for a significant share of this market, most of its sales are made to integrated tour operators belonging to the TUI group. ArkeFly currently sells only a limited amount of seats to third party tour operators as its main focus is to supply seats to tour operators belonging to the TUI group.

356. It should be noted that in calculating these market shares, the parties have excluded figures relating to Martinair's full charter services as Martinair has decided to no longer operate full charter flights as of the winter season 2008/2009. If these services were to be included in the market share calculations, the parties' combined market share on this market would presumably be much higher, possibly amounting to up to [60-70]*%. (147)

357. However, according to the parties, these market shares are inflated since they do not take into account flights included in holiday packages sold by Dutch TOs departing from airports in Belgium and Germany. Indeed, the results of the market investigation confirmed that Dutch TOs sell, to various extents, packages including flights with departure out of Brussels and Düsseldorf. Should these flights be included in the calculation of the market shares on this wider market, the parties' market share would necessarily be lower.

358. The parties argue that the transaction does not raise competition concerns with respect to this market. In support of this view, they argue that KLM and Martinair are not close competitors. As explained in more detail in recitals 184-189, there are clear differences in the parties' business models and market positioning as a result of which there is a significant asymmetry in the way the parties constrain each other pre-transaction. This limits any possible anti-competitive effects of the proposed concentration already at the outset. Whereas, according to the parties, Martinair is indeed a tour operator airline, KLM only sells about [10-20]*% of its total capacity on flights to holiday destinations to tour operators and it declared its intention to decrease its reliance on this type of sales.

359. Although the proposed concentration removes one of the main alternatives for tour operators wishing to purchase seats on long-haul flights departing from Amsterdam, there are several remaining alternative airlines (including leisure carriers) selling seats to tour operators for flights to various long-haul destinations out of Brussels and Düsseldorf. Indeed, LTU, a German airline based in Düsseldorf which operates both scheduled and charter flights, sells seats to Dutch TOs on long-haul flights to various holiday destinations (including Caribbean area) out of Düsseldorf. Similarly, Emirates sell seats to Dutch tour operators to destinations in Asia, the Pacific region and Africa. Several other airlines, for instance Virgin Atlantic and Jet Airways, sell seats to Dutch TOs on long-haul flights operated out of Brussels. In addition, should the opportunity arise, ArkeFly may decide to increase its capacity and its sales to third party tour operators. In light of this, Dutch tour operators selling packages to long-haul destinations will continue having alternative suppliers.

360. Several tour operators expressed concerns during the first phase market investigation that, given the removal of competitive pressure currently exerted by Martinair, the merged entity would have the ability to increase prices or to decrease capacity of seats sold to tour operators, thereby ultimately harming those end customers which value the additional services provided by tour operators as these customers may have to pay higher prices for their package holidays or have a limited choice of package holiday destinations available post-transaction.

361. However, Martinair's operations on all routes affected by the present transactions are highly dependent on sales to tour operators. While KLM on average sells approximately [10-20]*% of its capacity to tour operators, approximately [50-60]*% of Martinair's capacity is currently sold to tour operators (148). This percentage gets as high as [90-100]*% for some long-haul "sun and beach" destinations (149). Therefore, given the crucial importance of a high load factor for a leisure airline such as Martinair, as discussed in recital 179, and the fact that Martinair's load factor is highly dependent on the tour operator business, Martinair cannot afford losing many customers, especially not the ones purchasing large volumes of seats. (150)

362. In this respect, it seems that tour operators have some degree of bargaining power, at least with respect to the guaranteed seats sold by Martinair, which is dependent on tour operator business now more than ever. In support of this argument, parties submitted two very recent e-mail exchanges between Martinair and two of its tour operator customers in which these customers make specific requests with regard to the contents of their contracts with Martinair, for example, by explicitly asking for seats on flights to certain destinations noting that these destinations are absolutely necessary for the respective tour operator to purchase seats from Martinair. (151)

363. While KLM's own dependence on tour operators is much lower, it has indicated its intention that, post-transaction, it intends to maintain Martinair's market positioning as a leisure airline selling seats to tour operators. This intention is confirmed by internal documents of KLM. Accordingly the merged entity will, to a considerable extent, still rely on the tour operator business, in particular with respect to long-haul leisure destinations for which a substantial part of customers purchase a package holiday rather than an unbundled flight.

364. Moreover, as set out in section 6.2, the fact that KLM jointly controls Martinair pre-transaction is likely to reduce the anti-competitive effects of the proposed concentration. Finally, as discussed in section 8.4.1, the competitive pressure that Martinair currently exercises on KLM in passenger air transport in general is decreasing due to its financial situation and the need for the timely renewal of its fleet, which is ultimately controlled by KLM. Therefore the effects of the proposed concentration with respect to the parties' passenger air transport activities will in any event be limited.

365. In light of the situation described in recitals 354-364, it seems unlikely that the merged entity will have the ability to increase prices on the market for the wholesale supply of seats to tour operators with respect to long-haul destinations in general. As a result, it can be concluded that the proposed concentration would not significantly impede effective competition in the common market or a substantial part of it with respect to the supply of airline seats to tour operators for flights from Amsterdam, Brussels and Düsseldorf to all long-haul destinations.

8.4.4.2. The supply of airline seats to tour operators for flights from Amsterdam, Brussels and Düsseldorf to North American cities (Vancouver, Toronto and Miami)

366. The three North-American cities - Vancouver, Toronto and Miami - cannot in fact be considered as typical package holiday destinations and end customers therefore have a stronger tendency to purchase independent holidays, namely, various components separately. This is evidenced by the overall lower proportion of holiday packages sold with respect to these destinations as compared to other long-haul "sun and beach" destinations affected by the present transaction.

367. With respect to Miami, KLM sold [20-30]*% of its capacity to tour operators in 2007/2008 (representing only [...]*% of the total value of sales to this destination during this period). Similarly, in 2007 Martinair sold only [20-30]*% of its capacity on this route to tour operators. Concerning Toronto, which is a destination having a relatively high portion of business traffic, KLM sold only [5-10]*% of its capacity to tour operators in 2007/2008 (representing only [...]*% of the total value of sales to this destination during this period). Similarly, in 2007 Martinair sold only [10-20]*% of its capacity on this route to tour operators. As regards Vancouver, KLM sold [20-30]*% of its capacity to tour operators in 2007/2008 (representing only [...]*% of the total value of sales to this destination during this period). Similarly, in 2007 Martinair sold [30-40]*% of its capacity on this route to tour operators.

368. However, it should be noted that these sales also include sales to tour operators which will sell unbundled seats acting as travel agents. Therefore, with respect to both carriers, the actual wholesale supply of seats to be included in packages would be even lower.

369. The parties were not able to provide their own and their competitors' market shares with respect to this market. However, it can be inferred from the situation on the market for the sale of seats to end customers that, although the parties have a substantial combined market shares ([70-80]*% and [50-60]*% in case of AMS-YVR and AMS/DUS/BRU-YVR respectively, [70-80]*% and [50-60]*% in case of AMS-YYZ and AMS/DUS/BRU-YYZ respectively, and [70-80]*% and [50-60]*% in case of AMS-MIA and AMS/DUS/BRU-MIA respectively), the remainder of the market is rather dispersed (152) with several other airlines operating direct or indirect services out of Amsterdam to all of these destinations. More specifically, with respect to the Canadian destinations, namely Toronto and Vancouver, Air Transat flies directly and several other carriers, such as Air Canada, British Airways or United Airlines perform indirect services. As far as Miami is concerned, KLM performs only indirect services and many other carriers, such as British Airways, United Airlines etc. also offer indirect service. In addition, should the airports in Brussels and Düsseldorf be added in the relevant market, there would be even more alternatives.

