Comments Of The Department Of Justice

This document is available in two formats: this web page (for browsing content) and PDF (comparable to original document formatting). To view the PDF you will need Acrobat Reader, which may be downloaded from the Adobe site. For an official signed copy, please contact the Antitrust Documents Group.


BEFORE THE
DEPARTMENT OF TRANSPORTATION
WASHINGTON, D.C.



Joint Application of

American Airlines, Inc.
and
British Airways PLC

for approval of and antitrust immunity   
for alliance agreement


|
|
|
|
|
|
|
|
|
|
|
|         
Docket OST-97-2058

COMMENTS OF THE DEPARTMENT OF JUSTICE

The United States Department of Justice ("DOJ") hereby submits comments on the application of American Airlines, Inc. ("American" or "AA") and British Airways Plc. ("British Airways" or "BA") under 49 U.S.C. §§ 41308-09 for approval and antitrust immunity for an alliance agreement between the two air carriers.

The Alliance as proposed will significantly reduce competition in many U.S.-U.K. city pairs without producing sufficient efficiencies to outweigh the harm. Divestiture conditions, primarily slot divestitures at London Heathrow Airport ("Heathrow" or "LHR") can reduce that harm, but will not eliminate it in all affected city-pair markets. Hence, if DOJ were reviewing the Alliance under the antitrust laws, we would oppose it.

Even if a transaction violates the antitrust laws, it can nonetheless be approved by the Department of Transportation, but only if DOT finds that the transaction is necessary to achieve public benefits that clearly outweigh the competitive harm, and there is no less anticompetitive alternative for achieving those benefits. An open skies bilateral treaty might provide countervailing benefits to competition in some U.S.-U.K. markets. The Department of Justice has concluded, however, that the potential benefits of open skies are not sufficient to outweigh the harm of the Alliance as it is currently proposed, in large part because slot constraints at LHR create grave doubts that open skies alone will produce significant new entry and competition in U.S.-London markets. However, with a restructured transaction (to "carve out" operations on certain hub-hub routes) and with conditions designed to ensure access to LHR that would permit the level of airline service between LHR and the United States that would be expected in an open market, DOT may have a basis to conclude that the transaction is, on balance, in the public interest. The Department of Justice has concluded that this standard would be met if LHR slots and facilities sufficient to operate approximately 24 daily round trip flights were made available, through divestiture or some other means, in order for the public interest and competitive concerns to be satisfied.

I. Introduction

The proposed American/British Airways Alliance is the latest in a series of alliances formed between major U.S. airlines and foreign flag carriers. As both DOJ and DOT have recognized, such alliances can result in important benefits to consumers by linking networks to create new on-line connections to international destinations. In addition, open skies agreements reached in conjunction with the formation of alliances have allowed new entry and given airlines increased pricing freedom, all to the ultimate benefit of the traveling public.

There are some important differences between the AA/BA Alliance and earlier alliances reviewed by DOJ. First, the competitive losses threatened by the transaction affect a far larger number of passengers than were affected by any of the other alliances. American and British Airways are the two largest competitors in the United States' largest transatlantic air transportation market. They compete head-to-head in six London city pairs with large volumes of nonstop local traffic.(1) The Alliance is also likely to reduce competition for many passengers who currently benefit from one-stop and connecting competition between AA and BA.

Second, the potential consumer benefits from this Alliance are more limited than those associated with the prior alliances. The number of U.S. passengers likely to benefit from a new "online" service option where none existed before is smaller than was the case for other transatlantic alliances because many passengers already can fly online using either AA's or BA's existing networks. BA already serves 19 U.S. cities on its own, and proposes to add service to Denver. Those cities are the origin or destination point for a large percentage of the passengers who travel from the U.S. to the U.K. and beyond.(2) American also provides extensive service from U.S. points to a number of destinations beyond London. Thus, the number of passengers who might benefit from new service options is smaller than the number of passengers who are threatened by potential losses in competition on overlapping routes.

The anticompetitive effects of this transaction cannot be entirely remedied by conditions. The most obvious condition -- divestiture of slots and facilities at Heathrow -- will not likely result in entry in ORD-LHR or DFW-LON by new carriers capable of replicating the competition between American and BA. Moreover, while entry by a variety of carriers might be expected in the other markets where the two carriers compete on a nonstop basis (assuming slots are available), it is not clear that such a patchwork of replacement carriers will replicate the rivalry that currently exists between AA and BA across a large number of U.S.-U.K. city pairs.

Given the likelihood of competitive harm that cannot be completely remedied, DOT has two alternatives: (1) outright disapproval of the transaction, or (2) approval of the transaction on public interest grounds based on a finding that the transaction is needed to achieve consumer benefits in some markets (primarily as a result of open skies) that clearly outweigh the harm in other markets.

Before approving on public interest grounds, DOT should impose conditions designed to reduce the competitive harm in adversely affected markets. DOT should order slot divestitures where entry is likely and require restructuring of the transaction where slot divestitures alone are unlikely to facilitate sufficient post-alliance entry. Additionally, to ensure that the competitive benefits that open skies normally provide are achieved in U.S.-U.K. markets, DOT should make certain that sufficient Heathrow slots are available to allow for new entry in city-pairs where competition currently is artificially limited by the restrictive bilateral. New entry and competition on these city-pairs is a clear benefit to the public and, if widespread enough, could support a finding by DOT that the public interest is, on balance, enhanced by conditional approval, notwithstanding the unavoidable diminution of competition in some city pairs.

II. Analytical Standards

Under the statute, the Secretary must disapprove an application if the proposed agreement "substantially reduces or eliminates competition," unless the agreement is "necessary to meet a serious transportation need or to achieve important benefits" and there is no less anticompetitive alternative. 49 U.S.C. § 41309(b). The Secretary has broad authority to fashion conditions to remedy the anticompetitive effects of an agreement.

The appropriate framework for assessing the competitive effects of an alliance agreement such as this one is merger analysis. See United States v. Penn-Olin Chemical Co., 378 U.S. 158, 170 (1964). While the joint venture at issue here is not a merger, its effects are similar. Within the markets where the joint venture operates, it contemplates full integration between the two carriers. The joint venture will eliminate competition between the two carriers, consolidate their transatlantic operations, and create some efficiency gains from consolidated operations.

Our analysis of such transactions focuses solely on their likely effects in properly defined relevant markets. The central question is whether the proposed transaction is likely to result in increased prices to consumers by creating, enhancing, or facilitating the exercise of market power by the joint venture partners in any relevant markets.

In analyzing the likely competitive effects of such transactions, the Department of Justice uses the principles contained in the 1992 Horizontal Merger Guidelines. The Guidelines are widely accepted by economists and the courts as an analytical approach to assessing the competitive effects of a transaction. The starting point for that analysis is identifying the relevant markets in which the merging firms compete, identifying the firms that compete in those markets, and measuring concentration.

The relevant markets in airline mergers are at least as small as city pairs. However, as discussed in more detail below, there are often even smaller markets. For example, even between the same two cities, a significant number of travelers who are time-sensitive may not consider one-stop service to be a reasonable substitute for nonstop service.(3) In such cases, it is necessary to assess separately the impact of the transaction on time-sensitive passengers that require nonstop service.

