Rothery Storage & Van Co. v. Atlas Van Lines, Inc.

792 F.2d 210, 253 U.S. App. D.C. 142, 1986 U.S. App. LEXIS 25296

• Court: Court of Appeals for the D.C. Circuit
• Decided: June 3, 1986
• Docket: 84-5845
• Status: Published

Opinions

BORK, Circuit Judge:

Appellants, plaintiffs below, seek review of the district court’s decision dismissing their antitrust action against Atlas Van Lines, Inc. (“Atlas”). See Rothery Storage & Van Co. v. Atlas Van Lines, Inc., 597 F.Supp. 217 (D.D.C.1984). Appellants are five present and three former agents of Atlas. For convenience, we will frequently refer to them by the name of the first-named appellant, Rothery Storage & Van Co. (“Rothery”). Rothery claims that Atlas and several of the carrier agents affiliated with Atlas adopted a policy constituting a “group boycott” in violation of section 1 of the Sherman Act, 15 U.S.C. § 1 (1982), which prohibits “[e]very contract, combination ... or conspiracy ... in restraint of trade.” The trial court granted Atlas’ motion for summary judgment on several alternative grounds. See infra pp. 213-14. Because we find that Atlas’ policy is designed to make the van line more efficient rather than to decrease the output of its services and raise rates, we affirm.

I.

Atlas operates as a nationwide common carrier of used household goods under authority granted by the Interstate Commerce Commission. It contracts to provide moving services to individuals and to businesses transferring employees. Like most national moving companies, Atlas exercises its interstate authority by employing independent moving companies throughout the country as its agents. These companies execute a standard agency contract with Atlas, agreeing to adhere, when making shipments on Atlas’ authority, to such things as standard operating procedures, maintenance and painting specifications, and uniform rates. Typically, such an agreement will contain a provision barring an agent affiliated with a particular van line from dealing with any other line. The agency agreement is supplemented by Atlas’ bylaws, rules, and regulations governing the agents’ interstate operations.

Some of these independent moving companies, the “non-carrier agents,” have no interstate authority of their own and can move goods interstate only on Atlas’ authority. Until recently, other companies, the “carrier agents,” possessed their own interstate authority and could move goods to the extent of that independent authority as principals for their own accounts. Both types of agent may engage in intrastate carriage without Atlas’ permission or governance. A carrier agent, however, could act in interstate commerce both as an agent of the van line it serves and as a competitor of that van line. The carrier agents could, and some did, use Atlas equipment, training, and the like for interstate carriage under their own authorities and pay Atlas nothing.

A van line and its agents constitute an enterprise on a scale not easily obtainable by a single carrier. Atlas, which is the sixth largest van line in the nation, provides a network of 490 agents capable of carrying household goods between any two points in the nation. Atlas coordinates and supports the agents’ operations. The use of agents spares a van line the necessity of obtaining enormous amounts of capital to perform the same services and, quite possibly, avoids diseconomies of scale, i.e., the inefficiencies of a single management large and complex enough to perform all the functions that are now divided between the van line and its agents. The agents find customers and do the packing, loading, hauling, and storage. Atlas sets the rates, dispatches shipments, chooses routes, arranges backhauls so the agent’s truck need not return empty, arranges services at the origin and destination of shipments, collects all revenues and pays the agents, establishes uniform rules for the appearance and quality of equipment, trains salespeople and drivers, purchases and finances equipment for use by the agents, and maintains insurance on all shipments made under Atlas’ authority. In addition, Atlas conducts national advertising and promotional forums. With the assistance of agents, it handles customer claims. In short, Atlas, and its agents make up an enterprise or firm intergrated by contracts, one which is indistinguishable in economic analysis from a complex partnership.

The ability of the carrier agents to exercise their independent authority traditionally has been governed by “pooling agreements” that dictate the business relationship between a van line and a carrier agent affiliated with it. The van line could, of course, set the rates at which its agents carried goods on its authority. ICC regulations required that carrier agents use the same rates as their van line principal when carrying shipments under independent authority. Because the relationship between a carrier agent and a van line constitutes an agreement between competitors, Congress provided antitrust immunity for any such relationship governed by a pooling agreement approved by the ICC. See 49 U.S.C. §§ 11341-11342 (1982). In 49 U.S.C. § 10934(d) (1982), Congress also allowed agents to sit on the boards of their van lines without antitrust liability.

The deregulation of the moving industry, beginning in 1979, produced changes that had a profound impact on the relationship between van lines and their agents. Prior to the regulatory changes, independent moving companies had little ability to obtain their own interstate transportation authority. The ICC’s Policy Statement on Motor Carrier Regulation, 44 Fed.Reg. 60,-296 (1979), and the Motor Carrier Act of 1980, Pub.L.No. 96-296, 94 Stat. 793, greatly increased the ability of common carriers to obtain interstate moving authority. In 1981, moreover, the ICC repealed its requirement that carrier agents charge the same rate for agency shipments and shipments carried on their own accounts. See North American Van Lines, Inc. v. ICC, 666 F.2d 1087, 1094-96 (7th Cir.1981). Thus, agents could obtain interstate authority and could cut prices to attract business for their own accounts that otherwise might have constituted agency shipments for the van line’s account.

This increased potential for the diversion of interstate business to its carrier agents posed two potential problems for Atlas. Each of these problems is a version of what has been called the “free ride.” A free ride occurs when one party to an arrangement reaps benefits for which another party pays, though that transfer of wealth is not part of the agreement between them. The free ride can become a serious problem for a partnership or joint venture because the party that provides capital and services without receiving compensation has a strong incentive to provide less, thus rendering the common enterprise less effective. The first problem occurs because, by statute, a van line incurs strict liability for acts of its agents exercising “actual or apparent authority.” 49 U.S.C. § 10934(a) (1982). Thus, an increase of shipments made on the agents’ independent authority, but using Atlas’ equipment, uniforms, and services would create the risk of increased liability for Atlas although Atlas received no revenue from those shipments. Second, because carrier agents could utilize Atlas services and equipment on non-Atlas interstate shipments, the possible increase of such shipments meant that Atlas might make large outlays for which it received no return. We return to the free-ride problem in Part IV of this opinion.

To meet these problems, Atlas could have amended its pooling agreement to redefine the terms on which it allowed its carrier agents to compete with the principal company. Had Atlas chosen this course and obtained ICC approval of its amended pooling agreement, the new agreement would have enjoyed antitrust immunity under 49 U.S.C. §§ 11341-11342 (1982). Instead, on February 11, 1982, Atlas announced that it would exercise its statutory right to cancel its pooling agreement and would terminate the agency contract of any affiliated company that persisted in handling interstate carriage on its own account as well as for Atlas. Under the new policy, any carrier agent already affiliated with Atlas could continue to exercise independent interstate authority only by transferring its independent interstate authority to a separate corporation with a new name. These new entities could not use the facilities or services of Atlas or any of its affiliates.

II.

Because Atlas and its affiliates refuse to deal with any carrier agent that does not comply, several Atlas carrier agents, appellants here, charged that Atlas’ new policy constitutes a “group boycott.” They filed this action, and after the completion of discovery on the issue of liability, both sides filed cross motions for summary judgment.

The district court granted summary judgment to Atlas on alternative grounds. First, relying on Copperweld Corp. v. Independence Tube Corp., 467 U.S. 752, 104 S.Ct. 2731, 2741 n. 15, 81 L.Ed.2d 628 (1984), the court held that, because the challenged policy was promulgated by Atlas’ board of directors, the plurality of actors essential to a finding of conspiracy did not exist. See 597 F.Supp. at 225. The court rejected Rothery’s argument that Copperweld did not apply because some of the directors represented carrier agents and, therefore, might have a separate interest in adopting the new policy. See id. at 227-29; see also Greenville Publishing Co. v. Daily Reflector, Inc., 496 F.2d 391, 399 (4th Cir.1974) (stating that when a corporation’s officers have an “independent personal stake in achieving the corporation’s illegal objective,” it constitutes an exception to the rule that a corporation cannot conspire with its own officers) (citations omitted).

Second, the court rested on 49 U.S.C. § 10934(d)(4)(1982), which provides antitrust immunity for “discussions or agreements between a motor common carrier ... and its agents ... related solely to ... ownership of a motor common carrier ... by an agent or membership on the board of directors of any such motor common carrier by an agent.” See 597 F.Supp. at 226. Because the “[discussions, voting, and agreement” leading up to the challenged policy involves the terms under which Atlas agents themselves can own independent motor common carriers in interstate service, the court found that policy immune from antitrust liability. See id. at 227-28.