370. In addition, as set out in more detail in recitals 184-189, there are clear differences in the parties' business models and market positioning as a result of which there is a significant asymmetry in the way the parties constrain each other pre-transaction. This limits any possible anti-competitive effects of the proposed concentration already at the outset.

371. The market investigation did not reveal any competition concerns with respect to the wholesale supply of seats to tour operators on any of these routes. In light of the fact that the parties will continue to face competition from other airlines post-transaction and tour operators will therefore have alternative suppliers, while end customers can also decide to purchase unbundled flights from alternative airlines, it seems unlikely that the merged entity will have the ability to increase prices on the market for the wholesale supply of seats to tour operators with respect to Miami, Toronto and Vancouver.

372. Moreover, as set out in section 6.2, the fact that KLM jointly controls Martinair pre-transaction is likely to reduce the anti-competitive effects of the proposed concentration. Finally, as discussed in section 8.4.1, the competitive pressure that Martinair currently exercises on KLM in passenger air transport in general is decreasing due to its financial situation and the need for the timely renewal of its fleet, which is ultimately controlled by KLM. Therefore the effects of the proposed concentration with respect to the parties' passenger air transport activities will in any event be limited.

373. As a result, it can be concluded that the proposed concentration would not significantly impede effective competition in the common market or a substantial part of it.

8.4.4.3. The supply of airline seats to tour operators for flights from Amsterdam, Brussels and Düsseldorf to holiday destinations in the Caribbean region and Mexico (Punta Cana, Cancun and Havana)

374. As regards the wholesale supply of seats to tour operators for flights to Punta Cana, Cancun and Havana, the distinction must be made between typical "sun and beach" destinations, namely Punta Cana and Cancun on the one hand, and Havana on the other. While the first two are mainly package holiday destinations, Havana's visitor base seems to be more equally split between independent and package holidays.

375. Indeed, the results of customer survey indicate that with respect to Punta Cana and Cancun, respectively 95% and 91% of the total sample of end customers surveyed purchased a package holiday. The situation is different for Havana where only 49% of end customers purchased a package holiday while the remaining 51% of customers purchased an unbundled flight.

376. The parties were not able to provide their own and their competitors market shares with respect to the market for the wholesale supply of seats to tour operators to these destinations. In 2007/2008, KLM sold respectively [10-20]*%, [10-20]*% and [10-20]* % of its capacity on the flights to Punta Cana, Cancun and Havana to tour operators. Martinair sales to tour operators to these destinations accounted for respectively [80-90]*%, [90-100]*% and [30-40]*% of its 2007 capacity to Punta Cana, Cancun and Havana.

377. Martinair is the only airline operating a direct service to these destinations out of Schiphol, while several other airlines operate indirect services. In addition, in light of the results of the market investigation with respect to the scope of the relevant market, the Brussels and Düsseldorf airports should be included in the market definition. LTU, a German carrier performing both scheduled and charter flights and selling seats to Dutch tour operators operates a direct service to Punta Cana and Cancun out of Düsseldorf. With respect to Havana, Virgin Atlantic offers an indirect flight out of Brussels while Iberia is also active with an indirect service. Iberia also flies to Cancun out of Brussels.

378. Moreover, as explained in more detail in recitals 184-189, there are clear differences in the parties' business models and market positioning as a result of which there is a significant asymmetry in the way the parties constrain each other pre-transaction. This limits any possible anti-competitive effects of the proposed concentration already at the outset.

379. The market investigation did not reveal any substantiated competition concerns with respect to the wholesale supply of seats to tour operators for any of these destinations. In light of the fact that, post-transaction, the parties will continue to face competition from other airlines operating indirect services out of Amsterdam, Brussels and Düsseldorf and tour operators will therefore have alternative suppliers, it seems unlikely that the merged entity would have the ability to increase prices on the market for the wholesale supply of seats to tour operators with respect to Punta Cana, Cancun and Havana.

380. Moreover, as set out in section 6.2, the fact that KLM jointly controls Martinair pre-transaction is likely to reduce the anti-competitive effects of the proposed concentration. Finally, as discussed in section 8.4.1, the competitive pressure that Martinair currently exercises on KLM in passenger air transport in general is decreasing due to its financial situation and the need for the immediate renewal of its fleet, which is ultimately controlled by KLM. Therefore the effects of the proposed concentration with respect to the parties' passenger air transport activities will, in any event, be limited.

381. As a result, it can be concluded that the proposed concentration would not significantly impede effective competition in the common market or a substantial part of it.

8.4.4.4. The supply of airline seats to tour operators for flights to Curacao and Aruba

382. As regards the wholesale supply of seats to tour operators for flights to Curacao and Aruba, it must again be recalled that there are no flights to Aruba or Curacao from either Düsseldorf or Brussels and therefore the proposed transaction needs to be primarily assessed on the basis of the traditional O&D approach, with respect to the wholesale supply of airline seats to tour operators for flights to these two destinations (153).

8.4.4.4.1. Market positions on the Amsterdam-Curacao and Amsterdam-Aruba

Curacao routes

383. The table below shows the market shares of the parties and their competitors with respect to the wholesale supply of seats to tour operators for flights on the Amsterdam-Curacao route on a yearly basis for the year 2007/2008.

<table>

384. As noted in recital 308, KLM operates 7 weekly flights to Curacao, whereas Martinair serves this destination 5 times a week during the winter season and 4 times a week during the summer season. As the parties are the only air carriers active on this route which offer seats to tour operators on a large scale, their combined market share on this route is high, [reaching [90-95]*% in 2007/2008. The only competitor on this route, ArkeFly, operates 6 weekly flights to Curacao, but uses most of its capacity captively for inclusion in package holidays offered by TUI or other tour operators belonging to the same corporate group as TUI and ArkeFly. Therefore, when excluding these captive sales, ArkeFly reaches a market share of only [5-10]*% on the "merchant market" for the wholesale supply of seats to tour operators for flights to Curacao. If captive sales were included, ArkeFly's market share would be more than [4-11]* times as high, reaching [50-60]*%.

Aruba

385. The situation is similar on the AMS-AUA route. The table below shows the market shares of the parties and their competitors with respect to the wholesale supply of seats to tour operators for flights on the Amsterdam-Aruba route on a yearly basis for the year 2007/2008.

<table>

386. The parties' combined market share on this route will amount to [90-95]*%. As noted in recital 312, KLM operates 4 direct flights per week on this route, whereas Martinair serves this destination with direct flights twice a week during the summer season and 3 times a week during the winter season. ArkeFly, which operates 3 direct flights per week between Amsterdam and Aruba, again uses most of its capacity captively for inclusion in package holidays offered by TUI or other tour operators belonging to the TUI group and therefore has a market share of only [5-10]*% on the "merchant market" for the wholesale supply of seats to tour operators for flights to Aruba. Similar to the situation with respect to Curacao, ArkeFly's market share would be more than [3-9]* times as high, reaching [40-50]*%, if captive sales were included.

387. Given that ArkeFly's sales to other tour operators than those belonging to the TUI group are limited, with respect to these two routes the proposed concentration arguably reduces the number of competitors in the market for the wholesale supply of seats to (non-integrated) tour operators from two to one.

8.4.4.4.2. Asymmetry/closeness of competition

388. In spite of this, as already indicated in recitals 184-189, there are clear differences between KLM and Martinair in terms of their market positioning and business models. These differences between KLM and Martinair in terms of business model also translate into the wholesale market for the sale of seats to tour operators in that Martinair is much more a "tour operator airline" than KLM.

