Entry and Competition in the U.S. Airline Industry: Issues and Opportunities (1999)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Executive Summary This report updates a 1991 Transportation Research Board (TRB) re- port, Winds of Change: Domestic Air Transport Since Deregulation. The committee that authored Winds of Change endorsed deregulation. Con- sumers dearly had benefited from lower prices and new airline services during the 1980s, the first full decade after deregulation. More city-pair markets had three or more competitors at the end of the decade than at the beginning. Carriers had become more cost conscious and industry productivity was improving. All of this was achieved even though many avenues for competitive entry had not been fully opened nor exploited. Important vestiges of reg- ulation had continued, such as administrative limits on the use of some key airports and preexisting financial and contractual agreements between airport operators and established airlines. These holdover policies and practices gave the incumbent carriers advantages that helped offset the lower operating costs that some newer airlines might have enjoyed. Other post-deregulation developments favoring the incumbent carriers included the proliferation of frequent-flier programs—enticing travel- ers, particularly business travelers, to the services of a single carrier—as well as new ways for airlines to influence travel agents, such as offering extra commissions, known as overrides, for steering business their way. Partly for these reasons, few new carriers survived the decade follow- ing deregulation. Competition also diminished in some markets because of the U.S. Department of Transportation's (DOT's) acceptance of most

2 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY airline mergers, including several involving direct rivals. Although the Winds of Change committee gave the deregulated industry generally pos- itive marks, it expressed concern that with further concentration, too few airlines would survive to ensure the competition needed to preserve the gains of deregulation. The Winds of Change committee therefore urged several actions to encourage and safeguard competition: More vigilant reviews of mergers, particularly of airlines with over- lapping service areas; New rules to reduce the influence of airlines on travel agents, re- moving the contractual limitations on the use of competing computer reservation systems (CRSs); and Requirements that agents disclose to consumers any commission overrides and other booking incentives from individual carriers. The committee did not recommend further actions to overcome all pos- sible impediments to competition and indications of unchecked market power, recognizing that some of the developments that gave incum- bents a comparative advantage—such as hub-and-spoke operations, fre- quent-ffier programs, and CRSs—also bestowed benefits on consumers. Furthermore, the airline industry's return on invested capital, both before and after deregulation, had been scant, suggesting limited market power generally and raising serious concerns about the industry's ability to fi- nance itself over the long term. This study committee did not address the question of whether airline deregulation was sound national policy, as concluded in Winds of Change and in many other studies of the industry. Its emphasis is on preserving and enhancing the gains made since deregulation, primarily by changing public policy to foster more widespread and vigorous airline competition. UPDATE OF COMPETITION ISSUES Several early outcomes and features of deregulation reported in Winds of Change have carried over into the current decade. Trends in overall flight offerings, passenger traffic, and prices are evidence that consumers con- tinue to benefit. Scheduled departures have risen by nearly 20 percent

Executive Summa7y 3 since 1990, and the number of air travelers has grown by more than one- third. Adjusted for inflation, average fares fell by about 25 percent from 1990 to 1998, at about the same pace as the real decline in jet fuel and other airline costs—cost reductions that have stemmed in part from the competitive pressures ushered in by deregulation. Competition contin- ues to be particularly strong in medium- to long-haul markets, as airlines offer frequent flights through their hubs to an array of destinations. Be- cause of hub-and-spoke systems, travelers in such markets enjoy many airline choices as well as flight options, although transfers at hub airports are the norm. Certain characteristics of the industry, however, continue to raise con- cern. Travelers in many cities with major hub airports continue to have few airlines to choose from when flying in short-haul, nonstop markets, or their "spokes." As a consequence, hubbing carriers often account for 50 percent or more of the local passenger traffic in hub cities. Public discontent over airlines charging higher prices for unrestricted tickets— usually for purchases by time-sensitive business travelers—is most pro- nounced in these hub cities and other markets in which short-haul ser- vice is dominated by one or two carriers. Many new, low-fare carriers have entered such markets, often eliciting sharp price-cutting and other aggressive responses by incumbents, which the new entrants have protested as unfair. Another relatively recent development is that major airlines have formed partnerships to share frequent-flier programs and to coordinate flight schedules, fares, and services through codesharing. Although potential benefits and costs are associated with these develop- ments, the outcome remains unclear and worrisome. Airline Pricing Practices Over the past decade, public dissatisfaction with airline ticketing prac- tices has grown—particularly because of the many restrictions on lower- priced tickets and increases in the price of unrestricted tickets. Advance purchase requirements, weekend stay-over rules, penalties for changing reservations, and various other booking restrictions are aimed largely at distinguishing price-sensitive from time-sensitive travelers. The lowest fares usually are paid by travelers who are less concerned about sched- ules, are unlikely to cancel or modify their travel plans, and are more will-

4 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY ing to fly during off-peak periods and to use uncongested, but typically more remote, airports in return for cheaper fares. The highest fares usu- ally are paid by travelers who prefer unrestricted tickets that allow last- minute changes and who desire frequent and conveniently scheduled flights. These passengers, who tend to be least sensitive to price, are often referred to as price-inelastic, or nondiscretionary, travelers. The higher fares paid by nondiscretionary travelers—usually business travelers—are a reflection in part of the higher costs associated with pro- viding last-minute service and ffights during peak travel periods. More generally, the unrestricted fares have helped large airlines to cover both the operating and capital costs of maintaining networks that provide the frequent and convenient service desired by time-sensitive travelers and to earn the long-term profits necessary to attract capital and sustain these networks. At the same time, fare restrictions have allowed air- lines to fill seats by offering leisure travelers lower fares that at least cover the incremental, or marginal, cost of the service. The public's discontent with the many restrictions imposed on lower- priced tickets and with the sharply higher prices charged for unrestricted tickets reflects in part a misunderstanding of the role of fare differences in allowing airlines to provide the kind of service desired by time-sensitive travelers. Yet it also might reflect a sense that the higher fares charged to nondiscretionary travelers in some markets exceed the cost of effi- ciently providing schedule-intensive service and are to some extent indicative of market power achieved partly through limits on competi- tive entry and possibly by the threat of predatory response. The spread between the lowest and highest fares has widened during the 1990s. In 1992, for instance, travelers paid about two times the median ticket price for the highest fares (in this case, the 90th percentile); but in 1998, they paid nearly three times the median. Some of this in- crease in fare dispersion might be attributable to cyclical effects, possibly due to disproportionately expanded demand by business travelers and to higher use of aircraft and seating capacity. Under these conditions, air- lines take risks in not selling some seats in anticipation of late-booking travelers or last-minute itinerary changes, since these seats—which might otherwise have sold at lower fares earlier—might fly unsold. The highest fares, which are increasing the spread, might reflect these costs. New, low-fare carriers, which have reduced fares in many markets, likely have

Executive contributed to an overall increase in the spread between the lowest and highest airline fares. To some observers, however, the widening spread in fares reflects the increasing ability of airlines to segment price-inelastic travelers, and to charge them exceptionally high fares in markets where pricing is not adequately disciplined by competition. Emergence of Airline Alliances and Other Partnerships Several major U.S. carriers recently have formed partnerships. Most of these involve sharing frequent-ffier programs, although some also coor- dinate ffights in connecting markets through codesharing. While only one domestic alliance so far has involved financial integration, it is rea- sonable to worry that domestic airlines in looser marketing relationships will be less inclined to compete vigorously with one another. The pos- sibility that these relationships will strengthen and migrate toward mergers—de facto, if not de jure—also raises concern. Even more alarming to some is the proliferation of alliances between U.S. and international carriers, often facilitated by grants of antitrust immunity to coordinate fares, seating capacity, schedules, and marketing. Although some travelers in connecting markets might benefit from these alliances, the potential gains to travelers in mainline markets—gateway- to-gateway routes where allied airlines were once main competitors—are not evident, and it is possible that these travelers are losing out. Moreover, the longer-term effects of these alliances may be exclusionary, ultimately forcing some unaffiliated U.S. airlines out of international markets by di- verting their feed traffic and weakening their overall route structure to the detriment of domestic competition. An issue that deserves explicit atten- tion is whether these expanding alliances are compatible with longer-range international aviation goals, such as unrestricted entry and competition by the most efficient carriers on a multilateral or global basis. Resurgence in Low-Fare Entry and Concerns About Unfair Competition About one in five passenger trips today is on an airline that can be char- acterized as primarily a "low fare" operator. Southwest Airlines alone accounts for about two-thirds of these travelers, and has played an impor-

6 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY tant role in the industrywide decline in average fares since deregulation, not only because of its own low prices but because it has spurred lower prices by other airlines, even compelling some to create special low-fare divisions. Other low-fare startup airlines have had mixed success. Undercapital- ization, poor choices of markets, inexperienced management, and unex- pectedly vigorous competitive responses have contributed to some failures. Increased market demand during the decade has raised the cost of air- craft and labor, possibly inhibiting entry and expansion by new airlines that want to pursue a low-cost, low-fare strategy. Concerns about the safety of startup airlines—particularly after the 1996 crash of a Valujet flight—also might have slowed expansion during the latter half of the 1990s. Nevertheless, nonincumbent airlines, led by Southwest, have entered nearly 2,000 markets (nonstop segments) during the past six years, and net market entries—that is, entries minus exits—have been positive on balance, exceeding 500. Market exits and failures among new airlines are not necessarily a cause for alarm, since a high rate of failure by new businesses character- izes many—if not most—industries. But in addition to facing the normal difficulties of running a new business, as well as the longstanding imped- iments to entry described earlier, new airlines also have encountered aggressive responses by incumbents. DOT has expressed concern that these responses sometimes have strayed beyond the bounds of fair com- petition, aiming instead at improperly excluding competitors. It there- fore has proposed a method for detecting and for enforcing prohibitions against unfair, exclusionary conduct in the industry (Appendix A). UNFAIR COMPETITION AND DOT's ENFORCEMENT PROPOSAL DOT forwarded to the study committee 32 complaints of unfair com- petitive conduct, fried by new entrants between 1993 and 1999. The com- mittee did not review each in detail; however, it is apparent that some of the actions described are difficult to reconcile with fair and efficient com- petition. About one-third reported chronic difficulties obtaining gates and other facilities at airports dominated by incumbents. About half in- volved sharp price cutting and increases in capacity by incumbents in re-

Executive Summary 7 sponse to entry. Some complained of incumbents offering higher travel agent commissions and bonus frequent-flier miles in contested markets, allegedly to divert enough potential customers to make the new service unprofitable. Among the most troubling, in the committee's opinion, were four reports of incumbents not only sharply lowering fares but also temporarily scheduling many more flights, some in city-pair markets in which they previously had not offered nonstop service nor jet flights. Sharp reductions in price and increases in capacity are predatory if designed to drive out or suppress competition to gain higher future prices and profits through increased market power. Some economists have pos- tulated that firms employ predatory tactics not only to strengthen or pre- serve their monopoly position in the markets in which they cut prices, but also to deter competitive entry in their other markets. Therefore, a valid concern is that airlines might engage in predation, even on a lim- ited basis, with the broader aim of dissuading entry and increasing mar- ket power throughout their networks. Distinguishing predatory behavior from the kind of competition that benefits consumers and then proving the distinction empirically, how- ever, are difficult—particularly in the airline industry with its frequent fare wars and constant shifts in city-pair service. A common test pre- sumes that predation occurs when a firm charges prices that are below marginal costs, so that incremental revenues generated from the sale of one more unit are less than the incremental costs incurred making the sale. Determining the applicable marginal cost of providing an airline passenger trip, however, is not easy. Airlines can enter and exit markets without incurring large, unrecoverable costs, because their assets are mo- bile, and they can lease aircraft, airport gates, and terminal space, as well as contract for ground services. Under these circumstances, even short- run average variable costs—an accounting measure often used as a prac- tical substitute for marginal cost—can be difficult to quantify retrospec- tively and at the applicable unit level. The distinctive characteristics and practices of the industry also have effects on the likelihood of successful predation and therefore on the probability of airlines employing predatory tactics. On the one hand, the relative ease of competitive entry or reentry can limit the ability of a predator to hold onto the monopoly position and recover the losses incurred from predation. On the other hand, the fluidity of airline assets

ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY and the consequent ability of incumbent carriers, when challenged by price-cutting rivals, to add and withdraw capacity quickly, can facilitate predation when there are significant barriers to competitive entry. In proposing an enforcement plan, DOT warned that it would investi- gate instances in which an incumbent airline, challenged by a new entrant, had responded by lowering fares and increasing seating capacity so that total revenue generated in the market is lower than would have been likely with a more "reasonable alternative esponse." According to DOT, a reasonable response would be for the incumbent to match the low-fare offerings of the new competitor on a restricted basis, without sharply increasing seating capacity. Apparently, DOT regards the rev- enue that the incumbent sacrifices by shifting capacity as an opportu- nity cost of predation. Opportunity cost—that is, the value of the best alternative response that is forgone—is relevant for assessing the costs of suspected preda- tory activity. However, DOT's proposal for detecting predatory behav- ior raises valid concerns about administrative feasibility and about the potential for undesirable consequences. Defining and specifring a class of new entrants for protection risks arbitrary enforcement, possibly favoring inefficient carriers. DOT's proposed criteria for detecting possible predation, which would require a comparison of revenues actu- ally earned by the incumbent with hypothetical revenues from other possible responses, are inherently speculative and therefore likely to be difficult to administer. In general, industry-specific administrative agencies are prone to rely on prescriptive regulation rather than on market forces to achieve their goals; in particular, they sometimes seek to protect their constituencies from competition perceived as destructive or predatory. There is a risk, therefore, that DOT's efforts will become increasingly regulatory and tend to inhibit, or even prohibit, some legitimate competitive responses. Moreover, policy and program objectives can change over time. Twenty years after economic deregulation, political influences on the provision and regulation of air transportation remain strong. Other experiences recounted in this report—the 30-year history of federal limits on access to some key airports or the recent antitrust immunity extended to inter- national airline alliances—suggest that the potential for unintended pol- icy outcomes should not be underestimated.

Executive Summaiy 9 Because of these risks, a case can be made that federal assessments of anticompetitive conduct should be entrusted primarily to the Department of Justice (DOJ), which has a clear mandate—as well as the resources and expertise—to enforce the nation's antitrust laws. DOJ also can levy signif- icant penalties through the criminal and civil courts. The committee rec- ognizes that DOJ has been actively exploring how antitrust laws apply to predation in the airline industry and has been developing approaches that take into account the industry's characteristics. In the committee's opin- ion, DOJ's involvement in this area is a healthy development. The committee also believes, however, that DOT has an important role in preserving and enhancing opportunities for competitive entry in the airline industry. This must entail concerted action—as recom- mended in this and other reports—to remove the persisting impedi- ments to entry that are under DOT's authority. In addition, DOT should ensure that airlines are not exploiting their advantageous rela- tionships with airports, air traffic control access, CRSs, and travel agents to hinder competition and to limit entry opportunities. The committee harbors reservations, however, about DOT's proposal for identifying and forbidding predation in the airline industry. Many members are concerned that DOT's proposal could become increasingly regulatory, thus inhibiting genuine competition, and they worry that it designates certain classes of carriers for special consideration. All com- mittee members believe more work is necessary to develop meaningful tests for detecting and proving predation—tests that facilitate enforce- ment and compliance but that also protect the competitive process rather than specific competitors. For these reasons, some members of the study committee would pre- fer that DOJ—which is unencumbered by industry regulatory responsi- bility and has greater antitrust expertise—take the lead in enforcement, as it did in a recent action against a major carrier. However, other members—while sharing the committee's general concerns about regula- tory risks—judge the problem serious enough to warrant the more active involvement of DOT, exercising its own independent enforcement authority to prevent unfair methods of competition. 'While these com- mittee members are uncertain about the administrative feasibility of DOT's proposed guidelines, they believe DOT should be given the opportunity to develop and apply objective tests for predation, ensuring

10 ENTRY AND COMPETITION IN THE U.S: AIRLINE INDUSTRY that markups on unrestricted fares are subject to a competitive discipline devoid of exclusionary practices. They are optimistic that DOT can do this without becoming overly regulatory and without inhibiting the kind of competitive price cutting that provides lasting fare reductions. In short, the committee believes DOT's proposal as currently formu- lated has flaws. Committee members differ, however, on the seriousness of the flaws and whether the flaws can be resolved by DOT. Despite differences on these issues, committee members are unani- mous in believing that DOT's main focus should be on expanding oppor- tunities for more entry and competition. Freedom of entry, including freedom from the threat of anticompetitive behavior, is made possible in large part by the removal of barriers to competitive entry and reently; this is the best antidote to excessive fares stemming from too much market power. Expanding capacity and ensuring efficient use of the nation's airports and airways is critical for this. Ensuring that structural develop- ments within the industry, both domestically and internationally, are favorable to more competition is also important. The following recom- mendations emphasize these goals. RECOMMENDATIONS The purpose of this study has been to examine the state of airline com- petition and to offer recommendations for furthering and safeguarding it. The main reason for caring about the competitive process is that vig- orous rivalry promises consumers more product and service choices at lower prices and with better quality. However, in addition to consumer interests, public policy must take into account many other considerations that could not be examined in this study. The committee's récommen- dations therefore must be understood as limited to the goal of preserv- ing and fostering competition in the airline industry. To the extent that these recommendations raise additional considerations, integrating and reconciling them with the goals of competition must be left to others. System Capacity and Opportunities for Competition Congestion and delays caused by the inefficient provision of airway and airport capacity affect not only the on-time performance of airlines, but

Executive Summary 11 also the routes airlines choose to fly, how they schedule and design their networks, and the types of equipment they use. Capacity shortages that persist, unresponsive to increasing demand, can limit new competition, particularly entry and expansion by low-cost carriers. To compensate for shortcomings in the way airport and airway capacity are provided—and in the absence of proper pricing—many administrative limits have been imposed. These include regulatory controls on airline use of some of the nation's most important airports; air traffic control procedures that require queuing for access to navigable airways and airports; and federal restric- tions on the ability of airports to raise and invest firnds for expansion. Not only have these inefficient allocation mechanisms adversely affected competition, but they have permitted the deferral or neglect of more efficient and direct means of supplying needed infrastructure. These regulatory controls are poor models for meeting future demands on the system. The sooner they are replaced and reformed, the sooner steps can be taken to provide airport and airway capacity more effectively and with fewer negative side effects. This also would allow for more direct pricing and for specific solutions to noise, congestion, and other prob- lems that these administrative schemes address only indirectly. Although the committee recognizes the existence of many practical difficulties asso- ciated with the introduction of new allocation methods—and the uncer- tainties that arise—it believes the following actions should begin as soon as possible: Apply federal and other funds to expand airport and airway capac- ity, particularly by investing in capacity-enhancing technology. The goal should be to use pricing both to finance expansion and to allocate capacity more efficiently. Both technology and pricing should be employed to encourage additional ffights to and from underused sec- ondaiy airports in major metropolitan areas. Introduce pricing methods in place of administrative restrictions to manage airline access to some of the country's major airports. The emphasis should be on the early substitution of pricing for current slot controls and perimeter limits on long-haul flights, with the goal of allocating scarce airport and airway space more efficiently and fairly among competing airlines and taking into account other technical and operational factors.

12 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY Ensure that federal rules governing airport funding and spend- ing do not conflict with—but instead help to achieve—the goal of increasing the availability of gates and needed infrastructure at major airports. These steps should allow more airlines to serve critical markets, encour- age more competition in cities with dominated hubs, and increase indus- try entry and competition in general. In this same spirit of increasing competition, the committee urges an end to federal limits on foreigners owning and operating U.S.-based airlines. The main rationale for restrictions on foreign ownership and investment is that other nations have them. Yet the main consequence is to limit the flow of capital and expertise into the domestic airline industry, possibly denying U.S. travelers the opportunity for more com- petitive services. Reluctance to ending such restrictions unilaterally is understandable, because the restrictions are enmeshed in international trade strategies aimed at prompting other countries to end their limits on foreign entry and trade. While the committee was not able to weigh the importance of these other factors, it believes this change in policy would benefit U.S. consumers on balance, even if other countries retain their more restrictive policies to the potential detriment of their own citizens. Airline Alliances and Partnerships In overseeing domestic airline alliances, DOT has authority to limit domestic codesharing and to impose conditions on other partnerships. The committee worries that codesharing and other collaborative arrangements among major domestic carriers will facilitate undesirable consolidation among current or potential rivals. To ensure an early and thorough evaluation of the effects of partnerships on competition, the committee recommends that all collaboration plans among major U.S. airlines be subject to traditional, economic-based merger analy- ses by DOJ, and that these plans—even if they do not involve exchanges of equity or transfers of assets—be subject to advance notification requirements similar to those in the Hart-Scott-Rodino process. (This process requires advance notification of certain acquisitions of stock and other assets).

Executive Summary 13 In the case of international aviation agreements, the committee rec- ognizes the complexities involved in opening markets governed by restrictive, bilateral aviation treaties. Relaxing bilateral restrictions is important; however, granting antitrust immunity to a few U.S. and foreign carriers engaged in global alliances might prove too high a price. Although they might benefit some travelers in the near term, the long-term exclusionary effects of the alliances might do more harm than good, by reducing rivalry in mainline international routes and by mak- ing it more difficult for unaffihiated carriers to compete. To ensure a critical review of these potential effects, the committee recommends a two-part process for reviewing and approving applications for an- titrust immunity by international airline alliances. DOJ should per- form the initial review and then forward to DOT only those applica- tions acceptable on competitive considerations. DOT should review these applications with respect to other issues of public interest and international policy. In addition, DOJ should perform follow-up cr1- tiques of immunized alliances approaching renewal. With its focused, procompetitive mission, DOJ is more likely than DOT to execute a thorough scrutiny of the competitive effects of anti- trust immunity. DOT's role, on the other hand, is better suited to exam- ining the broader and prospective effects of alliances on consumers and the industry over the long term. DOT would be more effective in this charge if it did not have to defend a position on specific alliances and re- quests for immunity. Airline Ticket Distribution System Travel agents—and the CRSs they use—provide an important service to consumers by making information available about the fare and service offerings of competing airlines. They also offer small airlines and new entrants access to a national network for marketing their services and distributing their tickets. Continued improvements to this system and the advent of new means of ticket distribution by airlines and agents— including Internet options—should be encouraged, since the potential gains from advances in distribution are so large. Nevertheless, ensuring and instilling impartiality in the system, however it evolves, should remain a priority for DOT.

14 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY DOT should monitor and investigate airlines' aggressive and se- lective use of travel agent incentives—as well as disincentives—to di- vert customers away from rival carriers. Consideration also should be given to developing rules for CRS listings to limit the kinds of code- shared ffights posted and to avoid conferring unfair competitive ad- vantages. This study focuses on some well understood and recognized opportuni- ties to encourage airline competition, especially in larger markets. Many smaller airline markets, however, are potentially promising—and needy—candidates for increased entry and competition, particularly by carriers offering new nonstop jet service. Some suggestions and ideas for further exploration are offered at the end of this report to foster more competitive services in these markets, which could serve as points of entry into the industry. Indeed, encouraging and providing opportuni- ties for communities to attract and retain airline services is both sensible and timely, as smaller, regional jets are being introduced. This is a case in which government actions might be appropriate to promote new air- line competition. In general, the committee strongly and unanimously urges positive steps that would encourage entry and competition in the entire industry. More emphasis should be placed on providing infrastructure capacity responsively and efficiently, without impeding legitimate competition. This approach must be accompanied by vigilant oversight of airline mar- keting and ticket distribution, making certain they are fundamentally fair and do not predispose the industry to further unhealthy consolidation. Some committee members are uncertain, and others are skeptical, about the prospects of using administrative procedures such as those proposed by DOT to police airline predation. All believe, however, that DOT's basic aim of preserving and expanding opportunities for competition should remain the principal goal of aviation economic policy.

V Overview and Trends The Transportation Research Board (TRB) study committee that pro- duced Winds of Change held its final meeting in the spring of 1991. The committee had reviewed the general experience of the U.S. airline in- dustry during the more than a dozen years since legislation ended gov- ernment economic regulation of entry, pricing, and ticket distribution in the domestic market.1 The committee examined issues ranging from passenger fares and service in small communities to aviation safety and the federal government's performance in accommodating the escalating demands on air traffic control. At the time, it was still being debated whether airline deregulation was favorable to consumers. Once viewed as contrary to the public interest,2 the vigorous airline competition 'The Airline Deregulation Act of 1978 was preceded by market-oriented administra- tive reforms adopted by the Civil Aeronautics Board (CAB) beginning in 1975. 2 Congress adopted the public utility form of regtilation for the airline industry when it created CAB, partly Out of concern that the small scale of the industry and number of willing entrants would lead to excessive competition and capacity, ultimately having neg- ative effects on service and perhaps leading to monopolies and having adverse effects on consumers in the end (Levine 1965; Meyer et al. 1959). 15

16 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY spurred by deregulation now is commonly credited with generating large and lasting public benefits. Since then, deregulation and related market- oriented reforms have spread to other transportation industries as well as to many international aviation markets. The Winds of Change committee concluded that the removal of eco- nomic restrictions on airlines had benefited consumers, both generally and in most individual communities and city-pair markets. Average fares had fallen and flight frequencies had increased as most of the established airlines realigned and expanded their networks, becoming more inno- vative, cost conscious, and responsive to traveler demands than they had been when the government protected them from the rigors of competi- tion. Yet the committee also identified several means to enhance com- petition. The specific steps recommended in Winds of Change are reprinted in Appendix B. SCOPE AND ORGANIZATION OF THE REPORT This report reviews the current state of competition in the airline indus- try and offers recommendations for increasing competition and oppor- tunities for entry. Because Winds of Change presents a comprehensive review of the history and rationale for deregulation and its implications for passenger fares, productivity, and operations—as well as safety—no attempt is made here to review experience before 1990 or to reexamine the soundness of deregulation as public policy. While concurring with the favorable conclusions about deregulation that were reached in Winds of Change —as well as in many other studies3—the committee observes that opportunities remain for ftirthering competition, and little progress has been made in exploiting them. More than most industries—and certainly more than any other trans- port mode—the airline industry is followed closely by the national media and policy makers; this close interest makes it difficult to take a prospec- tive view of developments in the industry and to develop a more strategic See Morrison and Winston (1995) for a study of the airline industry since deregula- tion, induding a listing of other studies.

Overview and Trends 17 approach to aviation policy. The committee therefore began by review- ing the major developments in commercial aviation during the past decade, induding updated analyses of the passenger fares, competition, and market-entry trends presented in Winds of Change. These analyses are discussed in this first chapter, along with a description of the most recent wave of startup airlines. In Chapter 2, the discussion turns to the specifics of airline compe- tition. One longstanding and prominent issue is the ability of major airlines to dominate local passenger traffic at the hub airports where they concentrate their flights. Airlines typically carry a large share of local passengers in the city-pair routes from their main hubs. New car- riers entering these markets with low-fare service have complained that incumbents respond by sharply lowering fares, expanding capacity, and employing other "exclusionary" tactics to drive out the challenger and inhibit further competition. These complaints—as well as the Depart- ment of Transportation's (DOT's) proposal to prohibit such anticom- petitive conduct—are examined in Chapter 2. Positive steps to expand and create market conditions conducive to more competition—and thus to limit the opportunities for airlines to obtain and exploit market power—is the emphasis in Chapter 3. The focus is on steps that would ensure sufficient airway and airport capac- ity, providing more opportunities for new entrants and allowing current carriers to expand their services competitively. The importance of free- ing the flow of both capital and airline expertise into the industry and of ensuring an impartial ticket distribution system is also considered in Chapter 3. The emergence of alliances and other partnerships among major air- lines, both domestically and internationally, is reviewed in Chapter 4. These developments—ranging from partially-merged, frequent-flier pro- grams to highly integrated partnerships sharing codes for flights—have become controversial because of their potential for furthering anticom- petitive consolidation within the industry. Examined are the rationales for alliances and the assertions that these arrangements are incompatible with the goal of fostering competition. In Chapter 5, the report concludes with additional suggestions and ideas that deserve further exploration as means to promote new airline services and entry in smaller markets.

18 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY GENERAL DEVELOPMENTS AND TRENDS DURING THE 1990s The airline industry has evolved in many unanticipated ways during the past decade. Concern has fluctuated widely over the industry's financial and competitive situation. Early in the decade, its fragile financial con- dition was a central issue and the subject of policy proposals—mostly short-lived—to return the industry to profitability.4 By mid-decade, alarm over weak profitability had faded, transforming instead into wor- ries that some airlines had gained too much market power and were exploiting it by raising fares well above the cost of service. Although the industry's average fares declined during the 1990s, the widening of the range of fares became a more prominent issue. Fares varied greatly by market and even among individual travelers seated in the same row on the same ffights. The 1990s began with turmoil in the industry. Higher jet fuel prices and traveler fears of terrorism caused by Iraq's invasion of Kuwait com- bined with a global recession and sharply expanded industry capacity to increase operating costs, dampen travel demand, and reduce aircraft load factors. Concern over the competitive effects from the demise of two long- distressed airlines, Eastern and Pan American, developed into general alarm over the weakened financial condition of the industry as a whole. These anxieties were intensified as several debt-burdened airlines—Con- tinental, TWA, and America West—filed for bankruptcy protection. The recession and resulting industry operating losses from 1990 to 1992 spurred proposals to ease debt and cash-flow burdens—for instance, by allowing airlines to defer remittance of federal ticket-tax revenues (GAO 1991). As travel demand recovered and airlines began experi- menting with higher fares, the emergence of new low-fare airlines gained notice. Many of these startup carriers used the equipment and labor shed by the large airlines during the recession and earlier failures. However, in 1996, the highly publicized crash of a Valujet airliner damaged the reputation of low-fare carriers, contributing to the financial failure of some and discouraging startups generally. instance, see the 1993 report of the National Commission on Airline Competition.

Overview and Trends 19 Even before this, some startup airlines had complained about what they perceived as intentionally injurious, or predatory, tactics by incum- bents. By the mid-1990s, alarm over the major airlines' financial condi- tion was superseded by concern that financially strengthened incumbents were actively seeking the demise of startups and systematically sup- pressing competition. Meanwhile, public discontent with airline service and pricing practices, particularly over the higher fares charged to busi- ness passengers traveling on unrestricted tickets, was rising. Enjoying dramatic growth in travel demand, the airlines were increasingly viewed as disinterested in, and perhaps even disdainful of, their main cus- tomers.5 VVhatever the merits of these concerns, consumers have continued to benefit overall from a deregulated industry, which still is characterized by significant price competition, including frequent fare wars that attract many bargain-seeking travelers. Adjusted for inflation, average fares decreased 25 percent from 1990 to 1998 (Figure 1-1). However, indus- try operating costs, as well as fares, have been largely stable since the middle of the decade, perhaps drifting slightly upward. Evident from the graph in Figure 1-1 is that most of the reductions in fares during the 1990s occurred early in the decade. Since 1995, trends have been rela- tively flat. But how have these changes in average fares compared with changes in underlying production costs? One commonly used index of operat- ing costs is DOT's Standard Industry Fare Level (SIFL), which de- picts changes in airline operating costs per available seat-mile. Changes in the SIFL, as shown in Figure 1-2, suggest that much of the overall decline in average fares during the 1990s was caused by declining jet fuel costs. The trends also show, however, that there has not been a significant divergence—or a growing gap—in fare levels relative to costs, suggesting that the gains from deregulation have not eroded. Some of these cost reductions might be due to the competitive pressures ushered in by deregulation, supplying further evidence of the policy's continuing benefits. See Murray, M. Airfares: Fare or Foul? NationalJournal, April 4, 1999, PP. 950-954.

75-500 miles 501-1000 miles II0011 500 miles 1501-2000 miles 2001 or more miles 45 40 35 0) 30 E 25 a, 0 20 15 10 5 0 Year and Quarter NOTE: Average yield per passenger trip (average fare/miles flown) was calculated using the DOT 10 percent ticket sample accessed through Database Products. Figures were adjusted using the gross domestic product (GDP) price deflator. Trips with zero fares were excluded. (1 mile = 1.61 kilometers) Figure 1-1 Average yield for domestic air travel by mileage block, 1990 to 1998 (2nd quarter).

140 120 100 80 60 40 20 0 S'2 •? Month-Year Figure 1-2 Trends in Standard Industry Fare Level (SIFL) Index, 1990 to 1998.

22 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY As the decade of the 1990s progressed, airlines refined their ability to charge different fares to different groups of travelers: this evidently widened the spread in fares paid. Unable to take advantage of the heav- ily restricted low-price tickets, time-sensitive business travelers have increasingly expressed concern that the fares they are being charged are far above the cost of providing the service in an efficient manner. In this report, as in most other studies, reference often is made to "average fares," and average fare data are used in many of the data analy- ses. As discussed in the next section, however, the airlines serve two dis- tinct types of customers—business travelers and leisure travelers. Air- lines have become skilled at distinguishing these passengers and charging them widely differing fares by imposing purchase restrictions. Thus, in addition to average fares, the actual fares paid, sorted by re- striction type, would be helpful in examining trends. However, it is un- clear how such data could be gathered and analyzed. "Restricted" or "discounted" fares are terms that cannot be defined easily or uniformly. In an industry in which buyers are paying widely differing prices—and not necessarily for the same product—market averages are imperfect but still remain the most informative and reliable price measure available. PRICE DISPERSION IN THE AIRLINE INDUSTRY Fare dispersion has grown since deregulation, continuing into the past decade. In this section, trends in fare dispersion in the airline industry and their possible causes are discussed. There are two general reasons why buyers in any industry properly might be charged different prices for products that appear to be the same or nearly the same—these reasons are both efficiency- and welfare- enhancing. The first reason is that there is a difference in the direct costs of supplying the products to different groups of customers—such as the cost of delivering the product to the point of sale. The second reason is that there are significant economies of scale and scope in the products' supply, even though the products and the costs of supplying them are ef- fectively the same. 'V\Then a firm sells the same or almost identical prod- ucts at different prices to different buyers, or at different markups to the marginal cost, price discrimination has occurred.

Overview and Trends 23 In this circumstance, prices set uniformly at their marginal costs would not recover the total costs of supplying the products, including fixed and overhead expenses; and it then becomes necessary to charge buyers prices in excess of incremental cost, if the products are to continue to be supplied without external subsidy. In this situation, the seller can take advantage of buyers attaching widely different values to the same product, if the seller can sort the buyers on the basis of their different product valuations. The seller then can charge prices that are above mar- ginal cost to buyers who are least sensitive to price. This allows the seller to generate sufficient revenue to provide the product, but no more than that if competition is effective. A more general discussion of price dif- ferentiation is provided in the accompanying sidebar (Box 1-1). As discussed in Chapter 2, the major airlines maintain that fare dif- ferentials—that is, variations in fares between business and leisure trav- elers—are caused by direct differences in cost, such as those associated with flying during peak or congested periods. Unfortunately, the air- lines have offered little empirical evidence to support this explanation, perhaps because many shared costs make cost allocations among classes of service highly arbitrary. Evidence of growing fare dispersion also has not been accompanied by evidence of growing cost dispersion. Nevertheless, a substantial portion of the fare differences is explicable in terms of the di- rect costs of the two kinds of services. Certainly, the economies gained from a route's density—for example, denser routes permit larger planes, with correspondingly lower costs per seat mile—suggest that fares typi- cally will be lower in dense markets than in thin markets. Differing con- straints on airport and airway capacity, including terminal charges, also can contribute to differences in costs, and therefore to variations in fares across markets. Likewise, fares paid by travelers in the same markets can vary widely because of differences in the cost of traveling at different times of the day and week. For instance, higher fares should be expected for travel during peak times when demand is greatest and resources are tight (Gale and Holmes 1993; Borenstein and Rose 1994). Travelers booking early are less costly to serve than travelers booking much later, because holding unsold seats in anticipation of late-booking travelers in- creases the risk that the seats will fly empty. Business travelers pay higher fares in part because they tend to be late bookers; however, leisure trav- elers, who pay the lowest fares, tend to be early bookers. When demand

24 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY Box 1-1 V Ramsey Pricing Economists generally maintain that prices set equal to the mar- ginal costs of producing goods or services are optimal—that is, they are "welfare maximizing." However, problems can arise if the marginal costs are less than the average costs for the pro- ducer at levels of output that would prevail if prices were set uni- formly for all buyers. In this situation, the revenue from mar- ginal cost pricing would not cover the total costs and the producer would not be able to continue to provide the product— the product would not pay enough to replace equipment and other capital in the long term. The loss of the product in turn can create welfare losses, particularly for consumers who have valued it highly. A well-known theorem in economics, known as Ramsey pricing, after one of its earliest proponents, holds that the wel- fare losses created by deviating from marginal cost pricing are minimized if buyers with inelastic demands are charged the highest prices or markups above the marginal cost. In this way, the overhead, fixed, and other costs—which make up total costs but are otherwise excluded from marginal cost—can be in- cluded in the prices charged to buyers with inelastic demand. This minimizes the deviation in consumption patterns from what would have occurred if prices were set uniformly at the marginal cost. The airline industry's pricing practices are similar to Ramsey pricing. Leisure travelers who are sensitive to price (i.e., are price-elastic) are charged fares roughly equal to the marginal cost of serving them; but in the aggregate these are insuffi- cient to cover the ftill cost of maintaining the airline networks. U.S. airlines, competing for capital in private financial mar- kets, must cover their total costs. As Ramsey would suggest, the airlines can cover any shortfall in revenues from leisure tray-

Overview and Trends 25 Box 1-1 continued elers by charging higher prices—or whatever the traffic will bear—to price-inelastic, business travelers. In this way, the air- lines are able to finance a more extensive network than would have been possible otherwise. Of course, customers who make the extra contributions are seldom content with this role. Some business travelers believe they are being held captive to their inelastic demands because of anticompetitive actions taken by the airlines, such as holding back new entrants or new technologies—for example, smaller, re- gional jets. A traditional role of economic regulation, particu- larly in telecommunications and transportation, has been to smooth out the surcharges by making them uniform over broad classes of customers, and therefore more politically acceptable. However this is usually achieved with some sacrifice in welfare, by decreasing the markups to the most inelastic buyers while in- creasing them for buyers who are more sensitive to price. The politically acceptable solution usually has led to consumption patterns with greater deviation from the marginal cost opti- mum than would occur under a strict Ramsey solution. Although Ramsey pricing can be beneficial in the short term, it might be less desirable in the long term. The Ramsey opti- mum is static. Unless prices adjust quickly or producers have considerable foresight, the Ramsey optimum can be under- mined by any changes in the demand elasticities and other con- ditions that define it. Moreover, Ramsey pricing creates its own strong incentives for change. Buyers paying the highest prices, or markups, have incentives to seek alternatives to the high- priced service. Over time, they will find alternatives, or make ad- justments to their circumstances to rely less on the product; this would make much of the existing capacity redundant for some period of time or until the industry can adjust capacity properly. New technology and other changes and alternatives—in- cluding new competitors—that undermine an established Ram- continued

26 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY Box 1-1 continued sey pricing pattern are not welcome to the incumbent produc- ers. In the long term, sustaining such a pricing scheme usually requires government regulation or monopoly power to bar entry. Producers who price under a Ramsey scheme are likely to be hostile to change, even if the change is beneficial to consumers. Some of the most counterproductive experiences with industry economic regulation were attempts to maintain discriminatory pricing by delaying technological changes that otherwise were desirable (Geilman 1971). Price discrimination therefore should be allowed only if it is tested by open entry and competition, as recommended in this report. for seats is high, the cost of holding an empty seat in anticipation of late- booking or "walk-up" travelers—that is, the opportunity cost if the seat is not sold—can be high. This cost in turn can be reflected in higher fares for travelers who book late or who make last-minute itinerary changes. Although much of the observed spread in airline fares can be ex- plained by observable cost differences, not all can, and certainly some of the spread seems to be the result of price discrimination. Airlines have long been able to sort travelers according to their relative price elastic- ities by imposing various ticket restrictions. The types of restrictions are discussed in more detail below; the main point is that the general ability of airlines to price-discriminate may be advantageous to travelers, at least in the short run. The argument that some price discrimination in the airline industry can be desirable rests on the recognition that the prod- uct demanded by schedule-sensitive business travelers differs signifi- cantly from the product demanded by price-sensitive leisure travelers. The difference in preferences between the two kinds of travelers may be the greatest on short-haul flights. In these markets, leisure travelers would be expected to be sensitive to price because driving is a travel option. Meanwhile, flight frequency may be especially important to business travelers, since saving relatively small amounts of time is a main reason

