Justice Stevens

delivered the opinion of the Court.

The University of Oklahoma and the University of Georgia contend that the National Collegiate Athletic Association has unreasonably restrained trade in the televising of college football games. After an extended trial, the District Court found that the NCAA had violated § 1 of the Sherman Act1 and granted injunctive relief. 546 F. Supp. 1276 (WD Okla. 1982). The Court of Appeals agreed that the statute had been violated but modified the remedy in some respects. 707 F. 2d 1147 (CA10 1983). We granted certiorari, 464 U. S. 913 (1983), and now affirm.

I

The NCAA

Since its inception in 1905, the NCAA has played an important role in the regulation of amateur collegiate sports. It has adopted and promulgated playing rules, standards of amateurism, standards for academic eligibility, regulations concerning recruitment of athletes, and rules governing the size of athletic squads and coaching staffs. In some sports, such as baseball, swimming, basketball, wrestling, and track, it has sponsored and conducted national tournaments. It has not done so in the sport of football, however. With the *89exception of football, the NCAA has not undertaken any regulation of the televising of athletic events.2

The NCAA has approximately 850 voting members. The regular members are classified into separate divisions to reflect differences in size and scope of their athletic programs. Division I includes 276 colleges with major athletic programs; in this group only 187 play intercollegiate football. Divisions II and III include approximately 500 colleges with less extensive athletic programs. Division I has been subdivided into Divisions I-A and I-AA for football.

Some years ago, five major conferences together with major football-playing independent institutions organized the College Football Association (CFA). The original purpose of the CFA was to promote the interests of major football-playing schools within the NCAA structure. The Universities of Oklahoma and Georgia, respondents in this Court, are members of the CFA.

History of the NCAA Television Plan

In 1938, the University of Pennsylvania televised one of its home games.3 From 1940 through the 1950 season all of Pennsylvania’s home games were televised. App. 303. That was the beginning of the relationship between television and college football.

On January 11, 1951, a three-person “Television Committee,” appointed during the preceding year, delivered a report to the NCAA’s annual convention in Dallas. Based on preliminary surveys, the committee had concluded that “television does have an adverse effect on college football attendance and unless brought under some control threatens to seriously harm the nation’s overall athletic and physical *90system.” Id., at 265. The report emphasized that “the television problem is truly a national one and requires collective action by the colleges.” Id., at 270. As a result, the NCAA decided to retain the National Opinion Research Center (NORC) to study the impact of television on live attendance, and to declare a moratorium on the televising of football games. A television committee was appointed to implement the decision and to develop an NCAA television plan for 1951. Id., at 277-278.

The committee’s 1951 plan provided that only one game a week could be telecast in each area, with a total blackout on 3 of the 10 Saturdays during the season. A team could appear on television only twice during a season. The plan also provided that the NORC would conduct a systematic study of the effects of the program on attendance. Id., at 279. The plan received the virtually unanimous support of the NCAA membership; only the University of Pennsylvania challenged it. Pennsylvania announced that it would televise all its home games. The council of the NCAA thereafter declared Pennsylvania a member in bad standing and the four institutions scheduled to play at Pennsylvania in 1951 refused to do so. Pennsylvania then reconsidered its decision and abided by the NCAA plan. Id., at 280-281.

During each of the succeeding five seasons, studies were made which tended to indicate that television had an adverse effect on attendance at college football games. During those years the NCAA continued to exercise complete control over the number of games that could be televised. Id., at 325-359.

From 1952 through 1977 the NCAA television committee followed essentially the same procedure for developing its television plans. It would first circulate a questionnaire to the membership and then use the responses as a basis for formulating a plan for the ensuing season. The plan was then submitted to a vote by means of a mail referendum. Once approved, the plan formed the basis for NCAA’s negotiations *91with the networks. Throughout this period the plans retained the essential purposes of the original plan. See 546 F. Supp., at 1283.4 Until 1977 the contracts were all for either 1- or 2-year terms: In 1977 the NCAA adopted “principles of negotiation” for the future and discontinued the practice of submitting each plan for membership approval. Then the NCAA also entered into its first 4-year contract granting exclusive rights to the American Broadcasting Cos. (ABC) for the 1978-1981 seasons. ABC had held the exclusive rights to network telecasts of NCAA football games since 1965. Id., at 1283-1284.

