Reply Comments Of The Department Of Justice

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


BEFORE THE
DEPARTMENT OF TRANSPORTATION
WASHINGTON, D.C.



NOTICE OF PROPOSED RULEMAKING     
COMPUTER RESERVATION SYSTEM
REGULATIONS


|
|
|
|
|
|         
Docket Nos.     OST-97-2881
OST-97-3014
OST-98-4775
OST-99-5888

REPLY COMMENTS OF THE DEPARTMENT OF JUSTICE

    Communications with respect to this document should be addressed to:

R. Hewitt Pate
Acting Assistant Attorney General
Antitrust Division

Deborah P. Majoras
Deputy Assistant Attorney General
Antitrust Division

Craig T. Peters
Economist

June 9, 2003

Roger W. Fones
   Chief
Donna N. Kooperstein
   Assistant Chief

Robert D. Young
Jennifer L. Cihon
   Attorneys

Transportation, Energy, and Agriculture Section
U.S. Department of Justice
325 Seventh Street, N.W.
Washington, DC 20530
Telephone: 202/353-7131
Facsimile: 202/307-2784
E-mail: robert.young@usdoj.gov


TABLE OF CONTENTS

  1. Summary of Recommendation
  2. Background and History of CRS Rules
    1. CRS Market Power Over Airlines
    2. Airline Market Power In Passenger Air Transportation Markets
    3. Airline Ownership of CRSs
    4. Exclusionary Practices Engaged In By Vertically Integrated CRSs
    5. The Regulatory Response

      1. Rules Designed to Dissipate CRS Market Power Over Time

        1. Mandatory Participation
        2. System Tying
        3. Most Favored Nation (Parity) Clauses
        4. Travel Agency Contracts

      2. Rules Designed To Constrain the Exercise of CRS Market Power

        1. Display Bias
        2. Equal Functionality
        3. Non-Discriminatory Booking Fees

  3. The CRS Industry Today
  4. Comments on Proposed Regulations
    1. Potential Harm From Strategic Exercise of CRS Market Power
    2. Potential Harm From Nonstrategic Exercise of CRS Market Power
    3. Costs and Benefits of Regulatory Options

      1. Rules Intended to Dissipate CRS Market Power Over Time

        1. Mandatory Participation
        2. System-Tying Rules
        3. Most Favored Nation (Parity) Clauses in Airline-CRS Contracts
        4. Regulations Governing the Terms of Travel Agency Contracts

      2. Rules Aimed at Constraining the Exercise of CRS Market Power

        1. Display Bias
        2. Equal Functionality
        3. Non-Discriminatory Booking Fees

      3. Other Proposals

        1. Booking Data Sales
        2. Internet Distribution

  5. Conclusion

BEFORE THE
DEPARTMENT OF TRANSPORTATION
WASHINGTON, D.C.



NOTICE OF PROPOSED RULEMAKING     
COMPUTER RESERVATION SYSTEM
REGULATIONS


|
|
|
|
|
|         
Docket Nos.    OST-97-2881
OST-97-3014
OST-98-4775
OST-99-5888

REPLY COMMENTS OF THE DEPARTMENT OF JUSTICE

On November 15, 2002, the Department of Transportation (DOT) issued a notice of proposed rulemaking (NPRM) concerning the rules that govern airline computer reservations systems (CRS). DOT proposes to eliminate certain rules and to maintain or revise others in order to adapt the rules to the changes occurring in the air transportation industry. A round of comments on the NPRM was filed on March 17, 2003, and DOT held a public hearing on May 22, 2003. The United States Department of Justice (DOJ) submits the following comments in response to DOT's notice and the comments submitted to date.

I. Summary of Recommendation

DOJ recommends that DOT substantially reduce its regulatory supervision of the CRS industry for two reasons. First, many of the existing regulations have been ineffective in achieving their goals, and their retention may impose costs of their own on consumers. Second, two recent changes in the industry, divestiture of CRSs from their airline owners, which affects the CRSs' incentives to diminish airline competition, and increased airline use of the Internet to bypass CRSs, which creates some distribution competition, have lessened the need for industry-wide regulation. DOT should therefore rely largely on market forces to lower costs and improve service quality in the distribution of airline tickets to American consumers, recognizing that case-by-case law enforcement or administrative action may be necessary to prevent, investigate, and remedy future anticompetitive conduct and abuses in the CRS industry.

DOJ agrees with DOT that despite the recent growth of Internet distribution, CRSs continue to have market power over airlines. DOJ, however, has seen no evidence that regulations designed to erode that power have succeeded in the past or are likely to improve the situation in the future. Rather, experience over the past twenty years has shown that many of the existing regulations, such as the mandatory participation and subscriber contract rules, have been ineffective in reducing CRS market power. At the same time, however, these rules may have created their own inefficiencies, and some have been costly to enforce.

Rules aimed primarily at prohibiting the exercise of CRS market power against targeted airline rivals appear to have had greater success. The rules against display and functionality bias appear to have achieved their goals without significant enforcement problems or harmful side effects. Because these anti-bias rules appear to have been successful in ending prior abuses, because bias would likely occur in the absence of those rules, and because such bias harms consumers without countervailing efficiency benefits, DOJ recommends that DOT retain these rules.

The likely effect of continuing the non-discriminatory price rule is more ambiguous, particularly in light of changes in the industry since the last CRS proceeding. While the non-discriminatory price rule has been effective in eliminating the use of strategic CRS pricing (discriminatory fees to target specific airline competitors), it has not prevented supracompetitive, albeit uniform, CRS fees. Supracompetitive fees, even when not used to target specific airlines, are inefficient and harm consumers by artificially raising the cost of air travel. In addition, by forcing the CRSs to charge uniform prices, the rule itself can create inefficiencies and consumer harm.

DOT has not proposed a rule to remedy non-strategic supracompetitive pricing by CRSs. In the past, DOJ advocated the zero price rule to constrain both the strategic and nonstrategic exercise of CRS market power over price.(1) We pointed out the advantages of a structural rule, which relies on properly-aligned incentives. We also noted that because the zero price rule would prevent any airline from paying the CRS to disadvantage its competitors, it could eliminate the need for other rules designed to constrain the strategic exercise of CRS market power. It is clear from the NPRM, however, that DOT is unlikely to adopt the zero price rule or any other measure aimed generally at supracompetitive booking fees. 67 FR 69399.

Given the disadvantages of the nondiscrimination rule and given the potential for the Internet to diminish CRS market power, DOT should forgo any pricing rule at this time. Instead, DOT should assess, after some reasonable transition period, whether developments in the market continue to erode CRS market power or inhibit its exercise. If they have not, DOT should then reconsider the zero price or another pricing rule.

Finally, DOJ's recommendation assumes that the recent divestitures represent a permanent change in the ownership structure of the industry. DOT therefore should make clear that any attempt at reintegration into CRS by airlines will be closely scrutinized by the appropriate enforcement agencies.

II. Background and History of CRS Rules

The Civil Aeronautics Board ("Board") first imposed regulations on CRSs in November 1984, in an effort to remedy competitive problems in the CRS and airline industries. As successor to the Board, DOT revised and readopted the rules in 1992. The DOJ has conducted a number of investigations in the CRS industry, as well as participated actively in rulemaking proceedings and in a subsequent DOT proceeding that resulted in amendments on parity provisions in CRS contracts.(2) Although the DOJ has occasionally disagreed with DOT on particular rules, we have in the past generally supported regulation in this area.