Overview and Trends 27 for the decision to fly rather than drive. By being able to identif' leisure travelers through ticket restrictions, the price-discriminating airline can offer discounted fares to fill unsold seats on flights that might oth- erwise fly partially empty; at the same time, it does not permit business travelers to take advantage of these lower fares. 'VVithout the ability to restrict access to these low-fare tickets by time- sensitive business travelers, the airline might not be able to cover the total cost of providing frequent and extensive service. For example, at a hub, airline operations realize large economies of scale and scope, lowering av- erage costs; moreover, the number of routes that can be served increases disproportionately with an increase in the number of scheduled flights. It is possible that hub fares set uniformly at marginal cost would gener- ate revenue insufficient to recover total costs, necessitating discrimina- tory pricing if the hub airline wants to maintain its service. Price dis- crimination allows carriers to cover both the operating and capital costs of providing the schedule-intensive service desired by business travelers, while filling empty seats with leisure travelers whose low fares at least cover their incremental, or marginal, cost. However, the possibility that airlines have become too skilled at iden- tifying price-inelastic travelers and too eager to charge them excessive fares—above the level necessary to efficiently provide the service— has become an issue. One indicator of this would be that the price- discriminating airline is reaping monopoly profits—that is, excessive returns on capital. But this is difficult to determine, because it involves projections of airline profits and the effects of the business cycle. A pro- tection against this kind of exploitation is free entry. If entry is not arti- ficially impeded, competing services will ensure that the fares charged to price-inelastic travelers reflect, in the long run, the full or stand-alone cost of efficiently providing the service desired. As some uniformly low-fare carriers, such as Southwest Airlines, have discovered, business travelers are not completely insensitive to price and are sometimes willing to accept fewer schedule options in re- turn for substantial fare reductions. In those dense markets in which air- lines are charging excessive fares to price-inelastic travelers, it is pre- cisely this kind of nonnetwork, point-to-point, low-fare entry that might be most expected, especially given the high overhead and fixed

28 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY costs associated with creating a brand new, competing hub-and-spoke system. As noted later in this chapter, however, new low-fare service has tended to focus on dense markets only, where point-to-point ser- vice can be economical; service coverage in many thinner markets— often important to business travelers—remains the domain of the larger carriers with their larger and higher-fixed-cost networks. The intro- duction of smaller regional jets and other new technologies might change this situation, as noted later. Fare Restrictions as Means of Differential Pricing In their formative years, regulated airlines concentrated on business trav- elers. Unscheduled charter service was permitted for some low-fare travel (e.g., tour packages), but in general the price-sensitive, leisure travel mar- ket was neglected (Douglas and Miller 1974; Meyer and Oster 1984, 3). But by the mid-1960s, regulated airlines were increasing their use of dis- count coach fares to attract more leisure passengers to fill seats on their new larger-capacity jet aircraft. Senior citizens, students, and family of travelers flying on fttll fares were offered discounts of up to 25 percent. Using such differential methods of pricing, airlines could identif' and attract price-sensitive traffic without marking down the fares charged to regular business travelers. The airlines maintained—and the CAB reg- ulators presumed—that such discriminatory pricing was needed to en- sure the long-run profitability of the service. Though seldom able to pur- chase the discounted fares, business travelers were thought to have benefited because more frequent ffights were available in larger aircraft. Some economists predicted that deregulation would diminish air- lines' ability to differentiate among types of travelers, because of a gen- eral belief that monopoly power (or regulatory protection) was required for discrimination and a lack of understanding of the economies of scale and scope that would come to characterize a post-deregulation industry. According to this view, new and incumbent airlines would tai- lor their services to particular kinds of travelers. For example, single-class, service-intensive airlines would cater to business travelers by providing more spacious seating, frequent departure times, and generous in-flight amenities, but charging uniformly high fares to cover the cost of this pre- mium service. Meanwhile, no-frills airlines would emerge, offering off-

Overview and Trends 29 peak, low-fare services for mostly leisure passengers. Under these sce- narios, the variation in fares paid by travelers seated side-by-side on the same flight would be small. Although some new single-class carriers did emerge after deregulation, the incumbent airlines quickly realized that they could efficiently serve both business and leisure travelers on the same flights.6 Through hub- and-spoke networks and the economies of consolidating traffic at central hub airports, airlines could increase the number of scheduled flights from previous point-to-point schedules. The large increase in flight destina- tions and frequencies (i.e., the scope economies) were a boon to business travelers, particularly those traveling to cities with hub airports. Hub cities such as Charlotte, Cincinnati, and Detroit have experienced a 50 percent or more increase in scheduled jet departures sine 1985. Airlines recognized also that leisure travelers—self-identified by their willingness to accept booking restrictions—could be accommodated on many of the same flights as time-sensitive travelers. Seats that otherwise would fly empty could be sold at a significant discount, yet still cover incremental, or marginal, costs. Startup airlines focusing on leisure traffic—but unable to attract business travelers seeking frequent and convenient flights to numerous points—were at a significant disadvan- tage. Many failed within the first few years after deregulation. Moreover, it soon became evident that certain service amenities could be offered se- lectively to high-fare business travelers—for instance, accrual of free va- cation trips, access to airport club lounges, upgrades to first dass, and special boarding privileges. Since deregulation, airlines have fine-tuned their fares using a complex combination of purchase terms such as Saturday-night stay-over restric- tions, advance purchase requirements, and penalties for cancellations and exchanges. Yield-management practices—in which airlines reassess on a flight-by-flight basis how many seats are to be offered at a discount, by how much, and when—have proliferated and become more sophisti- cated with experience and advances in information technology. At any 6 An early recognition of this possibility appeared when regulated airlines requested approval to offer "part charters" on regularly scheduled flights. These figures were derived from a review of DOT traffic schedules.

30 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY one time, an airline might be offering a dozen or more different coach fares on a given ffight. Frequent—ffier programs, now offered by nearly all airlines, also identifj the most loyal customers—who generally are the most price-inelastic. These travelers receive additional bonus miles and other benefits, particularly when purchasing high-fare tickets. Airport security requirements that travelers must present personal identification before boarding also have enhanced the ability of airlines to price- discriminate, as low-fare travelers cannot resell their heavily restricted advance-purchase tickets (e.g., through intermediaries) to late-booking travelers, who must then pay the higher fares. Trends in Fare Dispersion Some insights into recent trends in fare dispersion can be gained by com- paring the median fare paid by passengers in short-, medium-, and long- haul markets in 1992, 1995, and 1998. As shown in Table 1-1, the median fare in these three aggregate market groupings, based on market dis- tance, has declined by about 15 to 20 percent over this 6-year period. The lowest fares (that is, the 10th percentile fares) also have fallen slightly in real terms—by about 5 percent. The higher-fare travelers (the 90th percentile), however, are now paying 5 to 25 percent more. Also evi- dent is that these travelers are paying fares much higher than the median, at least in comparison with earlier periods (1995 and 1992). For instance, travelers paying the highest fares in 1992 paid 2 to 2.1 times the median fare. In 1998, these high-fare travelers paid 2.7 to 2.9 times the median. It is possible that these increasing ratios are a misleading statistical artifact of increased low-fare service generally, which drives down the median fare and makes it smaller relative to the highest fares. Actually, the highest fares have risen in real terms and the medians have declined. As shown in Table 1-2, the top-fare travelers (95th percentile and above) accounted for 17 to 18 percent of airline revenue in 1998, compared with 8 to 13 percent in 1992. The effect of Southwest Airlines on such aggregate fare data can be seen in Table 1-3. Although Southwest charges different fares to trav- elers based on certain demand and cost characteristics—by offering dis- counts for advance purchases—it does not vary fares to the same extent

Table 1-1 Distribution of Fares in Short-, Medium, and Long-Haul Markets, 2nd Quarters of 1992, 1995, and 1998 (adjusted to 1998 dollars) 10th Percentile Fare ($) 25th Percentile Fare(S) 50th (Median) Percentile Fare(S) 75th Percentile Fare(S) 90th Percentile Fare (5) 95th Percentile Fare (5) 1998 Short Haul 45 64 96 164 260 316 (0.47 x median) (0.67 x median) (1.7 x median) (2.7 x median) (3.3 x median) Medium Haul 76 95 126 177 339 469 (0.60 x median) (0.75 x median) (1.4 x median) (2.7 x median) (3.7 x median) Long Haul 98 135 176 256 517 705 (0.56 x median) (0.77 x median) (1.5 x median) (2.9 x median) (4.0 x median) 1995 Short Haul 47 71 101 164 231 291 (0.46 x median) (0.71 x median) (1.6 x median) (2.3 x median) (2.9 x median) Medium Haul 94 112 145 217 364 437 (0.65 x median) (0.77 x median) (1.5 x median) (2.5 x median) (3.0 x median) Long Haul 130 160 200 278 512 662 (0.65 x median) (0.80 x median) (1.4 x median) (2.6 x median) (3.3 x median) 1992 Short Haul 50 78 114 173 227 287 (0.44 x median) (0.68 x median) (1.5 x median) (2.0 x median) (2.5 x median) Medium Haul 78 107 156 228 327 379 (0.50 x median) (0.68 x median) (1.5 x median) (2.1 x median) (2.5 x median) Long Haul 103 137 204 294 420 486 (0.S0xmedian) (0.6lxmedian) (I.4xmedian) (2.1 xmedian) (2.4xmedian) NoTEs: Fares were adjusted for inflation using the GDP price deflator. Fare data are from DOT's sample of every 10th ticket, computed by Database Products, Inc. Median fares were calculated based on all passenger traffic in each mileage block, excluding passengers with fares below $10. Other filters were applied to eliminate unusually high fares that were probably miscoded. Percentiles should be interpreted as the following example illustrates: travelers paying the "75th percentile fare" are paying fares higher than 75 percent of all travelers in the mileage block. The second quarters (May-June) were selected for each year because they avoid summer peaks and winter troughs, as well as major holiday travel periods. Short haul = 100 to 750 mi; medium haul = 751 to 1500 mi; long haul >1500 mi.

32 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY Table 1-2 Share of Total Airline Revenues Derived from Passengers Paying Lowest to Highest Fares for Short-, Medium-, and Long-Haul Markets, Second Quarters 1992, 1995, and 1998 Percentage Share of Total Industry Revenues Derived from: Passengers Paying Lowest Fares Passengers Paying Highest Fares 1st to 25th 26th to 50th 51st to 75th 75th to 95th >95th Percentile Percentile Percentile Percentile Percentile 1992 2nd Quarter Short Haul 14 19 30 29 8 Medium Haul 12 20 24 30 13 Long Haul 13 20 24 30 13 1995 2nd Quarter Short Haul 15 18 25 32 10 Medium Haul 14 18 23 32 13 Long Haul 14 18 23 31 14 1998 2nd Quarter Short Haul 10 27 24 21 18 Medium Haul 13 16 21 34 17 Long Haul 12 16 21 34 17 Note: See Table 1-1 for explanation of data sources, calculations, and definitions. as most other large airlines. High-fare travelers (90th percentile) on short-haul Southwest flights pay about twice the median fare for all Southwest passengers in those markets. By contrast, high-fare travelers paid three times the median on other larger incumbent carriers (Amer- ican, Continental, Delta, Northwest, TWA, United, and US Airways) examined collectively. Part of this variation is attributable to a statistical distortion caused by combining traffic from several airlines, encompass- ing a wide mix of qualitative service differences and market demand and

Table 1-3 Distribution of Fares in Short- and Medium-Haul Markets for Southwest Airlines and Incumbent Carriers, 2nd Quarters of 1992, 1995, and 1998 (adjusted to 1998 dollars) 25th Percentile 50th (Median) 75th Percentile 90th Percentile -- - 95th Percentile Fare ($) Percentile Fare ($) Fare ($) Fare ($) Fare ($) 1998 Short Haul Southwest 48 63 85 114 122 (0.76 x median) (1.3 x median) (1.8 x median) (1.9 x median) lncumbents 71 114 223 353 456 (0.62 x median) (2.0 x median) (3.1 x median) (4.0 x median) Medium Hall Southwest 93 104 126 148 163 (0.89 x median) (1.2 x median) (1.4 x median) (1.6 x median) Incumbents 100 131 190 400 522 (0.76 x median) (1.5 x median) (3.1 x median) (4.0 x median) 1995 Short Haul Southwest 48 77 82 117 130 (0.61x median) (lix median) (l.Sxmedian) (l.7xmedian) .lncumbents 88 140 240 349 428 (0.62 x median) (1.7 x median) (2.5 x median) (3.0 x median) Medium Haul Southwest 101 116 138 161 186 (0.88 x median) (1.2 x median) (1.4 x median) (1.6 x median) lncumbents 124 162 250 - 424 507 (0.76 x median) (1.5 x median) (2.6 x median) (3.1 x median) 1992 Short Haul Southwest 50 63 87 121 133 (0.79 x median) (1.4 x median) (1.9 x median) (2.1 x median) lncumbents 101 160 247 349 401 (0.63 x median) (1.5 x median) (2.2 x median) (2.5 x median) Medium Haul Southwest 107 126 154 200 219 (0.85 x median) (1.2 x median) (1.5 x median) (1.7 x median) lncumbents 129 184 272 386 451 (0.70 x median) (1.5 x median) (2.1 x median) (2.5 x median) Nois See Table 14 for explanation of data sources, calculations, and definitions. Incumbents Consist of American, Continental, Delta, Northwest, TWA, United, and US Airways. Long-haul markets were not considered because Southwest carries few passengers beyond 1,500 mi. Fares are adjusted to 1998 dollars using the GDP price deflator.

34 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY cost conditions, yielding a wider range of fares.' Still, a marked difference in the fare structure of Southwest Airlines and US Airways, two carriers of similar size and scope, is evident in Figure 1-3. While most short-haul travelers on Southwest Airlines pay fares close to the average, the differ- ential is much greater for low- and high-fare travelers on US Airways. A more precise means of measuring trends in fare dispersion is to use the Gini coefficient, first applied to airline fare analysis by Borenstein and Rose (1994). A Gini coefficient encompasses the entire distribution of fares, rather than specific percentiles. A Gini coefficient of zero indicates that all fares are the same; as it moves closer to 1, the fare distribution is more dispersed.9 Figure 1-4 shows the Gini coefficients for fares since deregulation. The three trend lines represent different origin—destination (O-D) comparison groups: "O-D routes," comprising all tickets for all routes for all carriers, regardless of the number of connections (or seg- ments); "O-D routes by segments" breaking down O-D traffic by the number of segments or nonstop routes (separating travelers on nonstop and connecting flights); and "O-D routes by segments and carriers" focusing on the dispersion at the carrier and segment levels (e.g, Amer- ican nonstop itineraries, Continental one-stop itineraries). As the mar- ket definitions become narrower, the dispersion is reduced, suggesting that some fare dispersion at the O-D level is caused by individual carrier and segment differences. As would be expected, the Gini coefficients are much higher today than in 1978, when CAB limited the amount of fare variation generally and on individual routes. A comparison of 1990 and 1998 levels shows 8 Airport congestion, for instance, has been found to correlate with more fare dispersion, as might be expected from peak-load pricing (Borenstein and Rose 1994). As an exam- ple, Southwest schedules flights to avoid peak activity and heavily congested airports. In addition, because of variations in airport expenses and other site-specific costs (such as terminal charges) average fares vary more among short-haul markets than among long- haul markets. In long-haul markets, fuel and labor expenses, which are less variable across markets, account for a higher portion of overall costs. Perfect equality is when 50 percent of the travelers account for 50 percent of the fare revenue. To illustrate, a Gini coefficient of 0.2 indicates that two fares drawn at random would vary by 40 percent (2 x 0.2) from the mean in the market, however defined. Thus, in a market with an average fare of$100 and Gini coefficient of 0.2, the expected absolute spread in two randomly drawn fares would be $40.

1,400,000 1,200,000 1,000,000 12 800,000 0 C a, U) C') 600,000 400,000 200,000 F, 6'0 'o Do Percentage of Average Fare NOTE: Short-haul trips are 750 miles or less. See Figure 1-1 for data sources. Average fares for Southwest and US Airways were $64 and $170, respectively. Figure 1-3 Fare dispersion for short-haul trips on Southwest Airlines and US Airways, related to average fare for each carrier, 2nd quarter of 1998.

0.2€ 0.2 O.2 0.2( 0.1€ 0.1€ 0.14 0.1 0.1 75 79 50 51 52 Wi 54 50 sb 151 IJU UW WU UI Year NOTE: Calculations by Steven Morrison. See text for description of Gini coefficient and its application to 0-0 routes by segment and carrier. Figure 1-4 Fare dispersion in U.S. domestic airline markets using Gim coefficient.

Overview and Trends 37 that the Gini coefficient has not changed significantly. Although there was a move upward from 1995 to 1997, these data do not, by themselves, offer evidence of airlines enjoying greater ability to price-discriminate during the 1990s. The discrepancy between the Gini coefficient and the median—fare analyses presented earlier needs to be reconciled, although time did not permit further evaluation in this study. Other Factors Contributing to Fare Dispersion The airline industry has been prone to wide cyclical swings. As dis- cussed at length in Winds of Change, excess capacity has been a recur- rent problem for the industry during and shortly after recessions. Fig- ure 1-5 shows industry trends in revenue passenger-miles and available seat-miles during the 1990s. The gap between the two graphs indicates empty seats, or underutilized capacity. Excess seat capacity has de- clined during the decade, as indicated by average load factors, which rose from 60 percent in 1990 to nearly 70 percent in 1998. Low load factors early in the decade reflect the drop-off in demand during the 1990 to 1991 recession, coupled with expanded capacity. More re- cently, higher load factors reflect an industry with high demand and intensive use of capacity. In the past—notably in the early 1990s—whenever failing airlines at- tempted to generate revenues from underused capacity by sharply lowering fares, other airlines followed suit. The resulting fare wars benefited travel- ers in the short term. However, for well-run private airlines to survive and prosper, fares eventually must rise to levels sufficient to recover long-term costs and keep capital in the industry. As shown in Figure 1-6, incumbent carriers experienced large losses during the early 1990s recession, but have since experienced positive operating profits, often at record levels. The increases in unrestricted fares observed in recent years have been justified at least in part as restoring the industry's total returns from the large losses it suffered in the early 1990s. This committee lacked the time and resources to make a comprehensive assessment of the industry's fi- nancial performance—both currently and historically—that would be necessary to reach a definitive judgment on this. Although observed in- creases in the fares charged to price-inelastic travelers might be evidence that airlines are exercising their market power, it is unclear that these fare

Billions of Miles 800 700 600 500 400 300 200 100 0 Load Factor Percentage 67.4 68.9 58.9 621 550 550 554 4 419 41 393 ASM RPM Load Factor ou 70 60 50 40 30 20 10 0 90 91 92 93 94 95 96 97 Year Source: FM 1998, IX-16. Figure 1-5 Trends in annual seat-miles (ASM), revenue passenger-miles (RPM), and load factors for all domestic jet carriers, 1990-1997.

5 4 3 2 1 > CL -2 -3 -4 -5 "vu 000 500 000 500 0 500 000 500 0perating profits 000 ____________________________________ Annualized GDP Growth Rate 19& ? 7 \? '7 '? 7 I 7 .? 7 '.1 19k7 Year and Quarter NarE: Incumbents consist of American, Continental, Delta, Eastern, Northwest, Pan Am, TWA, United, and US Airways. SOURCE: Dresner and Windle 1999. Figure 1-6 Operating profits of incumbent carriers (1998 dollars).

40 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY increases have been excessive, exploiting market power. Perhaps the only check against such exploitation is to ensure that opportunities for com- petitive entry are made available and not impeded. MARKET AND INDUSTRY ENTRY TRENDS There are several ways to examine entry activity in the airline industry. Because carriers compete for passengers in individual airport-pair markets—discussed in the next chapter—entry into these markets is of particular interest. An entrant in an airport- or city-pair market can be a new airline offering service on the route for the first time or an incum- bent expanding operations. To assess entry trends and influences, the study committee commissioned a paper on resource availability and new entry in the domestic airline industry from Martin Dresner and Robert Windle of the University of Maryland Robert H. Smith School of Busi- ness.10 Their paper serves as the basis for the following discussion on recent entry trends. Trends in Market Entiy Activity Focusing on airlines operating jet aircraft with 70 seats or more,11 Dresner and Windle examined both entry and exit patterns at the individual route level (segment or nonstop airport-pair market) for the period 1989 to 1998. As shown in Figure 1-7, the rate of carriers beginning service on nonstop routes12 was relatively stable throughout most of the 1990s, fluc- tuating between 100 and 200 entries per quarter. Route-level entry, however, began to decline in 1996. A notable subtrend was that quar- 10 The commissioned paper (Dresner and Windle 1999) is available from TRB or from the authors (Robert H. Smith School of Business, University of Maiyland, College Park, MD 20742). 11 This distinction was made to ensure consistency with this study, which focuses on competition among airlines operating large jet aircraft. Commuter airlines, which oper- ate smaller aircraft, are generally viewed as offering services that complement rather than compete with, those of larger commercial airlines. 12 Determined from DOT's traffic schedules, and excluding observations with 10 or fewer operations during the quarter.

Overview and Trends 41 terly entries by incumbent carriers (American, Continental, Delta, Northwest, TWA, United, and US Airways) had been falling through- out the decade. Entries by other, nonincumbent carriers—influenced by Southwest—generally increased during each quarter in the first half of the 1990s, as shown in Figure 1-8. Yet starting in 1996, entry activity by new carriers also began to decline. A review of total net entries—route exits subtracted from entries— shows negative figures in most quarters during the 1990s (with a large negative in late 1990, following the demise of Eastern and Pan Am— see Figure 1-9). Net entries by nonincumbents were positive in nearly all quarters from 1992 to 1996; however, since 1996 these carriers have tended to exit more routes than they have entered. Still, these airlines, led by Southwest, have entered nearly 2,000 markets (nonstop seg- ments) during the past six years, and net market entries—that is, en- tries minus exits—have been positive on balance, exceeding 500. Explanations for these entry trends vary. The notable decline in entries beginning in 1996 has led many observers to conclude that a primary cause was the Florida crash of a Valujet DC-9 in May of that year.13 Valujet's suspension and subsequent renewal of operations following the crash undoubtedly influenced the total number of market entries and exits. Negative public reactions to the crash, as well as the uncertainties of plans by the Federal Aviation Administration (FAA) to increase over- sight of newly certified airlines,14 also are viewed by some as adversely affecting the financial performance of other low-fare airlines, reducing their access to capital, and hindering their ability to expand.15 As dis- cussed later in this chapter, the number of brand-new carriers entering the industry, as well as new applications for certification, declined fol- lowing the crash. 13 On May11, 1996, Valujet flight 592 crashed in the Florida Everglades. All 110 peo- ple onboard were killed. On June 18, Valujet suspended its operations. With FAA ap- proval, Valujet restarted in late September 1996. 14 Following a 90-day review of its safety program, FAA announced its intention to heighten surveillance of newly certffied airlines (see GAO 1997). 15 For example, Nethercutt and Pruitt (1997) show statistically significant losses in the stock portfolios of low-cost airlines following the Valujet crash.

300 All Carriers lncumbent Garners - - - 4 quarter moving average 250 200 150 100 50 *04 - - - 0 % % % Year and Quarter NOTE: A nonstop route is a flight segment operated by a carrier. See Figure 1-6 for list of Incumbents. Source of data is DOT traffic schedules. Calculations by Database Products, Inc. SOURCE: Dresner and Windle 1999. Figure 1-7 Quarterly counts of new nonstop routes served by carriers using jet equipment with at least 70 seats.

140 120 100 80 60 40 20 0 S I IL I t 'I II I V I I' I It 1L Il4s j4_1 I'! 5- %I k ' * - - - 4 quarter moving average , ,', 'y 'V. \) ,D Year and Quarter NOTE: See Figure 1-7 for definitions and data sources. SOURCE: Dresner and Windie 1999 Figure 1-8 Quarterly counts of new nonstop routes served by nonincumbent carriers using jet equipment with at least 70 seats.

200 150 100 50 0 -50 .100 150 .200 .250 Year and Quarter NOTE: Net entries equals the number of new routes entered minus the number of routes exited. See FIgure 1-7 for definitions and data sources. SOURCE: Dresner and Windle 1999 Figure 1-9 Quarterly net new nonstop routes served by carriers using jet equipment with at least 70 seats.

Overview and Trends 45 Other Possible Influences on Entry Dresner and Windle did not examine the influence of the Valujet inci- dent on entry activity but focused on two other possible determinants— labor costs and aircraft availability. A commonly held but largely intuitive belief among industry observers is that when the supply of labor and ffight equipment is tight—and accompanied therefore by high prices— the expense of starting and operating a new airline will dissuade startups. A tightening in the supply of these resources might even cause some air- lines that depend on low-cost labor and equipment to cut back or fail. Conversely, when supply is ample, an increase in the number of startup airlines might be expected. Liquidation of several large airlines during the late 1980s and early 1990s presented significant opportunities for startups at the beginning of the decade. However, simple trends in the cost and availability of these produc- tion factors do not confirm such clear relationships. Figure 1-10 shows the average salary per employee, adjusted for inflation, for all airlines and for incumbents only. The trend lines indicate that employee compensation had been rising in real terms for incumbents, but not for other carriers. This difference would suggest a comparative advantage for nonincumbents, possibly generating more entries. On the other hand, a review of the supply of the three common kinds of narrow-body aircraft typically used by new entrants—Boeing 737, Boeing 727, and DC-9—suggests that reduced availability of, and higher prices for, flight equipment beginning in the mid-1990s might have hindered entries (Figure 1-11). During the latter half of the 1980s, many large airlines increased aircraft capacity significantly, driving up the price of these and many other used narrow-body aircraft (Figure 1-12). As these airlines began to shed their excess equipment during and following the 1990-1991 recession, prices fell sharply, providing an opportunity for startup carriers. Aircraft listed for sale or lease declined, however, during the early to middle 1990s, with a particularly sharp dropoff in 1995 to 1996 following the Valujet crash and the temporary removal of many narrow-body aircraft from the market. In recent years, the availability of these aircraft has risen, partly because of a large number of new aircraft deliveries (see Dresner and Windle paper for data). After the sharp decline in market prices during the early 1990s, however, prices for used narrow-body jets began to sta- bilize during the middle part of the decade (Figure 1-12).

80,000 70,000 60,000 50,000 U) 40,000 0 30,000 20,000 10,000 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 Year SOURCE: Dresner and W:ndle 1999 Figure 1-10 Average salary per employee for all carriers and for incumbent carriers only (1998 dollars).

1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1000 900 800 700 600 500 400 300 200 100 0 Year NoT1, Data provided by BACK Associates. SOURCE Dresner and Wmdle 1999. Figure 1-11 Monthlylistings ofused aircraft available for sale orleaseworidwide: DC-9s, B-727s, B-737s, and all aircraft types.

120 100 80 60 0) 0) 40 20 1-90 2-90 1-91 2-91 1-92 2-92 1-93 2-93 1-94 2-94 1-95 2-95 1-96 2-96 1-97 2-97 1-98 2-98 Half Year SOURCE: GRA Aviation Specialists, Inc. Figure 1-12 Indexed market valuations (adjusted for inflation) of used narrow-body aircraft, 1990 to 1998.

Overview and Trends 49 Analyses by Dresner and Windle showed some slight correlation between these two factors—employee compensation and aircraft avail- ability—and net entry activity. However, this was an area of inquiry the committee did not have time to consider fully. Nonetheless, the influence of underlying economic factors should be considered when monitoring entry and exit patterns in the industry and analyzing their possible causes. A general conclusion that can be reached from these assessments is that there has been much entry and exit activity in the airline industry during the past decade. Shifting in and out of markets is apparently com- mon in the airline industry, although entry activity during the latter part of the decade seemed to decline. "While several possible reasons have been presented for this recent trend, its overall significance and causes remain uncertain. RESURGENCE OF LOW-FARE SERVICE Probably the most significant development in the U.S. airline industry during the past decade has been the continued expansion of Southwest Airlines and the resurgence in low-fare entry generally. Southwest Airlines Established as a Texas intrastate carrier in 1971, Southwest started fly- ing to markets outside of Texas immediately after deregulation; it now serves hundreds of domestic city-pairs far from its original base at Dallas' Love Field. Although the number of city-pair markets served by Southwest has grown substantially during the past decade, the carrier has retained its basic operating strategy, which is to serve short-haul (usually less than 750 mi), dense routes (offering a potential of 500 passengers per day) with frequent flights and low fares. When serving the nation's largest markets, Southwest often chooses to fly from secondary airports, such as Midway in the Chicago area, Oakland in Northern California, and recently MacArthur Field in metropolitan New York. Unlike other major airlines, Southwest does not need to operate hub-and-spoke net- works to fill its aircraft, since it concentrates service in dense, short-haul

50 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY markets in regions with enough population to generate sufficient local traffic to sustain nonstop service. Although it specializes in point-to-point service, Southwest's significant presence in some airports (such as Love Field in Dallas, Hobby Field in Houston, St. Louis, Salt Lake City, Oakland, and Baltimore—Washington International) permits a fair amount of one-stop service. Nevertheless, by minimizing the idle time for its aircraft parked at gates to receive connecting passengers (a routine wait for airlines operating connecting banks), Southwest is able to keep its aircraft flying and generating revenues. Southwest is the archetypal—as well as in major respects a unique— low-cost, high-productivity airline. An important source of the airline's cost savings and efficiencies is its avoidance of congested primary air- ports in large cities. Delays are costly to airlines, undercutting optimal use of aircraft and labor. To minimize costs, Southwest has targeted many secondary, underused airports near large cities, encountering less air traffic control delays as well as less congestion in ground handling both for aircraft and passengers, and lower facility and service fees (see Table 1-4). However, because it relies on point-to-point service between secondary airports, often less convenient for travelers, Southwest must generate higher passenger volumes through lower prices. Low fares and high aircraft use are imperative to its operating strategy. When it does fly into the major airports of large cities—and in some cases, into the hubs of incumbents (e.g., Detroit, Salt Lake City, and St. Louis)—Southwest seldom schedules flights during the peak take-off and landing times, choosing the periods between connecting banks. By avoiding congested airports and the peak activity of other carriers, Southwest can use its fleet most productively. Southwest achieves other important efficiencies through its uniform, all-coach fleet of Boeing 737s. These aircraft are well suited to short-haul, dense markets, and using a single type of aircraft simplifies maintenance, in-flight service, pilot training, and spare parts inventory. Additionally, Southwest participates in only one computer reservation system (CRS) and emphasizes direct customer purchases of tickets, saving on travel agent commissions and other booking fees. It also eschews seat assign- ments to speed aircraft boarding. For the most part, Southwest does not achieve cost savings and efficiencies through lower wages (common for

Overview and Trends 51 Table 1-4 Passenger Traffic on Southwest Airlines in Busiest U.S. Airports and Largest Metropolitan Areas, 4th Quarter, 1998 25 Busiest Airports by Enplanements by Enplanements by Southwest in Enplanements' Southwest Secondary Airports Metropolitan Area ORD - 635,384 MDW Chicago ATh -- - Atlanta LAX 750,486 853,426 ONT/BURISNA Los Angeles DFW - 814,203 DAL Dallas SF0 99,196 764,092 OAK San Francisco MIA - 188,911 FLL Miami DEN - - Denver JFK - - NewYork DTW 122,383 - Detroit PHX 1,043,034 - Phoenix LAS 1,017,901 - Las Vegas EWR - - New York STh 418,513 - SLLouis MSP - - Minneapolis BOS - 240,464 PVD/MHT Boston IAN 32,443 842,740 HOU Houston MCO 229,377 - Orlando - SEA 182,668 - Seattle LGA - - NewYork PIT - - Pittsburgh SLC 205,920 - Salt Lake City PHL - - Philadelphia CVG - - Cincinnati DCA - 481,566 BWI Washington, D.C. SAN 563,337 - San Diego TOTAL 4,665,258 4,820,786 'Ranked according to 1996 enplanement totals, including international enpianements. See Appendix D for airport identification codes. other low-fare carriers), but through constant economizing and efficient and intense use of its labor and equipment. With its low-fare strategy, Southwest successfully has generated high traffic volumes in many secondaxy airport-pair markets (e.g., Ontario, California—Oakland, California) and has pressured other airlines to lower their fares not only in the same markets but in related airport-pair markets (e.g., Los Angeles—San Francisco). Table 1-5 illustrates South- west's effect on fares and traffic in the markets it enters. In 160 short- haul (less than 750 mi) airport-pairs that Southwest entered for the first time between 1990 and 1998, annual passenger traffic increased by more than 20 million (174 percent), and average yields adjusted for inflation

52 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY fell by 54 percent. Southwest has accounted for about two-thirds of the passengers and nearly all of the traffic growth in the more than 300 short-haul airport-pair markets it now serves, as shown in Table 1-6. No other airline operates in the same way on the same scale as South- west Airlines. This airline alone accounts for about 75 percent of the passenger traffic carried on low-fare airlines as defined by DOT (see Figure 1-13). Most other low-fare airlines do not pursue the strategy of inaugurating significant service at secondary airports.16 Yet this strat- egy has allowed Southwest considerable cost savings despite the risks of untested markets and the need to improve the secondary airports. Only once—with Detroit's City Airport—has Southwest abandoned its secondary airport strategy because of insufficient supporting infra- structure (e.g., ground services) and tepid interest from travelers. By offering frequent flights, Southwest also attracts a significant amount of business traffic, and therefore it is not typical of the low-fare airlines that have served secondary airports by focusing on leisure clientele. Unlike Southwest, many newer low-fare airlines seeking business traffic have elected to compete directly with major carriers on their hub routes. Some operate out of the main hubs of major airlines (e.g., Air- Tran in Atlanta, Frontier Airlines in Denver); others have established bases in spoke cities and concentrated service in dense hub-spoke routes (e.g., Vanguard, based in Kansas City, operates routes to Mm- Table 1-5 Change in Average Yield (Adjusted to 1998 Dollars) and in Passenger Traffic in Markets Entered by Southwest Airlines During the 1990s 1990 1997-98 Percent Change Passenger Trips 12,170,210 33,372,310 174% Average Yield (cents/mi) 32.40 14.80 -54% NOTES: See Table 1-I for explanation of data sources and definitions. Includes only markets that Southwest did not serve in 1990 but where it accounted for 10 percent or more of total market passenger trips in the full year from 1997 3rd Quarter to 1998 2nd Quarter. The GDP price deflator was used to adjust for inflation. 11 However, some low-fare carriers might enter the market after an entry by Southwest.

Overview and Trends 53 Table 1-6 Average Yield for Trips in Short-Haul Markets Served by Southwest Airlines, 3rd and 4th Quarters of 1997 and 1st and 2nd Quarters of 1998 Passengers per Total Share of Average Yield Day Each Way Carrier Passengers Carried Passengers (cents/mile) 2:500 Southwest 63 21,317,380 17.78 N=55 Other 37 12,284,830 16.28 Subtotal 100 33,602,210 17.23 250 to 499 Southwest 73 8,247,880 18.88 N=45 Other 27 3,021,230 16.84 Subtotal 100 11,269,200 18.33 125 to 249 Southwest 67 5,614,250 16.96 N70 Other 33 2,734,330 18.06 Subtotal 100 8,348,580 17.32 50to 124 Southwest 57 1,637,690 16.33 N=54 Other 43 1,256,320 17.93 Subtotal 100 2,894,010 17.02 20 to 49 Southwest 54 670,090 16.82 N=48 Other 46 563,620 16.03 Subtotal 100 1,233,710 16.46 5 to 19 Southwest 60 217,710 16.08 N44 Other 40 145,430 17.54 Subtotal 100 363,140 16.66 All Southwest 65 37,705,000 17.81 N316 Other 35 20,005,850 16.71 Total 100 57,710,850 17.43 NoTEs: See Table 1-1 for explanation of data source and calculations source. Markets are based on airport pairs in which Southwest accounted for 10 percent or more of total O-D trips. N=Number of markets. 1 mile = 1.61 kilometer. neapolis, Chicago, Pittsburgh, and Dallas—Ft. Worth). By compari- son, Southwest has not sought access to slot-controlled airports and has avoided busy hubs, even forsaking such large markets as Min- neapolis and Atlanta. Arguably, Southwest also has provided the impetus for many in- cumbent carriers to establish low-fare divisions, such as Metrojet (US Airways), Delta Express, and United Shuttle. Incumbents generally

20,000,000 18,000,000 16,000,000 14,000,000 12,000,000 10,000,000 I- 8.000,000 6,000,000 4,000,000 2,000,000 0 sk, 9ç 9 % , '9.k,, $.& Year and Quarter NOTE: DOT considers the following as low-fare carriers: AirTran, Air South, ValuJet, Western Pacific, American Transair, Reno, Spirit, Vanguard, Nations, Carnival, Kiwi, Morris, Tower, Pro Air, and Frontier. Several smaller carriers not included. Figure 1-13 Total passenger trips on low-fare carriers: Southwest Airlines and others.

Overview and Trends 55 have deployed these low-fare, no-frill brands as competitors for South- west and other low-fare airlines in secondary airport-pairs in major metropolitan markets. Seldom will a low-fare division fly between major airports already served by the incumbent. Because traffic and fare data for these low-fare divisions are aggregated with those of the main incumbent, it is difficult to assess their overall effects on prices and service. Startup Airlines Earlier Entrants In the decade following deregulation, 1978 to 1988, the airline industry was awash in new entries—not only new airlines, but also the expansions of major, regional, and intrastate carriers. Dissatisfied with the econom- ics of point-to-point service, most incumbents strengthened or adopted hub-and-spoke systems that allowed for extensive route networks. Free to realign their routes, incumbent carriers sought to balance traffic flow at their hub airports by expanding their networks into new city-pair mar- kets. For instance, Eastern Airlines—which traditionally had carried East Coast north-south traffic—expanded its network westward; simi- larly, United Airlines, which had focused primarily on east-west traffic, moved into many northern and southern city-pair markets from its Chicago and Denver hubs (Bailey et al. 1985). US Airways and Delta extended their networks by merging with Piedmont and Western Air- lines, respectively. Deregulation was expected to lead major airlines to realign their route networks, but also to offer opportunities for smaller regional and intra- state carriers to extend their systems, as well as for charter (or supple- mental) operators to move into interstate service. Local-service airlines such as Allegheny, Frontier, Piedmont, and Ozark previously had been prevented by CAB from entering longer-haul and larger markets. Mean- while, intrastate airlines not subject to CAB regulation—such as Air Florida, Southwest Airlines (Texas), Air California, and Pacific South- west Airlines (California)—now could offer interstate service; so too could such established charter airlines as World Airways and American Trans Air.

