^* A list of acronyms used in this article is provided on page 47. A professor of management at Texas A&M University, Michael W. Pustay is the author of International Business (with Ricky Griffin, 1996), "The Globalizing Airline Industry: The Need for a New International Policy Regime," in M. Kreinin, ed., Contemporary Issues in Commercial Policy (1995); "Six Myths of the Contemporary Airline Industry," Journal of the Transportation Research Forum vol. 32, 2 (March 1992); "Denmark and the Maastrict Treaty: A Market Analysis" (with L. Bierman and J. Kolari) Duke Journal of Comparative and International Law (Fall 1992), among other books and articles. International trade and investment have enjoyed unprecedented growth in the fifty years since the end of World War II. Prospects look bright for continuance of this trend during the remainder of the decade. The ratification of the Uruguay Round of the GATT in 1994 and the creation of the World Trade Organization in 1995 are indicative of the world-wide consensus regarding the importance of international commerce to national and world prosperity. The strengthening of regional trading alliances in Europe, North America, South America, and Asia in the 1990s reflects similar forces. But one area where the advocates of free trade have not been persuasive is in the provision of international airline services. Trade in airline services in most regions of the world is still governed by bilateral agreements that often restrict rather than promote competition. Liberalization efforts that have freed trade in other goods and services have bypassed this industry. The Uruguay Round, the Canadian-U.S. Free Trade Agreement, and the North American Free Trade Agreement, for example, did little to open trade in airline services.(1) Despite the general openness of and the large magnitude of trade between their two economies, these observations held true for air services between the United States and Canada from the end of World War II until early 1995. For most of the latter half of the post-war period, the bilateral air service agreement between Canada and the United States was among the most restrictive of all such pacts entered into by the United States. Leaders from both countries recognized the inappropriateness of that state of affairs and that promotion of trans-border air travel should be encouraged. For example, the Shamrock Summit Declaration, signed in March, 1985, by Prime Minister Mulroney and President Reagan, declared that liberalization of trans-border airline services was an important goal of both countries. Yet for over a decade Canadian and U.S. negotiators intermittently met and failed to come to a new accord. But in February, 1995, officials of the two countries announced the signing of a new air services agreement (ASA) that significantly liberalized trade in airline services between the two countries. This paper explores the evolution of trans-border airlines services between Canada and the United States and examines why liberalization was so difficult to achieve. It begins by reviewing the policy regime that governed the international airline industry: the Chicago Conference, the International Air Transport Association, and the utilization of bilaterally-negotiated ASAs to authorize service on individual routes. It then examines the domestic and international airline policies of the two nations. These policies have had a significant impact on the structure of the airline industries of the two countries, which in turn has affected the nature of existing trans-border competition. The competitive impact of these policies is then used to explain why liberalization of trade in airline services between the two countries was so difficult to achieve. Finally, the initial responses of carriers to the liberalization of trade in airline services are analyzed. I. THE POLICY REGIME OF THE INTERNATIONAL AIRLINE INDUSTRY Historically the policy regime of the international airline industry rested on three pillars.(2) The first was the right of nations to control the air space above their territory and thus aviation services to, from, and over them. This principle was originally established at Europe's 1919 International Convention for Air Navigation. Participants at the 1944 Chicago Conference reiterated and extended it to the other continents. Although representatives of the United States advocated "unrestricted international operating rights with market forces determining frequencies and fares,"(3) most of the other nations at the Chicago Conference, fearing the post-war domination of U.S. flag carriers, successfully blocked the adoption of the U.S. proposals.(4) The Chicago Conference set the stage for the creation of a protectionist, mercantilistic approach to governing trade in international airline services in the post-war period. The second pillar of the industry's institutional structure was the utilization of bilateral ASAs to control commercial airline service between each pair of countries. Typically, such bilateral air service agreements detail the routes that can be served between the two countries and the number of carriers from each country allowed to serve on each of these routes. Although they existed in the pre-World War II period, most post-war agreements were patterned after the 1946 Bermuda Agreement between the United States and the United Kingdom. Bermuda-type agreements facilitated the creation of an anti-competitive international airline industry. Nations jealously guarded the granting of traffic rights and extended most of them only on the basis of reciprocity. Competition in international markets was also restricted through the incorporation of secret codicils into these bilateral agreements. A common device, pooling agreements, divided the revenues earned in a country-pair market equally between two carriers (one from each nation), creating obvious incentives to reduce competitive behavior in that market.(5) The net effect was that entry into most international markets was severely restricted. The resulting network of routes reflected the political power and negotiating skills of the carrier's home government. For the most part, the international routes of a carrier were a set of bridges connecting a city in its home country to a city in a foreign country. Such route networks bore little resemblance to those that would be constructed by a carrier seeking to develop an efficient network over which to carry its passengers. Historically, the third pillar of the industry's institutional structure was the International Air Transport Association (IATA), a private trade association whose membership consists of most of the world's international scheduled airlines. Its members, meeting on a regional basis, attempted to agree on a common set of prices to charge. IATA's ability to establish a series of regional airline cartels and suppress intra-regional price competition was enhanced by support from numerous nations, many of whom as owners of international carriers explicitly empowered IATA to establish prices. Moreover, many bilateral agreements specifically endorsed the tariff-making activities of IATA's regional conferences.(6) A. The Collapse of the Old Policy Regime This policy regime created a pro-industry, anti-consumer environment in which international airlines competed. Surprisingly, the U.S. government, despite its desires to promote competition in international air services, resisted efforts to utilize multilateral solutions to end protectionism. For example, the United States chose not to use the GATT for this purpose. In the eyes of U.S. negotiators, reliance on the GATT would either slow down liberalization or allow protectionist nations to obtain a free ride on any aviation liberalizations agreed to by the United States. The former would occur because of the difficulty of gaining agreement among all GATT members to aggressively liberalize trade in air services. The latter would occur because, should the United States liberalize its air service through the GATT with one member, it would be compelled to liberalize service for all GATT members because of the GATT's use of the mostfavored-nation principle. Not only would this reward protectionist nations, it would also denude the United States of any negotiating leverage it might have to encourage such countries to liberalize their own policies. Another explanation for the reticence of the United States to utilize a multilateral approach was that it had accomplished significant changes in the international policy regime as a result of three actions it had unilaterally undertaken beginning in 1978. First, the U.S. Civil Aeronautics Board initiated a "show cause" order threatening the exemption of IATA's price setting actions from U.S. antitrust laws.(7) Without this exemption, international carriers faced the threat of triple damage lawsuits for violating the price-fixing provisions of U.S. antitrust laws. After much litigation and diplomatic acrimony, IATA's ability to establish prices in international markets involving the United States was weakened. IATA's control over pricing in other regions has subsequently atrophied as well.(8) Second, the United States signed liberal "open skies" bilateral air services agreements with over twenty countries that reduced governmental control over entry, exit, and pricing in these markets. Many of these agreements freed U.S. carriers to fly between any point in the United States and any point in the other country and charge any price they wished. Typically, carriers from the other country were given the right to expand the number of points they served in the United States, as well as obtaining similar pricing freedom. Third, the United States deregulated its domestic airline industry, granting domestic carriers pricing freedom in addition to freedom of entry and exit. These three policy changes were not the outcome of a carefully developed consensus among the world's leading aviation powers; rather, they resulted from unilateral actions by the United States. They had two important consequences. First, they shattered the existing international policy regime. Reduction in IATA's power over international prices meant that intra-regional price competition could (and did) increase. The open skies treaties created pockets of liberalized trade in airline services around the world. They led to new entry into the relevant country-pair markets and increased price and service competition among carriers serving these country-pairs. Diversion of international passengers from non-liberalizing countries within a region to liberalizing countries was substantial, pressuring carriers from non-liberalizing areas to cut their prices to staunch the loss of business.(9) Air Canada, for example, was forced to cut its transatlantic fares in order to deter Europe-bound Canadian tourists from originating their journeys in the United States.