Where a transaction results in high concentration in the relevant market, the Department of Justice proceeds to examine other factors to determine whether the remaining competitors will find it profitable to increase prices or reduce the quality of the service they provide. One factor DOJ considers is whether entry into the market is so easy that market participants, after the transaction, could not profitably maintain a price increase over pre-merger levels. It is not enough that entry be possible, that is, that others "can" enter -- rather, the relevant inquiry is whether entry would occur if there were a small but significant anticompetitive price increase. "Entry is that easy if entry would be timely, likely, and sufficient in its magnitude, character and scope to deter or counteract the competitive effects of concern." Merger Guidelines § 3.0.

Our competitive analysis assumes that bilateral restrictions on pricing and entry are removed concurrently with approval of the Alliance.

III. The Alliance Would Substantially Reduce Competition


  1. Current Competition Between AA and BA

American and British Airways are significant competitors in major U.S.-U.K. markets, as Table 1 sets forth. A combined AA/BA would be the dominant carrier on the six nonstop city-pairs where they currently compete (New York, Boston, Miami, Chicago, Los Angeles and Dallas to/from London). More than half of all U.S.-London traffic travels in these six gateway-gateway markets.

In addition to nonstop competition, American and British Airways, as the two carriers with the most U.S.-U.K. frequencies and the largest number of gateways, carry on a broad-based rivalry across a large number of U.S.-U.K. city pairs, competing in almost half of the largest U.S.-U.K. city pairs.(4) Because of their frequency advantage over other carriers and the number of gateways they serve, AA and BA are able to capture large shares of this connecting traffic. This is particularly true in the Northeast, where their strong positions at New York, Boston, and Philadelphia give AA and BA significant advantages in competing for connecting traffic.


  1. Relevant Markets

There are two separate relevant product markets in each city pair between the U.S. and the U.K.: a time-sensitive (or "business") market, and a leisure market that is not time sensitive. Business travelers will tend to choose nonstop service when it is available, while leisure travelers will consider one-stop alternatives. In addition to distinguishing between types of passengers based on their demand for nonstop and connecting service, market definition on U.S.-London routes also depends upon which London airport passengers prefer -- time sensitive business travelers overwhelmingly prefer Heathrow to Gatwick.

    1. Nonstop Service is a Relevant Market

Whether nonstop service is a separate market from one-stop service depends on whether one-stop service can discipline a supracompetitive price increase for nonstop service. Could a monopolist of nonstop service (but not all service) profitably impose a price increase on an identifiable group of buyers? Evidence shows that for an identifiable group of time-sensitive business passengers, one-stop service is not a reasonable substitute for nonstop service; they would not switch to one-stop service in response to a price increase in nonstop service. The airlines can and do charge these travelers different prices than leisure travelers, targeting time-sensitive passengers with fare restrictions and conditions. In addition, the practice of negotiating fares directly with large corporate purchasers is becoming increasingly widespread among airlines. This practice allows airlines to price discriminate even more precisely among purchasers with varying degrees of price and time preferences. Allowing the Alliance to eliminate competition for business passengers between American and British Airways threatens these passengers with the possibility of anticompetitive price increases.

The existence of a separate market for nonstop service is supported by data on U.S.-London service, by recent econometrics work that shows that time-sensitive, business type consumers have a strong preference for nonstop versus one-stop travel,(5) and by evidence regarding corporate travel policies.

The preference of business travelers for nonstop service is reflected in the total number of passengers choosing one-stop levels of service on routes where nonstop service is available. As the fare analyses in Appendix A show, average fares for nonstop service are significantly above average connecting fares. These average fare differentials indicate that higher fare (i.e., time-sensitive business) passengers prefer nonstop service, and thus strongly suggest that nonstop and connecting service are in different markets.

The value that business passengers place on their time would also make them unlikely to switch to onestop service in response to an increase in nonstop prices. If the competitive nonstop fare is $500, a 5% increase would amount to an extra $25 per one-way ticket. Assuming that a transfer caused a two hour delay,(6) any business that valued its executives' time at more than $12.50 per hour (i.e., almost any business) would pay the higher price for nonstop service rather than require a connection. Available evidence on corporate travel policies also supports the conclusion that increases in nonstop prices would not lead to substantial shifts of business passengers to nonstop service. According to the 1996 American Express Survey of Business Travel Management, 78% of all companies have a policy in place requiring employees to use the lowest "logical" airfare available, but only 25% of those corporations require the employees to use connecting flights to achieve the lowest fare.(7)

In Order 96-6-33, DOT concluded that connecting and nonstop service were in the same market. Id. at 13. Two of the factors relied on in that order, however, frequent flyer affiliation and the availability of negotiated discount fares, will tend to drive business passengers to choose the nonstop carrier in most cases where there is a hub at one or both endpoints. The most attractive frequent flyer program and corporate discounts will be offered by the hub carrier which, in turn, is likely to be the nonstop carrier. It is well recognized that an airline's frequent flyer program becomes more attractive as its presence at a city increases.(8) The effect of negotiated discount fares is similar. A significant and growing portion of business traffic travels subject to contract fares negotiated between the airline and a corporation's travel manager. Most travel managers appear to favor airlines with extensive route networks that are easily accessible from major company locations. Airlines that do not offer attractive route networks with adequate nonstop service will not be able to obtain a corporate contract even if they offer a higher discount level. Airlines with more extensive route networks also have an advantage over smaller airlines with less extensive networks at a given city because carriers often make any corporate deal contingent upon the corporation agreeing to meet target market shares in multiple city pair markets.

Hence, frequent flyer preferences and corporate discount programs will tend to reinforce business passengers' preference for the nonstop hub carrier in any given city pair market. For example, for business passengers originating in Dallas, American's hub strength gives it by far the most attractive frequent flyer program. Moreover, Dallas-based corporations with extensive travel requirements will prefer to negotiate with American for a corporate travel contract since American is the airline best positioned to serve their travelers' needs. American can and does reinforce that preference by requiring corporations to use American for all their travel needs as a condition to granting discounts. If an airline such as United wished to attract Dallas-based business passengers away from American's nonstop DFW-LON service onto its DFW-ORD-LON connecting service, United's frequent flyer program would be of little help, and United's corporate travel discounts would have to be significantly larger than American's.

Order 96-6-33 also suggested that the elapsed time advantage of nonstop service becomes less significant as the length of the journey increases. However, longer elapsed time is not the only drawback that frequent travellers associate with connecting service. The uncertainly and inconvenience involved in changing planes and, for westbound flights connecting to U.S. gateways, the need to retrieve baggage and clear customs in mid-journey are among the other difficulties. Moreover, with longer journeys, especially international journeys, employers become more willing to pay substantial premiums to minimize business travellers' inconvenience. Indeed, many corporations allow executives on international flights to upgrade to business or first class, despite the huge increase in expense.(9) Service factors (especially frequency and schedule convenience) and frequent flyer programs appear to be far more significant than price in determining what carrier a business traveler will use.(10) Thus, the available evidence strongly suggests that most time-sensitive passengers would not switch to one-stop service in the event of a price increase in nonstop service.(11)

    1. Service to Heathrow is a Relevant Market

London is served by two primary airports: Heathrow (LHR) and Gatwick (LGW). The current bilateral limits the number of U.S. carriers that can serve Heathrow to two (American and United) and limits the U.S. gateways that may be served from Heathrow. Most smaller U.S. gateways are served exclusively to Gatwick. The available evidence suggests that time-sensitive business travelers overwhelmingly prefer Heathrow to Gatwick and would not switch to Gatwick to avoid a price increase.