56 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY In October 1978, Midway Airlines became the first completely new airline in 38 years to be granted a CAB certificate; it was quickly followed by New York Air, Muse, People Express, and several others. Over the next half dozen years, CAB would certify an additional two dozen brand-new airlines and would approve the expansion of many established intrastate and charter carriers. Some of these new airlines entered main- line, short-haul routes, offering nonstop, point-to-point service in mar- kets with high passenger volumes—like New York Air, flying between LaGuardia and Washington National. Without expensive labor contracts, many of these new airlines had lower cost structures than incumbents. They could offer much lower fares to attract travelers away from incumbents and to induce demand for new leisure traffic. At first, some incumbents allowed the new airlines to maintain a price advantage, but as price-sensitive passengers switched to the new airlines, declining load factors compelled incumbents to match the lower fares. Without the price advantage, the growth rates of new entrants slowed considerably, travelers chose airlines with established reputations especially when offered higher levels of service. Faced with aggressive responses, as well as a recession and rising fuel prices, some new airlines failed quickly (e.g., Altair, Golden West), while others changed their operating strategies, concentrating service in secondary airports such as Midway and Newark. Some new entrants also began to stress connecting services in their route planning, to increase load fac- tors as well as ffight frequencies that might attract business travelers. For instance, the low-fare carrier People Express established hubs in Newark and—after purchasing Frontier—in Denver. America West Airlines, formed in 1983, established a hub in Phoenix. Retaining cost advantages was one of the challenges new airlines faced as they matured and expanded their networks, sometimes by purchasing established airlines (Meyer and Oster 1987). Escalating costs coupled with operating strategies like those of mainline carriers proved prob- lematic in competing for service-oriented traffic. Most notably, People Express switched from a strategy of serving niche markets with high leisure traffic volumes (e.g., Newark—Jacksonville) to serving more tra- ditional but heavily congested, large markets such as Newark—Chicago. Shortly before declaring bankruptcy, Midway Airlines redoubled its efforts to attract business travelers, expanding into Philadelphia by

Overview and Trends 57 obtaining airport gates from Eastern Airlines. Air Florida, which offered a mix of high and low fares on unconventional routes, such as White Plains to Chicago O'Hare, could not generate passenger volumes suffi- cient to compete with the major airlines offering service between primary airports, such as Chicago—LaGuardia (Meyer and Oster 1984, 127). By 1990, nearly all of the airlines created in the wake of deregulation had failed or had been purchased by incumbent airlines. In addition, many of the former intrastate carriers (e.g., Air Florida) and some larger airlines (e.g., Braniff) had failed, or merged with other incumbents ('Western and National), or were on the verge of failing (Eastern and Pan American). Most of the large local service (i.e., regional) airlines had merged with larger airlines (e.g., Ozark with TWA, Republic with North- west). Only two airlines formed soon after deregulation are still offering jet service today (America West, formed in 1983, and Midwest Express, formed in 1984), along with one former intrastate carrier (Southwest) and one former charter operator (American Trans Air). Only the last two are consistently pursuing low-fare strategies.17 Although there are many explanations for the survival of so few of the airlines formed immediately after deregulation, the demise of so many undoubtedly has contributed to concerns that recent startups will suffer similar fates. Latest Entrants Following a dearth in startup and a recession that dampened air travel demand in the early 1990s, new entries surged again later in the decade. More than 20 new airlines were certified for operation and achieved operating status from 1992 to 1996 (see Figure 1-14).11 No new airlines were certified in 1997, but three applicants were granted certification in 1998.19 During the 1990s, about half as many airlines 17 America West still retains a low-fare strategy in many markets, in comparison with incumbents. 18 America West still retains a low-fare strategy in many markets, in comparison with incumbents. 19 Four applicants were awaiting DOT certification reviews in the second quarter of 1999.

58 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY have exited the industry as have entered, including several startup air- lines, such as Air South (1994-1997) and Western Pacific (1995-1997). Two incumbents, Eastern and Pan American, also exited. Some startups merged with larger airlines, including Morris Air with Southwest in 1995, and more recently, Business Express and Reno Air with Amer- ican Airlines. Other startups, such as Nation's Air Express and Grand Airways, operated briefly but never managed to generate significant passenger traffic. Nevertheless, many of the startup airlines from the early to mid-1990s continue to operate, including AirTran Airways (formed from a merger of startups Valujet and AirTran Airlines), Spirit, Vanguard, Frontier, and Midway.20 Joining these are some others that began scheduled air service in the late 1980s, most notably American Trans Air and Tower Air. With the exception of Midway, all of these new entrants operate primarily as low-fare airlines. Similar to the first wave of startups after deregulation, the airlines in this second wave have adopted varying operating strategies. Some have con- centrated on leisure traffic in tourist markets like Florida, while others have sought a mix of business and nonbusiness travelers in nontourist markets. For the most part, however, they have expanded their opera- tions slowly, particularly compared with some of the low-fare airlines formed in the wake of deregulation, such as People Express. Though most operate point-to-point service, several have set up bases in the hub airports of major airlines, to take advantage of high traffic volumes and the relatively high fares offered by incumbents. Before its purchase by Southwest, Morris Air operated from Delta's hub in Salt Lake City. Frontier Airlines initiated service from United's hub in Denver, flying short-haul, nonstop, low-density routes such as Fargo, Grand Forks, and Bismarck. Frontier since has added flights to larger markets such as Albuquerque, Las Vegas, and El Paso. It has also added flights on major hub-to-hub routes such as Denver to Atlanta and Dallas—Fort Worth. The main base of operations for AirTran Airways 20 Frontier and Midway differ from the original Frontier—which was established as a local carrier under CAB regulations—and from the original Midway Airlines, formed shortly after deregulation.

1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 Year SOURCE: Dresner and Windle 1999 Figure 1-14 Number of newjet carriers entering the industry and number exiting (or merging with others).

60 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY is Delta's Atlanta hub, from which it serves numerous business and leisure markets in the East, South, and Midwest. Spirit Airlines began operations out of Northwest's hub at Detroit Metro Airport. Though it has sought business traffic, Spirit now concentrates on serving leisure markets—for instance, by offering direct service from Detroit, New York (Newark, LaGuardia, MacArthur Field), and Chicago to vacation destinations such as Myrtle Beach, Atlantic City, and Florida. Vanguard Airlines has taken a different approach by establishing its base in Kansas City, a spoke for most hubs, but concentrating service in nonstop hub routes (e.g., Minneapolis, Chicago, and Dallas). American Trans Airlines operates largely from Indianapolis and Chicago's Mid- way Airport, serving business markets such as Dallas, Philadelphia, and New York (LaGuardia). Eastwind Airlines, founded in 1992, has moved its base operations to Greensboro, North Carolina, and flies to business centers in the Northeast (Boston, New York, Philadelphia, and Pitts- burgh), as well as to Florida. Midway Airlines, which began its opera- tions in 1993 from Chicago's Midway Airport, has adopted American Airlines' former hub at Raleigh-Durham International Airport and focuses its operations on East Coast markets. Midway is service-oriented, marketing its wide, leather seats, and seeks to attract business travelers in a manner similar to the more established Midwest Express Airlines, which operates from Milwaukee.21 As discussed earlier, there was a sharp drop-off in industry and mar- ket entries in 1996, coinciding with the aftermath of the Valujet crash. The extent to which this decline was a reaction to the accident and to the response by government regulators remains unclear. DOT slowed administration of its certification reviews for safety and fitness while reassessing its rules and processes, and it dismissed several applications for certification in 1997 and 1998. Undoubtedly, public concern over the safety of low-fare airlines and investor uncertainties about government regulatory reactions contributed to a decline in new airlines seeking entry. Nevertheless, even before the Valujet crash, DOT had expressed 21 However, Midway uses smaller regional jet aircraft than Midwest Express, which has a fleet of DC-9s.

Overview and Trends 61 concern that entry was being threatened by the unfair competitive prac- tices of incumbent airlines, including "predatory behavior" (DOT 1996, 32). These allegations are examined in more detail in the next chapter. Three years after the Valujet accident, entry activity seems to be rebounding. From 1997 to 1998, several new low-fare airlines inaugu- rated service, including ProAir, centered in Detroit, and Access Air, based in Des Moines. ProAir, which began operations from Detroit City Airport in July 1997, has focused on attracting business travelers, sched- uling short-haul ffights between Detroit and other business centers such as Pittsburgh, Newark, Atlanta, Philadelphia, and New York LaGuardia. With financial backing from the Detroit-area business community, ProAir has signed multi-year agreements with General Motors Corpo- ration and Daimler-Chrysler Corporation to provide air transportation for company employees and their families. AccessAir, formed in late 1998, is currently offering long-haul flights on B-737 aircraft from Des Moines and Moline, Iowa, to New York LaGuardia and Los Angeles. Like ProAir, AccessAir has financial backing from local businesses. Several other airlines have applications pending with DOT or have been certified and have announced their intentions to start service in 1999, including Legend Airlines based in Dallas (Love Field), Sun Country Airlines in Minneapolis, and National Airlines in Las Vegas. A low-fare, short-haul airline temporarily known as New Air plans to operate from New York's JFK airport and has applied to DOT for take- off and landing slots beginning in late 1999.22 SUMMARY During the 1990s, the airline industry has continued to evolve, still seek- ing a long-term equilibrium after 40 years of economic regulation that ended in 1978. Nonetheless, consumer gains from the deregulation of the industry have not eroded. From 1990 to 1998, average air fares fell by more than 25 percent, adjusted for inflation. Reductions in key pro- duction costs, including jet fuel, have contributed to this decline. 22 Bringing New Air to New York. Business Week, May 3, 1999, pp. 182-184.

62 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY Spurred by these lower fares and costs, and boosted by a strong econ- omy, passenger travel has increased along with airline profitability. One reason for improved profitability is a growing spread in the fares paid during the 1990s. Travelers who paid the highest fares (the top 5 percent of fares) accounted for 15 to 20 percent of airline revenue in 1998, compared with 10 to 15 percent in 1992. Some fare differentials can be attributed directly to costs—such as higher fares charged for flights during peak travel times, when resource constraints are highest and productivity can be lowest because of delays from congestion. Other markups, however, almost certainly are related to the high value of the service to some passengers—especially schedule- sensitive business travelers—and to the ability of airlines to identifj these travelers and charge them higher fares. Nevertheless, these travelers may be made better off, on balance, by an airline's ability to price-discriminate, which allows it to cover the cost of providing more frequent ffights. The concern is that the same condition that allows price discrimination— market power—also allows airlines to charge price-inelastic travelers fares that exceed cost. In a deregulated environment, the main antidotes for excessive mar- ket power are assurances of free competitive entry coupled with prohi- bitions on behavior aimed at undermining the competitive process. Leading the way with regard to the former has been Southwest Airlines, which has entered scores of new city-pair markets during the 1990s. The airline industry also has experienced a resurgence in new airlines, partic- ularly low-fare startups. Southwest and some other low-fare startups have focused on serving secondary airports in dense city-pair markets. These markets generally are less costly to serve than the busier main airports of metropolitan areas, which are prone to traffic congestion and other con- straints that increase operating costs and serve as obstacles to entry. However, many startup airlines have challenged incumbents at their hub airports, often producing lower fares for travelers. During the mid- 1990s, low-fare carriers began service on hundreds of nonstop routes that previously had been dominated by hub carriers. The highly publi- cized crash of a jet operated by a low-fare carrier in 1996 coincided with a decline in both the number of startup airlines entering the industry and the expansion of these airlines into new markets. Moreover, this incident occurred when DOT began to suspect that major airlines were sup-

Overview and Trends 63 pressing low-fare competition by sharply reducing fares and then raising them again as soon as the competitor had left the market. REFERENCES ABBREVIATIONS DOT Department of Transportation FAA Federal Aviation Administration GAO General Accounting Office TRB Transportation Research Board Bailey, E., D. Graham, and D. Kaplan. 1985. Deregulating the Airlines. MIT Press, Cambridge, Mass. Borenstein, S., and N. Rose. 1994. Competition and Price Dispersion in the Airline In- dustry.Journal ofPolitical Economy, Vol. 102, Aug., pp. 653-683. Craun, J.M., and R. Bennet. 1993. The Airline Deregulation Evolution Continues: The Southwest Effict. Office of Aviation Analysis, Office of the Secretary, Department of Transportation, Washington, D.C. DOT. 1996. The Low Cost Airline Service Revolution. Office of Aviation and Inter- national Economics, Office of the Secretary, Washington, D.C. Douglas, G., and J. Miller III. 1974. Economic Regulation of Domestic Air Transport: Theory and Policy. The Brookings Institution, Washington, D.C. Dresner, M., and R. Windle. 1999. Resource Availability and New Entry in the US. Airline Industry, prepared for TRB, April. Available from the authors and TRB. FAA. 1998. Aviation Forecasts: Fiscal Years 1998-2009. Report FAA-APO-98-1, Office of Aviation Policy and Plans, U.S. Department ofTransportation, Washington, D.C. Gale, l.A., and T.J. Holmes. 1993. Advance—Purchase Discounts and Monopoly Allo- cation of Capacity. The American Economic Review, Vol. 83, pp. 135-146. GAO. 1991. Airline Competition: Weak Financial Structure Threatens Competition. Re- port RCED-91-110. Washington, D.C., April. GAO. 1997. Aviation Safety: New Airlines Illustrate Longstanding Problems in FAA's Inspection Program. Report RCED-97-2, Washington, D.C., Oct. Geilman, Aj. 1971. Surface Freight Transportation. In Technological Change in Regu- lated Industries (W.M. Capron, ed.), Brookings Institution, Washington, D.C. Levine, M.E. 1965. Is Regulation Necessary? California Air Transportation and National Regulatory Policy. Yale Law Journal, Vol. 74, July, pp. 1416-1447. Meyer, J.R., M. Peck, J. Stenason, and C. Zwick. 1959. The Economics of Competition in the Transportation Industries. Harvard University Press, Cambridge, Mass. Meyer, J.R., and C.V. Oster. 1984. Deregulation and the NewAirline Entrepreneurs. MIT Press, Cambridge, Mass. Meyer, J.R., and C.V. Oster. 1987. Deregulation and the Future of Intercity Passenger Travel. MIT Press, Cambridge, Mass. Morrison, S.A., and C. Winston. 1995. The Evolution oftheAirlinelndust.'y. The Brook- ings Institution, Washington, D.C.

64 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY Nethercutt, L.L., and S.W. Pruitt. 1997. Touched by Tragedy: Capital Market Lessons from the Crash of Valujet Flight 592. Economic Letters, Vol. 56, pp. 351-358. Ramsey, E.P.A. 1927. A contribution to the Theory ofTaxation. EconomicJournal, Vol. 37, March, pp. 47-61. TRB. 1991. Special Report 230: Winds of Change—DomesticAir Transport Since Deregu- lation. National Research Council, Washington, D.C.

V Airline Competition at Hub Airports and Complaints of Unfair Conduct High average fares in many of the city-pair markets involving the hub air- ports of major airlines have been a recurrent subject of public concern and policy debate during the past two decades. In recent years, these markets also have served as main entry points for many new, low-fare airlines. However, new entrants have reported highly aggressive responses by hub- bing incumbents, prompting the Department of Transportation (DOT) to propose criteria for identifying and taking enforcement action against unfair, exclusionary practices. Trends in fares and competition at hub air- ports, the competitive concerns expressed by new entrants, and DOT's enforcement proposal are considered in this chapter. COMPETITION IN CITY-PAIR MARKETS GENERALLY Spread of Hub-and-Spoke Systems Airlines compete for passengers at the city-pair level. There are thousands of combinations of origin and destination (O-D) points that constitute 65

66 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY the markets for air transportation; in these markets, rivalry and entry activity are most important. In some, such as Los Angeles—San Francisco, thousands of passengers travel each day in both directions. In others, only a handful fly each year. The densest city-pair markets can support multiple flights by large jet aircraft each day, with most travelers head- ing between the same points. For the most part, frequent point-to-point jet service is confined to large cities with substantial business traffic and to some popular tourist routes. In the majority of city-pair markets, trav- elers must connect to another ffight at a transfer point, usually a hub air- port. Except for passengers in the densest markets, which have sufficient traffic to support regular point-to-point service, most who fly nonstop are originating from or headed to a hub airport, accompanied on the ffight by many other connecting passengers flying to and from dozens of connecting, or spoke, cities. The spread of hub-and-spoke systems following deregulation greatly increased the frequency of flights in most city-pair markets (Morrison and Winston 1986). These systems allow an airline to use a single air- craft to seat travelers heading from a spoke city to several different final destinations, then consolidate these travelers at a central, hub airport with others from other spoke flights, and finally recombine traffic to var- ious connecting points on departing flights. This process—almost al- ways taking place "online," that is, through a single airline or its affili- ated commuter carriers—allows for many nonstop flights at the hub airport. Travelers going to and from spoke cities to cities with hub air- ports, therefore, benefit from frequent nonstop service. Travelers headed between spoke cities also benefit because there are more flight options, even though they usually must transfer at a hub. Meanwhile, travelers who live in hub cities gain not only from frequent nonstop flights but also from the increased availability of nonstop service between scores of spoke cities. Many medium-size hub cities could not support as much nonstop service without the densities created by this connecting traffic. A key finding described in Winds of Change was that the number of city-pairs with three or more effective competitors increased sharply between 1979 (immediately after deregulation) and the mid-1980s, but had diminished somewhat during the late 1980s. An important source of this initial increase in competition was the established, or incumbent, airlines realigning their routes and expanding their hub-and-spoke net-

Airline Competition at Hub Airports and Complaints of Unfair Conduct 67 works.1 Added to that, the increased entry from startup airlines and the expanded operations by formerly intrastate and charter operators made many city-pair markets competitive battlegrounds. The largest increases in competition occurred in the long-haul markets. Hub-and-spoke sys- tems multiplied the routing options for long-distance travelers, who could choose to fly using the hubs of several competing carriers to reach their final destinations. The late-1980s dip in competition, most evident in short-haul markets, was attributed to mergers and failures among air- lines that had started service soon after deregulation. Update of Competition Analyses Any update of trends in competition in the airline industry will be affected by the specific time periods selected for comparison. The base year selected, 1992, was right after Winds of Change was released, as the air- line industry was emerging from the effects of a national recession. The comparison year, 1997, was the last fill year for which complete data were available when the analyses were conducted in early 1999. It is likely that comparisons using other years as beginning and end points would have yielded somewhat different outcomes. The results presented here, therefore, should be viewed as a snapshot comparison of competi- tion levels. Market Distance An update of the analyses in Winds of Change reveals that the majority of travelers in long-distance markets continue to experience—and benefit from—the most competition. As shown in Table 2-1, about 35 percent of all travel occurred in city-pair markets with three or more effective competitors (as defined in Table 2-1). This percentage has declined somewhat since 1992, when the figure was 39 percent. However, of all air travelers in the longer-haul markets, slightly more than half had the 1 Hub-and-spoke systems were difficult to construct when the Civil Aeronautics Board limited the ability of airlines to enter and exit routes. However, some airports, such as Chicago O'Hare and Atlanta Hartsfleld operated as major hubs even before deregulation.

Table 2-1 Passenger Trips by Market Distance and Number of Effective Competitors, 1992 and 1997 Competitors in Market Market Distance 1992 1997 (miles, one way) One Two ThreeorMore One Two ThreeorMore <250 Passengers 4,188010 8,168,650 9,472,110 6,548,290 7,411,970 10,644,560 Percent 19 37 43 27 30 43 251 to500 Passengers 16,124,510 25,261,090 14,641,990 23,887,130 39,643,400 12,119,050 Percent 29 45 26 32 52 16 501 to 750 Passengers 7,797,780 19,316,630 11,200,100 8,206,580 35,837,550 15,589,130 Percent 20 50 29 14 60 26 751 to 1000 Passengers 5,496.280 17,448,370 12,625,720 9,301,150 26,935,660 14,774,580 Percent IS 49 35 18 53 29 1001 to 1500 Passengers 3,544,970 19,252,140 21,795,340 6,124,600 23,691,110 31,685,210 Percent 8 43 49 10 39 52 1501 to 2000 Passengers 3,487,630 9,460,030 9,500,020 5,309,720 9,653,090 15,685,210 Percent 16 42 42 17 31 51 2001 or more Passengers 2,485,400 5,353,550 15,852,040 2,591,510 9,498,690 17,658,870 Percent 10 23 67 9 32 59 All Passengers 43,124,581 104,262,454 95,087,320 61,968,980 152,671,470 118,156,580 Percent 18 43 39 19 46 35 NOTE: An "effective competitor" is any airline carrying 10 percent or more of local traffic in a city-pair market. Data and calculations by DOT per request of the committee based on the 10 percent ticket sample in Databank IA.

Airline Competition at Hub Airports and Complaints of Unfair Conduct 69 benefit of choosing among three or more competitors in 1997—similar to the level in 1992. Travelers in shorter-haul markets (under 1,000 miles) have had fewer competing airlines from which to choose. This should be expected, since the closer two points are geographically, the less economical or practical it is to offer connecting service. Few rival hubs are likely to be within sufficient proximity to offer practical transfer service, and few travelers are interested in paying the time penalty for making connec- tions on shorter-haul trips. There has been a decline in competition in these markets, however. About 75 percent of short-haul travelers flew in markets with fewer than three competitors in 1997, compared with 68 percent in 1992. Market Density The Winds of Change study also examined market-level competition with respect to market traffic density. During the 1980s, high-volume mar- kets experienced the most competition. More rivalry might be expected in dense city-pair markets, where the traffic volumes can support more and varied, competing services. The Winds of Change analyses combined data from airports in metropolitan areas to obtain a more complete pic- ture of the competitive options available to travelers in a region. A similar examination of competition levels in 1997 reveals single-carrier markets accounted for only 14 percent of travel when market densities surpassed 500 passengers per day (Table 2-2). A comparable level (12 percent) was found in 1992. As might be expected, travelers in low-density markets (under 20 passengers per day) continued to experience the least compe- tition; a slightly higher share (32 percent) of these travelers had only one carrier to choose from in 1997 than in 1992 (28 percent). Overall, the data presented in Table 2-2 show that a higher share of air travel was occurring in one- and two-carrier markets in 1997 than in 1992. As shown in Table 2-3, travelers in the densest markets (more than 100 passengers per day) have experienced declining average fares in recent years irrespective of the number of competitors. Passengers in more lightly-traveled routes, by comparison, have experienced rising fares. Although there are likely to be some markets that have migrated upward in density during the time period simply because of growing aggregate

70 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY Table 2-2 Passenger Trips in Markets by Density and Number of Competitors, 1992 and 1997 1992 1997 Market Percent of Total Percent of Total Density Competitors Percent Passengers Percent Passengers 20 or fewer One 28 3.2 32 2.9 Two 44 5.1 46 4.1 Three ormore 28 3.2 22 1.9 21to50 One 13 1.2 16 1.2 Two 34 3.1 50 3.6 Three or more 53 4.8 34 2.4 51to100 One 19 1.9 20 1.9 Two 26 2.7 34 3.4 Three or more 55 5.5 46 5.6 101 to 200 One 21 2.9 20 2.3 Two 36 4.9 45 5.1 Three or more 43 5.7 35 4.1 201 to 500 One 20 5.1 20 5.0 Two 50 13.2 45 13.0 Three or more 30 7.8 35 10.0 501 or more One 12 3.5 14 4.9 Two 47 14.0 49 17.0 Three or more 41 12.0 37 12.7 Total Passengers Total Passengers All One 18 43,124,580 19 61,968,980 Two 43 104,260,460 46 152,671,470 Three or more 39 95,087,320 35 118,156,580 Total 100 242,472,360 100 332,797,030 NOTES: Density measured by O-D passengers per day each way. City-pairs combine all airports in the metropolitan area. A competitor is any airline carrying 10 percent or more of O-D traffic in the city-pair market. See Table 2-I for data sources and definitions. travel demand, the observed differences may have other explanations as well. One possible reason is that Southwest Airlines and most other new low-fare challengers have entered only the densest city-pair markets. Of- fering mostly point-to-point service, these jet airlines are not well suited to more lightly-traveled routes. COMPETITION IN HUB MARKETS Concerns Over Hub Concentration Competition in hub markets was a topic of interest in Winds of Change. Throughout the 1980s, it became apparent that operations at a dozen or

Airline Competition at Hub Airports and Complaints of Unfair Conduct 71 Table 2-3 Average Fares Paid by Passengers in Markets with 1,2, and 3 or More Effective Competitors, 1992 to 1997 (adjusted to 1997 dollars) Average Fare ($) Market Percentage Change Density Competitors 1992 1997 1992-1997 20 or fewer One 211 217 +2.8 Two 213 201 -5.6 Three or more 197 216 +9.6 21to50 One 176 184 +4.5 Two 191 169 -11.5 Three or more 198 222 +13.1 51 to 100 One 185 218 +17.2 Two 179 190 +6.1 Three or more 200 186 -7.0 101 to200 One 197 176 -10.7 Two 201 160 -20.4 Three ormore 201 179 -10.9 201 to500 One 176 167 -5.1 Two 177 159 -10.2 Three or more 176 161 -9.5 501 or more One 182 154 -18.4 Two 178 147 -17.4 Three or more 171 157 -8.2 Nom: See Table 2-1 for definitions and sources. Data and calculations by DOT per request of committee. so of the country's largest hub airports were dominated by one or two airlines. Hub airlines accounted for increasing shares of the total flights and enpianements at their hubs as their networks and connecting traffic grew. In some instances, these dominant positions were enhanced by mergers between airlines that shared hubs (e.g., TWA-Ozark and Northwest-Republic).2 As enpianements increased, hub carriers also increased their share of local traffic—that is, of O-D travel. The high proportion of local traffic handled by hub airlines was in large measure the natural result of the superior service they could provide local travelers—particularly schedule- sensitive ones. Competing airlines not operating a hub in the market 2 GAO 1988.

72 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY could not achieve the high traffic densities to schedule the frequent nonstop ffights desired by local business travelers. This scheduling advan- tage, boosted by marketing innovations—such as offering commission overrides to local travel agents and frequent-ffier programs, which build brand loyalty—tended to strengthen the carrier's hold on local traffic, particularly for business travel (Levine 1987; Borenstein 1989). Concern over hub concentration has persisted. Anecdotal evidence of high fares at hubs has prompted many studies of the effect of hub-and- spoke systems on fares and service, including analyses of fare differen- tials—or "hub premia"—between hub and nonhub markets. Results from some of these earlier studies3 were reviewed in Winds of Change, for which additional analyses also were conducted. The findings led the ear- lier committee to comment favorably on the overall benefits of hub-and- spoke systems, including consumer gains from increased competition in most connecting markets. At the same time, however, the committee observed higher average fares in city-pair hub markets in comparison with other markets, when applying controls for distance and traffic density. For nearly two decades now, the literature consistently has shown higher fares in city-pair markets that include a concentrated hub as either the origin or destination point; this especially applies to short-haul mar- kets in which one or two hubbing carriers handle most of the local traf- fic. Such findings, persisting over time, have prompted observers to question whether adequate conditions exist for free entry in many hub markets, particularly considering the many advantages hubbing carriers enjoy. Although some of these advantages exploit the preferences of travelers (e.g., frequent—flier programs), others seem unrelated to the efficiency or service of the carrier, such as the ability to offer commission overrides to local travel agents.4 Moreover, Winds of Change commented extensively on the ability of hubbing airlines to benefit from longstand- ing financial agreements with hub airports and to affect entry by limiting the availability of critical infrastructure such as airport gates (as discussed in Chapter 3). In response, incumbent airlines—which operate most of the major hubs—have long held that higher average fares in hub markets mainly GAO 1990; DOT 1990; Borenstein 1989. For an insighthil review of these advantages, see Levine 1987.

Airline Competition at Hub Airports and Complaints of Unfair Conduct 73 are the result of differences in traveler preferences among hub and non- hub markets. They point out that hub airports are located in major busi- ness centers, which attract many more time-sensitive business travelers. The airlines claim that comparisons between hub and nonhub markets exaggerate fare differences, because many nonhub city-pairs include more leisure destinations. Thus, two main reasons have been offered by airlines for why analy- ses of aggregate fare data consistently show that major hub markets tend to have higher fares, on average, than nonhub markets. Both stem from the notion that hubs have a disproportionate share of business travelers. The first reason, which is not especially controversial, is that few, if any, major hub cities are important tourist destinations; any aggregate com- parisons of average fares at hub and nonhub markets are likely to include more leisure markets in the nonhub data. This alone could account for some of the fare differential at hubs, since leisure travelers tend to be less costly to serve than business travelers, who demand greater schedule fre- quency (as discussed in Chapter 1). Another possible reason is that most hub airports are in cities that are major centers of businesses, centrally located (e.g., Chicago), and attractive to large, travel-intensive busi- nesses, in part because of the superb nonstop service. To the extent that hubs serve a higher proportion of business travel- ers, higher average fares could be expected in hub markets because of the cost differences in serving time-sensitive travelers. By the same token, however, the higher proportion of price-inelastic business travelers in hub markets also provides greater opportunity for hubbing airlines to exercise their market power by price discriminating and possibly raising fares above the cost of efficiently providing the schedule-intensive service. As discussed in Chapter 1, fare differentials are necessary for airlines to pay for the frequent and convenient service desired by business travelers. The issue, which cannot be empirically assessed here, is whether higher fares at hubs are the result of major airlines exploiting market power in their hub markets. Update of Hub Analyses The Winds of Change study defined "concentrated hubs" as those large cities in which either a hubbing carrier accounted for more than 50

74 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY percent of all local traffic or two hubbing carriers together accounted for more than 75 percent.' The following 10 cities were dassifled as concen- trated hubs: Atlanta, Charlotte, Cincinnati, Dayton, Denver,6 Memphis, Minneapolis, Pittsburgh, Salt Lake City, and St. Louis. In addition, five other hub cities were defined as relatively unconcentrated, even though in each case a single hubbing carrier accounted for 40 to 50 percent of the local traffic. These were Dallas, Chicago, Detroit, Nashville, and Raleigh-Durham. Dallas and Chicago were deemed unconcentrated because the calculations included significant traffic from the two sec- ondary airports, Midway and Love Field. Comparisons of average fares in city-pair markets involving these ten concentrated and five unconcentrated hubs consistently found higher fares in the concentrated hubs, especially in short-haul markets (con- trolling for distance and density). Also, to control for the effects of price- sensitive leisure (i.e., tourist) markets, Winds of Change excluded from the city-pairs all O-D points involving California, Florida, Nevada, and Arizona. Nevertheless, average fares were higher in the concentrated hubs than in the other markets. Hub CityAnalyses To update some of these earlier analyses, the current committee tracked trends in the market shares of the largest carriers in those 15 hubs since 1990. The updated results in Figure 2-1 show that most of the hub car- riers continue to handle 50 percent or more of the airport's local traffic, although there is no discernible upward trend in concentration. In three It is important to avoid using a carrier's share of total enplanements as the main mea- sure of market share or to identify concentrated hubs. Carriers using an airport as a main transfer point can account for a large share of enplanements regardless of their share of local traffic. Delta, for instance, would still account for more than 35 percent of enplane- ments at Cincinnati even if it carried no local traffic. The share of local passenger traffic carried by the major carrier in a hub airport (or on an individual route) is a more infor- mative measure. 6 Denver was the only airport that was defined as a concentrated hub because it had two carriers (United and Continental) together accounting for more than 75 percent of local traffic.

100 90 80 e. 70 C) I— 60 0 50 (5 U 0 —J 40 30 20 C) 0 0 10 —NW at DTW --NWat MSP ---DL atATL —W--AA at DFW ---AA at RDU8 - - TWaISTL —G--AA at BNAb DL at SLC US at CLT —US at PIT UAat DEN US at DAYC NWatMEM DL at CVG UAatORD ? 19p ' , It, % 8Beginning in 1995, the market share totals are for US Airways. b Beginning in mid1 995, the market share totals are for Southwest Airlines. clfl 1992, US Airways abandoned its Dayton hub. NOTES: These 15 airports were selected for reasons explained in the text. The analysis does not reflect traffic at DAL and MOW, which would lower As and UAs market shares in Dallas and Chicago, respectively. Airport codes are listed in AppendIx D. Figure 2-1 Market share of largest carrier at 15 current and former hubs, 1990 to 1998.

76 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY instances—Dayton, Nashville, and Raleigh-Durham—the hubs since have been abandoned by the incumbent carriers and no airline accounts for more than 35 percent of local traffic. In Salt Lake City, the domi- nance of Delta Airlines has declined somewhat as Southwest has gained a significant share of many city-pair markets. Charlotte and Pittsburgh, main hubs of US Airways, continue to be among the most concentrated airports. In this case, the figures shown for Chicago and Dallas do not reflect competition from airlines in the secondaiy airports, Midway and Love Field. It is important to note, however, that whereas American Airlines accounts for nearly 60 percent of local traffic at DFW, it accounts for less than 40 percent of Dallas-area traffic when Love Field is included. Likewise, United accounts for 45 percent of traffic at Chicago O'Hare but less than 35 percent of Chicago-area traffic when Midway is included. Trends in average fares in these 12 current and 3 former hub airports, from 1990 to mid-1998, show a general decline when adjusted for infla- tion (Figure 2-2). Noteworthy, however, is that most of this decline occurred during the first half of the 1990s, before the general improve- ment in the economy. From the second quarter of 1995 to the second quarter of 1998, fares fell in only 5 of the 15 markets—Atlanta, Dayton, Detroit, Memphis, and Nashville. Two of these markets are former hubs (Dayton and Nashville) and the others have experienced significant low- fare new entry. The updates reveal other notable changes. The sharpest fare declines over the entire period occurred in Salt Lake City (-40 percent) and Atlanta (-33 percent), probably because of entries by Morris Air and Southwest in Salt Lake City and by Valujet and later by AirTran in Atlanta. When adjusted for inflation, average fares also have fallen by 20 percent or more in St. Louis (-24 percent) and Memphis (-22 percent), both served extensively by Southwest and some other low-fare airlines. Travelers in the three former hubs of Dayton, Raleigh-Durham, and Nashville also experienced average fares that were 20 percent lower; a decline that is at odds with the hypothesis that higher average fares at hubs reflect a naturally (and disproportionately) high percentage of busi- ness travel in hub cities. Meanwhile, in the highly concentrated hubs of Charlotte and Pittsburgh—neither a location for lasting, low-fare entry—average fares have declined by less than 5 percent.

50 275 250 225 200 175 0 150 125 100 75 0 D1W MSP ATL H DFW RDU - - STL —O--BNA .L SLC CLT S PIT I DEN DAY MEM CVG ORD Year and Ouarter NOTES: Average fares adjusted for inflation using the GDP price deflator. See note to Figure 2-1, and see Appendix D for airport codes. Figure 2-2 Average fares at 15 current and former hubs, 1990 to 1998 (adjusted to 1998 dollars).

78 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY Hub City -PairAnalyses More relevant are fare differentials in city-pair markets. A comparison of average fares in the most heavily traveled city-pair markets was made for the second quarter of 1998. City-pairs that involved one or more of the 12 concentrated hubs were defined as hub markets and all other pairs were defined as nonhubs. Average fares were calculated for more than 800 of the country's busiest city-pair markets. The results of this comparison, grouped by short-, medium-, and long-haul markets are shown in Figure 2-3. Each data point represents the average fare of a hub or nonhub city-pair market with distance indi- cated on the horizontal axis. In all three groupings, the hub markets tended to have higher fares than nonhub markets; the trend lines show the least—square fit. To control more precisely for the effects of distance and other possi- ble variables affecting average fares, several multiple regression analyses were performed for the 1,000 top airport-pair markets.' For a full year ending in the second quarter of 1998, average fares were regressed on distance, population, and income variables (in logarithmic and linear forms). As shown in Table 2-4, the highest-fare markets—irrespective of regression form—consistently involved hub or slot—controlled airports as either origin or destination points. The 12 concentrated hub airports were involved in 54 of the 75 highest-fare markets.9 Of the 75, 29 involved at least one of the four slot-controlled airports (see Chapter 3). Airports repeatedly in the high-fare markets were LaGuardia (19 per- cent of city pairs), Philadelphia (15 percent), Detroit (15 percent), Newark (13 percent), Atlanta (12 percent), Chicago O'Hare (12 per- cent), and Boston (11 percent). Seven other airports—Dallas—Ft. Worth, Minneapolis, St. Louis, Washington (Reagan National), Charlotte, Baltimore, and Cincinnati—also appeared in 5 to 10 percent of the 75 highest-fare markets. These were the most recent data available for the analyses. 8 To save time, the committee tasked DOT to run the regressions in accord with the committee's instructions. ° Newark and Philadelphia, which are important but unconcentrated hubs for Conti- nental and US Airways (respectively), also were involved in a large number of high-fare city-pair markets.

250 50 Average fares in 249 densest city-pair markets of less than 500 miles: 97 hub and 152 nonhub markets ________ 0 0 0 o65 o 0 0 0 o• 6 ' o on— .• cc j• 0e j%0 ' ' —• —nhub trendline - - -Rib trendhne 150 225 300 375 450 Average fares in 322 densest city-pair markets of 500 to 999 miles: 165 hub and 157 nonhub markets -o 0 - -0-o--0—. 0 S 0 0 0.00 00 0 0 • 00 C 0. 0 000 0 0 • o0% OA 00 Cb . 0 C - 0. 0 —nhubfrendtne - - -Rib trendine 600 700 800 900 1000 Average fares in 290 markets of 1000 to 1999 miles: 125 hub and 165 nonhub markets 0 ) 0 0 .cbcJ oo 00 S0 0- --Rib trendline —Rinhubtrendhr,e 1000 1200 1400 1600 1800 2000 Nonstop dtance (ne) . = Nonhub city-pair market o = Hub city-pair market NOTES: See text for definitions. 1 mile = 1.61 kilometer. Figure 2-3 Average fares in the densest hub and nonhub city-pair markets, 2nd quarter, 1998. 0 500 350 350 250 (5 ISO I: so

Table 2-4 Seventy-five Highest- and Lowest-Fare Markets Among Densest 1,000 in 1997, Controlling for Effects of Population, Income, and Distance Highest-Fare Concentrated Slot Lowest-Fare Concentrated Slot Markets Hub Controls Markets Hub Controls I. CLEDTW X I. SLC SMF 2. DTW PIT X 2. SF0 SNA 3. AU TYS X RNOSNA AU BNA X 4. MCI MDW 5. DTWMKE X 5. LA.SMAF 6. CVGORD X X 6. LIT SOP 7. AUGSO X 7. (ND SRQ 8. MSPORD X X 8. OAKSNA 9. All. CVO X 9. (ND PIE 10. DSMORD X X IS. BWI CLE it BWIEWR II. CMHLAS 12. CLE PHI. 12. ACYFLL 13. DFWLGA X 31 DALUT GSO lAD 14. GEGSMF IS. DIW MSP X IS. RNOTUS 16. DTW ND X 16. MCI Sit X I?. CVGPHL X 17. SIC SNA 18. CLTPHL X 18. DALMAF 19. JFK PHI. X OKCSU X DCADTW X X 20. CLE MDW 21. ORDPIT X X 21. PHXPVD 22. MIlE MSP X 22. DALLBB 23. CLE EWR 23. LAX SLC X 24. BWI DTW X 24. MCO MSY 25. DCA P.DU X 25. LAS SMF 26. DT%V EWR X 26. ELF PHX 27. CLI EWR 31 27. MCI SF0 28. DTW POlL X 28. GEG SEA 29. DCAEWR X 29. MCO SAT 30. CLI LGA X X 30. PDX SIC 31. CLIORD X X 31. MCI SDF 32. IAN LGA X 32. AMA LAS 33. LGA MSP X X 33. MCOPVD 34. AlL CLT X 34. 1040 LAS 35. DTW LGA X X 35. OAKPDX 36. HPN ORD X X 36. LAS SAT 37. 805 DTW X 37. SEA SMF 38. CVGEWR X 38. SEA SLC X 39. BOS EWR 39. LAS MDW 40. AU DCA X X 40. FLL MO 41. EWR MSP X 41. MDW SDF 42. PHI. Sit X 42. ABQPHX 43. CLI Sit X 43. HOULAS 44. LOARIC X 44. BO! SEA 45. IAN PHL 45. SAN SLC 31 46. (Al) ORD X X 46. AUS LAS 47. MEM ORD X X 47. KNOSAN 48. BOS MSP X 48. GEG SIC 49. BOS DCA X 49. LAX OAK 50. AU LGA X X SO. LOB POOl SI. CLE LGA X SI. BNAMCI 52. LGA PIT X X 52. FLL TPA 53. AlL RIC X LAS PDX CVGLGA X X 54. 801 SLC X 55. DFW SIC X 55. SAN SIC 56. BWI MIff 56. ONT SIC 57. LGA Sit X 31 57. BWI SLC X 58. PHI. RDU X 58. LAS OMA 59. BOS PHL 59. GEGOAI( 60. BOS Sn. X 60. MDW OMA 61. BOS IAN 61. OAK SLC X 62. DEN LGA X X 62. LAS SIC 63. GSO LGA X 63. RJ4OSJC 64. BOS DFW X 64. LAX SMF 65. CMHEWR 65. BNACLE 66. DFW MEM .X 64. LAS SEA 67. DFW SNA X 67. AMA DAL 68. EWR IAN OAK ONT CMH PHI. 69. OAXSEA 70. DFW LAX X 70. BOl COG 71. MSP PHL X 71. OAKSAI4 72. AU CHS X fl. BOl GEG 73. OWl LGA X 73. OAK SAN 74. DFW SF0 X PHX SAN BDL ORD X X 75. DNA MSY NoTE See Appendix D for sirport codes.