(10) The three unilaterally-imposed reforms also boosted the competitiveness and efficiency of the U.S. airline industry in both domestic and international markets. Most U.S. carriers quickly readjusted their domestic route systems from the linear, point-to-point operations they had operated under regulation to complex hub-and-spoke networks. These hub-and-spoke systems allowed them increased control over their on-line feed of both domestic and international passengers. Their development and structure reflected the forces of economic efficiency, not the diktat of regulators trying to balance the needs of politically powerful interest groups. More generally, the intense competitive pressures faced by U.S. carriers induced them to streamline their operations as a matter of survival. Naturally, the competitive lessons they learned in the domestic market were transferred to the international market. Changes in the competitive positions of carriers also caused a re-allocation of U.S international operating rights from weaker carriers like TWA and Pan Am to much stronger competitors like American, Delta, and United airlines. The noticeable improvement in the competitiveness of U.S. flag carriers vis-a-vis foreign flag carriers disequilibrated the competitive balance among the world's airlines, causing reactions throughout the international airline industry. Some countries tried to roll back the tide: France and Germany, for example, responded to the onslaught of U.S. flag carriers by trying to renegotiate their existing ASAs with the United States to make them more restrictive. Other countries tried to duplicate for their carriers the advantages created for U.S. flag carriers by the U.S. policies. Foreign carriers and governments undertook many new initiatives, including privatization of state-owned carriers (e.g., Canada, Mexico, the United Kingdom), deregulation of domestic airline industries (Australia, Canada, Chile, and the European Union), liberalization of non-U.S. bilateral treaties, strategic alliances, mergers, and cross-ownership arrangements. However, these policy changes have not been uniformly implemented, and as a result there are wide differences in the competitiveness of carriers and in the levels of competition in individual international country-pair and city-pair markets. II. DOMESTIC AVIATION POLICIES The domestic aviation policies of the United States and Canada today share several important similarities. Both countries rely on privately-owned carriers competing in deregulated environments to provide air transportation services. Still, the structures of the airline industries in the two countries differ substantially as a result of the disparate sizes of their domestic markets and their divergent past regulatory policies. As we will discuss, differences in past domestic regulatory policies have affected the international airline policies of the two countries, which in turn have shaped the positions that the two nations have taken in negotiating a new trans-border air services agreement. A. U.S. Domestic Aviation Policies The U.S. airline industry grew from a handful of barn-storming entrepreneurs in the early 1920s to a dozen or so corporations in the 1930s partly as a result of federal subsidies to carry airmail on routes between major U.S. cities. But a series of financial scandals, an airplane crash involving a U.S. senator, and aggressive bidding for airmail contracts--on some routes, carriers offered to carry the mail for free--led to calls for reforming the airmail subsidy program. Change came in 1938 with the passage of a Civil Aeronautics Act which initiated regulation of the domestic airline industry. A newly-created Civil Aeronautics Board was given responsibility for controlling entry, exit, and prices in domestic airline markets. Over time, the CAB created three classes: trunk carriers, which provided service between large U.S. cities using jet aircraft; local service carriers, which provided service between large U.S. cities and smaller communities using large turboprop equipment; and commuter carriers, which were freed from CAB regulation so long as they used aircraft of 19 seats or fewer. A feature of these carriers unique to the United States is that they were all privately-owned; indeed, the United States is the only major country that has relied completely on privately-owned carriers to provide airline services in its domestic and international markets. Over the forty-six years of its existence (1938-1984), the CAB tried to find an appropriate balance between promoting too much and too little competition. When the airlines were doing well financially and the economy was growing, the CAB held proceedings to allow new entry into existing city-pair markets or to authorize new service to previously unserved markets. When the industry's financial performance weakened, the CAB would refuse to hold any new route application hearings. As a result of this process, on most major U.S. domestic routes two or three carriers typically provided service. However, most of their competitive energies were focused on service competition, rather than price competition, as a result of the CAB's pricing regulations. (11) Another feature of CAB regulation warrants mentioning: it treated the airline industry as a closed club, allowing no new carriers to enter the trunk segment or the local service segment of the industry. More importantly, the CAB strove to ensure that no carrier in these two segments would fail financially. In choosing between competing applicants seeking to provide service on a newly-authorized route, the CAB was forced to make trade-offs among such factors as strengthening weaker carriers, maximizing single-carrier or single-plane service, promoting competition, and spreading awards among carriers so that all would maintain their political allegiance to the CAB's regulatory system. As a result of its balancing act, at the end of the regulatory era some twenty carriers comprised the trunk and local service segments of the industry. The United States began the statutory deregulation of its domestic airline industry in 1978. Over a period of years the Airline Deregulation Act freed U.S. air carriers to enter and exit domestic markets at will and to charge whatever prices they wished, subject to safety and fitness standards administered by the U.S. Department of Transportation. The experience has generally been positive: new carriers have entered the industry, inflation-adjusted prices have fallen, flight frequencies and passenger traffic have risen, and the intensity of competition has increased. But the process has not been costless. Many carriers have exited the industry as a result of bankruptcies and mergers, the number of major carriers in the industry has shrunk from its pre-deregulation level, and some communities have lost jet service. Although industry concentration measured by the four-firm concentration ratio has risen above its pre-deregulation level, concentration on individual routes has fallen. (12) The number of major players in the U.S. airline industry remains large by Canadian standards: American, United, Delta, Northwest, Continental, TWA, US Air, and Southwest have extensive systems and compete with one another on a nation-wide basis. All except Southwest operate large, complex hubbing operations and most have significant international operations. B. Canadian Domestic Aviation Policies Scheduled airline service on short-haul routes began in Canada in 1919. Transcontinental services were slower to develop. For many years, for example, travelers wishing to fly from Toronto to Vancouver were forced to fly via the United States. To overcome this problem, the Trans-Canadian Airlines Act of 1937 called for the establishment of a federally-owned carrier to provide transcontinental service. (13) Trans-Canada Airline, which became Air Canada in 1964, commenced service between Montreal and Vancouver in 1938. Canadian Pacific Airlines (later CP Air and Canadian Airlines), the carrier that would grow to become Trans-Canada's chief rival, developed from a series of acquisitions and mergers initiated in 1933 by its corporate parent, the Canadian Pacific Railway. After the Second World War a number of regional carriers also arose to serve local needs, including Quebecair, Nordair, Eastern Provincial Airlines (EPA), and Pacific Western Airlines (PWA). During the 1960s these regional airlines were assigned designated territories and given the mission of acting as feeders for the transcontinental services of CP Air and Air Canada. A number of charter carriers also emerged to offer non-scheduled service, particularly between central Canada and points in Florida. (14) Federal regulation of domestic air services commenced in Canada with the passage of the Transport Act of 1938. It empowered a Board of Transport Commissioners to issue new operating authority when it was required by the "public convenience and necessity." Governmental policy focused on creating a national network of routes. To cross-subsidize unprofitable services, regulators favored the establishment of route monopolies in profitable markets. Under this and successive legislation, Air Canada enjoyed a protected and favored position in many key markets. On the main transcontinental routes, for example, Air Canada was granted a monopoly position from 1937 to 1959. After CP Air was allowed to provide service on these routes, it labored under capacity restrictions that limited its flight frequencies to well below those offered by Air Canada until 1979. As a result of these policies, Air Canada enjoyed a dominant position in the domestic market. In 1984 the government of Canada began to relax its regulation of domestic airline service with the publication of New Canadian Air Policy. A year later Freedom to Move further outlined the government's transport policies, followed by the adoption of a National Transportation Act which was implemented at the beginning of 1988. These reforms were driven in part by a re-thinking of appropriate regulatory philosophy and in part by the diversion of traffic to such U.S. airports as Seattle, Buffalo, and Burlington as Canadians voted with their feet, seeking to benefit from cheaper U.S. fares brought on by U.S. deregulation. (15) The regulatory reforms divided the country into two zones, northern and southern. Existing regulations in the less populated northern zone remained intact, as the area's thin population, only 5 percent of Canada's total, was believed by policymakers to be unable to support competition. The southern zone, where it was thought market density was sufficient to support competition, experienced significant regulatory relaxation. Here most controls on airline service were removed, including restrictions on capacity, frequency of service, and equipment type, although controls on exit were retained in some circumstances. Carriers were also granted power to reduce prices as they saw fit, while their ability to raise prices automatically was limited by an inflation index. (16) But the airline industry in Canada, unlike that of the United States, was highly concentrated. Concentration increased in the late 1980s with the establishment of Canadian Airlines International Limited (CAIL), a new carrier created from the consolidation and merger of the operations of CP Air, Wardair, and four regional airlines--EPA, Quebecair, Nordair, and PWA. Most remaining regional carriers act as feeders either to CAIL or Air Canada. (17) Air Canada and CAIL are essentially a duopoly, controlling directly or through their affiliates 98 percent of domestic passenger revenues, 97 percent of domestic passengers, and 93 percent of all domestic passenger-kilometers in 1992. (18) As we will discuss later, the differences in the concentration and structure of the domestic airline industries in the two countries affects trans-border competition. The differing structures also affects the