DEPARTMENT OF JUSTICE
I appreciate the opportunity to be here today. The future of the airline industry is a timely and important topic, both for the industry and for government policy makers. Of course, it is always more interesting to talk of the future than of the past, but I am constrained not to speculate on any particular law enforcement decisions that the Division might make in the future with regard to the airline industry. But, if past is prologue, then I hope my remarks today about the Division s treatment of competition issues in the airline industry to date are relevant as you contemplate where the industry is likely to be at the turn of the century. I will talk mainly about how the Division analyzes mergers and code sharing agreements in the airline industry.
ENFORCEMENT ACTIONS IN THE CLINTON ADMINISTRATION
Before I begin, I want to reiterate what I have stated on many occasions: that antitrust enforcement is and has to be fundamentally nonpartisan and bipartisan. Indeed, our record in antitrust enforcement in the airline industry vividly demonstrates that there is great continuity from one Administration to another. For example, the airlines fares case (U.S. v. Airline Tariff Publishing Co.) which we settled in March 1994 with the remaining seven defendants, was the product of two Administrations: it was investigated and filed in the Bush Administration, and pursued vigorously and resolved in the Clinton Administration. That consent decree was a major victory for consumers and businesses -- it prohibits computer exchanges of information that, in our opinion, constituted price fixing and which resulted in excess travel costs to the public of up to $2 billion.
Given the importance of the airline industry to our economy and to the communities across our nation and the traveling public, it is essential that our enforcement efforts be consistent and even-handed. Our objective is to ensure that the competitive benefits from deregulation continue. And while the benefits of deregulation may not necessarily have been evenly distributed, deregulation has benefited everyone by bringing vigorous competition in pricing and a proliferation of upstart airlines. A study by Professor Morrison estimates that "[d]eregulation has led, on average, to fares 19.5 percent lower than they would have been if regulation continued. The annual saving is about $7.8 billion (in 1991 dollars)." It is important that this competition -- protected by enforcement of the antitrust laws -- remains vigorous.
In addition to the airlines fares case that I mentioned earlier, our enforcement efforts in the airline industry in this Administration have been active, involving a variety of matters. In March 1994, the Division entered into a consent decree with the International Association of Machinists (IAM) regarding the terms under which the IAM could appoint union members to the boards of two competing airlines, Northwest and TWA. The settlement mandated that a "fire wall" be established between the IAM representatives serving on the two boards, and prohibits the directors from exchanging confidential business and pricing information. Another significant case, which we filed in October 1994, involved the Association of Retail Travel Agents's boycott of travel providers such as airlines and car rental companies which did not follow the Association's prescribed commission levels and other policies. The Association entered into a consent decree that prohibited it from engaging in such activities and required it to conduct periodic reviews of antitrust requirements with its officers and directors.
In May 1995, the Division filed an amicus brief in connection with an agreement between TWA and travel agents to settle a private case brought by the travel agents over the issue of commission caps. In response to concerns expressed by the Division, TWA and the travel agents modified their settlement by removing those parts of the settlement that fixed the commission levels TWA would pay all competing travel agents and created a collective incentive among all travel agents to favor TWA over its competitors. The Division then filed its brief noting that it did not object to the modified settlement.
On the merger front, the Division in February 1995 filed an antitrust suit and a proposed consent decree regarding Sabreliner Corporation's acquisition of Midcoast Aviation, Inc. from TWA. The consent decree required Sabreliner to divest its transient general aviation fueling facilities at Lambert Field in St. Louis since the merger of Sabreliner's Lambert Field facilities with those of Midcoast would have created a monopoly in the sale of jet fuel to transient general aviation customers.
Speaking of mergers, this is, of course, an important time in the history of the airline industry, with the talk of a new wave of consolidation among major airlines and the growing number of cooperative agreements between domestic and international airlines. It is not surprising that there is talk of mergers in the airline industry; we have just ended one of the busiest years in terms of merger review in the history of the Antitrust Division. 1995 represented a record year in merger activity in the United States: 8,956 mergers worth a total of $457.88 billion. Throughout the economy, companies are consolidating. The majority of these ventures pose no antitrust concern; however, among the few mergers that present an antitrust question, some have been the largest in history, both in terms of the acquisition price and the commerce potentially affected.
Certainly, if it had occurred, the proposed sale of USAir to United would have been an example of one of these "mega-mergers." Between them, these two major airlines serve hundreds of cities -- a merger between them would have affected competition in many airline markets, and had the potential to provoke additional mergers or alliances in the industry. As a result, I think it would be useful to say a few words about the Department s merger enforcement policy in the airline industry.