One reason that Heathrow service is in a separate market for business travelers is geography. Heathrow is much closer to the central London business district (14 miles) than is Gatwick (25 miles). According to the Civil Aviation Authority, the average drive time from Heathrow to central London is 38 minutes, compared to 64 minutes from Gatwick.(12) In addition, many businesses have built offices in the London suburbs near Heathrow, in part to take advantage of the airport's proximity. As a result, Heathrow is much more conveniently located for many business travelers, and has a much larger catchment area than Gatwick.

The preference of business passengers for Heathrow translates into significantly higher yields for Heathrow service. In comparing its own routes at Gatwick relative to Heathrow, BA estimates that its yields at Gatwick are substantially less than those obtainable at Heathrow (CA p. 14, BAX 228515). Another internal BA memo examining the effect of moving service from Heathrow to Gatwick concluded that the impact of shifting flights from Heathrow to Gatwick depends on other competition on the city-pair.

(CA pp. 15-16, BAX 2227-2228).

American has also found a strong preference for Heathrow among business passengers. Indeed, the preference for Heathrow remains strong even where the passengers are traveling on connecting routings from interior points. (CA p. 17, AA 14810).

    1. Service to JFK Airport May Be a Relevant Market

The available evidence on relative prices from the New York area's two international airports, JFK and Newark, is not as clear. Many business passengers appear to prefer JFK, but average prices from those two airports to major international destinations do not exhibit as clear a divergence as do prices to Heathrow and Gatwick. For corporate purchasers of travel from New York, prices and service from one New York area airport may not discipline prices and service from the other. Such corporations will vary in the degree to which they view service from JFK and Newark as substitutes. Some corporations' employees may have a strong preference for JFK because they tend to live near that airport. Others' may lean toward Newark for analogous reasons. Still others may have employees with mixed preferences and so would prefer carriers that can offer service from both airports. Airlines can track large corporations' travel patterns and requirements by using CRS data. The airlines vary the discounts they are willing to offer each corporation depending on the corporation's needs and bargaining position.(13) Because of this ability to price discriminate, Newark appears to be an imperfect substitute for JFK for service to the U.K.


  1. Competitive Effects

A combination between AA and BA would have significant effects. In the six nonstop overlap markets, it would eliminate all nonstop competition in DFW-LON and BOS-LHR, reduce the number of competitors from three to two in ORD-LHR and MIA-LHR, and reduce the number of competitors from four to three in NYC-LHR and LAX-LHR. As noted, more passengers traveled in those markets combined than between the U.S. and any other European country.

American's aggressive growth since it acquired TWA's route rights in 1991 has made it BA's primary rival for most of its U.S.-U.K. service.

(14) It increased its JFK-LHR flights from 3 daily round trips in the summer of 1992 to as many as 7 in the summer of 1995. BA

(15) responded with significant frequency and capacity increases of its own. Thus, current frequency and capacity levels are largely a result of the AA-BA rivalry. AA and BA have also driven one another to increase corporate discounts and travel agent commissions.(16)

In markets such as these, price is not the only factor, or even the most significant factor, valued by time sensitive passengers. To that extent, the competitive effects of this combination should be analyzed in the context of differentiated products.(17)

In the New York market, a combination of American and BA would not only combine carriers with the lion's share of the frequency and traffic in the market, but would also combine two carriers with strong brand identities, the most attractive frequent flyer programs,(18) greatest level of frequency and, because of BA's reservoir of Heathrow slots, the greatest ability to add frequency. All of these factors have a significant effect on the brand preferences of time-sensitive passengers, and make the prospects for new entry to replicate the existing level of interbrand competition more difficult.

The competitive losses in connecting markets will be significant as well. As might be expected with a merger of the two most extensive U.S.-U.K. route networks, the Alliance could eliminate a significant competitor from a large number of U.S.-U.K. markets. Most of these markets have two to five other competitors, and the competitive loss in many of them is not great. However, in the aggregate, the losses are likely significant enough that they should not be ignored.


  1. Even With the Removal of Bilateral Restrictions on Entry, Significant Entry Barriers Will Remain

"A merger is not likely to create or enhance market power or to facilitate its exercise, if entry into the market is so easy that market participants, after the merger, ... could not profitably maintain a price increase above pre-merger levels." Merger Guidelines, §3.0. Such entry must be timely, likely and sufficient in magnitude, character and scope to counteract post-merger price increases.

As outlined above, the competitive losses accompanying approval of this Alliance would be substantial and immediate. The two carriers are the two largest competitors in each of the nonstop city pairs or airport pairs in which they overlap. Overall, they account for nearly sixty percent of all U.S.-U.K. service. To preserve competition, entry must be timely and substantial. It will not be timely and substantial in the circumstances here.

    1. Entry into Heathrow Airport is Severely Constrained

Heathrow is one of the busiest and most congested airports in the world. It handles over 58 million passengers annually, and operates at or near capacity throughout the day. Heathrow operates at capacity during all peak hours, and the demand for slots by airlines currently authorized to serve it far exceeds the supply of slots.(19) Heathrow is constrained both on the airside and on the groundside. Its runway capacity is constrained by the need to operate in a manner that minimizes noise for surrounding communities. On the ground, both terminal facilities and aircraft parking spaces are in very limited supply during peak periods. These constraints make it highly unlikely that the airport will be able within a reasonably short time frame to accommodate both the capacity demands that would be made on it as a result of an open U.S.-U.K. bilateral as well as the requirements of carriers responding to exercises of market power made possible by an AA-BA Alliance. See Answer of BAA PLC, May 7, 1998.

The U.K. Government has taken the position during the course of negotiations over the proposed open skies agreement that it has no ability to facilitate the creation or acquisition of slots to permit the growth of U.S.-LHR services.(20) Thus, the only available sources for slots for new U.S.-U.K. services are airport capacity growth and acquisition from carriers already serving LHR. These two sources are likely to be inadequate to support "timely, likely and sufficient" entry in the AA-BA markets after approval of the Alliance.

Prospects for Creation of New Slots are Limited. British Airports Authority plc ("BAA"), the airport operator for both LHR and LGW, estimated in 1996 that an average of one new arrival or departure slot per peak hour would become available at LHR each year during 1997 and 1998.(21) BAA defines the peak period as the 14-hour period from 6 a.m. to 8 p.m., so its estimated growth would yield 7 new daily round trip services each year. Accepting BAA's estimate, that level of added capacity would be far from adequate to satisfy the demand that can be anticipated after open skies and approval of the AA-BA Alliance. First, not all of the new slots will be appropriate for U.S.-U.K. service.(22) Second, even for slots that fall within desirable time periods, there is no assurance that they will be allocated to U.S.-U.K. carriers, let alone to carriers who would use them in markets where competition would be lost as a result of approval of the Alliance. The applicable slot allocation rules do provide some preference for "new entrants,"(23) but markets other than U.S.-U.K. markets will also have carriers with new entrant status applying for slots. Moreover, new entrant status provides no assurance that any slots reserved or allocated for that group will be commercially usable. The U.K. Civil Aviation Authority's study of slot allocation for the summer 1995 season(24) is illustrative. It concluded that, while many "new entrant" slot applicants received some allocation of slots, most of the slots they were allocated were far inferior to those requested and many were at unusable times. Id. at 10. It appears that, over time, new entrants such as Canadian Airlines International and Virgin Atlantic have been able to assemble commercially viable slot holdings. However, even assuming that such growth would continue to be possible at LHR, that kind of slow, long term, incremental growth is not the kind of timely entry that can be relied upon to discipline supra-competitive prices in key U.S.-LHR markets.