Airline Competition at Hub Airports and Complaints of Unfair Conduct 81 These results illustrate the magnitude of the fare differentials at hubs, which have spurred so much public concern. Of course, many of the highest-fare markets in Table 2-4 are among the nation's most popular for business travel; thus many of the unexplained fare residuals might be related not only to market power and the low price elasticity for busi- ness travelers, but also to the higher costs of meeting service-oriented demand. Also evident from the analyses is that many of the high-fare markets involve airports in the East and Midwest, where airway and airport capacity constraints contribute to costly traffic congestion, and conse- quently higher fares. Consistently missing from the highest-fare mar- kets are western cities, induding some city pairs (e.g., Los Angeles—San Francisco) that might be expected to generate a significant amount of business traffic. Presumably, the influence of business traffic on average fares should be manifest through higher average fares in some markets outside the East and Midwest. The geographic concentration of high- fare city-pair markets in the Midwest and East raises questions about the causes. The presence of Southwest and other low-fare airlines in many west- ern markets is cited frequently to explain this geographic pattern. More interesting, however, are the underlying factors that attract discount air- lines to these markets. Less airport congestion and fewer air traffic delays are possible reasons. Another possibility is the differing demand charac- teristics among eastern and western travelers—perhaps because the longer distances between population centers in the West can make flying the preferred mode for leisure travelers, leisure-oriented airlines can be more viable. Whatever the reasons for this pattern, they merit further explo- ration in future studies. ALLEGATIONS OF UNFAIR COMPETITIVE RESPONSES TO NEW ENTRY Growing concern over higher fares at hub airports was tempered some- what with the emergence of many low-fare carriers during the early 1990s and by the continued expansion of Southwest Airlines. As noted in Chapter 1, many start—up airlines began service at hub airports or at sec- ondary airports in hub cities. The new entrants would schedule a

82 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY few flights per day on dense routes, operating point-to-point service to achieve high load factors on narrow-body aircraft designed for short-haul, high-volume markets. Although not the only markets selected by the startups, nonstop routes from major hubs such as Chicago O'Hare, Dal- las—Fort Worth, Denver, and Atlanta were targets of much new entry. To fill planes without feed traffic from connecting flights, the new entrants offered fares below those of the hubbing airlines. The idea was to stimulate pent-up demand not being met by the hubbing airlines with their restricted offerings of discount fares, although diversion from incumbents also was anticipated. 'VVhatever the particular strategy, the effect of new entry in reducing fares and increasing leisure traffic in many hub markets quickly became evident, not only to incumbents but also to DOT, which began to promote the entry of low-fare airlines to challenge the dominant positions of hub carriers. In its April 1996 report, The Low—Cost Airlines Service Revolution, DOT characterized the resurgence of low-fare airlines as a "watershed development in domestic aviation" that was having "a profound effect on efficiency, competition, consumers, and industry structure" (DOT 1996, 1). The report estimated annual consumer savings of $6 billion from new airlines that had based their operations in hub airports and focused on service in city-pair routes having above-average fares. Complaints by these new entrants pointed to highly aggressive responses by incum- bents, prompting DOT to consider whether the larger carriers were try- ing to drive out their smaller rivals and exclude them from markets. In its 1996 report, DOT announced its intention to review carefully alle- gations of anticompetitive conduct and to cooperate with the Depart- ment ofJustice (DOJ) to enforce violations of antitrust law. In addition, DOT noted its own statutory authority to prohibit unfair methods of competition (49 U.S.C. §41712), stating that it would consider pro- ceeding independently to deter conduct that could be characterized as anticompetitive under antitrust principles. DOT's Proposed Enforcement Policy Against Unfair, Exdusionaiy Practices In the April 10, 1998, Federal Register, DOT published a proposed "Enforcement Policy Regarding Unfair Exclusionary Behavior in the

Airline Competition at Hub Airports and Complaints of Unfair Conduct 83 Air Transport Industry."10 The statement, contained in Appendix A, described DOT's perception of the problem, as well as its authority and its means to address it. In a preface, DOT maintained that its main inter- est was to discourage incumbents (referred to as "major carriers") from engaging in unfair conduct designed to exclude new entrants from competing in hub markets. Although DOT acknowledged that "unfair exclusionary conduct" might encompass various tactics airlines employ to suppress competition in other kinds of markets, its policy statement focused on the pricing and capacity responses by incumbents when challenged by new entrants in hub markets. It is in these markets that DOT indicated it had received the most complaints about aggressive price-cutting responses by incumbents. DOT singled out new entrants—which it defined as independent air- lines starting jet service during the past 10 years—as the most likely and susceptible targets for exclusionary conduct, since these new entrants fre- quently operated from hubs and had limited resources and staying power to withstand a prolonged offensive. DOT noted that the established low-fare airlines, namely Southwest Airlines, were seldom the targets for aggressive price-cutting responses in markets in which they competed against hubbing carriers. DOT stated that its informal investigations of complaints revealed that incumbent airlines had both the opportunity and the motive to engage in exclusionary conduct. The opportunity was provided by the availability of comprehensive and "real time" information on competitor prices (obtained through computer reservation systems [CRSs]) and by their ability to change prices quickly and to shift aircraft and seats among city-pair markets, without incurring significant, additional fixed or over- head costs. According to DOT, the incumbent's motive was to protect its dominant position in the hub market and its long-term ability to charge higher fares to price-inelastic business travelers. DOT identified three pricing and capacity responses to screen out potentially exclusionary conduct by an incumbent; if any of the follow- ing actions resulted in lower local revenue to the incumbent than would a "reasonable alternative response," DOT would investigate: 10 See Federal Register. 1998. Vol. 63, No. 9, (April 10) pp. 17919-17922.

84 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY The incumbent added capacity and sold a large number of seats at very low fares; The number of local passengers that the incumbent carried at the new entrant's low fares—or at fares substantially below its own previous pricing—exceeded the new entrant's total seat capacity; or The number of local passengers that the incumbent carried at the new entrant's low fares—or at fares substantially below its own previous pricing—exceeded the number of low-fare passengers carried by the new entrant. DOT explained that a reasonable alternative response could be the incumbent matching the low-fare offerings of its new competitor on a restricted basis—for instance, without greatly increasing the number of low-fare seats made available. This presumes that the incumbent responding in such a restrained manner could retain much of its high- fare, business traffic because of its service and marketing advantages. In DOT's view, incumbent airlines were protecting and strengthening their ability, gained through market power, to set prices well above cost when they responded to low-fare entry in the unreasonable manner of sharply reducing fares and increasing the number of unrestricted low-fare seats available. Using these three criteria to guide its investigations, DOT warned it would pursue cases that strongly suggested exclusionary behavior, through hearings before administrative law judges. It concluded its proposed pol- icy statement by noting that in addition to examining questionable pric- ing and capacity responses, it also would consider other indicators of unfair competition by airlines, such as actions reducing opportunities for— or raising the cost of—entry and competition. Reports of incumbent car- riers hoarding gate space, using their contractual agreements with hub air- ports to bar access by rivals or to increase the price of airport services, and offering travel agents extra commissions to discourage bookings on new entrants also were identified as actions that would prompt further inquiry. Rationale for DOT's Criteria In its proposed criteria, DOT's focus was on responses to new entry by major airlines involving pricing and capacity assignments that it be-

Airline Competition at Hub Airports and Complaints of Unfair Conduct 85 lieved could reasonably be construed as having predatory aims. There is no universally accepted theory about what constitutes predation or how best to detect and prove its occurrence. The fundamental concern is that the dominant firm will reduce prices specifically to drive out ri- vals or to discourage future entry or reentry, expecting to recoup any losses incurred by subsequently raising and keeping its prices above com- petitive levels.11 In general, predation is believed to have occurred when a firm with dominant market position has priced its products below mar- ginal cost, so that the additional revenue generated from the sale of one more unit would be less than the incremental cost of making the sale.12 It is presumed that a rational firm would not incur such avoidable losses for very long—for instance, by continuing to add capacity—unless it had some other aim such as strengthening its market power and its potential to raise fares charged to price-inelastic travelers in the future. The traditional view is that such losses may be an investment intended to generate higher future returns in those markets. Recent economic theories, however, point to the possibility that predation also has other recoupment objectives. For instance, a dominant firm might engage in predatory behavior to send a signal to current and prospective rivals that its costs are low and the potential for profitable entry is slight. 13 The pur- pose of predation also might be to develop the firm's reputation as a tough competitor and to suggest that entry will spur prolonged and costly price cutting. 14 The price-cutting incumbent also might be try- ing to build a larger wall to deter entry into the business generally—for instance by compelling prospective competitors to amass more credit or cash reserves to remain solvent until achieving profitability. According 11Joskow, P.L., and A.K. Klevorick. 1979. A Framework for Analyzing Predatory Pric- ing Policy. Yale Law Journal, Vol. 89, No. 2, Dec. 12 Exceptions include instances in which the firm cuts prices for promotional reasons, such as introducing a new product, and other reasons that bring benefits, such as an increase in the sale of a complementary product. 13 Milgrom, P., and J. Roberts. 1990. New Theories on Predatory Pricing. In Industrial Structure in the New IndustrialEconomics, Clarendon Press, Oxford. " See Kreps, D., and R. Wilson. 1982. Reputation and Imperfect Information. Journal ofEconomic Theory, Vol. 27, pp. 253-79. Also, Comanor, W.S., and H.E. Frech. 1993. Predatory Pricing and the Meaning of Intent. Antitrust Bulletin. Vol. 38, No. 2, Sum- mer, pp. 293-308.

86 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY to these theories, a single act of predation can have effects that transcend the specific market in which the price cutting has occurred. One significant difficulty in proving predation is determining when and how the firm is likely to recoup its investment in predatory losses. Theories about predatory tactics suggest a variety of possibilities, includ- ing many that would be formidable to quantify—such as the effect of predation in deterring entry in other markets or the industry generally. Moreover, a firm's marginal cost function may not be evident. Because of the many practical difficulties of quantifying marginal costs, the fed- eral courts in recent years have adopted the short-run average variable cost (AVC) as a proxy. 15 AVC is an accounting measure of the avoidable (i.e., nonfixed) costs of producing output during a period—such as ex- penses for labor, fuel, and material. These variable costs are totaled and divided by the output during the period to calculate the AVC. Economists, however, also view opportunity costs as an appropriate component of marginal cost. More profitable opportunities forgone by deploying resources in a particular way constitute a true cost. A firm that neglects opportunities for more profitable uses of its resources, and that has the information to ascertain these opportunities, is presumably act- ing against its own interest. Thus, to the extent that AVC mainly reflects the direct expenses incurred in production, it is an unsatisfactory proxy for marginal cost—since it does not account for more profitable oppor- tunities forgone. By emphasizing revenue "self diversion," DOT seemingly was trying to incorporate opportunity costs into its method of detecting predation. DOT's three screening criteria presumed that an incumbent that was substantially reducing fares and adding seating capacity to a route in re- sponse to new entry was sacrificing higher-fare sales in that market and possibly in other markets by diverting seats from connecting passengers or redeploying aircraft from other routes. DOT seemed to be implying that the profits sacrificed from these other possible uses should be in- cluded in calculations of marginal cost, since they represented opportu- nity costs. 11 The average variable cost measure, adopted by the federal courts to test for preda- tion, was first proposed by Areeda, P.E., and D.F. Turner. 1975. Predatory Pricing and Related Practices Under Section 2 of the Sherman Act, Harvard Law Review, Vol. 88, pp. 697-733. The test is commonly referred to as the Areeda-Turner Test.

Airline Competition atHubAirports and Complaints of Unfair Conduct 87 One of the difficulties with DOT's approach for detecting predation, however, is ascertaining opportunity costs. In the airline industry, a major airline can operate in thousands of markets that vary widely in profitability at any given time; it is likely, therefore, that the airline has excess or idle capacity somewhere in its system. Determining retrospec- tively if and where this capacity could have been employed more prof- itably can be speculative and hypothetical. Dynamic analyses of response options also might suggest many reasonable alternatives. An incumbent airline, challenged by a low-fare entrant, might find it unprofitable to continue its discriminatory pricing—for instance, by continuing to charge high fares to time-sensitive travelers while matching the new entrant's low fares on a limited basis. One reason is that time-sensitive business travelers might find the discounted—and often unrestricted—fares of the new entrant to be sufficiently low to be appealing, despite the sacri- fice of frequent ffier benefits and schedule intensity. In this regard, the incumbent may not have a realistic option—at least in the short-term— of continuing to offer higher unrestricted fares, as DOT presumes. Because of the complexities of airline pricing and network operations, distinguishing between legitimate and questionable competitive responses poses significant challenges, and raises the possibility of false charges of predation, and the risk that some genuine predatory conduct will go un- detected. While particular theories of predation that are more or less rele- vant to the airline industry could not be judged in this study (with its broader charge), the committee did consider specific complaints of ex- clusionary practices received by DOT and some of the risks and challenges involved in trying to spot and prohibit such conduct in the airline industry. Reports of Exdusionary Conduct DOT submitted to the committee a total of more than 40 complaints from new entrants alleging exclusionary or anticompetitive conduct by incumbents between March 1993 and May 1999; these are listed in Appendix C. A few of the complaints were not relevant. Some, for instance, concerned nonpassenger services. In addition, there were also multiple complaints by the same carrier against another. The first type was disregarded, while the repeated complaints were combined into single cases. In the end, there were 32 individual and combined cases. DOT believed this list was not comprehensive, and that other similar— but unreported—occurrences were likely.

88 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY Though possibly incomplete, the 32 complaints, summarized in Table 2-5, illustrate the various kinds of conduct that have prompted concern. Half of the complaints dealt with incumbents exploiting both their contractual relationships with airports and their marketing and dis- tribution advantages. The complaints involving contractual relationships pointed to questionable shortages in hub airport gates for leasing or sub- leasing, excessive subleasing rents charged by incumbents, the required use of an incumbent's ground-support service (e.g., baggage handling) Table 2-5 Nature of Informal Complaints About Unfair Practices Received by DOT, 1993 to 1999a Year and Number 1999 1998 1997 1996 1995 1994 1993 Total 2 2 3 3 3 0 3 16 1 4 3 0 2 0 0 10 0 1 0 02 0 3 6 3 7 6 3 7 0 6 32 Complaint Unfair pricing and capacity responses Impediments to gate access and other airport facilities or servicesC Other t Total All 32 reports presented to the committee by DOT are provided in Appendix c, along with the case num- bers that are referenced in the notes below. Reports were combined into tingle reports (cases) if they in- volved the same carriers and time periods. Five were excluded because they were not relevant (these com- plaints, however, are presented at the end of Appendix Q. DOT also provided the committee with a second Set of complaints (not shown in Appendix C) about competitive practices of airlines and airports submitted by members of congress and state and local public officials. The general problems identified are discussed in Chapters 2, 3. and 5—many concern fare and service levels in small- and medium-size communities. A third set of complaints, mainly from travel agents and major airlines, concerned travel agent commissions and computer reservation systems. Some of these issues are discussed in chapter 3. In addition to these complaints, DOT has identified other cases in which it believes unfair or exclusionary pricing and capacity responses have occurred, on the basis of its own review of fare and traffic data; several of these were also presented to the committee and some are examined in Table 2-6. Finally. DOT observes that the complaints listed in Appendix C are not comprehensive and that other similar, but unreported, cases are likely. All 16 of these complaints allege sharp fare cuts and large increases in seating capacity by incumbents. Four (#9, 10, 13, 16) allege that incumbents added flights to routes they did not previously serve, including nonstop flights that bypass the incumbents' respective hubs. Cases #27 and 28 (categorized as "Other") also involve complaints of unfair pricing and capacity responses. 'These 10 complaints vary, although most concern impediments to airport gate access. Four contend that gates were not made available (#17, 22, 23, 26) due to hoarding by incumbents. Three complain of excessively high lease charges or other unusually high airport or ground-handling fees (such as landing fees) for nonsignatoly carriers (#20, 21, 24). Two complain of the reluctance of incumbents to sell available slots at fair market value (#19,25). One (#22) involves frequent shifting of the gate locations offered to the new entrant. One asserts un- fair rules restricting use of an airport (Love Field) that favors an incumbent hubbing airline (#18). Most of these six complaints consist of assertions that incumbents used their marketing advantages in an unfair and highly selective manner to suppress competition by new entrants. Two contend that travel agents were offered higher commissions for booking flights on incumbents only on those routes challenged by the new entrant (#27, 28). Four claimed incumbents were unwilling to make arrangements for interlining and for joint ticketing and baggage handling services (#27, 29, 31, 32). One (#30) complained of an airport (Pittsburgh) being closed for operations in favor of a new airport that favored the main incumbent airline.

Airline Competition at Hub Airports and Complaints of Unfair Conduct 89 at excessive rates, and the frequent shifting of the gates available to new entrants. In two cases, new entrants complained that incumbents were unwilling to sell airport slots at fair market value. There were six complaints focusing on incumbents' use of marketing and distribution advantages. Two claimed that incumbents offered travel agents higher commission overrides for booking flights in markets chal- lenged by new entrants. Three complained that incumbent-affiliated CRSs listed new entrant flights in a biased manner. Others claimed that incumbents were unwilling to participate in joint ticketing and baggage transfer arrangements for passengers interested in interlining (i.e., trans- ferring between the two airlines for connecting service). Because of the limited information, it was not possible for the com- mittee to assess the validity of all these complaints. Certainly the vague and informal nature of some complaints weakened their credibility. Never- theless, the committee recognized that incumbents could use all of these tactics—limiting access to airports by restricting the availability of slots and gates; influencing CRS listings; and offering special travel agent incentives—to the detriment of smaller rivals, possibly denying them the opportunity to compete fully on the basis of relative costs and the attrac- tiveness of their offerings. Recommendations for correcting these par- ticular matters are offered in Chapters 3 and 4. The committee reviewed more closely the complaints that involved in- cumbents sharply reducing fares and increasing flights and seating capac- ity on specific routes challenged by new entrants. Such actions were the main target of DOT's proposed enforcement policy. About half of the for- warded complaints involved this general kind of response; however, the committee could examine only seven. DOT also forwarded to the com- mittee several additional examples of pricing and capacity responses by in- cumbents that appeared suspicious based on its informal application of the proposed enforcement criteria. The committee reviewed five of these cases in detail; making a total of 12 cases examined, as listed in Table 2-6. The review was by no means intended to be systematic or conclusive, but to provide insight into the kinds of problems that have caused concern. All 12 involved a short- to medium-haul market (less than 1,000 mi) and a major carrier's hub at one or both ends. All of the alleged aggres- sors were incumbent airlines and all of the alleged victims were airlines formed during the past 10 years, also characterized by DOT as low-fare

Table 2-6 Some Possible and Asserted Cases of Unfair Pricing and Capacity Responses by Incumbents to New Entry Quarter Before Entry Entry YR. Average Seats Avg. t.oad aIR City-Pair Fare Available Flights Factor Complaints received by DOT I 96-2 DIW-BOS Incumbent $ 257 227.400 848 57% New Entrant • 2 96-1 ATL-MOB Incumbent 8 186 202,900 700 73% New Entrant - - - 3 95-4 01W-PIlL Incumbent I $ 165 150.100 523 87% Incumbent 2 $ 179 48.800 242 56% New Entrant - - - 4 95-3 PIT-BOS Incumbent $ 130 209,400 858 12% New Entrant - - 5 95-2 MSP-MCI Incumbent $ 201 92.800 407 52% New Entrant • . . - 6 95-2 DFW-ICT Incumbent I $ 111 27,300 430 48% Incumbent 2 $ 128 22,100 482 32% New Entrant - - - - 7 94.4 DFW-MCI Incumbent 1 $ 111 99,500 132 70% Incumbent2 $ 114 60,200 537 57% New Entrant - - - - Other suspect cases identified by DOT using screening criteria 8 97-2 ATh-CIT Incumbent I $ 205 192,800 688 56% Incumbent 2 $ 190 136.200 574 54% New Entrant - - - - 9 96-4 AlL-MCI Incumbent $ 119 183,700 638 80% New Entrant - - - 10 96-1 AlL-PIT Incumbent 1 $ 168 153,600 538 59% Incumbent 2 8 161 90,300 434 59% New Entrant - - - - II 95.1 ATL-D1W Incumbent 1 $ 164 165.600 493 87% Incumbent 2 $ 177 131,900 527 57% New Entrant - - 12 94-3 ATL-DFW Incumbent I $ 185 213,000 179 59% Incumbent2 $ 217 425,000 1100 88% New Entrant • - . - NAn Informalion Is not available due to lack of reporting. a Most recent quarter If 8 quarters have not elapsed. b New entrant (ValuJet) permanenity exited market following suspensIon of operations In 2nd quarter 1998. NOTE: See Appendix D for airport codes.

Table 2-6 continued Second Quarter After Entry Eighth Quarter After Entry a Status of Average Seats Avg. Load Average Seats Avg. Load New Fare Available Flights Factor Fare Available Flights Factor Entrant $ 99 306,700 832 82% $ 232 273,800 675 71% $ 70 12,400 71 27% - - exited $ 112 207.000 725 70% 8 Be 209.800 712 70% $ 42 17,000 75 51% $ 54 45,000 205 34% competing $ 221 133.400 491 57% $ 189 153,700 516 52% $ 233 42.200 210 46% 8 208 81,500 394 53% $ 55 15,100 76 71% - - - - exited $ 135 228,400 788 69% $ 177 202.800 824 74% $ 84 NA NA NA - - - - exited $ 69 141,700 603 70% $ 78 150,300 676 75% $ 43 11,300 30 NA $ 80 43,000 160 55% competing $ 65 47,900 820 58% $ Be 54,200 873 88% $ 126 18,200 510 41% 8 100 15,700 525 66% $ 44 24,300 189 60% - exited $ 91 99,500 779 79% $ Be 139.400 1087 79% $ 68 20,000 176 53% - - $ 50 18,300 148 58% 8 63 21.700 173 70% competing in its proposed guidelines $ 227 203,800 717 70% $ 162 228,600 805 68% 8 87 129,000 592 87% $ 153 128,300 572 56% 8 55 50,800 221 37% - - - - exited $ 111 175,200 828 81% 8 141 188.300 642 78% $ 79 39,700 155 43% 8 129 65,000 289 82% competing $ 93 147,000 544 77% $ 217 144,100 530 58% $ Be 121,100 520 61% $ 203 87.800 395 58% $ 79 40,900 181 50% - - - - exitedb 3 108 204,209 515 64% $ 111 238,400 630 75% 8 93 172,600 894 88% $ 95 173,900 897 71% 8 Be 53.000 237 50% 8 99 31.100 143 55% competing $ 104 188,400 799 78% 8 100 346,800 1283 64% 8 158 471,500 1186 65% 8 115 545,500 1491 72% 8 88 74,400 329 84% $ 97 77,000 333 44% competing carriers. The committee reviewed the chronology of average fares, pas- senger traffic, load factors, ffights, and exit and entry activity. By the second quarter after the lower-priced entry, In 10 of the 12 cases, the new entrant's average fares in the market were at least 50 percent lower than the average fare of the incumbent with the highest market share during the quarter preceding entry.

92 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY In 9 of these cases, average fares for one or more of the incumbents in each market fell by one-third or more. In 4 cases the incumbents' total seats in the market increased by one-third or more. As might be expected, the sharply lower fares resulted in higher load factors for the incumbents, as demand was spurred. Eight quarters after the entry, The new entrant had exited in half the cases, although in one case because of an unrelated circumstance (Valujet's exit following the sus- pension of its operations in the summer of 1996). In the other 6 cases, the new entrant and incumbent were still competing. In 5 of these cases, average fares for both the incumbent and new entrant were much lower than they were during the quarter before new entry. It would thus appear that travelers in these 12 markets benefited, at least initially, from the low-fare new entry and subsequent price-cutting by incumbents. Of the five cases in which the new entrant exited within two years (exduding the Valujet case), at least two involved the incum- bent sharply reducing fares and increasing capacity during the entrant's challenge and then returning to much higher fares shortly after the entrant had exited. In these cases, the fare savings to travelers appear to have been fleeting. Though limited, this sample offers some insight into the problems likely to arise in detecting predation. For instance, it would have been dif- ficult to predict in which of the 12 markets competition would have been driven out and fares restored to previous levels and in which markets com- petition would have continued to the lasting benefit of consumers. Among all 32 complaints, the most difficult to reconcile with normal competition occurred when incumbents "overlaid" the routes of new entrants, particularly by introducing new nonstop jet flights. Four of the complaints forwarded by DOT involved such responses, as noted in Table 2-5. Some involved the incumbent (see complaints 10 and 16 in Appendix C), or its commuter affiliate, introducing point-to-point ser-

Airline Competition at Hub Airports and Complaints of Unfair Conduct 93 vice in a market by bypassing its own hub-and-spoke system. For in- stance, in March 1996, Air South complained that both Continental's and Delta's commuter affiliates had attempted to overlay its new service in three markets: Charleston—Newark, Columbia—Newark, and Myrtle Beach—Newark. Three years earlier, DOT had questioned the motives of Northwest Airlines when it announced plans for nonstop service between Reno and Seattle, Los Angeles, and San Diego. These plans were un- veiled shortly after Reno Air began service from Reno to Minneapolis, which is Northwest's main hub. Since point-to-point service is unusual for a hubbing carrier in moderate-density markets, the introduction of this service, coupled with low fares, certainly suggests exclusionary inten- tions and deserves further review by antitrust enforcers.16 Agency Enforcement Roles The nation's antitrust laws consist mainly of the Sherman Act—the first federal antitrust law—and the Federal Trade Commission (FTC) and Clayton Acts. Predatory pricing is most commonly analyzed under Section 2 of the Sherman Act, which makes it unlawful for a business to "monop- olize, or attempt to monopolize" trade or commerce. As this law has been interpreted, it is violated only if a firm tries to maintain or acquire a monopoly position through unreasonable methods. For a court, a key factor in determining what is unreasonable is whether the practice has a legitimate business justification. DOJ, FTC, state attorneys general, and private plaintiffs can bring suit under the Sherman Act. Civil rather than criminal suits have been the norm for suspected predation violations. The act allows treble damage awards for civil violations, penalties that are generally considered to be a significant deterrent to unlawful con- duct. Antitrust litigation, whether public or private, can take many years to pursue and at significant legal expense. To further deter anticompetitive conduct, FTC was established by Congress. FTC is a consumer protection agency with two basic mandates 16 For economic reasons, major carriers tend to add flights only through their hubs, because the additional flights will connect traffic and raise load factors on other flights. New service that bypasses the hub will divert some passengers from hub flights, reduc- ing load factors. The addition of bypass service, therefore, is unusual, and deserves scrutiny when coincidental with new entry.

94 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY under the FTC Act: to guard the marketplace from unfair methods of competition, and to prevent unfair or deceptive acts or practices that harm consumers. FTC can ftle cases both in federal court and in a special ad- ministrative forum. Section 5 of the FTC Act outlaws "unfair methods of competition" but does not define "unfair." The Supreme Court has ruled that violations of the Sherman Act also are violations of Section 5, which also can cover some practices that are beyond the scope of the Sherman Act. Section 5 empowers FTC to prevent unfair methods of competition and business practices that restrain competition. FTC also shares with DOJ (along with private plaintiffs and states) responsibility for enforcing Section 7 of the Clayton Act, which prohibits mergers and acquisitions when the effect "may be substantially to lessen competition, or to tend to create a monopoly." The two federal enforcement agencies therefore work together on many matters, for instance in developing guidelines for horizontal mergers and in enforcing Section 7A of the Clayton Act (called the Hart-Scott-Rodino Act of 1976), which requires that firms give both agencies prior notification of planned mergers. It is important to note that the airline industry is not subject to FTC oversight under Section 5 of the FTC Act. The Civil Aeronautics and Federal Aviation Acts gave comparable competition and consumer pro- tection authorities to the Civil Aeronautics Board (CAB). These author- ities later were transferred to DOT in 1984, when CAB was abolished. Specifically, the law (49 U.S.0 41712) now states that DOT, on its own initiative or after receiving a complaint, can investigate and decide whether an air carrier or ticket agent has been or is engaged in an unfair or deceptive practice or an unfair method of competition. If DOT, after notice and an opportunity for a hearing, finds that an air carrier is en- gaged in an unfair or deceptive practice or unfair method of competi- tion, it can order a stop to the practice or method. An issue that underlies the current debate over predation in the airline industry is whether enforcement should be handled primarily by DOJ under the Sherman Act. The controversy hinges in part on whether the actions targeted by DOT can and should be forbidden through adminis- trative procedures. As discussed in Chapter 3, DOT has used its admin- istrative authorities to prohibit unfair methods of competition by regu- lating the listing of competing fare and service offerings on CRSs. DOT has defined as unfair and illegal certain CRS information displays that favor particular carriers. DOT's proposed criteria for spotting potential

Airline Competition at Hub Air ports and Complaints of Unfair Conduct 95 predatoiy conduct, however, differ from this usual regulatory approach. The incumbent pricing and capacity responses that the criteria focus on are not types of behavior that are prohibited per se. They are instead de- scribed by DOT as screening criteria for triggering further inquiry. A concern that many major airlines have is that these criteria will evolve into, or have the practical effect of, regulatory standards, causing the pricing and capacity responses they describe to become the defini- tions of unfair methods of competition, possibly inhibiting some truly competitive pricing and capacity responses. Underlying this concern is a sense that administrative rules tend to become increasingly restrictive and rigid over time, in part because of the agency's well-intended ef- forts to be objective and evenhanded in applying the rules and to provide the industry and public officials with guidance about acceptable conduct. A related concern is that such bureaucratic specifications might not dis- tinguish sufficiently among different marketplace circumstances; also that those applying the rules might lose sight of their broader purpose and pursue other, narrower goals—for instance, protecting individual competitors rather than the competitive process. Consideration was not given in this study to FTC's possible role in pre- dation enforcement, since this option falls outside the current framework; however, DOJ's involvement was considered. DOJ's Antitrust Division employs hundreds of lawyers, economists, and industry experts to identif', investigate, and prosecute monopoly practices and other antitrust viola- tions. Because of the breadth of its coverage, it is also expected—and com- pelled—to remain distant from the day-to-day issues and policy concerns of individual industries and sectors of the economy.'7 This detachment is generally viewed as beneficial, reducing the chances that the agency will become too close to the industries it watches over, and making it less sus- ceptible to pressures for favorable treatment from industry constituencies. The involvement of industry- or sector-specific agencies—such as DOT—in antitrust matters is controversial in part because of this con- cern over industry influences. These agencies are immersed in the most routine industry matters, charged not only with regulation but often with dispersing federal aid, operating vital infrastructure, and sometimes with promoting the general welfare of the industry. While this closeness has 17 However, it can participate in regulatory hearings advising agencies on the antitrust implications of their proposed actions.

96 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY the benefit of allowing agency personnel to develop expertise in indus- try operations, a concern is that this same attribute, over time, will cause the agency to become overly deferential to industry viewpoints. The extent to which these risks might apply to DOT's proposed involvement in developing and enforcing prohibitions against predation is a matter of judgment. As explained in the Executive Summary, all of the committee members recognized the risks but differed in assessing them. aUJUVIUMMIM Domestic airlines compete for travelers in thousands of city-pair markets across the United States. The expansion of hub-and-spoke systems since deregulation has led to more competition in more city-pair markets, es- pecially in longer-haul markets. Because of the traffic densities created by hub networks, however, airlines have been able to dominate local traffic on many of the short-haul, nonstop routes, or spokes, emanating from their hubs. It is not unusual for hub-based, nonstop markets to account for two-thirds or more of an airline's total passengers. To be sure, various ef- ficiencies and service advantages are usually the main reasons hubbing air- lines have attained dominant positions. Nevertheless, other factors, such as marketing and ticket distribution advantages, as well as preferential con- tracts with airport operators, also have contributed. The possibility that incumbent carriers are exploiting their dominant positions at hubs and charging monopoly fares has been a concern of pol- icy makers for nearly two decades. Higher average fares in concentrated hub markets compared with unconcentrated hub and nonhub markets have been observed in several studies, including some simplified compar- isons made here. Whether the fare differential is related more or less to in- herent differences in market characteristics and costs or to dominant car- riers exploiting market power cannot be conclusively determined from the data. Nevertheless, the consistency with which hub markets appear among the highest-fare markets is noteworthy and raises the possibility that hub carriers are exploiting market power in ways that would not be sustained if they were subject to more effective competition. The reemergence of low-fare carriers early in the 1990s—and their targeting of high-fare hub markets—seemed to offer a timely check on this problem. Hub markets subjected to significant and lasting low-fare

Airline Competition atHubAirports and Complaints of Unfair Conduct 97 entry have experienced large reductions in fares during the past 10 years. DOT officials, concerned about high fares at hub airports, have viewed low-fare entry as an antidote to hub dominance, and have therefore ex- pressed alarm about incumbent behavior that might suppress or inhibit entry. This concern motivated DOT to propose a means of spotting predatory conduct in the airline industry. The committee could not conduct a thorough review of the complaints and evidence regarding anticompetitive conduct, although a cursory review revealed some actions that were difficult to reconcile with fair and efficient competition. Particularly difficult to reconcile were cases in which incumbent carriers added nonstop service in low- to moderate- density markets they had not previously served directly, coincident with a new entry. In some of these cases, the incumbent bypassed its own hub to initiate the service, a strategy seldom employed outside of high-den- sity markets. The logical inference is that such responses are probably temporary—possibly calculated to protect the incumbent's hub traffic and to dissuade similar challenges elsewhere—and would seem to war- rant additional scrutiny. Incumbents also have been charged with using frequent-ffier pro- grams, travel agent incentives, and other marketing advantages to target and disadvantage new entrants. Other complaints cited incumbents for using their preferential contracts with airports to restrict or withhold gates from challengers and to make airport operations otherwise expen- sive and burdensome. Some also have suggested that incumbents have manipulated CRS listings to put new entrants at a disadvantage in mar- keting and distributing their fare and service offerings. Although the committee was unable to assess the details and validity of these specific complaints, it believes they merit further investigation by DOT. In otherwise exercising its statutory authority to prevent unfair com- petition with regard to predation, DOT must be mindful of the chal- lenges involved in identif'ing and proving predation and of the risk that its enforcement efforts might become increasingly regulatory and pro- tective of individual competitors rather than of the competitive process. With respect to these risks, the committee is concerned that DOT's pres- ent enforcement proposal contains flaws. These include the arbitrary definitions of the specific types of new entrants that deserve special at- tention or scrutiny; moreover, DOT's empirical means of detecting or

98 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY testing for predation depend too much on hypothetical scenarios and speculation about alternative responses. Although consideration of op- portunity cost is important in testing for predation, the difficulty of de- veloping such tests should not be underestimated. Committee members' differing opinions on the seriousness of these risks and challenges, and the best enforcement role for DOT, are explained in the Executive Summary. Notwithstanding these differences, the com- mittee unanimously believes DOT's strategic role should be positive, fostering marketplace conditions that are conducive to entry and more competition. Chapter 3 identifies several such opportunities and recom- mends actions to exploit them. REFERENCES ABBREVIATIONS DOT Department of Transportation GAO General Accounting Office Areeda, P.E., and D.F. Turner. 1975. Predatory Pricing and Related Practices Under Section 2 of the Sherman Act, Harvard Law Review, Vol. 88, pp. 697-733. Borenstein, S. 1989. Hubs and High Fares: Dominance and Market Power in the U.S. Airline Industry. RAlVDJournalofEconomics, Vol. 20, Autumn, pp. 344-365. Comanor, W.S., and H.E. Frech. 1993. Predatory Pricing and the Meaning of Intent. Antitrust Bulletin. Vol. 38, No. 2, Summer, pp. 293-308. DOT. 1990. Secretary's Task Force on Competition in the U.S. Domestic Airline Industry: Pricing. Volumes I and II. DOT. 1996. The Low Cost Airline Service Revolution. Office of Aviation and Inter- national Economics. April, Washington, D.C. GAO. 1988. Airline Competition: Fare and Service Changes at St. Louis Since the TWA- Ozark Merger. Report RCED-88-217BR, Washington, D.C. GAO. 1990. Airline Competition: Industry Operating and Marketing Practices Limit Mar- o ketEntiy. Report RCED 90-147. Washington, D.C. Joskow, P.L., and A.K. Klevorick. 1979. A Framework for Analyzing Predatory Pricing Policy. Yale Law Journal, Vol. 89, No. 2, Dec. Levine, M.E. 1965. Is Regulation Necessary? California Air Transportation and National Regulatory Policy. Yale Law Journal, Vol. 74, July, pp. 1416-1447. Levine, M.E. 1987. Airline Competition in Deregulated Markets: Theory, Firm Strat- egy and Public Policy. Yale Journal on Regulation, Vol. 4, Spring, pp. 393-494. Milgrom, P., and J. Roberts. 1990. New Theories on Predatory Pricing. In Industrial Structure in the New Industrial Economics, Clarendon Press, Oxford. Morrison, S.A., and C. Winston. 1986. The EconomicEffects ofAirline Deregulation. The Brookings Institution, Washington, D.C.

'V Exploiting Opportunities for Airline Entry and Competition Increased opportunities for entry and competition in the airline indus- try depend on the efficient provision of airport and airway capacity to accommodate current as well as new and varied aviation services. Yet ob- stacles persist, including longstanding rules that curb access to some of the country's largest airports. There also are restrictions on how airport operators can raise and invest funds to meet demands for runways, ter- minal buildings, and gate space. Since deregulation, an imbalance has become increasingly evident. The speed and agility with which airlines have been able to respond to changing traveler demands surpass the ability of federal, state, and local governments to expand airport and airway capacity and to adopt pricing mechanisms that induce and ration that infrastructure more efficiently. The nation's airport and airway infrastructure are not the only untapped opportunities for airline entry and competition. Federal law limits foreign ownership and operation of U.S.-based airlines; this is a prominent example of competitive opportunities diminished by govern- 99

100 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY ment policy. Reconsideration of these restrictions—which reduce the flow of capital, management experience, and competing operators into the domestic airline industry—has been recommended by others and is again urged here. The airline ticket distribution system is another aspect of the indus- try that can open or dose more possibilities for competition. The ticket distribution system now consists largely of travel agents using airline- affiliated computer reservation systems (CRSs). However, the advent of Internet reservation systems and other new forms of distribution have spurred changes that could either reduce or augment the system's com- prehensiveness and impartiality. IMPROVING THE USE AND AVAILABILITY OF AVIATION INFRASTUCTURE Deregulation proved that the airline industry could be innovative and efficient when exposed to competition in the marketplace. Not all aspects of the industry, however, have been subject to competitive discipline. Although individual airports compete for flights and passengers, they remain almost entirely under the purview of the public sector. The air traffic control system—administered by the Federal Aviation Adminis- tration (FAA)—is even more insulated, as the only significant industry component that is not free to respond to marketplace demands. Without this freedom, constraints on the supply of airports and nav- igable air space have increased, and their adverse effects have been mag- nified. Chronic airport and airway congestion affect not only the on-time performance of airlines, but also where airlines choose to fly, how they design their networks, and the type of equipment they use. Infrastructure constraints are almost certain to have detrimental effects on competition, particularly on entry and expansion by low-cost carriers.1 Moreover, to compensate for capacity shortages, administrative measures have been adopted, such as hourly quotas limiting airline use of some high-demand airports. However, these administrative remedies have had the unintended effect of creating other obstacles to airline entry and competition. See Chapter 1 for a discussion of how Southwest and some other low-fare carriers depend on intense use of equipment and labor to achieve a competitive advantage.

