UNITED AIR LINES, INC., et al., Petitioners, British Airways PLC, et al., Intervening Petitioners, v. CIVIL AERONAUTICS BOARD, Respondent, Frontier Airlines, Inc. et al., Intervening Respondents.
Nos. 84-1877, 84-2351, 84-2719, 84-2835 and 84-3018.
United States Court of Appeals, Seventh Circuit.
Argued April 4, 1985. Decided July 3, 1985.
Edward L. Foote, Winston & Strawn, Chicago, Ill., for United Air lines.
William C. Clark, Barrett, Smith, Schapiro, Simon & Armstrong, New York City, for British Air Lines.
Thomas L. Ray, C.A.B., Washington, D.C., for C.A.B.
Frank T. Costello, Zuckert, Scoutt, Rosenberger & Johnson, Washington, D.C., for Republic Air Lines.
Before BAUER and POSNER, Circuit Judges, and MORTON, Senior District Judge.*
POSNER, Circuit Judge.
We have consolidated petitions to review two rules (containing three contested regulations) issued by the Civil Aeronautics Board concerning the computerized reservation systems that several airlines provide to travel agencies. 14 C.F.R. Parts 255, 256. Although the Board has since gone out of business as a result of the deregulation of the airline industry, the authority under which these rules were issued survived the deregulation and has been transferred to the Department of Transportation, which will be administering the rules if we uphold them. (To simplify exposition, we shall refer to the regulatory agency throughout as the Board.) The petitions raise difficult questions of administrative law.
Since the mid-1970s several major airlines have created computerized reservation systems, each consisting of computer terminals and printers in travel agents' offices plus telecommunications hook-ups to the airline's master computer that enable the travel agents to send as well as receive. The terminal in the agent's office displays information about flights, including fares, departure and arrival times, and seat availability. The travel agent can use the equipment to book a flight for the customer and print out the ticket. Although each airline that developed a computerized reservation system did so as a marketing tool for its own flights, each system contains flight information for other airlines as well--without which the system would have very limited value to the agent, since no airline serves all markets and few travel agents have more than one computerized reservation system. An airline that owns such a system will charge travel agents for the use of its system and will charge other airlines whatever they will pay to have their flights listed in the system. The charge to another airline generally is higher, the more competitive the other airline is with the airline owning the system.
There are six systems; five are owned by airlines. The largest system is American Airlines'--43 percent of all travel agencies in the United States (by revenue) use it. United's is the next largest, with 27 percent, followed by TWA with 10 and Eastern with 4. Delta and the one independent system have 2 percent each. Of all airline tickets sold, 57 percent (measured by revenues) are sold through computerized reservation systems. The Board's rules are limited to systems owned by airlines; it has no regulatory authority over the independent provider.
Besides the direct charges levied on travel agents and other airlines, airlines that own computerized reservation systems derive substantial revenue from the additional airline business that they get from "biasing" the system, that is, displaying flight information in a way that favors their own flights. The airline might for example impose a "penalty" in its computerized reservation system of 30 minutes on a competitor's flights. Suppose (to take a hypothetical case) United had a nonstop flight leaving Denver for New York at 12:25 p.m. and Frontier had an identical flight leaving at noon. If a customer phoned a travel agent who had United's computerized reservation system and said he wanted a flight from Denver to New York that was leaving around noon, and the travel agent punched this information into the computer, the computer would display United's flight first, ahead of Frontier's. The 30-minute penalty would have put Frontier's flight in second place, as if it really left after rather than before United's flight. Since the actual departure times appear on the screen, the moderately alert travel agent will not be fooled into thinking that United's flight really is more convenient for the customer than Frontier's. But many customers don't care much about the exact time of flight, and since the screen on the computer terminal has room for only eight flights, the devices (only one of which we have mentioned) that the system's owner uses to bias the display may push an otherwise more convenient flight onto the next screen, which will be displayed only if the travel agent presses a button. He may not press it if he doesn't think the customer has a strong preference for a particular departure or arrival time. Although the advantage that biasing confers on an owning airline would seem to be small, apparently it is big enough to generate millions of dollars in extra passenger revenues for such airlines.
The Board's first rule under attack in this case forbids airlines to bias their computerized reservation systems, except when the biasing is directed against certain foreign airlines, or to charge different prices to different airlines for listing in their systems (price discrimination). Other provisions of the rule are not challenged by anyone, so need not be discussed. In a separate rulemaking proceeding the Board issued a rule forbidding the deletion from the airlines' computerized reservation systems of information about connecting flights of two airlines listed under a single airline's code name. Sometimes two airlines will make an arrangement in which two connecting flights, one provided by each airline, are treated as far as possible as if they were connecting flights on the same airline (for example, by having the same or a close-by gate at the connecting terminal), and they are listed in the various computerized reservation systems under the name of the larger airline only. See Allegheny Substitute-Service Agreements, 80 C.A.B. 588, 592-94 (1979). United Air Lines announced that this was a deceptive practice and that it would delete information on such connecting flights from its computerized reservation system. The Board's rule forbids delisting, thus reserving to the Board the responsibility for preventing deception of airline passengers. Although prompted by United's announcement, the delisting rule, like the rule against biasing and price discrimination, applies to all airlines that own computerized reservation systems.
