Documentation

____" """!:'. ______. ...., cgMP[flt:o:.J tltlUNAl CT- 88// TRISt:tlAl DI LA co:KU!UtEilf't ~ . 0 F ) J l E D RIGISTIAR - llfGISTRAIRf THE COMPETITION TRIBUNAL -- OTTAWA, ON:r.- -i--~ f IN THE MATTER OF an Application by the Director of Investigation and Research under Section 64(1) of the

Competition Act, R.S. c. C-23 as amended.

AND IN THE MATTER OF a Limited Partnership formed to combine the operations of the Reservec and Pegasus computer reservation systems. AND IN THE MATTER OF The Gemini Group Automated Distribution Systems Inc.

B E T W E E N: THE DIRECTOR OF INVESTIGATION AND RESEARCH, Applicant - and -AIR CANADA INC. Place Air Canada 500 Dorchester Blvd. West Montreal, Quebec H2Z lXS - and -AIR CANADA SERVICES INC. c/o Air Canada Place Air Canada 500 Dorchester Blvd. West

Montreal, Quebec H2Z lXS

- and -PWA CORPORATION Suite 2800, 700-2nd Street, s.w. Calgary, Alberta T2P 2W2

- and -CANADIAN AIRLINES INTERNATIONAL LTD. Suite 2800, 700-2nd Street, s.w. Calgary, Alberta T2P 2W2 - and -PACIFIC WESTERN AIRLINES LTD. c/o Canadian Airlines International Ltd. Suite 2800, 700-2nd Street, s.w. Calgary, Alberta T2P 2W2

- and -CANADIAN PACIFIC AIR LINES, LIMITED c/o Canadian Airlines International Ltd. Suite 2800, 700-2nd Street, s.w. Calgary, Alberta T2P 2W2

- and -154793 CANADA LTD. c/o PWA Corporation Suite 2800, 700-2nd Street, s.w. Calgary, Alberta T2P 2W2

- and -153333 CANADA LIMITED PARTNERSHIP c/o The Gemini Group Automated Distribution Systems Inc. 1 First Canadian Place P.O. Box 84 5th Floor

Toronto, Ontario M5X 1K4 - and -THE GEMINI GROUP AUTOMATED DISTRIBUTION SYSTEMS INC.

1 First Canadian Place P.O. Box 84 5th Floor Toronto, Ontario

M5X 1K4

Respondents

REPLY OF THE DIRECTOR OF INVESTIGATION AND RESEARCH

A. INTRODUCTION 1. The Tribunal has received three responses from the Respondents, one from Air Canada (the "AC Response"), one from PWA Corporation, Canadian Airlines International Ltd., and 154793 Canada Ltd. (the "PWA Response") and one from 153333 Canada Limited Partnership, Air Canada Services Inc. and The Gemini Group Automated Distribution Systems Inc. (the "Gemini Response"). This reply will serve as the full reply of the Director to all three responses.

2. In reply to the Respondents, the Director states that this merger prevents or lessens competition or is likely to prevent or lessen competition substantially in the following respects:

(a) The merger eliminates Pegasus as an independent effective CRS competitor, reducing the number of CRS competitors from three to two in most urban markets and from two to one in many non-urban markets in Canada. Pegasus is the only other Canadian competitor and exclusively hosts CAIL and its affiliated carriers, the second largest airline network in Canada. The addition of Pegasus' market share to an already dominant Reservec results in a significant increase in market concentration. Moreover, the impact on competition of the removal of Pegasus is particularly important in light of the prospects for future growth of Pegasus, either by itself or as a joint venture partner with a CRS other than Reservec.

- 2 -(b) This merger significantly increases the barriers to entry for future competitors seeking to enter the CRS market in Canada. Gemini will control the most complete, timely and accurate information on AC, CAIL and their respective affiliates which are the carriers most frequently booked in Canada. As a result, Gemini will become a compulsory CRS for most travel agents even if other systems have superior functionalities.

(c) Gemini will also be able to prevent future competition and to exercise substantial market power in the CRS market in Canada because AC and CAIL and their affiliates dominate the airline passenger market in Canada. This airline dominance, coupled with the vertical integration of Gemini with AC and CAIL, will ensure that Gemini will be the CRS vendor of choice for travel agents in Canada.

(d) The joint ownership by AC and CAIL of Gemini will ensure that AC and CAIL retain control of the distribution of airline passenger seats to travel agents in Canada. This is the direct result of the decision by AC and PWA Corporation and their respective affiliates to eliminate competition between the two principal carrier groups in Canada, namely Air Canada and CAIL, in the provision of CRS services to travel agents in Canada.