The need for regulation arose from the prevalence of market power in two adjacent markets — the market for CRS services to airlines and the market for passenger air travel — and the vertical integration of firms participating in both markets.

  1. CRS Market Power Over Airlines

For decades, airlines have relied heavily on travel agents to provide consumers with information on their services and to sell their tickets, and travel agents, in turn, have overwhelmingly relied on CRSs to obtain the flight information, make the bookings and issue the tickets. At the time of the first CRS rulemaking, travel agents accounted for 60% of all airline revenues, and 90% of those sales were made by agents using a CRS.(3) By the time DOT began considering renewal of the rules in 1989, travel agents accounted for about 75% of airline ticket revenues, and over 95% of agents relied on a CRS.(4) Moreover, because the vast majority of agents used only one CRS, each CRS controlled access to a discrete group of agents, representing a large share of airline passenger revenue. Given these industry conditions, airlines did not consider different CRSs to be substitutes for one another. As a consequence, we concluded in our 1989 comments that "each CRS constitutes a separate market for air carriers, and each is a monopolist with market power over carriers that want to sell airline tickets in areas where the CRS has a significant number of travel agents."(5)

The CRSs also sold information services to travel agents. A travel agent only needed to subscribe to one CRS to obtain information on most airlines and, thus, in contrast to the airline side, the CRSs were effective substitutes for one another for their travel agent customers. As a consequence, that side of the industry enjoyed effective competition on both price and service. Indeed, there was considerable evidence that CRSs were sharing some of their high booking fees with subscribers by offering low and even negative prices to agents so that a CRS could increase its travel agent base and thus its market power over airlines.(6)

  1. Airline Market Power In Passenger Air Transportation Markets

The hub/spoke route systems that most airlines adopted since the industry was deregulated resulted in regional airline concentration, with most city pairs served by only a few airlines. Hubbing airlines dominated service on the city-pair routes to and from their hub cities, often operating well over half of the flights in and out of their hub cities. Travel agents in a region had a strong interest in maintaining strong relationships and good communication with the airline that provided the services most frequently used by their customers. Thus, travel agents in a region dominated by a particular carrier disproportionately used the CRS they perceived to provide the best service in booking the dominant airline — historically, the CRS in which the dominant airline had an ownership interest. Travel agents in a region were also more likely to have incentive commission deals with the dominant local airline.

  1. Airline Ownership of CRSs

In 1984, there were six CRSs, five owned by airlines: Sabre was owned by American Airlines; Apollo was owned by United Air Lines; System One was owned by Eastern Airlines; DATAS II was owned by Delta Airlines; and PARS was owned by TWA. Only one CRS, MARS PLUS, was unaffiliated.

By 1990, the number of CRSs had decreased to four, and all had domestic airline owners: Sabre was still owned by American Airlines; Apollo was now owned by United Airlines, USAir, and several foreign carriers; System One was now owned by Texas Air Corp.; and Worldspan was owned by Delta, Northwest and TWA.(7)

  1. Exclusionary Practices Engaged In By Vertically Integrated CRSs

The presence of CRS and airline market power gave airline-owned CRSs the incentive and ability to restrict competition in both markets. In airline markets, the CRSs were able to bias information displays to give greater prominence to their affiliated airline flights and to suppress information on competitors' flights. The CRSs found that even small changes in display prominence could have significant effects in diverting passengers to their flights from competitors.(8) In addition to enhancing the airline owner's profitability, this traffic diversion discouraged service from its competitors, thereby raising prices to consumers on concentrated routes.

The CRSs also biased the quality and timeliness of the information their systems displayed. For example, CRSs often contained more accurate information on the host carriers' schedules, fares, and fare rules because the vendor controlled the timing in which changes in that information were loaded into its computer.(9) Similarly, a CRS's information on seat availability and reservation confirmations for the owner-carrier tended to be more timely, accurate and reliable than similar information for other airlines, because the host carriers' internal reservations system was usually housed in the same computers with its CRS, while similar information for other carriers had to be transmitted between systems. Because CRSs provided more accurate and reliable information and bookings on the host carrier, travel agents were more confident in booking flights on the host carrier than on other carriers, and therefore tended to book disproportionately on the host carrier. In this way, functional bias, like display bias, led to the diversion of passengers to the CRS-affiliated carrier away from its competitors, ultimately reducing competitive options, particularly in cities dominated by the dominant airline's CRS.

Booking fees provided the CRS owner-carriers with another way to exercise their CRS market power against competing airlines. CRSs used booking fees as a strategic weapon to raise rival airlines' costs, charging higher access fees to airlines that were direct airline competitors than they charged to non-competitors ("discriminatory booking fees"). 1985 Report to Congress at 12.

The owner-carriers also exercised their airline market power to reduce CRS competition. For example, owner airlines would refuse to pay override commissions or provide other marketing benefits to local travel agencies that did not use their systems. 57 FR 12600, 12625. Further, to maintain the superiority of their CRSs in regions where the owner-carriers had strong airline presence, the owner-carriers would refuse to participate fully in competing CRSs.(10) In such cases, the revenues from maintaining CRS and airline dominance in their core markets more than compensated for any loss of airline sales through the competing CRSs. 57 FR 43800.

Hence, there was a mutually reinforcing effect between airline share and CRS share in regional markets. Because CRSs affiliated with the dominant air carrier in a region had a substantial advantage over other CRSs in obtaining travel agent subscribers, the CRSs tended to have dominant shares of travel agent business in cities where their owner-carriers had dominant shares of the airline service.(11) Once a CRS became dominant in a region, bias in favor of the owner-carrier helped it maintain its dominant airline position by diverting passengers from its competitors, decreasing their ability to enter or expand service in those markets. The dominant airline's CRS then shared these "incremental" airline revenues with travel agents by lowering subscriber fees, helping the CRS to maintain its dominant position.

  1. The Regulatory Response

The regulations adopted in response to these circumstances pursued both long-term and short-term remedies. The long-term goal was to dissipate CRS market power by facilitating multiple-CRS use by travel agents and by promoting new entry by non-airline CRS vendors, expansion by the smaller CRSs, and a decrease in regional dominance by CRSs. Airline Marketing Practices at 80; 57 FR 43817; 56 FR 12590-12591. The short-term goal was simply to prohibit directly the most obvious exercises of CRS market power to harm competition in passenger air transportation markets. See 49 FR 11657.

Rules aimed at dissipating CRS market power over time include the mandatory participation rule (requiring airlines that own a CRS system to participate in competing CRS systems at the same level at which they participate in their own systems) and a prohibition against parity clauses in CRS contracts (requiring airlines to participate in the contracting CRS at a level at least as high as they participate in any other CRS). Additional long-term efforts to dissipate CRS market power have included rules that impose restrictions on the terms of CRS vendors' contracts with travel agent subscribers, including the length of the contracts, minimum use clauses, clauses tying system use to commissions, clauses preventing the use of multiple systems by a subscriber, and clauses preventing subscribers from using their own hardware to access competing systems.

The short-term solution was to establish rules designed to limit directly the exercise of CRS market power to harm airline competition, including: (1) a prohibition against display bias, (2) a prohibition on functionality or "architectural" bias, and (3) a prohibition against discriminatory booking fees.