389. This difference also shows, albeit to a lesser extent, with respect to the AMS-CUR and AMS-AUA routes. During the IATA year 2007/2008, KLM sold [10-20]*% of its capacity on the AMS-CUR route and [20-30]*% of its capacity on the AMS-AUA route to tour operators, whereas Martinair's sales to tour operators amounted to roughly twice as much of its total capacity, namely [30-40]*% on the AMS-AUA route and [40-50]*% on the AMS-CUR route.

390. To some extent, the difference in business models between KLM and Martinair is also reflected in the way in which the parties contract with tour operators. In principle, KLM sells seats to tour operators in the same way as it sells seats to end-users (the only difference being that tour operators get special rates which are lower than end customer fares). Whereas Martinair operates in much the same way with respect to smaller tour operations, it sells the largest part of its capacity as "guaranteed seats" and makes package deals with some large tour operators, selling capacity on various routes in one package. (154) At least for those larger tour operators, the parties cannot be considered as close substitutes.

391. By way of contrast, from the perspective of the end customer buying package holidays from tour operators, Martinair seems to be a closer competitor to ArkeFly. Most customers purchase the package holiday either from a tour operator including a flight with Martinair or from TUI including a flight with ArkeFly. Therefore, pre-merger, with regard to the market for wholesale supply of seats to tour operators for inclusion in package holidays, Martinair seems to be more constrained by ArkeFly than by KLM.

8.4.4.4.3. Competitive constraints exercised by TUI/ArkeFly

392. Despite the fact that TUI/ArkeFly currently uses most its capacity captively on the AMS-CUR and AMS-AUA routes, it exercises a significant competitive constraint on the parties and will continue to do so post-transaction. More specifically, as will be explained in more details in the following recitals, the competitive constraint exercised by TUI/ArkeFly will be twofold: (i) ArkeFly constrains the parties as a competing supplier in the wholesale market for the supply of seats to tour operators; (ii) moreover, TUI exercises an indirect constraint on the parties in the downstream market for the supply of package holidays to end customers.

8.4.4.4.3.1. Constraint exercised by ArkeFly as a competing supplier of airline seats to tour operators

393. Although ArkeFly currently sells seats to third party tour operators only to a limited extent, the threat that it may change its policy and increase its sales to other tour operators already constitutes some, albeit limited, constraint on the parties' post-transaction market conduct towards tour operators. In fact, ArkeFly has confirmed during the market investigation that it would seriously consider starting to operate charter or scheduled services to a holiday destination out of Schiphol if a group of tour operators approached ArkeFly and gave support to ArkeFly's operations, for example, by purchasing seats on a guaranteed basis.

394. Taking into account that, in addition, ArkeFly has stated that it can increase its capacity within a relatively short time frame (see recitals 286-287 and 328-329), the threat that ArkeFly could actually increase its sales to third party tour operators on the AMS-CUR and AMS-AUA routes also appears reasonably credible.

8.4.4.4.3.2. Constraint exercised by TUI on the downstream package holiday market

395. More importantly, however, the parties also face significant indirect constraints from TUI on the downstream market for the sale of package holidays to end customers.

396. The second phase market investigation and the customer survey have shown that Dutch leisure travellers are generally very price-sensitive. This is arguably particularly true for package holiday travellers who often choose package holidays instead of independent holidays because the former tend to be not only "care-free", but also cheaper than the latter. Consequently, if the parties were to raise prices to tour operators post-transaction, a significant number of customers who would normally purchase a package from one of TUI's competitors would likely switch to TUI's packages. In that case, however, these passengers would be lost also for the merged entity which, as a result, would likely refrain from increasing the price.

397. As in the case of sales of seats to end customers, when considering a price increase, the merged entity would therefore have to take into account the loss of package holiday travellers to other destinations, the loss of customers that would decide not to travel at all and the loss of tour operators switching directly to TUI/ArkeFly. (155) This could significantly affect the load factor on the parties' flights to Aruba and Curacao and thus affect the profitability of their operations. In fact, KLM sells up to [20-30]*% of its seats on flights to these two destinations to tour operators, Martinair even up to [40-50]*%. The parties thus depend on sales to tour operators to a significant extent given that the minimum load factor for profitable operations is around 80%, even 90% in the case of Martinair (see recital 179).

8.4.4.4.4. Bargaining power of large tour operators in case of guaranteed seats on Martinair's charter flights

398. In addition, at least with respect to Martinair, large tour operators will also have at least some, albeit limited, bargaining power towards the merged entity, namely in any case insofar as they purchase guaranteed seats on Martinair's flights (see recital 153).

8.4.4.4.5. Concerns expressed by tour operators

399. During the Commission's first phase market investigation, several market participants indicated that they expected that prices might increase as a result of the proposed concentration, at least on some routes where the parties' activities overlap, notably the AMS-CUR and AMS-AUA routes. However, such concerns remained largely unsubstantiated.

400. Only two market participants expressed more substantiated concerns with respect to the effects of the proposed concentration, including particularly the effects on the AMS-CUR and AMS-AUA routes.

401. One of these complainants suggested that the proposed concentration would strengthen the dominant position of KLM and Martinair on the routes where both parties operate (direct) flights. More specifically, the transaction would make KLM's knowledge, marketing (including KLM's Frequent Flyer Programme) and funds available to Martinair, enhance the parties' possibilities of code sharing, interlining and other types of coordination and unite even more early morning and late night slots at Schiphol airport within the hands of only one player. According to this complainant, this will raise barriers to entry and expansion for the parties' only remaining competitor on these routes, ArkeFly. These concerns appear largely unrelated to the effects that the proposed concentration may have on the routes where both parties operate direct flights. As regards slots, the market investigation has confirmed that early morning and late night slots are not necessary for the operation of long-haul flights out of Amsterdam and that, outside of peak hours, slots are still widely available at Schiphol airport. Advantages in terms of marketing, funding and coordination of schedules or other market conduct are potentially efficiency-enhancing effects of the proposed transaction. Moreover, ArkeFly has explicitly confirmed its ability to increase capacity on the AMS-CUR and AMS-AUA routes if the parties were to increase prices on this route.

402. Another concern expressed by this complainant relates to the fear that the merged entity may engage in predatory pricing on the AMS-CUR and AMS-AUA routes thus driving ArkeFly out of the market. This suggests that, at least in the short term, prices for consumers would decrease. For this concern to be valid in the context of this merger control procedure, there would need to be clear indications that such practices are likely to occur as a result of the proposed transaction (despite the possible deterrent effect of Article 82 EC and the equivalent provisions in the national competition laws) and that they would also likely force ArkeFly to leave the market in the long term (instead of only causing ArkeFly to use its aircrafts on other routes for a limited period of time and to re-enter the AMS-CUR and AMS-AUA routes as soon as operations on those routes become profitable again because the merged entity raises prices). However, the market investigation has produced no significant evidence to this effect.

403. The other complainant expressed concerns essentially relating to past business practices of KLM such as the fact that KLM discriminates between tour operators by not offering equal conditions to all of them. The complainant was concerned that KLM may engage in similar practices post-transaction. These concerns, however, appear to be unrelated to the proposed concentration. Moreover, even if such concerns were merger-specific, there are no indications on the basis of the Commission's market investigation that the proposed concentration would cause such practices to become more likely in the future.

8.4.4.4.6. Conclusion on the assessment of the proposed concentration with respect to the supply of airline seats to tour operators for flights to Curacao and Aruba

404. With respect to the AMS-CUR and AMS-AUA routes, the merged entity will continue to be constrained by TUI/ArkeFly, both as a competing supplier in the market for the wholesale supply of seats to tour operators and as a tour operator competing with the merged entity's customers downstream. Considering the high price-sensitivity of Dutch leisure customers, it appears more likely than not that any price increases on the part of the merged entity towards tour operators competing with TUI would be unprofitable.