The Current Plan

The plan adopted in 1981 for the 1982-1985 seasons is at issue in this case.5 This plan, like each of its predecessors, recites that it is intended to reduce, insofar as possible, the adverse effects of live television upon football game attendance.6 It provides that “all forms of television of the football *92games of NCAA member institutions during the Plan control periods shall be in accordance with this Plan.” App. 35. The plan recites that the television committee has awarded rights to negotiate and contract for the telecasting of college football games of members of the NCAA to two “carrying networks.” Id., at 36. In addition to the principal award of rights to the carrying networks, the plan also describes rights for a “supplementary series” that had been awarded for the 1982 and 1983 seasons,7 as well as a procedure for permitting specific “exception telecasts.”8

In separate agreements with each of the carrying networks, ABC and the Columbia Broadcasting System (CBS), the NCAA granted each the right to telecast the 14 live “exposures” described in the plan, in accordance with the “ground rules” set forth therein.9 Each of the networks agreed to pay a specified “minimum aggregate compensation *93to the participating NCAA member institutions” during the 4-year period in an amount that totaled $131,750,000. In essence the agreement authorized each network to negotiate directly with member schools for the right to televise their games. The agreement itself does not describe the method of computing the compensation for each game, but the practice that has developed over the years and that the District Court found would be followed under the current agreement involved the setting of a recommended fee by a representative of the NCAA for different types of telecasts, with national telecasts being the most valuable, regional telecasts being less valuable, and Division II or Division III games commanding a still lower price.10 The aggregate of all these payments presumably equals the total minimum aggregate compensation set forth in the basic agreement. Except for differences in payment between national and regional telecasts, and with respect to Division II and Division III games, the amount that any team receives does not change with the size of the viewing audience, the number of markets in which the game is telecast, or the particular characteristic of the game or the participating teams. Instead, the “ground rules” provide that the carrying networks make alternate selections of those games they wish to televise, and thereby obtain the exclusive right to submit a bid at an essentially fixed price to the institutions involved. See 546 F. Supp., at 1289-1293.11

*94The plan also contains “appearance requirements” and “appearance limitations” which pertain to each of the 2-year periods that the plan is in effect. The basic requirement imposed on each of the two networks is that it must schedule appearances for at least 82 different member institutions during each 2-year period. Under the appearance limitations no member institution is eligible to appear on television more than a total of six times and more than four times nationally, with the appearances to be divided equally between the two carrying networks. See id., at 1293. The number of exposures specified in the contracts also sets an absolute maximum on the number of games that can be broadcast.

Thus, although the current plan is more elaborate than any of its predecessors, it retains the essential features of each of them. It limits the total amount of televised intercollegiate football and the number of games that any one team may televise. No member is permitted to make any sale of television rights except in accordance with the basic plan.

Background of this Controversy

Beginning in 1979 CFA members began to advocate that colleges with major football programs should have a greater voice in the formulation of football television policy than they had in the NCAA. CFA therefore investigated the possibility of negotiating a television agreement of its own, devel*95oped an independent plan, and obtained a contract offer from the National Broadcasting Co. (NBC). This contract, which it signed in August 1981, would have allowed a more liberal number of appearances for each institution, and would have increased the overall revenues realized by CFA members. See id., at 1286.

In response the NCAA publicly announced that it would take disciplinary action against any CFA member that complied with the CFA-NBC contract. The NCAA made it clear that sanctions would not be limited to the football programs of CFA members, but would apply to other sports as well. On September 8, 1981, respondents commenced this action in the United States District Court for the Western District of Oklahoma and obtained a preliminary injunction preventing the NCAA from initiating disciplinary proceedings or otherwise interfering with CFA’s efforts to perform its agreement with NBC. Notwithstanding the entry of the injunction, most CFA members were unwilling to commit themselves to the new contractual arrangement with NBC in the face of the theatened sanctions and therefore the agreement was never consummated. See id., at 1286-1287.

Decision of the District Court

After a full trial, the District Court held that the controls exercised by the NCAA over the televising of college football games violated the Sherman Act. The District Court defined the relevant market as “live college football television” because it found that alternative programming has a significantly different and lesser audience appeal. Id., at 1297-1300.12 The District Court then concluded that the NCAA *96controls over college football are those of a “classic cartel” with an

“almost absolute control over the supply of college football which is made available to the networks, to television advertisers, and ultimately to the viewing public. Like all other cartels, NCAA members have sought and achieved a price for their product which is, in most instances, artificially high. The NCAA cartel imposes production limits on its members, and maintains mechanisms for punishing cartel members who seek to stray from these production quotas. The cartel has established a uniform price for the products of each of the member producers, with no regard for the differing quality of these products or the consumer demand for these various products.” Id., at 1300-1301.