The court’s third reason went to the merits of the antitrust claim. In the absence of “ ‘[considerable’ and ‘unambiguous’ ” experience with relationships between carrier agents and van lines and the impact of deregulation on those relationships, the court declined to treat the Atlas policy as illegal per se. See 597 F.Supp. at 231. Shifting to a rule-of-reason analysis, the court held that Atlas acted reasonably in adopting a policy that ended the carrier agents’ ability to benefit from the “diversion” of Atlas’ “business infrastructure” to shipments made on their own accounts. See id. at 233. The court found it significant that Atlas adopted the less restrictive alternative of forcing affiliated carrier agents to exercise their independent authority through a separate company, rather than prohibiting any exercise of such authority by carrier agents. See id. at 234. The court also noted that any reduction in competition between Atlas and its carrier agents had to be weighed against procompetitive effects the new policy would have on inter-brand competition. See id. at 235.

While we do not agree that the challenged arrangement lacks the elements of a horizontal agreement, we uphold the trial judge’s conclusion that Atlas’ new policy does not offend the antitrust laws. The challenged restraint is ancillary to the economic integration of Atlas and its agents so that the rule of per se illegality does not apply. Neither are the other tests of the rule of reason offended since Atlas’ market share is far too small for the restraint to threaten competition or to have been intended to do so. Because the reasonableness of the restraint is so clear, we do not reach the question of whether the challenged arrangement falls within the statutory exemption from antitrust liability contained in 49 U.B.C. § 10934(d) (1982).

III.

Before turning to considerations we think dispositive, we deal with arguments that Atlas and Rothery each consider decisive. Atlas contends that there is no agreement between actual or potential competitors and hence no violation of section 1 of the Sherman Act. Rothery argues that Atlas’ policy constitutes a boycott that is per se illegal. We think neither argument well-founded.

A.

Section 1 of the Sherman Act condemns only those restraints of trade achieved by contracts, combinations, or conspiracies. Thus, only agreements between legally separate entities are covered. The argument that the Atlas board of directors was incapable of conspiring for Sherman Act purposes depends on the legal principle, laid down in Copperweld Corp. v. Independence Tube Corp., 467 U.S. 725, 104 S.Ct. 2731, 81 L.Ed.2d 628 (1984), that an agreement within a single enterprise is not an agreement contemplated by the Act. That principle is inapplicable here.

When the Atlas policy challenged in this case went into effect, every agent in the system was an actual or potential competitor of Atlas. The carrier agents were actual competitors and the non-carrier agents, because of the ICC’s increased willingness to grant interstate moving authority, were potential competitors. Every carrier that stayed in the Atlas network adhered to a policy of ending or lessening its competition with Atlas (by abandoning its interstate authority or transferring that authority to a separate company with a new trade name) or of not entering into full competition with Atlas (by not obtaining interstate authority). Agents required to ship on Atlas' interstate authority must, of course, abide by Atlas’ rates. Thus, all of these legally separate corporations agreed to a policy that restricted competition.

The Supreme Court dealt with an analogous arrangement in National Collegiate Athletic Association v. Board of Regents, 468 U.S. 85, 104 S.Ct. 2948, 82 L.Ed.2d 70 (1984) (“NCAA ”). In order to remain members of the NCAA in good standing, universities had to adhere to the NCAA’s policy with respect to the televising of football games. The Court said:

By participating in an association which prevents member institutions from competing against each other on the basis of price or kind of television rights that can be offered to broadcasters, the NCAA member institutions have created a horizontal restraint — an agreement among competitors on the way in which they will compete with one another.

Id. at 2959 (footnote omitted). That reasoning controls here.

If it is deemed important, it may be noted that the Atlas Board of Directors consisted of actual or potential competitors of Atlas and that is also sufficient to take this case out of the Copperweld rule. See United States v. Sealy Corp., 388 U.S. 350, 87 S.Ct. 1847, 18 L.Ed.2d 1238 (1967); see also Topeo Associates, Inc. v. United States, 405 U.S. 596, 92 S.Ct. 1126, 31 L.Ed.2d 515 (1972). The two non-carrier agents represented on the eleven-person Board when Atlas adopted the policy were capable of competing by acquiring interstate authority. The four carrier agents represented on the Board at the time were actual competitors of Atlas. Three of the remaining members of the board were officers of Atlas. Thus, all but two members of the board represented separate legal entities that competed in interstate commerce. This brings the case within the rule of Sealy and Topeo and shows the existence of a horizontal restraint. That conclusion does not mean, however, that the restraint is illegal.

B.

Since the restraint on competition within the Atlas system involves an agreement not to deal with those who do not comply with Atlas’ policy, and so may be characterized as a boycott, or a concerted refusal to deal, Rothery contends that Supreme Court decisions require a holding of per se illegality. It cannot be denied that the Court has often enunciated that broad proposition. See, e.g., Arizona v. Maricopa County Medical Society, 457 U.S. 332, 344 n. 15, 102 S.Ct. 2466, 2473 n. 15, 73 L.Ed.2d 48 (1982); Klor’s, Inc. v. Broadway-Hale Stores, 359 U.S. 207, 212, 79 S.Ct. 705, 709, 3 L.Ed.2d 741 (1959); Northern Pacific R.R. v. United States, 356 U.S. 1, 5, 78 S.Ct. 514, 518, 2 L.Ed.2d 545 (1958); Fashion Originators’ Guild of America v. FTC, 312 U.S. 457, 468, 61 S.Ct. 703, 708, 85 L.Ed. 949 (1941). Other cases, such as Associated Press v. United States, 326 U.S. 1, 65 S.Ct. 1416, 89 L.Ed. 2013 (1945), seem to indicate that not even the presence of a joint venture or economic integration to which the agreement against competition is ancillary can save a boycott from per se illegality.

Despite the seeming inflexibility of the rule as enunciated by the Court, it has always been clear that boycotts are not, and cannot ever be, per se illegal. To apply so rigid and simplistic an approach would be to destroy many common and entirely beneficial business arrangements. As one commentator put it, “[a]ll agreements to deal on specified terms mean refusal to deal on other terms,” and the literal application of per se illegality to any situation involving a concerted refusal to deal would mean in practical effect “that every restraint is illegal.” See Rahl, Per Se Rules and Boycotts Under the Sherman Act: Some Reflections and the Klor’s Case, 45 Va.L.Rev. 1165, 1172 (1959). For that reason, “any comprehensible per se rule for [group] boycotts ... is out of the question.” Id. at 1173.¹ Lower courts have long agreed with that assessment. See Phil Tolkan Datsun, Inc. v. Greater Milwaukee Datsun Dealers' Advertising Association, 672 F.2d 1280 (7th Cir.1982) (excluding an auto dealer from membership in an advertising association lacking market power did not constitute a per se violation of section 1 of the Sherman Act); United States Trotting Association v. Chicago Downs Association, Inc., 665 F.2d 781 (7th Cir.1981) (en banc) (finding that it was not per se unlawful for an organization of rival harness racing operations to prevent free riding on its informational and standard-promulgation services by sanctioning members who allowed their horses to participate in events sponsored by non-affiliates of the organization); United States v. Realty Multi-List, 629 F.2d 1351 (5th Cir.1980) (refusing to apply the per se rule where membership requirements had served to foreclose plaintiff from participating with rival realtors in a service providing multiple listing of properties); Smith v. Pro Football, Inc., 593 F.2d 1173 (D.C.Cir.1978) (eschewing a per se approach in a challenge to the NFL player draft because of a lack of “purpose to exclude competition”); E.A. McQuade Tours, Inc. v. Consolidated Air Tour Manual Committee, 467 F.2d 178 (5th Cir. 1972) (rejecting per se analysis of the refusal of airlines to list plaintiff air tour operator in air tour manual after plaintiff failed to meet listing criteria), cert. denied, 409 U.S. 1109, 93 S.Ct. 912, 34 L.Ed.2d 690 (1973); Deesen v. Professional Golfers’Association, 358 F.2d 165 (9th Cir.) (upholding the exclusion of a professional golfer with poor scores from tournament participation), cert. denied, 385 U.S. 846, 87 S.Ct. 72, 17 L.Ed.2d 76 (1966); Molinas v. National Basketball Association, 190 F.Supp. 241 (S.D.N.Y.1961) (holding that a league could adopt and enforce a rule prohibiting any of its teams from employing a player suspended for gambling).²

The Supreme Court has now made explicit what had always been understood. In Northwest Wholesale Stationers, Inc. v. Pacific Stationery & Printing Co., — U.S. -, 105 S.Ct. 2613, 86 L.Ed.2d 202 (1985), the plaintiff, a stationer, challenged as per se illegal its expulsion from a wholesale purchasing cooperative for violating the group’s bylaws. The Court said that “not all concerted refusals to deal should be accorded per se treatment.” Id. at 2621. We analyze Pacific Stationery below at somewhat greater length. See infra pp. 228-29. It is sufficient for present purposes to note that appellants’ contention about the per se illegality of all boycotts has now been squarely rejected by the Supreme Court.