    1. Rules Designed to Dissipate CRS Market Power Over Time
      1. Mandatory Participation

The existing mandatory participation rule (14 C.F.R. § 255.7) was intended to prevent an airline with a dominant position in local airline markets from obtaining a similar dominant position for its affiliated CRS. 56 FR 12608. The mandatory participation rule applies only to system owners.(12) It was also hoped that the rule would help diminish the already dominant CRS shares in numerous areas. 57 FR 43795. The greater the local dominance by a CRS, the better able it is to favor its affiliated airline by foreclosing rival airlines from access to local travel agents.

      1. System Tying

The existing ban on "system tying" prohibits an airline owner of a CRS from tying travel agent commissions to CRS use. 14 C.F.R. § 255.8 (c) and (d). Like the mandatory participation rule, this rule also is intended to prevent a locally dominant airline from gaining a similarly dominant CRS share in a region. For example, the dominant airline at a hub city might otherwise require agents in that city to use its affiliated CRS in order to receive commission overrides. 56 FR 12600, 12625.

      1. Most Favored Nation (Parity) Clauses

In 1997, DOT adopted a provision that prohibited a CRS from requiring an airline to participate in its system at the same level as or at a higher level than the airline participates in other CRSs. 14 C.F.R. § 255.6(e). The ban does not apply to airlines that own or market a system (indeed, the mandatory participation rule imposes a parity requirement on such airlines.) The parity rule was designed to promote competition among the CRS for airlines and to enable airlines to use alternatives to CRSs. 62 FR 59784-59785.

      1. Travel Agency Contracts

The travel agency contract rules are generally intended to reduce CRS market power by reducing the need for airlines to subscribe to all systems. Limits on contract length (14 C.F.R. § 255.8(a)) and a prohibition against "rollover clauses" (14 C.F.R. § 255.8(a)) were designed to facilitate agents' replacing one system with another. DOT later adopted a prohibition against "minimum use" clauses (14 C.F.R. § 255.8(b)), and a prohibition of contracts restricting agents' use of equipment to access alternative systems (14 C.F.R. § 255.9) to encourage agencies to simultaneously subscribe to and use multiple systems or other distribution channels.

    1. Rules Designed To Constrain the Exercise of CRS Market Power
      1. Display Bias

The current display bias rules (14 C.F.R. § 255.4) are relatively detailed in their proscriptions. In general, the display bias rules were designed to prevent CRSs from limiting downstream airline competition by offering preferential treatment to affiliated airlines. 67 FR 69395. They require each CRS to maintain at least one integrated display containing the schedules, fares and availability of all participating carriers. The rules do not prescribe the criteria that a CRS must use to order the information in its integrated displays, but they do prohibit the use of carrier identity as a factor, limit the manner in which the displays may select and rank connecting flights, and require that the display criteria be disclosed to participating carriers.

      1. Equal Functionality

The equal functionality rules were designed to ensure that the CRSs could not bias the accuracy, reliability and accessibility of information for any airline. The current rules require equal access to enhancements and prohibit the use of default features that favor the airline owner of a CRS. 14 C.F.R. § 255.5. In addition, a provision contained in the section regulating display bias requires the CRSs to apply the same standards of care and timeliness to loading information concerning participating carriers as they apply to their owners. 14 C.F.R § 255.4(d).

      1. Non-Discriminatory Booking Fees

The prohibition on discriminatory booking fees (14 C.F.R. § 255.6(a)) is primarily intended to prevent system-owning airlines from imposing higher booking fees on their competitors. Thus, the primary purpose of the rule was to constrain strategic exercises of CRS market power. 49 FR 32522. Secondarily, it was originally hoped that the rule would limit the non-strategic exercise of CRS market power by giving smaller airlines with less bargaining power vis-a-vis CRSs the benefits of any lower fees negotiated by larger airlines with relatively greater bargaining power. Id.; 54 FR 38873.

III. The CRS Industry Today

Perhaps the most significant change that has occurred in this industry since the last review of the rules is the divestiture of the CRSs by their domestic airline owners. Sabre became independent of American in March 2000; Cendant purchased full ownership of Galileo in October 2001; and the airline owners of Worldspan announced an agreement to spin off their ownership interests to a private equity firm in March 2003.(13) Amadeus has not had a U.S. carrier equity owner since 1997, when Continental sold its interest, although Amadeus continues to be partly owned by three foreign airlines (Air France, Lufthansa and Iberia). Assuming, therefore, that the sale of Worldspan is completed, the vertical ownership relationships that made it likely that CRS market power would be used strategically to distort competition in airline markets and that airline market power would be used strategically in CRS markets, will have been eliminated.

In another significant development since the last CRS proceeding, airlines have begun to change the way they distribute their products, particularly by using the Internet. The existence of the Internet as a low-cost distribution outlet has enabled some airlines to put competitive pressure on travel agencies and CRSs to reduce costs through traditional channels as well. Although the percentage of airline revenues generated through CRSs remains higher today than it was in 1982, it is lower than in 1988, due largely to increased efforts by airlines to bypass CRSs. Almost all airlines have expanded their direct sales to consumers through their own Internet sites, and a few highly successful low-cost airlines have developed distribution strategies that rely less heavily on travel agents (and therefore CRSs).

Recent airline data reveal this distinct, though still limited, movement toward Internet-based distribution. Brick and mortar travel agents' share of sales for Orbitz airline owners has decreased from 76% in May 2000 to 67% in March 2002,(14) with most of the change attributable to airlines' website sales and on-line travel agent sales. (15) The airlines' own sites do not depend on CRSs. Although online travel agencies depend on CRSs for some or all of their information, a base of online agencies adds little to CRS market power over airlines. As DOT has noted, passengers using online agencies frequently search multiple Internet sources before making a travel purchase.(16) Hence, an airline excluded from one online agency site is more likely to be considered in the purchase decision of a passenger using the Internet unless many of the passenger's information sources (including airline websites) do not offer that airline for sale.

At this point, the evidence suggests that leisure passengers with relatively simple itineraries are best suited to using the Internet. Comments of American Airlines, Inc. at 16-17. Online agencies sell most airline tickets to price-sensitive leisure passengers.(17) Business passengers, especially those traveling on behalf of smaller businesses, are likely increasing their use of the Internet as well.(18) However, the National Business Travel Association reports that today less than 10% of corporate travel is booked through the Internet and that many corporations forbid their employees from booking travel on the Internet, even if the employees find a lower fare through that channel.(19) Comments of the National Business Travel Association at 11, 15; cf. Comments of American Airlines, Inc. at 16-17 (large corporations have complex air travel needs and prefer to use travel professionals).