The District Court found that competition in the relevant market had been restrained in three ways: (1) NCAA fixed the price for particular telecasts; (2) its exclusive network contracts were tantamount to a group boycott of all other potential broadcasters and its threat of sanctions against its own members constituted a threatened boycott of potential competitors; and (3) its plan placed an artificial limit on the production of televised college football. Id., at 1293-1295.

In the District Court the NCAA offered, two principal justifications for its television policies: that they protected the gate attendance of its members and that they tended to preserve a competitive balance among the football programs of the various schools. The District Court rejected the first justification because the evidence did not support the claim that college football television adversely affected gate attendance. Id., at 1295-1296. With respect to the “competitive balance” argument, the District Court found that the evidence failed to show that the NCAA regulations on matters such as recruitment and the standards for preserving amateurism were not sufficient to maintain an appropriate balance. Id., at 1296.

*97 Decision of the Court of Appeals

The Court of Appeals held that the NCAA television plan constituted illegal per se price fixing, 707 F. 2d, at 1152.13 It rejected each of the three arguments advanced by NCAA to establish the procompetitive character of its plan.14 First, the court rejected the argument that the television plan promoted live attendance, noting that since the plan involved a concomitant reduction in viewership the plan did not result in a net increase in output and hence was not procompetitive. Id., at 1153-1154. Second, the Court of Appeals rejected as illegitimate the NCAA’s purpose of promoting athletically balanced competition. It held that such a consideration amounted to an argument that “competition will destroy the market” — a position inconsistent with the policy of the Sherman Act. Moreover, assuming arguendo that the justification was legitimate, the court agreed with the District Court’s finding “that any contribution the plan made to athletic balance could be achieved by less restrictive means.” Id., at 1154. Third, the Court of Appeals refused to view the NCAA plan as competitively justified by the need to compete effectively with other types of television programming, since it entirely eliminated competition between producers of football and hence was illegal per se. Id., at 1155-1156.

Finally, the Court of Appeals concluded that even if the television plan were not per se illegal, its anticompetitive limitation on price and output was not offset by any *98procompetitive justification sufficient to save the plan even when the totality of the circumstances was examined. Id., at 1157-1160.16 The case was remanded to the District Court for an appropriate modification in its injunctive decree. Id., at 1162.16

II

There can be no doubt that the challenged practices of the NCAA constitute a “restraint of trade” in the sense that they limit members’ freedom to negotiate and enter into their own television contracts. In that sense, however, every contract is a restraint of trade, and as we have repeatedly recognized, the Sherman Act was intended to prohibit only unreasonable restraints of trade.17

*99It is also undeniable that these practices share characteristics of restraints we have previously held unreasonable. The NCAA is an association of schools which compete against each other to attract television revenues, not to mention fans and athletes. As the District Court found, the policies of the NCAA with respect to television rights are ultimately controlled by the vote of member institutions. 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.18 A restraint of this type has often been held to be unreasonable as a matter of law. Because it places a ceiling on the number of games member institutions may televise, the horizontal agreement places an artificial limit on the quantity of televised football that is available to broadcasters and consumers. By restraining the quantity of television rights available for sale, the challenged practices create a limitation on output; our cases have held that such limitations are unreasonable restraints of trade.19 Moreover, the District Court found that the minimum aggregate price in fact operates to preclude any price negotiation between broadcasters and institutions, *100thereby constituting horizontal price fixing, perhaps the paradigm of an unreasonable restraint of trade.20

Horizontal price fixing and output limitation are ordinarily condemned as a matter of law under an “illegal per se” approach because the probability that these practices are anti-competitive is so high; a per se rule is applied when “the practice facially appears to be one that would always or almost always tend to restrict competition and decrease output.” Broadcast Music, Inc. v. Columbia Broadcasting System, Inc., 441 U. S. 1, 19-20 (1979). In such circumstances a restraint is presumed unreasonable without inquiry into the particular market context in which it is found. Nevertheless, we have decided that it would be inappropriate to apply a per se rule to this case. This decision is not based on a lack of judicial experience with this type of arrangement,21 on the fact that the NCAA is organized as a nonprofit entity,22 or on *101our respect for the NCAA’s historic role in the preservation and encouragement of intercollegiate amateur athletics.23 Rather, what is critical is that this case involves an industry-in which horizontal restraints on competition are essential if the product is to be available at all.

