R. HEWITT PATE
SUBCOMMITTEE ON ANTITRUST, COMPETITION, AND BUSINESS RIGHTS
INTERNATIONAL AVIATION ALLIANCES:
NOVEMBER 7, 2001
Mr. Chairman and members of the Subcommittee, I am pleased to present this written statement on how the Antitrust Division analyzes international aviation alliances. I know Subcommittee members are aware of the proposed alliance between American Airlines and British Airways currently under consideration by the Department of Transportation. It would not be appropriate for me to comment specifically on this proposed alliance while it is pending before us. But I am happy to discuss in general the Antitrust Division's analytical approach.
Antitrust Enforcement in the Airline Industry
Since Congress enacted the Airline Deregulation Act of 1978 and placed greater reliance on antitrust enforcement in lieu of regulation, the Antitrust Division has maintained an active enforcement program in the airline industry to ensure that the benefits of airline competition sought by Congress are realized by consumers. During the 1980s, the Division recommended that the Department of Transportation (which had authority over airline mergers until 1989) disapprove two mergers, TWA/Ozark and Northwest/Republic, which involved the merger of the only two hub carriers at St. Louis and Minneapolis respectively. The merging carriers were the only airlines providing nonstop service between the hub city and smaller cities in the surrounding region.
More recently, in October 1998, the Division filed suit to block Northwest Airlines from buying a controlling stake in Continental Airlines. They were the fourth- and fifth-largest U.S. airlines, competing on hundreds of routes across the country, and the proposed acquisition would have substantially diminished their incentives to compete against each other. The Division rejected Northwest's plan to put its Continental stock in a "voting trust" for six years as insufficient to prevent the competitive harm likely to result from the acquisition. Last November, after trial had begun, Northwest announced it was selling Continental the shares that would have given it control, and would retain only a five-percent share. Because the sale of control back to Continental remedied the competitive harm, the Division dropped its lawsuit.
And this summer, we announced our intent to challenge the United Airlines/ US Airways merger, the second- and sixth-largest airlines, after concluding that the merger would reduce competition, raise fares, and harm consumers on airline routes throughout the United States and on a number of international routes, including giving United a monopoly or duopoly on nonstop service on over 30 routes. We concluded that United's proposal to divest assets at Reagan National Airport and American Airlines' promise to fly five routes on a nonstop basis were inadequate to replace the competitive pressure that a carrier like US Airways brings to the marketplace, and would have substituted regulation for competition on key routes. After our announcement, the parties abandoned their merger plans.
The Division has also moved aggressively to challenge various proposed acquisitions of gates or slots that would eliminate existing or potential hub competition, including Eastern's proposal to sell a block of gates to USAir at the gate-constrained Philadelphia International Airport, and Eastern's proposed sale of slots and gates at Reagan Washington National Airport to United, which operated a significant hub out of nearby Dulles Airport.
The Division has also challenged transactions involving international route authority. For example, the Division brought a civil action under Clayton Act § 7 against the 1991 investment agreement between British Airways and USAir, after concluding that the transaction threatened competition in gateway city pairs and certain connecting city pairs -- in particular, service between Northeast and Mid-Atlantic cities and London.
In addition to challenging transactions that we concluded would adversely affect the structure of the airline industry, we have investigated and challenged practices as collusion in violation of section 1 of the Sherman Act. In 1992, we sued Airline Tariff Publishing Co. and eight major airlines, alleging that the airlines used the ATPCO electronic fare submission and dissemination system to fix prices, which we concluded had cost consumers up to $2 billion in travel expenses. The consent decrees ultimately entered into banned improper signaling of future pricing intentions.
Other conduct that the Division has challenged includes agreements on international fares undertaken outside the scope of the International Air Transport Association, such as the criminal indictment of Air Florida and others in 1984; and the solicitation by American Airlines President Robert Crandall of a price increase from one of his chief rivals, which we challenged in the early 1980s as attempted monopolization. More recently, the Division has challenged the conduct of American Airlines in responding to new entry by low-cost airlines in its Dallas-Fort Worth hub as predation in violation of section 2 of the Sherman Act. We are now appealing the district court's grant of summary judgment for American.