The history of these rulemaking proceedings has now to be described. In 1982 the Board, along with the Justice Department's Antitrust Division, had, at the request of a Congress besieged with complaints from travel agents and from airlines that do not own computerized reservation systems, begun to investigate biasing, price discrimination, and related practices. After completing its report to Congress, and after the Justice Department had completed its own investigation, which concluded (though the Department did not file an antitrust suit) that airlines which own computerized reservation systems use them to weaken competition from other airlines, the Board issued an advance notice of proposed rulemaking on September 9, 1983. The notice invited interested persons to submit written comments by October 17, but the deadline was later extended to November 7. Reply comments were due two weeks later, but this deadline was later extended to December 16. In March 1984 the Board issued a formal notice of proposed rulemaking with a comment and reply period that closed on May 11, and on July 27 the Board issued its final rule. The other rulemaking proceeding, which was limited to the matter of the delisting of connecting flights, proceeded similarly. In neither proceeding was there an evidentiary hearing, that is, a hearing in which witnesses testified "live" and were subject to cross-examination. United Air Lines--the principal opponent of the proposed rules--had requested that the proceedings be treated as adjudication rather than rulemaking, so that there would be such a hearing.
Only United has challenged the Board's two rules in all the respects we have mentioned--biasing, price discrimination, and delisting of certain connecting flights. Republic Airlines challenges the prohibition of price discrimination as having led to higher charges to Republic for listing its flights in computerized reservation systems, while British Airways makes a similar objection and also objects to one aspect of the exemption, in the anti-bias prohibition, regarding certain foreign airlines.
United challenges first of all the Board's power to issue rules, as distinct from cease and desist orders in adjudicative proceedings, to enforce the prohibition in section 411 of the Federal Aviation Act against "unfair or deceptive practices or unfair methods of competition in air transportation or the sale thereof." 49 U.S.C. Sec. 1381. The provision of the Act that empowers the Board to make rules, section 204(a), 49 U.S.C. Sec. 1324(a), empowers it to make only rules "pursuant to and consistent with the provisions of" the Act. Section 411 creates an adjudicative procedure for enforcing its substantive provisions. It authorizes the Board to "investigate and determine whether any air carrier ... has been or is engaged in unfair or deceptive practices or unfair methods of competition...." If the Board finds, "after notice and hearing, that such air carrier ... is engaged in such unfair or deceptive practices or unfair methods of competition, it shall order such air carrier ... to cease and desist from such" behavior. How, United asks, can a rule forbidding a deceptive practice or an unfair method of competition be pursuant to and consistent with section 411, when that section specifies cease-and-desist-order proceedings, traditionally akin to equity trials, for enforcing its prohibitions?
But a more natural reading of section 204(a) is that the Board can make only rules designed to carry out policies set forth elsewhere in the Act--in section 411, for example. Section 411 announces a policy against unfair or deceptive practices and unfair methods of competition, and while at the same time it creates an adjudicative procedure for enforcing that policy, nothing in the Act indicates that it is the exclusive procedure. In this area, moreover, a page of history is worth a volume of textual explication. The Board has been issuing rules based on section 411 since 1960. See, e.g., National Airlines, Inc., Enforcement Proceeding, 31 C.A.B. 390, 392 (1960) (overbooking); 14 C.F.R. Part 250 (same); 14 C.F.R. Part 253 (notice of passenger contract terms); 14 C.F.R. Part 254 (liability for lost luggage). These rules are well known to every air traveler and no court has ever questioned the Board's authority to issue them. Nader v. Allegheny Airlines, Inc., 426U.S.290, 294-95, 303, 306-07, 96 S.Ct. 1978, 1982-83, 1986, 1987-88, 48 L.Ed.2d 643 (1976), is illustrative of the many cases that assume the existence of such authority. A similar exercise of rulemaking authority by the Federal Trade Commission under section 5 of the Federal Trade Commission Act, 15 U.S.C. Sec. 45, on which section 411 of the Federal Aviation Act was modeled (American Airlines, Inc. v. North American Airlines, Inc., 351U.S.79, 82, 76 S.Ct. 600, 603, 100 L.Ed. 953 (1956)), was upheld in National Petroleum Refiners Ass'n v. FTC, 482F.2d672 (D.C.Cir.1973), a case that has long been regarded as authoritative, and whose validity United does not challenge. Although the language of the corresponding sections of the two statutes is not identical, none of the differences seems deliberate, let alone material.
It is true that the Trade Commission's exercise of what had been a dormant rulemaking power (other than to make rules of procedure) proved to be extremely controversial, with the result that in 1975, and again in 1980, Congress placed specific limitations on the use of the power to deal with unfair or deceptive practices. See Magnuson-Moss Act of 1975, adding 15 U.S.C. Sec. 57a; Federal Trade Commission Improvements Act of 1980, amending 15 U.S.C. Sec. 57a. But it was assumed that the Trade Commission had the power to issue substantive rules; Congress just wanted to limit the exercise of the power in particular ways. This hardly suggests either that the Civil Aeronautics Board lacks power to make rules to administer section 411 or that it must conform its rulemaking to procedures prescribed in amendments to the Federal Trade Commission Act. On the contrary, the legislative history of the 1984 "Sunset Act," which transferred the Board's rulemaking powers to the Department of Transportation upon the Board's demise, reveals that one reason why Congress refused to transfer those powers to the Federal Trade Commission instead of the Department of Transportation was "the prolonged rulemaking procedures which FTC is required to undertake under the Magnuson-Moss Act." H.R.Rep. No. 793, 98th Cong., 2d Sess. 6 (1984).
Congress, looking forward to the period after abolition of the Board, was very concerned to preserve (in the Department of Transportation) authority to enforce section 411. See Civil Aeronautics Board Sunset Act of 1984, Pub.L. 98-443, Secs. 3, 7, 98 Stat. 1703, 1706 (Oct. 4, 1984); H.R.Rep. No. 793, supra, at 3-6; H.R.Conf.Rep. 1025, 98th Cong., 2d Sess. 14-15 (1984). And Congress was well aware that the Board had used rulemaking to enforce the section. See H.R.Rep. No. 793, supra, at 4. It is too late to inquire whether, as an original matter of interpretation of sections 204(a) and 411, rulemaking can be used to prevent unfair or deceptive practices or unfair methods of competition. To hold that it cannot be so used would pull the rug out from under Congress's restructuring of airline regulation.