Exploiting OpportunitiesforAirline Enhy and Competition 101 Market-Based Approaches for Allocating Airport and Airway Capacity A key point made in Winds of Change was that the performance of com- mercial airlines is interrelated with the capacity of airports and naviga- ble airspace. The government's provision of this basic transportation in- frastructure was not keeping pace then and is not keeping pace now. Over the past decade, several other reports—most recently the 1997 Na- tional Civil Aviation Review Commission report—have conduded that capacity shortfalls are likely to worsen unless changes are made in the way this vital infrastructure is provided. When demand for airport and airway capacity exceeds supply, queues and delays develop. Chronic air traffic delays are indicative not only of demand straining against insufficient capacity, but also of inefficiency in rationing infrastructure use. Throughout its history, FAA has handled air traffic control at most airports on a first-come, first-served basis. Its current method of controlling peak traffic demand is through "flow con- trol," which keeps aircraft on the ground awaiting openings in traffic flow. Most commercial airports handle demand in a similar first-come, first-served manner. Little distinction is made among aircraft of differ- ing types, whether a large jet with more than one hundred passengers, or a smaller general aviation aircraft. The exceptions are the major air- ports that are subject to federal quotas on hourly operations, as discussed below. About 400 airports in the United States have FAA towers to provide air traffic control (ATC). Of these, the top 75 airports account for the vast majority of operations (FAA 1998). During the 1990s, the number of commercial aircraft departures increased more than 15 percent (Fig- ure 3-1). Missing from the aggregate data, however, are the operations for- gone because of insufficient capacity or shifted to time periods and airports that are less desirable for travelers. On the basis of FAA's demand projec- tions, such undesired effects can be expected to increase in the years ahead. FAA expects the number of operations by major commercial domestic carriers to grow by 25 percent by 2010 (Figure 3-1). These projections suggest that better use of existing capacity and more responsive means of supplying it are essential. While airport runway expansions and air traffic control modernization could increase capacity, it is also important to take advantage of the underutilized parts of the system.

20,000 18,000 16,000 14,000 12.000 0 0 - - 10,000 - - CL o 8,000 - - 6,000 - - 4,000 - 2.000 0 SOURCE: FM 1998 (Table 27) 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Year Figure 3-1 Scheduled air carrier operations, historical and projected.

Exploiting OpportunitiesforAirline Entiy and Competition 103 For instance, secondary airports in many major metropolitan areas have idle capacity, and even many hubs have extra capacity between connect- ing banks. Regarding this as an opportunity, Southwest Airlines con- centrates its operations at secondary airports and during lulls in hubbing activity at the primary airports where it chooses to operate. Given the ex- pense and practical difficulties of expanding runways and terminals at many congested major airports—partly because of limited space and community opposition to noise—the importance of using existing capac- ity more intensively and wisely is apparent. The fees imposed on airlines and other users of airways and runways do not vary with congestion. Air traffic control is paid for largely through pas- senger and fuel taxes, which have little relation to the cost of supplying and using this infrastructure. For instance, most airport landing fees are calcu- lated based on aircraft weight with no consideration given to the operation's incremental effect on traffic congestion. These fees do not reflect the opportunity cost of using airspace and airports during peak periods, a situ- ation that has contributed to excessive peak-hour use and little direct incentives for airport and air traffic control managers to enhance capacity in the face of high demand. Economists long have maintained that the un- derpricing of scarce resources like navigable airways and runways will result in overuse and undersupply and that airport congestion could be reduced by setting prices that better reflect total costs for the use of this infrastruc- ture.2 For instance, Morrison (1987) and Morrison and Winston (1989), in modeling airport runway pricing from an economic weWare perspective, demonstrated large efficiency gains from the adoption of congestion pric- ing. Other simulations, including Daniel (1991), have found that conges- tion-based airport pricing would encourage airlines to shift their operations away from the peak, reducing congestion. In the committee's view, pricing the use of airways and airports is the most suitable approach for rationalizing these operations, and probably the only long-term solution to ensuring efficient use and supply of vital infra- structure. Precisely how user costs should be calculated and prices estab- lished remains an unresolved issue, although certainly congestion, noise, and other undesirable external effects of operations should be included 2 See VIckeiy 1963 for an early discussion of congestion pricing for roads and other trans- portation facilities. See Levine 1969 for an early discussion of congestion pricing for airports.

104 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY to the extent possible. Winds of Change urged experimentation with— and greater reliance on—pricing to spur more efficient use and provision of airport and airway capacity. This committee agrees that pricing is the most efficient means of rationing scarce airway and airport capacity and prompting more supply. Charging for access to heavily used airports likely will encourage greater use of secondary and reliever airports. Setting fees that reflect the true marginal cost of using congested airports during high-demand pe- riods would encourage those peak users who place the lowest value on flying during these periods to shift either to off-peak times or to nearby secondary airports. General aviation operators also would be more in- clined to use smaller, reliever airports, of which there are more than 3,000 in the United States. This shift could have a significant effect on the availability of capacity at some major commercial airports such as Washington Reagan National, where general aviation accounts for about 20 percent of total airport operations. FAA has tried to encourage gen- eral aviation operators to use reliever airports; charging extra for the use of major airports during congested periods would further this goal. There are many practical difficulties to gaining acceptance of such market-based strategies. Past attempts to raise landing fees to reflect congestion costs have been impeded by legal, political, and contractual obstacles. The most notable example is the effort by Boston's Logan Air- port in the 1980s to impose higher landing fees based on the cost of the resources required to handle the operation (excluding congestion costs, however). The traditional landing fee had been based on an aircraft's weight, which mainly affects runway wear. Boston's experimental fee structure would have generated lower charges for operators of large jet aircraft—which accounted for 60 percent of the operations and carried 94 percent of the passengers at the airport—and higher charges for op- erators of smaller general aviation and commuter aircraft.3 Federal aid to airports, however, has long been conditioned on an open-access policy that prohibits airports from discriminating against classes of users, such as private and small commercial operators. Contending that cost-based See Massport's Program for Airport Capacity Efficiency, prepared by the Massachu- setts Port Authority, Dec. 11, 1987.

Exploiting OpportunitiesforAirline Enhy and Competition 105 landing fees would be discriminatory against users of smaller aircraft, FAA vetoed the Logan Airport proposal. Another practical obstade to runway pricing is that some of the nation's largest airports have lease clauses with their major airline tenants that have the effect of limiting the airport operator's ability to raise fees, including landing charges. One of the reasons for these clauses—as well as FAA's open access policy—is a longstanding concern that major air- port owners, benefiting from their local monopoly position, will raise congestion fees excessively and use the proceeds for purposes other than enhancing capacity for air operations. Developing and implementing appropriate market-based approaches will require some experimentation. This committee could not examine specific pricing options and methods, but it does believe this is the ftin- damentally correct approach and that experimentation should begin. Without pricing to induce supply and manage demand, airport and airway capacity will remain poorly allocated and increasingly rationed through inefficient administrative procedures, queuing, and delays. With air traffic demand growing, these allocation methods inevitably will have to be replaced. Their competition-inhibiting side effects also threaten to become more severe. Airport Perimeter Rules Federal and local rules that limit long-haul ffights to and from three major U.S. airports—Washington Reagan National, New York LaGuardia, and Dallas Love Field—should be eliminated. The rules no longer serve their original purpose and have produced too many adverse side effects, induding barriers to competition. The rules arbitrarily prevent some air- lines from extending their networks to these airports; they discourage competition among the airports in the region and among the airlines that use these airports; and they are subject to chronic attempts by special in- terest groups to obtain exemptions. Perhaps most significantly, the rules have had the undesirable effect of discouraging concerted efforts to find direct and efficient solutions to traffic congestion and noise concerns, undertaken at similar airports elsewhere.

106 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY Background Two major U.S. airports—Washington Reagan National and NewYork LaGuardia—are subject to limits on nonstop arrivals and departures that exceed certain distances. In the case of LaGuardia, the Port Authority of New York and New Jersey has prohibited nonstop flights exceeding 1,500 mi, with the exception of flights to and from Denver. In the case of Reagan National, federal law has prohibited nonstop flights that exceed 1,250 mi. 4 These constraints—known as perimeter rules—have been modified periodically. When first introduced by the Civil Aeronautics Board (CAB) during the mid-1960s,5 the perimeter distance for Reagan National was 650 mi, although some flights to farther cities (e.g., Minneapolis, St. Louis, Memphis, and Miami) were exempted under grandfather provi- sions. Complaints from other communities outside this prescribed limit prompted the Department of Transportation (DOT) to raise the thresh- old to 1,000 mi in 1981. Five years later, Congress codified the basic concept of a perimeter limit at the airport extending it to 1,250 mi under the Washington Metropolitan Airport Act of 1986. The original purpose of these rules was to encourage the use of Dulles for Washington D.C.—area flights and JFK for New York—area flights; both are international airports,' with longer runways and more spacious passenger facilities that could accommodate the larger jet aircraft intro- duced during the 1960s. In addition, this would reduce jet noise at Reagan National and LaGuardia, a growing concern of nearby residents. Under these plans, the two restricted airports would serve primarily local traffic, a higher portion of which involved smaller aircraft suited to short runways and smaller terminals. Moreover, travelers flying longer dis- tances presumably would suffer less inconvenience as a result of the extra ground transportation necessary to reach the more remote Dulles and JFK airports, since the added ground transportation time would account for a smaller portion of the total travel time. Public Law 99-591. 5 The plan was introduced in May 1966 by voluntary agreement among the regulated carriers and was approved as a rule by CAB soon after. 6 The promotion of Newark International Airport also was a consideration in adopting this rule.

Exploiting Opportun itiesfor Airline Entiy and Competition 107 Another major airport affected by perimeter-flight restrictions was Love Field in Dallas. Several years before the construction of Dallas- Fort Worth International Airport in 1974, all the major interstate car- riers promised to abandon commercial service from Love Field. Federal legislation later codified this agreement, confining ffights of large com- mercial aircraft from Love Field to points within the state of Texas and in contiguous states.7 Over the past two decades, access was extended to additional midwestern and southern states. Changing Purpose of the Rules The need to promote JFK, Dulles, and Dallas—Fort Worth airports has diminished as these airports have grown substantially. JFK is the largest international gateway to the United States. Not only has Dulles also become an important gateway, but its use as a hub by United and its proximity to residential and commercial development in Northern Vir- ginia has made it convenient for local travel. As a main gateway and hub for American Airlines,' Dallas—Fort Worth has become one of the most heavily used airports in the country. The Port Authority of New York and New Jersey, which runs JFK, LaGuardia, and Newark airports, and the Metropolitan Washington Airports Authority (which runs Reagan National and Dulles airports) continue to support the perimeter rules to boost domestic traffic at Dulles and JFK. The most prominent advocates of the rules, however, are residents near the airports, who are concerned that relaxed limits will increase jet operations, creating additional noise, especially from heav- ier jet aircraft loaded with more fuel and passengers for longer trips. Criticisms of Perimeter Rules Critics have contended that the perimeter rules misallocate capacity at key airports, leading to higher fares and less convenient service for travelers. Federal legislation codifying this agreement was not necessary before deregulation because Southwest Airlines, an intrastate airline that centered its operations at Love Field beginning in 1972, was not authorized to operate outside of Texas. 8 Newark International Airport is also heavily used for both international and domestic travel as the major eastern hub for Continental Airlines.

108 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY They also maintain that the rules might have industry-wide effects on competition by preventing airlines with hubs located beyond the perime- ter from serving key business markets with nonstop flights. For example, because it operates its main hub in Phoenix, beyond the perimeters, America West Airlines has objected to the rules. In contrast, all of the major incumbent airlines can reach both Reagan National and LaGuardia from major hubs within the perimeters. Restrictions on long-haul ffights at Reagan National, LaGuardia, and Love Field limit competition with other airports in their regions and also among the main carriers that use them. The main airlines at Dulles (United) and Dallas—Fort Worth (American) benefit from diverted traf- fic. Without restrictions, some maintain, Love Field would appeal to more airlines, increasing competitive options for Dallas-area travelers and serving as a versatile second airport for the region, much as Midway Airport does in Chicago. Critics also point to other problems. From time to time, the perime- ter rules have been modified under pressure from special interests. This continuing pressure is likely to generate additional piecemeal exemptions and modifications. These patchwork changes typically have had little to do with achieving a balance of ffights among the Washington, D.C., New York, and Dallas area airports and more to do with the political influence of the airlines and communities seeking airport access. Administrative rules like these provide a standing opportunity for the government to decide which airlines and markets deserve access to key airports—decisions that can be made far more efficiently and equitably on market criteria. Users and operators of other airports close to downtowns and neighbor- hoods, such as Boston, St. Louis, and San Diego, apparently have found ways to accommodate heavy demand and traffic concerns without resorting to administrative schemes for limiting access. Another adverse consequence of the perimeter rules is that they have discouraged the search for similar accommodations in the New York, Washington, D.C., and Dallas regions. Recommendation on Perimeter Rules To enhance opportunities for airline entry and competition, the com- mittee urges the removal of restrictions on long-distance ffights from

Exploiting OpportunitiesforAirline Entry and Competition 109 airports in favor of pncing controls that create incentives to find and adopt more direct means of accommodating traffic demands and other technical and operational factors. Slot Controls Federal rules that set hourly quotas on take-offs and landings at four of the nation's most popular airports hinder competition and should be re- placed with more direct and efficient means of allocating access. Con- tinuing reservations—known as slots—implement these quotas. Because the slots effectively serve as obstacles to competitive entry, they have had the perverse effect of motivating some major airlines to advocate con- tinued constraints on capacity and service at these airports. Like the perimeter rules, these administrative restrictions also have become fre- quent targets for special interests seeking slot exemptions at these air- ports. The three-decade-old slot controls—a makeshift means of ra- tioning airway and airport capacity—have delayed the introduction of more efficient, market-like means of providing and rationing this essen- tial infrastructure. Background on Slot Controls Under the Federal Aviation Act of 1958, FAA has authority to regulate the use of navigable air space in the United States. During the summer of 1968, rising demand, coupled with work slowdowns by air traffic con- trollers, led to long delays and congestion at airports throughout the coun- try, but particularly at Chicago O'Hare and the major airports of the Northeast (DOT 1995, 21-23). Late that year, FAA adopted the so-called high density rule to limit the number of take-offs and landings at Reagan National, LaGuardia, JFK, and O'Hare.9 No longer open to any and all aircraft, these four airports were subjected to hourly quotas on the 9 Newark Airport also was designated for quotas, but the restrictions were suspended be- cause its congestion was primarily due to landside constraints at the airport, not to air- way capacity limitations (DOT 1995, 1).

110 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY number of instrument flight rules (IFR) reservations they could accept.10 The continuing reservations were allocated among three classes of users: Scheduled commercial air carriers operating jet aircraft; Scheduled commuter airlines or air taxis; and All other users—primarily general aviation and charter aircraft. The predominant IFR capacity for each airport determined the num- ber of slots, as prescribed by FAA. Some of the air carrier slots at JFK and O'Hare were set aside for international flights. In all cases, the slot quotas purportedly were based on air traffic management considerations, not on landside constraints such as ta.xiways, gates, and terminals. Though originally a temporary measure to relieve air traffic conges- tion until more permanent system enhancements could be made, the high density rule became permanent in 1973 (DOT 1995, 24-25). At that time, the airline industry was highly regulated, so that the fixed number of slots could be allocated among the handful of airlines with the most extensive, CAB-prescribed route authorities at the affected air- ports. Individual slots were allocated by the airlines, which formed vol- untary committees that decided scheduling by unanimous agreement.11 Although some airlines complained about the quotas, the voluntary process worked fairly well for about 10 years (DOT 1995, 23). The deregulation of the airline industry disrupted this voluntary allo- cation process. The lifting of route franchises rapidly increased the demand for slots by competing airlines, including new carriers seeking to enter mainline markets as well as established carriers seeking to pro- tect and expand their networks. Disputes and deadlocks during the slot allocation process became common and FAA often was called in to reallocate slots through administrative flat. Meanwhile, the 1981 air traf- 10 lnstrument ffight rules define procedures for aircraft operations during low-visibility conditions; most notably, lengthening the distances between aircraft. Because these are conservative allowances for low-visibility conditions, most airports have a greater capac- ityto accept flights during high-visibility conditions. 11 Likewise, voluntary committees of commuter carriers allocated the slots for "air taxis," though general aviation and charter slots were made available on an ad hoc basis through advance reservation.

Exploiting OpportunitiesforAirline Enhy and Competition 111 fic controllers strike prompted FAA to introduce slots at 18 other con- gested airports. During this period, FAA allowed airlines to trade, buy, and sell slots in order to maintain networks and ensure efficient use. Slots at the 18 airports were removed in 1984, but bolstered by this experi- ence, FAA promulgated rules allowing the sale of slots at the four orig- inal and continuing slot-controlled airports, beginning in 1985 (DOT 1995, 26). In adopting the "buy/sell" rule, FAA explicitly acknowledged that scarce slots had become valuable assets to airlines.12 Although the new rule would not make the slots any more or less scarce, it would distrib- ute them more efficiently among carriers that valued them most. Econ- omists had long argued for slot trading, since it would allow the market, rather than government, to decide who should use these valuable resources. It was doubtful that FAA, or voluntary airline committees, could make these determinations in a fair and efficient manner. Allow- ing airlines to buy and sell slots was viewed generally as consistent with the efficient outcomes and competitive processes introduced by deregulation. Though permitting the trades, FAA emphasized that commercial air- line slots remained federal property and could be recalled at any time. Nevertheless, the agency expected an active market. Though perhaps not evident to FAA at the time, one of the reasons for the high demand for slots is the economic value of the price-inelastic business traffic that is accessed by having these key airports in an airline's network. Rather than selling them through auction, FAA elected to give—or grandfather— the slots to airlines that held them at the time of the rulemaking; how- ever, it withheld a small number to reallocate among new entrants.13 Although it acknowledged it had conferred windfall gains on the grand- fathered holders, FAA argued that this imperfect allocation method was the least disruptive way to achieve an outcome that would benefit travelers. 12 See High Density Traffic Airports; Slot Allocation and Transfer Methods, Federal Register, Vol. 50, No. 245, Dec. 20, 1985, pp. 52180-2201. This notice of final rule- making reviewed in detail the rationale for the buy/sell rule and the public comments in favor and opposition. 13 The rule provided 15 percent of slots to new entrants, international operators, and air- lines operating with subsidies from the Essential Air Service program.

112 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY The agency anticipated significant slot turnover as airlines with the most beneficial uses became the highest bidders. However, as a pre- caution against hoarding—that is, acquiring slots but not using them fully—the rulemaking included "use or lose" provisions requiring that the slots be used 65 percent of the time; it later raised this to 80 per- cent. FAA did not anticipate a need to enforce these use-or-lose pro- visions, believing that the slots would be reallocated quickly to the highest-value, most intensive users. The agency discounted the no- tion that a small number of airlines might amass a large number of the slots, inhibiting competitive entry. In response to concerns that air- lines might hoard slots for anticompetitive reasons, FAA argued that the high opportunity cost of underused slots would preclude such behavior, and that even the smallest airline with sufficient use could procure a slot through leasing or secured financing—since the slots could be used as collateral. Competition from the other, non-slot—controlled airports in the New York, Washington, D.C., and Chicago areas was viewed as an additional deterrent to the hoarding of slots by incumbent airlines trying to safe- guard their market power. These market deterrents, combined with the use—or—lose provisions and active antitrust enforcement, were deemed sufficient to prevent anticompetitive consequences. As expected, the rules led to active buying and leasing of slots. Many slot holders today, including financial institutions, do not operate them but lease them to others. The last major change in the slot control rules occurred in 1994. In the FAA Authorization Act of 1994, Congress allowed DOT to grant exemptions from the high density rule to provide essential air service to small communities, international service, and service by new en- trant carriers based on public interest and under other circumstances determined by the Secretary as "exceptional." Reagan National Air- port was excepted from most of these rule changes. Significantly, Congress cited the rules in the legislation, instructing DOT how to grant exemptions. What had started out two decades earlier as a tem- porary measure to alleviate traffic problems at four major airports had now been transformed into a semipermanent rule, incorporated into federal legislation.

Exploiting OpportunitiesforAirline Entry and Competition 113 Effects of Slot Controls on Airline Efficiency and Competition Despite the repeated modifications of slot controls, many fundamental concerns remain. Perhaps the most enduring criticism is that they allo- cate access to key airports on arbitrary distinctions among operators of large jet, commuter, and general aviation aircraft rather than on the most efficient use of the slots. About 20 percent of the operations at LaGuardia and one-third at Reagan National are by nonjet aircraft (DOT 1995). This is because the class-based assignments of slots were left unchanged by the 1985 buy/sell reforms, even though the majority of air travelers fly in larger commercial jets and would benefit from the shifting of an increased share of slots to larger aircraft. Another often-cited concern over slot controls is that incumbent air- lines can hoard these assets to restrict entry and expansion by rivals. New York, Chicago, and Washington, D.C., are among the largest air mar- kets in the country. Direct access to these airports can be important for developing a large hub-and-spoke network, since they are popular ori- gins and destinations for business travelers. The high price paid by air- lines for jet slots—often exceeding 61 million—reflects the willingness of air travelers to pay higher fares to fly to and from these convenient airports. The concentration of slots among a few major carriers also reflects efficient use because of network economies of scope. For instance, United and American Airlines hold the majority of slots at Chicago O'Hare, which serves as a hub for both carriers. US Airways, the major network carrier in the Northeast, holds the most jet slots at Reagan National and LaGuardia. Nevertheless a slot is also more likely to be valuable to a carrier protecting a dominant market position than to a new entrant seeking to compete with a market-dominating incumbent. Therefore the problem might not be that incumbents are hoarding slots and not using them, but that the slot controls themselves have restricted opportunities for significant airport access by rival airlines. For instance, a slot-controlled airport could not become the base for an expanding low-fare entrant, as Atlanta did for Valujet. DOT identified several sim- ilar problems with slots and with possible reforms in its 1995 study of the high density rule (DOT 1995). One common claim, for example, is that slot holders add unprofitable ffights on weekends to satisfr the

114 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY use-or-lose rule; another is that they lease slots to affiliated commuter flights on a short-term basis for the same purpose (DOT 1995). Recent Changes in the Slot-ControiRules In first applying the standard of exceptional circumstances for adding slots, as specified in the 1994 FAA Authorization Act, DOT originally granted exemptions only when there was a significant void in service, such as in markets without previous nonstop service. It adopted this strict standard to minimize increases in air traffic. Yet in response to re- quests from members of Congress, new entrant airlines, and individual communities, in 1997 DOT loosened its criteria for approving exemp- tion.14 Under its new criteria, exceptional circumstances involved Carriers with a "demonstrated potential for low-fare entry"; Single-carrier routes that could support additional entry; and Markets without "meaningful" price competition. Once again, the exemption process had compelled the government— this time, DOT rather than CAB—to make distinctions about which airline markets can support entry, where price competition is "meaning- fiji," which carriers merit special treatment as "new entrants," and which communities deserve additional service. DOT also must entertain filings from other carriers and interested parties stating why applicants should or should not be approved, including assertions by incumbent airlines that exemptions would generate excess capacity or put current slot hold- ers (including grandfathered holders) at a competitive disadvantage. However, approved slot exemptions seldom have been revoked, even after new-entrant recipients had merged with or transferred their slots to larger incumbent airlines. For example, the slots received by Reno Air under DOT's exemption program were transferred to American Airlines when Reno and American merged in 1999. In such cases, it is difficult to avoid the inference that the value of the slots was a factor in these merg- 14 DOT's criteria and rationale for granting exemptions—and the reasons why these cri- teria have been changed—are explained in Order 97-10-16 (October 24, 1997), Office of the Secretary, DOT.

Exploiting OpportunitiesforAirline Entry and Competition 115 ers, or that dispensing slots to spur new competition and expanded ser- vice might be futile. Use of slots for other purposes—for example, to promote new entry— reflects the way the high density rule has evolved beyond its original purposes—to control traffic congestion. In this evolution, DOT's Office of the Secretary has been charged with administering the new slot exemptions, not FAA, the agency responsible for air traffic control and for administering the slot control rules. The effect of added slots on air traffic congestion apparently is no longer an exclusive or even significant concern. However, Congress authorized DOT to grant slot exemptions only in three of the four airports (O'Hare, LaGuardia, and JFK). By not authorizing slot exemptions at Reagan National, where local residents favor controls to decrease jet noise, Congress has acknowledged that noise abatement is a central reason for retaining these quotas. Evolving Purpose of Slot Controic Noise relief now might be the overriding reason for retaining restric- tions. In the case of Reagan National and LaGuardia, virtually the entire day is subject to slot restrictions—even on weekends—mostly because of community concerns about noise, not because of air traffic and airport congestion. At the same time, federal regulations are requir- ing the use of quieter, Stage 3 jet aircraft. By comparison, decreased air safety is not a genuine concern, since safety has never been the main rea- son for slot controls. According to DOT, slot controls do not play a role in air safety, because the air traffic management system employs such conservative practices as "ground delay" flow control (DOT 1995, 3). However, slot controls have influenced the mix, number, and timing of operations at the four affected airports. A DOT study of the high den- sity rule indicated that the early elimination of slot controls would increase operations by 5 to 20 percent at these airports (DOT 1995). The study, which did not assess the dynamic or long-term effects, also predicted that the mix of aircraft operations would change, but the re- sult of lifting the quotas would depend on the new methods of alloca- tion put into effect. It remains unclear, therefore, how the removal of slot controls would affect changes in the number, timing, and mix of operations, and how aircraft operators, air traffic controllers, and airports would respond to

116 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY demand. Airlines and other users certainly could change schedules and types of aircraft. The public's willingness to accept or adjust to varying degrees of delay and inconvenience would affect the search for new methods of allocation. If take-offs and landings at these slot-controlled airports are governed by the first-come, first-served queuing process used by other major U.S. airports, it is conceivable that many travelers would accept additional delays in exchange for increased access to these airports during peak periods. Recurrent delays from heavy demand, however, would prompt direct responses to relieve congestion, including some that would reduce demand during congested travel times. It is unclear, too, whether the removal of slots would disclose addi- tional, previously hidden capacity. The extent to which runways, gates, and other physical and landside limitations preclude higher usage, and whether these constraints can be eased, and how, remain open questions. Like the perimeter rules, slot controls have enabled these airports to postpone the pursuit of other means to manage and accommodate jet operations. Lifting slot controls would perhaps be the only way to as- certain the size of the prpblem and to compel more efficient measures— such as congestion pricing—to expand supply and ration its use. As discussed in Chapter 2, slot-controlled airports consistently are among the highest-priced markets in the country. These airports are main destinations for business travel, especially from many nearby cities such as Rochester, Richmond, and Des Moines, which have repeatedly expressed concern to DOT and Congress about high air fares. In the committee's view, the detrimental effects of slot controls on airline efficiency and com- petition are well documented and are too far-reaching and significant to continue. Studies by the General Accounting Office (1996), the National Commission to Ensure a Strong Airline Industry (1993), and others have urged phasing out slot controls or increasing the number of slots. In its proposal for FAA reauthorization, the Clinton Administration called for the end of slots after a 5-year period at LaGuardia, JFK, and O'Hare air- ports. Other similar proposals are circulating in Congress.15 15 Clinton Administration FAA Reauthorization Bill, unveiled Feb. 8, 1999. This bill proposed the elimination of slots at three of the four slot—controlled airports—Reagan National was the exception—within five years, although regional jet aircraft would be exempted immediately.

Exploiting OpportunitiesforAirline Entry and Competition 117 As discussed earlier, direct economic means of allocating access to these key airports—for example, by peak-period pricing for use of airports or air traffic control services and by supply-side investments in reliever airports—are preferable to slot controls and other administrative schemes for rationing airport and airway access. Recommendation on Slot Controls The committee recommends the early elimination of slot controis, to be replaced by pricing and other market-based methods for allocating and supplying airport and airway capacity to control congestion and other undesirable effects from air traffic. Supply of Gates and Other Airport Facilities Though airports are mostly owned and operated by state and local gov- ernments, they receive significant funding from the federal govern- ment. Partly due to concern that state and local owners might misuse their airport monopolies—for instance, by restricting access by some operators, raising user fees excessively, or diverting revenues to non- airport uses—federal aid has been contingent on adherence to an array of rules about how airport operators can generate, invest, and spend revenues. Adopted piecemeal to address particular concerns, these rules have contributed to longstanding financial relationships between many large U.S. airports and their largest airline tenants. In short, many airports depend on these tenants for financial assistance in build- ing facilities. The availability of these facilities, critical for competitors, can be limited by the financial and contractual agreements between the airport operators and their airline tenants. Coupled with other federal rules stipulating equal access to airports by all prospective users, these agreements have made it difficult for airports to charge fees that reflect the full costs of using runway space. The overall effects of these federal-aid rules on competition merit thorough review. A topic for consideration is whether travelers would be better off if airports had more freedom to raise and spend revenues for en- hancing capacity and managing demand. At the same time, airports that are chronically short of gates and other passenger facilities for use by potential competitors should be prompted by the federal government—

118 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY and even compelled through the withholding of federal aid—to make sufficient facilities available. Background on Gates and Facilities The allocation of airport gates and aircraft parking positions—necessary for enpianing and deplaning passengers, loading and unloading baggage and supplies, and refueling and servicing the aircraft—would seem to be straightforward and uncontroversial. Yet there have been repeated com- plaints that shortages of available gates at some major airports— especially hubs—are an obstacle to airline competition. As with slots, there is concern that incumbent airlines are dominating scarce gate space to the detriment of rivals and potential entrants. Although they might own the passenger loading bridge and aircraft service equipment, airlines typically lease their gates from the airports. Many of these leases are exclusive-use contracts, under which the airport operator can renegotiate terms for gate usage, but day-to-day use and as- signment of ffights is the airline's decision. Airlines can sublease gates to other airlines, with or without airport permission, depending on their contracts. Some airports promote preferential- and joint-use arrangements. Under the preferential-use arrangement, when a gate is not being used by the lessee, the airport operator retains the right to assign it temporarily to another airline. Under the joint-use strategy, gates are rented to more than one airline. A few major airports and many smaller commercial airports operate most of their gates on a common-use basis, under which the airport operator makes all gate assignments. Many airports have a mix of arrangements. Table 3-1 lists the share of gates at major U.S. airports that are subject to these different agreements, based on a 1998 survey by the Air Transport Association. Because of scheduling differences, one airline may have temporarily idle gates when another has high demand and a shortage. Seldom, how- ever, will a gate—holder offer temporary use of its idle, leased gates to another airline—but this reluctance can be legitimate, because if the gate is not vacated in time, it could disrupt a later arrival. Although gate scarcity is to be expected during peak activity, a fre- quent concern is that some airlines might be using their exclusive-use

Table 3-1 Gate Leasing and Use Arrangements at Major U.S. Airports Airport Exclusive Use (% of Total) Preferential Use (% of Total) Common Use (% of Total) ATL 81 0 19 AUS 0 100 0 BOS 80 0 20 BWI 0 80 20 CLE 98 0 2 CLT 80 0 20 CVG 53 47 0 DAL 94 0 6 DCA 0 100 0 DEN 0 71 29 DFW 93 0 7 DTW 71 23 6 EWR 84 0 16 FLL 51 8 41 HNL 0 0 100 HOU 0 100 0 LAD 0 90 10 IAH 84 0 16 JFK 80 0 20 LAS 0 92 8 LAX 58 0 42 LGA 83 0 17 MCI 0 76 24 MCO 89 4 7 MDW 87 0 13 MEM 97 0 3 MIA 0 0 100 MSP 19 81 0 OAK 0 90 10 ORD 87 0 13 PDX 0 70 30 PHL 81 8 11 PHX 95 0 5 PIT 80 12 8 SAN 0 62 38 SEA 79 0 21 SF0 82 5 13 SJC 37 60 3 SLC 96 3 1 STL 14 86 0 TPA 31 44 25 SOURCE: Air Transport Association. NOTES: See text for description of exclusive, preferential, and common use. See Appendix D for list of airport codes.

120 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY contracts to hold onto gates that otherwise sit idle for much of the day—for instance, during the periods between connecting banks. When few or no extra gates are reserved by the airport operator for other airlines, recurrent gate scarcity can become an impediment to entry. To prevent a new rival from gaining a foothold, it is possible that an incumbent airline might withhold gates or offer unattractive terms, such as requiring use of its ground support services (e.g., baggage handling) at premium rates, as discussed in Chapter 2. Other hindrances, such as the relocation of available gates to remote and inconvenient terminal areas, also have been reported. Concern over Leasing and ContractuaMgreements Airports have entered into exclusive, long-term gate leases with airlines for various reasons. Some are old agreements carried over from before deregulation, when a limited number of airlines had authority to oper- ate from individual airports. At that time, the leases might have made sense to airports and airlines alike. Airports were assured long-term ten- ancy, while the airlines could predict and manage their airport costs more easily. Many exclusive-use contracts encompassed more than gates; some included proprietary rights to entire terminals. Airports precluded by federal-aid rules from accumulating reserve funds could use the leases to obtain favorable revenue bond financing for capital investments. In return for being a guarantor of these debts, the airlines demanded and usually received exdusive-use agreements for gates and other passenger facilities. Many of these agreements included "majority in interest" clauses, giving signatory airlines special rights to approve airport capital expenditure plans, including plans for new gates and terminals. This authority to limit increases in operating costs effectively allowed signa- tory airlines to approve new projects and limit landing fees and other air- port user-charges. As backers of substantial airport debt, airlines have maintained that restrictive contractual clauses were essential for minimizing their liabil- ity for excess airport debts unrecoverable through nonairline airport rev- enues, such as concessions. Although these kinds of contractual arrange- ments were more common before deregulation, they continue to be used

Exploiting OpportunitiesforAirline Entry and Competition 121 by some airlines and airports.16 Meanwhile, some of the older agree- ments have many years left before they expire. Other Funding Sourcesfor Gates and Passenger Facilities A source of funds for airport development not tied to airlines is the fed- eral aid provided under the Airport Improvement Program (ALP), ad- ministered by FAA. AlP grants account for 5 to 25 percent of the cap- ital improvement expenditures by most major airports. Aid rules, or grant assurances, restrict the use of these funds, partly out of concern that state and local authorities might apply them to nonairport projects and programs. Airports that receive AlP funds are also required by grant as- surances to avoid unjust discrimination against specific classes of aircraft operators—such as general aviation—and they cannot give individual operators exclusive or preferential rights to fly into and out of the airport. In the committee's judgment, federal-aid rules that narrowly pro- scribe the ability of airport operators to generate and invest funds for air- port facilities have produced outcomes that warrant review as possibly adverse to competitive entry. Perhaps the most illuminating, and trou- bling, example of how these rules can be harmful is FAA's refusal to allow airports to assess fees that ftilly reflect their costs incurred for air- craft operations, as noted earlier in the Logan Airport example. In 1991, Congress recognized that the federal-aid rules had hampered the ability of airports to regulate the use and supply of their facilities. It passed legislation allowing airports to charge airlines a fee known as a "passenger facility charge" (PFC) for each passenger enplanement. FAA previously had prohibited such charges, concerned that airports, and the state and local authorities that run them, would use their monopolies to raise user charges indiscriminately and to divert the revenues to nonair- port uses. The legislation capped PFCs at $3 per head and limited the diversion of the funds to nonairport uses. Airports can use the revenues for a variety of purposes, from improving runways to adding terminals 16 Such arrangements can assure local communities of the continued presence of a car- rier, especially hubbing carriers that provide local employment opportunities and fre- quent nonstop service.

122 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY and gate capacity. The rules stipulate that gates and terminals Rinded through PFC revenues cannot be subject to exclusive-use contracts or to long-term leases. Not all airports have imposed PFCs, and the overall effect of these revenues in increasing gate capacity remains unclear. DOT is studying their effect and considering ways to encourage the use of PFC ftinds for more gate construction.17 Gate Space Problems atHubAirports A 1996 GAO report on barriers to entry in the airline industry identified six major U.S. airports—Charlotte, Cincinnati, Detroit, Minneapolis, Newark, and Pittsburgh—as having a majority of their gates under exclu- sive leases, usually to one or two tenant airlines (GAO 1996). Many of the leases would not expire for several years. In all but one of these air- ports (Newark), the dominant hubbing carriers had exclusive rights to more than two-thirds of the airport's jet gate capacity. GAO suggested that these findings indicated how incumbent airlines were able to achieve and retain a dominant position at some airports, especially at their hubs. Airlines responded that intense use of gate space is a consequence of hub operations, which entail a large number of flights. Nevertheless, GAO recommended that FAA should consider re- ducing federal AlP grants to airports that lease a majority of their gates on a long-term basis to a single airline and that do not have sufficient gates for potential entrants. In the committee's view, limited access to airport gates can be an obstacle to entry that warrants close monitoring; DOT should take reme- dial action when airport operators fail to ensure that gates are being used and supplied efficiently. Recommendations on Access to Gates Airport operators should take steps to ensure sufficient gate supply for competitors, including buying back gates from dominant incum- 17 See DOT. 1998. Request for Public Comment on Competitive Issues Affecting the Domestic Airline Industry. Docket OST-98-4025, Federal Register, Vol. 63, p. 37612 (July 13).

Exploiting OpportunitiesforAirline Enhy and Competition 123 bents, if necessary. DOT, which can identify airports where gate availability is a recurrent problem, should monitor them closely; moreover, federal aid should be contingent on the airport having well- defined plans to ensure sufficient gate supply. At the same time, DOT should review thoroughly its own rules affecting the ability of airports to obtain and spend funds for passen- ger facilities and other capital improvements. OTHER OPPORTUNITIES TO SPUR COMPETITION Foreign Ownership of Domestic Airlines It is a longstanding requirement that airlines operating or based in the United States must be owned principally by U.S. citizens, who must account for at least 75 percent of the voting shares. Many other coun- tries have similar requirements. In the committee's view, removing these strictures on investment and participation in the U.S. airline industry would benefit domestic travel- ers, who currently bear the costs caused by any impediments to compet- itive entry. Any influx of foreign capital and management experience into the domestic industry, affecting small and large carriers alike, would foster more competition, benefiting travelers through more aviation ser- vices and options. In the committee's opinion, the arguments in favor of retaining these limits are not convincing. The most common argument is that owner- ship by U.S. citizens is essential to the military's emergency airlift capa- bilities. Since 1951, the Department of Defense has had agreements with most large domestic airlines to maintain a Civil Reserve Air Fleet program. The concern is that foreign-owned or managed airlines would be less willing to participate. The other main argument against foreign ownership is that the United States must retain these limits on foreign ownership as long as other countries do the same. In this way, ownership limits can be lever- age in negotiating for open access for U.S. airlines and investors in foreign internal markets. In the committee's view, preserving these limits to ensure civil airlift ca- pabilities is unwarranted. Other options are available to provide this im- portant capability, including direct payments to carriers. Moreover, it is

124 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY unclear why U.S. ownership affects the integrity and performance of the civil airlift program. In addition, retaining these limits to prompt foreign countries to open their domestic markets to U.S. competition is similarly questionable. U.S. domestic travelers would be the beneficiaries of more open entry into U.S. markets, just as foreign travelers would benefit from more open entry into their domestic markets. Paradoxically, U.S. policy is withholding the benefits of competition from domestic travelers to bar- gain for open domestic markets abroad, which primarily would benefit foreign travelers. Recommendation on Foreign Ownership Restrictions against foreign citizens owning and operating U.S.- based airlines should be lifted. Airline Ticket Distribution System The system for distributing air fare information has changed signifi- candy since deregulation. Travel agents using CRSs are now the pre- dominant source of fare information, reservations, and ticketing. This system is considered beneficial to travelers, providing a comprehensive and impartial channel of information on competing fare and service op- tions. Over the past two decades, federal regulators have recognized this benefit and sought to preserve the neutrality and integrity of the system. Certain practices, however, such as any extra commissions provided by airlines to agents generating a large volume of business, remain subjects for concern. The possibility that airlines are becoming more adept at in- fluencing the advice travel agents give to their clients deserves public poi- icy attention. Yet the overall distribution system is dynamic, and the re- lationships among agents, airlines, and reservation systems are changing fast. The advent of Internet reservation websites and other computer- ized, alternative forms of distribution have presented opportunities and risks. These changes can reduce distribution costs and increase system efficiency to the benefit of travelers. However, the changes also could have negative effects, reducing the completeness and impartiality of the distri- bution system and raising consumer search costs. With so many changes ongoing, the committee is reluctant to offer specific advice. Monitoring these developments, nevertheless, seems in order and appropriate.