As Judge Bork has noted: “[S]ome activities can only be carried out jointly. Perhaps the leading example is league sports. When a league of professional lacrosse teams is formed, it would be pointless to declare their cooperation illegal on the ground that there are no other professional lacrosse teams.” R. Bork, The Antitrust Paradox 278 (1978). What the NCAA and its member institutions market in this case is competition itself — contests between competing institutions. Of course, this would be completely ineffective if there were no rules on which the competitors agreed to create and define the competition to be marketed. A myriad of rules affecting such matters as the size of the field, the number of players on a team, and the extent to which physical violence is to be encouraged or proscribed, all must be agreed upon, and all restrain the manner in which institutions compete. Moreover, the NCAA seeks to market a particular brand of football — college football. The identification of this “product” with an academic tradition differentiates *102college football from and makes it more popular than professional sports to which it might otherwise be comparable, such as, for example, minor league baseball. In order to preserve the character and quality of the “product,” athletes must not be paid, must be required to attend class, and the like. And the integrity of the “product” cannot be preserved except by mutual agreement; if an institution adopted such restrictions unilaterally, its effectiveness as a competitor on the playing field might soon be destroyed. Thus, the NCAA plays a vital role in enabling college football to preserve its character, and as a result enables a product to be marketed which might otherwise be unavailable. In performing this role, its actions widen consumer choice — not only the choices available to sports fans but also those available to athletes — and hence can be viewed as procompetitive.24

*103 Broadcast Music squarely holds that a joint selling arrangement may be so efficient that it will increase sellers’ aggregate output and thus be procompetitive. See 441 U. S., at 18-23. Similarly, as we indicated in Continental T. V., Inc. v. GTE Sylvania Inc., 433 U. S. 36, 51-57 (1977), a restraint in a limited aspect of a market may actually enhance marketwide competition. Respondents concede that the great majority of the NCAA’s regulations enhance competition among member institutions. Thus, despite the fact that this case involves restraints on the ability of member institutions to compete in terms of price and output, a fair evaluation of their competitive character requires consideration of the NCAA’s justifications for the restraints.

Our analysis of this case under the Rule of Reason, of course, does not change the ultimate focus of our inquiry. Both per se rules and the Rule of Reason are employed “to form a judgment about the competitive significance of the restraint.” National Society of Professional Engineers v. United States, 435 U. S. 679, 692 (1978). A conclusion that a restraint of trade is unreasonable may be

“based either (1) on the nature or character of the contracts, or (2) on surrounding circumstances giving rise to the inference or presumption that they were intended to restrain trade and enhance prices. Under either branch of the test, the inquiry is confined to a consideration of impact on competitive conditions.” Id., at 690 (footnotes omitted).

Per se rules are invoked when surrounding circumstances make the likelihood of anticompetitive conduct so great as to *104render unjustified further examination of the challenged conduct.25 But whether the ultimate finding is the product of a presumption or actual market analysis, the essential inquiry remains the same — whether or not the challenged restraint enhances competition.26 Under the Sherman Act the criterion to be used in judging the validity of a restraint on trade is its impact on competition.27

I — ( 1 — I I — I

Because it restrains price and output, the NCAA’s television plan has a significant potential for anticompetitive effects.28 The findings of the District Court indicate that this *105potential has been realized. The District Court found that if member institutions were free to sell television rights, many-more games would be shown on television, and that the NCAA’s output restriction has the effect of raising the price the networks pay for television rights.29 Moreover, the *106court found that by fixing a price for television rights to all games, the NCAA creates a price structure that is unresponsive to viewer demand and unrelated to the prices that would prevail in a competitive market.30 And, of course, since as a practical matter all member institutions need NCAA approval, members have no real choice but to adhere to the NCAA’s televison controls.31