United has somewhat more persuasive arguments against the particular exercises of the Board's power that are challenged in this case. We find it helpful in appraising these arguments to distinguish between the Board's prohibition of biasing, on the one hand, and of price discrimination and the delisting of certain connecting flights, on the other. Although all three prohibitions were based on the Board's antitrust analysis, the prohibition of biasing was also, and as we read the Board's order alternatively, based on the Board's power to prohibit deceptive, as distinct from anticompetitive, practices. Insofar as the Board believed that a rule against biasing was necessary simply to protect travel agents and air travelers from being deceived, the Board's proceeding clearly was adequate, and we uphold the rule without hesitation.
Remember that section 411 is essentially a copy of section 5 of the Federal Trade Commission Act. Long before there were rulemaking proceedings under section 5, the Trade Commission, in a series of cease and desist order proceedings dealing with false advertising and false labeling, had established, and the courts had accepted, the principle that the tendency of a practice to deceive consumers is a question the Commission can decide on the basis of common sense and experience, without taking testimony about consumers' actual behavior. See, e.g., FTC v. Colgate-Palmolive Co., 380U.S.374, 391-92, 85 S.Ct. 1035, 1046-47, 13 L.Ed.2d 904 (1965); E.F. Drew & Co. v. FTC, 235F.2d735, 740-41 (2d Cir.1956); Zenith Radio Corp. v. FTC, 143F.2d29, 31 (7th Cir.1944); but cf. Dayco Corp. v. FTC, 362F.2d180, 186-87 (6th Cir.1966). This was how the Commission in 1964 justified the use of informal rulemaking to create a rule against the deceptive advertising and labeling of cigarettes, see Statement of Basis and Purpose of Trade Regulation Rule, 29 Fed.Reg. 8325, 8350, 8372 (July 2, 1964); and it is mainly in the area of deceptive practices that the Civil Aeronautics Board has promulgated rules under section 411. American Airlines, Inc. v. North American Airlines, Inc., supra, 351 U.S. at 86, 76 S.Ct. at 605, suggests that the Board has the same freedom to base a finding of deceptive practice on its general experience as the Trade Commission does.
A traveler buys a ticket. He thinks it guarantees him a seat. He arrives at the airport and finds that he has no seat, because the airline overbooked. He is deceived, because the airline has never disclosed its practice of overbooking, and is angry. No trial is necessary to validate the conclusion that undisclosed overbooking is deceptive and that something should be done about it. Informal rulemaking is a sensible method for dealing with problems of deception in general, and in any event one sanctified by long usage and broad acceptance and by the parallel to proceedings under the Federal Trade Commission Act.
The biasing of computerized reservation systems is a problem of deception. There is no dispute that it occurs and that it diverts business to the biasing carriers. Few air travelers know that when they call a travel agent and ask him to book them on the most convenient flight the agent will in all likelihood be using a reservation system tilted in favor of the carrier that sold him the system. Loyal and skillful travel agents no doubt correct for the tilt, but not all travel agents are either. No doubt other tactics--such as commission "overrides"--are used to achieve the same type of diversion that biasing achieves, and they have not (not yet anyway) been forbidden by the Board; but we cannot prescribe the Board's enforcement priorities. Maybe biasing can be defended as a method by which airlines that spent hundreds of millions of dollars to develop computerized reservation systems, at considerable risk of failure, can recoup their investment with a profit commensurate with the amount of the investment, the length of time it has been outstanding, and the risk of loss--though of course an airline can, and does, charge both the travel agents and other airlines for access to its computerized reservation system.
But whether or not the case against biasing is open and shut is not the question. All that matters is that the relevant considerations can be fully developed in an informal, speedy, nonoral, notice-and-comment rulemaking proceeding of the sort the Board conducted. The basic issue that the Board had to decide--the tendency of a marketing practice to deceive a significant number of consumers, whether ultimate or intermediate (here both groups are involved), see Great Lakes Airlines, Inc. v. United States Overseas Airlines, Inc., 34 C.A.B. 59, 61 n. 7 (1961)--has traditionally been treated as judgmental rather than evidentiary. The agency that in informal, speedy, nonoral, notice-and-comment rulemaking proceedings has regulated overbooking and compensation for lost luggage and many other facets of airline marketing was authorized to use the same type of proceeding to deal with biasing.
A harder question is presented by the Board's antitrust rules, which is how the prohibitions against price discrimination and against delisting certain connecting flights must be regarded; for they, unlike the prohibition of biasing, rest entirely on the Board's antitrust analysis. The questions about the Board's antitrust analysis are more procedural than substantive, but United raises questions of both sorts.
Substantively, the Board's analysis and conclusions, although far from airtight, can hardly be thought arbitrary or capricious, which is the relevant statutory standard. See 5 U.S.C. Sec. 706(2)(A). It is difficult, though as the recent success of People Express shows not impossible, to market air transportation without the help of travel agents. And it is difficult for an airline to enlist that help effectively if its flights are not included in the travel agent's computerized reservation system. It is also difficult (risky and costly) to create a computerized reservation system from scratch. A large initial investment is required; there is no assurance of success; and having your own system does not liberate you from having to get access to other systems. An airline with its own system must still buy into several others. Even the largest system, American Airlines', is used by fewer than half of the travel agencies (weighted by ticket revenues). Unless an airline limits its operations to one small region, it must, whether or not it has its own computerized reservation system, persuade several of the largest airlines to list its flights in their systems if it is to have a fair chance of competitive success. It thus is dependent for an essential facility on what may be its principal competitors; and while the airlines that own computerized reservation systems do not refuse to include their competitors' flight information, they sometimes charge substantial fees for inclusion.