Exploiting OpportunitiesforAirline Entiy and Competition 125 Background on the Ticket Distribution System With the deregulation of the airline industry, complex fare and service offerings have proliferated. Faced with an unprecedented variety of choices, travelers have relied on travel agents to book their trips. Before deregulation, travel agencies accounted for about one-third of the tick- ets purchased for air travel; today more than 80 percent are purchased through these intermediaries (ASTA 1997). To the consumer, the travel agent ideally represents a neutral source of information on airline fares and schedules. Although agents custom- arily have been paid a commission based on the ticket price—ostensibly an incentive to sell higher-priced tickets—most travelers assume they are acting in the best interest of the buyer.18 Travel agents portray them- selves as objective intermediaries, or even as consumer advocates, search- ing out the most convenient flights and lowest fares within the limits defined by the traveler. In a highly competitive and fragmented indus- try relying on repeat business, the agents maintain that consumers can expect to receive even-handed service. In an airline's view, however, a travel agent sells its product, becoming an integral component of its marketing and distribution system. Because agents influence consumer choices, airlines have reason to induce agents to promote their fare and service offerings. At the same time, airlines want to avoid paying commissions when they can. To the consumer, the commissions paid by airlines to travel agents are not important—travelers usually can obtain the same ticket at the same gross price from the agent as from the airline. However, this situation is changing. Using Internet websites, airlines are increasing efforts to sell their products directly to consumers by offering discounted fares not ob- tainable through travel agents. Meanwhile, more agents are charging customers a booking fee to compensate for declining commissions and the smaller commissions from low fares, which nonetheless can be time- consuming to find and procure. ' The commissions—ranging from 8 to 10 percent of the ticket price—are now capped at $50 to $100, which lessens the incentive to sell higher fares. Caps and commissions differ, depending on the airline and whether the travel is domestic or international.

126 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY Role of Computer Reservation Systems Because of these sometimes-conflicting incentives, the relationship between the travel agent and airline industry has often been questioned. A main concern has been that airlines can use CRSs to bias the informa- tion given to consumers and to facilitate incentive plans for travel agents. These computer systems—there are several in competition—allow agents to check airline schedules and seat availability, book flights, issue tickets, and change and cancel reservations. With the notable exception of Southwest, most airlines participate in all of the CRSs. The airlines pro- vide their fare and schedule offerings to central databases that distribute the information to the CRS vendors and then to the thousands of travel agencies around the country. Virtually all agents, including Internet-based agencies, use CRSs. The distribution system has bestowed large benefits on consumers. An agent using a CRS has up-to-date information on the fare and flight offerings of nearly all airlines in a given market, displayed in a format that can be quickly scanned for competing options. Not only does the CRS provide the fares of major airlines, but also those of smaller partic- ipating airlines. Although airlines pay a booking fee to the CRS vendor and a commission to the travel agent, access to this extensive distribu- tion system is critical to smaller airlines with minimal name recognition and limited resources for other distribution and marketing methods. Consumers gain from the ready, comprehensive information on com- petitive fare and flight options. For the most part, domestic CRSs were developed by the major domestic airlines, and most continue this affiliation. Travel agents gen- erally subscribe to a single vendor. After deregulation, the host airlines used these systems to influence traffic. In particular, the algorithms used to display ffight offerings on CRS screens listed the host airline's offer- ings ahead of competitors'. Likewise, the booking fees charged by ven- dors often varied significantly by airline. Carriers that did not own a CRS frequently were charged the highest fees, especially in markets where they competed head-on with the airline that owned the CRS. This prac- tice was particularly harmful to the new and smaller airlines that had unfamiliar brand names and modest advertising and marketing capabil- ities, depending on a CRS to reach customers.

Exploiting OpportunitiesforAirline Entry and Competition 127 In response to these problems, in 1984, CAB—in its last major action—issued rules prohibiting unfair and anticompetitive CRS vendor practices. These requirements included unbiased displays of fare and flight information and availability to all airlines on a nondiscriminatory basis. In a 1990 review, DOT conduded that the new CRS rules had lessened the biases and anticompetitive effects of CRS ownership by airlines sig- nificantly. DOT noted, however, that certain problems remained, partic- ularly the so-called "halo" effect—several studies had demonstrated that airline owners of CRSs benefited from higher bookings by travel agents who were their subscribers. This was attributed to several causes—some innocuous—including the tendency of agents to have more confidence in the information provided by the host airline (especially on last-seat avail- ability) and also to subscribe to the CRSs of the airlines they traditionally had booked. Another suspected cause was the cooperative business rela- tionships that developed between the airline-owners of a CRS and the agents who subscribed; possibly these relationships induced the agents to emphasize the host airline's offerings. To help remedy this tendency, DOT prescribed additional rules in 1992 to make it easier for agents to use third—party hardware and software to access CRSs. Recent Developments in Distribution Concerns about the halo effect and CRS bias in general have declined in recent years as attention has shifted to the many other changes in the dis- tribution system. For instance, airlines are selling tickets through their Internet websites or through auctions conducted on the Internet; offer- ing fare discounts and rebates by working directly with corporations and other frequent purchasers; and forming alliances to reduce interlining and make it easier to book international trips and other complicated itin- eraries through a single airline. Because travel agent commissions represent the second or third largest operating cost for airlines after labor and sometimes friel, carriers are searching for alternative forms of distribution. By reducing commissions, they have compelled many agents to charge consumers directly for their services, perhaps inducing some travelers to purchase their own tickets directly. Southwest Airlines, for instance, traditionally has avoided travel agent sales, using in-house distribution methods that target repeat buyers.

128 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY Website distribution also offers the prospect of reducing labor- intensive, in-house reservation systems. "Net" fares—usually offered to a large company, net of travel agent commissions—transfer the distri- bution costs to the customer, who might be able to reduce the transac- tion costs through highly-tailored, contracted, or in-house corporate travel management services. Airlines can segment demand by offering fares on websites, since trav- elers attracted there would be loyal customers. Last-minute Internet pro- motions and auctions of empty seats further separate price-sensitive travelers, who must plan in advance, from travelers who are able and willing to travel on impulse or on short-notice if the fare is right. Either way, with direct Internet bookings, the airline also avoids having to pay travel-agent commissions. A possible outcome of these developments is that retail travel agents— and the CRSs they employ and rely on—could become less important in the distribution system. For example, carriers distributing tickets through their websites and other direct means might list a smaller por- tion of their available seats on CRSs; this in turn would reduce the com- pleteness of the information provided in these systems and possibly the value to consumers. This loss of market share also could make CRSs less appealing to airlines—which might have some positive effects on com- petition, but also might hasten the obsolescence of CRSs if the airlines fail to maintain the databases. Although they might benefit from the many new forms of distribution, consumers also might lose an impor- tant source of impartial and comprehensive information if the CRS infrastructure was allowed to deteriorate. TraveMgent Incentives Winds of Change expressed concern that consumers typically are unaware of the financial relationships between many travel agents and airlines, particularly the practice of rewarding agents with extra commissions, or overrides, for meeting sales quotas on particular routes or overall sales levels.19 Although recognizing that some of these volume-based over- 19 The payments are in addition to the base commissions, which usually are established as a percentage of the ticket price.

Exploiting OpportunitiesforAirline Entiy and Competition 129 rides are returned as fare rebates to the traveler, the Winds of Change committee nevertheless recommended a requirement that agents must disclose any financial incentives to their customers. This committee agrees with that recommendation. Maintaining the neutrality of the travel agent ticket distribution sys- tem is important. There are indications, however, that airlines continue to search for ways to influence agents. According to a recent report by the DOT Inspector General, customized CRS software add-ons, or enhancements, have been developed by independent vendors to restruc- ture screen displays, emphasizing sales on airlines offering overrides.20 Some carriers also are purchasing and using CRS marketing information and sales data to track the percentage of a travel agent's business directed to rivals, permitting the airline to offer additional commissions and other compensation to agents who generate or surpass a specific market share. Smaller and low-fare airlines presumably cannot offer similar incentives. For example, incumbent airlines can limit their commission incentives to incremental travelers only, while the smaller, nonhubbing carriers must raise commissions on travel agent sales for all passengers. Although a case can be made that volume-based commission overrides, if passed along, bestow some benefits on consumers, the benefits from airlines moni- toring or policing compliance by travel agents seem much more ques- tionable. Recommendations on the Ticket Distribution System Aggressive efforts by airlines to police travel agent sales deserve further scrutiny, and might warrant new rules requiring public disdosure of extra commissions and other targeted incentives that can prejudice agents. In general, however, changes in the distribution system should be viewed as opportunities to enhance the system's overall benefit to consumers, and should not be dissuaded unless the neutrality and completeness of the distribution system is fundamentally threatened. DOT should remain alert to the possibility of such erosion. 20 See DOT 1999. Report on Travel Agent Commission Overrides, Report CE-1999- 0609, Office of the Inspector General, March 2.

130 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY SUMMARY Increasing airport and airway capacity is essential to opportunities for competition and entry in the airline industry. In the long—term, ensuring adequate capacity will require an emphasis on charging users a cost-based price—including congestion costs—for access to this infrastructure, both to manage demand and to provide sufficient supply. This also should prompt the development and introduction of capacity-enhancing tech- nologies. Part of the supply shortage must also be met through invest- ments in additional physical capacity and in secondary airports, to relieve the congestion at airports that can neither expand nor provide sufficient opportunities for competing services. The current system of queuing for access to airports and air traffic control services results in delays and is likely to get worse, unless more responsive ways are found to regulate demand and supply. Administra- tive schemes to limit and allocate use of some high-demand airports, without user pricing, have had the adverse effect of limiting competi- tive opportunities. Similarly, federal rules intended to prevent airports from charging monopoly rates for services have contributed to the undesirable outcome of curbing airports' sources of revenue, causing some airports to depend on financial relationships with their main air- line tenants, to the detriment of their tenants' potential competitors and of consumers. The committee urges the application of federal and other funds to expand airport and airway capacity, particularly by investing in capac- ity-enhancing technology. The goal should be to use pricing both to finance expansion and to allocate capacity more efficiently. Both technology and pricing should be employed to encourage additional ffights to and from underused secondary airports in major metropoli- tan areas. The following complementary actions are also recommended to enhance competitive opportunities: Introduce pricing methods in place of administrative restrictions to manage airline access to some of the country's major airports. The emphasis should be on the early substitution of pricing for current slot controls and perimeter limits on long-haul ffights, with the goal of al-

Exploiting OpportunitiesforAirline Entry and Competition 131 locating scarce airport and airway space more efficiently and fairly among competing airlines and taking into account other technical and operational factors. Ensure that federal rules for airport funding and spending au- thority do not conflict with the goal of increasing gate availability at major airports but are used positively to achieve it. End limits on foreigners owning and operating U.S.-based air- lines. These recommended actions will create an overall environment in the airline industry that is more conducive to market entry and competitive services. However, these actions must be accompanied by government vigilance fostering fair and vigorous competition. A vital part of this gov- ernment role is to ensure a well-frmnctioning, efficient, and unbiased sys- tem for airlines to distribute their fare and schedule offerings and for consumers to obtain comprehensive and impartial information. In this regard, the committee urges further scrutiny of aggressive efforts by airlines to police travel agent sales, which might necessitate new fed- eral rules requiring public disclosure of extra commissions and other targeted incentives that can prejudice agents. REFERENCES ABBREVIATIONS ASTA American Society of Travel Agents DOT Department of Transportation FAA Federal Aviation Administration GAO General Accounting Office ASTA. 1997. Competition in Distribution of Travel Services. The Future of Travel Agents, Alexandria, Va. Daniel, J. 1991. Peak-Load-Congestion Pricing of Hub Airport Operations with Endogeneous Scheduling and Traffic-Flow Adjustments at Minneapolis-St. Paul Airport. In Transportation Research Record 1298, TRIB, National Research Coun- cil, Washington, D.C., pp. 1-13. DOT. 1999. Report on Travel Agent Commission Overrides, Report CE-1999-0609, Office of the Inspector General, March 2. DOT. 1995. Study ofthe High Density Rule: Report to Congress. Washington, D.C., May. FAA. 1998. FAA Aviation Forecasts: Fiscal Years 1998 to 2009. Report FAA-APO-98-1. Washington, D.C.

132 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY GAO. 1996. Airline Deregulation: Changes in Aiifares, Service, and Safety at Small, Medium-Sized, and Large Communities. Report RCED-96-79. Washington, D.C., April. Levine, M.E. 1969. Landing Fees and the Airport Congestion Problem. Journal ofLaw and Economics, Vol. 12, pp. 79-108. Morrison, S. 1987. The Equity and Efficiency of Runway Pricing. Journal ofPublic Eco- nomics. Vol. 34, pp. 6 1-85. Morrison, S., and C. Winston. 1989. Enhancing the Performance of the Deregulated Air Transportation System. Brookings Papers: Microeconomics 1989. The Brookings Institution, Washington, D.C. National Civil Aviation Review Commission. 1997. AvoidingAviation Gridlock A Con- sensusfor Change. Dec. National Commission to Ensure A, Strong Competitive Airline Industry. 1993. Change, Challenge, and Competition:A Report to the President and Congress. Washington, D.C., Aug. Vickery, W. 1963. Pricing in Urban and Suburban Transport. American Economic Review, Vol. 53 (Papers and Proceedings), pp. 452-465.

fl V Effects of Airline Alliances and Partnerships on Competition The series of mergers and failures of domestic carriers from the mid- 1980s to the early 1990s led to ominous predictions that the decade would end with a more concentrated industry consisting of a few large airlines exploiting their market power. Winds of Change, released shortly after several major airline mergers and failures, expressed concern that these trends might continue, possibly jeopardizing the many consumer benefits achieved from deregulation. These concerns may have been exaggerated or premature. The indus- try did not become significantly more concentrated during the 1990s, despite the failure of some large incumbent carriers (e.g., Eastern and Pan American) early in the decade. Unanticipated developments—such as the expansion of Southwest Airlines and the emergence of many other low-fare carriers—have counteracted any industry tendency toward concentration. In 1992, 95 percent of domestic passenger trips were on the top 10 airlines, and 70 percent on the top five. These percentages have declined slightly. By 1998,91 percent of trips were on the top 10 air- 133

134 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY lines, and 67 percent on the top 5 (Table 4-1). Likewise, 78 percent of the revenue was generated by the top 5 carriers in 1992, compared with 71 percent in 1998. These declines occurred mainly during the first half of the decade; levels have been stable since. Though industry-wide trends do not indicate increased concentra- tion, trends at the airport and city-pair levels suggest slight increases. As discussed in Chapter 2, many hub markets remain highly concentrated and the number of medium- to high-density city-pair markets with three or more competitors declined during the 1990s. Concentration at the city-pair level—where airlines compete for passengers—raises the pos- sible exertion, or exploitation, of market power. For various reasons—induding the difficulties of integrating union- ized work forces—future consolidations of airlines might eschew out- right mergers and take more indirect forms, such as strategic "codeshar- ing" alliances, joint ventures, and other collaborations and partnerships. Codesharing agreements among major carriers, commonplace in inter- national aviation, also have taken root in the domestic sector, as have cer- tain other partnerships, such as shared frequent-flier programs. These developments, both international and domestic, suggest how the indus- try might evolve in the next decade. BACKGROUND AND KEY COMPETITION IS SUES Airline alliances are not a new phenomenon. Major carriers have shared their codes with regional and commuter airlines for more than 30 years. Table 4-1 Share of Domestic Passenger Trips and of Revenues for Top 10 U.S. Carriers 1992, 1995, and 1998 Percentage of Industry Total Trips Revenue 1992 1995 1998 1992 1995 1998 Top 3 Carriers 63 41 45 53 49 49 Top5 Carriers 70 65 67 78 72 71 Top7 Carriers 85 78 81 91 90 85 Top 10 Carriers 95 91 91 98 95 95 SOURCE: Database Products. Inc.

Effects ofAirlineAiiances and Partnerships on Competition 135 These "vertical" affiances benefited consumers and did not involve air- lines that competed or were likely to compete in the same markets. What is new over the last half-dozen years is that codeshare agreements and other airline partnerships have become common in international avia- tion, and now are being adopted by some large domestic carriers. The competitive implications of these partnerships, sometimes involving rival airlines, has become an issue only recently; the long-term implications for competition in the airline industry remain open questions. Early Commuter Codeshare Agreements In the codeshare's earliest and simplest form, a commuter carrier would adopt the two-letter identifier of a major carrier to list its flights in the OfficialAirlines Guide (OAG), and later in computer reservation systems (CRSs). By relinquishing its own code, the commuter carrier could have its offerings listed prominently as online itineraries in the CRS, increas- ing its traffic from travel-agent bookings. The major carrier would gain from an additional online source of feed traffic for its hub-and-spoke system without having to serve low—volume commuter markets directly at its own higher cost structures. Commuter codesharing started in 1967, when Allegheny Airlines, obli- gated by the Civil Aeronautics Board (CAB) to serve some unprofitable short-haul routes, shared its codes with commuter airlines operating smaller, less costly propeller aircraft. Since then, codesharing agreements between major and commuter airlines have become more comprehen- sive and common. Just about every commuter carrier now has a code- sharing agreement—often exclusive—with a major carrier. Frequently, the commuter airline adopts a variation of the major carrier's name (e.g., Delta Connection, US Airways Express, American Eagle), along with common aircraft color schemes, flight attendant uniforms, and ticket stock. Commuters participate in the major carrier's frequent—ffier pro- grams; marketing, advertising, and reservations are often handled by the major airline; and revenues are allocated among the partners according to an agreed formula. Many commuter airlines have become—or act as—subsidiaries of their larger codeshare partners. Such integrated relationships offer economies of traffic density and allow for tighter connections and coordinated baggage handling, check-

136 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY in, and ticketing, as well as other less tangible benefits, such as the en- hanced service level and safety requirements major carriers can impose on their codeshare partners. Because only some of these results could be achieved under the once-common "interline" agreements between inde- pendent commuter and major carriers, these more integrated relation- ships generally are viewed as favorable to travelers. Only rarely did the commuter and major carrier compete with one another for traffic in the same markets. As codeshare arrangements proliferated during the 1980s, however, some observers expressed concern that carriers without these agree- ments—especially unaffied commuters and larger start—up and secondary carriers at major hub airports—effectively would be denied flow traffic, not only resulting in increased dominance by the hubbing carrier but also in a decline in independent commuters (Winds of Change, 288-289). The latter has certainly occurred—unaligned commuter airlines have become rare, and the total number of commuter airlines has declined by more than half since 1981 (FAA 1998, IV-6). International Codeshare Affiliations and Partnerships In much the same manner as domestic codeshare affiliations, arrange- ments between U.S. and foreign airlines can benefit travelers in smaller markets by connecting nonrival networks. But foreign airline code- sharing also has raised many of the same concerns as those of domestic affiliations—plus some others. Nearly all large foreign carriers offering international service—as well s many smaller foreign carriers that focus on domestic and connecting service—have entered into codesharing or other marketing agreements, such as shared frequent—flier programs, with U.S. carriers. Some of these are highly integrated affiliations, involving not only schedule coordination but sharing aircraft, ground support, reservations, and marketing programs. As with the major U.S. carriers and their affiliated commuters, many of the participants in international alliances do not normally compete with one another. For instance, a U.S. carrier might share its codes with a foreign airline that offers connecting service within its own country or region. Denied an opportunity to carry this connecting traffic itself—perhaps because of

Effects ofAirlineAiiances and Partnerships on Competition 137 national restrictions on ownership and entry (or cabotage)—the U.S. carrier might find it advantageous to coordinate with foreign airlines. There are some fundamental differences, however, between the inter- national and commuter codeshare agreements. Besides scale, the most obvious difference is that many international alliances involve airlines that compete with one another or have done so in the recent past. Neither car- rier relinquishes its own code. The international codeshare arrangement described earlier—in which a U.S. international carrier shares its code with a foreign airline serving internal domestic routes—is rare, because most foreign airlines also operate internationally, sometimes in the same gateway-to-gateway markets as their U.S. partners. An undesirable con- sequence, therefore, might be a decline in rivahy on some overlapping routes, especially on the dense, gateway-to-gateway, intercontinental segments. Large international operators also can become significant global competitors in the future, especially if national restrictions on for- eign ownership and entry are substantially eased or lifted. In contrast to airlines in a typical commuter code-sharing arrangement, partners in the international arena are more likely to be actual or potential competitors. Domestic Airline Partnerships In most initial cases of domestic codesharing involving major U.S. air- lines, the partners were not competitors in many markets—and certainly not in the markets where codes were shared. The codeshare routes have consisted mainly of lightly traveled city-pairs, requiring transfers at two hub airports, each operated by one of the partners. Continental and America West entered into a limited codeshare arrangement in 1994, when the two linked their networks to serve small, peripheral city-pair markets such as Tucson, Arizona, to Portland, Maine. However, the 1998 agreement between Northwest and Continental—the fifth and sixth largest in the United States in terms of passenger trips—to share codes on many routes introduced a much more comprehensive arrangement (GAO 1999). Whether other carriers will pursue similar codeshare alli- ances is unclear. After the Northwest-Continental plan was announced, several other major U.S. carriers unveiled intentions to share codes on ffights (e.g., United-Delta and American-USAirways). However, none of these other codeshare plans involved equity transfers, unlike the

138 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY Northwest-Continental plan—and none has developed ftirther; yet each opens up the possibility of two major competitors joining on routes between their hubs. Many partners share their frequent-flier programs. Typically, recip- rocal programs allow members of both frequent-ffier programs to earn miles on either, although restrictions vary on whether miles can be com- bined or redeemed on all flights, including international service. Joint availability of some other services, such as airport dub lounges, is also common. Key Competition Issues As noted, adverse effects on travelers in nonstop, gateway-to-gateway and hub-to-hub markets—where now-allied carriers once competed—are a potential drawback of codesharing, and have received much attention. However, less attention is given to the possible cumulative effects of these alliances, both domestically and internationally, on the U.S. domestic industry, particularly on the viability of unaffihiated U.S. airlines. Although alliances forged with equity transfers are routinely subject to merger review by the Department ofJustice (DOJ), there are no dear guidelines for when domestic codeshare agreements that do not involve such holdings—or when other domestic collaborations, such as recipro- cal frequent-ffier programs—should be reviewed under the merger prin- ciples. A concern is that some of these domestic affiliations might lead to outright or de facto mergers with potentially negative consequences for consumers. A longer-range issue that perhaps has received the least attention is how the global alliances among some of the world's largest carriers will affect industry competition if national and international restrictions on ownership are lifted. ALLIANCES AMONG MAJOR DOMESTIC CARRIERS Airline mergers were commonplace in the 1980s. At that time, the Department of Transportation (DOT) had authority to review mergers and allowed many to proceed, induding several involving failing airlines as well as carriers competing in many of the same airports, cities, and city-pair markets. The 1986 mergers of TWA with Ozark and North-

Effects ofAirline Alliances and Partnerships on Competition 139 west with Republic are prominent examples involving competitors— TWA and Ozark shared a hub in St. Louis, and Northwest and Republic shared hubs in Detroit and Minneapolis. DOJ, which had objected to those two mergers, since has been given authority to review airline mergers and acquisitions. Although DOJ has not disapproved of mergers between carriers that have no significant car- rier network overlap—for example, the purchase of Reno Air by Amer- ican Airlines—it generally has opposed mergers when networks overlap or when the two carriers could develop into significant competitors in the future. Winds of Change endorsed this approach, recommending that "DOJ oppose mergers or asset acquisitions in which the carriers offer substantial parallel service or share a hub airport; however, DOJ should not necessarily oppose mergers or asset acquisitions of carriers with com- plementary or end-to-end routes" (p. 17). Early Limited Codesharing Agreements The &st significant codeshare affiliation between two large jet operators— the arrangement between Continental and America West—involved sev- eral dozen low-volume routes in which neither carrier previously had competed against the other. Alaska and Northwest subsequently entered into a similar partnership on a small number of low-density routes. In these codeshare agreements, neither partner previously had operated over the entire route, making the alliances essentially vertical, or end-to-end, with no overlaps in service and limited potential for competition. Trav- elers flying on codeshare itineraries normally transferred twice, once at each of the partners' hubs, benefiting from the link between the two hub- and-spoke networks. These agreements did not raise antitrust concerns. Proponents of limited codesharing point to two kinds of consumer benefits. First, codeshare partners can coordinate connections, baggage handling, and other services more effectively than under interline arrangements. Second, the codeshared itinerary would appear on CRS displays as an online service, preferred by travelers and listed ahead of interline offerings. In markets where there already was true, single-carrier online service, the creation of another ostensible online option, it is argued, could spur competition among airlines, benefiting travelers. Of

140 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY course, some travelers might feel misled by the portrayal of interline ser- vice as online. From the standpoint of the codesharing airlines, other advantages are possible. A preferential online listing for each of the partners could dis- place other options from schedule displays and lead to more bookings from travel agents. More Extensive Domestic Codesharing Agreements Northwest's planned ownership position in Continental in 1998 raised the possibility of more extensive codesharing among the nation's fifth (Northwest), sixth (Continental), and ninth (America West) largest air- lines in terms of passengers carried. Starting in 1999, Northwest, Con- tinental, and America West began sharing codes on additional domestic and international routes. However, these arrangements have not involved nonstop routes between the hubs of the three airlines.1 Unlike earlier codeshare agreements, the Northwest-Continental- America West arrangements have involved many markets, usually one- stop destinations, where one or more of the carriers already had offered through-service. In these cases, no completely new online service has been created, and the codeshare has not introduced new competition. Table 4-2 shows the scheduled OAG listings created by a codeshare among Continental, Northwest, and America West in a one-stop mar- ket. In this example, Continental already had offered service in the mar- ket through its Houston hub. Rather than create new online services, the codeshare simply tripled the number of schedule listings for Continen- tal's flights. Only one new online connection was created (the 1:30 p.m. ffight) through a combination of Continental and Northwest flights. In the committee's opinion, it is plausible that consumers benefit from codesharing by major carriers on low-volume routes in which neither partner previously had offered service over the entire route and was not likely to do so. Such agreements might lead to more convenient con- 1 The carriers may have abstained from agreements partly because of a suit by DOJ seek- ing to block Northwest's acquisition of Continental and also because of concerns raised by DOT.

Effects ofAirline Alliances and Partnershps on Competition 141 Table 4-2 Example of Domestic Codeshare Schedule Listing in OfficiaMirline Guide OAG FlightDisk - Worldwide Edition Effective 5/1/99 thru 6/14/99 CONNECTIONS Outbound Date: From: Corpus Christi, Texas 5/19/99 To: Wayne County Airport - Detroit Airlines: All PAU Connectina Fliehts ]st Arrive Connect Duration XThSaSu - CO 1888/CO 174 6:00a CRP 11:24a DTW IAN 04:24 XThSaSu HP*1888/HP* 174 6:00a CRP 11:24a DTW IAH 04:24 XThSaSu NW*7888/NW*774 6:00a CRP 11:24a DTW 11tH 04:24 Daily DL 4571/DL 294 6:00a CRP 11:52a DTW ATL 04:52 Su WN 8/WN 487 6:55a CRP 3:50p DTW LIT 07:55 Daily CO 3712/CO 174 9:00a CRP 2:22p DTW tAll 04:22 Daily HP*3712/HP* 174 9:00a CRP 2:22p DTW lAB 04:22 Daily NW*7748/NW*774 9:00a CRP 2:22p DTW lAB 04:22 XSa WN 34/WN 875 1:20p CRP 9:35p DTW HOU 07:15 XSaSu CO*3674/CO* 683 1:30p CRP 7:54p DTW lAB 05:24 XSaSu CO*3674/NW 183 1:30p CRP 7:54p DTW IAN 05:24 XSaSu NW69741NW 1836 1:30p CRP 7:54p DTW 11tH 05:24 XSa CO 3704/CO 1752 5:00p CRP 10:44 DTW 11tH 04:24 XSa HP*3704/HP* 1752 5:00p CRP 10:44 DTW IAH 04:24 XSa NW*72041NW* 7752 5:00p CRP 10:44 DTW IAN 04:24 XSa CO 4198/CO 1712 8:10p CRP 1:35a+1 DTW lAB 04:25 XSa HP*41981HP* 1712 8:10p CRP 1:35a+1 DTW I IAN 04:25 XSa NW*74981NW*7712 8:10p CRP 1:35a+1 DTW I TAll 04:25 Codeahared flight (excluding commuter codeshares with affihinted major carriers). DL = Delta Airlines NW = Northwest Airlines CO = Continental Airlines HP = American West Airlines WN = Southwest Airlines See Appendix D for key to airport codes. nections and create some additional competitive routes for travelers in smaller markets. Less plausible, however, are the consumer benefits from codesharing that major carriers have claimed for single-connect or nonstop markets in which one or both of the partners already operates through-service. The concern is that such arrangements will reduce competition in these mar- kets, because it is likely that the partners have—or had—competing

142 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY ffights in the market, or that they had the potential to become rivals in the market, because most nonstop and one-stop routes involve the part- ners' hubs either as an end or transfer point. These arrangements also might aim at increasing market share by diverting traffic from competitors through the preferential listings of online itineraries in CRSs. This problem can be corrected as travel agents learn to sort through the displays to locate competing itineraries. Close scrutiny of the intent, fairness, and competitive effects of such practices, however, is warranted. As discussed in Chapter 3, it is critical that the ticket distribution system present unbiased and reliable informa- tion. DOT should evaluate these listings through its oversight of CRSs. Recommendation on CRS Listings of Codesbares The committee recommends that DOT consider revising existing CRS rules to prohibit listings of the same itinerary under more than one carrier's code when one of the major codeshare partners—or its commuter affiliate—serves the entire itinerary. Shared Frequent-Flier Programs and Other Partnerships Reciprocal frequent-flier programs have not been restricted by either DOT or DOJ, and are generally viewed as less potentially harmflul to competition than codesharing (GAO 1999). Reciprocal programs ben- efit some travelers by allowing them to accumulate more frequent—flier credits and redeem them for free travel to a larger number of possible destinations. Participating airlines benefit by making their frequent—flier programs more enticing and more difficult to match. It is unclear whether joint frequent—flier programs spur collaboration among partners in other ways or cause them to compete less aggressively, or make them reluctant to compete directly. The net benefits to consumers depend on these competitive effects and on the creation or enhancement of market power. Whether shared frequent—flier programs and limited codesharing relationships might evolve into more comprehensive and substantive re- lationships, undermining the partners' incentives to compete and in- creasing the potential of de facto mergers, deserves attention.

Effects ofAirline Alliances and Partnerships on Competition 143 DOT has authority to review the competitive effects of partnerships among major U.S. airlines. New legislation, passed in October 1998, gives it authority to impose waiting periods on certain newly created or proposed partnerships—for example 30 to 60 days for reciprocal frequent-flier programs and up to 150 days for codeshare agreements.2 However, the legislation does not offer DOT guidance on how to assess these partnerships nor on the bounds of its authority to oppose them. Because DOT's authority does not limit DOJ's authority to review alli- ances, the waiting period can be used by DOJ to evaluate joint venture proposals under the merger standards. In the committee's opinion, early notification and evaluation of domestic airline affiances and partnerships by DOJ's Antitrust Division should be formal and obligatoty, because the affiliations can be irre- versible and possibly precursors of de facto mergers. Recommendation on DOJReview ofAirline Collaboration Plans The committee recommends that all collaboration plans among major U.S. airlines be subject to traditional, economic-based merger analyses by DOJ, and that these plans—even if they do not involve exchanges of equity or transfers of assets—be subject to advance notification require- ments similar to those required under the Hart-Scott-Rodino process.3 INTERNATIONAL ALLIANCES Although the primary purpose of this study is to assess airline competi- tion in U.S. domestic markets, most large U.S. domestic carriers also 2 Public Law 105-277 Sec. 4176 (a). 3 The Hart-Scott-Rodino Antitrust Improvements Act of 1976 requires that certain pro- posed acquisitions of stock or assets be reported to the Federal Trade Commission (FTC) and to the Antitrust Division of DOJ before consummation. The parties then must wait a specified period—usually 30 days—before they complete the transaction. The pre-merger notification program, with its filings and waiting-period requirements, provides the antitrust enforcement agencies with the time and the information to con- duct the review. DOJ's and FTC's principles of merger analysis are contained in the 1992 Horizontal Merger Guidelines (revised in 1997), which is widely accepted by economists and courts as an analytical approach to assessing the competitive effects of a transaction, including joint ventures that fall short of outright mergers.

144 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY have important international operations. International traffic headed to and from major gateways such as JFK and O'Hare account for a signif- icant share of passengers on some carriers' domestic routes, affecting their overall route structure, revenues, competitive position, and possi- bly ability to survive. During the past half dozen years, many U.S. and foreign airlines have formed codesharing alliances. As already discussed, these arrangements allow carriers to reach more passengers by linking two or more hub-and- spoke systems serving different markets and geographic regions. Schemat- ically, this creates a barbell-shaped dual-hub network, depicted in Figure 4-1. Such partnerships can be appealing to airlines because of the con- straints on airline entry and expansion imposed by national governments and embodied in aviation treaties as well as national citizen ownership laws. Aviation treaties generally are negotiated country-by-country in a bilateral framework Most countries still limit entry by foreign carriers in international markets and prohibit entry in domestic markets. Gateway hub 0 Behind-gateway city Gateway-to-gateway (code sharing, could be flown by either code-sharing partner) Behind-gateway feeder routes (always flown by hubbing airline) SOURCE GRA 1994. Figure 4-1 Schematic of an alliance.

Efficts ofAirline Alliances and Partnerships on Competition 145 DOT generally has favored the formation of international alliances, granting several of them immunity from antitrust laws, in the belief that the public will benefit from the network efficiencies as well as from the new competition, and will suffer relatively little from the allied airlines cooperatively setting fares and capacity. In its early approvals, DOT rea- soned that immunity would enhance competition in international markets by allowing airlines with small market shares to combine their networks and become more effective in competing against larger airlines. Some of these larger international carriers operated from countries with signifi- cant restrictions on market entry and competition; competition from the immunized allied airlines was a way to prompt these countries to open their markets. In approving later alliances, DOT's emphasis has changed, focusing instead on the benefits of creating alliances that could com- pete against one another, rather than against individual airlines. This new emphasis has compelled the approval of antitrust immunity for more alliances. A concern of the committee is that in advocating international alli- ances and also granting antitrust immunity, DOT might not have given sufficient consideration to the potential effect of international alliances on the competitive structure of the domestic airline industry. Requests for Antitrust Immunity As noted earlier, travelers tend to prefer online, single-carrier service, believing that connections are likely to be smoother, baggage service more dependable, and check-in faster and more convenient. For travel- ers requiring several transfers, the appeal of online service—or service that works similarly—is even stronger. These travelers, moreover, sel- dom care about which hub they must travel through to transfer to or from their origin and destination points. Recognizing this, Northwest and KLM Royal Dutch Airlines formed the first significant international codeshare agreement in 1993,4 connect- ing Northwest's main hubs in Minneapolis and Detroit with KLM's main Some codeshare itineraries between U.S. and foreign airlines had existed before 1993, but only on a small number of route extensions.

146 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY hub in Amsterdam. In this way, passengers traveling to and from many interior points in the U.S. and Europe could fly through these gateways, using one carrier for an internal segment and the Atlantic crossing, and then transferring to the other for the remaining itineraly. The traffic den- sities created by linking hub-and-spoke systems had the added benefit of supporting new nonstop, international service for travelers in what once had been much smaller gateway cities, such as Minneapolis and Detroit. Although certain features of these agreements—such as joint ticket- ing, single check-in, and coordinated baggage handling—could be achieved through traditional interline agreements, Northwest and KLM maintained that enhanced service in low- to moderate- volume markets would require a highly integrated partnership (GRA 1994, 9). In earlier, more limited agreements, fare and capacity levels were set independently by each carrier and the revenue from passengers flying on codeshare itin- eraries was divided on a predetermined, prorated basis. Northwest and KLM claimed that intensive coordination of operations, pricing, and marketing would be needed, induding yield management, scheduling, utilization, seat inventories, gates, ground support, marketing, reserva- tions, and advertising. As a practical matter, the two airlines proposed a merger of their transatlantic operations. To facilitate this integration, the two carriers asked DOT for an exemption—or immunity—from antitrust prohibitions.5 DOT granted the exemption, citing potential public benefits that, in its judgment, out- weighed potential costs. DOT's assessment was that the alliance would falter without antitrust immunity, but there also was an additional con- sideration—the infusion of KLM capital would increase the likelihood of Northwest's survival. The Northwest—KLM alliance's immunity— initially granted for six years6—was seen as a way to strengthen the corn- 5 As a result of the Airline Deregulation Act of 1978, CAB's authority to grant immunity in the domestic industry was substantially curtailed, and DOJ was given an expanded role in enforcing domestic antitrust proscriptions and, later, responsibility to review airline merger plans. The Deregulation Act, however, did not affect CAB's longstanding author- ity, granted in the 1938 Civil Aeronautics Act, to approve agreements involving interna- tional air transportation or to confer antitrust immunity on an agreement, if necessary for diplomacy and the public interest. In 1985, DOT inherited these authorities from CAB. 6 The immunity order called for the parties to resubmit an application in May 1999. Pro- cedures have not been defined, however, for reviewing resubmissions.

Efficts ofAirline Alliances and Partnershps on Competition 147 petitive position of two relatively minor transatlantic airlines and to introduce international service to several U.S. cities. DOT was persuaded that the alliance and its activities could not be open to antitrust pro- scriptions or to legal challenges from other airlines without being inher- ently unstable. The antitrust immunity would encourage the two airlines to invest in service improvements that might take several years before returning a profit. Since this first grant, other alliances with international airlines have applied for antitrust immunity, including large transatlantic carriers like United with Lufthansa; Delta with SwissAir, Sabena, and Austrian; and American with British Airways. DOT has approved the immunity requests involving United and Delta. In some important respects, these alliances—and the rationale for approving them—differed from the Northwest—KLM case. The United—Lufthansa alliance certainly would affect competition on some large, international, gateway-to-gateway routes—such as Frankfttrt—Washington Dulles—where United and Lufthansa had been transatlantic rivals. However, in 1996, when these alliances were approved, DOT's emphasis was on creating competing alliances to offer fare and service alternatives for connecting passengers. Under its original grant of immunity, the Northwest—KLM alliance had diverted traffic from larger international carriers. The effect made it more difficult for DOT to oppose subsequent alliances, and each successive grant of immunity has increased expectations of more approvals. As party to the public review of carrier requests for antitrust immu- nity, DOJ has emphasized the potential adverse effects on travelers in mainline, gateway markets where the partner airlines had been rivals. DOJ recommended that DOT should condition antitrust approval on "carving out" overlapping routes from the immunity agreements. Allied carriers still could share codes and coordinate other activities on the carve-outs, but would not have immunity for highly coordinated pric- ing, inventory, and yield management. DOT has required carve-outs in nonstop markets, such as Chicago—Frankftirt, before conferring immu- nity on some alliances. In general, the carve-outs apply only to nonstop passengers who purchase full-fare tickets in the United States; the joint development of corporate, group, promotional, and discounted fares still is sanctioned and accepted.