The anticompetitive consequences of this arrangement are apparent. Individual competitors lose their freedom to comp*107ete.32 Price is higher and output lower than they would otherwise be, and both are unresponsive to consumer preference.33 This latter point is perhaps the most significant, since “Congress designed the Sherman Act as a ‘consumer welfare prescription.’” Reiter v. Sonotone Corp., 442 U. S. 330, 343 (1979). 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.34 Restrictions on price and output are the paradigmatic examples of restraints of trade that the Sherman *108Act was intended to prohibit. See Standard Oil Co. v. United States, 221 U. S. 1, 52-60 (1911).35 At the same time, the television plan eliminates competitors from the market, since only those broadcasters able to bid on television rights covering the entire NCAA can compete.36 Thus, as the District Court found, many telecasts that would occur in a competitive market are foreclosed by the NCAA’s plan.37

*109Petitioner argues, however, that its television plan can have no significant anticompetitive effect since the record indicates that it has no market power — no ability to alter the interaction of supply and demand in the market.38 We must reject this argument for two reasons, one legal, one factual.

As a matter of law, the absence of proof of market power does not justify a naked restriction on price or output. To the contrary, when there is an agreement not to compete in terms of price or output, “no elaborate industry analysis is required to demonstrate the anticompetitive character of such an agreement.” Professional Engineers, 435 U. S., at 692.39 Petitioner does not quarrel with the District Court’s *110finding that price and output are not responsive to demand. Thus the plan is inconsistent with the Sherman Act’s command that price and supply be responsive to consumer preference.40 We have never required proof of market power in such a case.41 This naked restraint on price and output requires some competitive justification even in the absence of a detailed market analysis.42

*111As a factual matter, it is evident that petitioner does possess market power. The District Court employed the correct test for determining whether college football broadcasts constitute a separate market — whether there are other products that are reasonably substitutable for televised NCAA football games.43 Petitioner’s argument that it cannot obtain supracompetitive prices from broadcasters since advertisers, and hence broadcasters, can switch from college football to other types of programming simply ignores the findings of the District Court. It found that intercollegiate football telecasts generate an audience uniquely attractive to advertisers and that competitors are unable to offer programming that can attract a similar audience.44 These findings amply support its conclusion that the NCAA possesses market power.45 Indeed, the District Court’s subsidiary finding that advertisers will pay a premium price per viewer to reach audiences watching college football because of their demographic characteristics46 is vivid evidence of the uniqueness of this product.47 Moreover, the District Court’s market *112analysis is firmly supported by our decision in International Boxing Club of New York, Inc. v. United States, 358 U. S. 242 (1959), that championship boxing events are uniquely attractive to fans48 and hence constitute a market separate from that for nonchampionship events. See id., at 249-252.49 Thus, respondents have demonstrated that there is a separate market for telecasts of college football which “rest[s] on generic qualities differentiating” viewers. Times-Picayune Publishing Co. v. United States, 345 U. S. 594, 613 (1953). It inexorably follows that if college football broadcasts be defined as a separate market — and we are convinced they are — then the NCAA’s complete control over those broadcasts provides a solid basis for the District Court’s conclusion that the NCAA possesses market power with respect to those broadcasts. “When a product is controlled by one interest, without substitutes available in the market, there is monopoly power.” United States v. E. I. du Pont de Nemours & Co., 351 U. S. 377, 394 (1956).50

*113Thus, the NCAA television plan on its face constitutes a restraint upon the operation of a free market, and the findings of the District Court establish that it has operated to raise prices and reduce output. Under the Rule of Reason, these hallmarks of anticompetitive behavior place upon petitioner a heavy burden of establishing an affirmative defense which competitively justifies this apparent deviation from the operations of a free market. See Professional Engineers, 435 U. S., at 692-696. We turn now to the NCAA’s proffered justifications.

IV

Relying on Broadcast Music, petitioner argues that its television plan constitutes a cooperative “joint venture” which assists in the marketing of broadcast rights and hence is procompetitive. While joint ventures have no immunity from the antitrust laws,51 as Broadcast Music indicates, a joint selling arrangement may “mak[e] possible a new product by reaping otherwise unattainable efficiencies.” Arizona v. Maricopa County Medical Society, 457 U. S. 332, 365 (1982) (Powell, J., dissenting) (footnote omitted). The essential contribution made by the NCAA’s arrangement is to define the number of games that may be televised, to establish the price for each exposure, and to define the basic terms of each contract between the network and a home team. The NCAA does not, however, act as a selling agent for any school or for any conference of schools. The selection of individual games, and the negotiation of particular agreements, are matters left to the networks and the individual schools. Thus, the effect of the network plan is not to eliminate individual sales of broadcasts, since these still occur, albeit subject to fixed prices and output limitations. Unlike Broadcast Music’s blanket license covering broadcast rights *114to a large number of individual compositions, here the same rights are still sold on an individual basis, only in a noncompetitive market.