IV.

Appellants contend, however, that Atlas’ restraints include horizontal price maintenance since the agents must ship on rates established by Atlas. We take this to be a claim that the horizontal elimination of competition within the system is illegal per se or, failing that, is nevertheless unlawful under a rule-of-reason analysis.

Before turning to the case law, we analyze the economic nature and effects of the system Atlas has created. It will be seen to be a system of a very familiar type, one commonly used in many fields of commercial endeavor.

Atlas has required that any moving company doing business as its agent must not conduct independent interstate carrier operations. Thus, a carrier agent, in order to continue as an Atlas agent, must either abandon its independent interstate authority and operate only under Atlas’ authority or create a new corporation (a “carrier affiliate”) to conduct interstate carriage separate from its operation as an Atlas agent. Atlas’ agents may deal only with Atlas or other Atlas agents.

The result of this is an interstate system for the carriage of household goods in which legally separate companies integrate their activities by contract. In this way the participants achieve many of the same benefits or efficiencies that would be available if they were integrated through ownership by Atlas. At the outset of this opinion, supra pp. 211-12, we set out the functions performed by Atlas and by the agents and stated that the system is a contract integration, one identical, in economic terms, to a partnership formed by agreement. Analysis might begin and end with the observation that Atlas and its agents command between 5.1 and 6% of the relevant market, which is the interstate carriage of used household goods.³ It is impossible to believe that an agreement to eliminate competition within a group of that size can produce any of the evils of monopoly. See, e.g., 3 J. Von Kalinowski, Antitrust Laws and Trade Regulation § 8.02[3], at 8-34 to 8-34.2 & n. 71 (1986). A monopolist (or those acting together to achieve monopoly results) enhances its revenues by raising the market price. It can do that only if its share of the market is so large that by reducing its output of goods or services the amount offered by the industry is substantially reduced so that the price is bid up. If a group of Atlas’ size reduced its output of services, there would be no effect upon market price because firms making up the other 94% of the market would simply take over the abandoned business. The only effect would be a loss of revenues to Atlas. Indeed, so impotent to raise prices is a firm with a market share of 5 or 6% that any attempt by it to engage in a monopolistic restriction of output would be little short of suicidal.

Appellants argue that Atlas’ 6% national market share understates the market power of Atlas and its agents to impose an anticompetitive result because appellants introduced evidence of the existence of distinct product geographic submarkets and because of evidence that the national market “approximates a tight oligopoly.” Reply Brief of Appellants at 19-20. Each of these propositions is wrong. With respect to the existence of submarkets, plaintiffs conceded the existence of a nationwide market and did not offer the district court any evidence that created a genuine issue of material fact as to the existence of sub-markets. The district judge, therefore, had only the national market before him. Indeed, so clear is the state of the evidence that we would affirm on this basis even if we thought it was not the rationale of the district court’s decision. See Jaffke v. Dunham, 352 U.S. 280, 281, 77 S.Ct. 307, 308, 1 L.Ed.2d 314 (1957) (per curiam); United States v. General Motors Corp., 518 F.2d 420, 441 (D.C.Cir.1975) (Leventhal, J.); 10 C. Wright, A. Miller & M. Kane, Federal Practice and Procedure: Civil 2d § 2716, at 658 (1983).

The criteria for defining markets are well-known. Because the ability of consumers to turn to other suppliers restrains a firm from raising prices above the competitive level, the definition of the “relevant market” rests on a determination of available substitutes. As Professor Sullivan has stated:

To define a market in product and geographic terms is to say that if prices were appreciably raised or volume appreciably curtailed for the product within a given area, while demand held constant, supply from other sources could not be expected to enter promptly enough and in large enough amounts to restore the old price and volume.

L. Sullivan, Antitrust § 12, at 41 (1977); see United States v. E.I. du Pont de Nemours & Co., 351 U.S. 377, 395, 76 S.Ct. 994, 1007, 100 L.Ed. 1264 (1956); Satellite Television & Associated Resources, Inc. v. Continental Cablevision of Virginia, Inc., 714 F.2d 351, 356 (5th Cir.1983), cert. denied, 465 U.S. 1027, 104 S.Ct. 1285, 79 L.Ed.2d 688 (1984). The degree to which a similar product will be substituted for the product in question is said to measure the cross-elasticity of demand, while the capability of other production facilities to be converted to produce a substitutable product is referred to as the cross-elasticity of supply. The higher these cross-elasticities, the more likely it is that similar products or the capacity of production facilities now used for other purposes are to be counted in the relevant market.

Brown Shoe Co. v. United States, 370 U.S. 294, 82 S.Ct. 1502, 8 L.Ed.2d 510 (1962), introduced into merger law the concept of submarkets within the relevant market. The Supreme Court identified several “practical indicia” that may be used to delineate submarkets, such as “industry or public recognition of the submarket as a separate economic entity, the product’s peculiar characteristics and uses, unique production facilities, distinct customers, distinct prices, sensitivity to price changes, and specialized vendors.” Id. at 325, 82 S.Ct. at 1524. These indicia seem to be evidentiary proxies for direct proof of substitutability. Brown Shoe said as much: “Because § 7 of the Clayton Act prohibits any merger which may substantially lessen competition ‘in any line of commerce’ (emphasis supplied), it is necessary to examine the effects of a merger in each such economically significant submarket to determine if there is a reasonable probability that the merger will substantially lessen competition.” Id. at 325, 82 S.Ct. at 1524.⁴ That view of submarket analysis is also mandated by the purpose of the antitrust laws: the promotion of consumer welfare. See Reiter v. Sonotone Corp., 442 U.S. 330, 343, 99 S.Ct. 2326, 2333, 60 L.Ed.2d 931 (1979). When submarket indicia are viewed as proxies for cross-elasticities they assist in predicting a firm’s ability to restrict output and hence to harm consumers.

Under no recognized definition of sub-markets did appellants offer evidence sufficient to raise an issue of material fact. Plaintiffs did not offer evidence that prices in alleged submarkets move independently of prices in surrounding areas or that, if prices were appreciably raised or volume appreciably curtailed in some areas, supply from other sources would not promptly restore the original price and volume. Nor did plaintiffs offer any evidence that any of the Brown Shoe “practical indicia” of sub-markets are present. Plaintiffs offered instead only assertions that Atlas had high shares of the market in a few cities. Such testimony does not show low cross-elasticities of demand or supply. It does not serve as evidence that these cities were segregable markets in which Atlas could raise prices appreciably without attracting competitors’ trucks from adjacent territories. Aside from this, plaintiffs offered only two casual remarks by carrier agents, one stating that its various offices constitute “distinct market area[s]” and the other merely alluding to “geographic and product markets and submarkets.” These general comments were not evidence of anything, and, in particular, they were certainly not evidence that the industry recognized some specific submarket as a “separate economic entity.”⁵ It is clear that these conelusory remarks, which are all plaintiffs offered, would not suffice to create an issue for a jury, and the trial judge would have to direct a verdict for defendant on the question of submarkets. See 10 C. Wright, A. Miller, & M. Kane, supra, § 2713.1, at 616. The plaintiffs in this case simply did not furnish any evidence “ ‘fairly arguable and of a substantial character,’ ” General Motors Corp., 518 F.2d at 442, of Brown Shoe factors indicating the existence of distinct submarkets.⁶

All that was before the district court, and all that is before us, therefore, is a nationwide market. Atlas, it is agreed, did 6% of the business in that market. And the relevance of this figure as a measure of market power is in no way diminished by appellants’ fanciful claims of “tight oligopoly” in the national market.

Given the fact, shown in note 3, that this industry consists of 1100 to 1300 interstate carriers, employing about 8000 agents, it would seem to be impossible to entertain any notion of market power. What plaintiffs offered to support their theory of “tight oligopoly” was a list of van lines and market shares that affirmatively proved their market power contention to be chimerical. Market concentration, and hence presumed power, is commonly measured according to the Herfindahl-Hirschman Index (“HHI”). As the U.S. Department of Justice Merger Guidelines explain, “[t]he HHI is calculated by summing the squares of the individual market shares of all the firms included in the market____ Unlike the traditional four-firm concentration ratio, the HHI reflects both the distribution of the market shares of the top four firms and the composition of the market outside the top four firms.” U.S. Department of Justice Merger Guidelines at 13-14 (1984) (footnote omitted) (“Merger Guidelines”).