In most other respects, the CRS industry is not substantially different than it was twelve years ago. Notwithstanding increased use of the Internet, traditional network carriers are still largely dependent on CRSs and "brick-and-mortar" travel agents.(20) CRS market power over airlines derives primarily from the inability of most airlines to withdraw from any CRS. For most airlines, travel agents continue to be a critical distribution outlet. Travel agents, in turn, rely heavily on CRSs, even though almost all travel agencies have Internet access.(21) Further, most travel agents rely primarily on one CRS.(22) It is often costly and inefficient for agents to use multiple CRSs because it requires agencies to incur additional training costs and to implement accounting, billing and recordkeeping systems that consolidate all of the CRS transactions. Some airlines have developed CRS bypass programs for agents, but so far, these programs have not been heavily used.(23) Some airlines assert that increasingly large payments from competing CRSs to travel agencies prevent agencies from using lower-cost channels.(24)

The number of CRSs decreased from six in 1984 to four in 1990, where it remains today. It is clear, therefore, that those CRS rules designed to encourage new CRS entry have not succeeded. Nor have CRS rules prevented booking fees from rising and remaining a significant component of airline distribution costs.(25)

IV. Comments on Proposed Regulations

Regulation should be considered (or extended) only when (1) market participants have substantial and durable market power that will likely harm consumers directly, or will be exercised in ways that exclude or limit competition in contiguous markets, and (2) the regulation will likely be effective and enforceable without imposing significant costs of its own. Prohibiting practices whose potential effects are ambiguous, or that would produce efficiencies, can have the effect of raising consumer prices more than the practices themselves.

When the need for regulation is in doubt, a less regulatory approach will generally be preferable. While there will inevitably be some uncertainty about the consequences of eliminating particular rules, regulation should be maintained only when the likely benefits outweigh the likely harms. DOJ's mandate to enforce the antitrust laws can be relied upon to prohibit monopolization and anticompetitive agreements. As DOT has stated, "creating rules designed to eliminate all risk of possible illegal conduct would likely interfere with legitimate business practices." 67 FR 69389. When the risk level is uncertain, it will often be most appropriate to rely on law enforcement on a case-by-case basis to deal with anticompetitive conduct, rather than on industry-wide regulation. At the same time, when a particular form of conduct poses a severe risk of adverse competitive consequences with minimal offsetting efficiencies, explicit rules prohibiting such conduct are justified.

  1. Potential Harm From Strategic Exercise of CRS Market Power

Despite the airlines' CRS divestitures, some incentives and ability to engage in strategic conduct to limit airline competition remain. On the one hand, CRSs retain their ability to exercise their market power in ways that favor one airline over another, but, on the other hand, without airline ownership, CRSs have no direct incentive to do so. Conversely, airlines have a clear incentive to use CRS market power to disadvantage competitors, but, without ownership of CRSs, the airlines have no ability to do so. The airlines' and CRSs' respective incentives and abilities to exercise market power can be aligned through contract to their mutual advantage -- an airline can pay a CRS to use its market power to disadvantage the airlines' competitors in the airlines' hub markets. As before, the CRS could bias against the targeted airlines in display and functionality. And, as before, display and functionality bias would divert passengers without regard to airlines' prices or quality In each case, the effect would be to deter expansion and entry by potentially more efficient competitors and perhaps even cause their exit from some markets.(26)

While the likelihood of "bias buying" cannot be predicted with certainty, CRSs apparently are already planning on its sale.(27) Experience shows that bias is easy to implement and effective in limiting competition. Experience also shows that CRSs and airlines are able to estimate the value of incremental rents that could be shared through bias. As DOJ noted in its 1989 Comments, CRS vendors were able to estimate and document the incremental earnings achieved through bias. Indeed, CRS vendors used these estimates to determine how to price their systems to different subscribers. The vendors would assume that as a result of placing their system with an agent, their affiliated carriers would earn a certain percentage of additional revenues by booking more passengers on their flights than they otherwise would, and the vendors would then use those expected revenues in determining the amount of the discount to give the agent.(28) In 1988, CRS vendors' estimates of the additional airline revenues earned from their subscribers as a result of bias ranged from nine to fifteen percent of airline revenues sold through the CRS, amounting to approximately $900 million to $1.5 billion of airline revenue.(29)

The continued disproportionate strength of CRSs in the hubs of their former airline owners makes for natural partners. If the dominant CRS and airline in a city can reach an acceptable bargain, both can profit. And if bias buying does occur, consumers will be the ultimate losers — paying higher fares in the "protected" markets. For these reasons, DOT should make it clear that future contracts and transactions between CRSs and airlines will be monitored closely to ensure that they do not result in re-integration. Also, as discussed below, some continued regulation may be advisable as well.

  1. Potential Harm From Nonstrategic Exercise of CRS Market Power

The airlines' CRS divestitures leave unaffected the incentive and ability of CRSs to fully exercise their market power in nonstrategic ways. The CRSs may still have incentives to charge supracompetitive booking fees and, absent a price rule, the only constraint on their ability to do so would be any countervailing airline bargaining power.

DOT has not proposed a rule to remedy nonstrategic supracompetitive pricing by CRSs. In the past, DOJ advocated the zero price rule to constrain both the strategic and nonstrategic exercise of CRS market power over price. We pointed out the advantages of a structural rule, which relies on properly-aligned incentives. We also noted that because the zero price rule would prevent any airline from paying the CRS to disadvantage its competitors, it could eliminate the need for other rules designed to constrain the strategic exercise of CRS market power. It is clear from the NPRM, however, that DOT is unlikely to adopt the zero price rule or any other measure aimed generally at supracompetitive booking fees. 67 FR 69399.

Instead, DOT will be relying on countervailing market power by airlines to constrain CRS booking fees. Although airline bargaining power has not in the past been sufficient to produce competitive booking fees, bargaining power of airlines could increase if their ability to shift sales to the Internet and other alternative channels continues to increase significantly. DOT should assess, after some reasonable transition period, whether the alternative distribution channels have continued to dissipate CRS market power. If they have not, DOT should then reconsider the zero price or any other pricing rule.

  1. Costs and Benefits of Regulatory Options

After twenty years, it is apparent that the rules aimed at dissipating CRS market power over the long term have not produced the desired results. There are fewer CRSs today than in 1984, and together they represent an equally large portion of carriers' revenues. On average, therefore, each of the remaining CRSs represents an even larger percentage of passenger revenues for airlines. Travel agents continue to be wedded to use of a single system. If it were clear that regional CRS shares had become more reflective of national shares, that might suggest a partial success of such rules; however, the evidence on this is conflicting.(30) Where there is little evidence that a rule has produced the desired benefits, and there is reason to believe the rule imposes costs of its own, it should not be readopted.

    1. Rules Intended to Dissipate CRS Market Power Over Time
      1. Mandatory Participation

The NPRM proposes either eliminating the mandatory participation rule, or expanding it to cover airlines that market a system in addition to those that own a system (67 FR 69393-69395). There is no clear evidence that the mandatory participation rule is reducing CRS market power, either nationally or locally.

In addition, practices prohibited by the mandatory participation rule are potentially efficiency-enhancing. If an airline dealing at arm's length with its CRS suppliers were free to reduce its level of participation in some systems, or to induce agents to use a low-cost system, the airline's bargaining power would be enhanced. The airline would therefore be in a better position to negotiate lower booking fees or to drive bookings toward lower-cost outlets.

The extent to which the mandatory participation rule has, in fact, prevented airlines that own systems from negotiating better CRS access terms or otherwise lowering their CRS costs is unclear. Nonetheless, because the mandatory participation rule does not appear to have contributed to any lessening of CRS market power, and may impose costs or create inefficiencies of its own, DOJ recommends elimination of the rule.