The District Court did not find that the NCAA’s television plan produced any procompetitive efficiencies which enhanced the competitiveness of college football television rights; to the contrary it concluded that NCAA football could be marketed just as effectively without the television plan.52 There is therefore no predicate in the findings for petitioner’s efficiency justification. Indeed, petitioner’s argument is refuted by the District Court’s finding concerning price and output. If the NCAA’s television plan produced pro-competitive efficiencies, the plan would increase output and reduce the price of televised games. The District Court’s contrary findings accordingly undermine petitioner’s position. In light of these findings, it cannot be said that “the agreement on price is necessary to market the product at all.” Broadcast Music, 441 U. S., at 23.53 In Broadcast Music, the availability of a package product that no individual could offer enhanced the total volume of music that was sold. Unlike this case, there was no limit of any kind placed on the volume that might be sold in the entire market and each individual remained free to sell his own music without restraint. Here production has been limited, not enhanced.54 *115No individual school is free to televise its own games without restraint. The NCAA’s efficiency justification is not supported by the record.

Neither is the NCAA’s television plan necessary to enable the NCAA to penetrate the market through an attractive package sale. Since broadcasting rights to college football constitute a unique product for which there is no ready substitute, there is no need for collective action in order to enable the product to compete against its nonexistent competitors.55 This is borne out by the District Court’s finding that the NCAA’s television plan reduces the volume of television rights sold.

V

Throughout the history of its regulation of intercollegiate football telecasts, the NCAA has indicated its concern with protecting live attendance. This concern, it should be noted, is not with protecting live attendance at games which are shown on television; that type of interest is not at issue in this case. Rather, the concern is that fan interest in a televised game may adversely affect ticket sales for games that will not appear on television.56

Although the NORC studies in the 1950’s provided some support for the thesis that live attendance would suffer if *116unlimited television were permitted,57 the District Court found that there was no evidence to support that theory in today’s market.58 Moreover, as the District Court found, the television plan has evolved in a manner inconsistent with its original design to protect gate attendance. Under the current plan, games are shown on television during all hours that college football games are played. The plan simply does not protect live attendance by ensuring that games will not be shown on television at the same time as live events.59

There is, however, a more fundamental reason for rejecting this defense. The NCAA’s argument that its television plan is necessary to protect live attendance is not based on a desire to maintain the integrity of college football as a distinct and attractive product, but rather on a fear that the product will not prove sufficiently attractive to draw live attendance when faced with competition from televised games. At bottom the NCAA’s position is that ticket sales for most college games are unable to compete in a free market.60 The *117television plan protects ticket sales by limiting output—-just as any monopolist increases revenues by reducing output. By seeking to insulate live ticket sales from the full spectrum of competition because of its assumption that the product itself is insufficiently attractive to consumers, petitioner forwards a justification that is inconsistent with the basic policy of the Sherman Act. “[T]he Rule of Reason does not support a defense based on the assumption that competition itself is unreasonable.” Professional Engineers, 435 U. S., at 696.

VI

Petitioner argues that the interest in maintaining a competitive balance among amateur athletic teams is legitimate and important and that it justifies the regulations challenged in this case. We agree with the first part of the argument but not the second.

Our decision not to apply a per se rule to this case rests in large part on our recognition that a certain degree of cooperation is necessary if the type of competition that petitioner and its member institutions seek to market is to be preserved.61 It is reasonable to assume that most of the regulatory controls of the NCAA are justifiable means of fostering competition among amateur athletic teams and therefore procompetitive because they enhance public interest in intercollegiate athletics. The specific restraints on football telecasts that are challenged in this case do not, however, fit into the same mold as do rules defining the conditions of the contest, the eligibility of participants, or the manner in which members of a joint enterprise shall share the responsibilities and the benefits of the total venture.

The NCAA does not claim that its television plan has equalized or is intended to equalize competition within any *118one league.62 The plan is nationwide in scope and there is no single league or tournament in which all college football teams compete. There is no evidence of any intent to equalize the strength of teams in Division I-A with those in Division II or Division III, and not even a colorable basis for giving colleges that have no football program at all a voice in the management of the revenues generated by the football programs at other schools.63 The interest in maintaining a competitive balance that is asserted by the NCAA as a justification for regulating all television of intercollegiate football is not related to any neutral standard or to any readily identifiable group of competitors.