In addition to these law enforcement efforts that I have described, the Antitrust Division has also engaged in competition advocacy in various matters before the Department of Transportation. One type of this competition advocacy has been filing comments in DOT proceedings to consider whether to approve proposed international airline alliances and to decide whether to grant antitrust immunity for all or part of any such alliance, matters over which DOT retains authority under 49 U.S.C. §§ 41309 and 41308, respectively. As that is the subject of your hearing today, I will now turn to a more detailed description of how we approach the analysis of these international alliances, with particular emphasis on what is referred to as "code sharing."
Antitrust Analysis of International Alliances
Absent an express grant of antitrust immunity by DOT, the antitrust laws apply to international airline alliances just as they do to domestic airline alliances.
Airline marketing alliances are essentially joint ventures between airlines. These alliances fall somewhere between an outright merger and a traditional arm's-length interline agreement. They come in all shapes and sizes. Some may involve simply sharing frequent-flyer programs or airport lounges. Others may involve "code sharing," in which each carrier uses its partner's two-letter airline designator code for listing its own flights in computer reservation systems, in which case the alliance probably includes some effort to coordinate travel logistics such as check-in and gate locations. Occasionally, an alliance is accompanied by a stock investment by one airline in its partner. Alliances can involve commuter carriers, domestic carriers, foreign carriers, or a combination.
Most of the Division's experience with alliances between major airlines has been in the international marketplace, between airlines of different nationalities who may be restricted from serving each other's domestic markets. Alliances between major U.S. carriers, as distinct from alliances between hub carriers and commuter carriers that serve those hubs, are a relatively recent phenomenon; there have been only two significant alliances entered into between major U.S. carriers thus far, the one between Continental and America West entered into several years ago, and the more recent one between Northwest and Continental. But in most respects our analytical approach is the same whether the airline alliance is domestic or international, although there are a few possible differences, which I'll mention in a minute.
The term "code sharing" can mean as little as one airline allowing another airline to use its computer reservation system codes to sell seats on its planes on routes in which the second airline cannot compete, or as much as comprehensive integration of marketing and operations that involves joint decisions on price, capacity, schedules, and other competitively sensitive matters.
Alliances involving code sharing can have significant procompetitive as well as significant anticompetitive potential. On the procompetitive side, they can create new service, improve existing service, lower costs, and increase efficiency, all to the benefit of consumers. On the anticompetitive side, they can result in market allocation, capacity limitations, higher fares, or foreclosure of rivals from markets, all to the injury of consumers.
The antitrust investigation of a code share involves a case-by-case analysis of the specific terms of the agreement to assess its effect on competition. The first step is to define the relevant market, which may be one city-pair route, or a set of such routes, and then to measure that market in terms of its participants and concentration. One important consideration is whether the code-sharing partners are actual or potential horizontal competitors. Generally, the greatest threat to competition comes when two of very few airlines that compete in a market enter into a code-share agreement in that market.
In contrast, when a code share is proposed to link a city-pair market served by one carrier with a city-pair market served by the other, rather than to cover a city-pair market in which both carriers are actual or potential competitors, the proposed code share would create what is referred to as an "end-to-end efficiency," which is generally procompetitive.
After the relevant market has been defined and measured, the next step is to assess the potential adverse competitive effects of the code share. If the code-share partners will both operate flights in the market, the Division considers whether the agreement is structured in a way that the partners' capacity, scheduling, and pricing decisions will remain independent -- that is, whether it is structured in a way that gives each carrier the strongest possible incentive to sell seats on the flights it operates rather than on those of its code-share partner, and to cut its prices and improve its service to gain market share against its partner.
One approach taken in some code shares to preserve some independence in pricing and marketing of seats on the shared flights has been to use a block-seat arrangement, where the non-operating carrier purchases a fixed number of seats and bears the risk of loss if those seats are not sold. The block-seat arrangement is not an ideal solution, because the cost of the block of seats to the non-operating carrier, which is the key determinant of the ultimate fare to the consumer, is set by agreement between competitors. But the block seat arrangement is an improvement over joint sales and marketing, because it can create some additional incentive for each partner to market its own seats aggressively.
In cases in which independent operations by the two partners are not contemplated or considered likely under their proposed code-share agreement, and the Division concludes that the code-share agreement would reduce or eliminate competition between the code-share partners in certain city-pair markets, the next step in the Division's analysis is to consider how likely it would be that new competitors would enter these markets in response to any anticompetitive behavior by the code-share partners. If sufficient and timely entry could be expected to neutralize any anticompetitive behavior, then the Division would conclude that the code-share agreement would not be likely to create or facilitate the exercise of market power by the code-share partners.