      1. System-Tying Rules

DOT is considering retaining the current ban on CRS owner-carriers tying agent commissions to CRS use. DOT is also considering extending the system-tying ban to airlines that market a CRS and expanding it to prohibit the tying of marketing benefits or access to corporate discount fares. (67 FR 69409-69410). The primary purpose of the system-tying rules has been to prevent a carrier with local airline passenger market dominance from using that dominance to increase the local CRS dominance of its owned (or chosen) system.

DOT notes, however, that the system-tying ban has presented significant enforcement problems. In the past, system owners have found alternative inducements to accomplish the same results, and have at times simply ignored the ban with impunity. Airline Marketing Practices at 91; 54 FR 38872; 56 FR 12625. Given the wide-ranging sales and marketing contacts between airlines and travel agents and DOT's limited enforcement resources, prohibitions of this type will continue to be costly and difficult to enforce, regardless of expansion or refinement.

With the elimination of CRS-airline ownership links, the likelihood that the system tying rules will help dissipate CRS market power becomes even more attenuated. The basis for considering such rules is that, even without ownership ties, airlines and CRSs might contractually agree to use system-tying for anticompetitive purposes. For example, a CRS might contractually agree to bias its service in favor of a partner airline in exchange for the airline's efforts to exclude a rival CRS from a local market. The likelihood of such developments is not clear, and in any event, there is no evidence that this rule reduced CRS market power, even when the systems were airline-owned. Concerns over bias are best dealt with directly. See infra at 30-32. For all of these reasons, the DOJ recommends eliminating the system tying prohibition.

      1. Most Favored Nation (Parity) Clauses in Airline-CRS Contracts

DOT proposes to prohibit several types of "most-favored-nation" (MFN) clauses in airline-CRS contracts. The proposed rules include the existing prohibition of parity clauses ("parity MFN"); a prohibition of clauses requiring "most-favored" access to fares and inventory ("fare MFN"); and a prohibition of clauses requiring nondiscriminatory treatment of a CRS's subscribers ("subscriber MFN"). 67 FR 69393.

MFN clauses can be used for anticompetitive purposes. DOJ filed comments in support of DOT's parity clause prohibition in 1996, based in large part on evidence it collected in a civil antitrust investigation of parity clauses. Similarly, when circumstances have warranted, DOJ has initiated antitrust enforcement actions against the use of similar clauses in other industries.(31) In the current Orbitz investigation, DOJ has been closely examining an MFN provision that — unlike those at issue here — involves agreement to the MFN by horizontal competitors that jointly own a distribution venture. The use of MFN clauses can, however, be an efficient contracting practice, so that a prohibition of such clauses should not be adopted without a clear basis for concluding that they are likely to have anticompetitive consequences.

One potential benefit of permitting MFN clauses is that they may encourage investments that would not otherwise be made. A CRS has a legitimate business interest in ensuring that it does not compete at a disadvantage relative to its rivals. In a changing competitive environment with significant uncertainty about future market conditions, it may be reasonable for a CRS to seek a commitment from participating airlines not to offer more favorable treatment to rival CRSs (or other distributors). If it is unable to negotiate such a commitment, it may be deterred from making quality-enhancing or cost-reducing investments.

Moreover, a prohibition of MFN clauses can be harmful if, by preventing airlines and CRSs from freely negotiating mutually agreeable contracts, the prohibition encourages the adoption of less efficient terms. In particular, permitting contracts with MFN clauses may promote lower booking fees, because a CRS may need to offer a discount in order to induce an airline to accept the clauses. For example, USAir recently agreed to provide all its web fares to Sabre in exchange for lower booking fees. By preventing such agreements, the proposed rules may prove to be an impediment to airlines' efforts to reduce their distribution costs.

On the other hand, MFN clauses may reinforce CRS market power over airlines, particularly if they discourage the development of alternative distribution channels. For example, a low-cost distribution channel on the Internet may not offer the same level of functionality as a CRS, but may nonetheless be able to attract usage by travel agents if it has preferential access to desirable fares and inventory from a significant number of airlines. By negotiating MFN clauses with a sufficient number of airlines, a CRS may be able to prevent the growth of this alternative, effectively imposing a barrier to entry into the air travel distribution marketplace. Individually, each airline may be willing to accept an MFN clause, particularly if it anticipates that a sufficient number of other carriers will also accept the clauses, so that the alternative outlet is unlikely to be successful. But collectively, the airlines may face higher distribution costs as a result of the MFN clauses. While it is uncertain whether this exclusionary strategy would be successful, the risk should not be dismissed lightly. An additional risk is that, if airlines agree to MFN clauses governing access to fares and inventory, they will have a reduced incentive to offer selective discounts through their websites or other distribution channels. This may have the effect of dampening competition among airlines.

Given the benefits of permitting more flexible negotiations and given that anticompetitive MFN clauses may be challenged in enforcement actions, it would be reasonable for DOT to conclude that the MFN prohibitions should be withdrawn. On the other hand, several of the concerns expressed above could argue against complete and immediate withdrawal of the MFN clause prohibitions. In view of the benefits of permitting more flexible negotiations, it would be undesirable to maintain the rules for an extended period, at least without further evidence that they are necessary. DOJ believes that DOT could reasonably decide either to withdraw the prohibitions on fare MFNs and parity MFNs, or retain such restrictions for a limited transition period.

There is less justification for DOT's proposed rule prohibiting subscriber MFNs that prevent airlines from offering different treatment to a CRS's subscribers. While this type of MFN clause might, in principle, make it more difficult for an airline to shift traffic to a system that offers reduced booking fees, there is no evidence that such clauses have actually had this effect. Moreover, the risk that such clauses will have an exclusionary effect on the development of alternative distribution channels has not been persuasively established. It is particularly telling that some of the airlines that this rule is intended to benefit have argued against its adoption.(32) Hence, DOJ recommends that this rule not be adopted.

      1. Regulations Governing the Terms of Travel Agency Contracts

DOT proposes to retain most current restrictions on subscriber contracts, and is considering expanding the current restrictions by restricting liquidated damages for early subscriber termination of contracts, further restricting contract length, and prohibiting "productivity pricing," i.e., CRS payments to subscribers for producing airline bookings. 67 FR 69406-69409.

Regulations designed to promote competition among CRSs by encouraging travel agencies to use multiple CRSs have been ineffective, and DOJ has expressed its skepticism of these rules in the past.(33) Even without the contract clauses, it is costly for travel agents to use multiple CRSs. Even if some increase in switching occurred as a result of the proposed rules, it seems unlikely that switching will be sufficiently prevalent to reduce most airlines' need to participate in all four CRSs. By contrast, to the extent they actually were effective in limiting CRSs' ability to bind subscribers to them, the rules could have a negative effect on CRSs' incentives to make investments that enhance their value to travel agencies, including increased automation, customized features and other functionality enhancements, and the provision or upgrade of equipment.

It also appears that competition among systems for travel agency business has resulted in changes in the kinds of restrictive terms of subscriber contracts that are permissible under the rules, but have concerned DOT in the past.(34) This suggests that behavioral rules that regulate the terms of CRS-subscriber contracts may be unnecessary because competition among CRSs for subscribers is apparently eliminating contracts that limit subscriber options.