ANTITRUST MERGER ANALYSIS IN THE AIRLINE INDUSTRY
As you know, the Division has had the authority to review, and since 1989 to challenge in court if necessary, mergers, acquisitions and intercarrier agreements concerning domestic air transportation. However, the last wave of significant airline mergers occurred prior to 1989 and were reviewed by the Department of Transportation. Although the Division participated in merger review proceedings before the Department of Transportation, DOT did not always follow our advice. Now that the decision to challenge mergers in courts is in our hands, many wonder what the Division s view would be concerning any potential merger between domestic airlines today. While I won t speculate on the outcome of any potential merger review, I would like to set out certain principles that we apply in all merger cases, and comment on the specific framework that we have applied to airline mergers and joint ventures in the past.
In analyzing the likely competitive effects of airline mergers and acquisitions, we use the principles contained in the 1992 Horizontal Merger Guidelines. The Merger Guidelines provide a clear framework for determining whether a transaction is likely significantly to reduce competition. The starting point for merger analysis is identifying the relevant markets in which the merging firms compete, and identifying the firms that compete in those markets.
We have consistently found that relevant airline markets are generally no larger than city-pair routes. A passenger who needs to travel from Washington to Kansas City will not go to Cleveland instead if fares to Kansas City increase. As in any merger, market definition focuses on identifying those products or services that consumers view as reasonable substitutes for the products or services of the merging firms.
We have also found that relevant airline markets may be narrower than city pairs. For many air travelers, not all flights on a given city-pair route would be viewed as adequate substitutes for each other. Depending on the traveler s time-sensitivity, for example, he or she might not switch to a one- or two-stop flight, even if the fares on all nonstop flights on the city-pair route were increased significantly. Similarly, while some travelers may be indifferent as to which of several airports in a city they use, others might not switch to a flight using an alternative airport even if the fares on flights using the most convenient airport were to increase significantly. As a result, there can be instances in which the relevant market is limited to the nonstop flights between two cities, or to the flights using a specific airport in a large city.
If a proposed merger would produce a carrier with an undue market share in a city pair or collection of city pairs, or would result in high concentration in the relevant market, the Division then proceeds to examine other factors relevant to whether increased market concentration, together with such factors, would enable competitors in that market to raise prices to consumers or reduce the quality of the service they provide. These factors include entry conditions.
There is much attention given these days to successful entry by new, low-cost airlines, and as competition advocates, we are gratified by their success. As a result, competition has increased and consumers have benefitted from this new entry, receiving more service and lower fares. I would like to strike a few cautioning notes, however, regarding entry conditions in the industry.
First, incumbent airlines have pursued efficient route structures to more effectively serve a large number of cities -- basically the "hub-and-spoke" systems operated by major carriers today. Many of the new entrants, however, have pursued point-to-point entry on heavily traveled routes. Consequently, many city-pair routes remain highly concentrated, with little prospect of new entry. Second, airlines use sophisticated marketing techniques -- including computer reservation systems, code-sharing, frequent flyer programs and commission overrides -- to develop loyalty to their airline among consumers and travel agents and to minimize the effect of entry from new, start-up airlines. Finally, through the use of computer reservations systems, yield-management technology, and selective pricing, airlines can quickly respond to changes in demand and supply on each of their routes, minimizing the effect of the entry on other fares, routes or flights. In short, the recent success of upstart airlines in a number of important city pairs does not mean that entry in all airline city pair markets is easy or likely. And, in fact, the record since deregulation shows that most upstart airlines have either failed or been merged into existing carriers. As a result, the Division has been, and continues to be, vigilant in monitoring entry conditions.
The Division has maintained an active merger enforcement program in the airline industry for many years, before and after deregulation. For instance, in the mid 1980's, the Division recommended that the Department of Transportation (which at that time had antitrust authority for domestic air transportation) disapprove two mergers, TWA/Ozark and Northwest/Republic, which involved the merger of the only two hub carriers at St. Louis and Minneapolis respectively. In both cases, the merging carriers were the only airlines providing nonstop service between the hub city and smaller cities in the surrounding region -- cities like Bismarck, North Dakota and Cedar Rapids, Iowa, which do not generate a large volume of traffic. For those city pairs, we concluded that only a hub carrier could generate sufficient traffic to achieve minimum efficient scale and that entry by any non-hubbing carrier was unlikely, even if prices went up, because they would be a significant cost disadvantage with respect to the incumbent carrier. The Department of Transportation, however, declined to follow our advice, and approved both the TWA/Ozark and the Northwest/Republic transactions.