bargaining needs and positions of the two countries' negotiators. U.S. negotiators had to juggle the conflicting needs and interests of the eight major U.S. carriers, as well as accommodate the interests of smaller, new entrants such as ValuJet, Kiwi, and Reno Air. Canadian negotiators had to balance the needs of just two carriers. But these two carriers' interests were by no means homogeneous, thereby complicating the development of a bargaining strategy by Canadian negotiators. III. INTERNATIONAL AVIATION POLICIES Variations in the structure of the two countries' domestic airline industries and the evolution of their domestic regulatory policies and philosophies have resulted in substantive differences in the evolution of their international aviation policies. These differences in turn have affected the ability of their carriers to exploit the new Canada-U.S. air services agreement. A. U.S. International Aviation Policies As noted earlier, U.S. negotiators at the 1944 Chicago Conference favored open competition in international airline markets, no doubt because they expected U.S. flag carriers to dominate the world market. Although U.S. policymakers have retreated from their pro-competitive stance from time to time--most notably in the early 1970s--in general in negotiating new ASAs the United States has sought to implement this philosophy. It has attempted to expand the number of U.S. gateway cities from which service can be offered as well as to increase the number of carriers from each country that may offer service on any city-pair route. Such an approach is not surprising, since relative to other aviation powers the United States has had more potential international gateways and more carriers capable of providing international service. While benefiting U.S. consumers, this policy also makes it easier to satisfy the competing interests of various U.S. communities and carriers seeking new international operating authority. (19) This pro-competitive policy has been incorporated into the international route allocation decisions of U.S. regulators. For example, in the North American Routes Case, one of its first such post-war rulings, the Civil Aeronautics Board (CAB) granted transatlantic authority to three carriers--Pan American, TWA, and American Export Airlines, which later was merged into American Airlines. (20) By 1950, the CAB had created head-to-head competition between Pan Am and TWA at London, Frankfurt, Paris, and Rome. Subsequent route awards allowed additional carriers, such as National, Continental, and Delta, to enter the transatlantic market. The pro-competitive philosophy of the United States was reinforced in 1978 with the Carter administration's announcement of its new "Open Skies" policy and the passage of the International Air Transportation Competition Act of 1979. As a result of these two initiatives, the U.S. Department of Transportation has sought in negotiating new bilateral air service agreements to: 1. reduce entry barriers to international service by increasing the number of international gateways; 2. increase price competition among carriers and reduce government control over carrier pricing decisions; 3. allow multiple designation of carriers; 4. eliminate restrictions on capacity and frequency; 5. increase carrier flexibility in providing intermediate and beyond service; 6. increase flexibility in offering charters; and 7. eliminate unfair competitive practices that discriminate against U.S. carriers. (21) These objectives obviously promote the pro-competitive policies of the United States. They also accommodate the unique needs of the United States that result from its geographic size, the number of its potential gateways, and the number of U.S. flag carriers. No other country has as many major carriers as the United States--hence the emphasis on multiple designation. No other country has as many potential gateways as the United States, given its geographic size and the dispersion of its population among major cities. Thus U.S. negotiators seek international agreements that allow them to accommodate these diverse corporate and geographic interests. B. Canadian International Aviation Policies Initially, Canada's international aviation policies were strongly influenced by Air Canada's status as a Crown corporation. Air Canada enjoyed a monopoly on all-Canadian flag international routes from 1937 to 1948, when CP Air was designated as the nation's flag carrier in the Pacific. CP Air's domain expanded in 1965 when it became the sole Canadian carrier to serve Latin America, Southern Europe, Australia, New Zealand, and Asia. Air Canada was given monopoly rights to the rest of Europe (including the United Kingdom), the Middle East, and the Caribbean. (22) An order by the minister of transportation in 1987 resulted in a realignment of this "division of the world" policy by reallocating international operating authority between CP Air and Air Canada. (23) Rather than receiving exclusive rights to serve regions, the two carriers were given exclusive rights to serve individual countries within these regions. In most cases, both carriers served each of the world's regions. The right to serve Spain and Portugal was shifted from CP Air to Air Canada, for example, while traffic rights to Scandinavia and parts of South America were shifted to CP Air from Air Canada. Wardair's entry into the transatlantic market and its subsequent purchase by CAIL eroded the boundary lines established in the 1987 order. As a result, both CAIL and Air Canada may provide service to London, Frankfurt, and Paris. While Canada has negotiated the right to designate multiple carriers in 39 of the 61 ASAs it has signed, it has exercised this right only in a few cases. (24) A pair of policy statements issued in late 1994 and early 1995 indicated an increased willingness by Canadian transport officials to liberalize Canada's international air transport policy. Under a new "use-it-or-lose-it" proviso, any Canadian carrier is free to apply for international operating authority that has been underutilized by an incumbent Canadian airline. Similarly, once traffic in an existing market reaches a minimum threshold--currently defined as 300,000 one-way origin and destination passengers--a second Canadian carrier will be eligible to apply for and receive authority to serve such a market, as long as such service is allowed by the relevant bilateral air services agreement. (25) To date, however, the 1994-95 policy statements have had little impact on competition in Canadian international airline markets. As a result of these policies, Air Canada possessed a monopoly on all trans-border routes until 1967. In that year CP Air received its first U.S.-Canada route, between San Francisco and Vancouver, as a result of the 1966 air service agreement between the two nations. However, most of the routes authorized by the 1966 ASA and the 1974 amendments to it were allocated to Air Canada. Thus CP Air, and its corporate successor, Canadian Airlines, had little opportunity to expand its trans-border service under the 1966 ASA. Accordingly, when the new ASA was signed between Canada and the United States in February, 1995, Canadian Airlines was in a far weaker position to exploit the agreement than Air Canada. It was but a minor player in the trans-border market, serving only the U.S. west coast cities of San Francisco, Los Angeles, and Honolulu. Air Canada had a far greater physical presence (i.e., gates, check-in terminals, etc.) and brand name recognition in most U.S. cities than Canadian Airlines. Canada's regulatory policy of dividing the world into regional (and later country) spheres of influence suppressed intra-flag competition among Canadian carriers in international country-pair markets. Presumably this policy was adopted to reduce price competition in these markets and to allow Canadian-flag airlines to focus their competitive energies on foreign carriers. But this policy had two negative by-products. First, by limiting the number of international cities that each carrier could serve, it artificially reduced the ability of Air Canada and CAIL to construct efficient route networks to control on-line passenger flows on their domestic and international flights. For example, had Canadian regulators granted Air Canada all of the country's international operating authorities ex-Toronto, and CAIL all international operating authorities ex-Vancouver, both carriers would have been able to capture additional economies of scope at their primary airports, thereby enhancing their efficiency and competitiveness. (26) In contrast, the open skies treaties signed by the United States expanded the number of international cities that each of its carriers could serve and incorporate into their hubbing systems. Second, because they were arbitrarily precluded from competing in certain international markets, each Canadian carrier had an incentive to link up with foreign carriers that possessed such rights. For example, locked out of much of the Asian market by government policy, Air Canada had an incentive to enter into a strategic alliance with United Air Lines in order to access United's Asian flights. The net result is that instead of competing with Air Canada on Canada-Asia routes, CAIL is now competing with the strategic alliance of Air Canada and United. Ironically, by suppressing intra-flag competition, the division of the world policy may have actually increased it and diverted some of its benefits to foreign carriers. IV. AIR SERVICE AGREEMENTS BETWEEN THE UNITED STATES AND CANADA Until February, 1995, the bilateral air services agreement (ASA) between the United States and Canada was among the most restrictive of all ASAs to which the United States was a party. The basic ASA governing trans-border air services was signed in 1966. It has been amended several times, most importantly in 1974 when additional point-to-point routes were added to those agreed to in 1966. As amended, the 1966 ASA delineated 83 separate point-to-point routes between the two countries. Only 19 of these routes--so-called "double track routes"--were open to competition between carriers of the two nations. Thirty-eight were reserved for U.S. carriers and the remaining 26 were limited to Canadian airlines. (27) On most of these routes, each nation could designate more than one airline only with the permission of the other government. While capacity was left to the determination of the designated carriers, either government was allowed to reject proposed fares on trans-border routes. U.S. domestic and international airline policies underwent significant revision during the Carter administration. As part of its efforts to reduce regulatory control over the industry, in 1979 the U.S. government proposed revision of the ASA to expand the number of trans-border city-pairs receiving service. It was particularly interested in expanding service to the newly-developing hubbing complexes established by most U.S. carriers as a result of domestic deregulation. Because they perceived that most of the benefits would accrue to U.S. airlines, neither Canadian policymakers nor Canadian airline executives were particularly enthusiastic about a major expansion of the ASA at that time. (28) Two agreements were signed in 1984 that liberalized trans-border air services to a limited extent. The Regional, Local, and Commuter Services Agreement granted automatic approval