The Department of Justice divides market concentration into three categories and characterizes a market with an HHI below 1000 as “unconcentrated." Merger Guidelines at 15. When the HHI is between 1000 and 1800, the market is “moderately concentrated,” and above 1800, “highly concentrated.” Id. One would suppose that a “tight oligopoly” is a market that is “highly concentrated.” In fact, when the market shares plaintiffs provided are squared, the van line market has an HHI of approximately 520.⁷ It is thus low on the range of unconcentrated markets. The Guidelines state that “the Department will not challenge mergers falling in this region [below 1000], except in extraordinary circumstances.” Id. No extraordinary circumstances have been shown, or even suggested, in this case.

We do not mean to suggest that if the HHI were higher and within one of the more concentrated categories, the arrangement would necessarily be illegal. It must be recalled that the Guidelines apply to mergers tested under section 7 of the Clayton Act, a statute aimed at halting “incipient monopolies and trade restraints outside the scope of the Sherman Act,” Brown Shoe, 370 U.S. at 318 n. 32, 82 S.Ct. at 1520 n. 32, and which therefore applies a much more stringent test than does rule-of-reason analysis under section 1 of the Sherman Act. It must also be recalled that the Guidelines apply to mergers between firms that ordinarily have no internal competition. Here we are dealing with firms that are merely limiting internal competition and are not merging to eliminate competition between firms. Indeed, if every van line in this industry eliminated all internal competition and then the largest two van lines merged, the resulting HHI would be only 868, still well below the top border for “unconcentrated markets.” ⁸ Plaintiffs did not offer evidence that this market approximates a tight oligopoly. They offered evidence that makes such a characterization merely absurd and which demonstrates the absence of market power in any van line, much less in Atlas, the sixth largest van line.

We might well rest, therefore, upon the absence of market power as demonstrated both by Atlas’ 6% national market share and by the structure of the market. If it is clear that Atlas and its agents by eliminating competition among themselves are not attempting to restrict industry output, then their agreement must be designed to make the conduct of their business more effective. No third possibility suggests itself. But we need not rely entirely upon that inference because the record made in the district court demonstrates that the challenged agreement enhances the efficiency of the van line. The chief efficiency, as already noted, is the elimination of the problem of the free ride.

A carrier agent can attract customers because of Atlas’ “national image” and can use Atlas’ equipment and order forms when undertaking carriage for its own account. Plaintiffs’ Statement of Genuine Issues 1126, at 40 (“PSGI”). The carrier agents “benefit from use of the services of moving and storage firms affiliated with Atlas, for origin or destination work at remote locations, when operating independently of Atlas.” Id. K 27, at 41. This benefit involves not only the availability of a reliable network of firms providing such services, but also includes the benefit of Atlas’ “mediating collection matters” among its affiliates. Id. 1127, at 42. To the degree that a carrier agent uses Atlas’ reputation, equipment, facilities, and services in conducting business for its own profit, the agent enjoys a free ride at Atlas’ expense. The problem is that the van line’s incentive to spend for reputation, equipment, facilities, and services declines as it receives less of the benefit from them. That produces a deterioration of the system’s efficiency because the things consumers desire are not provided in the amounts they are willing to pay for. In the extreme case, the system as a whole could collapse.

By their own assertions, appellants establish that the carrier agents in the Atlas organization have benefited from Atlas’ business infrastructure in carrying shipments made for their own accounts. Rothery suggests free riding does not occur, and that the district court erred in concluding that it did. See Brief for Rothery at 29-30. That argument, however, cannot withstand scrutiny, for Rothery has conceded that the carrier agents associated with Atlas do derive significant benefits from Atlas in dealing with customers for their own profit.⁹ We find the district court’s conclusion that free riding existed to be amply supported and by no means clearly erroneous.

A few examples will suffice. Plaintiff-appellants conceded below that the carrier agents “benefited” from their association with Atlas’ “national image.” PSGI 1126, at 40. We cannot rationally infer that this consumer identification advantage did not benefit the carrier agents in operating on their own accounts while using Atlas equipment and personnel trained by Atlas. Rothery also allowed that, while the carrier agents bore the bulk of costs associated with their operations, Atlas did make “some small contributions” to the group advertising programs and “some contributions” to the painting of trucks on which the Atlas logo appeared. See Statement of Material Undisputed Facts in Support of Plaintiffs’ Motion for Partial Summary Judgment as to Liability 1TV-C, at 15 (“Plaintiffs’ Statement of Material Undisputed Facts”).

Rothery also credited Atlas with providing a dispatching service, a clearinghouse service for the settlement of accounts among its affiliates, assistance in settling claims among affiliates, certain written forms, sales meetings to provide exposure to national customers, driver and employee training programs, and the screening of the quality and reliability of affiliated firms that provided origin and destination services for the carrier agents. See Plaintiffs’ Statement of Material Undisputed Facts 11Y-D, at 15-17.

Rothery did not assert in its Statement of Material Undisputed Facts, nor may we infer, that the carrier agents could not avail themselves of the benefits derived from these services when operating for their own accounts. Many of these services confer intangible advantages that redound to the benefit of the carrier agent as a whole, and do not admit of easy segregation as between shipments on Atlas’ interstate authority and shipments on the carrier agent’s authority. For example, if Atlas provides superior training to the employees of its carrier agents, that training improves the quality of work not only on shipments undertaken for Atlas but also on shipments made on the carrier agent’s own interstate authority. And because carrier agents may elect to use their own or Atlas’ interstate authority for a given shipment, see Plaintiffs’ Statement of Material Undisputed Facts f I-E, at 3, exposure to national clients at Atlas’ sales meetings can provide them with interstate customers for their own, as well as for Atlas’, accounts.

These examples are not exhaustive, but they illustrate the point. Even though entitled to every favorable inference, see United States v. General Motors Corp., 565 F.2d 754 (D.C.Cir.1977), Rothery, in light of the facts it deemed undisputed below, could not seriously contend that the carrier agents’ association with Atlas did not provide them with benefits that aided them in conducting business in competition with Atlas. Thus, because the plaintiff-appellants asserted that the carrier agents paid Atlas only for its clearinghouse service and for the provision of written forms, we agree with the district court’s finding that many of the services supplied as part of Atlas’ arrangement with the carrier agents’ arrangement resulted in Atlas subsidizing its competitors. See 597 F.Supp. at 233.

If the carrier agents could persist in competing with Atlas while deriving the advantages of their Atlas affiliation, Atlas might well have found it desirable, or even essential, to decrease or abandon many such services. See Continental T. V., Inc. v. GTE Sylvania Inc., 433 U.S. 36, 55, 97 S.Ct. 2549, 2560, 53 L.Ed.2d 568 (1977) (“Because of market imperfections such as the so-called ‘free rider’ effect, [certain] services might not be provided ... in a purely competitive situation ____”). Of that tendency there can be no doubt. When a person or business providing goods or services begins to receive declining revenues, then, other things being equal, that person or firm will provide fewer goods or services. As marginal revenue drops, so does output. Thus, when Atlas’ centralized services, equipment, and national image amount to a subsidy of competing carrier agents, this cuts down the marginal revenue derived from the provision of such things so that less will be offered than the market would reward.

On the other side, the firm receiving a subsidized good or service will take more of it. As cost declines, then, other things being equal, demand increases. Carrier agents, that is, will increase the use of Atlas’ services, etc., on interstate carriage for their own accounts, over-consuming that which they can obtain at less than its true cost. In this way, free riding distorts the economic signals within the system so that the van line loses effectiveness in serving consumers. The restraint at issue in this case, therefore, is a classic attempt to counter the perceived menace that free riding poses. By compelling carrier agents to transfer their interstate authority to a separate entity, Atlas can continue providing services at optimal levels, confident that it will be paid for those services.

The Atlas agreements thus produce none of the evils of monopoly but enhance consumer welfare by creating efficiency. There seems no reason in the rationale of the Sherman Act, or in any comprehensible policy, to invalidate such agreements. Nevertheless, at one, intermediate, point in the history of antitrust, Supreme Court decisions seemed to require just that result. It seems clear, however, that the law has returned to the original understanding so that the agreements before us are plainly lawful. Current decisional law is best understood if placed in historical context.

V.