If one airline owned all of the computerized reservation systems, we would have a conventional antitrust monopolization problem of the "bottleneck" variety. See, e.g., MCI Communications Corp. v. American Tel. & Tel. Co., 708F.2d1081, 1132-33 (7th Cir.1983). It would not make a critical difference whether the monopolist refused to list the flights of competing airlines altogether or merely charged high prices to list them in order to slow the growth and weaken the competition of those airlines. Although none of the airline owners of computerized reservation systems has a conventional monopoly position in the market for that service, and they are not accused of colluding, the Board found that some of them, anyway, had substantial market power. This finding, if sustained, would bring their competitive practices within the broad reach of section 411. We know from many decisions under both that section and its progenitor, section 5 of the Federal Trade Commission Act, that the Board can forbid anticompetitive practices before they become serious enough to violate the Sherman Act. See, e.g., Pan American World Airways, Inc. v. United States, 371U.S.296, 303-07, 83 S.Ct. 476, 481-84, 9 L.Ed.2d 325 (1963); Atlantic Refining Co. v. FTC, 381U.S.357, 369-70, 85 S.Ct. 1498, 1506-07, 14 L.Ed.2d 443 (1965); E.I. Du Pont de Nemours & Co. v. FTC, 729F.2d128, 136-37 (2d Cir.1984). The resemblance of the conduct in this case to what has traditionally been regarded, whether rightly or wrongly, as monopolistic behavior is close enough to avoid the Second Circuit's recent criticisms of the Trade Commission's attempt to stretch the antitrust prohibition of collusive pricing to cover parallel but independent pricing practices (such as basing-point pricing) for which there was a legitimate business justification. See id. at 140. Price discrimination, and denying a competitor access to an essential facility on equal terms (one way of describing the delisting of connecting flights of two airlines listed under the name of one), are traditional methods of illegal monopolization. See, e.g., United States v. Aluminum Co. of America, 148F.2d416, 436-38 (2d Cir.1945); MCI Communications Corp. v. American Tel. & Tel. Co., supra, 708 F.2d at 1132-33. Though no airline has a monopoly market share, that is not required by section 411.
Although no more than a modest extrapolation from conventional antitrust thinking, the Board's analysis that led it to forbid price discrimination and the deletion of connecting-flight information is open to various objections. If the owner of a computerized reservation system used the system to weaken competition from other airlines, it is a little hard to see why those airlines would not simply switch their patronage to a competing system that was not biased against them. Competition would (one might have thought) force at least some of the owners of competing systems to offer unbiased listings in order to expand the market for their systems. Even if every airline owner refused, because of the impact on its air transportation revenues, to give equal prominence to a competitor's flights, there is nothing to stop independent companies from offering a computerized reservation system with no such inhibitions--and one does. But the one independent is small. And we have said that an airline would need to be listed, if not in all systems, at least in the largest ones, which means in the systems of its principal competitors. Of course if the owner of a system charges such a high price that no competing airline will pay it, the owner is hurt. It not only loses revenue from that airline; its system will be worth less to travel agents if it contains less information. But the owner may be able to extract a high enough price from competitors to slow their growth; indeed, that may be a purpose of the high prices. As for delisting connecting flights of different airlines listed under a single airline's name, United may have hoped to gain more business by thus complicating the marketing of competing services than it would lose to travel agents by providing them with less flight information.
Analyses less rigorous than the Board's in this case have supported many decisions condemning practices thought to create a danger of foreclosing competitors from sources of supply or marketing channels, sometimes in a setting of lower market shares than in this case. See, e.g., Ford Motor Co. v. United States, 405U.S.562, 92 S.Ct. 1142, 31 L.Ed.2d 492 (1972); FTC v. Brown Shoe Co., 384U.S.316, 86 S.Ct. 1501, 16 L.Ed.2d 587 (1966). Moreover, national market share figures may understate the actual market power of the individual airlines, though perhaps not significantly. From a consumer standpoint the airline market consists of a series of discrete city pairs, each consisting of the consumer's desired origin and destination; and there are of course fewer airlines serving any particular city pair than there are airlines in the nation as a whole. However, the airplanes and other capital equipment of the airline industry are highly mobile, and now that the industry has been deregulated there no longer are legal barriers to airlines' redeploying their equipment, and swiftly too, to any city pair in which ticket prices are above marginal costs. See Bailey & Panzar, The Contestability of Airline Markets During the Transition to Deregulation, 44 Law & Contemp.Prob. 125 (Winter 1981). When the elasticity of supply is high (meaning that a slight increase in price will evoke a large increase in output, in this case of airline service between particular city pairs), potential supply must be included in estimating market shares. Thus airplanes flying between Des Moines and Salt Lake City are in the "market" for air transportation between Newark and Atlanta. Nevertheless there is some evidence that city-pair markets are not entirely meaningless--that high concentration in some of those markets leads to high prices. See Graham, Kaplan & Sibley, Efficiency and Competition in the Airline Industry, 14 Bell J. Econ. 118, 128-37 (1983).
All this assumes of course that the structure of the market for air transportation provides a useful guide to competitive conditions in what is, at least analytically, the distinct market for computerized reservations (or perhaps for all reservations). Obviously there is some relationship between the markets. The record shows that in cities like Denver, where United accounts for a very large fraction of departing and arriving flights, United is able to persuade most travel agents (72 percent, weighted by revenue) to subscribe to its computerized reservation system. This in turn makes a listing in that system a must for airlines that want to compete in Denver, and so enables United to charge a high price for a listing, thereby impeding (no one knows by how much) the growth of competing airlines in the Denver market. Such a strategy may well be short-sighted; it may accelerate the rise of competing computerized reservation systems; but we cannot call the Board irrational for thinking, with some factual basis, that United was pursuing such a strategy. There might seem to be grave doubt whether the Board could reach the same conclusion about airlines that have only modest shares of the market for computerized reservation systems, such as Delta with its 2 percent. But no other airline besides United is challenging the Board's findings with respect to market power.