of route applications under specific circumstances. Such approval was restricted to routes served by aircraft of 60 or fewer seats between a U.S. city of less than one million population or a Canadian city of less than 500,000 persons. (29) In addition, stage lengths could not exceed 400 miles for routes serving central Canada and 600 miles for other routes. A second agreement in the same year granted automatic approval to trans-border routes between Montreal's Mirabel Airport and points in the United States, excluding seven U.S. gateways (Boston, Chicago O'Hare, Los Angeles, Miami, New York JFK, San Francisco, and Seattle). As part of this agreement, the U.S. government designated the San Jose, California airport to receive similar treatment as Mirabel. Furthermore, these air services operated under a double disapproval pricing regime--that is, carriers were free to establish whatever prices they wished unless both the governments of Canada and the United States disapproved of them. (30) While billed as an experiment to assess the impact of entry and pricing freedom in the trans-border market, the agreement was also motivated in part by the Canadian government's desire to stimulate activity at Mirabel, which had failed to match its supporters' traffic forecasts after its opening in conjunction with the 1976 Montreal Olympics. (31) As part of the Shamrock Summit Declaration in March, 1985, both countries promised to examine the possibility of creating free trade in trans-border aviation services. To facilitate the negotiations, the countries agreed to exchange concept papers outlining their respective visions of a new agreement. The concept paper issued by the United States favored liberalization of trans-border air services along the lines of its "open skies" philosophy: complete freedom of carriers to provide service between any city in Canada and any city in the United States; double disapproval pricing; and liberalized charter regulations. In contrast, Canada's concept paper sought to create a common market in airline services. The key difference in the two proposals involved cabotage. (Cabotage occurs when a carrier from country A carries local traffic between two cities in country B.) U.S. negotiators dismissed the Canadian proposal, fearful that if they granted Canadian carriers cabotage rights in the United States, then Asian and European airlines would soon clamor for the same privilege. (32) Canadian negotiators similarly rejected the U.S. position, believing that the proposal would benefit U.S. carriers to the detriment of Canadian airlines. Discussions continued, but with little success. Twelve negotiating sessions in 1991-92, for example, failed to create an agreement acceptable to both sides. (33) A. Breaking the Deadlock Yet both sides were unhappy with the status quo. The 1966 ASA seemed to displease virtually everyone. Its anti-competitive philosophy ran contrary to the aviation policies of both countries. Service on many routes was restricted to a single carrier, reducing actual and potential competition below that found in either country's domestic airline market. Moreover, either country was allowed to disapprove a fare, further strengthening the anti-competitive biases of the agreement. Almost two-thirds of the 100 largest cities in the United States lacked nonstop service to Canada. Many major city-pair combinations, such as Montreal-Atlanta or Toronto-Washington, did not receive service because of the limited number of routes detailed in the ASA. (34) Both sides agreed that the 1966 ASA was suppressing trans-border travel and economic activities between the two nations. For example, between 1980 and 1993, trans-border air travel grew only 1.8 percent annually, well below growth rates experienced in country-pairs involving their other leading trading partners. (35) Community groups on each side of the border--particularly representatives of local airports, the United States Airports for Better International Air Service (USA-BIAS) coalition and the Association of Canadian Airport Communities (ACAC)--complained bitterly that the existing ASA was hindering economic development of their areas. (36) The arrangement also worked to the particular disadvantage of the Canadian carriers. U.S. carriers had direct access to some 90 percent of Canada's population, while Canadian carriers had equivalent access to only 30 percent of the U.S. population. (37) Because of the limited services authorized by the 1966 ASA, much of this traffic was forced to funnel through a handful of U.S. cities. Hubbing systems operated by U.S. carriers at these sites allowed them to control the flow of passengers from non-gateway U.S. cities on to their flights to and from Canadian cities. (38) The competitive value of these hubs was enhanced by the availability of pre-clearance facilities at several major Canadian airports. They allow U.S.-bound travelers to clear U.S. customs in Canada, making it easier for them to utilize the connecting services offered at U.S. hubs. (39) One Canadian study indicated, for example, that only 11 percent of passengers flying a Canadian carrier had a connection in the United States, compared to 55 percent of passengers on trans-border flights of U.S. carriers. (40) In the eyes of Canadian airline executives and policymakers, Canadian flag carriers were not doing very well under the old ASA. By Canadian estimates, U.S. citizens accounted for only 40 percent of the trans-border market, yet U.S. carriers routinely carried over two-thirds of scheduled trans-border traffic, as Table 1 reports. (41) Table 2, which presents information about trade in international passenger transportation (IPT) services between the two countries, depicts a similar situation. Over an extended period of time the United States has exported more IPT services to Canada than it has imported from Canada. Despite their unhappiness with the 1966 ASA, Canadian negotiators were very apprehensive about the ability of Canadian carriers to survive in an open fight with the U.S. carriers in the trans-border market. Such concerns were of great significance to them: a 1991 policy statement issued by the Ministerial Task Force on International Air Policy concluded that protecting the interests of Canadian airlines should be given higher priority than community development and consumer welfare. (42) Thus the carriers' support for or against a new ASA was a major consideration for Canadian negotiators. While both stood to gain opportunities to enter new trans-border markets, the interests of Canada's two primary flag carriers in the creation of a new, liberal ASA were asymmetric. Air Canada had a strong position in the trans-border market under the old ASA, while CAIL had a very weak one. Should a liberal ASA be signed, CAIL had little existing market share to lose; the reverse was true for Air Canada. CAIL was thus generally supportive of a new agreement, although it preferred an incremental liberalization of the old ASA rather than an open skies approach. (43) Because it had much to lose as well as much to gain under a new ASA, Air Canada's support for a new ASA in the 1991-92 round of negotiations was "tepid," (44) and its lack of enthusiasm no doubt contributed to the unsuccessful conclusion of these talks. Table 1 U.S.-Canada Scheduled Transborder Passengers (in thousands) U.S. Market Total U.S. Flag Other Flag Share 1994 10,655 7,194 3.461 68% 1993 10,810 7,545 3,265 70% 1992 10,260 7,064 3,196 69% 1991 9,871 6,649 3,222 67% Source: U.S. Department of Transportation, U.S. International Air Passenger and Freight Statistics, various issues. The strategic position of the two Canadian carriers changed significantly in the two-year period that followed the 1991-92 negotiating round. CAIL had been losing money steadily since 1989. Canadian transport officials were concerned that a potential failure of CAIL could create a domestic monopoly. Their fears were lessened, however, by the 1992 announcement by American Airlines and Canadian Airlines that the two carriers would seek to establish a comprehensive strategic alliance, with American buying a minority interest in the Canadian carrier. A primary benefit of such an alliance to American was the ability to feed its U.S. passengers on to CAIL's transpacific routes, allowing it to compete with the transpacific services of its U.S. rivals, including those of United, its chief nemesis. American would also be able to feed passengers into CAIL's domestic network, similarly raising the attractiveness of its U.S. domestic flights vis-a-vis those of its U.S. domestic rivals for trans-border passengers using connecting services. Canadian Airlines in turn would benefit by flowing its Canadian passengers over American's U.S. network, thereby raising the attractiveness of Canadian Airlines' domestic flights vis-a-vis those of its bitter rival, Air Canada. (45) Air Canada fought a fierce battle against the proposed arrangement, attempting to derail it by offering to buy Canadian Airlines' Asian, European, and South American routes. (46) Despite this opposition, American Airlines succeeded in purchasing a one-third interest in Canadian Airlines in 1994, although by law its voting rights were limited to 25 percent. Canadian Airlines and American Airlines then entered into an extensive code-sharing agreement and coordinated their flight offerings, helping both boost their competitiveness against their respective chief domestic rivals. As part of this arrangement, Canadian Airlines shifted its computerized reservations from the Gemini system, which it had shared with Air Canada and which was allied with United's Apollo system, to American's Sabre system. (47) Table 2 U.S.-Canada Trade in International Passenger Transportation (IPT) Services, 1986-1995 (in millions of U.S. dollars) US IPT Exports US IPT Imports US Share of IPT Trade 1995 $1118 $314 78.1% 1994 $1133 $302 79.0% 1993 $1191 $260 85.3% 1992 $1099 $227 82.9% 1991 $1040 $249 80.7% 1990 $979 $256 79.3% 1989 $811 $224 78.4% 1988 $742 $254 74.5% 1987 $663 $204 76.5% 1986 $551 $212 72.2% Source: Survey of Current Business, various issues Air Canada's strategic position had also changed. It began taking delivery in September, 1994, of the new 50-seat Canadair CL-65 regional jets. It had earmarked 24 of these aircraft for trans-border service under the terms of the 1984 Regional, Local, and Commuter Services Agreement, and held options for an additional 24 aircraft which it could use effectively only if a new liberal ASA were signed. (48) With these small, "hub-busting" jets it hoped to steal a march on its competitors. It believed that these aircraft were uniquely suited to exploiting smaller, underdeveloped trans-border business-oriented routes and would allow such passengers to bypass the fortress hubs operated by its U.S. competitors, thereby weakening an important competitive advantage they enjoyed vis-a-vis Air Canada. Air Canada also adopted an "if you can't beat 'em, join 'em" strategy. It gained access to several important U.S. hubbing complexes by negotiating two strategic alliances with U.S. carriers. It purchased a 24 percent ownership interest in Continental Airlines in 1993 and began to coordinate its flight schedules and code-share with Continental at its Houston and Newark hubs. (49) Yet in early 1996 Air Canada decided to liquidate its investment in the U.S. carrier. It sold the last of its Continental shares