Transfers From Current Slot Holders Can Not Be Relied Upon to Provide Sufficient Slots. AA and BA argue that there is a thriving market for the purchase and sale of slots at Heathrow, and that carriers wishing to enter U.S.-LHR markets will be able easily to purchase such slots or acquire them from their alliance partners who already operate at LHR.(25) The evidence does not support their assertion.

Although slot transfers have occurred at Heathrow, the transfers do not demonstrate that a de facto buy/sell market has emerged under current British interpretations of E.C. slot rules. Moreover, whether current holders would be willing to sell to U.S. carriers at prices equivalent to those that would prevail in an open and competitive slot market is doubtful. Certainly, AA and BA (which hold about 42% of all LHR slots) will be unwilling to sell any slots to potential competitors at prices the competitor would be willing to pay. AA and BA know that any U.S. carrier would use a slot in competition with them on some U.S. route. As long as AA and BA can use the slot themselves, there is no possibility that a U.S. carrier can buy a slot from AA/BA. In essence, the U.S. carrier would not only have to pay AA/BA the scarcity rent (or opportunity cost) of the slot, but also the expected loss to AA and BA from additional competition, which would make buying a slot a poor investment for a U.S. entrant.

While other national flag carriers at Heathrow might be more willing to consider sales to U.S. carriers, there are several obstacles to such sales. First, most of those carriers would not be willing to trade slots that form an integral part of their own hub and spoke networks. Even if a carrier has some "less essential" services that it could conceivably move to Gatwick while retaining others at Heathrow, a carrier that currently operates only at Heathrow would substantially increase its costs and the complexity of its operations by moving a few flights to Gatwick.(26) Moreover, their ability to obtain slots at Gatwick is not at all clear. Slot availability has become much more constrained at Gatwick since AA and BA announced their Alliance, and recent entrants, including USAirways, have had difficulty obtaining needed slots.

The second obstacle to effective buy/sell implementation is that existing terminal and parking restrictions may also impede slot transfers to more efficient users. Some of the proposed "solutions" to slot constraints, therefore, are unlikely to provide significant sources for slots. For example, AA and BA have suggested that daily turboprop service by KLM from LHR to Rotterdam would qualify as sufficiently low value to be worth selling to U.S. carriers. Although AA and BA probably are correct about the value of the service, it is not at all clear whether or how quickly the BAA would be able to find the necessary stands and terminal space to meet requirements for a Boeing 747 if a slot is currently being used by a turboprop operator.

BAA's comments in this proceeding make it clear that such constraints could prevent or delay significant slot transfers. Answer of BAA. The OFT proposal contemplated the divestiture of slots sufficient to operate 12 daily round trip flights. The transfer of slots from the AA/BA Alliance to other carriers would be phased in over three slot scheduling seasons. BAA's comments make it clear that it would have great difficulty accommodating transfers at a rate faster than 4 per six-month period. Moreover, in making that assessment, it has indicated that the new services might have to be scattered among several terminals, and implied that it would be very difficult for the airport to accommodate a more ambitious or more accelerated conversion of small aircraft services to large ones. As discussed further below, 12 new daily round trips introduced over three scheduling seasons would not allow for the timely and sufficient entry needed to prevent priceincreases.

    1. Additional Barriers Will Deter Entry Into Some U.S.-U.K. Markets

All of the nonstop overlaps between American and BA have alliance hubs at one or both endpoints. It has been well documented that nonstop entry against a dominant hub carrier is difficult, and that a hub carrier facing no nonstop competitors can raise prices to local passengers in routes emanating from the hub.(27) The ability of hub carriers to "dominate" service and pricing in city pairs originating at their hubs stems from three factors.

First, the hub enables the hub carrier to transport substantial numbers of connecting passengers in the city pair in addition to local passengers. For example, on its DFW-LGW service, American carries more London passengers traveling to and from U.S. points behind Dallas than it does local Dallas-London passengers. The ability to carry connecting or feed traffic permits carriers to operate much more capacity and frequency on a route than the local market would otherwise merit. This added frequency can tend to block carriers without access to their own sources of feed traffic from entering to serve the local market, since the market already has more flights than the local market would support on its own. The availability of this "extra" capacity does not necessarily result in low prices for local passengers in hub markets, however. The hub carrier can employ pricing and yield management techniques to restrict the number of seats available to local passengers when it is not faced with competition. When faced with entry, however, the hub carrier can instantly increase the number of seats available in the local market by changing its yield management allocations.

The second deterrent to entry in hub markets stems from the observed relationship between an airline's scale at a city (and in a city pair) and its share of passengers, variously referred to as the "s-curve," or "origin point presence" effect. Passengers appear to disproportionately prefer airlines with the most service to and from local airports. This preference is reinforced by airline marketing and sales tools, such as frequent flyer programs, travel agent commission overrides, and corporate discount programs, which are structured by the airlines to better reward passengers and travel agents as the airline's share of local service increases. Airlines recognize that they will tend to obtain a disproportionately large share of passenger revenue originating at points where they have a large presence.(28) Some carriers quantify the effect of the s-curve and factor it into their route entry decisions.(29) Others factor it into their strategic resource allocation decisions, and tend to withdraw from cities and regions where their presence is weak and build where their presence is strong relative to other airlines. A significant contributor to the financial benefits the parties expect from this transaction is caused by this "presence" effect.(30)

Third, the existence of substantial connecting traffic on the route and the local market advantage conferred by origin point presence enable the hub carrier to respond strategically to new service aimed at the local market. To stimulate new demand, entrants often cut prices in the local market. Without feed, however, an entrant cannot attract connecting traffic to help fill its plane and cover its costs. The incumbent airline, with its established feed traffic and (generally) higher frequency, can often respond by matching the entrant's prices and still operate profitably on the route. If a simple price match is insufficient to deprive an entrant of enough local traffic to survive, the hub carrier can shift enough capacity to the local markets to accomplish that goal. Moreover, where service in a market is constrained by slot availability, a hub carrier with access to a large pool of slots has even greater ability to respond to entry in this way because the entrant will be unable to add capacity on its own. American's president has referred to such strategic responses as "predatory scheduling."(31) The net result of "predatory scheduling" is to discourage new entry in the first place, or to render it unprofitable where it occurs.

In Order 96-6-33, DOT expressed faith that entry would occur in markets that had New York as an endpoint. It also expressed the belief that low cost carriers could succeed in the international marketplace. Order 96-6-33 at 15. DOT's own recent experience in investigating possible exclusionary conduct by airlines in domestic markets, as well as recent carrier exits from transatlantic markets affected by immunized alliances, should give cause to reassess those conclusions.

It is clear that in a number of cases, hub carriers have shown the inclination and ability to respond to entry with strategies that have rendered the entry non-viable.(32) Even if such strategies are lawful, the reality that hub carriers can and will use them successfully must be factored in to an analysis of the likelihood of entry.