Since the Board considered the various arguments bearing on its antitrust analysis and reached a conclusion that is plausible, if perhaps not compelling, its rules cannot be set aside as arbitrary and capricious. But a serious question remains of whether the Board gave United Air Lines all the procedural safeguards that United was entitled to.
The Administrative Procedure Act makes clear that notice of the scope and general thrust of the proposed rule, and an opportunity to submit written comments, are all the procedure that an agency engaged in "informal rulemaking" is required to provide. See 5 U.S.C. Sec. 553(c); United States v. Allegheny-Ludlum Steel Corp., 406U.S.742, 758, 92 S.Ct. 1941, 1951, 32 L.Ed.2d 453 (1972). An evidentiary hearing is not required. See, e.g., United States v. Florida East Coast Ry., 410U.S.224, 240-46, 93 S.Ct. 810, 818-21, 35 L.Ed.2d 223 (1973); FCC v. WJR, The Goodwill Station, Inc., 337U.S.265, 281-82, 69 S.Ct. 1097, 1106-07, 93 L.Ed. 1353 (1949). A movement in the courts of appeals, particularly the Court of Appeals for the District of Columbia Circuit, to require where appropriate additional procedures in informal rulemaking beyond those required by the Administrative Procedure Act (see, e.g., cases cited in O'Donnell v. Shaffer, 491F.2d59, 62 (D.C.Cir.1974)), was cut short by the Supreme Court's holding and blunt language in Vermont Yankee Nuclear Power Corp. v. Natural Resources Defense Council, Inc., 435U.S.519, 543-48, 98 S.Ct. 1197, 1211-14, 55 L.Ed.2d 460 (1978).
But an agency is not allowed to conduct what is in fact an adjudicative proceeding without giving the affected parties an opportunity for an evidentiary hearing, just by calling the proceeding informal rulemaking. That would be too facile an evasion of the procedures that the Administrative Procedure Act requires for adjudicative proceedings such as (we may assume) adjudicative proceedings under section 411. See especially 5 U.S.C. Sec. 556(d). Hence if the Board here had, without bothering with an evidentiary hearing though requested by United Air Lines to conduct one, found that United had violated section 411, and issued a rule restricting United's marketing of its computerized reservation system, the rule would be invalid. But this is not what the Board did--quite. Admittedly it came close.
Although we can set aside, as quite unfounded, United Air Lines' expression of concern that these rules might be given collateral estoppel effect in private antitrust suits that are pending against United in another circuit, see Delamater v. Schweiker, 721F.2d50, 53 (2d Cir.1983) (per curiam); cf. United States v. Utah Construction & Mining Co., 384U.S.394, 421-22, 86 S.Ct. 1545, 1559-60, 16 L.Ed.2d 642 (1966), the parts of the Board's statements explaining its prohibitions against price discrimination and the delisting of connecting flights do read like excerpts from an antitrust opinion issued at the conclusion of an evidentiary hearing. The Board finds that United and the other airline owners of computerized reservation systems have market power, and while studiously, perhaps too studiously, disclaiming any conclusion that a particular airline has abused that power and discriminated against competing airlines, leaves the reader in no doubt that this is precisely what it believes. In its notice of proposed rulemaking on biasing and price discrimination, the Board, while on the one hand saying that it has "made no accusations of illegal conduct," 49 Fed.Reg. at 11660, on the other hand states that "CRS carriers [i.e., the airlines that own computerized reservation systems] are engaged in unfair methods of competition" and in "conduct producing competitive harms that may be analogous to conduct that would be labeled an abuse of monopoly power under section 2 of the Sherman Act," are "impeding [travel] agents' ability to use other systems in conjunction with their own," "clearly restrict the output of their product," "have the ability and strong incentives to exercise this [market] power to reduce competition in air transportation, ... [and] some CRS owners may in fact be exercising this power today." Id. at 11646-47, 11655, 11657. In issuing the final rule the Board noted "example after example in which the major [CRS] vendors have: dictated fees that were discriminatory, not based on cost, not affected by the prices charged by other vendors and unchanged by pleas that they were excessive; forced other carriers to pay substantial net ticketing fees as a condition of CRS access; steadily increased the degree of bias in their system, with no adverse effect on their CRS market penetration; and reduced or withheld service on a selective basis to discipline air transportation competitors. Both the CRS vendors and their air carrier customers are acutely aware of the vendors' market power, and they act accordingly." 49 Fed.Reg. at 32543 (citations omitted); see also id. at 32547.
These findings were based primarily on the comments submitted by the Justice Department, summarizing the results of its investigation into allegations that airlines that own computerized reservation systems were violating the antitrust laws. Although the airlines had and took several months to prepare written rebuttals to the Justice Department's report, they did not have the full opportunity normally given persons accused of antitrust violations, including the "incipient" violations punishable under statutes such as section 5 of the Federal Trade Commission Act and section 411 of the Federal Aviation Act, to test the truthfulness of the allegations. They had no opportunity to use subpoenas or other methods of pretrial discovery, or to cross-examine the witnesses on whose depositions the Department's comments were based, or to explore the background, authenticity, and meaning of various corporate documents on which those comments relied. For example, one of the juicier tidbits in the Justice Department's comments is a letter from the president of American Airlines to Continental Airlines turning down a request to buy computer-reservation software on the ground that Continental had begun competing with American in air fares. United Air Lines could not depose the president of American Airlines, or officers of Continental, in an attempt to rebut the damaging implication of the letter regarding the anticompetitive uses to which American Airlines had allegedly put its computerized reservations capability.