in January, 1997, reaping a handsome profit on its investment. Despite the termination of the ownership arrangement, the strategic alliance between the two carriers remains in place. Currently Air Canada and Continental operate code-sharing services on 30 flights in six trans-border markets. (50) Air Canada also entered into a similar code-sharing and flight coordination program with United Airlines to benefit from United's hubs at Chicago, Denver, and San Francisco. While these strategic alliances gained Air Canada quick access to many important U.S. domestic airlines markets, they also served another strategic purpose. Coupled with the entry freedom permitted under the new ASA, they allowed Air Canada to expand its "home" traffic base from that provided by Canada to that provided by North America. They thus improved Air Canada's ability to compete for transatlantic, transpacific, and trans-border traffic in the rapidly globalizing airline industry. (51) B. Provisions of the 1995 ASA As a result of these pressures, the United States and Canada agreed to a new, much liberalized ASA in February, 1995. The new ASA allows each country to designate as many carriers as it wishes to provide trans-border services, a clause consistent with the open entry and multiple designation policies of the United States. Neither country may unilaterally limit the capacity offered by any of these carriers. Carriers are free to charge any prices they wish; such fares remain in effect unless both governments disapprove of them. The grounds for disapproval are limited to preventing unreasonable discrimination or exploitation of a dominant position (fares too high) and protecting carriers from competing against low fares resulting from government subsidies or from low fares designed to eliminate competition. Canadian carriers were given access to scarce slots at LaGuardia and O'Hare and allowed to purchase slots at Washington National airport. (52) The pact offers the Canadian carriers a head start, as they are immediately free to fly between any U.S. and any Canadian city of their choice, with no regulatory restrictions on aircraft size, capacity, or frequency of service. They may also combine service to two U.S. points, although they may not carry local traffic between the two points (e.g., Air Canada may fly Toronto-Kansas City-Houston, but it may not carry local passengers between Kansas City and Houston.) While similar rights will ultimately be granted to U.S. airlines, in the short run their ability to offer service to Vancouver and Montreal (Dorval) was constrained for two years and for three years at Toronto (Pearson). During the first year of the agreement, the U.S. government was allowed to designate six carriers offering two frequencies a day to provide new service to Montreal and Vancouver, and two carriers offering two frequencies a day to provide new service to Toronto. (53) At the beginning of year two, it could make a second series of similar designations. At the beginning of year three, limits on new U.S.-flag service to Vancouver and Montreal end, and the U.S. government may authorize four carriers to provide two new daily frequencies to Toronto. Under certain circumstances, U.S. carriers may increase the frequency of service to the three cities during this transition period to match that offered by Canadian carriers not operating under such constraints. After the transition period is over--twenty-four months in the case of Vancouver and Montreal and thirty-six months in the case of Toronto--U.S. carriers will be able to offer whatever services they wish to any Canadian city they choose. Carriers from both countries will be allowed to code-share their flights as they see fit. The desire of Canadian negotiators to exempt temporarily the primary Montreal, Toronto, and Vancouver (MTV) airports from the ASA's open entry provisions is understandable given the concentration of trans-border traffic at those three airports. Toronto (Pearson), Montreal (Dorval), and Vancouver accounted for 70 percent of the commercial trans-border aircraft movements between the two countries, as Table 3 indicates. Table 4 presents the 25 largest trans-border scheduled markets in 1995. The three MTV airports account for the twelve largest of these markets and 24 out of the 25 largest. Accordingly, the phase-in provides Canadian carriers with some temporary protection at these key airports and would be expected to funnel new U.S.-flag service to other Canadian cities, who were prime lobbyists for a change in the ASA in the first place. V. THE IMPACT OF THE NEW AGREEMENT The impact of the new air services agreement can be measured in a variety of ways. Changes in passenger traffic in the total trans-border market and in individual city-pair markets are perhaps the most direct measure of the new ASA's impact. Unfortunately, data detailing traffic between international city-pairs involving the United States are available to the general public only with a three-year lag. Accordingly, in order to provide current information regarding this rapidly evolving market, this paper will measure the impact of the new ASA on the trans-border market by analyzing the number of new trans-border services introduced by individual carriers. The new agreement has led to a significant increase in the number of trans-border city-pairs receiving service, as indicated by Table 5, which reports the Canadian cities involved in these services, and Table 6, which shows their U.S. counterparts. These tables depict, by flag of carrier, the number of trans-border city-pairs receiving nonstop scheduled service under the old and new ASAs using jet aircraft of 50 seats or more, a definition designed to include the Canadair CL-65 jet but exclude commuter-type turboprop aircraft. "Old" services are defined as those provided in January, 1995; services authorized by the old ASA but not actually being performed during January, 1995, are not included in these tables. "New" services are defined as any added after the new ASA was signed through October, 1996, even if the new service was subsequently abandoned. (54) These tables are based on the number of trans-border city-pairs served by each carrier. Accordingly, they overstate the number of city-pairs actually receiving service because some received service from more than one carrier. (55) Table 3 Trans-border Aircraft Movement 1995 Canadian U.S. Airport Carriers Carriers Total Toronto (Pearson) 56,041 56,220 112,261 Montreal (Dorval) 17,375 27,028 44,403 Vancouver 19,031 26,093 45,124 All others 22,067 63,154 85,221 Total 114,514 172,495 287,009 Source: Statistics Canada A. Impact of the New ASA on Cities As Tables 5 and 6 indicate, under the aegis of the old ASA nonstop jet service was provided in only 67 trans-border markets in January, 1995, and only nine Canadian cities received such service. The majority was concentrated at four Canadian cities: 87 percent of the old trans-border city-pairs involved either Toronto, Montreal, Vancouver, or Calgary. In contrast, under the old ASA the four U.S. cities with the most trans-border services (Chicago, Los Angeles, New York, and San Francisco (or Boston)) accounted for only 25 routes, or 37 percent of the trans-border services. Moreover, 25 U.S. cities received trans-border service. Table 4 Scheduled Trans-border Origin and Destination Air Passengers, 1995 Rank City-Pair Annual 1995 1 New York-Toronto 802,400 2 Montreal-New York 346,700 3 Chicago-Toronto 342,800 4 Los Angeles-Vancouver 299,900 5 Los Angeles-Toronto 265,700 6 Boston-Toronto 228,800 7 San Francisco-Toronto 221,100 8 Miami-Toronto 205,000 9 San Francisco-Vancouver 180,900 10 Tampa/St. Petersburg-Toronto 154,100 11 Miami-Montreal 132,700 12 Atlanta-Toronto 128,700 13 Calgary-Los Angeles 126,100 14 Chicago-Montreal 117,700 15 Boston-Montreal 116,500 16 Dallas-Toronto 115,200 17 Los Angeles-Montreal 114,900 18 Philadelphia-Toronto 113,700 19 Toronto-Washington/Baltimore 105,100 20 Orlando-Toronto 94,500 21 New York-Vancouver 85,700 22 Montreal-Washington/Baltimore 81,100 23 Chicago-Vancouver 72,300 24 Detroit-Toronto 71,900 25 Seattle/Tacoma-Vancouver 71,700 Total of above city-pairs 4,595,200 Total of all city-pairs 10,135,000 Source: Statistics Canada Table 5 Trans-border Nonstop City-Pair Service (By Canadian Terminus) Old Services Total New Services Total Canadian U.S. Old Canadian U.S. New Grand Flag Flag Services Flag Flag Services Total Toronto 9 15 24 24 9 33 57 Montreal 6 9 15 4 9 13 28 Vancouver 3 7 10 9 18 27 37 Calgary 4 5 9 6 5 11 20 Edmonton 2 2 1 1 2 4 Regina 1 1 1 1 2 Saskatoon 0 1 1 1 Winnipeg 1 1 2 3 1 4 6 Halifax 1 1 4 1 5 6 Ottawa 1 2 3 4 2 6 9 Total 26 41 67 55 48 103 170 N.B.: includes nonstop scheduled service using jet aircraft of 50 seats or more a service is defined as a carrier offering at least one weekly nonstop flight in the city-pair market Source: Official Airline Guide, January 1, 1995 and subsequent issues As a result of the new ASA, U.S. and Canadian airlines have added nonstop scheduled services to 103 trans-border city-pairs through October, 1996. These new routes serve ten different Canadian cities. Among Canadian cities, the primary beneficiaries of the new ASA have been Toronto and Vancouver, which have received over half (60 of 103) of the new services. In absolute terms, Toronto appears to be the big winner, benefiting from approximately one-third of the new flights. Services to Vancouver, however, more than tripled, so that in relative terms Vancouver has done better than Toronto. These two cities plus Montreal and Calgary received 84 new services, or 82 percent of the total. Accordingly, under the new ASA there has been a slight decrease in the concentration of trans-border services at these four cities. Table 6 Trans-border Nonstop City-Pair Service (By U.S. Terminus) Old Services Total New Services Total Canadian U.S. Old Canadian U.S. New Grand Flag Flag Services Flag Flag Services Total Atlanta 0 2 3 5 5 Baltimore 2 2 0 2 Boston 3 2 5 1 1 6 Chicago 4 4 8 3 3 6 14 Cleveland 1 1 0 1 Cincinnati 2 2 1 1 3 Dallas-FW 2 2 2 2 4 Denver 1 1 3 2 5 6 Detroit 2 2 3 3 5 Fort Lauder 0 3 3 3 Ft. Meyers 0 2 2 2 Honolulu 2 2 1 1 3 Houston 1 1 1 2 3 4 Kansas C 0 1 1 1 Las Vegas 0 5 1 6 6 Los Angel 4 2 6 2 2 4 10 Maui 0 1 1 1 Miami 2 2 4 3 3 7 Milwaukee 0 1 1 1 Minneapol 1 2 3 1 5 6 9 Nashville 1 1 1 1 2 New York 4 2 6 1 2 3 9 Newark 0 3 3 3 Orlando 0 7 7 7 Palm Spr 0 3 3 3 Philadelph 2 2 1 1 3 PortlandO 1 1 0 1 Phoenix 0 1 1 2 2 Pittsburg 3 3 1 1 4 Raleigh 0 1 1 1 Reno 0 1 1 2 2 Rochester 1 1 0 1 San Diego 0 1 1 1 San Franc 3 2 5 1 2 3 8 San Jose 1 1 0 1 St. Louis 0 1 1 2 2 St. Petersb 0 1 1 1 Salt LakeC 2 2 1 1 3 Sarasota 0 1 1 1 Seattle 1 1 2 2 3 Spokane 2 2 0 2 Tampa 2 1 3 4 1 5 8 Wash-Nat 0 2 2 4 4 Wash-Dull 0 2 1 3 3 W Palm B 0 2 2 2 Total 26 41 67 55 48 103 170 N.B.: includes nonstop scheduled service using jet aircraft of 50 seats or more a service is defined as a carrier offering at least one weekly nonstop flight in the city-pair market Source: Official Airline Guide, January 1, 1995 and subsequent issues In contrast, 38 U.S. cities received new trans-border flights under the new ASA. The four U.S. cities with the most new trans-border services (Chicago, Las Vegas, Minneapolis, and Orlando) accounted for only 25 routes, or 24 percent of the new trans-border city-pairs. Because of the economic geography of the two countries, the fact that more U.S. cities than