Overlapping hubs need not necessarily be at the same airport. Some cities, like Washington, New York and Chicago, have more than one airport in their metropolitan regions. For example, in 1991 we opposed Eastern s proposed sale of 67 slots at Washington National to United because United operated a significant hub out of nearby Dulles International Airport. The sale of the Eastern slots to United would have combined previously competing hubs providing nonstop service from the Washington area to a number of Northeastern and Florida cities.
Our competitive concerns are not limited to mergers that eliminate existing hub competitors, however. We also have moved aggressively to block acquisitions that would have eliminated potential hub competitors. For instance, in 1989 when Eastern proposed to sell eight gates to USAir at the gate-constrained Philadelphia International Airport, we announced our intention to prevent the proposed deal because the sale would have prevented the establishment of another hub operation potentially capable of competing with USAir s Philadelphia hub operation.
Finally, we would likely examine whether merging airlines operate what we call "alternate" hubs. The concern with an alternate hub merger generally relates to competition on smaller city pairs that do not have nonstop service, but do have competing connecting flights over alternative hubs in the same general region of the country. In such a case, the alternative connecting service between such city pairs within that region might all require transfer at hubs controlled by a very small number of carriers, or require unreasonably circuitous routing to a hub further outside the region. Although this concern has not led us to challenge a transaction in the past, it is an issue we will continue to examine.
The Division has also reviewed numerous proposed transfers of international route authority. In route acquisition cases, if concentration is high and entry by new carriers or the addition of capacity by existing carriers is precluded by international agreement, the Division is likely to object to a transfer of route authority that significantly increases concentration in the relevant markets. For example, the Division in 1991 opposed TWA s proposed sale to American Airlines of international route authority to London from six U.S. gateways. .ANTITRUST ANALYSIS CONCERNING CODE SHARING
A topic that I know is of great interest today is the Division s review of marketing alliances between airlines, particularly those involving international travel. During the 1980s, in the wake of airline deregulation, code sharing developed in the United States between our major trunk airlines and so- called "regional" or "commuter" airlines. During the 1990s, code sharing spread well beyond its early uses. The practice has proliferated between U.S. and foreign carriers serving international city-pair markets. Internationally, dozens of code share agreements now exist between U.S. airlines and their foreign counterparts covering hundreds of city pairs throughout the world. Some of the most prominent international code share agreements include those between USAir and British Air, Northwest and KLM, United and Lufthansa, and Delta s alliances with several European airlines.
The term "code share" can mean as little as allowing another airline to use its code when its sells seats on your plane on a route in which it cannot compete, or as much as comprehensive integration of marketing and operations that involves joint decisions on price, capacity, schedule and other competitively sensitive matters. In conducting an antitrust investigation, we always analyze the specific terms of each agreement on a case-by-case basis -- and the specific terms of agreements vary widely.
The antitrust laws fully apply to all domestic code sharing agreements, and, absent a grant of express statutory antitrust immunity by the Department of Transportation, the antitrust laws fully apply to the international code shares as well. Yet nearly all existing code shares have gone into effect without objection from the Department of Justice. The reason is simple -- most code sharing agreements have been procompetitive and pro-consumer, and do not violate U.S. antitrust laws.
Those of us in the Department of Justice believe that an important reason that U.S. and foreign carriers have structured their code sharing agreements in ways that promote competition and consumer welfare is the knowledge that we closely review these intercarrier agreements and will act decisively in the event a code sharing or other joint marketing agreement unreasonably restrains airline competition. So far, carriers have demonstrated a high sensitivity to the antitrust laws, and have structured their agreements accordingly.
To antitrust law enforcement authorities, code sharing agreements are simply forms of corporate integration that fall somewhere between outright merger and traditional arm's length interlining agreements. As with mergers and acquisitions, code share agreements may raise traditional horizontal, as well as vertical, concerns. And as with mergers and other intercorporate agreements, code sharing has the potential to be significantly procompetitive -- it can create new service, improve existing service, lower costs and increase efficiency, all to the benefit of the traveling public. By the same token, code sharing can also be a mask for anticompetitive arrangements between actual or potential competitors to allocate markets, limit capacity, raise fares, or foreclose rivals from markets, all to the ultimate injury of consumers. The ability to distinguish the latter from the former is crucial for aviation policy-makers and antitrust enforcement authorities.