(e) Furthermore, AC, CAIL and Gemini are anxious to limit competition in Canada between existing CRS vendors both for the reasons

- 3 ­stated above and in a conscious attempt to ensure that other CRS vendors do not further penetrate the market in Canada. This is the basis for their refusal to provide a direct access link to other CRS vendors. In addition, the refusal is intended to protect Gemini's position in the CRS market as long as it is considered or perceived to be less attractive than its competitors.

3. Through Gemini, AC and CAIL will substantially control the distribution of airline passenger seats to travel agents and consumers and thereby protect their respective positions as the dominant carriers in Canada. AC and CAIL, in the absence of competition in the CRS market, will have the capacity to maintain and expand their control of the distribution of airline passenger seats, which is unregulated, and their control of the domestic airline market which is limited in size and largely protected from encroachment by foreign carriers.

4. In the circumstances, this merger will not only lessen or prevent competition substantially in the CRS market, but it will also have significant anticompetitive effects on airline markets. This is not in the public interest no matter how beneficial it may be in serving the private interests of the Respondents.

B. CRS MARKET (a) Relevant Market 5. In paragraph 41 of the PWA Response, and paragraphs 20, 21 and 28 of the AC Response, the

- 4 -Respondents take issue with the Director's analysis of the relevant geographic market. Despite acknowledging that the relevant geographic market "for most CRS issues" is the Canadian market, the Respondents make the vague assertion that the North American market is the relevant market for some purposes, and compare the size of Gemini to the size of other CRS vendors in North America. This analysis is inappropriate. 6. The use of a North American market definition implies that the u.s. CRS vendors have the same incentive and ability to provide a nationwide network of CRS services in Canada as do the Canadian CRS vendors. This is simply not the case. The primary motivation for the investment in the national CRS networks developed by Pegasus and Reservec was the significant participation of their airline parents in Canadian airline markets. The CRS forms an important means by which CAIL and AC distribute their own airline services to travel agents and ultimately to the traveling public. However, u.s. and international airlines cannot fly point to point in Canada and therefore do not participate in many of the airline markets in Canada. u.s. and international CRS vendors, without the ability to vertically integrate with carriers flying into airline markets in Canada, do not have the same strong incentives to develop a network of CRS locations across Canada competitive with those of Pegasus and Reservec. Similarly, the Canadian CRS vendors have demonstrated a lack of incentive to establish nationwide networks of CRS locations in the U.S. Moreover, other considerations that indicate that U.S. CRS systems would face a much different

- 5 ­competitive environment in Canada than that faced in the u.s. are discussed in paragraph 30 below. These factors and those described above suggest that the Canadian CRS market is distinct from the U.S. CRS market. Thus, the Director rejects the notion that there is a North American CRS market.

7. In order to effectively market their products to the Canadian public, airlines, including international and transborder carriers, must be able to communicate accurate, timely, and complete information on their services, and provide the ability to make reservations and issue tickets. This is accomplished by the listing of all of this critical information on CRS's that have nationwide networks of travel agent locations throughout Canada. Similarly, travel agents in Canada require information and reservations capability in relation to the airlines of most interest to their clients. Thus, there clearly is an important national aspect to the CRS market in Canada.

8. In evaluating the prevention or lessening of competition that results from this merger it is useful and appropriate to consider local geographic markets throughout Canada as well as the national market. Airlines attempt to distribute and market their product to all points where they may gain customers, and are therefore concerned about distribution to specific local or regional markets in their network. Travel is purchased locally because travel agents in one local market are not generally accessible to customers located in other local markets. In addition, the information on airlines of most

- 6 ­interest to travel agents and their customers may vary depending on their geographic location.

(b) Market Share Measurement/Concentration 9. In paragraph 9(a) of the PWA Response and paragraphs 2(b), 32, and 33 of the AC Response, the Respondents claim that the Director has overstated the degree of concentration by measuring market shares by travel agent locations and number of CRT terminals. The Respondents also claim that flight segments booked by Gemini and rival CRS vendors are the most appropriate measure of market share. Nevertheless, it is instructive that both AC and PWA, in analyzing the CRS market in presentations to their respective Boards of Directors prior to the merger, relied on the number of travel agent locations and CRT terminals in Canada to measure the market share of CRS vendors. These measures indicate that both AC and PWA Corporation believed that Gemini would have an overwhelming share of the market after the merger no matter how concentration is measured.

10. Travel agent locations and CRT terminals are appropriate measures of market share because a large proportion of revenues for a CRS come from the fees for terminals and other charges paid for by travel agents. Moreover, travel agent locations are important because they represent access points for the public and the airlines. They measure all geographic points at which travel can be booked and are an important indicator of the geographic penetration of the distribution channel employed by the airlines.