405. Moreover, as set out in section 6.2, the fact that KLM jointly controls Martinair pre-transaction is likely to reduce the anti-competitive effects of the proposed concentration. Finally, as discussed in section 8.4.1, the competitive pressure that Martinair currently exercises on KLM in passenger air transport in general is decreasing due to its financial situation and the need for the timely renewal of its fleet, which is ultimately controlled by KLM. Therefore the effects of the proposed concentration with respect to the parties' passenger air transport activities will, in any event, be limited.

406. It must therefore be concluded that, also with respect to the wholesale supply of seats to tour operators on the AMS-CUR and AMS-AUA routes, the proposed concentration would not significantly impede effective competition in the common market or a substantial part of it.

8.4.4.5. Conclusion on the assessment of the transaction with respect to the wholesale supply of airline seats to tour operators

407. As regards the wholesale supply of seats to tour operators, it must therefore be concluded that the proposed concentration would not significantly impede effective competition in the common market or a substantial part of it.

9 ASSESSMENT OF EFFICIENCIES

408. According to recital 29 of the preamble to the EC Merger Regulation156 and the Horizontal Merger Guidelines (157), it is possible that efficiencies brought about by a merger counteract the effects on competition and in particular the potential harm to consumers that it might otherwise have. Parties to a concentration may thus detail the efficiency gains generated by the concentration that are likely to enhance the ability and the incentive of the merged entity to act pro-competitively for the benefit of consumers. Typical examples of such efficiencies include cost savings, new product introduction and service or product improvement. Efficiency claims need to be verifiable (namely reasoned, quantified and supported by internal studies and documents if necessary). To counteract the anticompetitive effects of a merger, such efficiencies must be likely to benefit consumers and could not have been achieved to a similar extent by means that are less anti-competitive than the proposed concentration (the so called "merger specificity"). The three conditions - verifiability, merger specificity and consumer benefit - are cumulative.

409. In the present case, the parties have submitted a study which provides indications on the type and the size of supply side efficiencies which would arise if the parties merge. (158) According to the study, supply side efficiencies would consist mainly of joint purchasing benefits, overhead reductions, rationalisation of frontline operations, systems integration and network efficiencies. At least some of these cost savings appear relevant and the study provides an order of magnitude of total annual cost savings (EUR 16.4 million) However, for the purpose of this merger investigation, amongst other issues, it would be necessary to examine to what extent such efficiencies would be passed on to consumers. As the study does not distinguish between efficiencies affecting parties' fixed cost and efficiencies affecting variable cost, it is difficult to infer from it the expected reduction of the parties' operating cost (and a possible pass-on to consumers).

410. The study also identifies a number of demand-side efficiencies which would arise as a consequence of the merger. These would consist of the elimination of the double marginalization effect for feeder passengers, an increase in flight bundles available to passengers, a reduced risk of delay, frequent flyer benefits and improved check-in facilities. Even if these efficiencies are likely to exist to some extent, however, they are very difficult to be quantified and the study's attempt to do so relies on rather strong assumptions. In particular, it is assumed that KLM and Martinair would not change their flight schedule post-merger. This is, however, rather unlikely as rationalization of networks and schedules is a natural effect of any transaction. Moreover, in another part, the study itself abandons this assumption, namely in the part where supply side network efficiencies are claimed to arise as a consequence of rescheduling of the merging parties' flights.

Conclusion on the assessment of efficiencies

411. To conclude, the study is helpful in pointing out a number of efficiencies which may arise as a result of the merger. However, the underlying assumptions are too strong to allow any inference on their ultimate quantification, and no clear-cut indication can be drawn from it. It follows that the efficiencies presented by the parties cannot be taken into account in assessing the effects of the proposed concentration on competition. However, regardless of this conclusion, the proposed concentration would not significantly impede effective competition in the common market or a substantial part of it.

10 CONCLUSION

412. For the reasons set out above, the proposed concentration would not significantly impede effective competition in the common market or in a substantial part of it. It should therefore be declared compatible with the common market pursuant to Article 8(1) of the EC Merger Regulation and with the EEA Agreement pursuant to Article 57 thereof.

Has adopted this decision:

Article 1

The notified concentration whereby KLM Royal Dutch Airlines NV (the Netherlands) acquires sole control within the meaning of Article 3(1)(b) of Regulation (EC) No 139-2004 of the undertaking Martinair Holland NV (the Netherlands) is declared compatible with the common market and the EEA Agreement.

Article 2

This Decision is addressed to

KLM Royal Dutch Airlines NV

Amsterdamseweg 55 Amstelveen

Netherlands

Notes:

1 OJ L 24, 29.1.2004, p. 1. 6

2 Until very recently, Martinair also operated passenger flights from Amsterdam to Paramaribo. However, Martinair has decided to stop operations on this route from 1 September 2008; cf. http://www.parool.nl/parool/nl/30/ECONOMIE/article/detail/15996/2008/06/03/Vlucht-op-Paramaribo-weer-duur.dhtml, http://www.reisrevue.nl/managementnieuws/id/393/martinair_stopt_met_paramaribo.html, and http://news.caribseek.com/Curaçao/article_66106.shtml.

3 Air France-KLM: EUR 24 110 million for the financial year 2007/2008 (KLM's financial year runs from 1 April until 31 March); and Martinair: EUR 951 million in 2007 (Martinair's financial year corresponds to the calendar year). Both Air France-KLM and Martinair have an EU-wide turnover exceeding EUR 250 million, irrespective of whether the "place of departure" or the "50/50"-method is applied. Martinair applied the "50/50"-method to geographically allocate its turnover. However, its customers for both cargo and passenger services are predominantly based in the Netherlands. Therefore, also on the basis of the "place of departure"-method Martinair has an EU-wide turnover exceeding EUR 250 million. Air France-KLM applied the "place of departure"-method in order to geographically allocate its turnover. With a worldwide turnover exceeding EUR [...]* billion and the majority of its flights departing from or returning to the Netherlands or France, it is clear that also on the basis of the "50/50"-method the EU-wide turnover of Air France-KLM exceeds EUR 250 million.

4 KLM EUR 12 918 million for the fiscal year 2007/08 and Martinair EUR 467 million in 2007. 8

the serious doubts identified by the Commission in the course of its first phase investigation.

5 In this respect, the parties refer to Cases COMP/M.3733 Dow/DDE, COMP/M.3385 - Cargill BCA and COMP/M.2908 - Deutsche Post/DHL II.

6 In addition to the cases quoted by the parties (see footnote 5), see Cases COMP/M.3155 - Deutsche Post/Securior, COMP/M.3448 - Electricidade de Portugal/Hidroelectrica del Cantabrico, COMP/M.3379 - P&O/Royal Nedlloyd/P&O Nedlloyd, and COMP/M.2679 - EDF/TXU Europe/24 Seven.

7 For example, one of the reasons that led the Commission to enter a phase II investigation in case COMP/M.794 - Coca-Cola/Amalgamated Beverages GB was the fact that the acquiring party and the target had increasingly become direct competitors. See paragraph 203 of the decision.

8 Case IV/M.1328 KLM/Martinair. In this case, no decision was published as the transaction was eventually abandoned and the notification was withdrawn on 25 May 1999.

9 Form CO, p. 14; Commission decision of February 2004 in Case COMP/M.3280 - Air France/KLM, at paragraph 137.

10 Both the members of the Executive Board and the members of the Supervisory Board are appointed by the shareholders. KLM and Maersk each have the right to nominate two members of the Supervisory Board. The fifth member of the Supervisory Board is nominated by the other Supervisory Board members (Form CO, paragraph 26; and Article 24 of Martinair's Articles of Association attached to the Form CO as Annex 4).