Whatever the likelihood of effective entry on a hub carrier's spoke routes, that likelihood is further reduced by alliances where the partners have hubs at one or both endpoints. The same reasons that drive carriers to enter an alliance to gain traffic (increased revenue from code sharing, presence, and combined frequent flyer program strength) will tend to deprive less-favored entrants of traffic for the same reasons that the alliance gains traffic. Indeed, even a carrier with significant feed or presence at one endpoint may not be able to maintain viable service against a rival with hubs at both endpoints. The effect of combining the strengths of two carriers, each with a hub on one side of the Atlantic, is evident in recent decisions by some U.S. carriers to withdraw from transatlantic routes. In New York-Zurich and New York-Brussels, American Airlines, despite its substantial presence in New York, no longer provides nonstop service in competition with the Delta/Swissair/Sabena Alliance recently immunized by the DOT. A similar withdrawal also occurred between Washington and Frankfurt, where Delta (despite its Frankfurt hub) no longer competes with nonstop service against the United/Lufthansa Alliance.(33) This experience suggests that entry on routes dominated by American and British Airways will be difficult, and in some cases, existing service may be cut back by U.S. carriers unable to compete effectively against an AA/BA Alliance. The only airlines that should be seriously considered as likely entrants in these routes are those with significant operations at one or both endpoints, or other marketing or operational strengths that can act as counterweight to the powerful origin point advantages that will be wielded by the AA/BA Alliance.


  1. Slot Availability Alone Will Not Eliminate All of the Anticompetitive Effects of the Transaction

The most obvious entry barrier, Heathrow slot constraints, can be addressed through conditions requiring that slots be made available to entrants. However, slots alone will not fix all the competitive losses.

    1. Entry is Unlikely in Hub-Hub Markets

In the ORD-LHR and DFW-LGW markets, there is very little likelihood that any new entry would occur after American and BA join forces. At O'Hare, United is the only other airline positioned to serve the market, and it is already present.(34) British Midland, which has a relatively small connecting complex at Heathrow, has publicly expressed interest in serving a number of U.S.-Heathrow markets, including Chicago. However, it does not appear that British Midland has evaluated in detail the likely profitability of any of the routes it listed, and thus its ability to survive in competition against carriers operating three of the largest hub complexes in the world is subject to serious doubt. Nonetheless, expansion by United in the Chicago market would benefit consumers, although not as much as entry by a new carrier. Slot divestitures are necessary to permit such expansion, but would still fall short of a remedy that would preserve the pre-transaction status quo -- competition among three independent carriers.

On the DFW-London route, Delta's small and shrinking connecting complex at DFW would likely be insufficient to support entry against a combined entity with large hubs at both endpoints of the route. Thus, the proposed Alliance will be a merger to monopoly for time-sensitive DFW-London passengers.

The MIA-LHR market is one with Alliance hubs at both endpoints. However, several factors lead to the conclusion that there is at least one plausible entry candidate into that market. First, according to information provided by the parties, the Miami market is more leisure-oriented than the other nonstop markets affected by this transaction. There is evidence that the origin point presence effect that normally favors hub carriers is not as strong among travelers in leisure-oriented markets as in business-oriented in markets.(35) Second, it appears that a large proportion of the traffic in that market originates at the European end. Those two factors may make it possible for a carrier such as British Midland to viably to enter that route, assuming of course that slots are available.(36)

    1. Entry Sufficient to Replace the Competition Between American and BA In The New York-London Market Will Be Difficult

American currently operates 6 daily nonstops and BA operates 5 daily subsonic nonstops between JFK and LHR. The two carriers also operate a total of 3 daily round trips between Newark and LHR. The two airlines are far larger than the other two competitors on the route, and they compete intensely for the business travel market, for which they are both very well positioned. British Airways' LHR hub and dominant presence in London obviously place it in a strong position to capture European originating traffic. The LHR hub can also offer New York originating passengers service to a vast array of points beyond London, in addition to the most frequent service to London. Similarly, American can offer New York business travelers the largest frequent flyer program, as well as service in many markets that are important to business travelers, including Chicago, Los Angeles, San Francisco, Miami, Latin America and numerous European destinations. The loss of competition between closely matched and well positioned competitors will be difficult to replace.

Business travel accounts for a large portion of this market. In the third quarter of 1995, there were approximately 450,000 local passengers traveling between New York and London. Approximately of the passengers traveling on American and BA were paying $800 or more for a round trip journey, and those higher fares accounted for approximately of AA's and BA's revenue generated on that route. American and British Airways were particularly successful in attracting this group of passengers; while AA and BA account for approximately of all the seats sold to local passengers traveling between New York and London, AA and BA together earned approximately of the revenue generated by these local passengers during the third quarter of 1995.(37)

For business travelers, factors other than price often determine their purchase decision. Schedule convenience, particularly frequency, is one of the most important of these factors.

(38) In the U.S., airlines have found that hourly shuttle-type service is very successful in business markets. While shuttle-type service would be less feasible and advantageous in transatlantic markets because of distance and time differences, the convenience of frequent, well-timed schedules is nonetheless vital for serving much of the New York-LHR business market.(39) Moreover, as discussed above, there is evidence that the level of frequency required to provide meaningful competition is relative. Carriers with a dominant share of the frequency in a market tend to attract a disproportionate share of the high-yield business traffic in the market.(40) This evidence suggests that a minimum of three to four daily round trips would be necessary in New York-LHR to make an airline a viable and effective competitor against AA/BA for business travelers.

The combination of a large BA hub at one endpoint of the route, the Alliance partners' longstanding position in the market, and the large frequency advantage they are likely to enjoy(41) will render entry on a meaningful scale by carriers positioned to compete effectively with the Alliance for business traffic in the New York-LHR market difficult. The number of airlines positioned as likely entrants is small. Delta, with a fairly large transatlantic operation at New York, and the "presence" it enjoys as a result of its operation of the Delta Shuttle has the ability to attempt a competitive offering. Continental's Newark hub might enable it to mount frequent service to LHR from Newark, which would provide some competitive discipline to JFK-LHR services. British Midland's LHR connecting service might also support entry on a small scale. However, whether such entry can take place, and whether it can replicate the current rivalry between AA and BA, is not clear. At a minimum, building a comparable schedule and marketing presence will take considerable time(42), and until a comparable presence is achieved, the competitive constraint provided by a new entrant may not be as effective as the constraints that American and BA currently provide to each other.

IV. Possible Remedies and Public Interest Conditions

As discussed above, the Alliance will significantly reduce (and in some cases completely eliminate) competition in a number of important U.S.-U.K. city pairs. While some of this competitive harm can be remedied by providing other carriers with the slots and facilities necessary to enter, it is not possible to fully replace the competition lost through the combination of the two largest carriers in the United States' largest transatlantic air transportation market. A patchwork of substitute carriers can be put together to operate flights in most of the nonstop gateway-gateway markets in which American currently operates, but there is simply no way to completely replicate the competitive significance of the American network. Furthermore, the efficiencies created by the transaction are unlikely to outweigh the serious competitive harm. Under a strict antitrust analysis the Department would therefore oppose the transaction.

The Department recognizes, however, that DOT can approve an application if it determines that the transaction is necessary to advance important public benefits, and that, on balance, the transaction advances the public interest. Using this standard, DOT could determine the transaction is in the public interest if conditions were imposed that (1) minimized the anticompetitive effects of the Alliance in the overlap markets, and (2) provided sufficient additional public benefits in the form of increased competition in U.S.-U.K. air transportation markets to outweigh the competitive harms of the transaction. We estimate that LHR slots sufficient to operate roughly 24 additional daily round trips would have to be made available as a result of the transaction and open skies in order to tip the balance in favor of approval.