The law concerning contract integrations and the restraints of trade that augment their effectiveness has been a variable growth. At times, courts have thought integrations or partnerships with restraints like Atlas’ obviously not only lawful but desirable. At other times, courts have treated the restraints as illegal per se, beyond any possibility of justification. The question is the state of the law today. We begin with the law’s earliest analysis of such restraints, then discuss later cases that implicitly repudiated that analysis, and, finally, seek to discover the degree to which the law has now returned to its original formulation.

From the inception of antitrust policy, the Supreme Court has recognized that the elimination of rivalry by the joinder of rivals into a larger economic unit is not, per se, an unlawful restraint of trade. After framing a rule of per se illegality for naked price fixing among competitors in United States v. Trans-Missouri Freight Association, 166 U.S. 290, 17 S.Ct. 540, 41 L.Ed. 1007 (1897), the Supreme Court in United States v. Joint Traffic Association, 171 U.S. 505, 19 S.Ct. 25, 43 L.Ed. 259 (1898), responded to the objection that condemning every restraint would destroy the most ordinary and indispensable contracts and integrations since they restrained the competition of the parties. The Court majority replied that neither the formation of a corporation or a contract of partnership had ever “been regarded in the nature of a contract in restraint of trade or commerce.” 171 U.S. at 567, 19 S.Ct. at 31. Though the verbal formulation of the law has changed, it remains true that the lessening of actual or potential rivalry inherent in the formation of every corporation and every partnership has never been held or said to be illegal per se by the Supreme Court. As Justice Holmes put it in dissent in Northern Securities Co. v. United States, 193 U.S. 197, 411, 24 S.Ct. 436, 472, 48 L.Ed. 679 (1904) (misconstruing the rule applied by the majority), any such interpretation of the Sherman Act “would make eternal the bellum omnium contra omnes and disintegrate society so far as it could into individual atoms.” No such construction of the Act is thinkable.

Appellants here do not challenge the formation or existence of the Atlas system. They complain only of the agreement that prevents them from maintaining an independent interstate carrier operation in the same corporation that acts as an agent of Atlas. The law has had more difficulty with agreements of this sort than it has had with the underlying integration or partnership. We begin by examining the Sherman Act’s earliest position with respect to such agreements.

In United States v. Addyston Pipe & Steel Co., 85 F. 271 (6th Cir.1898), affd, 175 U.S. 211, 20 S.Ct. 96, 44 L.Ed. 136 (1899), Judge (later Chief Justice) William Howard Taft framed a rule of per se illegality for “naked” price-fixing and market-dividing agreements, i.e., agreements between competitors who cooperated in no other integrated economic activity. But Taft recognized that such a rule would not do where fusions or integrations of economic activity occurred and, further, that agreements eliminating rivalry within such an enterprise were means of enhancing the firm’s efficiency. He explained the reasons for the validity of an agreement “by a partner pending the partnership not to do anything to interfere, by competition or otherwise, with the business of the firm”:

[W]hen two men became partners in a business, although their union might reduce competition, this effect was only an incident to the main purpose of a union of their capital, enterprise, and energy to carry on a successful business, and one useful to the community. Restrictions in the articles of partnership upon the business activity of the members, with a view of securing their entire effort in the common enterprise, were, of course, only ancillary to the main end of the union, and were to be encouraged.

United States v. Addyston Pipe & Steel Co., 85 F. at 280.

To be ancillary, and hence exempt from the per se rule, an agreement eliminating competition must be subordinate and collateral to a separate, legitimate transaction. The ancillary restraint is subordinate and collateral in the sense that it serves to make the main transaction more effective in accomplishing its purpose. Of course, the restraint imposed must be related to the efficiency sought to be achieved. If it is so broad that part of the restraint suppresses competition without creating efficiency, the restraint is, to that extent, not ancillary. Taft added the further obvious qualification that even restraints ancillary in form are illegal if they are part of a general plan to gain monopoly control of a market. 85 F. at 282-83.

If Taft’s formulation is the law today, it is obvious that the Atlas agreements are legal, for Addyston Pipe & Steel’s analysis of ancillary restraints fits this case exactly.¹⁰ The Atlas network involves a union of the parties’ enterprise to carry on a useful business, the challenged agreements are ancillary in that they enhance the efficiency of that union by eliminating the problem of the free ride, and, given Atlas’ small market share, the agreements cannot be part of a plan to gain monopoly control of the market.

The argument that horizontal eliminations of competition among legally independent persons or companies are automatically illegal, even though the restraint is ancillary to a partnership or a joint venture, rests primarily upon United States v. Topeo Associates, Inc., 405 U.S. 596, 92 S.Ct. 1126, 31 L.Ed.2d 515 (1972). The business arrangement in Topeo very closely resembles Atlas’ policy.

Topeo was a cooperative association of twenty-five small and medium-sized regional supermarket chains operating stores in thirty-three states. Topeo functioned as a purchasing agent for its members, ensured quality control on purchased products, developed specifications for certain types of products, and performed other tasks that gave members the efficiencies usually attainable only by large chains. Topco’s stock was owned by its member chains. They were geographically dispersed and, in their various areas, had market shares ranging from 1.5% to 16%, with the average being about 6%.

The legal challenges centered on Topco’s private-label program. Topeo members had difficulty competing with larger chains and this problem was to some degree attributable to the larger chain’s ability to develop their own private labels. The Court’s opinion set out some of the efficiencies of private labeling:

Private-label products differ from other brand-name products in that they are sold at a limited number of easily ascertainable stores. A&P, for example, was a pioneer in developing a series of products that were sold under an A&P label and that were only available in A&P stores. It is obvious that by using private-label products, a chain can achieve significant cost economies in purchasing, transportation, warehousing, promotion, and advertising. These economies may afford the chain opportunities for offering private-label products at lower prices than other brand-name products. This, in turn, provides many advantages of which some of the more important are: a store can offer national-brand products at the same price as other stores, while simultaneously offering a desirable, lower priced alternative; or, if the profit margin is sufficiently high on private-brand goods, national-brand products may be sold at reduced price. Other advantages include: enabling a chain to bargain more favorably with national-brand manufacturers by creating a broader supply base of manufacturers, thereby decreasing dependence on a few, large national-brand manufacturers; enabling a chain to create a “price-mix” whereby prices on special items can be lowered to attract customers while profits are maintained on other items; and creation of general goodwill by offering lower priced, higher quality goods.

United States v. Topeo Associates, 405 U.S. at 599 n. 3, 92 S.Ct. at 1129 n. 3.

Thus, Topeo, like Atlas, was a contractual integration of legally independent businesses designed to achieve efficiencies unavailable to its members separately. Also like Atlas, however, Topeo had an ancillary horizontal restraint designed to make the integration more effective. Members could sell Topco-brand products only in designated, and usually exclusive, territories. Member chains were free to expand into each other’s territories, and did, but they could not sell the Topeo brand in the new territory if another member held the rights there. This restraint had a clear relationship to marketing effectiveness. As Topeo said, “ ‘[pjrivate label merchandising is a way of economic life in the food retailing industry, and exclusivity is the essence of a private label program; without exclusivity, a private label would not be private.’ ” 405 U.S. at 604, 92 S.Ct. at 1132. It noted that every national and large regional chain had its own exclusive private-label products. “ ‘Each such chain relies upon the exclusivity of its own private label line to differentiate its private products from those of its competitors and to attract and retain the repeat business and loyalty of consumers.’” Id. at 604-05, 92 S.Ct. at 1132. Smaller chains had to have their own lines to compete with larger chains, which accounted for the Topeo program, and needed similar exclusivity for advertising and promotional purposes, which accounted for the market division with respect to the Topeo label.

The Supreme Court said, however, “[w]e think that it is clear that the restraint in this case is a horizontal one, and, therefore, a per se violation of § 1 [of the Sherman Act].” 405 U.S. at 608, 92 S.Ct. at 1134. The Topeo opinion also clarified United States v. Sealy Corp., 388 U.S. 350, 87 S.Ct. 1847, 18 L.Ed.2d 1238 (1967), which had appeared to hold illegal a very similar set of restraints among mattress manufacturers that wished to market a national brand because both price fixing and market division were involved. Sealy was now said to hold that horizontal territorial limitations by themselves were per se unlawful. See 405 U.S. at 609, 92 S.Ct. at 1134.

If Topeo and Sealy, rather than Addyston Pipe & Steel, state the law of horizontal restraints, the restraints imposed by Atlas would appear to be a per se violation of the Sherman Act. An examination of more recent Supreme Court decisions, however, demonstrates that, to the extent that Topeo and Sealy stand for the proposition that all horizontal restraints are illegal per se, they must be regarded as effectively overruled.

C.