In assessing the effect on competition, the first question we ask is whether code sharing partners are actual or potential horizontal competitors, and if so, in what city pairs. From an antitrust viewpoint, the greatest threat to competition comes when two of very few airlines that compete on a city pair enter into a code sharing agreement on that city pair. This is the same concern that would be present if the two carriers were planning to merge. Any time two of very few airlines on a city pair act jointly, whether in a domestic U.S. market or an international city pair, we are concerned about the effect on competition.
Most code sharing agreements, whether domestic or international, have raised relatively few such horizontal concerns because the parties by and large are not direct competitors, and are unlikely to become competitors in the foreseeable future on the city pairs affected by the agreement. Examples in the United States of domestic code sharing agreements that tend not to raise significant horizontal issues are those between commuter and jet carriers. Our investigations over the years have shown that, because they operate different types of aircraft, jet and commuter airlines seldom serve the same city-pair markets, and are not likely to enter very many of the same markets.
In many international city-pair markets, entry is restricted by legal barriers, although there is an increased trend, which we applaud, towards open skies for international service, especially in Europe. But under current laws, foreign airlines are not potential entrants in U.S. city pairs, and U.S. airlines are not permitted to serve local traffic on European city-pair markets. The vast majority of markets affected by international code sharing agreements have been behind-gateway markets that neither code sharing airline can serve directly. Code share agreements thus create new "on-line" (same carrier) service in behind gateway cities. If behind gateway markets are the only ones affected by a code share agreement, then the agreement is not likely to be challenged on horizontal grounds.
But there are numerous situations where potential code share partners are actual or potential competitors. In deregulated U.S. markets, the most likely nonstop competitors on any city pair are carriers with a hub at one of the endpoints. Accordingly, the most serious threat to competition is presented when two carriers enter a code share or other joint marketing agreement that includes service between their hubs. For instance, we investigated a proposed code share agreement between two major airlines that would have included code sharing on nonstop flights between several of the hubs of the two partners. Although not a merger, the proposed arrangement went well beyond minimal code sharing, and contemplated joint decisions on schedule and capacity in the hub-to-hub markets. Fortunately, that transaction was abandoned after we informed the carriers of our antitrust objections.
The same hub-to-hub concerns exist for international code share agreements, even those that relate primarily to connecting traffic between behind gateway cities. If a code share arrangement includes hub-to-hub markets, there is an increased antitrust concern, and the Department will conduct a careful competitive analysis of such markets. To do this, we consider whether the partners will both operate flights in the market, and whether their capacity, schedule and pricing decisions will remain independent. And by independent I mean much more than retaining the legal right to act alone -- I mean that the agreement 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 increase its operating capacity to gain market share. In other words, the more the agreement resembles a traditional interline pact limited to how passengers will be transferred and accounted for between connecting airlines, the less likely is any competitive harm present.
If independent operations are not contemplated, we look for evidence that one of the partners is not likely to enter (or is likely to exit) absent the code share agreement, or that coordination on the hub-to-hub market is necessary to achieve significant procompetitive efficiencies in serving beyond-hub city pairs on a code share basis. The evidence that such efficiencies outweigh the potential competitive harm in the hub-to-hub market must be clear.
An important consideration in assessing the effect of international code sharing on competition is the sweeping antitrust immunity that has been granted to the International Air Transport Association ("IATA") that permits otherwise competing airlines collectively to discuss and set passenger fares between the United States and foreign destinations. Under the Merger Guidelines, we examine the extent to which a merger or joint venture arrangement will increase the likelihood that the firms remaining in the market will be better able to coordinate their behavior in a way that harms consumers. This inquiry tends to focus on market characteristics that make it possible for firms tacitly to reach consensus on the terms of coordination, and detect and punish deviations from those terms. IATA tariff conferences of course make it easy (and legal) for member carriers to agree expressly on prices in markets where they compete. Thus, the presence of IATA tariff coordination in affected markets may lead the Department to challenge code sharing between horizontal competitors in situations where otherwise it would not. Moreover, if a proposed code share agreement has both procompetitive and anticompetitive effects, the Department considers, as part of its overall competitive analysis, whether continued IATA membership is necessary to achieve any benefits and whether withdrawal from IATA would reduce any harm. In particular, we evaluate whether a code share alliance setting its fares independent of IATA would constitute a less anticompetitive alternative means by which the benefits of the alliance can be achieved.