In addition, since they were first promulgated, the subscriber contract rules have presented significant enforcement problems. The CRSs have created new ways to try to lock agencies into their systems more quickly and creatively than DOT has been able to revise its rules.(35) Travel agents and CRSs both benefit from the supracompetitive booking fees generated under the current structure of the market. They will naturally have incentives to develop and enter into contractual arrangements to share and extend those benefits. If productivity-type payments continue to be in the mutual interest of CRSs and subscribers, the parties will have incentives to tailor their arrangements to hide their true effects and achieve the same result, while complying with the letter of a prohibition against productivity pricing.(36)

Because the CRSs aggressively compete for travel agency business, and because the regulations governing subscriber contracts have been largely ineffective and may impose their own inefficiencies, DOJ recommends that DOT allow those rules to lapse.

    1. Rules Aimed at Constraining the Exercise of CRS Market Power

In contrast to the rules aimed at dissipating CRS market power, the various "anti-bias" rules aimed at prohibiting the exercise of CRS market power to harm downstream airline competition and consumers appear to have been generally successful. As noted above, the decisions of the major airlines to divest their CRS ownership may be in part due to the effectiveness of the anti-bias rules in constraining the strategic exercise of CRS market power downstream.

      1. Display Bias

DOT proposes to continue the prohibitions against display bias. 67 FR 69396. These rules appear to have been successful in eliminating display bias without producing significant enforcement problems. On the other hand, the fact that the rules are so detailed reflects the fairly large number of complaints over the years about particular display practices by CRSs. See discussion of display issues at 57 FR 43802-43808.

Display bias is the most obvious and direct form of system bias for an airline to purchase. It is also probably the easiest to "fine tune" to target particular airlines in particular markets. As discussed above, CRS owner-carriers discovered early that small changes in display position could produce large shifts in revenue. In addition, airlines that are dominant in a particular market have the greatest incentive and ability to purchase bias to undercut competitive offerings in those markets.(37) Thus, there is a strong likelihood that airlines would purchase CRS display bias to buttress the airlines' positions in local markets.

On the other hand, the potential costs of imposing regulations prohibiting display bias is not as obvious as it is for some other regulatory proposals. Unlike advertising, display bias does not assist the consumer in making an informed purchase decision. Moreover, because CRS display bias is aimed at the travel agent, the consumer may have limited ability to discern the bias and to obtain objective information. Given the consumer deception problems with display bias,(38) the potential for legitimate, pro-consumer benefits from this practice is much smaller than for other CRS practices under consideration, and the possible harm from regulation is less significant.

Because there are clear benefits from continued regulation of display bias, and the potential harm from the regulation is small, we recommend that DOT retain these regulations.

      1. Equal Functionality

Bias in the accuracy, reliability or accessibility of information available for airlines that competed with CRS-affiliated airlines was one of the primary ways in which CRS market power was used strategically to distort airline competition. The current rules, which DOT proposes to continue, require equal access to enhancements, prohibit the use of default features that favor the airline owner of a CRS, and require equal treatment with respect to the standards of care and timeliness used in the loading of airlines' information. 67 FR 69398. These rules appear to have succeeded in limiting vendor-carriers' ability to extract incremental revenue from functionality bias, and do not appear to have presented significant enforcement difficulties. Id.

Where price discrimination is permitted, however, both airlines and CRSs are more likely to have legitimate reasons to agree on distinctions with respect to access to enhancements than they would in regard to display bias, because enhanced or differential functionality and access to information depends in part on the desires, cooperation and investments of the airline. Thus, disparate treatment with respect to enhancements and pricing could allow CRSs to better tailor efficient price and service packages to airlines. Such freedom might also allow CRSs greater leeway to share with airlines the development cost and risk of new functions. For example, an airline might be made the "launch partner" for a new CRS function and be granted a certain period of exclusivity in exchange for sharing in the development and testing cost for that function. However, the same analysis does not apply to the current requirement that CRSs provide equal treatment to non-owner participating airlines in the care and timeliness with which they load information (14 C.F.R. § 255.4(d)). It is difficult to imagine a legitimate business reason for differential treatment in that regard.

Because the current regulation requiring equal access to defaults and service enhancements would be inconsistent with our recommendation to eliminate non-discriminatory pricing regulations, we also recommend that DOT eliminate these rules. However, we recommend that DOT retain an equal functionality provision similar to current section 255.4(d) prohibiting CRSs from discriminating among participants in the care and timeliness with which they load information.

      1. Non-Discriminatory Booking Fees

DOT is considering extending or withdrawing this rule. 67 FR 69398-69401. The rule was primarily intended to prevent system-owning airlines from imposing higher booking fees on their competitors. The rule was also intended to limit CRSs' ability to extract supracompetitive booking fees.

The rule appears to have been successful in preventing the strategic use of booking fees to harm airline competition, and there do not appear to have been significant enforcement or evasion problems. See 1989 Comments at 48. The elimination of airline ownership of CRSs, however, reduces the incentives for CRSs to engage in discriminatory conduct, and thus reduces the benefit of keeping the rule. Even in the absence of vertical integration, there remains a risk that airlines will contract with CRSs to pursue such cost-raising strategies to potentially anticompetitive effect. However, because offering more favorable terms to certain customers is not necessarily anticompetitive and because it does not appear that the current rule has been effective in constraining booking fees to competitive levels, any use of discriminatory booking fees for exclusionary or predatory purposes can and should be addressed on a case-by-case basis.

It is even unclear whether the rule has kept average booking fees lower than they would be otherwise. By preventing discriminatory price increases to carriers with limited bargaining power, the rule could reduce average booking fees. On the other hand, the reverse may also be true: by preventing selective discounts to carriers with greater bargaining power, the rule may actually increase average booking fees. While removing the prohibition against discriminatory fees would inevitably result in carriers with less bargaining power having higher CRS costs than others, it is uncertain how great these disparities would be and whether even greater disparities would likely be introduced by the "purchase" by airlines of booking fee bias against their competitors.

The long-term effects on efficient investment of removing the prohibition against discriminatory fees are also ambiguous. Increased booking fees for certain airlines may increase their incentives to invest in alternative distribution outlets. On the other hand, the incentives of carriers whose booking fees decrease to invest in CRS bypass may also decrease.

Continuing the non-discriminatory price rule has potential efficiency costs because it bars use of a possibly efficient form of pricing.(39) Although the CRSs can and do have fees that reflect cost differences for transactions, the rule may limit CRSs' ability to offer selective discounts, which can in some cases be an efficient response to the market.

In sum, continuing the non-discriminatory price rule is likely to prevent strategic use of booking fees to harm downstream airline competition, but be ineffective in forcing the uniform fees to competitive levels. Although there do not appear to be significant enforcement costs or difficulties, the rule does have potential efficiency costs. Accordingly, the DOJ recommends that the prohibition on discriminatory booking fees be eliminated.