Also, despite much talk by the Board about "market power," the only effort made to establish this elusive force beyond the market-share statistics mentioned earlier in this opinion consisted of evidence that the owners of computerized reservation systems had engaged in price discrimination and other practices symptomatic of monopoly or market power. This illustrates the important role that findings of abuses played in the Board's analysis, but the broader point is that United Air Lines did not have the opportunity it would have had in an antitrust trial to show that the "market power" that it and other owners of computerized reservation systems appear to have really is trivial, and was not (could not have been) abused. Maybe United could have shown that the deregulation of the airline industry, and resulting entry of new firms into the industry and old airlines into new (to them) city pairs, has created attractive opportunities for non-airline vendors of computerized reservation systems, opportunities that will place tight limits on the power of the airlines that own such systems to oppress their competitors. Although such evidence would also weaken the case for prohibiting biasing, we do not think that is important. The Board can forbid a deceptive practice without finding that it is widespread or can substantially distort competition, and we saw that the Board's alternative basis for the prohibition was independent of the market-wide effects, if any, of the practice.
On the question whether price discrimination may in fact be a reasonable method of pricing so unconventional a service as a computerized reservation system--especially in the era, about to begin, when biasing, another method of recouping the investment in creating such a system, will be forbidden--United did introduce an economist's statement that if an airline were forced to charge uniform prices to all airlines wanting to be listed in its computerized reservation system the uniform price might have to be so high that some airlines could not afford to pay it. Everyone might be worse off: some airlines might not be served at all and the providers' net revenues might be lower too. The Board rejected the economist's statement as based on incorrect assumptions. And in fact the statement is too remote from the particulars of computerized reservations systems to carry much weight. But the economist did not have as much time to immerse himself in those particulars and present a convincing statement as he would have had in an adjudicative proceeding.
Thus if we thought ourselves free to do so we might hold that the Board had used rulemaking procedures to make adjudicative factfindings; in the accepted jargon, that it had crossed the line that separates "legislative" facts--the kind that can be found reliably without an evidentiary hearing--and "adjudicative" facts, which cannot be. Cf. Alaska Airlines, Inc. v. CAB, 545F.2d194, 200-01 (D.C.Cir.1976). The distinction is a well established one, though variously expressed. See, e.g., United States v. Florida East Coast Ry., supra, 410 U.S. at 244-46, 93 S.Ct. at 820-21; Broz v. Schweiker, 677F.2d1351, 1357 (11th Cir.1982), modified on other grounds under the name of Broz v. Heckler, 721F.2d1297 (11th Cir.1983); Association of National Advertisers, Inc. v. FTC, 627F.2d1151, 1161-63 (D.C.Cir.1979); Independent Bankers Ass'n v. Board of Governors, 516F.2d1206, 1217 (D.C.Cir.1975); 2 Davis, Administrative Law Treatise Sec. 12:3 (2d ed.1979). It survived Vermont Yankee. See 435 U.S. at 542, 98 S.Ct. at 1210. It is, as we have suggested, implicit in 5 U.S.C. Sec. 556(d), setting forth procedures for certain adjudications before federal agencies. As Professor (now Judge) Scalia has written, even after Vermont Yankee the "judicial importation of cross-examination requirements, where an ancient and well-established adjudicatory issue has been kidnapped into rulemaking," may still be proper. Vermont Yankee: The APA, the D.C. Circuit, and the Supreme Court, 1978 Sup.Ct.Rev. 345, 395.
However, the distinction between legislative and adjudicative facts is not easy to apply, or facts such as market power and price discrimination--abstractions far removed from the simple observations of the eyewitness to a car crash--easy to classify with respect to it. In support of the conclusion that they are adjudicative facts, at least when they are contested, as they are here, and findings on them used to support a far-reaching regulation of a substantial market, we could point out that the Federal Trade Commission; while active in promulgating rules to prohibit deceptive and otherwise fraudulent practices, has promulgated only one rule (and that back in 1967) to prohibit an antitrust violation. And that rule was of the simplest kind; it forbade the discriminatory provision of advertising allowances. See section 2(d) of the Clayton Act, as amended by the Robinson-Patman Act, 15 U.S.C. Sec. 13(d); 16 C.F.R. Part 412 (Trade Regulation Rule Against Discriminatory Practices in Men's and Boys' Tailored Clothing Industry); 16 C.F.R. Ch. 1, at pp. 4-5 (table of contents of Subchapter D, Trade Regulation Rules). Although the Commission has long been urged to do more in the way of antitrust rulemaking, see, e.g., Elman, Rulemaking Procedures in the FTC's Enforcement of the Merger Law, 78 Harv.L.Rev. 385 (1964), the urgings have fallen largely on deaf ears. And apart from the rules challenged in this proceeding and the tariff regulation at issue in National Small Shipments Traffic Conference, Inc. v. CAB, 618F.2d819 (D.C.Cir.1980), the Civil Aeronautics Board has only one rule dealing with an antitrust type of subject--a rule implementing the statutory regulation of interlocking directorates and similar interlocking relationships. See 14 C.F.R. Part 251 (as modified by Part 287). The use of rulemaking to enforce antitrust principles remains exceptional, for reasons well illustrated by this case.