Canadian cities received new service is not surprising. In light of the strong support for liberalization of the old ASA by various Canadian community economic development groups, it is somewhat surprising that trans-border service has remained concentrated in the four Canadian cities. As we will discuss later in the paper, an explanation for this result lies in the entry strategies that individual carriers adopted. But a different picture of changes in the trans-border market emerges if we exclude charter conversions from the analysis. Over half of the new Canadian flag service under the new ASA represents conversion of charter services to scheduled services. (56) Much of this is provided on a less-than-daily basis, and some of it is seasonal as well. Moreover, there is some doubt whether all such services are truly "new"; operationally, in many cases little has changed except their regulatory classification. Tables 7 and 8 report data similar to that found in Tables 5 and 6, except that charter conversions are excluded. Canadian flag airlines--Air Canada, Canadian Airlines, and numerous small charter specialists--dominated the trans-border charter market which primarily served Canadian residents who wished to escape to the warmer climates of Arizona, Florida, and Hawaii in winter. (57) When these exclusions are made, we observe that new services in Canada are as concentrated as the old services. Toronto, Vancouver, Montreal, and Calgary received 87 percent of the new services (62 of 71 new routes) offered under the new ASA, a percentage equal to their share under the old ASA. In contrast, the four U.S. cities receiving the most new service under the new ASA (Chicago, Minneapolis, Atlanta, and Denver) accounted for only 22 of the 71 new routes, or 31 percent of the total, when charter conversions are excluded. B. Analysis of the Impact by Flag of Carrier As Tables 5 and 6 indicate, under the old ASA Canadian flag carriers provided service on only 26 routes, or 39 percent of the total served, while U.S. flag carriers served 41 city-pair routes, or 61 percent of the total. Of the 103 new routes, U.S. carriers provided service on 48 (47 percent of the total new routes) and Canadian carriers on 55 (53 percent). Thus Canadian carriers appear to have been more aggressive than U.S. carriers in adding new services and have done so at a rate disproportionate to their share of routes under the old ASA. But again the picture changes if one excludes the 32 charter conversions reported in Tables 5 and 6. As Tables 7 and 8 indicate, the share of new services offered by Canadian carriers is only 32 percent absent charter conversions. Whether Canadian flag carriers have been more aggressive than U.S. carriers (or vice versa) in exploiting the new ASA thus depends on whether charter conversions are included in the analysis or not. C. New Service to the MTV Airports According to Table 5, 73 percent (49 of 67) of the city-pairs served under the old ASA involved Montreal, Toronto, and Vancouver (MTV). As noted earlier, the ability of the Canadian flag carriers to exploit the new ASA was unrestricted, while limitations were placed on U.S. flag carriers in trans-border markets involving the MTV airports. (58) Despite the restricted access of U.S. flag carriers to the MTV airports, 73 of the 103 new services, or 71 percent, involved the MTV airports. Thus it appears that the restrictions did little to disperse trans-border services. Canadian flag carriers clearly gained a head start at Toronto by exploiting this restriction, as they offered 24 of the 33 new services at that city, or 72 percent. Conversely, the head start at the other two restricted airports appears to have been less important. U.S. carriers offered 69 percent of the new services at Montreal and 67 percent of the new services at Vancouver. But many of these new services at the MTV airports by Canadian carriers represent conversion of charter services to scheduled services. As Table 7 indicates, if charter conversions are excluded, concentration of new services at the MTV airports increases: 77 percent of the new trans-border routes involve the MTV airports, while only 73 percent of the old trans-border routes served these three cities. Moreover, the dominance of Canadian flag carriers at Toronto is reduced and the dominance of U.S. flag carriers at Montreal and Vancouver is increased: U.S. carriers account for 82 percent of the new services involving Montreal and Vancouver, and 41 percent of those involving Toronto. Table 7 Trans-border Nonstop City-Pair Service Excluding Charter Conversions (By Canadian Terminus) Old Services Total New Services Total Canadian U.S. Old Canadian U.S. New Grand Flag Flag Services Flag Flag Services Total Toronto 9 15 24 13 9 22 46 Montreal 6 9 15 2 9 11 26 Vancouver 3 7 10 4 18 22 32 Calgary 4 5 9 2 5 7 16 Edmonton 2 2 1 1 3 Regina 1 1 1 1 2 Saskatoon 0 1 1 1 Winnipeg 1 1 2 1 1 3 Halifax 1 1 1 1 2 Ottawa 1 2 3 2 2 4 7 Total 26 41 67 23 48 71 138 N.B.: includes nonstop scheduled service using jet aircraft of 50 seats or more a service is defined as a carrier offering at least one weekly nonstop flight in the city-pair market Source: Official Airline Guide, January 1, 1995 and subsequent issues D. New Service to non-MTV Airports In eastern and central Canada, new trans-border services were initiated in 11 city-pair markets using airports other than Toronto or Montreal. On a percentage basis, this area seemingly was a major beneficiary of the new ASA, as it received service on only four city-pairs under the old ASA. And it also appears that the Canadian flag carriers moved more aggressively than U.S. flag carriers in exploiting the new ASA at unrestricted airports in the eastern half of the country. Of the 11 new scheduled services, eight were initiated by Canadian flag carriers and only three by U.S. flag carriers. Yet both these assertions are somewhat misleading. Because so much of the "new" service involves charter conversions, the seemingly dramatic improvement of trans-border air services at cities in eastern and central Canada other than Montreal and Toronto would appear to be overstated. For the same reason, the perception created by Table 5 that Canadian carriers have moved more aggressively than U.S. flag carriers to serve non-MTV city-pairs in this region may be somewhat misleading, as a comparison with Table 7 indicates. Of the eight new non-MTV Canadian flag nonstop routes involving eastern and central Canada, six serve Orlando and Tampa. On these six routes, less than daily service is provided, and most of these services are offered only seasonally. Only two of the newly-initiated Canadian flag services (Air Canada's Ottawa-Chicago and Ottawa-Dulles routes) provide daily, year-round service to trans-border travelers. Conversely, the three new routes established by U.S. flag carriers provide trans-border travelers the daily opportunity to conveniently connect to other flights at these carriers' hubs throughout the year. In western Canada (excluding Vancouver), 19 new trans-border routes were added as a result of the new ASA. Calgary has been the primary beneficiary of this new service: eleven of the new routes have Calgary as their Canadian terminus, although four of them represent charter conversions. Five western Canadian cities have been the recipients of new trans-border service, a broader spreading of the benefits of the new ASA than has been the case in eastern and central Canada. These new western routes are split evenly between Canadian flag and U.S. flag carriers. But eight of the new Canadian flag services represent conversion of charter service to scheduled service. It appears then, that U.S. flag carriers moved more aggressively into the western Canadian trans-border market than Canadian carriers if charter conversions are excluded from the analysis. E. Individual Canadian Flag Carrier Responses The strategies adopted by Air Canada and Canadian Airlines to exploit the new ASA differed dramatically. These differences reflect in part their relative strengths and weaknesses resulting from Canadian domestic and international airline policies and in part the precarious financial situation of CAIL at the time the new air services agreement was signed. Table 8 Trans-border Nonstop City-Pair Service Excluding Charter Conversions (By U.S. Terminus) Old Services Total New Services Total Canadian U.S. Old Canadian U.S. New Grand Flag Flag Services Flag Flag Services Total Atlanta 0 2 3 5 5 Baltimore 2 2 0 2 Boston 3 2 5 1 1 6 Chicago 4 4 8 3 3 6 14 Cleveland 1 1 0 1 Cincinnati 2 2 1 1 3 Dallas-FW 2 2 2 2 4 Denver 1 1 3 2 5 6 Detroit 2 2 3 3 5 Fort Lauder 0 0 0 Ft. Meyers 0 0 0 Honolulu 2 2 0 2 Houston 1 1 1 2 3 4 Kansas C 0 1 1 1 Las Vegas 0 1 1 1 Los Angel 4 2 6 1 2 3 9 Maui 0 0 0 Miami 2 2 4 3 3 7 Milwaukee 0 1 1 1 Minneapol 1 2 3 1 5 6 9 Nashville 1 1 1 1 2 New York 4 2 6 1 2 3 9 Newark 0 3 3 3 Orlando 0 0 0 Palm Spr 0 0 0 Philadelph 2 2 1 1 3 PortlandO 1 1 0 1 Phoenix 0 1 1 1 Pittsburg 3 3 1 1 4 Raleigh 0 1 1 1 Reno 0 1 1 1 Rochester 1 1 0 1 San Diego 0 1 1 1 San Franc 3 2 5 1 2 3 8 San Jose 1 1 0 1 St. Louis 0 1 1 2 2 St. Petersb 0 0 0 Salt LakeC 2 2 1 1 3 Sarasota 0 0 0 Seattle 1 1 2 2 3 Spokane 2 2 0 2 Tampa 2 1 3 1 1 2 5 Wash-Nat 0 2 2 4 4 Wash-Dull 0 2 1 3 3 W Palm B 0 0 0 Total 26 41 67 23 48 71 138 N.B.: includes nonstop scheduled service using jet aircraft of 50 seats or more a service is defined as a carrier offering at least one weekly nonstop flight in the city-pair market Source: Official Airline Guide, January 1, 1995 and subsequent issues Air Canada Under the old ASA, Air Canada provided service to more city-pairs than any other carrier in the trans-border market. As Table 9 indicates, it served 21 nonstop scheduled markets prior to the signing of the new ASA. Most of its service was focused in eastern and central Canada (16 of 21 routes); most of its service was concentrated at Toronto and Montreal as well (14 of 21 routes). This service pattern reflected, of course, Canadian domestic regulatory policies which gave Air Canada the primary role in serving the eastern half of the country and CP Air, the corporate predecessor to Canadian Airlines, the primary role in the west. As Table 10 indicates, Air Canada's trans-border routes under the old ASA served 8 U.S. cities, most of which can be characterized as being traditional U.S. gateways. Air Canada has been by far the most aggressive carrier in exploiting the new aviation accord in both words and deeds. Two weeks prior to the agreement's signing, Air Canada publicly announced that it would offer new service from Ottawa and Halifax to Chicago and expand service between Toronto and New York if the Canadian government granted it the landing slots at La Guardia and O'Hare promised by the pending agreement. On the day of signing (February 24), Air Canada said it would provide service on at least 20 new routes within 18 months. Two days after the signing an Air Canada press release proudly trumpeted that "Air Canada is first off the mark for open skies service" with its March 6 commencement (i.e., ten days after the signing) of daily service between Toronto and Atlanta. (59) But Air Canada's seemingly optimistic projections proved understated. By October, 1996, it had newly entered 31 nonstop trans-border markets. As Table 9 shows, of its 31 new routes, 24 involve the restricted cities of Montreal, Toronto, and Vancouver. Most of its