- 7 -11. The segments booked are another component of CRS revenue. Accordingly, the Director provided an estimate in paragraph 28 of the Application based on the information provided by AC and CPAL as to the segments booked for 1986 on these airlines. However, the information does not include segments booked on the other Canadian carriers and it includes segments which were booked in the United States on other CRS systems. For these reasons, updated information directly from Reservec, Pegasus and Sabre on segments booked will be required in order to compile an accurate picture of segments booked by these systems.

C. AIRLINE MARKET 12. The Respondents also complain in paragraphs 2(b) and 2(c) of the AC Response and paragraph 11 of the PWA Response, that the Director has overstated the degree of concentration in the airline market. However, they have not indicated the extent of the actual concentration, even though AC and CAIL have the best information on their respective market shares. AC and CAIL dominate the market for domestic airline travel and Gemini will permit AC and CAIL to maintain or increase their share of this market. While it is true that AC and CAIL do not dominate the transborder and international markets in which they compete, they have a significant presence in these markets. This merger will enable AC and CAIL to use Gemini's market power to increase their market share in transborder and international airline markets.

- 8 -D. NATURAL MONOPOLY/CRITICAL MASS 13. In paragraphs 2(a), 24(a), 24(c) and 46 of the AC Response and paragraphs 21 and 27 of the PWA Response, the Respondents claim that there are substantial economies of scale in CRS technology and that a critical mass in terms of the number of subscriber travel agencies is necessary to enable a CRS to be viable. The Respondents further allege in paragraphs 26(a), 26(b), 26(c), and 71 of the AC Response and in paragraphs 5, 31, and 74 of the PWA Response, that the merger provides the only opportunity for a Canadian CRS to achieve the critical mass necessary for full cost recovery.

14. The Respondents have not provided any information which would indicate what these economies of scale are or what the critical mass is in terms of travel agents or otherwise. With 2900 travel agent subscribers, Reservec was very profitable. Reservec is nearly the same size as Datas II (3000 travel agents) and Pars (3000 travel agents), two of the six U.S. CRS vendors. Datas II and Pars are both profitable.

15. Moreover, when CPAL decided to enter the CRS market in 1984, it estimated that breakeven would occur at 1000 travel agents despite the ongoing presence of Reservec. The Respondents now claim they need 90% of the Canadian CRS market, or about 3600 travel agent locations in order to break even. This claim also assumes that the number of travel agents and CRT's in Canada will remain static and that Reservec, Pegasus or Gemini would not participate in the growth of the

- 9 -market. This is an unlikely scenario in a deregulated environment and is contrary to the projections of the Respondents.

16. The revenues and costs associated with measuring economies of scale in CRS relate to more than travel agent locations. Revenues are also dependent on fees based on the number of terminals, and on booking and hosting fees. CRS revenue can therefore increase with deregulation, with or without travel location growth. Deregulation of airlines in Canada will likely produce an increase in the number of persons travelling, the number of travel agents and the number of bookings. Not only will the volume of business increase, but the complexity of information will increase as the number of fare types and new entrants increases. This increased volume and complexity will require Canadian agents to increase their level of automation by adding additional CRT terminals and related equipment. This in turn will decrease the share of market that must be obtained to become profitable.

17. The economies of scale associated with operating a CRS are not simple and linear because costs do not always decline as the number of travel agents served by a specific CRS increases. Serving more travel agents above a certain minimum can require substantial additional investment in hardware and software, thereby increasing the costs of production. It is not clear, therefore, that the minimum efficient scale for CRS operation requires a firm to hold all or nearly all of the Canadian market.

- 10 -E. PEGASUS AS A FAILING FIRM 18. The Respondents claim in paragraphs 36(a), 44 to 47 of the AC Response, and paragraphs 3(b), 27, 46 to 49 and 73 of the PWA Response, that Pegasus was a failing firm. This is not the case. As with any capital intensive venture, it requires a period of some years before a return on investment can be expected. In the case of the U.S. vendors, several years were required before profits were realized. In the case of Pegasus, CPAL estimated in 1984 that it required approximately 1000 travel agencies and 3 years to achieve breakeven. It had achieved approximately 72% of this goal by June 1987.

19. In fact, Pegasus achieved a 16 to 18% CRS market share based on locations/terminals in three years. This compares to Sabre which, notwithstanding its superior technology, achieved only a 10 to 13% share in four years. Pegasus accomplished this growth in a short period of time despite the fact that AC refused to participate in Pegasus, refused to pay booking fees and refused to provide a direct access link to Pegasus.

20. It is also important to consider the impact of significant changes in the air transport market that have taken place when considering the competitive significance and marketing prospects for Pegasus. Pegasus started operations when its owner, CPAL, had no more than 21% of the revenue passenger kilometers of level one carriers in Canada. As a result of the CPAL acquisition of Eastern Provincial Airways and other carriers, and

- 11 -PWA Corporation's takeover of CPAL in late 1986, the underlying air transport share of the carriers that are hosted only in Pegasus has grown substantially. According to CAIL, it had an estimated 44% of the domestic passengers in January 1988. This air transport share of the combined carriers suggests that Pegasus would have substantially increased its market penetration by adding on more travel agents and CRT's and making more bookings from its subscriber travel agencies.