But unfortunately for United Air Lines, the weight of authority, much of it in the Supreme Court and therefore beyond our power to reexamine, is overwhelming against forcing an administrative agency to hold an evidentiary hearing to resolve disputed questions of antitrust fact, though we can assume there would be an exception for a fact that could not rationally be found without providing an opportunity for cross-examination or some other trial-type procedural safeguard. Subject to this qualification, and provided that the agency issues what is genuinely a rule, which is to say a prospective regulation of general applicability, it is free to base the rule on the kind of findings normally made in an adjudicative proceeding, even if it conducts no evidentiary hearing. An important, perhaps controlling, precedent is United States v. Florida East Coast Ry., supra, 410 U.S. at 241-46, 93 S.Ct. at 819-21, where the Supreme Court held that the ICC could use informal rulemaking to establish incentive per diem charges for boxcars. "[T]he incentive payments proposed by the Commission in its tentative order, and later adopted in its final order, were applicable across the board to all of the common carriers by railroad subject to the Interstate Commerce Act. No effort was made to single out any particular railroad for special consideration based on its own peculiar circumstances.... Though the Commission obviously relied on factual inferences as a basis for its order, the source of these factual inferences was apparent to anyone who read the order.... The factual inferences were used in the formulation of a basically legislative-type judgment, for prospective application only, rather than in adjudicating a particular set of disputed facts." Id. at 245-46, 93 S.Ct. at 821. As explained by Judge Friendly, "The opinion [in Florida East Coast] seems to say that 'hearing' provisions in regulatory statutes, which had long been regarded as requiring trial-type hearings, have been modified by the Administrative Procedure Act so that nothing more than notice and written comment is required if the action falls within the APA's expansive definition of rulemaking, and implicitly, of course, that this comports with due process. The definition of rulemaking is exceedingly broad, about the only limitation being that a rule can have only future effect." "Some Kind of Hearing ", 123 U.Pa.L.Rev. 1267, 1307 (1975) (footnotes omitted). See also 1 Davis, Administrative Law Treatise Sec. 6:23 (2d ed. 1978).
The rules in this case were applied to all providers (within the Board's regulatory domain) of computerized reservation systems, and thus the findings of fact on which they were based "were used in the formulation of a basically legislative-type judgment, for prospective application only...." Moreover, no effort to resolve a "set of disputed facts" was made. Although past abuses were used illustratively, to show that the Board was not just chasing phantoms, it did not try to determine to what extent any airline had actually violated section 411. It did not rely on the juicy tidbits of which an antitrust jury might have made much; these are not mentioned, and we are not authorized to assume that the Board relied covertly on evidence that would have been unreliable if not tested by cross-examination or supplemented by the fruits of pretrial discovery. And the Board did not come down on just one or two firms; its rules apply to the whole industry of airline-provided computerized reservations systems, the big and the small providers alike. This distinguishes American Bancorporation, Inc. v. Board of Governors of Fed. Reserve System, 509F.2d29, 37 (8th Cir.1974), which involved a single acquisition by a bank holding company. Cf. SEC v. Frank, 388F.2d486, 491-92 (2d Cir.1968) (Friendly, J.). When government acts across the board, the danger of official arbitrariness is reduced. See Philly's v. Byrne, 732F.2d87, 92 (7th Cir.1984).
Another example of the power that agencies have to make antitrust-like regulations by means of informal rulemaking is the Federal Communications Commission's chain broadcasting rules (limiting network ownership of radio stations) upheld in National Broadcasting Co. v. United States, 319U.S.190, 63 S.Ct. 997, 87 L.Ed. 1344 (1943). The Commission had based these rules, which were not antitrust rules but had an antitrust flavor, on an investigation that had revealed abuses by the three networks of the economic power created by their ownership of radio stations. The Supreme Court said that its "duty is at an end when we find that the action of the Commission was based upon findings supported by evidence." Id. at 224, 63 S.Ct. at 1013. No one seemed troubled by the fact that the evidentiary basis of the findings had been laid in a rulemaking rather than an adjudicative proceeding, though the regulations that resulted were farther-reaching than anything involved in this case. Admittedly the case preceded the enactment of the Administrative Procedure Act, but many years later the Supreme Court upheld a similar regulation, limiting newspaper ownership by broadcasters, even though the regulation required some divestiture and the rulemaking proceeding did not conclusively establish the factual premises of the Commission's action. See FCC v. National Citizens Comm. for Broadcasting, 436U.S.775, 796-97, 98 S.Ct. 2096, 2112-13, 56 L.Ed.2d 697 (1978); see also United States v. Storer Broadcasting Co., 351U.S.192, 202-05, 76 S.Ct. 763, 770-72, 100 L.Ed. 1081 (1956); Recording Industry Ass'n of America v. Copyright Royalty Tribunal, 662F.2d1, 8-9 (D.C.Cir.1981).
Agencies, without having to conduct an evidentiary hearing, have been allowed to decide such antitrust questions as whether a particular firm or group of firms has or is abusing or is likely to abuse market power, which are the very questions involved in the present case. An example is the FCC's rule on the provision of customer equipment through separate subsidiaries of the former Bell operating companies, which we upheld last summer. See Illinois Bell Tel. Co. v. FCC, 740F.2d465 (7th Cir.1984). Like the rules in this case--more clearly than the chain broadcasting rules--the separate-subsidiary rule was based on antitrust analysis, and was as here a rule made after notice-and-comment rulemaking. The cases in which reviewing courts have allowed agencies, over objection, to make judgments on market power and related competitive issues without adjudicative hearings include Belenke v. SEC, 606F.2d193, 198-99 (7th Cir.1979); Brae Corp. v. United States, 740F.2d1023, 1038-40 (D.C.Cir.1984) (per curiam); United States v. FCC, 652F.2d72, 91-96 (D.C.Cir.1980) (en banc) ("expert witnesses and cross-examination would not add appreciably to [the FCC's] already-expert knowledge of the field," id. at 94); National Small Shipments Traffic Conference, Inc. v. CAB, supra, 618 F.2d at 828-31 ("The Air Carriers argue that the Board's decision must be supported by substantial evidence in the record. However, in concluding that abolishing the tariff filing requirement will increase price competition without depriving shippers of adequate rate information or precluding enforcement of the statutory ban on discriminatory rates, the Board made a determination that was 'primarily of a judgmental or predictive nature,' " id. at 829); Bell Tel. Co. of Pennsylvania v. FCC, 503F.2d1250, 1268 (3d Cir.1974); Phillips Petroleum Co. v. FPC, 475F.2d842, 844 (10th Cir.1973); cf. Trans World Airlines, Inc. v. CAB, 637F.2d62, 71 (2d Cir.1980).