new routes (23 of 31) are concentrated in eastern and central Canada. But Air Canada has also expanded its trans-border services into western Canada, offering eight new routes--five from Vancouver, two from Calgary, and one from Winnipeg. It also significantly broadened the pattern of U.S. cities that it served, adding 16 new U.S. cities under the 1995 ASA, as Table 11 indicates. If charter conversions are excluded, Air Canada's expansion appears somewhat less extensive; it offered only 19 new services under the new ASA and expanded into nine new U.S. cities, as Tables 11 and 12 demonstrate. But the general pattern of entry remains the same. Most of Air Canada's new services involve the MTV markets, and the carrier's expansion into trans-border markets serving western Canada remains noteworthy. Table 9 Trans-border Nonstop City-Pair Service By Canadian Flag Carriers Air Canada Total CAIL Total Grand Old New Services Old New Services Total Toronto 8 15 23 1 9 10 33 Montreal 6 4 10 0 10 Vancouver 5 5 3 4 7 12 Calgary 4 2 6 4 4 10 Edmonton 0 1 1 1 Regina 1 0 1 1 1 Saskatoon 0 0 0 Winnipeg 1 1 2 2 2 4 0 Halifax 1 2 3 2 2 5 Ottawa 1 2 3 2 2 5 Total 21 31 52 5 24 29 81 N.B.: includes nonstop scheduled service using jet aircraft of 50 seats or more a service is defined as a carrier offering at least one weekly nonstop flight in the city-pair market Source: Official Airline Guide, January 1, 1995 and subsequent issues Regardless of whether charter conversions are excluded or not, Air Canada has added more new trans-border service than any other North American carrier. Of the 103 new routes added subsequent to the signing of the new ASA, nearly one-third were attributable to Air Canada. Examination of the pattern of new services offered by Air Canada suggests that it has adopted several strategies: Table 10 Old Trans-border Services By Canadian Carriers Air Total Old Canada CAIL Services Atlanta 0 Baltimore 0 Boston 3 3 Chicago 4 4 Cleveland 0 Cincinnati 0 Dallas-FW 0 Denver 0 Detroit 0 Fort Lauder 0 Ft. Meyers 0 Honolulu 2 2 Houston 1 1 Kansas C 0 Las Vegas 0 Los Angel 3 1 4 Maui 0 Miami 2 2 Milwaukee 0 Minneapol 1 1 Nashville 0 New York 4 4 Newark 0 Orlando 0 Palm Spr 0 Philadelph 0 PortlandO 0 Phoenix 0 Pittsburg 0 Raleigh 0 Reno 0 Rochester 0 San Diego 0 San Franc 2 1 3 San Jose 0 St. Louis 0 St. Petersb 0 Salt LakeC 0 Sarasota 0 Seattle 0 Spokane 0 Tampa 2 2 Wash-Nat 0 Wash-Dull 0 W Palm B 0 Total 21 5 26 N.B.: includes nonstop scheduled service using jet aircraft of 50 seats or more a service is defined as a carrier offering at least one weekly nonstop flight in the city-pair market Source: Official Airline Guide, January 1, 1995 and subsequent issues Table 11 New Trans-border NONSTOP Services By Canadian Carriers All Services Services Excluding Charter Conversions Air Total New Air Total New Canada CAIL Services Canada CAIL Services Atlanta 2 2 2 2 Baltimore 0 0 Boston 0 0 Chicago 1 2 3 1 2 3 Cleveland 0 0 Cincinnati 0 0 Dallas-FW 0 0 Denver 3 3 3 3 Detroit 0 0 Fort Lauder 2 1 3 0 Ft. Meyers 1 1 2 0 Honolulu 1 1 0 Houston 1 1 1 1 Kansas C 1 1 1 1 Las Vegas 1 2 5 0 Los Angel 1 1 2 1 1 Maui 1 1 0 Miami 0 0 Milwaukee 0 0 Minneapol 1 1 1 1 Nashville 1 1 1 1 New York 1 1 1 1 Newark 0 0 Orlando 4 3 7 0 Palm Spr 3 3 0 Philadelph 1 1 1 1 PortlandO 0 0 Phoenix 1 1 0 Pittsburg 0 0 Raleigh 1 1 1 1 Reno 1 1 0 Rochester 0 0 San Franc 1 1 1 1 San Jose 0 0 St. Louis 1 1 1 1 St. Petersb 1 1 0 Salt LakeC 0 0 Sarasota 1 1 0 Seattle 0 0 Spokane 0 0 Tampa 1 3 4 1 1 Wash-Nat 2 2 2 2 Wash-Dull 2 2 2 2 W Palm B 1 1 2 0 Total 31 24 55 19 4 23 N.B.: includes nonstop scheduled service using jet aircraft of 50 seats or more a service is defined as a carrier offering at least one weekly nonstop flight in the city-pair market Source: Official Airline Guide, January 1, 1995 and subsequent issues 1) Exploited the temporary constraints imposed on U.S. carriers at the three phase-in airports by introducing extensive new services there. This is particularly noticeable at Toronto, which historically has been the most important city in Air Canada's route network. 2) Captured first-mover advantages in U.S. cities unserved or underserved under the old ASA. For example, the Washington, D.C. area became an important target for Air Canada. Trans-border service to the Washington area was quite limited prior to the signing of the ASA. But in May, 1995, Air Canada introduced thrice daily service between Ottawa and Washington (Dulles); in June it added five daily Toronto-Washington (National) flights and three Montreal-Washington (National) flights. In October it commenced thrice daily service between Toronto and Washington (Dulles). Air Canada has also introduced service to smaller metropolitan areas such as Raleigh-Durham, Nashville, and Kansas City. Air Canada has focused on providing these new services from Toronto, where the ability of U.S. carriers to respond is restricted by the phase-in features of the new ASA. (60) 3) In these short-haul business markets, it offered frequent service by using small jet aircraft (i.e., 50-seat Canadair CL-65 jet aircraft). By so doing, Air Canada hopes to build market identity, brand loyalty, and market share before U.S.- flag carriers can legally enter these markets. 4) Utilized code-sharing agreements to broaden the network of cities to which it can transport trans-border travelers. The carrier's code-share agreement with United Airlines in particular appears to have offered Air Canada a low-cost way to gain broad market identity to compete with CAIL in western trans-border markets. Canadian Airlines This firm was substantially disadvantaged under the old ASA. As Table 9 indicates, it operated only 5 trans-border services under that agreement. Four of these five routes served Canadian's traditional bailiwick of western Canada. CAIL has added non-stop scheduled service to 24 city-pairs under the new ASA, including 13 new U.S. destinations, and more than half of these markets involve eastern and central Canada. It would appear that as a result of the new ASA CAIL has made a major commitment to the trans-border market and to eastern and central Canada. But much of this expansion is the result of charter conversions. For example, of its nine new routes from Toronto, six represent charter conversions as a comparison of Tables 9 and 12 indicates. If charter conversions are excluded, CAIL's new trans-border markets shrink to four routes, three serving Toronto and one serving Vancouver, as Table 12 shows. These new routes serve only 3 U.S. cities, all of which are new to CAIL's scheduled route system. Table 12 Trans-border Nonstop City-Pair Service By Canadian Flag Carriers Excluding Charter Conversions Air Canada Total CAIL Total Grand Old New Services Old New Services Total Toronto 8 10 18 1 3 4 22 Montreal 6 2 8 0 8 Vancouver 3 3 3 1 4 7 Calgary 4 2 6 0 6 Edmonton 0 0 0 Regina 0 1 1 1 Saskatoon 0 0 0 Winnipeg 1 1 0 1 0 Halifax 1 1 0 1 Ottawa 1 2 3 0 3 Total 21 19 40 5 4 9 49 N.B.: includes nonstop scheduled service using jet aircraft of 50 seats or more a service is defined as a carrier offering at least one weekly nonstop flight in the city-pair market Source: Official Airline Guide, January 1, 1995 and subsequent issues CAIL has been far less aggressive than Air Canada in exploiting the window of opportunity granted by the new ASA. This is particularly noticeable if charter conversions are excluded: CAIL's addition of 4 new city-pairs pales in comparison to Air Canada's 19 new services. CAIL's behavior is the result of several factors. First, CAIL was able to provide very little service under the old ASA. It operated in only five trans-border city-pairs. Accordingly, its start-up costs for entering most new trans-border markets were much higher than Air Canada's. It lacked gate space, check-in facilities, landing slots, and, above all, market visibility in most U.S. cities. Second, CAIL is suffering from severe financial difficulties that no doubt have hindered its ability to exploit the new ASA. (61) CAIL has attempted to circumvent these difficulties by entering into a strategic alliance with American Airlines. CAIL operates relatively few trans-border flights itself, but its customers have access to numerous U.S. cities by flying on trans-border code-share flights operated by American Airlines. (62) F. Individual U.S. Flag Carrier Responses Five U.S. flag carriers offered nonstop trans-border jet service under the old ASA: American, United, Delta, US Air, and Northwest. Seven new U.S. flag carriers, in addition to these five, have taken advantage of the new ASA to provide trans-border service. American Airlines Traditionally the number two carrier in the trans-border market in terms of passengers carried, American Airlines has not moved as aggressively as Air Canada in exploiting the new ASA. As Tables 13 and 15 indicate, it served only seven city-pairs under the old bilateral agreement, six of which involved the MTV airports. But the routes it did serve are among the largest in the trans-border market. For example, American Airlines serves two of the three largest trans-border city-pair markets, Toronto-Chicago and Toronto-New York, under terms of the 1966 ASA. American added service to ten new city-pairs under the new agreement, as Tables 14 and 16 report. Of its ten new routes, seven involved MTV service. Outside of these cities, American added service from Chicago to Ottawa, Calgary, and Winnipeg, although the Chicago-Winnipeg service was subsequently abandoned. Eight of its ten new city-pairs involved American hub cities: Miami (Montreal, Toronto, and Vancouver); Dallas-Fort Worth (Montreal and Vancouver); and Chicago (Ottawa, Calgary, and Winnipeg). An important component of American Airlines' trans-border strategy is its code-share alliance with Canadian Airlines, a carrier partly owned by American Airlines. American appears to be the dominant carrier in this alliance. CAIL is the actual provider of code-share services in only three trans-border markets. American's primary benefit from this strategic alliance is its ability to feed its passengers on to CAIL's transpacific flights, as well as on to CAIL's extensive array of intra-Canada domestic flights. [Part 1 of 2] Table 13 Old Trans-border Nonstop Services By U.S. Carriers Amer. Contin- Delta NW United US Air Other Airlines ental Airlines Airlines Airlines Toronto 4 4 1 2 4 Montreal 1 4 1 3 Vancouver 1 3 1 2 Calgary 1 3 1 Edmonton 1 1 Regina Saskatoon Winnipeg 1 Halifax Ottawaa 2 Total 7 0 15 5 5 9 0 [Part 2 of 2] Table 13 Old Trans-border Nonstop Services By U.S. Carriers Grand Total Toronto 15 Montreal 9 Vancouver 7 Calgary 5 Edmonton 2 Regina 0 Saskatoon 0 Winnipeg 1 Halifax 0 Ottawaa 2 Total 41 N.B.: includes nonstop scheduled service using jet aircraft of 50 seats or more a service is defined as a carrier offering at least one weekly nonstop flight in the city-pair market Source: Official Airline Guide, January 1, 1995 and subsequent issues United Airlines American's main domestic rival, United Airlines, has followed a similar strategy. It operated in only five trans-border city-pairs under the old ASA. Two of these routes served its Chicago hub, while the other three provided service from western Canada to West Coast cities where United has a large presence. United added five routes under the new ASA, all of which served western Canada and major cities on its route structure in the western United States. [Part 1 of 2] Table 