Another important factor we consider is whether an "open skies" bilateral aviation agreement has been negotiated with the relevant foreign governments. If an "open skies" bilateral applies to the affected markets, then new entry by a non-hub carrier is possible, and we will investigate how likely such entry would be in the event the code share partners attempted to raise fares or reduce service.
The threat of a code share to competition on international gateway-to- gateway city pairs is increased where entry is governed by a restrictive bilateral, particularly if only two authorized carriers are involved. That was the case in the British Air/USAir arrangement, which applied to Philadelphia- London, Washington-London and Charlotte-London. Not only were these hub-to-hub markets for these airlines, but new entry was prohibited under our bilateral with the United Kingdom so long as BA and USAir both continued to hold route authority. The BA/USAir code share agreement contemplated common ownership and such significant collaboration that the carriers would no longer have the incentive to compete aggressively with each other on those three routes. Accordingly, in March 1993 we filed a lawsuit and negotiated a consent decree that required USAir promptly to sell its right to operate those routes to a new carrier, or surrender the routes to the U.S. Department of Transportation for reassignment. This solution mitigated the competitive harm on the hub-to-hub routes, and permitted the transaction, which was procompetitive in numerous other markets, to proceed.
In contrast, where an open skies policy prevails, route authority is not needed to facilitate new entry; indeed where there are open skies, there is nothing to divest. Thus, a liberal bilateral can reduce antitrust concerns that might otherwise be presented by international code share arrangements. That does not mean, however, that open skies are necessarily a complete solution to the loss of competition that can be caused by some hub-to-hub code share arrangements.
A new development in international code share agreements is the increased trend of code share partners to request that the Department of Transportation grant antitrust immunity to their agreement. Delta and three of its European partners, Sabena, Austrian and SwissAir, have asked for immunity, as have American and Canadian Airlines. More may follow. Our view is simply stated: It is not necessary for code share partners to receive antitrust immunity for any agreement that would not violate the antitrust laws; and conduct that would violate the antitrust laws should not be permitted, much less immunized. From our perspective, we will continue to review all code share agreements, including those where the parties seek immunity, and seek to prevent any anticompetitive agreements from being implemented.
Like mergers, code sharing can also raise vertical competitive issues. Most airlines are both competitors and customers of other airlines. Airlines engage primarily in the business of selling seats on their airplanes directly to passengers, and when two airlines are both offering to sell a seat to the same passenger to travel between the same cities, they are clearly horizontal competitors. But airlines also interline passengers to one other. One way to think of interline air service is as airlines trading passengers with each other at the connecting hub. The prices for these passengers are the prorates or divisions offered by the buying and selling airlines.
In the absence of competitive restraints, we would expect these prices to allocate interline passengers to the most efficient carriers, who in turn offer the best price and service options to the ultimate consumer, the passenger. When two airlines that supply one another with interline passengers enter into a code sharing agreement, however, other airlines that had previously obtained interline passengers from one of the parties to the agreement might find that it can no longer obtain interline passengers at the same cost as it had before. The code share agreement might in essence constitute an exclusive dealing, or at least a "preferred provider," relationship among the two airlines. While such an agreement may improve the efficiency of the two code sharing airlines, it could also impede the ability of other airlines in those markets to compete.
In the international context, for example, it is possible that in some gateway-to-gateway markets, carriers must be able to obtain a minimum number of interline passengers from beyond gateway markets for service to be economically viable. If, as a result of a code share agreement, a competitor of one of the code sharing partners could no longer effectively interline with the other code share partner, the competitor might be unable to continue profitably to serve the route. If the non-code sharing carrier likely would be forced to exit the route, and as a result fares on the gateway- to-gateway route increase (or service decrease), then the code share agreement could violate the antitrust laws.
Whether a code share agreement will in fact foreclose competing carriers from access to interline passengers, and whether such foreclosure would in fact lessen competition and raise fares in a market, will depend on many factors, including demand in the market, the relative costs of carriers serving the market, the availability of other sources of interline passengers (that is, the presence of other non-code sharing carriers on the route who could supply passengers), and the efficiencies resulting from the code share. To date, we have not challenged any code sharing agreements on the basis of this type of vertical foreclosure, but it is a concern we keep in mind as we review proposed code share agreements under the antitrust laws.
I hope that these comments give you an idea of how the Division analyzes antitrust issues in the airline industry. With your help, we will continue our efforts to preserve and enhance the beneficial role of competition and to safeguard the competitiveness of the American airline industry for the 21st Century.