None of these cases suggests that an agency has carte blanche to resolve antitrust issues without an evidentiary hearing. United States v. FCC, supra, 652 F.2d at 91, expressly disclaims such a proposition--but goes on to uphold the agency's action. For only "if a reviewing court finds that the procedures followed by the agency in adopting a rule have not produced a body of evidence enabling it to" uphold the rule may it remand for further proceedings under different procedures. Friendly, supra, 123 U.Pa.L.Rev. at 1313. It will rarely be possible to say, and it is not possible to say here, that merely because the agency denied cross-examination and the other procedural safeguards of the adjudicative process, a rulemaking proceeding in which interested persons had ample notice of the agency's proposals and sufficient if not generous time to prepare and submit comprehensive written comments provided so few procedural safeguards that the resulting rules lack an adequate factual basis.
More than authority is against United; though we are sympathetic to its position, we appreciate the arguments, as well as the cases, against it. The biggest practical difference between adjudicative and rulemaking procedure is that cross-examination is available in the former but not--not generally anyway and not here--in the latter. But cross-examination is perhaps not a terribly useful tool for extracting the truth about what are at bottom complex economic phenomena. Antitrust trials, whether judicial or administrative, have long been criticized for their inordinate length, cost, and complexity. See, e.g., Bok, Section 7 of the Clayton Act and the Merging of Law and Economics, 74 Harv.L.Rev. 226, 258-73 (1960). It is a fair question whether the additional scrutiny of evidence in an antitrust trial elucidates more than it confuses the issues, and hence whether the Board would really have come up with substantively sounder rules if it had given United Air Lines all the procedural rope that United sought. Moreover, the Board's rules for adjudicative proceedings are skimpy; perhaps in practice such proceedings differ little from rulemaking proceedings. See 14 C.F.R. Secs. 302.200 et seq. Although the comment period was fairly compressed, the airlines had been on notice since 1982, when the Board and the Justice Department began their investigations, that they might very well have to prepare a defense to a charge of anticompetitive practices. They could have put together a better economic defense than they attempted; they had two years. Furthermore, the fact that the Board is (more precisely, until its recent abolition was) engaged full-time in the regulation of the airline industry could rationally persuade it that it could dispense with procedural safeguards designed primarily to guide the lay judges we call jurors and the generalist judges who decide bench trials. Thus we need not decide whether the Federal Trade Commission, whose antitrust powers are not limited to a particular industry but range over much of the economy, is equally free to dispense with the conventional safeguards.
Another point in the Board's favor, and another ground for a narrow interpretation of our holding, is the legislative history of the Sunset Act, discussed earlier in this opinion. That history indicates not only that Congress approves of expeditious rulemaking proceedings to enforce section 411 but that Congress was aware of the present proceeding. The House Report mentions it specifically. See H.R.Rep. No. 793, supra, at 4-5. That report by the way was published after the Board had issued the rulemaking notices that contain its essential findings, and though it would be unrealistic to think that Congress implicitly approved the denial of an evidentiary hearing in this matter when it passed the Sunset Act, the legislative history at least makes clear that the use of informal rulemaking to deal with competitive problems in the airline industry is not so outre as United Air Lines (alone among the members of the industry) contends.
In the face of all this, United, to show entitlement to an evidentiary hearing in this matter, would have to show with some particularity that cross-examination or other procedures that only such a hearing would enable was necessary to resolve specific factual disputes that are critical to the soundness of the Board's rules. It has not attempted to show this. The rules are judgmental and are based primarily on the Board's accumulated knowledge of the airline industry and on inferences from general economic data that either were undisputed or could be probed adequately in a purely documentary proceeding.
Thus the rules must be upheld in their entirety against United's challenge. Republic Airlines' and British Airways' challenges to the price discrimination feature of the Board's rules do not require extended consideration. The basis of these challenges is that eliminating discrimination hurts those who pay the lower of the discriminatory price. This is true by definition but hardly shows that the elimination of the practice is arbitrary or capricious.
That leaves for discussion only British Airways' challenge to the foreign-carrier exclusion in the anti-bias rule. 14 C.F.R. Sec. 255.9(6). This challenge is a very narrow one. The Board allowed the airlines subject to the rule, all of which are U.S. airlines, to employ bias against any foreign airline that provides a computerized reservation system to travel agents overseas that is biased against U.S. airlines. British Airways does not question the propriety of this provision for self-help but complains that under the Board's rule the right of self-help may not be used--by British Airways, for example--against a domestic, as distinct from foreign, carrier that uses a biased computerized reservation system overseas. Suppose, for example, that TWA used such a system in Europe. The Board's rule would not forbid it to do so; the rule is limited to domestic service. And the foreign-carrier exclusion would not permit United Air Lines, say, to bias its computerized reservation system in the U.S. against TWA. TWA is not a foreign carrier, and the exclusion is for use against foreign carriers. British Airways' interest in all this, as near as we can figure out from its excessively cryptic briefing of the issue, comes from the fact that it is allied with United, and would therefore like United to be able to bias its computerized reservation system in British Airways' favor in the U.S., which United is not allowed to do under the Board's rule even if TWA is engaged in biasing overseas.
This is too minor and speculative a quibble to warrant judicial intervention. British Airways' challenge to the Board's rule came very late, and though it was nevertheless entertained by the Board, British Airways failed to put in evidence showing that U.S. carriers make extensive use of biased computerized reservation systems overseas or that a good way to prevent such bias would be to allow other U.S. carriers to retaliate against them in the U.S.--to the detriment of domestic travelers. We cannot say that the Board was arbitrary or capricious in rejecting British Airways' plea for a further exception to the anti-bias rule.