148 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY In the case of American Airlines' proposed alliance with British Air- ways, however, DOJ expressed concern that competition in too many U.S.—U.K. city-pairs would be affected. DOJ therefore suggested other countermeasures, such as the divestiture of slots at London Heathrow Airport; nevertheless, it conduded that the alliance should not be granted immunity on competitive grounds.7 The American Airlines—British Airways immunity request is still pending before DOT. Open Skies Initiatives In negotiating open skies agreements with foreign countries, the U.S. has sought, among other things, the lifting of limits on The number of airlines that can operate between the U.S. and the other country; The frequency and capacity of their service; and The fares they can charge.8 All of these goals represent significant alterations of longstanding bilat- eral agreements. The State Department formally negotiates the bilateral agreements, but DOT has final approval. During the 1990s, more than 30 open skies agreements were signed, with such nations as Germany, Canada, The Netherlands, Belgium, Norway, Italy, Austria, Sweden, and Switzerland. Other negotiations continue but have not achieved all of the market- oriented goals; for example, France, Australia, Japan, and the United Kingdom have not agreed to open skies treaties. DOT's dual roles in approving bilateral treaties and reviewing requests for antitrust immunity have raised concerns that immunity might be Comments of the Department of Justice Before the Department of Transportation, Joint Application of American Airlines, Inc. and British Airways, PLC for approval of antitrust immunity for alliance agreement, Docket OST-97-2058 (May 21, 1998). important element of open skies initiatives is the right of an airline from one country to carry traffic between two countries outside of its own country of registry, as long as the flight originates or terminates in its own country. Another is the right of an airline to carry traffic between two countries via its own country.

Effects ofAirlineAiiances and Partnerships on Competition 149 used as an enticement or concession in negotiating agreements, without regard to direct competitive effects. Certainly, DOT must recognize the important role that antitrust immunity has played in inducing foreign governments to renegotiate bilateral treaties under the open skies regime. Since the 1993 approval of immunity for Northwest—KLM, three dozen new treaties have been negotiated. Concerns about the increasingly explicit link between antitrust immu- nity and open skies negotiations were most evident following the Amer- ican Airlines—British Airways application. The application was filed with DOT in January 1997; 10 weeks later the department began processing it and accepting comments from interested parties, even though the open skies negotiations with the United Kingdom had not yet been concluded.9 Several parties commenting on the American Airlines— British Airways application noted that by openly associating antitrust immunity with open skies goals, DOT was creating the expectation among foreign countries and their airlines that completing an open skies treaty was not only necessary but sufficient for antitrust immunity. Benefits and Concerns from Immunized Alliances In its International Air Transportation Policy Statement,10 DOT has professed the strong belief that expanding cooperative arrangements among international carriers is desirable and beneficial to travelers. It has identified two main benefits, as well as several other positive side effects. First, end-to-end alliances develop connecting markets that have histor- ically suffered from poor service, stimulating travel in these low—volume markets. Second, the alliances can form competing networks, providing travelers from spoke—nongateway—cities with more competing options for international service. A significant side benefit is that, by funneling traffic from both ends, alliances allow cities such as Memphis and Cincin- nati to obtain nonstop international service. Both DOT and DOJ recognize, however, some possible disadvantages to alliances. First, they can diminish competition in some traditional or See AuBuchon 1999 for a more complete chronology. '°DOT. 1995. Statement of U.S. International Air Transportation Policy. Federal Reg- ister, Vol. 60, p. 21841, May 3.

150 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY mainline gateway routes, if the allied carriers previously had been rivals or might have competed directly after liberal open skies agreements were implemented. If accompanied by barriers to competitive entiy—such as slots—this possible negative effect could be more severe. Even when nonstop, overlapping routes are carved out of the immunity grant, the risk to competition remains. As DOJ has noted, codesharing airlines might compete less aggressively in price or capacity in overlapping mar- kets, to avoid undermining the agreement on connecting traffic.11 Early effects of alliances on international passenger traffic, fares, and service can be assessed empirically. Analyses by DOT and others [e.g., Brueckner and 'Whalen (1998)] suggest that alliances have produced net gains in traffic and service and reductions in fares for millions of over- seas travelers. Other DOT analyses provided to the committee show that nonstop, gateway-to-gateway traffic accounts for about one-third of all transatlantic trips, and that transatlantic traffic from connecting markets accounts for the remaining two-thirds. Transatlantic traffic from con- necting airports grew by more than 2 million passengers between 1992 and 1997—the period when the immunized alliances were established. Because of this evidence, DOT has adopted a generally favorable posi- tion on international alliances. However, some of this observed growth in traffic also might stem from the more market-oriented open skies agreements. Since the two developments—open skies and immunized alliances—coincided, it is unclear whether the liberalized bilateral agreements, given more time to take hold, would have spurred competition without the grants of anti- trust immunity. More generally, given the haste with which the alliances were formed, it is reasonable to question whether the intent of immu- nity requests was protectionist, producing open skies nominally, but with the underlying aim of protecting foreign national carriers from free competition. A longer-term issue is whether the number of international alliances— each of which is developing into a large, multicarrier alliance—will ulti- 11 Comments of the Department of Justice on the Order to Show Cause Before the Department of Transportation, American Airlines and the TACA Group Reciprocal Antitrust Services Proceeding, Docket OST-96-1700, January 28, 1998, p. 6.

Effects ofAirline Alliances and Partnershps on Competition 151 mately dictate the number of U.S. carriers that can maintain international operations, possibly reducing the number of domestic airlines capable of survival. DOT has acknowledged the potential for unfavorable domes- tic effects if a small number of large, international alliances cannot accom- modate all of the current U.S. carriers offering international service. DOT's 1995 International Aviation Policy Statement noted that "global systems and the growing use of codesharing will put significant pressure on carriers whose strategy does not include participation in such systems or in codeshare alliances or whose options to participate are limited. We expect these responses to lead to restructuring of service and airlines, similar to the U.S. domestic experience in 1980s."12 The absence of unaffiliated carriers abroad, or even of carriers willing to interline, raises the possibility that independent U.S. airlines might be weakened and possibly would not survive. In general, the committee favors the development of a broad-based, multinational framework governing international aviation. Even better would be an open and unrestricted international market in which any carrier could serve any market without restrictions on services and fares. If this remains the long-term goal of the deregulation policy initiated 20 years ago, then the emergence of global alliances linking some of the world's largest airlines could be either helpful or obstructive. On the pos- itive side, highly integrated global alliances might prompt countries to harmonize their bilateral treaties to facilitate adoption of a multilateral framework. On the negative side, once accustomed to alliance arrange- ments, airlines and transportation policy makers alike might be reluctant to advocate additional necessary market reforms. While the long-term net benefits from international alliances remain unclear, the committee is concerned that DOT has endorsed such arrangements openly while reviewing requests for antitrust immunity and also approving aviation treaties. A neutral position on international alliances would be preferable while negotiating for open skies reforms, giving the market-oriented reforms an opportunity to take hold. A concern of the committee is that DOT has taken a position of promoting these alliances while also being responsible for objectively 12 DOT. 1995. Statement of U.S. International Air Transportation Policy. Federa/Reg- ister, Vol. 60, p. 21841, May 3.

152 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY assessing the competitive effects of antitrust immunity requests, a potentially conflicting set of consequences. Although it is not evident that DOT overtly has exchanged approval of antitrust immunity for open skies, it clearly and repeatedly has linked the two issues, giving weight to open skies as a precondition for antitrust immunity. It is rea- sonable to infer that by making such a direct association, DOT has found it easier to convince foreign countries to agree to bilateral reforms. Many foreign countries are determined to protect their national airlines; DOT has eased their concern over more liberal bilateral agreements— which might explain the recent surge in open skies treaties. It is likely that successive approvals of immunity after the signing of open skies treaties has raised the expectations of foreign carriers and nations that immunity will be conferred after meeting that prerequisite. As immu- nity grants begin to expire in the next few years, DOT should anticipate significant diplomatic pressure to renew them. However, DOT has not established a formal process for reviewing alliances nearing the end of their approved terms—typically 5 years. In the committee's opinion, applications for renewal of immunity require careful consideration of the competitive effects on both domestic and international air transportation. Recommendation on Antitrust Immunity The committee recommends a two-part process for reviewing applica- tions for antitrust immunity by international airline alliances. DOJ should perform the initial review and then forward to DOT only those applications acceptable on competitive considerations. DOT then should review these applications for approval with respect to other issues of public interest and international policy. In addition, DOJ should per- form follow-up critiques ofiinmunized alliances approaching renewal. SUMMARY Contrary to what some had predicted 10 years ago, the domestic airline industry has not consolidated to the point that competition is threat- ened. Nevertheless, the emergence of airline alliances and other part- nerships, both domestically and internationally, bears close watching.

Effects ofAirlineAlliances and Partnerships on Competition 153 Outright mergers—widespread 10 to 15 years ago—have become the exception. Other collaborations among airlines have become more com- mon, such as sharing codes and frequent—ffier programs. Though not as straightforward to identify and assess as mergers, these partnerships can be steps toward horizontal consolidation and should not be overlooked. To assess the potentially adverse competitive effects from these domes- tic alliances and partnerships, the committee recommends the following: DOT should revise CRS rules to prohibit listings of the same itin- erary under more than one carrier's code when one of the partners already serves the entire itinerary. All collaboration plans among major U.S. airlines should be subject to traditional, economic-based merger analyses by DOJ. Even if the col- laborations do not involve exchanges of equity or transfers of assets, they should be subject to advance notification requirements similar to those required under the Hart-Scott-Rodino process. International airline alliances and partnerships also present issues involving the competitive structure of the domestic airline industry. One issue that deserves explicit attention is whether a small number of global airline alliances will result in a similarly small number of U.S. carriers capable of offering international service and therefore only a small num- ber capable of surviving domestically as well. Airlines with reduced oppor- tunity for international traffic and revenues might become weaker domes- tic competitors, less likely to emerge or survive as challengers to the major U.S. airlines. An aim of open skies aviation agreements was to increase competition in international markets. Public policy favoring international alliances, strengthened by exemption from antitrust laws, assumes that the world's largest airlines will be the main competition. In the committee's opinion, this assumption should be examined skepti- cally. In addition, the process for reviewing antitrust immunity should be modified to emphasize both the short- and long-term competitive effects of these agreements and their impacts on domestic, as well as inter- national, travelers. The following procedural change is recommended: A two-part process should be established for reviewing and approv- ing applications for antitrust immunity by international airline alli- ances. DOJ should perform the initial review and forward to DOT only

154 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY those applications acceptable on competitive considerations. DOT then should review these applications with respect to other issues of public interest and international policy. In addition, DOJ should per- form follow-on critiques of immunized alliances approaching renewal. Competitive and economic circumstances are subject to change, pos- sibly requiring new conditions for an immunized alliance or even for the alliance to be discontinued. A ftill review by DOJ is desirable for each immunized alliance seeking renewal, with attention to its effects on competition generally. REFERENCES ABBREVIATIONS FAA Federal Aviation Administration GAO General Accounting Office GRA GRA, Inc. AuBuchon, M.J. 1999. Testing the Limits of Federal Tolerance: Strategic Alliances in the Airline Industry. Transportation Law Journal. Vol. 26, No. 2, PP. 219-246. Brueckner, J.K., and W. T. Whalen. 1998. The Price Effects of International Airline Alliances. Working Paper, Department of Economics, University of Illinois at Ur- bana-Champaign. FAA. 1998. FAA Aviation Forecasts: Fiscal Years 1998 to 2009. Report FAA-APO-98-1. Washington, D.C. GAO. 1999. Aviation Competition: Effects on Consumers from Domestic Airline Alliances Vary. Report RCED-99-37. Jan. GRA. 1994. A Study ofInternational Codesharing. Prepared for Office of Aviation and International Economics, Office of the Secretary, U.S. Department ofTransportation. Washington, D.C.

V Competition and Entry in Smaller Markets This study has focused on some of the most significant and well recog- nized means of enhancing airline entry and competition, especially in larger markets. However, smaller markets are promising and deserving candidates for competitive activity. The Department of Transportation (DOT) can and should take the lead in seeking creative ways to increase competitive options for travelers in the small- to medium-size commu- nities that often serve as spoke cities for one or two carriers operating from larger hub airports. Although low traffic densities in many of these city-pair markets have made competitive jet service uneconomical, new technologies and changing market conditions are presenting new possi- bilities. As noted earlier, these markets have been the subject of complaints by new entrants about aggressive and anticompetitive responses by incumbents. Startup airlines also have alleged that incumbents are offer- ing travel agents extra commissions on ticket sales in markets challenged by smaller carriers. They also have alleged that larger carriers sometimes refuse to allow them to buy miles from frequent-ffier programs to award 155

156 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY to their customers, even when other airlines are purchasing miles. Highly selective uses of marketing advantages might require scrutiny and action by DOT; for example, possibly requiring airlines that sell miles to one airline to make similar offers to others. Small- to medium-size cities—particularly those that are spokes for single-hub carriers—are potential battlegrounds for competition in the airline industry. Cities such as Knoxville, Tennessee, and Grand Rapids, Michigan, have nonstop jet service only to two or three hubs in their regions. Although the hubs are important destination points for travel- ers from these medium-size communities, traffic to other large, but more distant, nonhub cities—such as NewYork and Washington, D.C.—also might be considerable and capable of supporting nonstop service on smaller jet aircraft. During the past 5 years, smaller, regional jet aircraft that can seat 35 to as many as 100 passengers have been introduced by commuter affiliates of major carriers, primarily to provide feed traffic to hubs. With their added speed and comfort, these aircraft are more popular, particularly with business travelers, than turboprop aircraft. They also have significantly greater range and potentially lower seat- mile costs on medium-length, low-density routes (i.e., flights of 400 mi or more). Some industry observers believe these aircraft will create many more competitive choices for travelers in small- and medium-size markets.1 The committee did not study the economics of regional jets; however, it is reasonable to assume that the new jets could increase the opportu- nities for some small- to medium-size cities to receive more nonstop ser- vice between more large destination cities than has been possible with mainline jet and turboprop aircraft. Regional jets are an example of how new technologies continue to affect the economics of the airline indus- try; they also might have implications for competition generally since regional jets could prompt incumbent airlines to compete directly with one another in more markets—for example, in serving the spoke cities 1 Most recently, these opportunities have been discussed in Poole and Butler 1999. DOT also has prepared a white paper available on its website, profiling regional jets and their emerging roles in U.S. markets (June 1998).

Competition andEnhy in Smaller Markets 157 of their competitors.2 They also might provide possibilities for new entrants to develop niche markets and services. In all these cases, they would enhance service and competitive choices in moderate-size markets. But spurring competition and service improvements in medium-size cities might require policymaking as well. For example, it will be impor- tant that limited airport and air traffic capacity do not impede the intro- duction and use of the new aircraft and technologies. Lifting slot controls at some key airports, promoting more gate availability, and providing more service opportunities at secondary and reliever airports—as rec- ommended in Chapter 3—should increase competition, even in smaller markets. Some major airlines and their pilots have contractual agree- ments, known as "scope clauses," that can limit the introduction and use of regional jet aircraft by the major airlines. The committee did not examine the potential institutional and contractual impediments to the use of these jets, but if the economics are right, pressures to relax these constraints will emerge. In addition, DOT should be sure that its own policies and practices are not among the unintended impediments. For example, its economic fitness determinations for a new airline's certification require an array of information describing the carrier's business plan, its equipment, fares, and intended markets. The committee was not able to discern the need for this specific information, but recognized that the required public fil- ings could help an incumbent. If some of these filing requirements are no longer necessary to ascertain fitness, they should be lifted or relaxed as vestiges of the regulated era. Competitively sensitive information should be treated as confidential. Additional, innovative government actions also might induce more rivalry in smaller markets and more nonstop service to more varied des- tinations. DOT should work with small- to medium-size communities 2 As an example, Northwest might add service between Minneapolis and Oklahoma City, now served mainly by TWA and American through their hubs in St. Louis, Chicago, and Dallas; or Delta might add service between Atlanta and Manchester, New Hampshire, now served mainly by USAirways through BWI, Philadelphia, and Pitts- burgh. Retaliation by incumbent carriers to hub "poaching," however, could dissuade major carriers from using regional jets in this competitive manner—a possibility that underlines the importance of new entrants spurring competition.

158 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY to consider positive actions for increasing competitive services. One con- cept worth exploring is that communities could offer airlines exclusive, but time-limited, rights to provide nonstop service in city-pair markets that have none. Brief "patents" of this kind might allay concerns that hub carriers will challenge them aggressively to protect their own hubs. Local residents also might be assured that the service will be sustained, although this might warrant local subsidies or other financial induce- ments, at least initially. A few communities, often with the support of local businesses, are now using subsidies to attract and retain desired air- line service (see Chapter 1). The committee is encouraged by DOT's determination to curb unfair competition in the airline industry. However, DOT also must remain alert to positive actions to encourage competition broadly, in large and small markets, long haul and short haul, and with new technologies and old. REFERENCE Poole, R.W.,Jr., and V. Butler, 1999. Airline Deregulation: The Unfinished Revolution. Regulation, Vol. 22, No. 1, pp. 44-51.

Appeildix A Statement of Enforcement Policy Regarding Unfair Exclusionary Conduct Department of Transportation Office of the Secretary April 1998 (Docket No. OST-98-3 713, Notice 98-16) Congress has put a premium on competition in the air transportation in- dustry in the policy goals enumerated in 49 U.S.C. §40101. The Department of Transportation thus has a mandate to foster and encour- age legitimate competition. We believe that legitimate competition encompasses a wide range of potential responses by major carriers to new entry into their hub markets1—responses involving price reductions or capacity increases, or both, or even neither. Some of the responses we 1 We use the term "new entrant" to mean an independent airline that has started jet ser- vice within the last ten years and pursues a competitive strategy of charging low fares. We use the term "major carrier" to mean the major carrier that operates the hub at issue. 159

160 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY have observed, however, appear to be straying beyond the confines of legitimate competition into the region of unfair competition, behavior which, by virtue of 49 U.S.C. §41712, we have not only a mandate but an obligation to prohibit. Following Congress's deregulation of the air transportation indus- try in 1978, all of the major air carriers restructured their route systems into hub-and-spoke networks. Major carriers have long charged con- siderably higher fares in most of their spoke city-pairs, or the "local hub markets," than in other city-pairs of comparable distance and density. In recent years, when small, new-entrant carriers have instituted new low-fare service in major carriers' local hub markets, the major carriers have increasingly responded with strategies of price reductions and ca- pacity increases designed not to maximize their own profits but rather to deprive the new entrants of vital traffic and revenues. Once a new entrant has ceased its service, the major carrier will typically retrench its capacity in the market or raise its fares to at least their pre-entry lev- els, or both. The major carrier thus accepts lower profits in the short run in order to secure higher profits in the long run. This strategy can benefit the major carrier prospectively as well, in that it dissuades other carriers from attempting low-fare entry. It can hurt consumers in the long run by depriving them of the benefits of competition. In those instances where the major carrier's strategy amounts to unfair compe- tition, we must take enforcement action in order to preserve the com- petitive process. We hereby put all air carriers on notice, therefore, that as a matter of policy, we propose to consider that a major carrier is engaging in unfair exclusionary practices in violation of 49 U.S.C. §41712 if, in response to new entry into one or more of its local hub markets, it pursues a strategy of price cuts or capacity increases, or both, that either (1) causes it to forego more revenue than all of the new entrant's capacity could have diverted from it or (2) results in substantially lower operating profits— or greater operating losses—in the short run than would a reasonable alternative strategy for competing with the new entrant. Any strategy this costly to the major carrier in the short term is economically rational only if it eventually forces the new entrant to exit the market, after which the major carrier can readily recoup the revenues it has sacrificed to achieve this end. We will therefore be focusing our enforcement efforts

Statement ofEnforcement Policy Regarding Unfair Exclusionary Conduct 161 on this strategy while continuing our scrutiny of any other strategies that may threaten competition. Our policy represents a balance between the imperative of encourag- ing legitimate competition in all of its various forms and the imperative of prohibiting unfair methods of competition that ultimately deprive consumers of the range of prices and services that legitimate competi- tion would otherwise afford them. This policy does not represent an attempt by the Department to reregulate the air transportation industry: we are neither prescribing nor proscribing any fares or capacity levels in any market. Rather, we are carrying out our statutory responsibility to ensure that if a new-entrant carrier's entry into a major carrier's hub mar- kets fails, it fails on the merits, not due to unfair methods of competition. BACKGROUND The competitive benefits of deregulation have been exhaustively docu- mented in numerous studies. Among other things, the major carriers' development of hub-and-spoke networks has brought most domestic air travelers more extensive service, more frequent service, and lower fares. Also widely documented are the competitive advantages in serving local markets that a major carrier enjoys at its hub. Flow traffic, or the pas- sengers that the major carrier is transporting from their origins to their destinations by way of its hub, typically accounts for more than half of the traffic in local hub markets. Flow traffic thus allows the major car- rier to operate higher frequencies in local markets than the local traffic alone would support. In turn, in local markets served by more than one carrier, the major carrier's higher frequency attracts a greater share of the local traffic than that carrier would otherwise carry.2 Due to its more extensive route network, the major carrier is also able to offer a frequent- flier program and commission overrides—i.e., higher commissions to travel agents for a higher volume of sales—that are more effective. These factors, too, confer competitive advantages on the major carrier in local hub markets. 2 This phenomenon, called the "S-curve" effect, reflects the value that time-sensitive travelers place on schedule frequency.

162 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY These advantages have translated into the power to charge higher local fares. A major carrier usually provides all of the service in most of its local hub markets, the exceptions being mainly city-pairs whose other end- points are hubs of other major carriers or city-pairs served by low-fare carriers. Many local hub markets that have enough traffic to support com- petitive nonstop service are nonetheless served only by the major carrier. In the absence of competition, the major carrier is able to charge fares that exceed its fares in nonhub markets of comparable distance and density by upwards of 40 percent, or at least $100 to $150 per round trip. Even in those local hub markets in which the major carrier competes with another major carrier, load factors may be relatively low, but fares are rel- atively high. We have observed, in fact, that low-fare service has provided the only effective price competition in major carriers' local hub markets. Major carriers use sophisticated yield-management techniques to price-discriminate and thereby maximize their revenues. They can mon- itor sales and fine-tune fares, change fare offerings for individual ffights as frequently as conditions may warrant, and segment each city-pair market so that those passengers needing the greatest flexibility pay the highest premiums while passengers needing progressively less flexibility pay progressively lower fares. The lowest fares, which typically carry heavy restrictions, provide revenue for seats that the carrier would other- wise fly empty. It is in the carrier's interest, of course, to sell each seat at the highest fare that it can. Generally, major carriers find it most prof- itable to focus on high-fare service, leaving much of the demand for low- fare service in many local hub markets unserved. Both these unserved consumers and travelers paying fare premiums in local hub markets stand to reap substantial benefits from new competi- tion. Southwest, a low-fare carrier certificated before deregulation, and various new-entrant carriers have shown that a nonhub carrier can com- pete successfully with a major carrier in the latter's hub markets.' By Southwest has scored the broadest and longest-lived success with this strategy, having established a strong presence in numerous local markets at a number of hubs. New- entrant carriers such as Valujet (now AirTran Airlines), Morris Air (before being acquired by Southwest), and Frontier have entered local markets at Atlanta, Salt Lake City, and Denver, respectively. Vanguard, another new-entrant carrier, has pursued a strategy of providing direct service between Kansas City and several hubs.

Statement of Enforcement Policy Regarding Unfair Exclusionaiy Conduct 163 charging lower fares, the new entrant can profitably serve that portion of a local market's demand which the major carrier has mostly not been serving; the resultant competition can bring fares down for most travel- ers. Traffic stimulation and reductions in average fares can both be dra- matic. According to a study by this Department, low-fare competition saved over 100 million travelers an estimated $6.3 billion in the year that ended September 30, 1995 .4 At Salt Lake City, for example, local mar- kets served by Morris Air and Southwest saw their traffic triple and their average fares decrease by half, while local markets served only by the dominant carrier saw their fares increase. By late 1995, the average fares in local markets served by Morris Air and Southwest were only one-third as high as fares in other local Salt Lake City markets. THE PROBLEM The major carriers view competition by new entrants as a threat to their ability to maximize revenues through price discrimination. As noted, not only will the previously unserved consumers take advantage of a new en- trant's low fares, but so, too, will at least some of the consumers that have been paying the major carrier's higher fares. Regardless of how the major carrier chooses to respond to the new entry, the more low-fare capac- ity available in the market, the less of its high-fare traffic the major car- rier will retain. The stakes are high: a major carrier's fare premiums in its local hub markets can mean revenues of tens of millions of dollars annually over its revenues in markets where fares are disciplined by competition. In some instances, a major carrier will choose to coexist with the low- fare competitor and tailor its response to the latter's entry accordingly. For example, at cities like Dallas and Houston, the major carriers tolerate Southwest's major presence in local markets by not competing aggres- sively for local passengers. Instead, they focus their efforts on carrying flow passengers to feed their networks. At the other extreme, the major carrier will choose to drive the new entrant from the market. It will adopt The Low Cost Airline Service Revolution, April 1996. A good portion of the savings occurred in local hub markets.

164 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY a strategy involving drastic price cuts and flooding the market with new low-fare capacity (and perhaps offering bonus frequent-ffier miles and higher commission overrides for travel agents as well) in order to keep the new entrant from achieving its break-even load factor and thus force its withdrawal. Before the new entrant does withdraw, the major carrier, with its higher cost structure, will cariy more low-fare passengers than the new entrant, thereby incurring substantial self-diversion of revenues— i.e., it will provide unrestricted low-fare service to passengers who would otherwise be willing to pay higher fares for service without restrictions. Consumers, for their part, enjoy unprecedented benefits in the short term. After the new entrant's withdrawal, however, the major carrier drops the added capacity and raises its fares at least to their original level. By accepting substantial self-diversion in the short run, the major car- rier prevents the new entrant from establishing itself as a competitor in a potentially large array of markets. Consumers thus lose the benefits of this competition indefinitely.5 We propose to consider this latter extreme to be unfair exclusionary conduct in violation of 49 U.S.C. §41712. We have been conducting informal investigations in response to informal allegations of predation, and we have observed behavior consistent with the behavior described above. The following hypothetical example involving a local hub market serves to illustrate the problem. Originally, the major carrier is able to charge one-third of its local passengers a fare of $350. These passengers generate revenue of $3 million per quarter, which constitutes half of the major carrier's total local revenue. After new entry, the major carrier ini- tially continues to price-discriminate, continues to sell a large number of seats at $350, and sustains little revenue diversion. Then the major car- rier changes its strategy and offers enough unrestricted seats at the new entrant's fare of $50 to absorb a large share of the low-fare traffic. It sells far more seats at low fares than the new entrant's total seat capacity. Consequently, virtually all of the passengers who once paid $350 now Economists have recognized that consumers are harmed if a dominant firm eliminates competition from firms of equal or greater efficiency by cutting its prices and increasing its capacity, even if its prices are not below its costs. See Ordover and Willig, An Eco- nomic Definition of Predation: Pricing and Product Innovation, YaleLawJournal, Vol. 91, No. 8, 1981.

Statement ofEnforcement Policy Regarding Unfair Exclusionary Conduct 165 pay just $50, and instead of $3 million, these passengers now account for revenue of less than $0.5 million per quarter. To make up the difference, the major carrier would have to carry six more passengers for each pas- senger diverted from the $350 fare to the $50 fare. The major carrier loses more revenues through self-diversion than it lost to the new entrant under its initial strategy. DEPARTMENT'S MANDATE Our mandate under 49 U.S.C. §41712 to prohibit unfair methods of competition authorizes us to stop air carriers from engaging in conduct that can be characterized as anticompetitive under antitrust principles even if it does not amount to a violation of the antitrust laws. The unfair exclusionary behavior we address here is analogous to (and may amount to) predation within the meaning of the federal antitrust laws.6 Although the Supreme court has said that predation rarely occurs and is even more rarely successful, our informal investigations suggest that the nature of the air transportation industry can at a minimum allow unfair exclusionary practices to succeed. compared to firms in other industries, a major air carrier can price-discriminate to a much greater extent, adjust prices much faster, and shift resources between markets much more readily. Through booking and other data generated by com- puter reservations systems and other sources, air carriers have access to comprehensive, real time information on their competitors' activities and can thus respond to competitive initiatives more precisely and swiftly than firms in other industries. In addition, a major carrier's ability to shift assets quickly between markets allows it to increase service frequency and capture a disproportionate share of traffic, thereby reaping the compet- itive advantage of the S-curve effect. These characteristics of the air transportation industry allow the major carrier to drive a new entrant from a local hub market. Having observed this behavior, other potential new entrants refrain from entering, leaving the major carrier free to reap greater profits indefinitely. 6 We will continue to work closely with the Department ofJustice in evaluating allega- tions of anticompetitive behavior, but we will take enforcement action under 49 U.S.C. §41712 against unfair exclusionary practices independently.

166 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY ENFORCEMENT ACTION We will determine whether major carriers have engaged in unfair exdu- sionary practices on a case-by-case basis according to the enforcement procedures set forth in Subpart B of 14 CFR Part 302. We will investi- gate conduct on our own initiative as well as in response to formal and informal complaints. 'Where appropriate, cases will be set for hearings before administrative law judges. We will apply our policy prospectively, and we expect to refine our approach based on experience. We anticipate that in the absence of strong reasons to believe that a major carrier's response to competition from a new entrant does not violate 49 U.S.C. §41712, we will institute enforcement proceedings to determine whether the carrier has engaged in unfair exclusionary practices when one or more of the following occurs: The major carrier adds capacity and sells such a large number of seats at very low fares that the ensuing self-diversion of revenue results in lower local revenue than would a reasonable alternative response, The number of local passengers that the major carrier carries at the new entrant's low fares (or at similar fares that are substantially below the major carrier's previous fares) exceeds the new entrant's total seat capacity, resulting, through self-diversion, in lower local revenue than would a reasonable alternative response, or The number of local passengers that the major carrier carries at the new entrant's low fares (or at similar fares that are substantially below the major carrier's previous fares) exceeds the number of low-fare pas- sengers carried by the new entrant, resulting, through self-diversion, in lower local revenue than would a reasonable alternative response. As the term "reasonable alternative response" suggests, we by no means intend to discourage major carriers from competing aggressively against new entrants in their hub markets. A major carrier can minimize or even avoid self-diversion of local revenues, for example, by matching the new entrant's low fares on a restricted basis (and without significantly increas- ing capacity) and relying on its own service advantages to retain high- fare traffic. We have seen that major carriers can operate profitably in the same markets as low-fare carriers. As noted, major carriers are competing

Statement ofEnforcement Policy Regarding Unfair Exclusionaiy Conduct 167 with Southwest, the most successful low-fare carrier, on a broad scale and are nevertheless reporting record or near-record earnings.7 We will consider whether a major carrier's response to new entry is consistent with its behavior in markets where it competes with other new-entrant carriers or with Southwest. Conceivably, a major carrier could both lower its fares and add capacity in response to competition from a new entrant without any inordinate sacrifice in local revenues. If the new entrant remained in the market, consumers would reap great benefits from the resulting competition, and we would not intercede. Conceivably, too, a new entrant's service might fail for legitimate competitive reasons: our enforcement policy will not guarantee new entrants success or even sur- vival. Optimally, it will give them a level playing field. The three scenarios set forth above reflect the more extreme and most obviously suspect responses to new entry that we have observed in our informal investigations. We do not intend them as an exhaustive list: we will analyze other types of conduct as well to determine whether to insti- tute enforcement proceedings.8 Besides examining service and pricing behavior, we will consider other possible indicia of unfair competition: for example, allegations that major carriers are attempting to block new entrants from local markets by hoarding airport gates, by using contrac- tual arrangements with local airport authorities to bar access to an air- port's infrastructure and services, or by using bonus frequent flyer awards or travel agent commission overrides in ways that appear to target new entrants unfairly. In an enforcement proceeding, if the administrative law judge finds that a major carrier has engaged in unfair exclusionary practices in vio- lation of 49 U.S.C. §41712, the Department will order the carrier to cease and desist from such practices. Under 49 U.S.C. §46301, violation of a Department order subjects a carrier to substantial civil penalties. We have crafted our policy not to protect competitors but to protect competition. We hope that it will provide consumers with the benefits One major carrier's internal documents that we reviewed as part of an informal inves- tigation of alleged predation show strong profits on individual flight segments where it competes with Southwest. Moreover, our statutory responsibility to prohibit unfair methods of competition is not limited to the unfair exclusionary practices addressed here. We will continue to monitor the competitive behavior of all types of air carriers.

168 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY of competition in increasing numbers of local hub markets over the long term. INITIAL REGULATORY FLEXIBILITYANALYSIS: The Reg- ulatory Flexibility Act of 1980, 5 U.S.C. §601 et seq., was enacted by congress to ensure that small entities are not unnecessarily and dis- proportionately burdened by government regulations or actions. The Act requires agencies to review proposed regulations or actions that may have a significant economic impact on a substantial number of small entities. For purposes of this policy statement, small entities include smaller U.S. airlines. It is the Department's tentative determination that the proposed enforcement policy would, as explained above, give smaller airlines a better opportunity to compete against larger airlines by guard- ing against exclusionary practices on the part of the larger airlines. To the extent that the proposed policy results in increased competition and lower fares, small entities that purchase airline tickets will benefit. Our proposed policy contains no direct reporting, record-keeping, or other compliance requirements that would affect small entities. Interested persons may address our tentative conclusions under the Regulatory Flexibility Act in their comments submitted in response to this request for comments. PAPERWORK REDUCTION ACT: This policy statement contains no collection-of-information requirements subject to the Paperwork Re- duction Act, P.L. No. 96-5 11, 44 U.S.C. Chapter 35. FEDERALISM IMPLICATIONS: This policy statement would have no substantial direct effects on the States, on the relationship between the national government and the States, or on the distribution of power and responsibilities among the various levels of government. Therefore, in accordance with Executive Order 12812, we have tentatively deter- mined that this policy does not have sufficient federalism implications to warrant preparation of a Federalism Assessment. (Authority Citation: 49 U.S.C. §41712.) Issued in Washington, D.C., on April 6, 1998. Rodney E. Slater, Secretary of Transportation

Appeildix B Recommendations from Winds of Change * Concerning Competition and Other Relevant Topics ANTITRUST POLICY The Committee recommends that the Department of Justice (DOJ) op- pose mergers or acquisitions in which carriers offer substantial parallel service in city-pair markets or share a hub airport. However, DOJ should not necessarily oppose mergers or asset acquisitions of carriers with complementary or end-to-end routes. Such mergers or asset acqui- sitions often may not harm competition. Given the importance of hav- ing at least three effective competitors in city-pair markets involving a connection at a hub, the maintenance of this level of choice for con- * TRB. 1991. Special Report 230: Winds of Change: DomesticAir Transport Since Deregu- lation. National Research Council, Washington, D.C., 399 pp. 169

170 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY sumers should be used as a test for the adequacy of competition in "over hub" traffic when merger and acquisition proposals are considered, in- cluding acquisitions of individual assets, such as gates. COMPUTER RESERVATION SYSTEMS (CRSs) The committee recommends that travel agents be allowed to use their own equipment or desk equipment leased from the host carrier to access multiple CRSs. This would require a Department of Transportation (DOT) regulation to prohibit contract or lease terms that restrict the ability of travel agents to use equipment that is connected to a CRS to switch freely among CRSs, along with continuing the prohibition against display and function bias and extending it beyond CRS owners to any software used in the interface allowing multiple access. CONSUMER INFORMATION The committee recommends improving consumer information by requir- ing agents to disclose the incentive commissions (commission overrides) that they receive from carriers. This could be accomplished by a DOT regulation that has the effect of making consumers aware of the existence of override programs and the identity of the carriers favored by the agent. CONGESTION PRICING DOT should permit and encourage airports to experiment with conges- tion pricing and invite evaluation of the effectiveness of these efforts by independent researchers. In the development of these proposals, DOT should consider how to avoid the potential exercise of monopoly power by airports or airlines and their customers and how the revenues earned by congestion pricing will be used to provide needed additional capacity.

Appeildix Informal Complaints to DOT by New Entrant Airlines About Unfair Exclusionary Practices March 1993 to May 1999 UNFAIR PRICING AND CAPACITY RESPONSES 1. Date Raised: May 1999 Complaining Party: AccessAir Camp lainedAgainst. Northwest Airlines Description: AccessAir, a new airline headquartered in Des Moines, Iowa, began service in the NewYork—LaGuardia and Los Angeles to Mo- line/Quad Cities/Peoria, Illinois, markets. Northwest offers connecting service in these markets. AccessAir alleged that Northwest was offering fares in these markets that were substantially below Northwest's costs. 171

172 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY Date Raised: March 1999 Complaining Party: AccessAir ComplainedAgainst Delta, Northwest, and TVVA Description: AccessAir was a new entrant air carrier, headquartered in Des Moines, Iowa. In February 1999, AccessAir began service to New York—LaGuardia and Los Angeles from Des Moines, Iowa, and Moline/ Quad Cities/Peoria, Illinois. AccessAir offered direct service (nonstop or single-plane) between these points, while competitors generally offered connecting service. In the Des Moines/Moline—Los Angeles market, Ac- cessAir offered an introductory roundtrip fare of $198 during the first month of operation and then planned to raise the fare to $298 after March 5, 1999. AccessAir pointed out that its lowest fare of $298 was substantially below the major airlines' normal 14- to 21-day advance pur- chase fares of $380 to $480 per roundtrip and was less than half of the major airlines' normal 7-day advance purchase fare of $680. Similarly, in the Des Moines/Moline/Peoria—New York market, AccessAir had of- fered an introductory roundtrip fare of $198 that was to expire on March 20, 1999. In all of these markets, Delta, Northwest, and TWA have con- tinued to offer fares at the level of AccessAir's introductory fares and did not match AccessAir's fare increases. AccessAir alleged that these airlines continued to charge fares that were substantially below their costs in order to drive AccessAir out of these markets and ultimately out of business. Date Raised: August 1998 Complaining Party: AirTran Airlines ComplainedAgainst Delta Air Lines Descriftion: AirTran said that Delta has been surreptitiously cutting fares on AirTran's routes to walkup passengers within a few days of departure by substantially opening lower fare, advance-purchase fare tickets to these unrestricted passengers, thereby charging substantially lower fares without restriction. AirTran said that Delta has attempted to hide this practice, which has targeted AirTran. Date Raised: February 1998 Complaining Party: Kiwi Airlines ComplainedAgainst Continental Airlines Descrzption: In February 1998, Kiwi announced that it would be inaugurating Niagara-Newark service with fares between $99 and $129.