11 This was confirmed by the parties' lawyers by e-mail of 13 August 2008.

12 Martinair provided copies of the minutes of the meetings of all main decision-making bodies of Martinair, namely Martinair's Shareholders Meeting, Supervisory Board, Executive Committee, Management Team and Top Team, for the period from January 2004 until May 2008. KLM is represented only in the Shareholders and Supervisory Board Meetings. In the minutes of these meetings there is no indication that KLM actively influences Martinair's market conduct to any significant extent. Decisions on Martinair's market conduct rather appear to be taken by internal decision-making bodies, in particular Martinair's Management Team and Top Team, but also internal Working Groups, where KLM is not represented. In particular, the minutes of Martinair's Management Team Meetings and the minutes of Martinair's Working Group for Paramaribo and Curacao show how Martinair carefully monitors KLM's prices and anticipates KLM's reactions to Martinair's pricing policy.

For example, in the minutes of Martinair's Management Team Meeting of 23 May 2006 which report on the discussion of Martinair's business case for its entry on the route Amsterdam-Curacao, it says, "We must be prepared for a reaction from KLM - will they match, will they lower their price significantly?" (translated from the Dutch original which reads as follows: "Wij moeten rekening houden met een reactie van de KLM - gaan zij matchen, gaan zij fors met de prijs naar beneden? ").

Similarly, the minutes of the meeting of Martinair's Working Group Paramaribo (PBM) / Curacao (CUR) of 25 August 2006 which report on possible promotions with respect to passenger flights on the route Amsterdam-Curacao, contain the following phrases: "Further reduction of the price to get more volume is being discussed, but we are of the opinion that we cannot get greater volume this way and that, as a result of this, the match-chance by KL will become too big..." (translated from the Dutch original which reads as follows: "Er wordt gesproken over het verder verlagen van de prijs om meer volume te pakken maar we zijn van mening dat we daarmee niet een groter volume zullen pakken en dat daarmee de match-kans door KL te groot word... ").

13 SkyTeam alliance has 11 full members: Aeroflot, Aeromexico, AirFrance, Alitalia, China Southern, Continental Airlines, CSA Czech Airlines, Delta airlines, KLM, Korean Airlines and Northwest Airlines. Associate members of SkyTeam are: AirEuropa, Copa Airlines and Kenya Airways. It should be noted that SkyTeam alliance provides for a framework within which various members have concluded bilateral and multilateral cooperation agreements whereby they cooperate more or less extensively. On 19 June 2008, Continental Airlines has publicly announced its plans to leave SkyTeam and join the Star Alliance in a closely coordinated marketing and alliance relationship with United Airlines and Lufthansa.

14 SkyTeam Cargo Alliance comprises Aeromexico, AirFrance, Alitalia, China Southern, Delta Airlines, KLM, Korean Airlines and Northwest Airlines.

15 KLM is responsible for the sale of all cargo on Kenya Airways out of Amsterdam. Those sales are reported as KLM sales.

16 In line with the market definition with respect to air cargo transport.

17 See Case COMP/M.3280 - Air France/KLM and Case COMP/M.3770 - Lufthansa/Swiss.

18 See Case COMP/M.3280 - Air France/KLM and Case COMP/M.5181 - Delta Airlines/Northwest Airlines.

19 Questionnaire to customers cargo, question 10; questionnaire to competitors cargo, question 10.

20 See Case COMP/M.3280 - Air France/KLM; COMP/M.3770 - Lufthansa/Swiss and COMP/M.5181 Delta Airlines/Northwest Airlines.

21 See Case COMP/M.3770 - Lufthansa/Swiss, para 19.

22 See Case COMP/M.3280 - Air France/KLM; Case COMP/M.3770 - Lufthansa/Swiss and Case COMP/M.5181 Delta Airlines/Northwest Airlines.

23 See previous footnote for reference to relevant Commission decisions.

24 Questionnaire to customers cargo, question 17.

25 Questionnaire to customers cargo, question 17.

26 See Case COMP/M.3280 - Air France/KLM; COMP/M.3770 - Lufthansa/Swiss and COMP/M.5181 Delta Airlines/Northwest Airlines.

27 Questionnaire to competitors cargo, question 42. Questionnaire to customers cargo, question 33.

28 Questionnaire to competitors cargo, question 39/40.

29 Questionnaire to customers cargo, question 30.

30 "KLM's strategic interest lies primarily with the cargo business of Martinair. The cargo business has an established position in the market and is sustainable in the long term". Form CO, at paragraph 68.

31 The full freighters operated by KLM (3) and Air France (10) are dedicated to specific routes, mostly from Europe to Asia and back.

32 Air France ([...]* Kilotons carried), KLM ([...]* KT), Martinair Cargo ([...]* KT) will together represent [...]* KT carried (ranking second after FedEx which carried [...]* KT of cargo; ahead of the third player United Airlines, which carried [...]* KT) - Source IATA 2007, 2007 WATS, cited in the Form CO at paragraph 351.

33 All market share tables in section 7.4 are based on the parties' estimates for those air cargo carriers for which IATA data was available. The Commission's market investigation has revealed that, on many routes, several (other) air cargo carriers operate air cargo transport services whose market shares are not listed or for which no market shares are entered in the market share tables in section 7.4.

34 * Parts of this text have been edited to ensure that confidential information is not disclosed; those parts are enclosed in square brackets and marked with an asterisk.

35 In particular questionnaire to competitors cargo, question 22. This reference is valid for the entire recital.

36 The parties, however, note that FemkeFly, an airline based in Panama, recently suspended its operations due to high fuel costs, Form CO, p. 146, at paragraph 480.

37 See in particular questionnaire to competitors cargo, question 38.

38 But see earlier footnote indicating that FemkeFly, recently suspended its operations, Form CO, at paragraph 480.

39 Questionnaire to customers cargo, question 17.

40 Questionnaire to competitors cargo, question 22.

41 See Form CO, at paragraphs 428 and 429.

42 In particular questionnaire to competitors cargo, question 22.

43 As indicated in footnote 2, until very recently Martinair also operated passenger flights from Amsterdam to Paramaribo.

44 For instance, question 15 of the questionnaire asks which other destination would have been chosen by the respondent if the price of the trip had been increased by 5-10%. To that question, only the subsample of customers who replied that they would switch to another destination after a price increase (question 14) has answered. This subsample amounts to only 137 respondents. It becomes thus difficult to further disaggregate the sample by routes or carriers in order to be able to get statistically significant results.

45 Throughout this decision, when reporting the results of the customer survey, reference is made to the sample of customers who expressed a definite answer to the specific question asked in the questionnaire, unless otherwise specified. That means that customers who expressed an indefinite answer (such as "don't know") or that did not answer at all are excluded from the sample of respondents for the observed question. This approach facilitates the interpretation of the results and does not alter the conclusions which can be inferred from them, provided that the difference between the underlying distribution of the subsample of respondents who expressed a definite answer and the distribution of the subsample of those who did not is not statistically significant.

46 Form CO, paragraph 52.

47 COMP/M.4439 - Ryanair/Aer Lingus

48 In any case, unless they would sell these seats at a loss, TOs cannot set a lower price for these seats than the one for which they purchased them from KLM or Martinair. As a consequence, these seats are unlikely to exert significant competitive pressure on seats sold directly by the parties.

49 Only a limited number of TOs (currently only one) purchase guaranteed seats from Martinair. With respect to guaranteed seats the price is paid in advance and TOs support the risk of non-selling. Therefore, TOs are contractually allowed to sell these seats without a land arrangement at the price they determine. One TO indicated that these seats are often sold at a loss. For regulatory restrictions applying to the sale of "seats only" on non-scheduled flights, please refer to footnote 50 and recitals 249-250.