  1. Minimizing Anticompetitive Effects in Overlap Markets

Entry in Overlap Markets. In some of the overlap city pairs affected by the Alliance, new entry will likely occur following open skies provided that entrants have access to well-timed slots (and related airport facilities) at LHR. The availability of slots sufficient to operate at least 14 slots is necessary to minimize the anticompetitive effects of the Alliance in overlap city pairs.

Four of the round trips are needed (and sufficient) to replace existing competition in BOS-LHR (2), MIA-LHR (1), and LAX-LHR (1). In each of these city pairs it is likely that new competitors would enter the London markets if slots and facilities were available. USAirways, for example, has a significant presence at Boston, and has expressed an intent to enter that market after open skies.(43) British Midland has also expressed interest in entering some of these markets, which do not appear to be as feed-dependent as LHR-ORD.

Slots sufficient for two round trips are needed to reduce the harm in ORD-LHR by creating expansion opportunities for United -- the other incumbent and the only well-positioned competitor. Two round trips would replace the current schedule of the smaller of the two Alliance carriers. It should be noted, however, that this will not replace the competition lost in the Chicago market as United is not a new competitor, and the number of carriers in the market will be reduced from three to two.

Finally, at least 8 round trips are needed for NYC-LHR. Currently American operates seven daily round trips and BA operates six (plus two daily Concorde round trips) in that market. It appears highly unlikely that any other carrier would immediately enter at that level of service, as no potential entrant has the established marketing presence and brand identity of AA or BA in the New York-LHR market. If only six slots were made available in the NYC-LHR market, the likely result is that they would be divided between two or more carriers, none of which would come close to matching the combined service of AA/BA. Moreover, one of the most likely entrants, Continental, would likely operate its service from Newark, which is an imperfect substitute for JFK. Under these conditions, there is a danger that simply spinning off the slots operated by the smaller of the two Alliance partners would leave AA/BA significantly less constrained in that market than they are today when the two carriers, each with similar levels of service, compete vigorously. In order to minimize the harm, the total number of slots available in New York should be at least 8 daily slot pairs -- enough to fund well-timed service by two new carriers at a level of four round trips each.

While providing LHR access to entrants would thus partially remedy the anticompetitive effects of the transaction, significant harm would remain unameliorated: (1) No carrier is likely to enter DFW-London, a hub-hub route where AA and BA are the only competitors; (2) The number of competitors on the Chicago-LHR route will be reduced from three to two; and (3) Current connecting competition in markets that the two carriers' networks are well-placed to serve will not be replaced.

Slot Allocation. The exact number of slots that would be required to achieve sufficient new entry in the overlap markets depends on the mechanism that will be used to distribute them. DOJ generally favors market-oriented solutions that allow would-be entrants to purchase assets with few strings attached, as the marketplace generally does a better job than regulators do of selecting the most efficient competitors. Because this approach does not ensure that the assets will be used in the markets most directly harmed by the transaction, more slots may have to be made available to increase the likelihood that the affected markets receive service. If, on the other hand, a carrier selection process were in place that awarded the assets to designated carriers for designated uses, relatively fewer slots would be needed. The discussion above assumes the latter -- that all of the slots are used to provide new service for the adversely affected city pairs.

American and BA argue that LHR slots for entrants are available from a number of sources -- slot allocation to new entrants, trading with incumbents, alliance partners -- and that AA/BA need not divest all of the slots required to remedy the anticompetitive effects of this transaction. While DOJ is far less confident that entrants will be able to obtain the slots necessary to compete with the Alliance from other sources, we agree that from a competitive standpoint the source of the slots is irrelevant. In structuring the conditions to be placed on the Alliance, DOT should ensure that any risk that slots are not available elsewhere should be borne by AA/BA rather than by consumers. For example, it may be appropriate to "credit" slots obtained through the new entrant procedures against the total number of slots necessary to remedy the anticompetitive effects of the Alliance.(44) To the extent slots are not available from other sources, however, DOT must require divestiture by AA/BA as a condition of approval.

Other Conditions. DOT should impose other conditions that might further ameliorate those anticompetitive effects for DFW-London and ORD-London. In other immunity cases, DOJ has advocated and DOT has imposed conditions "carving out" certain hub-hub markets from the joint operations. While such conditions are a second-best alternative to competition between unaffiliated, independent competitors, DOT should not approve the Alliance without similar conditions for DFW-London and ORD-London.


  1. Public Interest Conditions

As discussed above, entry in the overlap markets will only partially remedy the anticompetitive effects of the Alliance. DOT may nonetheless approve the application if it satisfies itself that there are other public interest benefits that outweigh the harm.

The primary benefit claimed for the transaction is open skies. The current U.S.-U.K. bilateral agreement undoubtedly restricts competition and raises prices for consumers in U.S.-U.K. markets. The bilateral limits the number of gateways, competitors and flights between the two countries. The bilateral also limits service to Heathrow to two carriers per country. There would likely be more service (and lower prices) in U.S.-London markets today if the restrictive bilateral had not been in place. In addition, the U.K. government has frequentIy used the fare disapproval authority it has under the bilateral to prevent airlines from charging low fares to interior U.S. passengers for travel via their hubs. Cargo service to and beyond the U.K. has also been restricted. Elimination of those restrictions is a public benefit. If DOT concludes that a grant of immunity to the proposed transaction would be "necessary" to the achievement of open skies, then it would not be inappropriate to "credit" this transaction with the public benefits of open skies under a public interest analysis.

Measuring that benefit is difficult, but DOT should not accept the residual harm unless the countervailing benefits are clear and substantial. In this case, the balance could not tip in favor of the public interest unless open skies were a meaningful benefit to U.S. consumers. This in turn requires the introduction of meaningful levels of additional capacity into U.S.-Heathrow city pairs as a result of the "trade." In other words, consumers would be the losers if the trade were open skies without meaningful prospects for entry at Heathrow in return for a merged AA and BA (even with divestiture of American's Heathrow "plant"). If this transaction is to be viewed as the price for open skies, it is a high price. For many city pairs, this transaction would remove American, the competitor that is best positioned to challenge the largest carrier, and add its marketing strength and market share to the largest carrier. The Alliance would thereby widen the gap between the first and second largest carrier in the market. The risks to competition inherent in such a combination should only be accepted if the benefits of open skies are not illusory.

In that connection, DOT has stated that de facto access to Heathrow for U.S. carriers will be a prerequisite for a grant of immunity.(45) DOT's concern that real prospects for entry and expansion at Heathrow are necessary to make open skies meaningful is appropriate. As discussed above, capacity growth and the operation of slot trading mechanisms are unlikely to provide significant opportunities for carriers to initiate U.S.-U.K. service at Heathrow. Hence, a reservoir of slots additional to those needed to address competitive problems would have to be available (perhaps by further divestiture from the parties) if DOT wants to assure that there will be meaningful access.

Deciding on the number of slots that would be necessary to make access "meaningful," or the number of slots that would be needed to outweigh the residual competitive losses resulting from the transaction, is not a exercise that lends itself to the same precision as an analysis of the competitive harms of the transaction. One approach might seek to ensure that there were enough slots available to permit meaningful and immediate LHR service (2 round trips) from major carriers' large U.S. hubs such as Atlanta, Philadelphia, Houston, Detroit and St. Louis, and would require that 10 slots be made available. It would be reasonable for DOT to conclude that this service would exist in an unrestricted market.(46) Immediate implementation of service to Heathrow from large hubs such as those, in addition to possible Heathrow entry by Continental from its Newark hub would provide clear public benefits by opening up Heathrow access to almost all passengers located at interior U.S. points.