The Supreme Court reformed the law of horizontal restraints in Broadcast Music, Inc. v. Columbia Broadcasting System, 441 U.S. 1, 99 S.Ct. 1551, 60 L.Ed.2d 1 (1979) (“BMI”), National Collegiate Athletic Association v. Board of Regents, 468 U.S. 85, 104 S.Ct. 2948, 82 L.Ed.2d 70 (1984) (“NCAA”), and Northern Wholesale Stationers, Inc. v. Pacific Stationery & Printing Co., — U.S. -, 105 S.Ct. 2613, 86 L.Ed.2d 202 (1985) (“Pacific Stationery ”).

In BMI, CBS brought suit to challenge the blanket licenses to perform copyrighted musical compositions issued by BMI and by the American Society of Composers, Authors and Publishers (“ASCAP”). Each blanket license gave the licensee the right to perform any and all of the compositions owned by members and affiliates of the performing rights organization for a stated period of time. Since BMI and ASCAP negotiated the price of their blanket licenses and distributed royalties to the copyright owners, the charge, upheld by the court of appeals, was that the blanket license was a form of price fixing between the copyright owners and so was a horizontal restraint illegal per se under the Sherman Act.

The Supreme Court rejected a “literal approach” to price-fixing because that approach, by itself, does not establish whether a particular practice is of a type that is plainly anticompetitive and very likely without redeeming virtue.

Literalness is overly simplistic and often overbroad. When two partners set the price of their goods or services they are literally “price fixing,” but they are not per se in violation of the Sherman Act. See United States v. Addyston Pipe & Steel Co., 85 F. 271, 280 (CA6 1898), aff’d 175 U.S. 211 [20 S.Ct. 96, 44 L.Ed. 136] (1899).

BMI, 441 U.S. at 9, 99 S.Ct. at 1557. That observation is significant for it shows that the BMI Court recognized that partnerships, one form of integration by contract, involve horizontal restraints that are not per se illegal. It also demonstrates, contrary to appellants’ claims here, that “price fixing” within a contract integration does not necessarily violate the antitrust law. But the Court’s analysis of whether the per se label was appropriate in BMI has even more significance for the present case.

The Court began by noting the existence of consent decrees the government had worked out with ASCAP and BMI, said the decrees, though they did not immunize the blanket license from third parties’ suits, did indicate that “the challenged practice may have redeeming competitive virtues,” so that the decrees could not be ignored completely in analyzing the practice of blanket licensing. See 441 U.S. at 13, 99 S.Ct. at 1559. Moreover, Congress, albeit in other performances contexts, had created compulsory blanket licenses, reflecting an opinion that blanket licenses “are economically beneficial in at least some circumstances.” Id. at 16, 99 S.Ct. at 1560.

These observations have their analogues in the present case. That the challenged practice of agent exclusivity “may have redeeming competitive virtues” is shown by the fact that all van lines use the practice, see 597 F.Supp. at 222, and that the Interstate Commerce Commission routinely approves the practice, without even requiring a hearing. See Three Way Corp. v. United States, 792 F.2d 282 (D.C.Cir. 1986). The constraints the consent decrees placed on ASCAP, BMI, 441 U.S. at 24, 99 S.Ct. at 1564, which has a great majority of the most desired music played in the United States, are analogous to the restraints the free market places on Atlas, which does less than 6% of the business in its market. Moreover, just as Congress recognized the utility of the blanket license, so Congress recognized the utility of restraints of the sort Atlas imposes. That is shown by the congressional provision of antitrust immunity for ICC-approved pooling agreements containing such restrictions and is further evidenced by the Senate Committee on Commerce, Science and Transportation statements concerning restrictions imposed outside a pooling agreement:

During the course of its deliberations, the Committee considered a proposal to include within the immunity section [49 U.S.C. § 10934(d) ] language that would grant immunity to agreements concerning exclusive agency representation, or the fiduciary duty of loyalty of an agent not to compete with the principal concerning the subject matter of the agency. The Committee determined that this type of relationship is not a violation of the antitrust laws and is standard agency law as expressed in Section 393 of the Restatement of the Law of Agencies section. The Committee felt that inclusion of such language could imply that such relationships are a violation of the antitrust laws in the absence of immunity. Thus, the Committee determined not to include such language in the bill.

S.Rep. No. 497, 96th Cong., 1st Sess. 8 (1979) (emphasis added). Though, like the Court in BMI, we are not bound by Congress’ opinion, it does suggest that agency exclusivity is “economically beneficial in at least some circumstances.”

These parallels between BMI and Atlas’ practice are specific to the two cases. They may be enough to take this particular practice out of the category of per se illegality. But there is in BMI reasoning of more general application which indicates that Topeo does not state the modern rule as to horizontal restraints. There is, first, the Court’s favorable citation of Addyston Pipe & Steel’s example of partners who eliminate price competition between themselves. If Topeo meant, as it seemed to, that the existence of a joint venture could not justify an agreement eliminating competition between the joint venturers, the BMI Court must be read as overruling Topeo to that extent.

The opinion pointed out that the practices of ASCAP and BMI were necessary to secure for copyright owners the right to control and profit from the public performance of their musical compositions. The Court said, “we would not expect that any market arrangements reasonably necessary to effectuate the rights that are granted would be deemed a per se violation of the Sherman Act.” BMI, 441 U.S. at 19, 99 S.Ct. at 1562. In the present case, there is authority to conduct interstate carriage granted by the ICC. Atlas holds that authority but would have a difficult time using it but for its arrangements with its agents. Atlas is not in a position to conduct all of the carriage involved itself. The restraints it imposes are reasonably necessary to the business it is authorized to conduct. One may quibble about whether the ASCAP blanket license is more necessary to the conduct of that business than the agent exclusivity arrangement is to the conduct of Atlas’ business. We do not believe, however, that, in choosing the words it did, the Supreme Court intended that lower courts should calibrate degrees of reasonable necessity. That would make the lawfulness of conduct turn upon judgments of degrees of efficiency. There is no reason in logic why the question of degree should be important.

That conclusion is buttressed by the Court’s next observation, that

inquiry must focus on whether the effect and, here because it tends to show effect, ... the purpose of the practice are to threaten the proper operation of our predominantly free-market economy — that is, whether the practice facially appears to be one that would, always or almost always tend to restrict competition and decrease output ...or instead one designed to “increase economic efficiency and render markets more, rather than less, competitive.”

BMI, 441 U.S. at 19-20, 99 S.Ct. at 1562 (emphasis added) (citations omitted). This inquiry implements the Court's designation of consumer welfare as the policy goal of the Sherman Act. Reiter v. Sonotone Corp., 442 U.S. 330, 343, 99 S.Ct. 2326, 2333, 60 L.Ed.2d 931 (1979). The Court concluded that the blanket license, though it involved price fixing in a literal sense, was not per se unlawful and remanded the case for further rule-of-reason analysis.

The Supreme Court’s NCAA decision confirms the analytical approach adopted in BMI. The NCAA, as part of its regulation of intercollegiate athletics, adopted a plan for the televising of member institutions’ football games. The plan limited the number of intercollegiate contests that could be televised and the number of times any one college could televise. No college could sell television rights independently. NCAA negotiated with two networks and arranged a specified price for categories of games. Colleges dissatisfied with the plan’s limitations on their ability to sell television rights brought suit, challenging the restraints under section 1 of the Sherman Act. The Supreme Court ruled that section 1 was violated but, significantly for the present case, refused to apply a per se rule.

In language we have already quoted, supra p. 215, the Court held that the NCAA plan created a horizontal restraint on price competition, 104 S.Ct. at 2959, but nevertheless explicitly refused to apply the per se rule, id. at 2960. It did so on the ground that horizontal restraints were necessary if the product was to be available. Id. at 2961. Moreover, the Court cited Continental T. V., Inc. v. GTE Sylvania Inc., 433 U.S. 36, 51-57, 97 S.Ct. 2549, 2558-61, 53 L.Ed.2d 568 (1977), for the proposition that “a restraint in a limited aspect of a market may actually enhance marketwide competition.” 104 S.Ct. at 2961-62. Sylvania held a rule of per se illegality inappropriate to a manufacturer’s division of its dealers’ territories. The Court reasoned that the restraint addressed the problem of the free ride, see 433 U.S. at 55, 97 S.Ct. at 2560 — if dealers could locate in each other’s territories, those who spent on advertising and promotion might not be able to recover their investment. Some dealers, with no advertising and promotion costs, could charge lower prices and capture the customers created by others’ expenditures. By applying GTE Sylvania in a horizontal case as “requiring] consideration of the NCAA’s justifications for the restraints, 104 S.Ct. at 2962, the Supreme Court made it clear that elimination of the free ride is an efficiency justification available to horizontal restraints that are ancillary to a contract integration. Following BMI, the Court reserved the per se rule for practices that, on their face, appear to be of the type that “ ‘would always or almost always tend to restrict competition and decrease output.’ ” Id. at 2960 (citing BMI, 441 U.S. at 19-20, 99 S.Ct. at 1562). When per se treatment is inappropriate, the court must consider the justifications advanced for the restraint. 104 S.Ct. at 2962. After noting that consumer welfare is the goal of the Sherman Act, the Court said:

A restraint that has the effect of reducing the importance of consumer preference in setting price and output is not consistent with this fundamental goal of antitrust law. Restrictions on price and output are the paradigmatic examples of restraints of trade that the Sherman Act was intended to prohibit.