    1. Other Proposals
      1. Booking Data Sales

DOT has proposed regulations restricting participating airlines' access to the detailed marketing and booking data generated and sold by the CRSs. 67 FR 69401-69404. Airlines use this data for marketing research, route planning, pricing, and revenue management decision-making purposes. Some airlines also apparently use the data to identify agencies and corporations for individualized sales efforts, to measure performance under their override commission programs, and for competitive intelligence. Id. Thus, the data is a powerful competitive tool, and because of its cost, volume and complexity, it may be a tool that can be used more effectively and efficiently by large airlines. DOT proposes to limit the use of this data because it is concerned that airlines that are dominant in a region may be using the data to discourage entry, by tracking the competitive initiatives of the entrants, and by punishing travel agencies who sell seats on the entrants. Id.(40)

Because carriers use the data for a variety of legitimate business purposes, there are clear benefits associated with the sale of this data. The NPRM recognizes this, and the proposed rule attempts to preserve some of the utility of the data "for legitimate marketing purposes." 67 FR 69403. On the other hand, the record before DOT does not contain enough information to evaluate the extent to which the data really is used for purportedly "illegitimate" uses, and whether such uses are, on balance, harmful to competition and consumers.(41) The harm from the sale of MIDT booking data has not been sufficiently well-documented to justify DOT's proposed restrictions on data that has uses that are clearly beneficial. Hence, DOJ recommends that the proposed rules governing the sale of booking data not be adopted.

      1. Internet Distribution

We agree with DOT's assessment that it would be undesirable to adopt any formal regulation of Internet-based distribution systems. 67 FR 69410-69412. As DOT notes, display bias does not appear to have presented significant problems so far. More importantly, unlike travel agents who are wedded to a single system, consumers who use Internet travel sites apparently can and do frequently switch among multiple sites to comparison shop, even for the same itinerary. 67 FR 69411. Moreover, regulation of such sites could discourage the development of the promising potential vehicle for the dilution of CRS market power, as DOT notes. 67 FR 69412. We therefore recommend to DOT that any additional regulation of Internet distribution is neither necessary nor justified at this time.

V. Conclusion

Recent industry changes have eliminated the need or utility for most of the regulations that currently govern CRSs. Therefore, with the exception of rules prohibiting display bias, bias in the loading of airline information and possibly certain types of MFN clauses, DOT should allow the rules to sunset as scheduled in January 2004.

Respectfully submitted,

R. Hewitt Pate
Acting Assistant Attorney General
Antitrust Division

Deborah P. Majoras
Deputy Assistant Attorney General
Antitrust Division

Craig T. Peters
Economist

Dated: June 9, 2003

_______________/s/________________
Roger W. Fones
Chief

_______________/s/________________
Donna N. Kooperstein
Assistant Chief

_______________/s/________________
Robert D. Young
Attorney

_______________/s/________________
Jennifer L. Cihon
Attorney

Antitrust Division
U.S. Department of Justice
Washington, DC 20530


FOOTNOTES

1. Such a rule would prohibit CRS vendors from charging or receiving payments from airlines for CRS-related services. Instead, CRSs would have to recover the full cost of airline bookings from travel agents.

2. See Comments of the Department of Justice, Notice of Proposed Rulemaking — Computer Reservation System Regulations, Docket No. 49812, September 19, 1996 ("1996 Comments"); Reply Comments of the United States Department of Justice, Notice of Proposed Rulemaking — Computer Reservation System Regulations, Docket No. 46494, filed August 8, 1991 ("1991 Reply Comments"); Comments of the United States Department of Justice, Notice of Proposed Rulemaking — Computer Reservation System Regulations, Docket No. 46494, filed July 9, 1991 ("1991 Comments"); Comments of the United States Department of Justice, Advance Notice of Proposed Rulemaking — Computer Reservation System Regulations, Docket No. 46494, filed November 22, 1989 ("1989 Comments"); Comments and Proposed Rules of the Department of Justice, Advanced Notice of Proposed Rulemaking — Airline Computer Reservation Systems, EDR - 466, Docket No. 41686 (before the Civil Aeronautics Board), filed November 17, 1983 ("1983 Comments"). The Department also provided a report to Congress on the CRS industry in 1985. See 1985 Report of the Department of Justice to Congress on the Airline Computer Reservation System Industry ("1985 Report to Congress").

3. 1985 Report to Congress at 7. See also, 49 FR 11644 at 11647: Travel agents sold 38% of domestic airline revenue in 1977, 53% in 1978, and 60-65% in 1982.

4. 1989 Comments at 10. Today 91% of agents rely on CRSs. Comments of the American Society of Travel Agents, Inc. at 23.

5. 1989 Comments at 11.

6. See Airline Marketing Practices: Travel Agencies, Frequent Flyer Programs, and Computer Reservations Systems, Secretary's Task Force on Competition in the U.S. Domestic Airline Industry (Feb. 1990) at 23-24 ("Airline Marketing Practices"). This phenomenon has grown to the point that CRSs now pay $1.00-$1.50 per booking to travel agents to use CRS services rather than the other way around. Comments of Sabre, Inc. at 7.

7. Worldspan was formed in 1990 by a merger between PARS and DATAS II.

8. For example, a 1981 study by Sabre found that more than half of all Sabre sales came from the first line on the screen displayed, and almost 92% came from the first screen. 1985 Report to Congress at 12 (citing 1983 Comments).

9. 1989 Comments at 18. For example, the vendor-carrier would often load its own fare decreases first and its own fare increases last. In extreme cases, CRSs would simply delete some of the sale fares offered by the vendor's competitors. 1985 Report to Congress at 12 n.15.

10. 56 FR 12608; 57 FR 43801. Similarly, when U.S. systems attempted to penetrate foreign markets, the national carrier would sometimes refuse to participate in the U.S. system in order to make the U.S. system unattractive to local agents. See 57 FR 43800; 67 FR 69421.

11. For example, in 1988, Sabre's share of all bookings in Dallas (a hub for its owner American) was 87.3%. Apollo's booking share in Denver (a United hub) was 56.6%. System One's share in Miami (a hub for its owner Eastern) was 79.7%. 1989 Comments at 24.

12. When the sale of Worldspan is completed, the only airlines to which they will apply are Lufthansa, Air France and Iberia.

13. Steve C. Salop & John Woodbury, Economic Analysis of the NPRM Proposals at 11-12 (attached as Comments of Sabre, Inc., app. 1) ("Sabre Expert Report").

14. The estimates are based on confidential data obtained for the Department's Orbitz investigation, and aggregated to preserve confidentiality. The estimates are consistent with public data. Delta estimated last year that sales through brick and mortar travel agents comprised about 64% of its revenues, even though they comprised only 47% of all of Delta's ticket sales. National Commission to Ensure Consumer Information and Choice in the Airline Industry, (June 26, 2002) ("NCECIC Hearings") (testimony of Scott Yohe), available at http://www.astanet.com/about/govaffairs/ncecic.asp). United has estimated that sales through brick and mortar travel agents account for more than 70% of its revenue. NCECIC Hearings (testimony of Greg Taylor), available at http://www.ncecic.dot.gov/hearings.asp). America West stated that in 2002, 65% of its revenue was generated by transactions completed through CRSs. Comments of America West, Inc. at 7. Northwest reported that same percentage in its comments. Comments of Northwest Airlines, Inc. at 10. Alaska Airlines reported that it receives 56% of its revenues from travel agents. Comments of Alaska Airlines, Inc. at 5.

By contrast, airlines such as Southwest and JetBlue are much less dependent on travel agents and CRSs. It has been reported that only 20% of Southwest's passenger revenues are generated through sales by travel agents, and only 10% of Jet Blue's total revenues are generated through that channel. Upheaval in in Travel Distribution: Impact on Consumers and Travel Agents, Report to Congress and the President by the National Commission to Ensure Consumer Information and Choice in the Airline Industry at 26 (Nov. 13, 2002)("NCECIC Report"). See also Southwest Airlines, 10-K (2002) (reporting that for the year ended December 31, 2002, approximately 49% of Southwest's passenger revenues came through its own Internet site).