50 In 2007 ArkeFly sold a limited number of unbundled seats on its non-scheduled flights on the routes Amsterdam-Punta Cana and Amsterdam-Cancun. According to Dutch regulatory rules, more specifically the so-called "Decree Inclusive Tours", non-scheduled airlines are not allowed to sell unbundled seats. However, in practice airlines often sell the seat with a one day accommodation voucher which de facto amounts to the sale of a seat only (see paragraphs 249-250 below for further details).

51 Parties reply to the 6(1)(c) decision, paragraph 7.

52 On the routes connecting Amsterdam to Curacao and Aruba, ArkeFly operates only scheduled services.

53 See case M.1524 - Airtours/First Choice, paragraphs 34-50 (insofar not affected by the CFI's annulment decision); M.157 - Air France/Sabena, paragraph 25; see also cases M.1494 - Sair Group/AOM, paragraph 11 and M.3770 - Lufthansa/Swiss, paragraph 20.

54 See case M.4439 - RyanAir/Aer Lingus, paragraph 299.

55 VFR travellers represent a substantial portion of the customer mix with respect to the destinations in the Netherlands Antilles, namely Aruba and Curacao.

56 Commission decisions in Cases COMP/M.3280 - Lufthansa/Swiss, COMP/M.4439 - Ryanair/Aer Lingus and COMP/M.3280 - Air France/KLM. See also CFI, case T-177/04 easyjet v Commission, of 4 July 2006 ECR (2006), II-1913, at paragraph 56.

57 Commission decision in Case COMP/M.3280 - Air France/KLM.

58 Commission decision in Case COMP/M.3280 - Air France/KLM.

59 If customers who asserted that they did not know the answer are excluded from the sample and the proportions are calculated only on the basis of customers who expressed a definite answer, the proportion of customers who would not switch to any of the Brussels or Düsseldorf airports is 42%, while the proportion of those who would switch to any of the two airports is 58%.

60 The Schiphol-survey is a survey held on a continuous basis by the marketing research department of Schiphol. Passengers departing from Schiphol airport are asked about various aspects of their flight, such as the purpose of their flight and their destination. The Schiphol survey covers approximately 100 000 passengers.

61 These customers are, however, not affected by the present transaction as Martinair does not offer business class. Therefore this part of the traffic can be assumed to be travelling with KLM both pre- and post-transaction.

62 The parties provided similar figures for Miami, South Africa and Australia. Given the substantially different customer mix and different geographic area, the Commission could not consider these figures as a benchmark for its assessment.

63 As explained in footnote 45 above, these proportions refer to the sample of customers who expressed a definite answer. That is: respondents who replied that they did not know their answer are excluded from the sample. In the specific case of the question related to substitutability at the point of destination, the proportion of "don't know" replies is particularly relevant: they amount to approximately 20% of the sample. However, the analysis of the distribution of the two subsamples (customers who expressed a definite answer and customers who did not) and the Chi-square tests reveal that such distributions do not differ significantly. It is thus possible to focus on the subsample of customers who expressed a definite answer. For the sake of completeness, though, the proportions of customers over the whole sample of respondents who indicated that they would switch to another destination are: 13% for Aruba, 10% for Curacao, 10% for Havana, 27% for Punta Cana, 20% for Cancun.

64 The figure is simply calculated by taking the ratio between the expected loss of customers to other destinations (13%) and the suggested price increase (7.5% - the average of the suggested 5-10% price increase). In this case the elasticity is computed only with reference to respondents who expressed a definite answer (see footnote above) and it does not include customers who would be lost by the merging parties for other reasons (for instance because they would completely refrain from travelling).

65 Provided that the elasticity of respondents does not significantly differ from that of the whole population of travellers from Amsterdam to Havana.

66 Commission decisions in Cases COMP/M.2041 - United/US Airways, COMP/M.2672 - SAS/Spanair, COMP/M.3280 - Air France/KLM and M.5181 - Delta/Northwest.

67 See case COMP/M. 4601 Karstadt Quelle/My Travel.

68 Commission decision in Case COMP/M.3280 - Air France/KLM.

69 Parties reply to the 6(1)(c) decision, paragraphs 14.

70 Commission decisions in Case COMP/M.4601 - Karstadt Quelle/My Travel, paragraph 44 and Case COMP/M.4439 Ryanair/Aer Lingus.

71 According to the parties, as of summer 2008, Martinair no longer allocates guaranteed seats to tour operators.

72 See the submission from the parties' economic advisors dated 26 October 2008 entitled "Counterfactual and efficiencies arising in the long haul passenger business", page 8.

73 See the minutes of the meetings of Martinair's Management Team Meeting of 7 January 2004 - 28 August 2007, Annex 5a to Form CO; and the minutes of the meetings of Martinair's Top Team of 18 September 2007 - 6 May 2008, Annex 5c to Form CO.

74 See Annex 1 to Form CO: VS Consulting Report.

75 See Slide 14 of the VS Consulting Report quoted in the previous footnote.

76 In this context, Martinair terminated its passenger operations to Orlando, Edmonton, Calgary, Jamaica and Tobago at the end of the 2007/2008 winter season.

77 See the submission from the parties' economic advisors dated 26 October 2008 entitled "Counterfactual and efficiencies arising in the long haul passenger business", pages 9-10.

78 See the submission from the parties' economic advisors dated 26 October 2008 entitled "Counterfactual and efficiencies arising in the long haul passenger business", pages 13-14.

79 Martinair provided a translation of an amendment of a loan agreement of [...]* from 2005, which includes the following provision: [...]* See the submission from the parties' economic advisors dated 26 October 2008 entitled "Counterfactual and efficiencies arising in the long haul passenger business", pages 43-44.

80 Martinair has provided details of five loans which it has taken out in recent years at rates ranging between Libor [...]* pts and Libor [...]* pts and between Euribor [...]* pts and Euribor [...]* pts (extended by a year to [...]* pts). KLM, on the other hand, achieved average lending rates between 2006 and 2008 of Euribor [...]* pts for secured financing and Euribor [...]* pts for unsecured financing.

81 A flight cycle includes one take-off and one landing.

82 See Annex 1 to Form CO: VS Consulting report, slide 67.

83 See Article 17 of Martinair's Articles of Association, Annex 4 to Form CO.

84 In this respect, it must be taken into account again that KLM has a veto right with regard to these investments.

85 Martinair is co-hosted in the KLM revenue accounting system Impala. Martinair also uses KLM's inventory system Corda and departure control system Codeco.

86 See paragraphs 14 and following of this decision.

87 In addition, for the purpose of the assessment of the effects of the proposed concentration with respect to the parties' passenger air transport activities, it can also be acknowledged that the degree of competitive constraint exercised by Martinair is limited to the current type and overall scale of Martinair operations given that KLM's consent is required to expand or invest in new Martinair fleet and/or to change Martinair's strategic business (such as a decision to re-enter short-haul passenger business).

88 According to the customer survey, 72% of the sample of respondents flying with Martinair indicated "leisure/holiday" as their trip purpose, while only 38% of respondents flying with KLM did so.

89 Two leisure carriers indicated that a load factor as high as 90% is necessary to make flights to long-haul leisure destinations sustainable. Another market participant indicated that, with a mixed business model (business + leisure traffic), a load factor of at least 75% is needed to achieve profitable operations to long- haul destinations. It should, however, be noted that these statements are especially valid with respect to a particular flight. When considering profitability of a service to a particular destination during a season, carriers always have a choice to decrease capacity (by putting smaller aircraft, withdrawing a frequency for example) to improve their load factor.

90 During the market investigation one respondent provided an example whereby a defection of just one passenger paying 500 EUR would eliminate the benefits of a fare increase of 5 EUR to 100 passengers. Therefore, given the low marginal cost of an extra passenger and the fact that airline seats are perishable goods, airline business is defined by the chase for the marginal passenger.