Id. at 2964 (footnote omitted). Upon considering the justifications advanced by the NCAA, the Court concluded that the restraint did not increase efficiency but decreased output and so violated section 1 of the Sherman Act.

Pacific Stationery applies to boycotts the general formula stated in BMI and NCAA — that the per se rule is confined to practices of the type that almost always decrease output rather than increasing efficiency — and so confirms, if confirmation were needed, that the formula applies to all horizontal restraints. Pacific Stationery, as already noted, involved the expulsion of a stationer from a wholesale purchasing cooperative for violation of the bylaws. The Court noted the efficiencies attainable through such cooperative ventures and said that a joint venture “must establish and enforce reasonable rules in order to function effectively.” 105 S.Ct. at 2620. This is a recognition that ancillary restraints are essential to the efficiency of a contract integration. An anticompetitive effect is to be presumed only if the plaintiff makes a “threshold showing” that the group “possesses market power or exclusive access to an element essential to effective competition.” Id. This statement of the law of ancillary restraints is so close to that of Addyston Pipe & Steel as to be virtually indistinguishable.

BMI, NCAA, and Pacific Stationery dictate the result in this case. All horizontal restraints are alike in that they eliminate some degree of rivalry between persons or firms who are actual or potential competitors. This similarity means that the rules applicable to all horizontal restraints should be the same. At one time, as we have seen, the Supreme Court stated in Topeo and Sealy that the rule for all horizontal restraints was one of per se illegality. The difficulty was that such a rule could not be enforced consistently because it would have meant the outlawing of very normal agreements (such as that of law partners not to practice law outside the firm) that obviously contributed to economic efficiency. The alternative formulation was that of Judge Taft in Addyston Pipe & Steel: a naked horizontal restraint, one that does not accompany a contract integration, can have no purpose other than restricting output and raising prices, and so is illegal per se; an ancillary horizontal restraint, one that is part of an integration of the economic activities of the parties and appears capable of enhancing the group’s efficiency, is to be judged according to its purpose and effect. In BMI, NCAA, and Pacific Stationery, the Supreme Court returned the law to the formulation of Addyston Pipe & Steel and thus effectively overruled Topeo and Sealy as to the per se illegality of all horizontal restraints.

The application of these principles to Atlas’ restraints is obvious because, as we have seen, supra pp. 211-12, 221-23, these restraints are ancillary to the contract integration or joint venture that constitutes the Atlas van line. The restraints preserve the efficiencies of the nationwide van line by eliminating the problem of the free ride. There is, on the other hand, no possibility that the restraints can suppress market competition and so decrease output. Atlas has 6% or less of the relevant market, far too little to make even conceivable an adverse effect upon output. If Atlas should reduce its output, it would merely shrink in size without having any impact upon market price. See supra p. 217. Under the rule of Addyston Pipe & Steel, BMI, NCAA, and Pacific Stationery, therefore, it follows that the Atlas agreements do not violate section 1 of the Sherman Act.¹¹

A joint venture made more efficient by ancillary restraints, is a fusion of the productive capacities of the members of the venture. That, in economic terms, is the same thing as a corporate merger. Merger policy has always proceeded by drawing lines about allowable market shares and these lines are based on rough estimates of effects because that is all the nature of the problem allows. If Atlas bought the stock of all its carrier agents, the merger would not even be challenged under the Department of Justice Merger Guidelines because of inferences drawn from Atlas’ market share and the structure of the market. We can think of no good reason not to apply the same inferences to Atlas’ ancillary restraints.

The judgment of the district court is Affirmed.

1 . The truth of this may be easily demonstrated. When a law firm refuses to hire an applicant there is a concerted refusal to deal since the lawyers in the firm are separate legal entities and capable of practicing law independently. It is also a boycott if the Ivy League refuses to admit a new college to membership or the American League refuses to admit a baseball team. It is no less a boycott if any of these groups refuses to deal because the applicant’s grades are too low, or its football program has standards unacceptable to the Ivy League, or the would-be baseball franchise is currently a slow-pitch softball team in an industrial league. A ruling that concerted refusals to deal are per se illegal would mean that Atlas not only must retain carrier agents that compete with it but must admit as an agent any trucker who applied regardless of the need for an additional agent, the trucker’s financial condition, its safety record, or its ability to serve customers. That is what a per se rule means: no group may impose a standard of any kind as a condition of dealing. That nonsensical requirement would destroy all of the groups concerned or force them into one ownership in order to claim the immunity of the Copperweld rule.

2 . Accord Brenner v. World Boxing Council, 675 F.2d 445, 455 (2d Cir.) (suspension of a fight promoter from promoting World Boxing Council fights for "failure to comply with his commitments") (applying the rule of reason unless the restraint at issue can “serve no purpose beyond the stifling of competition), cert. denied, 459 U.S. 835, 103 S.Ct. 79, 74 L.Ed.2d 76 (1982); Las Vegas Sun, Inc. v. Summa Corp., 610 F.2d 614, 619 (9th Cir.1979) (cancellation of newspaper advertising by those who had received unfavorable coverage by that newspaper) (holding that the rule of reason applies "[wjhere an agreement between competitors ‘having a primary purpose and direct effect of accomplishing a legitimate business objective is also alleged to have had an incidental and indirect adverse effect upon the business of some competitors’”), cert. denied, 447 U.S. 906, 100 S.Ct. 2988, 64 L.Ed.2d 855 (1980). The rationales employed have varied, with some cases relying upon recent Supreme Court precedent concerning the per se rule, see, e.g., Realty Multi-List, 629 F.2d at 1367; see also Part V-C infra at pp. 33-41, some relying upon the dicta in Silver v. New York Stock Exchange, 373 U.S. 341, 348-49, 83 S.Ct. 1246, 1252,10 L.Ed.2d 389 (1963), to the effect that a group boycott may be justified by reference to the policies of another federal statute "or otherwise,” see, e.g., Brenner, 675 F.2d at 454, and some relying upon both rationales, see, e.g., Phil Tolkan Datsun, 672 F.2d at 1284-87. The proposition that not all concerted refusals to deal are to be treated as per se illegal has found considerable support among commentators as well. See, e.g., L. Sullivan, Antitrust § 89, at 253-56 (1977); 2 J. Von Kalinowski, Antitrust Laws and Trade Regulation § 6C.02[2], at 6C-25 to 6C-27 & n. 68 (1986); Bauer, Per Se Illegality of Concerted Refusals to Deal: A Rule Ripe for Reexamination, 79 Colum.L.Rev. 685 (1979).

3 . The interstate household goods industry consists of 1100 to 1300 interstate carriers, of which the 15 largest constitute 70% of the market. These carriers employ roughly 8,000 agents. Based on data compiled for 1981, Atlas was the sixth largest interstate carrier, with a 5.86% market share. The market share tapered gradually, from the largest firm’s 13.3% share to Atlas’ position, after which shares dropped precipitously to 3.19% and 1.99% for the seventh and eighth largest firms. See Plaintiffs' Statement of Genuine Issues fl 69, at 97 (“PSGI"). Thus, the market cannot be said to he heavily concentrated and Atlas is by no means a dominant force in the market.

4 . The first group of indicia mentioned in Brown Shoe relates to the ability of the consumer to obtain substitutes for a product and, therefore, goes directly to the economic criteria that make one market distinct from another. One factor is "unique production facilities.” If a product requires unique production facilities, and the producer raises the price above the competitive level, the ability of other producers to shift resources to make the product would be limited, and the market definition should be likewise limited. "[D]istinct prices" and "sensitivity to price changes" also relate directly to the economic definition of a market. The first suggests that cross-elasticity of demand is low, the second that it is high.