21. The growth of CAIL would also create more incentive for AC to agree to provide a reciprocal direct access link to Pegasus because of the growing importance of having complete, timely and accurate CAIL information in Reservec. The implementation of such a link would have removed a major barrier to the growth of Pegasus.

22. The Pegasus losses and revenues described in paragraphs 3(b) and 47 of the PWA Response are questionable. First, since the internal CRS of CAIL and the external system used for travel agent distribution share common facilities, hardware and people, any allocation of costs as between these two systems is arbitrary and must be closely examined. Second, the revenues of Pegasus have been understated because they do not include the additional revenues that CAIL earned as a result of increased CAIL ticket sales by having Pegasus terminals in travel agencies. U.S. experience indicates that such incremental revenues can be significant. As noted in paragraph 47 but omitted in paragraph 3(b) of the PWA Response, the revenues must also include an estimate of the

- 12 ­benefit that CPAL and CAIL gained as a result of not paying booking fees that they otherwise would pay if they did not have Pegasus. Such cost avoidance benefits have, of course, increased with the growth of the CAIL network.

23. The agreement between PWA Corporation and AC indicates that Pegasus had substantial value. Pegasus assets transferred to the joint venture were valued at $18.7 million. In addition, PWA Corporation acquired a 50% ownership interest in Gemini but supplied only 27% of the initial capital (Pegasus assets plus $10 million in cash).

24. In any event, if the Tribunal finds that Pegasus is a failing firm, little weight should be given to this factor in this case because CAIL had a clear alternative which was less anti-competitive than the merger with AC, namely a joint venture proposal with Sabre.

25. In addition to the Sabre option, there are a number of other possible joint ventures that would be less anti-competitive than the current proposal. System One for example, has recently concluded a deal with the Amadeus group in Europe while Apollo is involved in the Gallileo consortium. CAIL, together with Pegasus, would be an attractive joint venture partner because of its large Canadian inventory (44% of domestic passengers), sizeable Canadian travel agent penetration (16-18%), and large Canadian sales force.

- 13 -26. The failing firm argument should also be given little weight because the difficulties that Pegasus did have in achieving market growth were the result of AC exercising its market power by refusing to participate in Pegasus, refusing to pay CPAL for AC bookings on Pegasus and refusing to agree to a direct access link between AC and Pegasus. The fact that Pegasus did as well as it did in these circumstances is further proof of its competitive significance. Moreover, it underscores the importance of the underlying air transport share of the carriers that are hosted in a CRS.

F. EFFECTIVE COMPETITION REMAINING 27. The Respondents, in paragraphs 12, 22, 38 to 41 in the AC Response and in paragraphs 43 and 44 of the PWA Response, claim that Sabre is an effective competitor and point, inter alia, to its growth in the market since 1983. Nonetheless, the ability of Sabre to remain a competitor is contingent on Gemini not exercising the market power it has gained from this merger.

28. The Respondents have refused to provide a direct access link to Sabre even though AC and CAIL would increase their distribution of airline seats by making the same information on their flights available in Sabre and Gemini. In reality, AC and CAIL seek to ensure their domination of the distribution network in Canada at the expense of other CRS vendors by means of artificial barriers to entry which are intended to protect Gemini's market share pending full implementation of its

- 14 ­revised system. The Respondents are well aware that the merger, and the merger alone, by eliminating competition between Reservec and Pegasus, affords Gemini an opportunity to consolidate its position in Canada and thereby forestall competition from other CRS vendors.

29. As a result of this merger, in some local markets in Canada Gemini has a monopoly or near monopoly. In larger centers where there is CRS competition from Sabre, there are many travel agents of small size that cannot afford duplicate systems. For these agents, Gemini is compulsory because only Gemini can provide the most complete, timely and accurate information on AC and CAIL. For these travel agents, neither Sabre nor other potential entrants provide a competitive alternative.

30. In paragraphs 36(b), 42, 49, and 56 of the AC Response and paragraphs 23 and 43 of the PWA Response, the Respondents assume that other u.s. vendors can easily enter the Canadian market. This is not true for a number of reasons. First, none of these other vendors is able to provide the most complete, timely and accurate information on AC, CAIL and their respective affiliates. Second, U.S. vendors do not serve Canada at only incremental costs. In some cases vendors may have to increase their facilities and telecommunication networks at considerable expense to add on Canadian locations. Third, because a CRS does not have unlimited capacity, there is an opportunity cost to serving a Canadian travel agent as opposed to a u.s. based travel agent who is closer at