15. Online agencies accounted for 5 percent to 13 percent of revenue over the period, and the carriers' own websites accounted for between 2 percent and 7 percent.

16. 67 FR 69411.

17. The Department's review of confidential documents received from a recent airline investigation revealed that one major airline estimated that last year the cost of the average round trip fare booked through a traditional agency was more than $250 higher that the average round trip fare booked through an online agency.

18. See American Express Survey of Business Travel Management 2000-2001, at C-12.

19. Large purchasers of travel such as corporations and government agencies can and do change travel agencies, but the process takes time. Such purchasers typically go through a bidding process that results in the award of their business to designated agencies for a certain period of time. See, e.g., American Express Survey of Business Travel Management 2000-2001, at C-1 to C-17. Changing agencies requires a transition that is not unlike that required for changing systems, including transfer of billing records, passenger history and reservations information, travel tracking data and corporate travel policy enforcement processes.

20. DOT reported in the NPRM that airline industry respondents reported that 70% of revenues flowed through a CRS in 2001. 67 FR 69380. Sabre also reports that 70% of all airline revenue currently goes through "brick and mortar" travel agencies. Sabre Expert Report at 8.

21. Comments of the American Society of Travel Agents, Inc. at 10-11 (citing ASTA Agency Automation, 2002 at 22).

22. See Comments of the American Society of Travel Agents, Inc. at 3-4, 22-23 (calling multiple system use a "pipedream" and reporting that 94% of its survey respondents use only one system); Comments of the Large Agency Coalition at 20.

23. Delta has in place an "Online Agency Service Center" that allows Delta-accredited travel agents to book Delta's published fares by bypassing the CRSs. NCECIC Hearings (testimony of Scott Yohe at 12). American Airlines has introduced its "EveryFare" program by which American pays travel agents an allowance in exchange for agents agreeing to pass through a portion of their booking incentive payment to American. Comments of American Airlines, Inc. at 21. However, American has stated that productivity payments restrict the EveryFare program from being able to combat CRS market power. Id. at 22.

24. E.g., Comments of American Airlines, Inc. at 12 (citing 67 FR 69383); Comments of America West, Inc. at 20 ("By acceding to contractual terms in their subscriber agreements which make it impossible for airlines to gain access to agents except through a particular CRS (rather than multiple CRSs or alternative distribution systems), travel agents ensure airlines must contract with all CRSs").

25. See Comments of American Airlines, Inc. at 13 (average growth rate of between 5.2% to 6.5% from 1995-2002 although the Consumer Price Index only grew 2.4% during that same period); Sabre Expert Report at 13 (approximate annual increase of 4.8% from 1993-2002); (Galileo Expert Report at 50) (average booking fee increase from 1992 to 2002 of 3.5% for Galileo, 4.8% for Sabre, and 5.8% for Worldspan); Comments of United Air Lines, Inc. at 24 (between 1990 and 2000, booking fees increased at an annual rate of approximately 7%).

26. Costs of providing airline service in any market are lumpy. The variable costs that an airline incurs depend, in large part, on the number of flights, rather than the number of passengers flown in the market. Thus, an airline will still incur most of its costs even if it carries a few less passengers. In addition, the number of flights in a market often cannot be decreased without jeopardizing profitability. Therefore, even if an airline carries only slightly fewer passengers, it might not be able to cover its costs, and the market would no longer be profitable.

27. See Comments of Amadeus Global Travel Distribution, S.A. at 53-54 (arguing that if DOT deregulates booking fees and mandatory participation, then DOT should allow CRSs to bargain with airlines for display bias); Comments of Sabre, Inc. at 141-142. Bias already occurs in Internet travel websites. For example, Delta's agreement with Priceline prevents other carriers from offering seats on Priceline on routes to and from Delta's Atlanta hub. Scott Thorston, Northwest-Delta Feud over Priceline.com Goes Public, Atlanta Journal Constitution, Mar. 3, 2000; see also Comments of Midwest Airlines, Inc. at 12-17 (discussing bias in online travel agency websites).

28. 1989 Comments at 15-16.

29. 1991 Reply Comments at 3.

30. Data reported by Galileo's expert shows a significant decline in regional dominance. Margaret Guerin-Calvert et al., Economic Analysis of DOT Proposals to Change the CRS Rules at 18 ("Galileo Expert Report") (attached as app. A to Comments of Galileo International). However, data provided confidentially to DOJ by Sabre shows persistence of large regional shares.

31. See, e.g., United States v. Medical Mutual, 7 Trade Reg. Rep. (CCH) para. 50,846 (N.D. Ohio 1998); United States v. Delta Dental Plan, 1997-2 Trade Cas. (CCH) para. 71,860 (D.R.I. 1997); United States v. Vision Service Plan, 1996-1 Trade Cas. (CCH) para. 71,404 (D.D.C. 1996); United States v. Delta Dental Plan of Arizona, Inc., 1995-1 Trade Cas. (CCH) para. 71, 048 (D. Ariz. 1995); United States v. Primestar Partners, L.P., 1994-1 Trade Cas. (CCH) para. 70562 (S.D.N.Y 1994); United States v. General Electric Co., 1977-2 Trade Cas. (CCH) para. 61,660 (E.D. Penn. 1977).

32. See, e.g., Comments of Delta Airlines, Inc. at 40-41.

33. See, e.g., 1991 Comments at 13-17.

34. See, e.g., Comments of Large Agency Coalition at 7-14 (reporting that CRSs are offering subscribers contracts with somewhat shorter duration, less onerous booking quotas and penalties, and more attractive terms for purchasing their own equipment).

35. See, e.g., discussion and comments cited at 67 FR 69405, 57 FR 43822-43823

36. Current contracts are already complex, typically containing provisions for up-front payments, equipment rental, productivity fees, and various payments from the subscriber to the system. See Comments of the Large Agency Coalition at 7-14.

37. Delta's agreement with Priceline preventing other carriers from offering low opaque fares on Atlanta routes through that Internet outlet illustrates the kinds of bias purchases that could become fair game in the absence of regulation. Scott Thorston, Northwest-Delta Feud over Priceline.com Goes Public, Atlanta Journal Constitution, Mar. 3, 2000.

38. The DOJ defers to DOT on the consumer protection issues in this proceeding.

39. Offering preferential terms to certain customers is not in itself necessarily anticompetitive. As DOT notes, "[a] supplier's agreement to charge one firm prices that are lower than those charged the firm's competitors does not normally constitute a violation of antitrust principles." 67 FR at 69399.

40. The data does not include any information on price, and unlike many other data exchange cases, including United States v. Airline Tariff Publishing Co., 836 F.Supp. 9 (D.D.C. 1993), the primary competitive concern here does not involve the potential use of the data to facilitate oligopolistic coordination of price or output among competitors.

41. For example, the NPRM reports one commenter's concern that the data allows airlines to "obtain up to the minute analysis of competitors' sales, market share and customer information, even on a pre-flight basis," and that this information could be used by a carrier, if "so disposed," to target entrants with predatory pricing and other anticompetitive activity. 67 FR 69403.

Updated June 25, 2015