The significant constraint on Heathrow (and Gatwick) slots means that carriers using normal slot allocation procedures will likely have to work patiently over a long time frame to build a portfolio of slots to support service at attractive times. If new Heathrow carriers are positioned to commence meaningful Heathrow service from a number of major U.S. hubs at the outset of the Alliance and open skies, this could provide a reasonable basis for U.S.-U.K. embarking upon an era of competition under open skies. It will enable these carriers to compete effectively with AA/BA, which will enure to the ultimate benefit of U.S. business travelers and other consumers.

CONCLUSION

This transaction will result in significant competitive harm that would be only partially remedied by new entry. Therefore, DOT should only approve the immunity application of the Alliance transaction if (1) it adopts conditions to minimize the harm in adversely affected U.S.-U.K. markets and (2) is convinced that public benefits in the form of increased competition in other U.S.-U.K. markets clearly outweighs the residual competitive harms.

Joel I. Klein
Assistant Attorney General
Antitrust

Constance K. Robinson
Director of Enforcement

Respectfully submitted,

______________________________
Roger W. Fones
Chief

______________________________
Donna N. Kooperstein
Assistant Chief

______________________________
Robert D. Young
Attorney

______________________________
Michael D. Billiel
Attorney
Transportation, Energy and
Agriculture Section

______________________________
William H. Gillespie
Economist
Antitrust Division
U.S. Department of Justice
Washington, DC 20530

Dated: May 21, 1998


APPENDIX A

Nonstop vs. Onestop. Among passengers paying for unrestricted coach tickets (generally between $400 and $900 one-way), there is a strong preference for nonstop service over one-stop alternatives. Table II, CA p.1, shows the number of business passengers in several fare categories that chose nonstop over connecting service in a number of U.S. gateways to London in the third quarter of 1995. Business passengers in these city-pairs consistently choose nonstop service.

The results from three representative markets are graphically represented in Charts 1A - 3B, CA pp. 2-7, which compare average fares for nonstop and connecting traffic for the third quarter of 1995. Chart 1A , CA p. 2, shows AA and BA passengers in the Philadelphia-London city pair. As noted above, BA was the only nonstop competitor and carried most of the passengers in all fare categories. American, the primary one-stop competitor, captured very few passengers in the business range ($400 and up). This is reflected in the difference in average coach fares, shown in Chart 1B.(47) The average BA passenger pays , compared to an average American fare of just . A similar result can be observed in the Miami-London charts, 2A and 2B, CA pp. 4-5. Once again, Chart 2A demonstrates that relatively few time sensitive business passengers chose connecting service over nonstop service in this market. Since AA offered both nonstop and connecting service from Miami, it is possible to make a comparison between the average AA fares for these two alternative services that should exclude factors relating to airline preference. As shown in Chart 2B, nonstop AA coach passengers in the Miami-London market paid an average of , compared to an average connecting fare of .

The third set of charts (3A and 3B, CA pp. 6-7) compares nonstop and connecting fares in the Los Angeles-London market. The percentage of connecting passengers in the $400 and over range is the highest of any major city pair at 23%. Calculation of the average fares in Chart 3B, however, still shows a significant difference between nonstop and connecting fares. On AA, for example, a connecting passenger paid one-way, compared to for a nonstop passenger.(48)


Table I

Direct Overlaps Between American Airlines and British Airways
in Routes from U.S. Gateway Cities to London
U.S. Gateway AA Flights/Day BA Flights/Day Other Airline Flights/Day
Boston 2 3 1 Virgin (LGW)
New York (JFK) 6 5
1 (LGW)
3 United
2 Virgin
New York (EWR) 1 2 1 United
1 Virgin
1 Virgin (LGW)
2 Continental (LGW)
Miami 1
1 (LGW)
1
1(LGW)
1 Virgin
Chicago 4 2 2 United
Dallas 2 (LGW) 1 (LGW) None
Los Angeles 1 2 1 United
2 Virgin

* Based on OAG, May 1998. All flights are to London Heathrow (LHR) unless denoted as flights to London Gatwick (LGW). BA's two daily Concorde Supersonic flights between JFK and LHR are excluded.


FOOTNOTES

1. The six U.S. cities are New York, Chicago, Miami, Boston, Dallas/Ft. Worth and Los Angeles. The six nonstop routes where AA and BA compete had more than 50% of all U.S.-U.K. gateway-gateway passengers in 1995, and more passengers than traveled in all U.S.-Germany nonstop markets combined. The number of local passengers transported by American alone in the six overlap routes was roughly 13 times greater than the number transported by United in the two routes where it overlapped with Lufthansa (ORD-FRA and IAD-FRA), and more than 30 times greater than the number transported by Northwest in the markets where it overlapped with KLM (MSP-AMS and DTW-AMS). Delta and its European alliance partners competed on ten nonstop city pairs, but these were all very thin routes that prior to the request for antitrust immunity had been operated under code-share/block space arrangements rather than independently operated service.

2.

Confidential Appendix ("CA"), pp. 8-10.

3. As discussed at Section 1.12 of the Merger Guidelines, the Department of Justice considers a product (such as nonstop service) to be a separate product market if the sellers of that product can price discriminate between the buyers that require that product and the buyers that are willing to buy other products (here, one-stop or connecting service). Since time sensitive consumers are usually also unwilling or unable to meet minimum stay limitations and advance purchase requirements, airlines can readily separate leisure from business customers by imposing these restrictions on tickets. Leisure customers in excess of an airline's projected capacity can then be diverted to one-stop alternatives or offered a higher fare since the airlines carefully control inventory available at discount levels.

4. Reliable data on international traffic from interior U.S. points is difficult to find. DOJ had obtained traffic data directly from significant U.S.-U.K. competitors, but the data is from the 1995 time period. While such data remains useful for some purposes (see, e.g., Appendix A), events such as the cancellation of the BA-US Airways Alliance make it less helpful for purposes such as measuring market shares. Some airlines use consolidated CRS booking data to measure market shares, but based on a comparison of a sampling of CRS booking data to Origin and Destination data, DOJ is not convinced that the CRS data accurately reflect carriers' market shares.

5. See, Berry, Steven, Carnall, Michael and Spiller, Pablo, "Airline Hubs: Costs, Markups and the Implications of Customer Heterogeneity," NBER Working Paper 5561, May 1996.

6. An examination of airline schedules suggests that one to one and a half hours in the typical minimum connect time, and that the difference in total elapsed time between nonstop and connecting flights (relative to nonstop flights) is about two hours for a route like LAX-London with numerous, well located connecting points and frequent service.

7. Id., at B25 and B27. For passengers on the East Coast, including those in the huge New York market, there are in any event few connecting opportunities to London that do not involve backtracking to a more distant gateway -- a highly undesirable option for a time-sensitive traveler.

8. See, e.g., Borenstein, Severin, "The Evolution of U.S. Airline Competition," Journal of Economic Perspectives, Spring 1992, p. 62; Levine, Michael E., "Airline Competition in Deregulated Markets: Theory, Firm Strategy and Public Policy," Yale Journal on Regulation, Spring 1987, 453-454.

9. See 1996 American Express Survey of Business Travel Management, p. B26 (36% of all corporations, and over 50% of corporations with annual travel budgets in excess of $1 million, allow employees to fly business class on international flights). See also, survey by Runzheimer International reported in Travel Weekly, August 11, 1997, p. 29 (56% of corporate respondents permit the use of business class when their trip exceeds certain time limits).