14 New Trans-border Services By U.S. Carriers Amer. Contin- Delta NW United US Air Other Airlines ental Airlines Airlines Airlines Toronto 2 2 1 2 2 Montreal 2 1 1 1 3 1 Vancouver 3 2 3 2 3 5 Calgary 1 1 2 1 Edmonton 1 Regina 1 Saskatoon 1 Winnipeg 1 Halifax 1 Ottawa 1 1 Total 10 5 5 8 5 5 10 [Part 2 of 2] Table 14 New Trans-border Services By U.S. Carriers Grand Total Toronto 9 Montreal 9 Vancouver 18 Calgary 5 Edmonton 1 Regina 1 Saskatoon 1 Winnipeg 1 Halifax 1 Ottawa 2 Total 48 N.B.: includes nonstop scheduled service using jet aircraft of 50 seats or more a service is defined as a carrier offering at least one weekly nonstop flight in the city-pair market Source: Official Airline Guide, January 1, 1995 and subsequent issues An important component of United's trans-border strategy is a code-sharing agreement with Air Canada which is far more balanced than American's arrangement with CAIL. Most of United's code-share flights operated by Air Canada provide trans-border service from eastern and central Canada to United's hubs at Chicago O'Hare and Washington (Dulles) airports. This arrangement reduces United's disadvantage relative to American Airlines in eastern trans-border services. Similarly, most of Air Canada's code-share flights operated by United serve western trans-border markets where Air Canada is at a disadvantage relative to CAIL. [Part 1 of 2] Table 15 Old Trans-border Services By U.S. Carriers Amer. Contin- Delta NW United US Air Other Airlines ental Airlines Airlines Airlines Atlanta Baltimore 2 Boston 1 1 Chicago 2 2 Cleveland 1 Cincinnati 2 Dallas-FW 2 Denver 1 Detroit 2 Fort Lauder Ft. Meyers Honolulu Houston Kansas C Las Vegas Los Angeles 2 Maui Miami 2 Milwaukee Minneapol 2 Nashville 1 New York 1 1 Newark Orlando Palm Spr Philadelph 2 PortlandO 1 Phoenix Pittsburg 1 2 Raleigh Reno Rochester 1 San Franc 1 1 San Jose 1 St. Louis St. Petersb Salt LakeC 2 Sarasota Seattle 1 Spokane 1 1 Tampa 1 Wash-Nat Wash-Dull W Palm B Total 7 0 15 5 5 9 0 [Part 2 of 2] Table 15 Old Trans-border Services By U.S. Carriers Grand Total Atlanta 0 Baltimore 2 Boston 2 Chicago 4 Cleveland 1 Cincinnati 2 Dallas-FW 2 Denver 1 Detroit 2 Fort Lauder 0 Ft. Meyers 0 Honolulu 0 Houston 0 Kansas C 0 Las Vegas 0 Los Angeles 2 Maui 0 Miami 2 Milwaukee 0 Minneapol 2 Nashville 1 New York 2 Newark 0 Orlando 0 Palm Spr 0 Philadelph 2 PortlandO 1 Phoenix 0 Pittsburg 3 Raleigh 0 Reno 0 Rochester 1 San Franc 2 San Jose 1 St. Louis 0 St. Petersb 2 Salt LakeC 2 Sarasota 0 Seattle 1 Spokane 2 Tampa 1 Wash-Nat 0 Wash-Dull 0 W Palm B 0 Total 41 N.B.: includes nonstop scheduled service using jet aircraft of 50 seats or more a service is defined as a carrier offering at least one weekly nonstop flight in the city-pair market Source: Official Airline Guide, January 1, 1995 and subsequent issues [Part 1 of 2] Table 16 New Trans-border Services By U.S. Carriers Amer. Contin- Delta NW United US Air Other Airlines ental Airlines Airlines Airlines Atlanta 3 Baltimore Boston 1 Chicago 3 Cleveland Cincinnati 1 Dallas-FW 2 Denver 2 Detroit 3 Fort Lauder Ft. Meyers Honolulu Houston 2 Kansas C Las Vegas 1 Los Angeles 1 1 Maui Miami 3 Milwaukee 1 Minneapol 5 Nashville New York 1 1 Newark 3 Orlando Palm Spr Philadelph PortlandO Phoenix 1 Pittsburg 1 Raleigh Reno 1 Rochester San Diego 1 San Franc 2 San Jose St. Louis 1 St. Petersb Salt LakeC 1 Sarasota Seattle 2 Spokane Tampa 1 Wash-Nat 2 Wash-Dull 1 W Palm B Total 10 5 5 8 5 5 10 [Part 2 of 2] Table 16 New Trans-border Services By U.S. Carriers Grand Total Atlanta 3 Baltimore 0 Boston 1 Chicago 3 Cleveland 0 Cincinnati 1 Dallas-FW 2 Denver 2 Detroit 3 Fort Lauder 0 Ft. Meyers 0 Honolulu 0 Houston 2 Kansas C 0 Las Vegas 1 Los Angeles 2 Maui 0 Miami 3 Milwaukee 1 Minneapol 5 Nashville 1 New York 1 Newark 3 Orlando 0 Palm Spr 0 Philadelph 0 PortlandO 0 Phoenix 1 Pittsburg 1 Raleigh 0 Reno 1 Rochester 0 San Diego 1 San Franc 2 San Jose 0 St. Louis 1 St. Petersb 0 Salt LakeC 1 Sarasota 0 Seattle 2 Spokane 0 Tampa 1 Wash-Nat 2 Wash-Dull 1 W Palm B 0 Total 48 N.B.: includes nonstop scheduled service using jet aircraft of 50 seats or more a service is defined as a carrier offering at least one weekly nonstop flight in the city-pair market Source: Official Airline Guide, January 1, 1995 and subsequent issues Northwest Airlines This airline has followed a slightly different strategy than other U.S. carriers in exploiting the new ASA. Like other U.S. flag carriers, its new trans-border services represent new spokes at its existing U.S. hubs. However, unlike other U.S. carriers, Northwest has chosen to serve smaller Canadian cities as well as larger ones. It has provided the only new U.S. flag service to Regina, Saskatoon, and Halifax. Northwest served but five city-pairs under the old ASA, and four of these involved its two hubs at Detroit and Minneapolis. Under the new ASA, Northwest added service on eight new city-pairs, three of which serve its hub at Detroit and five of which serve Minneapolis. Unlike most other trans-border carriers, the majority of Northwest's new routes do not serve the three restricted airports. New service on city-pairs involving western Canada included Calgary, Regina, Saskatoon, and Vancouver to Minneapolis, and Vancouver to Detroit. Northwest expanded its service in central and eastern Canada to include Halifax-Detroit, Ottawa-Detroit, and Montreal-Minneapolis. (63) Clearly, Northwest's strategy is to incorporate its new trans-border services into its existing hubs at Minneapolis and Detroit. Moreover, as geography dictates, it has chosen to maintain its existing strategy of adding service from western Canada to Minneapolis and from central and eastern Canada to Detroit. By so doing, Northwest hopes to feed trans-border passengers onto its domestic, transpacific, and transatlantic flights. Delta Airlines Transporter of the third largest number of trans-border passengers under the old ASA, Delta operated in more trans-border city-pairs than any other U.S. carrier under that agreement. Eleven of its 15 old routes served the MTV airports. Eleven different U.S. cities were as the termini of these routes. Despite its extensive existing trans-border service, Delta has moved very cautiously in exploiting the new ASA. It has added only five new routes, all of which serve restricted airports in Canada and Delta hub cities in the United States. It has added new service between Atlanta and Toronto, Montreal, and Vancouver. It has also added service between Vancouver and two of its other hubs, Cincinnati and Salt Lake City. US Air Under the old ASA, US Air had the second most extensive trans-border route network among the U.S. flag carriers. Consistent with its focus in the northeastern United States, all of US Air's nine routes under the 1966 agreement served eastern and central Canada. US Air adopted a similar strategy in exploiting the new ASA. Its five new routes complement US Air's existing trans-border routes, tying Montreal and Toronto to key northeastern cities in US Air's domestic route network. Continental Airlines Among the U.S. carriers that provided no trans-border service in January, 1995, Continental Airlines has offered the most new flights. But its commitment to the trans-border market has been ambivalent. In the initial allocation of routes in the restricted MTV markets by the U.S. Department of Transportation in spring, 1995, Continental received several plum route authorities. It was granted permission to fly between its Newark hub and Montreal and Vancouver. It also received permission to fly between its Houston hub and Toronto and Vancouver. (64) It commenced these services in June, 1995. A year later it received permission to serve the Toronto-Newark city pair. In spite of these regulatory blessings, Continental abandoned both of the new Houston routes as well as the Vancouver-Newark market in fall, 1995, after they failed to meet the carrier's profitability targets. To date it has not chosen to enter any unrestricted city-pairs. Continental's behavior in part may be explained by its code-sharing arrangement with Air Canada as well as the latter's substantial minority interest in Continental. Either of these factors may have reduced its incentive to exploit its new trans-border authorities aggressively. Moreover, this code-share alliance was somewhat imbalanced. Initially, the two carriers code-shared 30 flights, 26 of which were operated by Air Canada or its commuter airline affiliates and 4 by Continental. (65) By deferring to its code-share partner and allowing Air Canada to provide these services, Continental reduced its start-up and operating costs on these trans-border routes. However, it left itself vulnerable to changes in Air Canada's corporate strategy. In April, 1996, Air Canada announced it would dispose of its minority interest in Continental, arguing that with the advent of the new ASA its alliance with Continental was of lessened strategic value. (66) Air Canada sold its last remaining shares of Continental in January, 1997. Other U.S. Carriers Six other U.S. airlines have introduced new service to ten trans-border city-pairs. Services by these carriers, the majority of which are post-deregulation "start-up" companies, are concentrated in the western half of the continent (7 of 10 new routes). Most of this service also involves the carriers' hub cities. For example, Midwest Express introduced Milwaukee-Toronto service, while TWA developed new service between Toronto and its St. Louis hub, and Reno Air initiated Vancouver- Reno service. In summarizing the initial carrier response to the new air accord, several factors stand out. Air Canada's rapid and broad exploitation of the trans-border market distinguishes it from every other U.S. and Canadian flag carrier. While most of its new entry reflects the firm's traditional strengths in central and eastern Canada, it did use the new ASA and its code-share alliance with United Airlines to expand its participation in western trans-border markets where it had had little previous market presence. Air Canada's behavior also supports the argument raised earlier that its acquiescence was critical to the successful negotiation of the 1995 ASA. Conversely, the response of other carriers to the new accord has been far less rapid and extensive than Air Canada's. For example, American Airlines, which has added more trans-border city-pairs to its route structure than any other carrier except Air Canada if charter conversions are excluded, has added less than half the new routes of Air Canada. Moreover, most of these other carriers have focused on leveraging their existing strengths when entering new trans-border markets. Most new services by U.S. carriers, for example, tie Canadian cities to their U.S. hubs. Such changes do benefit consumers, who are better able to make convenient connections at these hub cities. Lastly, the lack of interest by U.S. and Canadian flag carriers in adding service to unrestricted cities is somewhat surprising. Several carriers have chosen to rely on code-share partners to gain them access to these markets. While Continental, for example, advertises new flights to Calgary, Quebec City, Ottawa, and Halifax, this service is actually performed by Air Canada or commuter carriers affiliated with Air Canada. Among unrestricted cities, only Ottawa and Calgary have attracted much new service. Only one carrier, Northwest Airlines, has utilized a strategy that is predicated