In the case of an international code share, an important threshold factor in assessing likelihood of new entry is whether the market is covered by an "open skies" bilateral agreement. Open skies means that new entry by a carrier is legally possible, although we would still need to investigate how likely such entry actually would be in the event the code-share partners attempted to raise fares or reduce service. On the other hand, where new entry is legally constrained by a restrictive bilateral agreement, the threat to competition of a code share on that city pair could be substantial, particularly if the code-share partners were the only two carriers authorized under the bilateral agreement.
And finally, if independent operations by the code-share partners in the relevant city-pair markets are not contemplated, and if sufficient and timely entry is not considered likely, we would look to see if there is persuasive evidence that one of the partners is likely to exit the market absent the code share. If so, because the partner would cease to be a competitive factor in the market no less if it exited than if it merged, this would be relevant to the Division's assessment of whether disapproving the code share, or withholding antitrust immunity, would prevent competitive harm. Similarly, we would also look to see if there were persuasive evidence that the code-share agreement would result in significant procompetitive efficiencies in serving other city pairs on a code-share basis -- efficiencies that could not otherwise be obtained except through the code share. If so, we would assess whether the procompetitive effect of these efficiencies would outweigh the potential competitive harm in the overlap city pair.
In short, we examine all of the facts and circumstances surrounding each code-share agreement and make our competitive assessment on a case-by-case basis.
As I mentioned, although our analytical approach to domestic alliances is similar to the approach taken with international alliances, there are a couple of differences. The first is that U.S. carriers have virtually unlimited rights to expand their operations within the U.S. -- subject to landing slot ceilings at a few airports -- and thus are, at a minimum, potential competitors of one another. In contrast, an international code-share agreement may be the only way in which a U.S. carrier can gain entry to serve foreign markets. The second is that major U.S. carriers -- even those with different regional strengths -- often compete with one another in significant markets, and sometimes are the only competitors in those markets, such as hub-to-hub-markets. In contrast, in many international alliances, laws and treaties often preclude U.S. carriers and their alliance partners from competing broadly against one another.
The Antitrust Division has applied this analytical approach to a number of proposed international code-share agreements. In the majority of them, we have found no horizontal competitive concerns.
Where a proposed code share for which antitrust immunity is sought combines certain horizontal overlaps with significant end-to-end efficiencies, our policy has been to seek to limit the immunity for any city pairs on which the proposed alliance partners are two of very few current or likely future competitors. As I mentioned, for an international code-share agreement, the Department of Transportation has the authority to confer antitrust immunity, after consulting with us. When antitrust immunity has been sought, we have recommended that DOT "carve out" certain unrestricted fares involving these city pairs from the order granting antitrust immunity, provided that the carve-out could reasonably be done without sacrificing important consumer benefits created by the code share. For example, the Division recommended that seven city pairs be carved out of the Delta/Swissair/Sabena/ Austrian alliance (Atlanta-Zurich, Atlanta-Brussels, Cincinnati-Zurich, New York-Brussels, New York-Geneva, New York-Vienna, and New York-Zurich); one for the American/ Canadian Air alliance (New York-Toronto); two for the United/ Lufthansa alliance (Washington-Frankfurt and Chicago-Frankfurt); and two for the United/ Air Canada alliance (Chicago-Toronto and San Francisco-Toronto).
We designed this carve-out approach to permit U.S. airline passengers to obtain the benefits of increased efficiency and enhanced beyond-gateway service provided by a code-share agreement, while avoiding possible diminutions in gateway-to-gateway service or increased air fares as a result of the agreement. Where a carve-out approach would be insufficient, and a proposed code-share agreement presents the potential for significant diminutions in gateway-to-gateway service while providing little likelihood for enhanced beyond-gateway service, the Division is prepared to recommend that DOT grant no antitrust immunity whatsoever, or even that it disapprove the code-share proposal in its entirety.
To date, DOT has accepted all of the carve-outs the Justice Department has proposed, with the exception of the four New York/ Europe carve-outs the Division sought for the Delta alliance. In that instance, as in other approval orders, DOT required the foreign alliance partners to report fare and other data of the sort already reported by U.S. carriers, in order to enable a review of the effect of the alliance on price and service on these routes and consideration of additional carve-outs if they appeared warranted.