10. See Official Airlines Guide 1997 Business Travel Lifestyle Survey, quoted in Travel Weekly, September 8, 1997 ("most convenient schedule" is the most influential factor determining international business travelers' choice of airlines. "Cheapest available fare" is listed as ninth in importance). See also BAP 041142

11.

CA pp. 11-13.

12. There is express train service at Gatwick with faster service to central London, but this appears unattractive to time-sensitive travelers because it does not take them directly to their hotel or business destination.

13. Large, sophisticated airlines can frequently track large corporations' travel patterns by using CRS data. Our review of the terms of American's and BA's corporate contracts shows fairly wide variation in the discount terms from customer to customer.

14. See AA/BA DOT AA 2288.

15. See, e.g., BAX 229936-229945, CA pp. 18-27.

16. See CA p. 28, BAX 611868.

17. See, Merger Guidelines, §2.2. In markets with differentiated products, a merger may place the merged company in a position to raise prices unilaterally to purchasers who find the products offered by the merging firms to be their first and second choices. The degree to which the merged entity will have power to raise price to such purchasers will depend on how close the merging products are as substitutes and on how well other firms in the market can reposition themselves to replicate the lost competition for those customers who are losing their second choice.

18. American's frequent flyer program is already very strong in New York, and would become unrivaled if it combined with USAirways' program.

19. See, Slot Allocation: A Proposal for Europe's Airports, U.K. Civil Aviation Authority Publication CAP 644, February, 1995.

20. Among other problems, giving preference to carriers wishing to serve U.S.-U.K. markets might violate applicable European Community slot allocation rules.

21. See CA pp. 29-37, December 4, 1996 letter from Alastair McDermid of BAA to Chris Allen of BA.

22. Carriers wishing to provide service between LHR and the U.S. are somewhat constrained in the departure and arrival times they can use. Time differences and preferred departure and arrival times mean that there are "windows" or "time channels" for LHR departures and arrivals. The preferred times for arrivals at LHR from points in the eastern U.S. are in the morning, from roughly 6 a.m. to noon. (There is also a brief evening window, between about 9 p.m. and 11 p.m., corresponding with early morning departures from the U.S. Such slots are only used in high frequency business markets such as New York-LHR by carriers that also operate at the more commonly preferred times). The preferred times for departures from LHR to the eastern U.S. are in the late morning, although departures are feasible between approximately 7 a.m. and 5 p.m.

23. Regulation on Slot Allocation at Community Airports, Council Regulation (EEC) 95/93. Under the rules, 50% of available slots are earmarked for new entrants. Article 10, §7.

24. Slot Allocation: A Proposal for Europe's Airports, supra.

25. See Joint Application, Exhibit JA-3, Appendix A.

26. BAP 020711-24.

27. See, e.g., Severin Borenstein, "Airline Mergers, Airport Dominance, and Market Power," American Economic Review, Volume 80, (May 1990), pp. 400-04, and Greg Werden, Andrew Joskow, and Richard Johnson, "The Effects of Mergers on Price and Output: Two Case Studies from the Airline Industry," Managerial and Decision Economics, Vol. 12 (1991), pp. 1-12.

28. See, e.g., AA/BA DOT AA 27736-38, 27872-75, 27879.

29. All airlines use some variation of the old Quality of Service Index ("QSI") formula developed by the Civil Aeronautics Board for route proceedings. The basic formula weighs all service in a market using factors such as type of service (nonstop vs. direct vs. connecting), aircraft type and frequency. The models currently in use by major airlines have become much more complex, and can include variables to account for presence (both in the city pair and at the end points), relative fares of carriers in the market, and time of day. See, e.g. BAP 140901-12, and CA p. 15 BAX 2227, estimating the value of the s-curve.

30. See, e.g., AA/BA DOT AA 7865, 27720, 27736-8, and 27872.

31. Testimony of Donald J. Carty, p. 40, AMR Corporation and American Airlines, Inc. v. UAL Corporation et al, USDC for the SDNY, 91 Civ 7773.

32. Statement of Department of Transportation's Policy Regarding Unfair Exclusionary Conduct in the Air Transportation Industry. Notice 98-16, April 6, 1998.

33. Exits by United from Washington-Zurich and American from Miami-Frankfurt suggest that a hub at one endpoint may not always provide enough support to compete against an alliance with a larger hub at the other endpoint.

34. United is likely to increase its frequency in ORD-LHR assuming there were an open bilateral and available slots, since the current bilateral restricts its operations.

35. See Severin Borenstein, The Dominant-Firm Advantage in Multi-Product Industries: Evidence from the U.S. Airline Industry, 106 Q.J. Econ. 1237 (1991).

36. British Midland is the second-largest slot holder at Heathrow, with approximately 14% of all Heathrow slots. All of those slots are used to operate its Heathrow network, however, and its public announcement makes clear that it would not enter U.S.-U.K. markets if it had to use its current slots to do so.

37. Source: Confidential data provided to DOJ by carriers serving New York-London market.

38. CA pp. 38-40. See also BAP 041142.

39. Indeed, the reason why both AA and BA provide the high level of frequency they do today is because each recognizes the critical importance of frequency, and will not allow the other to gain a significant edge. CA pp. 41-51, (BAX 700358-60, 229420-25).

40. The parties are likely to argue that relative frequency is important in short-haul domestic routes, but not on longhaul routes. Those arguments are undercut by their competitive actions on the JFK-LHR route as well as by planning documents such as BAP 140901-12.

41. It appears that the Alliance will operate at least 10 daily subsonic round trips in JFK-LHR, supplemented by supersonic service, and service in the overlapping markets of EWR-LHR and JFK-LGW. See CA p. 52-55 (BAX 306630-306633). BA's LHR slot holdings would also give the Alliance the flexibility to add more frequency in response to any challenge to its position of primacy.

42. It took American four years to build its presence to its current level.

43. This discussion assumes that American and USAirways remain independent competitors. Should the two carriers follow through on their stated intention to negotiate a comprehensive code-sharing alliance, USAirways' potential as an independent competitor to the Alliance may be called into question. If that occurs, DOT may have to reconsider the question of immunity for the AA/BA Alliance.

44. It is important to note again that to provide effective competition with the Alliance, slots made available to new entrants must be well-timed for transatlantic service.

45. E.g., Order 97-9-4, p. 17.

46. Had there been open skies over the last ten years -- when there has been some capacity expansion at Heathrow -- these markets might well be served today from Heathrow.

47. CA p. 3. Average fares were calculated for all passengers paying between $200 and $999 one-way for a ticket. Passengers paying less than $200 were excluded to eliminate the effect of frequent flyer tickets and passengers above $999 are often paying for a different class of service (business or first class). Very few passengers actually pay the average fare in most transatlantic city-pair markets. Leisure passengers pay less, while business passengers pay more for unrestricted tickets. The average fare is thus indicative of the mix of passengers in a given market. The more attractive the service is to business passengers, the higher the average fare will be.

48. These results are consistent across other city-pairs examined. One of the few exceptions is the Dallas-Ft. Worth to London market, where average nonstop and connecting fares are similar. This result appears to be due to the fact that the DFW nonstop service is to Gatwick rather than Heathrow. As discussed below, business travelers have a strong preference for Heathrow service, and a connecting passenger must be sufficiently remunerative to replace a high-yielding local or connecting passenger.

Updated June 25, 2015