91 Ethnic traffic in this context is to be understood as traffic caused by travellers which have roots or family ties on the Netherlands Antilles and which travel between the Netherlands and the Netherlands Antilles for this reason. This concerns, for example, Dutch residents with roots on the Netherlands Antilles which travel to the Netherlands Antilles to visit family members or to explore their country of origin.

92 According to an analysis of the purpose of the travel submitted by the parties, the proportion of business traffic on the routes from Amsterdam to Aruba, Curacao, Cancun, Havana and Punta Cana varies between 0% and 8%, while holiday remains the main purpose of the travel to these destinations.

93 See LECG, Critical loss analysis and evidence of passenger behaviour, A report for KLM and Martinair prepared for the purpose of the present transaction.

94 For instance with respect to travellers from Australia to the USA, price elasticity is 1.83; from Australia to overseas locations price elasticity is 1.70.

95 Amongst the 6 papers quoted by the LECG study, only one is an official paper published in an economic journal [Taplin 1997 - Journal of Transport Economics and Policy]. The other 5 sources are either unpublished reports or working papers.

96 According to a survey that Martinair carried out in the period June-October of 2007 to investigate the reasons for booking a flight with Martinair, more than 90% of all passengers mentioned "holidays" or "visiting friends" as the purpose of their journey. Only 3% to 4% stated "business" as the purpose of their journey. Martinair has an economy class and a comfort class. Passengers in the comfort class pay EUR 49 extra for 10% extra leg room and a superior coffee service. There is not a significant difference between passengers travelling in the economy and the comfort class as regards their motives for travelling. (Of the passengers in the economy class, 68% stated "holidays" as the purpose of their travel, and 23% "visiting friends and family". For passengers in the comfort class this was 73% and 20% respectively).

97 It is important to stress that only passengers travelling with unbundled flight were asked this question.

98 OJ L 134, 25.05.2007, p.1.

99 Judgments of the Court of Justice in Case C-466/98 and others of 5 November 2002, ECR [2002] I-9427.

100 According to the Düsseldorf airport authority, the airport is congested during peak hours which are on Monday to Friday from 6.00-12.00am and from 16.00-22.00pm; on Saturday from 6.00-7.00am and 9.00-14.00pmand on Sunday from 16.00-22.00pm.

101 With respect to the Havana airport, the parties were not able to provide information about the level of its congestion.

102 As already noted in section 8.3 on market definition, there are no direct or indirect flights from Düsseldorf and Brussels to Aruba and Curacao. With respect to these two destinations, the assessment will therefore include only flights from Amsterdam.

103 See the comment made in the previous footnote.

104 This estimate is based on figures in the 2007 Schiphol survey as well as figures of the 2007 NIPO survey "ContinuVakantieOnderzoek" ("the NIPO survey"). According to the Schiphol survey, in 2007 approximately 1.709 million passengers flew out of Schiphol airport to long-haul destinations. According to the NIPO survey approximately 93% of the Dutch leisure passengers to long-haul destinations departed from Schiphol, whereas - on average - approximately 4% of the Dutch leisure passengers departed from the airports of Brussels and Düsseldorf. This means that in total approximately 1.783 million Dutch leisure passengers departed from the airports of Amsterdam, Brussels and Düsseldorf. Of this group around [...]* Dutch leisure passengers were carried by KLM. In 2007 Martinair carried nearly [...]* passengers to long-haul destinations, approximately [...]*% of which were business passengers. Therefore, in 2007 Martinair carried approximately [...]* leisure passengers. It should be noted that the parties' estimates relate to all long-haul destinations and fail to indicate which of these should be regarded as holiday destinations.

105 For the source of these market share estimates see the previous footnote. These estimates include both sales of "seats only" to end customers and sales of seats to tour operators. According to the parties, however, their market shares would not be essentially different if these market shares were limited to sales of "seats only", see paragraph 167 of the Form CO.

106 This estimate is again based on the 2007 Schiphol survey and the 2007 NIPO survey. According to the Schiphol survey approximately 376 000 passengers flew out of Schiphol airport to the Netherlands Antilles and the Greater Caribbean area (including Surinam) in 2007. According to the NIPO survey 93% of the Dutch passengers to this area departed from Schiphol, whereas approximately 4% departed from the airports of Brussels and Düsseldorf. This implies that in total 392 000 Dutch passengers departed from the airports of Amsterdam, Brussels and Düsseldorf to the Netherlands Antilles and the Greater Caribbean area (including Surinam). Of this group, based on the Schiphol survey, approximately [...]* leisure passengers (all departing from Schiphol) were carried by KLM. In 2007 an estimated number of [...]* passengers to this area were carried by Martinair. The latter number includes Martinair 2007 passengers that flew to Aruba ([...]*), Curacao ([...]*), Havana ([...]*), Paramaribo ([...]*), Punta Cana ([...]*), Puerto Plata ([...]*) and Varadero ([...]*). Since on average approximately [...]* % of Martinair's passengers travel for business purposes, it is estimated that in 2007 Martinair carried [...]* leisure passengers.

Moreover, the market shares include passengers to Paramaribo. Should Paramaribo be excluded, the parties' market shares would be slightly lower.

107 For the source of these market share estimates see the previous footnote. These estimates include both sales of "seats only" to end customers and sales of seats to tour operators. According to the parties, the market shares would be respectively [20-30]*% and [30-40]*%, if these market shares were limited to sales of "seats only", see paragraph 173 of the Form CO.

108 In addition to these destinations, British Airways also serves Antigua.

109 That is to say the IATA summer season 2007 and the IATA winter season 2007/2008. The IATA summer season covers the period from 1 April until 31 October. The IATA winter season covers the period from 1 November until 31 March.

110 That is to say the IATA summer season 2006 and the IATA winter season 2006/2007.

111 pax = passengers

112 The data in the tables in this section concerns total passengers, both time-sensitive and non time-sensitive; however the market shares are almost identical if using total passengers or only non time-sensitive passengers, as there are very few time-sensitive passengers on the routes concerned (e.g., based on MIDT data, for the year 2007/2008, 8 437 out of a total of 155 290 passengers on the AMS-CUR route, 3 210 out of a total of 52 047 passengers on the AMS-AUA route; 3 877 out of a total of 67 426 passengers on the AMS-YVR route, 8 362 out of a total of 114 683 passengers on the AMS-YYZ route; 1 631 out of a total of 27 007 passengers on the AMS-MIA route; 190 out of a total of 19 069 passengers on the AMS-CUN route; 417 out of a total of 11 598 passengers on the AMS-HAV route; 50 out of a total of 12 595 passengers on the AMS-PUJ route).

113 As indicated in paragraph 25, in light of the existing cooperation between Air France/KLM, Delta Airlines/Northwest Airlines, Air France/Delta Airlines and KLM/Northwest Airlines, the Commission has on a fail-safe basis, and without prejudice to any future assessment, taken into account market shares of all four players.

114 MIDT data are prepared and issued on a monthly basis. Strictly speaking, these data record bookings, not the number of passengers actually flying on a given city pair, and usually do not include "direct" sales, that is sales made by the airlines themselves, and sales of seats on non-scheduled flights. However these flaws in the data appear to apply to all airlines in more or less the same way (with the exception of (i) low cost carriers which generally tend to sell a higher proportion of seats directly through their website instead of through travel agents and (ii) airlines which are vertically integrated with a tour operator with respect to seats that are included in package holidays of the respective tour operator).

115 One of these two flights makes a short stop on its way to Vancouver.

116 Only United Airlines exceed this duration on the return flight from Vancouver to Amsterdam. The stopover duration of this flight is 170 min. The information on stopover durations is based on the Official Airline Guide (OAG).