Informal Complaints to DOTby New EntrantAirlines 173 Continental immediately lowered its lowest unrestricted fare in the Buffalo-Newark market to $79, which was matched by USAirways. Kiwi says that Continental took this step solely to defend its "fortress hub" at Newark, where Continental controls 65 percent of the gates, rather than due to economics of the market. Kiwi alleged that the Greater Buffalo area, including both Buffalo and Niagara was the relevant market for an- alyzing competitive practices. Kiwi noted that "when Kiwi temporarily ceased operations in the fall of 1996, fares from Newark on routes that Kiwi competed with Continental surged 172 percent." Date Raised: March 1997 Complaining Party: Valujet Airlines ComplainedAgainst: Northwest Airlines Description: ValuJet alleged that Northwest responded to Valujet's entry into the Atlanta-Memphis market in late 1993 with an immedi- ate, drastic cut in fares. According to Valujet, the airline began service in the fourth quarter of 1993 and by the first quarter of 1994, Northwest had reduced fares to the point that its yield from the Atlanta-Memphis route was less than half ofwhat it had been six months earlier. Valujet also claimed Northwest increased capacity on the Atlanta-Memphis by more than 50 percent beginning in late 1994. Date Raised: February 1997 Complaining Party: Valujet Airlines ComplainedAgainst Delta Air Lines Description: Valujet alleged that Delta began targeting Valujet by adding substantial capacity in markets Valujet reentered subsequent to its fatal crash in June 1996. There were 14 Atlanta markets which Valu- Jet reentered after the crash, and Delta increased capacity, despite the fact that Valujet reentered with much less capacity than before the crash. Valujet asserted that Delta's pricing in non-Valujet markets was signif- icantly different from that in Valujet markets. In markets where Valujet had not resumed service, Delta significantly raised leisure fares. Valujet's most telling example of Delta's conduct was the Atlanta-Mobile market. On January 6, 1997, Valujet announced discontinuance of service. On January 7, 1997, Delta raised its fares substantially. The city of Mobile prevailed upon Valujet to return to Mobile by offering a joint marketing

174 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY arrangement, and Delta's leisure market fares reverted to match the ValuJet leisure fare level. Date Raised: January 1997 Complaining Party: Frontier Airlines ComplainedAgainst United Airlines Description: Frontier complained that United added capacity and matched Frontier's fares in several larger Denver markets served by Frontier, including Denver—Los Angeles. Frontier also asserted that United was frustrating Frontier's marketing efforts with corporate trav- elers by offering corporate customers in Denver discount fares that required the customers to use United in markets where it competed with Frontier in order to obtain discounts in markets where Frontier did not compete. Date Raised: November 1996 Complaining Party: Spirit Airlines ComplainedAgainst Northwest Airlines Description: Spirit began operating one nonstop Detroit-Philadelphia roundtrip flight in December 1995 and added a second flight in June 1996. Initially, Northwest did not increase capacity or sell large numbers of seats at low prices in response to Spirit's entry until after Spirit added the second daily flight. Northwest thereafter added service, slashed its fares, and, by eliminating most of the restrictions applicable to its low- est discount fares, sold almost all of its seats at low fares and few seats at higher fares, thereby sharply reducing its revenues in the market. When Spirit began flying between Detroit and Boston, Northwest reacted in a similar fashion to Spirit's introduction of Detroit-Philadelphia service. Spirit complained that Northwest's increased availability of discount fares and increased capacity represented a deliberate sacrifice of short- term profits with the intent of forcing Spirit out of the markets. Date Raised: October 1996 Complaining Party: Vanguard Airlines ComplainedAgainst American Airlines Description: Vanguard complained that American's responses to Van- guard's introduction of nonstop Kansas City-Dallas/Fort Worth (DFW)

Informal Complaints to DOTby New EntrantAirlines 175 flights and entry into the DFW-Phoenix and Cincinnati markets were designed to force Vanguard to withdraw from those markets and did not constitute a legitimate competitive response to Vanguard's entry. First, American substituted jet service for its commuter airline service in the DFW-Wichita market, which had been a major Vanguard market. American also added flights in the DFW-Kansas City market, even though it already heavily outscheduled Vanguard, and matched or un- dercut Vanguard's fares in that market. American began operating non- stop flights in the DFW-Cincinnati market, a market from which American had withdrawn two years earlier. American matched Van- guard's fares in that market and the DFW-Phoenix market. Vanguard complained that American's response to Vanguard's entry into the DFW-Phoenix market was designed to force Vanguard to withdraw from this market, and thus did not constitute a legitimate competitive response to Vanguard's entry. American matched Vanguard's fares in the DFW-Phoenix market. Vanguard also complained that American's re- sponses to Vanguard's entry into the DFW-Cincinnati market were de- signed to force Vanguard to withdraw from this markets and did not con- stitute a legitimate competitive response to Vanguard's entry. American began operating nonstop flights in the DFW-Cincinnati market, a mar- ket from which American had withdrawn two years earlier. American matched Vanguard's fares in that market. Vanguard also complained that American substituted jet service for its commuter airline service in the DFW-Wichita market, which had been a major Vanguard market after Vanguard's institution of nonstop Kansas City-DFW flights and entry into the DFW-Phoenix and Cincinnati markets. These actions were designed to force Vanguard to withdraw from those markets and did not constitute a legitimate competitive response to Vanguard's entry. 10. Date Raised: March 1996 Complaining Party: Air South ComplainedAgainst Continental Airlines Description: In early March 1996, Air South complained to the De- partment that Continental had attempted to "overlay" Air South's pro- posed new service offerings in three markets: Charleston-Newark, Co- lumbia-Newark, and Myrtle Beach—Newark Air South also alleged that Continental had made its discount fare offerings more generally available

176 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY than Air South's fares, thereby effectively undercutting Air South's fares. Air South asserted that, since Continental did not have a strong economic reason for this action, other than to drive a competitor from its market region, Continental's actions were in violation of the antitrust laws. Date Raised: August 1995 Complaining Party: Vanguard Airlines ComplainedAgainst Northwest Airlines Description: Vanguard complained that Northwest had responded to Vanguard's entry into the Chicago Midway—Minneapolis/St. Paul market with fare cuts and capacity increases designed to force Vanguard to exit the markets in August 1995. This action was similar to that taken by Northwest in May 1995 when Vanguard entered the Kansas City— Minneapolis/St. Paul market. Date Raised: June 1995 Complaining Party: Vanguard Airlines Camp lainedAgainst Northwest Airlines Description: Vanguard complained that Northwest had responded to Vanguard's entry into the Kansas City—Minneapolis/St. Paul market with fare cuts and capacity increases designed to force Vanguard to exit the market. In May 1995 Vanguard entered the Kansas City—Minneapolis/ St. Paul market, where Northwest's market share had exceeded 90 per- cent. Northwest immediately increased capacity and matched Van- guard's low fares on almost all of its six daily roundtrip flights, even though Vanguard operated only one nonstop and two one-stop roundtrip flights in the market. In addition, Northwest added a roundtrip ffight to its daily Kansas City—Minneapolis/St. Paul service. Date Raised: March 1995 Complaining Party: Valujet Airlines Camp lainedAgainst- USAirways Description: Valujet alleged that USAir engaged in predatory behav- ior in several specific markets: Washington Dulles—Florida, Washington Dulles—Hartford, and Washington Dulles—Boston. According to Valujet, after it began operating flights from Dulles to Boston, Hartford, and sev- eral Florida points, USAir cut its fares in the Dulles-Florida markets to

Informal Complaints to DOTby New Entrant Airlines 177 levels that could not have been economic for US Airways, and it began operating flights from Dulles to Boston and Hartford that undercut Valujet's fares. Valujet claimed that USAir within the last 10 years had not inaugurated service from Dulles to any city that was not a USAir hub. Valujet alleged that USAir entered the Dulles routes and under- cut its fares in order to regain its dominance of the routes between the Baltimore-Washington area and Boston, Hartford, and Florida. Date Raised: December 1993 Complaining Party: ValuJet Airlines ComplainedAgainst Delta Air Lines Description: Valujet alleged that Delta had instituted a campaign to prevent the new entrant from "getting off the ground." According to Valujet, after it began operating at Atlanta, Delta sharply reduced its fares in the markets served by ValuJet and expanded its capacity in a number of those markets, including Atlanta-Jacksonville and Atlanta-Memphis. Date Raised: August 1993 Complaining Party: MarkAir ComplainedAgainst Alaska Airlines Description: MarkAir complained that Alaska was charging below- cost fares in several markets to force MarkAir to exit those markets and weaken MarkAir's ability to compete against Alaska in other markets. Date Raised: March 1993 Complaining Party: Reno Air ComplainedAgainst Northwest Airlines Description: Reno Air inaugurated service in the summer of 1992 between Reno, Nevada, and West Coast cities. In early 1993, the carrier announced plans to initiate service in the Reno—Minneapolis/St. Paul market. Northwest, which then operated no Reno—Minneapolis/St. Paul flights, soon announced that it would begin operating nonstop flights between Reno and Minneapolis/St. Paul and would enter three of Reno Air's nonstop markets: Reno—Seattle, Reno—Los Angeles, and Reno—San Diego. Northwest matched Reno Air's fares in all markets in- cluding Reno Air's Minneapolis/St. Paul—West Coast connecting ser- vices. Reno Air complained to both DOT and DOJ.

178 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY GATE ACCESS AND AIRPORT IMPEDIMENTS Date Raised: April 1999 Complaining Party: AirTran Airlines Complained Against: Newark International Airport and several airlines Description: AirTran wanted to start service from Atlanta to Newark International Airport but was unable to obtain any gate facilities for its flights, even though several gates were underused. Although airport of- ficials tried to help AirTran obtain gates, the airlines that had gates avail- able refused to allow AirTran to use them, and the airport thus far had failed to require any airline to provide gates to AirTran. Date Raised: November 1998 Complaining Party: Legend Airlines ComplainedAgainst Dallas—Fort Worth International Airport Board and the City of Fort Worth Description: Legend, a start-up airline, planned to operate long-haul service from Love Field as permitted by Congressional legislation enacted in 1997. Legend intended to use large aircraft reconfigured to satisfy the maximum passenger capacity restrictions established by that legislation. In its complaint (FAA 16-98-20), Legend alleged that the DFW Board and Fort Worth were attempting to bar Legend from operating at Love Field by filing a state court suit against Dallas, Love Field's owner, that sought to bar Dallas from allowing Legend to operate its proposed flights (the DFW Board and Fort Worth claimed that a 1968 contract between Dallas and Fort Worth required Dallas under state law to limit Love Field service). Legend additionally complained that the DFW Board and Fort Worth were discriminating against Legend by not ob- jecting to services operated by other airlines at Love Field and Alliance Airport, a Fort Worth cargo airport, that also violated the 1968 contract between the cities. FAA dismissed the complaint because it was not within FAA's jurisdiction, since the commitments of Fort Worth and the DFW Board to the FAA covered only their own airports and did not cover Love Field, an airport owned by Dallas. The Department, how- ever, issued orders interpreting federal law which held that Legend's pro- posed services could not be restricted by Dallas, notwithstanding Fort Worth's contract daims (Orders 98-12-27 and 99-4-13).

Informal Complaints to DOTby New EntrantAirlines 179 Date Raised: October 1998 Complaining Party: Colgan Airways ComplainedAgainst TWA Description: Colgan stated that it was being treated unfairly by TWA at LaGuardia Airport. Colgan had been leasing 12 LaGuardia slots from TWA for an extended period. According to Colgan, TWA recalled six of the leased slots with virtually no notice, so that they could be leased out to another airline. Colgan was not given the opportunity to bid on them or to match the other carrier's offer. Date Raised: March 1998 Complaining Party: AirTran Airlines ComplainedAgainst Greater Rochester International Airport Description: AirTran said that it operated only one flight per day at Rochester and that nonsignatory airlines were unfairly subjected to in- creased rental rates, increased landing fees and a $4 facility charge for each passenger travelling in and out of the facility. AirTran contended that this nonsignatory charge was not imposed at other facilities. Date Raised: March 1998 Complaining Party: AirTran Airlines ComplainedAgainst New Orleans International Airport Description: AirTran contended that the major carriers at New Or- leans attempted to amend the facilities lease agreement to the detriment of AirTran, the only new entrant serving New Orleans. According to AirTran the new agreement would have required AirTran and any other new entrant airline to lease all available square footage within the bag make-up area rather than an area proportional to the actual operational needs. AirTran was the only carrier objecting to the change. AirTran said that the revised lease agreement would have increased its costs by $75,000. Date Raised: October 1997 Complaining Party: Kiwi Airlines ComplainedAgainst Delta Air Lines and Hartsfield—Atlanta Inter- national Description: In August 1997, Kiwi was notified by Hartsfield- Atlanta that it would be required to move its gates from D-South concourse

180 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY to D-North in order to provide additional space for Delta. Although Kiwi proposed an alternate, lower cost relocation to Hartsfield—Atlanta, it was rejected by the airport. Previously, Kiwi had been required by the airport to move its facilities from Concourse C to Concourse D-South, in order to provide additional space for ASA, Delta's regional affiliate. In addition, Kiwi had earlier surrendered a significant portion of its of- fice space on Concourse D to Delta. The airport then informed Kiwi that it had to relinquish the remainder of its office space to provide an additional break room for Delta's employees. In requiring Kiwi to move, the airport failed to give the carrier adequate lead time to plan the opera- tion, unlike the airport's treatment of other air carriers. Although Delta supposedly needed additional space, with the concurrence of the airport, Delta supposedly held two gates from which it operated no service and another gate from which it operated less service than required. Date Raised: October 1997 Complaining Party: Kiwi Airlines Camp lainedAgainst Continental Airlines and Newark International Airport Descrlption: Until October 1997, Kiwi had subleased space from TWA at Terminal A. In October 1997 Continental purchased the TWA gates (as well as three gates at Terminal A from another carrier). Conti- nental, which already held the majority of gates at Newark, had no oper- ations at Terminal A and purchased the gates for future expansion. After Continental acquired the TWA gates, Continental substantially raised Kiwi's charges for gate use and ground handling services required by Con- tinental's sublease. Kiwi alleged that Continental's acquisition of TWA's gates was an attempt to monopolize Newark with the consent of the Port Authority of New York and NewJersey. Kiwi also contended that the "tie- in" arrangement requiring Kiwi to purchase ground handling services from Continental was a violation of Section 1 of the Sherman Act. Date Raised: May 1997 Complaining Party: Reno Air Camp lainedAgainst Northwest Airlines and Detroit Wayne County Airport Description: Reno Air contended that Northwest and Detroit Wayne County Airport (DTW) colluded to deny Reno Air access to the

Informal Complaints to DOTby New Entrant Airlines 181 DTW domestic terminal even though gate space was available. Reno Air ftirther alleged that DTW, in concert with Northwest, offered gate po- sitions at the more expensive international terminal. When Northwest later needed international gates, it offered Reno Air domestic terminal gate positions in exchange for Reno Air relinquishing its international gates. Reno Air also said that DTW charged Reno Air 150 percent of normal landing fees because Reno Air was not a signatory carrier, al- though Reno Air had attempted to become a signatory carrier. Reno Air also alleged that Northwest entered the Detroit-Reno market only after Reno Air had indicated that it would serve the market. After Reno Air initiated service, Northwest substantially undercut Reno Air's fares, offered bonus WorldPerks miles for the market and then increased ser- vice to become the dominant carrier in the market. Then when Reno Air reduced service in the market, Northwest also reduced service. Date Raised: December 1995 Complaining Party: Valujet Airlines ComplainedAgainst Delta Airlines Description: Valujet complained that Delta preempted Valujet's efforts to obtain the LaGuardia slots needed for Valujet to begin oper- ating Atlanta-LaGuardia ffights in order to maintain Delta's monopoly in that market. When Valujet was ready to buy slots from TWA, which wanted to sell slots, Delta outbid Valujet for the slots. Delta assertedly did not need the slots for its own operations and bought them only to keep Valujet from obtaining them. Date Raised: January 1995 Complaining Party: Spirit Airlines ComplainedAgainst Northwest Airlines Description: Spirit stated that Northwest had exclusive use of 51 of the 80 gates at Detroit Metropolitan Airport and partial use of addi- tional gates. Spirit, in contrast, could not obtain adequate gate facilities and so could offer only a limited number of flights in competition with Northwest. Spirit complained that when it attempted to purchase two gates from USAir, Northwest frustrated Spirit's offer for the gates by outbidding it.

182 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY OTHER AND COMBINED COMPLAINTS Date Raised: November 1998 Complaining Party: Pro Air ComplainedAgainst Northwest Airlines Description: Pro Air contended that (1) Northwest had aggressively matched Pro Air's fares in all of the city-pair markets that Pro Air served, including Baltimore—Washington, Newark, Indianapolis, and Philadelphia from Detroit; (2) Northwest increased low-fare capacity in the Pro Air markets and, in some cases, added flight segments; (3) Northwest matched Pro Air's lower changed-reservation fees only in markets Pro Air served; in Pro Air markets, Northwest charged $25, whereas in other markets, Northwest charged $75; (4) Northwest refused to work with Pro Air in de- veloping interline agreements even when such agreements would benefit Northwest passengers, such as during a then-recent Northwest strike; and (5) Northwest might be paying significant commission overrides to travel agents who booked with Northwest on routes in which Pro Air competed. Date Raised: March 1995 Complaining Party: Nations Air ComplainedAgainst. USAirways Description: Nations Air complained that USAir engaged in preda- toiy conduct and "dirty-tricks" to cause Nations Air to exit the Pittsburgh- Philadelphia-Boston markets where USAir had had a monopoly. After Nations Air began operations, USAir allegedly undercut Nations Air's fares. The incumbent airline also used its travel agency override commis- sion programs and related programs to discourage travel agencies from booking customers on Nations Air. In addition, according to Nations Air, USAir discouraged some other firms from doing business with Nations Air, and its pilots at Philadelphia maneuvered their aircraft to block Nations Air ffights, delaying the ffights and inconveniencing Na- tions Air's passengers. Date Raised: February 1995 Complaining Party: Frontier Airlines ComplainedAgainst United Airlines Description: Frontier complained that United used various unfair tactics to undermine Frontier's ability to compete with United in several

Informal Complaints to DOTby New Entrant Airlines 183 Denver markets, particularly on Frontier's Denver—North Dakota routes and including Denver to Billings, Bismarck, Bozeman, Fargo, Minot, and Missoula. Frontier alleged that United unreasonably refused to sign a full ticketing and baggage agreement with Frontier and refused to codeshare with Frontier. Frontier additionally charged that United caused Apollo, the computer reservations system (CRS) controlled by United, to use display criteria that gave Frontier's competitive services a poor display position so that travel agents would be less likely to book Frontier's ffights. Date Raised: September 1993 Complaining Party: Nations Air ComplainedAgainst Allegheny County Description: Nations Air, then an applicant for certificated author- ity, planned to operate from Allegheny County Airport rather than Pittsburgh International Airport. It complained that the County had violated the antitrust laws by closing the airport and forcing Nations Air to use Pittsburgh International Airport. Date Raised: May 1993 Complaining Party: UltrAir ComplainedAgainst Continental Airlines Description: UltrAir, a start-up airline that flew from Houston to Los Angeles and New York, complained that Continental, which used Hous- ton as a hub, was trying to force UltrAir out of business. Continental allegedly intimidated travel agencies so they would not book UltrAir, undercut UkrAir's fares, refused to interline, and caused SystemOne, the CRS offered by Continental to travel agencies, to display inaccurate and incomplete information on UltrAir's schedules and fares. Date Raised: April 1993 Complaining Party: Kiwi Airlines ComplainedAgainst Continental Airlines Description: Kiwi, a start-up airline based at Newark, a Continental hub, complained that Continental took steps to undermine Kiwi's ability to operate successfully. Continental matched or undercut Kiwi's fares, added flights, and probably caused SystemOne, the CRS offered by

184 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY Continental to travel agencies, to display inaccurate and incomplete in- formation on Kiwi's services. COMPLAINTS NOT CONSIDERED BY THE COMMfl1I'EE Date Raised: January 1999 Complaining Party: USAirways ComplainedAgainst United Description: USAirways' Chairman Stephen Wolf wrote Secretary Slater complaining about United's announced intention to increase service at its Washington Dulles hub by 60 percent following USAirways' an- nouncement to increase its own operations at Dulles. Wolf described United's action as a "visible sign of dominant carrier predatory action. . and included comments by a Wall Street analyst that concluded that United's expansion "is about beating US Airways out of Dulles, not about maximizing profits" and that "because United intends to put too much capacity too quickly into these markets, its new flying is unlikely to be profitable." Wolf wrote, "As you know, predatory actions often sac- rifice short-term profits to protect dominance." He also maintained that "The unrelenting attempts of the major trunk carriers to undermine the operations and expansion of smaller carriers, both domestically and in- ternationally, is a dear and present danger to free market competition." Date Raised: 1998 Complaining Party: A low-fare carrier (requested confidentiality) ComplainedAgainst American Airlines Description: The low-fare carrier charged that American engaged in anticompetitive conduct to discourage the low-fare carrier from begin- ning service that would compete with American's service in one of American's hub markets. After American learned that the low-fare car- rier intended to enter the market, American signaled the low-fare car- rier that it was also planning to begin flying on the route and announced that it would match all of the low-fare airline's fares. Date Raised: October 1998 Complaining Party: United Airlines ComplainedAgainst Minneapolis—St. Paul International Airport

Informal Complaints to DOTby New EntrantAirlines 185 Description: UAL complained that MSP turned down its request to lease airport Gate 43 to (1) accommodate UAL's existing operations and future expansions of its schedule, and (2) improve the gate location of Great Lakes Aviation Ltd., which, doing business as United Express, was subsidized by DOT for essential air service between Minneapolis— St. Paul and six Midwestern communities. IV.Date Raised: March 20,1996 (Docket: OST-96-1172) Complaining Party: Alaska Central Express, Inc., BidzyTa HotAana, Inc., dlb/a Tanana Air Service, and Yute Air Alaska, Inc. ComplainedAgainst Warbelow's Air Ventures, Inc., and Hageland Aviation, Inc. Description: Complainants charged respondents with engaging in joint operations in certain Alaskan markets for the purpose of gaining an unfair advantage in mail tender from the U.S. Postal Service. Disposition: By Order 96-8-8, the Department dismissed the com- plaint on the grounds that complainants had failed to provide any evi- dence that respondents had engaged in activities constituting unfair or deceptive trade practices or unfair methods of competition within the scope of 49 U.S.C. §41712. V. Date Raised: October 1997 Complaining Party: Great Lakes Aviation ComplainedAgainst Northwest Airlines and the Minneapolis Airport Commission (MAC) Description: Great Lakes has asserted that it (and other carriers) have been and continue to be routinely denied access to attractive gate and terminal accommodations at Minneapolis Airport. Great Lakes had tried for years to gain access to a gate near United Airlines, its major code-share partner, to allow for the type of easy connections that are part and parcel of codesharing relationships. During a recent round of nego- tiations with MAC for an improved gate location, according to Great Lakes, MAC said it would first have to check with Northwest before agreeing to anything.

Ilpp011diX D Airport Identification Codes U.S. commercial service airports are identified by a three-letter code as- signed by the International Air Transport Association. Codes for air- ports mentioned in this report are listed below. ABQ Albuquerque International Airport, Albuquerque, New Mexico ACY Atlantic City International Airport, Atlantic City, New Jersey AMA Amarillo International Airport, Amarillo, Texas ATL William B. Hartsfield Atlanta International Airport, Atlanta, Georgia AUS Austin Robert Mueller Municipal Airport, Austin, Texas BDL Hartford/Springfield Bradley International Airport, Windsor Locks, Connecticut BHM Birmingham Airport, Birmingham, Alabama BIL Billings Logan International Airport, Billings, Montana BNA Nashville Metropolitan Airport, Nashville, Tennessee BOl Boise Air Terminal (Gowen Field), Boise, Idaho BOS General Edward Lawrence Logan International Airport, East Boston, Massachusetts 186

Airport Identjfication Codes 187 BUF Buffalo International Airport, Buffalo, New York BUR Burbank-Glendale-Pasadena Airport, Burbank, California BWI Baltimore-Washington International Airport, BWI Airport, Maryland CAE Columbia Metropolitan Airport, West Columbia, South Carolina CHS Charleston International Airport, Charleston, South Carolina CLE Cleveland Hopkins International Airport, Cleveland, Ohio CLT Charlotte/Douglas International Airport, Charlotte, North Carolina CMH Columbus International Airport, Columbus, Ohio COS Colorado Springs Municipal Airport, Colorado Springs, Colorado CRP Corpus Christi International Airport, Corpus Christi, Texas CVG Cincinnati International Airport, Covington, Kentucky DAL Dallas Love Field, Dallas, Texas DAY Dayton (James M. Cox) International Airport, Dayton, Ohio DCA Ronald Reagan Washington National Airport, Arlington, Virginia DEN Denver Stapleton International Airport, Denver, Colorado DFW Dallas/Fort Worth International Airport, Dallas—Ft. Worth, Texas DSM Des Moines International Airport, Des Moines, Iowa DTW Detroit Metropolitan (Wayne County) Airport, Detroit, Michigan ELP El Paso International Airport, El Paso, Texas EWR Newark International Airport, Newark, New Jersey FLL Ft. Lauderdale-Hollywood International Airport, Ft. Lauderdale, Florida GEG Spokane International Airport, Spokane, Washington GRR Grand Rapids (Kent County) International Airport, Grand Rapids, Michigan GSO Piedmont Triad International Airport, Greensboro, North Carolina GSP Greenville-Spartanburg Airport, Greer, South Carolina HNL Honolulu International Airport, Honolulu, Hawaii

188 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY HOU Houston William P. Hobby Airport, Houston, Texas TAD Washington Dulles International Airport, Prince William, Virginia IAH Houston Intercontinental Airport, Houston, Texas ICT Wichita Mid-Continental Airport, Wichita, Kansas IND Indianapolis International Airport, Indianapolis, Indiana ISP Ronkonkoma (Long Island) MacArthur Airport, Islip, New York JAX Jacksonville International Airport, Jacksonville, Florida JFK John F. Kennedy International Airport, Jamaica, New York LAS Las Vegas McCarran International Airport, Las Vegas, Nevada LAX Los Angeles International Airport, Los Angeles, California LBB Lubbock International Airport, Lubbock, Texas LGA New York LaGuardia Airport, Flushing, New York LGB Long Beach Airport (Daugherty Field), Long Beach, California LIT Little Rock Regional Airport, Little Rock, Arkansas MAF Midland International Airport, Midland, Texas MCI Kansas City International Airport, Kansas City, Missouri MCO Orlando International Airport, Orlando, Florida MDW Chicago Midway Airport, Chicago, Iffinois MEM Memphis International Airport, Memphis, Tennessee MHT Manchester Airport, Manchester, New Hampshire MIA Miami International Airport, Miami, Florida MKE Milwaukee General Mitchell International Airport, Milwaukee, Wisconsin MSN Madison (Dane County) Regional Airport (Truax Field), Madison, Wisconsin MSP Minneapolis-St. Paul International Airport (Wold-Chamberlain Field), St. Paul, Minnesota MSY New Orleans International Airport, New Orleans, Louisiana OAK Oakland International Airport, Oakland, California OKC Oklahoma City Will Rogers World Airport, Oklahoma City, Oklahoma OMA Omaha Eppley Airfield, Omaha, Nebraska ONT Ontario International Airport, Ontario, California

Airport Ident/1cation Codes 189 ORD Chicago-O'I-Iare International Airport, Chicago, Illinois ORF Norfolk International Airport, Norfolk, Virginia PBI Palm Beach International Airport, West Palm Beach, Florida PDX Portland International Airport, Portland, Oregon PHL Philadelphia International Airport, Philadelphia, Pennsylvania PHX Phoenix Sky Harbor International Airport, Phoenix, Arizona PIE St. Petersburg International Airport, St. Petersburg, Florida PIT Pittsburgh International Airport, Pittsburgh, Pennsylvania PVD Providence Theodore Francis Green State Airport, Warwick, Rhode Island PWM Portland International Jetport, Portland, Maine RDU Raleigh-Durham Airport, Morrisville, North Carolina RIC Richmond International Airport (Byrd Field), Richmond, Virginia RNO Reno Cannon International Airport, Reno, Nevada ROC Rochester International Airport, Rochester, New York SAN San Diego International Airport (Lindbergh Field), San Diego, California SAT San Antonio International Airport, San Antonio, Texas SDF Louisville Standiford Field, Louisville, Kentucky SEA Seattle-Tacoma International Airport, Seattle, Washington SF0 San Francisco International Airport, San Francisco, California SJC San Jose International Airport, San Jose, California SLC Salt Lake City International Airport, Salt Lake City, Utah SMF Sacramento Metropolitan Airport, Sacramento, California SNA Santa Ana (Orange County) John Wayne Airport, Santa Ana, California SRQ Sarasota-Bradenton Airport, Sarasota, Florida STL St. Louis—Lambert International Airport, St. Louis, Missouri SYR Syracuse Hancock International Airport, Syracuse, New York TPA Tampa International Airport, Tampa, Florida TUL Tulsa International Airport, Tulsa, Oklahoma TUS Tucson International Airport, Tucson, Arizona TYS Knoxville McGhee Tyson Airport, Alcoa, Tennessee

Study Committee Biographical Information John R. Meyer, Chairman, is James W. Harpel Professor Emeritus for Capital Formation and Economic Growth of the John F. Kennedy School of Government, Harvard University. He joined the faculty of Harvard in 1953, becoming professor of economics in 1959. From 1968 to 1973, he was professor of economics at Yale University. His publica- tions in transportation economics include Competition in the Transporta- tion Industries (1959); The Urban Transportation Problem (1965); Autos, Transit, and Cities (1981); Airline Deregulation: The Early Experience (1981); Deregulation and the New Entrepreneurs (1984); and Deregulation and the Future oflntercity Passenger Travel(1987). He was a member of the National Research Council's (NRC's) Committee for the Study of Air Passenger Safety and Service Since Deregulation (1989-1991). He serves or has served on the boards of directors for several corporations. He was vice chairman of the Union Pacffic Corporation and recently re- tired as a director. He also has served as a trustee for the Mutual Life In- surance Company of New York. Dr. Meyer earned his Ph.D. in eco- nomics from Harvard University. Elizabeth E. Bailey is John C. Hower Professor of Public Policy and Management and chair of the Public Policy and Management Depart- ment, University of Pennsylvania. She previously served as dean and pro- fessor of economics, industrial administration, and public policy at Carnegie Mellon University. From 1977 to 1983, she was a commis- sioner on the Civil Aeronautics Board (CAB). Before joining CAB, she was research head in the Economic Research Department of Bell Lab- oratories. Dr. Bailey serves on the board of directors of CSX Corpora- tion, Philip Morris Companies, and Honeywell Corporation. She is vice chairman of Bancroft NeuroHealth and on the board of trustees of the Brookings Institution, the College Retirement Equity Fund, and the National Bureau of Economic Research. She earned her Ph.D. in eco- nomics from Princeton University. 190

Study Committee Biograp hical Information 191 Jonathan B. Baker is an associate professor of law at American Univer- sity's Washington College of Law. He was director of the Federal Trade Commission's (FTC's) Bureau of Economics from 1995 through 1998. As a senior economist at the President's Council of Economic Advisors from 1993 through 1995, he worked on matters involving regulation, in- dustrial organization, and law. He also has been special assistant to the deputy assistant attorney general for economics in the Antitrust Divi- sion of the Department of Justice, assistant professor at Dartmouth's Amos Tuck School of Business Administration, attorney advisor to the acting FTC chairman, and an antitrust lawyer in private practice. He has published widely in the fields of antitrust law and policy and empirical industrial organization economics. He earned his J.D. from Harvard University and Ph.D. in economics from Stanford University. Roden A. Brandt is an independent air transportation consultant. His clients have included airlines in Europe, Asia, and the Americas, as well as global financial institutions. His services have related primarily to acquisitions and restructurings involving carriers, ranging from major international airlines to new entrants. He has held senior executive posi- tions at well established airlines as well as smaller entrepreneurial airlines that began service following deregulation. From 1996 to 1997 he was senior vice president of marketing and planning at TWA and from 1995 to 1996 was president of Air South. He was founding president and chief executive of Air Atlanta from 1982 to 1984; before that, he held senior executive positions at National Airlines and Pan American, including se- nior vice president, planning. He earned his bachelors degree in aero- nautical engineering and economics from Massachusetts Institute of Technology. Darius W. Gaskins is a partner in the consulting firm Norbridge, Inc., and in the investment group High Street Associates, Inc. From 1989 to 1991 he was visiting professor in the Center for Business and Govern- ment, John F. Kennedy School of Government, Harvard University. From 1985 to 1989 he was president and CEO of Burlington Northern Rail- road, and he was senior vice president for marketing and sales from 1982

192 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY to 1985. During 1981 and 1982, he was senior vice president of Natomas North America. Among several government positions, Dr. Gaskins has served as chairman of the Interstate Commerce Commission (1980-1981), deputy assistant secretary for policy analysis in the Depart- ment of Energy (1978-1979), director of the department of economic analyses of CAB, and director of the office of economics of FTC (1976-1977). From 1970 to 1973 and 1975 to 1976, he taught econom- ics at the University of California at Berkeley. In 1994, he was part of the original investment group that founded Eastwind Airlines. Dr. Gaskins is chairman of Resources for the Future and serves on the boards of Anacomp, Inc., Northwestern Steel and Wire Company, and Sapient Corporation. He is an engineer (M.S.E., University of Michi- gan) and earned his Ph.D. in economics from the University of Michigan. Jose A. Gómez-Ibánez is professor of public policy and urban planning in the Graduate School of Design and in the John F. Kennedy School of Government, Harvard University. Dr. Gómez-Ibáñez has authored or co-authored many publications on transportation policy and planning, including Autos, Transit, and Cities; Regulationfor Revenue: The Political Economy of Land Use Exactions; Cases in Applied Microeconomics; and Going Private: The InternationalExperience with Transport Privatization. He has served on several Transportation Research Board study commit- tees and was chairman of the Committee for Study of Public Policy for Surface Freight Transportation. While on leave from the Harvard faculty, he was an economist for the President's Council of Economic Advisors. Dr. Gómez-Ibánez earned his Ph.D. in public policy from Harvard Uni- versity. Cornish F. Hitchcock is an attorney in private practice. In 1998, he stepped down from his position as an attorney specializing in aviation issues at Public Citizen Litigation Group. During the preceding 20 years, he testified before Congress on more than 40 occasions concern- ing a wide range of aviation, transportation, and consumer issues. He has been a member of the Federal Aviation Administration's Regulatory Negotiation Advisory Committee and the International Foundation for

Study Committee Biograp hical Information 193 Airline Passengers. He has published several artides on airline deregu- lation and air safety and was a member of the NRC Committee for the Study of Air Passenger Safety and Service Since Deregulation. Mr. Hitchcock earned his J.D. from Georgetown University. Aifted E. Kahn is the RobertJulius Thorne Professor of Political Econ- omy, Emeritus, at Cornell University and special consultant with National Economic Research Associates, Inc. (NERA). He joined the faculty of the Cornell Economics Department in 1947 and was chairman from 1958 to 1963. He was Dean of the College of Arts and Sciences from 1969 to 1974. His expertise is in public utility regulation. He was chair- man of the New York State Public Service Commission from 1974 to 1977. From 1977 to 1978 he was chairman of CAB, a term that coin- cided with passage of the Airline Deregulation Act of 1978. From 1978 to 1980, he was advisor to the President on inflation. Dr. Kahn has writ- ten extensively on public utility regulation—most prominently, the two- volume Economics of Regulation (MIT Press, reprinted 1988) and Let- ting Go: Deregulating the Process of Deregulation, 1998—and on airline competition, telecommunications deregulation, and antitrust policy. He earned his Ph.D. in economics from Yale University. Randall Malin, a 31-year veteran of the airline industry, is an air trans- portation consultant in Los Gatos, California. From 1996 to 1998 he was president and CEO ofTravelNet, Inc. which developed software for corporate travel management. From 1980 to 1992, he was executive vice president of marketing for USAir, and was vice chairman of the board of directors from 1989 to 1992. At USAir he was responsible for sched- uling, pricing, yield management, sales, distribution, and ground and inffight services. He also served on the board of directors of the Apollo computer reservations system, which was partly owned by USAir, and was chairman of the board of directors of the Airline Tariff Publishing Company. During his 20 years at American Airlines (1961-1980), Malin held various executive titles, induding vice president of resource planning and vice president of sales and advertising. In the latter posi- tion, he was responsible for marketing Sabre, American's computer reservations system. 'While at American and USAir, he was involved in

194 ENTRY AND COMPETITION IN THE U.S. AIRLINE INDUSTRY a wide variety of regulatory matters and testified frequently at CAB, De- partment of Transportation, and Congressional hearings. From 1974 to 1975 he was American's spokesman on airline deregulation and testified before committees of Congress. Malin graduated from Dartmouth Col- lege and earned his M.B.A. from Dartmouth's Amos Tuck School of Business. Steven A. Morrison is professor of economics and chair of the Depart- ment of Economics of Northeastern University and managing editor of the Journal of Transport Economics and Policy. Before coming to North- eastern in 1982, he was an assistant professor at the University of British Columbia. He has held visiting positions at Harvard, Massachusetts Institute of Technology, the London School of Economics, and the Brookings Institution. Dr. Morrison specializes in transportation eco- nomics and has written extensively on the changes in fares, service, and competition in the airline industry since deregulation in 1978. His pub- lications include The Evolution of the Airline Industry (Brookings Insti- tution, 1995) and The Economic Effects ofAirline Deregulation (Brook- ings Institution, 1986). He has testified before Congress on airline competition matters on several occasions. Dr. Morrison earned his Ph.D. in economics from the University of California at Berkeley. Sherwin Rosen is Edwin A. and Betty L. Bergman Distinguished Ser- vice Professor of Economics at the University of Chicago. From 1988 to 1994 he was chairman of the Economics Department. He joined the fac- ulty of the University of Chicago in 1977. Before that, he was Kenan Professor of Economics at the University of Rochester. Dr. Rosen is editor of the Journal ofPolitical Economy and has served on the editorial boards of several economic journals. Dr. Rosen is president-elect of the American Economics Association and previously served as vice presi- dent. He is a fellow of the Econometric Society and American Academy of Arts and Sciences and a member of the National Academy of Sciences. Dr. Rosen has written extensively on product differentiation and price discrimination and his main fields of expertise are industrial organiza- tion and labor economics. He earned his Ph.D. in economics from the University of Chicago.

The Transportation Research Board is a unit of the National Research Council, which serves the National Academy of Sciences and the National Academy of Engineering. The Board's mission is to promote innovation and progress in transportation by stimulating and con- ducting research, facilitating the dissemination of information, and encouraging the imple- mentation of research results. The Board's varied activities annually engage more than 4,000 engineers, scientists, and other transportation researchers and practitioners from the public and private sectors and academia, all of whom contribute their expertise in the pub- lic interest. The program is supported by state transportation departments, federal agencies including the component administrations of the U.S. Department of Transportation, and other organizations and individuals interested in the development of transportation. The National Academy of Sciences is a private, nonprofit, self-perpetuating society of distinguished scholars engaged in scientific and engineering research, dedicated to the furtherance of science and technology and to their use for the general welfare. Upon the authority of the charter granted to it by the Congress in 1863, the Academy has a mandate that requires it to advise the federal government on scientific and technical matters. Dr. Bruce M. Alberts is president of the National Academy of Sciences. The National Academy of Engineering was established in 1964, under the charter of the National Academy of Sciences, as a parallel organization of outstanding engineers. It is autonomous in its administration and in the selection of its members, sharing with the National Academy of Sciences the responsibility for advising the federal government. The National Academy of Engineering also sponsors engineering programs aimed at meeting national needs, encourages education and research, and recognizes the superior achieve- ments of engineers. Dr. William A. WuIf is president of the National Academy of Engineering. The Institute of Medicine-was established in 1970 by the National Academy of Sciences to secure the services of eminent members of appropriate professions in the examination of policy matters pertaining to the health of the public. The Institute acts under the responsi- bility given to the National Academy of Sciences by its congressional charter to be an advis- er to the federal government and, upon its own initiative, to identify issues of medical care, research, and education. Dr. Kenneth I. Shine is president of the Institute of Medicine. The National Research Council was organized by the National Academy of Sciences in 1916 to associate the broad community of science and technology with the Academy's purpose of furthering knowledge and advising the federal government. Functioning in accordance with general policies determined by the Academy, the Council has become the principal operating agency of both the National Academy of Sciences and the National Academy of Engineering in providing services to the government, the public, and the scientific and engineering communities. The Council is administered jointly by both the Academies and the Institute of Medicine. Dr. Bruce M. Alberta and Dr. William A. Wuif are chairman and vice chairman, respectively, of the National Research Council. -S R T C iiAI ACADM . .1 Advisers to the Nation on Science, Engineering, and Medicine National Academy 01 Sciences National Academy of Engineering Institute of Medicine National Research Council

C C rIS B N[OD09IO06 C C C