The second set of indicia bear less directly upon the economic definition of a market, representing observations about what one ordinarily observes when a market is distinct. The “industry or public recognition of the submarket as a separate economic” unit matters because we assume that economic actors usually have accurate perceptions of economic realities. The "product’s peculiar characteristics" refers to the general truth that substitutes in a market often have a strong physical and functional relationship. Both "distinct customers” and "specialized vendors” may indicate unique product attributes, which refers again to the fact that products with distinct physical and functional attributes tend to be priced differently. These factors may be helpful where the other indicia are ambiguous. See generally 3 J. Von Kalinowski, Antitrust Laws and Trade Regulation § 8.02[2], at 8-27 (1986) (stating that Brown Shoe does not provide “a new test” for determining the relevant market, but merely provides "several new factors” in discovering "interchangeability between different products”).

5 . Indeed, other statements by plaintiffs demonstrate that there could hardly be such an industry recognition since plaintiffs as a group had no common recognition of submarkets. Thus, in support of its claim of a nationwide market, the defendant cited, among other things, the deposition of the owner of one of the plaintiff carrier agents. See Defendant’s Statement of Material Facts as to Which There Is No Genuine Dispute fl 69, at 60; see also Memorandum of Points and Authorities in Support of Defendant’s Motion for Summary Judgment at 62. This plaintiff testified that the carrier agent is licensed to operate throughout the United States, that is trucks will go wherever there is a "good shipment coming back,” and that shorter hauls are preferred but "it doesn’t work out that way.” Deposition of "Z” at 34. That plaintiff further stated:

By geography!,] [a] great part of our business is generated in the local marketplace. My sales market area is the whole world, however.

Id. at 120.

6 . Though we do not rely upon this observation in reaching our decision, it seems apparent why plaintiffs did not offer evidence of real sub-markets and why they barely mentioned the point and then in their reply brief only. In an industry in which the supply of the product, space in truck trailers, is among the most mobile factors of production imaginable, and the nature of the business causes these factors of production to be constantly moving throughout the country, it is inconceivable that any showing of submarkets could be made. Any attempt in one city to raise prices above competitive levels would be met by other van lines sending in trucks and trailers at a lower price. This view is supported by a plaintiffs deposition testimony. See supra note 5.

7 . The 1981 national market shares for the top fifteen van lines were as follows:

SHARE

1. Allied Van Lines, Inc. 13.30%

2. North American Van Lines, Inc. 13.07

3. United Van Lines, Inc. 11.41

4. Aero-Mayflower Transit Co. 8.46

5. Bekins Van Lines Co. 6.76

6. Atlas Van Lines, Inc. 5.86

7. Global Van Lines, Inc. 3.19

8. Burnham Van Services, Inc. 1.99

9. Wheaton Van Lines, Inc. 1.81

10. American Red Ball Transit 1.18

11. Neptune World Wide Transit 1.06

12. National Van Lines, Inc. .76

13. Pan American Van Lines, Inc. .56

14. Cartwright Van Lines, Inc. .52

15. Interstate Van Lines, Inc. .43

PSGI ¶ 69, at 97. We base our calculation of the HHI on these figures. Although we do not have at our disposal the market share of every firm in the industry, it is only the marginal or "fringe" firms for which we lack data, and the Guidelines explicitly and correctly state that "lack of information about small fringe firms is not critical because such firms do not affect the HHI significantly.” Merger Guidelines at 14 n. 14.

8 . The Supreme Court has shown that interbrand competition, which is intense in the van line industry, can prevent the manifestation of anti-competitive effects from a firm's internal restraints. In Continental T. V., Inc. v. GTE Sylvania Inc., 433 U.S. 36, 52 n. 19, 97 S.Ct. 2549, 2558 n. 19, 53 L.Ed.2d 568 (1977), the Court stated:

Interbrand competition is the competition among the manufacturers of the same generic product ... and is the primary concern of antitrust law. The extreme example of a deficiency of interbrand competition is monopoly, where there is only one manufacturer. In contrast, intrabrand competition is the competition between the distributors — wholesale or retail — of the product of a particular manufacturer.

The degree of intrabrand competition is wholly independent of the level of interbrand competition confronting the manufacturer. Thus, there may be fierce intrabrand competition among the distributors of a product produced by a monopolist and no intrabrand competition among the distributors of a product produced by a firm in a highly competitive industry. But when interbrand competition exists, ... it provides a significant check on the exploitation of intrabrand market power because of the ability of consumers to substitute a different brand of the same product.

9 . Rothery’s brief has substantially mischaracterized flV-C of its Statement of Material Undisputed Facts below. Rothery cites this paragraph for the proposition that “[w]hen a carrier agent makes a shipment on its own authority, it uses its own trucks, employees, packing materials, and other needed facilities.” Brief for Rothery at 29-30. In fact, the paragraph cited by Rothery does not support this sweeping assertion. It concedes that the carrier agents pay only "the bulk ... costs associated with their operations,” and it admits that Atlas made contributions covering some of the costs of operation. Nowhere does the cited paragraph suggest that contributions made by Atlas to the carrier agents benefited only Atlas shipments and not those undertaken for the carrier agents’ own accounts. Indeed, the only part of the cited paragraph that even resembles anything for which Rothery cites it is a quotation from the Atlas rules stating that Atlas’ operations were to be “conducted primarily with the facilities, drivers, and equipment of the agents of the Company.” See Statement of Material Undisputed Facts in Support of Plaintiffs’ Motion for Partial Summary Judgment as to Liability f V-C, at 15. This amounts to no more than a non-sequitur. It describes the manner by which shipments on Atlas’ interstate authority are to be made, and in no way precludes the conclusion that Atlas contributes to the carrier agents’ business facilities and that those contributions benefit the carrier agents in carriage on their own authority and for their own accounts. It goes beyond acceptable advocacy for Rothery to cite 1f V-C for the proposition that appears in its brief.

10 . A hypothetical was posed as oral argument which seems instructive. If there were a number of law firms in a given location of a size distribution like that of the van lines in this case, the extraction of a promise by each partnership that its members and associates not compete with the firm could, by no one’s estimation, be construed as running afoul of § 1 of the Sherman Act. Indeed, it is with respect to the legal profession that the law has properly distinguished between lawful ancillary and unlawful naked restraints. Although literally price fixing among competitors, fee schedules imposed by a law partnership are accepted, even taken for granted. Yet in Goldfarb v. Virginia State Bar, 421 U.S. 773, 95 S.Ct. 2004, 44 L.Ed.2d 572 (1975), the Supreme Court held that a minimum fee schedule applicable to the entire bar of a state violated § 1 of the Sherman Act. Thus, in its actual operation, the law has recognized the propriety of horizontal restraints ancillary to an efficiency-producing economic integration as distinct from the imposition of such restraints by competitors who have integrated none of their productive endeavors.

11 . Two additional points should be made. First, we do not think it significant to the outcome that Atlas’ policy allowed agents to exercise their own interstate authority through separate corporations. Once it is clear that restraints can only be intended to enhance efficiency rather than to restrict output, the degree of restraint is a matter of business rather than legal judgment. Second, though it is sometimes said that, in the case of restraints like these, it is necessary to weigh procompetitive effects against anticompetitive effects, we do not think that a useable formula if it implies an ability to quantify the two effects and compare the values found. Here, there are no anticompetitive effects and so there is nothing to place on that side of the scale. If the underlying contract integration is lawful, i.e., not of such size as to violate the Sherman Act, restraints ancillary to the integration, in the sense we have described, should be lawful. Weighing effects in any direct sense will usually be beyond judicial capabilities but predictions about effects may be reflected in rules about allowable size. The concurrence appears to suggest that the district court conducted a balancing of effects in some fashion other than by drawing inferences from market share and structure. If so, the district court did not explain its alternative method and made no findings on the subject. Nor does the concurrence articulate an alternative means of weighing procompetitive and anticompetitive effects. Antitrust adjudication has always proceeded through inferences about market power drawn from market shares. See, e.g., Consultants & Designers, Inc. v. Butler Service Group, Inc., 720 F.2d 1553, 1562-63 (11th Cir.1983) (holding under § 1 that when a defendant has "a relatively small portion of the ... [relevant] market,” imposition of a restraint cannot be construed "impermissibly [to] hurt either ... competitors or competition”); Smith v. Pro-Football, Inc., 593 F.2d 1173, 1185-86 (D.C.Cir. 1978) (applying § 1 and finding the "predictable effect" of a restraint imposed by 100% of the teams was "significantly anticompetitive”); United States v. Aluminum Company of America, 148 F.2d 416, 424 (2d Cir.1945) (L. Hand, J.) (a § 2 case holding that "[ninety] percent[] is enough to constitute a monopoly; it is doubtful whether sixty or sixty-four percent would be enough; and certainly thirty-three percent is not").