117 Communication from the Commission 2006 (871) of 9 January 2007 - Developing a Community Civil Aviation policy towards Canada.

118 As indicated in paragraph 25, in light of the existing cooperation between Air France/KLM, Delta Airlines/Northwest Airlines, Air France/Delta Airlines and KLM/Northwest Airlines, the Commission has on a fail-safe basis, and without prejudice to any future assessment, taken into account market shares of all four players.

119 See the comment in the previous footnote.

120 British Airways, United Airlines and Continental each operate at least one indirect flight between Amsterdam and Miami per day with a stopover duration of less than 150 min. US Airways also offers several indirect flights per week between Amsterdam and Miami with stopover durations of less than 150 min.

121 Other carriers operating flights on this route include ArkeFly and Jetairfly. For the reasons set out in footnotes 114 and 122, their data is not (fully) recorded in MIDT.

122 As regards ArkeFly, the parties explicitly noted that, on those routes where ArkeFly is active, the MIDT data underestimate the position of ArkeFly since the package holidays of Arke (a tour operator belonging to the TUI group) are not recorded in the Computer Reservation Systems ("CRS") (see paragraph 177 of Form CO).

123 Note, however, that also seats sold by Martinair to tour operators are only partly included in MIDT data.

124 Decree of 5 February 1981, Official Journal (Staatsblad) 1981, 33, as lastly amended on 23 March 1998, Official Journal (Staatsblad) 1998, 61.

125 For example, Thomas Cook offers flights with Jetairfly from Brussels to Punta Cana on its website, indicating that the price includes hotel accommodation for the first night.

126 The customer survey also enquired which ones of the suggested destinations would have been picked up by those customers who would be willing to switch destination as a response to a price increase. Unfortunately, responses are too few for inferring any statistical indication at the route level. However, at an aggregate level, the customer survey shows that respondents are rather evenly distributed across the different suggested options: 27% would pick a destination in the Dutch Antilles, 35% another destination in the Caribbean, 28% would pick an intercontinental destination outside the Caribbean.

127 As mentioned in footnote 45, the percentage terms for this and the following routes refer only to customers who expressed a definitive answer.

128 It should be noted that no Air France/KLM customers was surveyed on this route as Air France/KLM serves this destination only with an indirect flight via Paris.

129 As discussed in more detail in section 9, carriers may decrease the capacity in which case the need for a high load factor will be adapted to the actual scale of operations.

130 One of the other carriers operating flights on this route is ArkeFly. For the reasons set out in footnotes 114 and 122 and paragraph 247, ArkeFly's data is not (fully) recorded in MIDT.

131 Delta Airlines operates 6 indirect flights per week to Cancun via Madrid from both Amsterdam and Düsseldorf.

132 Northwest Airlines sell only very few seats on this route. Accordingly, its market share is close to 0% and therefore Northwest Airlines is not listed in table 11.

133 Thomas Cook offers flights from Brussels to Cancun on its website, indicating that the price includes hotel accommodation for the first night.

134 Note that no Air France/KLM customer was surveyed on this route as Air France/KLM serves this destination only with an indirect flight via Paris.

135 As discussed in more detail in section 9, the carriers may decrease the capacity in which case the need for a high load factor will be adapted to the actual scale of operations.

136 Connecting flights to London Gatwick are operated by British Airways.

137 Most of the stopover duration for Air France, Iberia and Virgin for AMS-HAV and HAV-AMS is around 150 min.

138 ArkeFly currently operates a fleet of three Boeing 767-300 aircraft, one Boeing 737-800 aircraft, two seasonal damp leased Boeing 737-900 aircraft and one seasonal wetleased Airbus A320. Two more Boeing 737-800 aircraft are on order and will enter the fleet in December 2008 respectively June 2009. ArkeFly operates leisure flights to 43 destinations in 17 countries, both medium haul (Mediterranean area and North Africa) and long-haul (Aruba, Dutch Antilles, Dominican Republic, Jamaica, Mexico, Cuba, Venezuela and Brazil). All destinations (except for Aruba, Curacao, Bonaire and Saint Maarten) are operated as non-scheduled flights.

139 TUI Travel Plc. is represented in over 20 countries. It deploys a complete travel concept, whereby airlines are fully integrated in the business of its respective daughter tour operators. TUI Travel Plc.'s total passenger fleet currently consists of approximately 160 aircraft.

140 In simple terms, this means that the customer survey indicates that there is 95% chance that the true percentage of passengers that would either switch destination or refrain from travelling in response to a price increase falls within the extremes defined by the confidence interval (The 95% chance is a standard threshold used by statisticians to assess whether a result has statistical value). The results thus indicate that the minimum statistically significant proportion of customers who would be lost in case of price increase is 8% and the maximum is 32%.

141 LECG - Critical loss analysis and evidence of passenger switching behaviour - 12 October 2008.

142 It should be stressed that the question on whether an alternative carrier was considered by the respondent could be asked only to customers who purchased an unbundled flight.

143 On the route AMS-PUJ, ArkeFly sells [significantly less]* seats to end customers, on the route AMS-CUN even [less]* [in both cases less than one fourth of the proportion of seats that ArkeFly sells to end customers on the routes AMS-CUR and AMS-AUA]*

144 Again, it should be stressed that the question on whether an alternative carrier was considered by the respondent could be asked only to customers who purchased an unbundled flight.

145 LECG - Critical loss analysis and evidence of passenger switching behaviour.

146 On the basis of the Schiphol survey, the parties estimate the total volume of this market (based on seats sold to tour operators) to amount to approximately 803 000 passengers in 2007. In that same year KLM sold approximately [...]* seats on long-haul flights to tour operators. This amounts to a market share of approximately [30-40]*%. In 2007, Martinair sold approximately [...]* seats on long-haul flights to tour operators. It should be noted that the parties claim that these market share estimates are inflated because they do not include seats on package tours departing from Brussels and Düsseldorf.

147 The parties indicated that Martinair sold [...]* seats on long-haul full charter flights in 2007. Depending on whether these seats are included or not in the total market of 803 000 passengers, Martinair's actual market share in 2007 would amount to [20-30]*% or even [30-40]*%.

148 Form CO, paragraph 296.

149 In 2006, Martinair sold [90-100]*% of its capacity on Amsterdam/Cancun route to tour operators.

150 Of the total seats sold by Martinair to tour operators, approximately [90-100]*% is accounted for by guaranteed seats, with respect to which tour operators negotiate the terms and conditions and for which they pay in advance and support the risk of not selling.

151 See the attachment to the parties' e-mail dated 27 August 2008 containing an e-mail of 21 April 2008 from Thomas Cook Nederland to Martinair.

152 In the airline industry it is but natural that carriers have large market shares on routes out of their home airports, while usually several other players are present with smaller market shares.

153 Martinair and KLM are the only two airlines currently operating flights on the routes AMS-CUR and AMS-AUA which are not vertically integrated with a tour operator.

154 During the IATA summer season 2008, OAD, Van der Valk and Thomas Cook, for the IATA winter season 2008/2009 only Thomas Cook purchased "guaranteed seats from Martinair.

155 As specified in paragraph 337, according to the customer survey, the proportion of respondents who bought an unbundled flight that could be lost by the merging parties because of these three effects would amount to 32% with respect to Aruba and 44% with respect to Curacao. The loss of package holiday travellers should not be smaller since the demand for package holidays is likely more elastic.

156 See Article 2(1)(b) and Recital 29.

157 Guidelines on the assessment of horizontal mergers under the Council Regulation on the control of concentrations between undertakings, OJ C 31, 5.2.2004, p. 5.

158 LECG - "Counterfactual and efficiencies arising in the passenger business" - 26 October 2008.