In addition, DOT has prohibited alliance partners from participating in "fare coordination" activities under the auspices of the International Air Transport Association. The Antitrust Division has for years raised concerns to DOT about this type of international cartel activity, and supports DOT's efforts in this regard.
One fairly recent international code-share agreement in which the Division provided comments to DOT was betweenAmerican Airlines and the TACA group, composed of six Central American airlines serving Costa Rica, El Salvador, Guatemala, Honduras, Nicaragua, and the Republic of Panama. American and some TACA carrier operated overlapping nonstop flights on virtually all routes between Miami -- the principal Latin American hub in the United States -- and the gateway cities in the Central American countries just mentioned, so that American and TACA had combined market shares ranging from 88 percent to 100 percent on those overlap city pairs.
At the same time, the number of passengers traveling between interior points in the United States beyond the Miami gateway and interior points beyond the Central American gateways -- the only passengers who could not already obtain full on-line service available from either American or the TACA group -- was an extremely small fraction of passengers flying gateway-to-gateway. So we found this to be an almost exclusively horizontal agreement, in contrast to the largely end-to-end international code-share agreements we had previously reviewed.
We concluded that the claimed efficiency benefits that were specific to the agreement were very slight, while some potential risks to competition would inevitably persist despite the best efforts to eliminate them through imposing conditions. The carriers did not seek antitrust immunity in their original proposal, but we nonetheless urged DOT to weigh these factors carefully before making a determination under 49 U.S.C. §41309 as to whether approval of this code-share agreement would be in the public interest. In the end, DOT approved the agreement with conditions, including requiring block-seat arrangements for the gateway-to-gateway part of the code share; requiring that the code share be non-exclusive, with an inference of noncompliance if there wasn't a code share with another carrier within two years; and a shorter duration before DOT would reconsider the agreement.
Subsequently, American and TACA applied for antitrust immunity for their code share. After some delays, they provided DOT with information it had requested, and that application is now pending before DOT.
Another fairly recent international code-share proposal in which the Division provided comments to DOT was betweenAmerican Airlines and British Airways -- not the pending proposal, but a proposal filed with DOT in 1997 that was dismissed in 1999 because of lack of progress between the United States and the United Kingdom toward an open skies bilateral. We filed comments in May 1998, urging that the alliance not be approved unless it was significantly restructured. The two air carriers were the largest ones offering service between the U.S. and the UK, and we concluded that the alliance would significantly reduce airline competition in that service. We stated that, while an "open skies" agreement would help provide an important competitive benefit in creating a possibility for new entry, and would be essential in our view in order to justify DOT approval of the proposed alliance under the standard in its statute, an open skies agreement would in the Division's view not be enough to realize that possibility for new entry, due to severe constraints on new access at London's Heathrow Airport. We advised DOT to also require that sufficient slots and related facilities be made available to the two carriers' competitors -- enough to operate at least 24 additional daily round-trips between the U.S. and Heathrow. And we also urged DOT, if it approved the alliance, to carve out the Dallas/London and Chicago/London city-pair markets from the alliance; both carriers operated hubs at all three airports.
To summarize, aviation alliances can and do take many different shapes and forms, and the our assessment of an alliance depends both on the terms of the alliance and on the carriers involved. Certain kinds of alliances have dealt with matters that were not competitively troublesome to the Division. Even those alliances that involved matters that the Division ordinarily considered competitively sensitive -- such as code sharing -- sometimes have involved carriers that did not have significant competitive overlap. Alliances concern us most when they involve carriers that are substantial competitors of each other's, and code sharing that could be used as a means for anticompetitively coordinating service and fare offerings.
The Antitrust Division assesses on a case-by-case basis -- and market-by-market basis -- whether a proposed code-share alliance is likely to act as a disincentive for the alliance partners to enter markets served by the other or to compete vigorously in markets that they both serve. We look to see whether the alliance is likely to divide and allocate markets, or to produce high fares. We place critical importance on carefully reviewing the actual terms of each alliance agreement.
I hope this written statement has helped the Subcommittee understand how the Antitrust Division evaluates international code-share agreements. I would be happy to respond to any written questions that you or other members of the Subcommittee may wish to submit.