*119The television plan is not even arguably tailored to serve such an interest. It does not regulate the amount of money that any college may spend on its football program, nor the way in which the colleges may use the revenues that are generated by their football programs, whether derived from the sale of television rights, the sale of tickets, or the sale of concessions or program advertising.64 The plan simply imposes a restriction on one source of revenue that is more important to some colleges than to others. There is no evidence that this restriction produces any greater measure of equality throughout the NCAA than would a restriction on alumni donations, tuition rates, or any other revenue-producing activity. At the same time, as the District Court found, the NCAA imposes a variety of other restrictions designed to preserve amateurism which are much better tailored to the goal of competitive balance than is the television plan, and which are “clearly sufficient” to preserve competitive balance to the extent it is within the NCAA’s power to do so.65 And much more than speculation supported the District Court’s findings on this score. No other NCAA sport employs a similar plan, and in particular the court found that in the most closely analogous sport, college basketball, competitive balance has been maintained without resort to a restrictive television plan.66

Perhaps the most important reason for rejecting the argument that the interest in competitive balance is served by the television plan is the District Court’s unambiguous and well-supported finding that many more games would be televised in a free market than under the NCAA plan. The hypothesis that legitimates the maintenance of competitive balance as a procompetitive justification under the Rule of *120Reason is that equal competition will maximize consumer demand for the product.67 The finding that consumption will materially increase if the controls are removed is a compelling demonstration that they do not in fact serve any such legitimate purpose.68

VII

The NCAA plays a critical role in the maintenance, of a revered tradition of amateurism in college sports. There can be no question but that it needs ample latitude to play that role, or that the preservation of the student-athlete in higher education adds richness and diversity to intercollegiate athletics and is entirely consistent with the goals of the Sherman Act. But'consistent with the Sherman Act, the role of the NCAA must be to preserve a tradition that might otherwise die; rules that restrict output are hardly consistent with this role. Today we hold only that the record supports the District Court’s conclusion that by curtailing output and blunting the ability of member institutions to respond to consumer preference, the NCAA has restricted rather than enhanced the place of intercollegiate athletics in the Nation’s life. Accordingly, the judgment of the Court of Appeals is

Affirmed.

Justice White,

with whom Justice Rehnquist joins, dissenting.

The NCAA is an unincorporated, nonprofit, educational association whose membership includes almost 800 nonprofit public and private colleges and universities and more than *121100 nonprofit athletic conferences and other organizations. Formed in 1905 in response to a public outcry concerning abuses in intercollegiate athletics, the NCAA, through its annual convention, establishes policies and rules governing its members’ participation in college sports, conducts national championships, exerts control over some of the economic aspects of revenue-producing sports, and engages in some more-or-less commercial activities. See Note, Tackling Intercollegiate Athletics: An Antitrust Analysis, 87 Yale L. J. 655, 656-657 (1978). Although some of the NCAA’s activities, viewed in isolation, bear a resemblance to those undertaken by professional sports leagues and associations, the Court errs in treating intercollegiate athletics under the NCAA’s control as a purely commercial venture in which colleges and universities participate solely, or even primarily, in the pursuit of profits. Accordingly, I dissent.

I-H

“While it would be fanciful to suggest that colleges are not concerned about the profitability of their ventures, it is clear that other, non-commercial goals play a central role in their sports programs.” J. Weistart & C. Lowell, The Law of Sports § 5.12 (1979). The NCAA’s member institutions have designed their competitive athletic programs “to be a vital part of the educational system.” Constitution and Interpretations of the NCAA, Art. II, § 2(a) (1982-1983), reprinted in App. 216. Deviations from this goal, produced by a persistent and perhaps inevitable desire to “win at all costs,” have in the past led, and continue to lead, to a wide range of competitive excesses that prove harmful to students and institutions alike. See G. Hanford, Report to the American Council on Education, An Inquiry into the Need for and Feasibility of a National Study of Intercollegiate Athletics 74-76 (1974) (Hanford); Marco, The Place of Intercollegiate Athletics in Higher Education: The Responsibility of the Faculty, 31 J. Higher Educ. 422, 426 (1968). The fundamental policy *122underlying the NCAA’s regulatory program, therefore, is to minimize such deviations and “to maintain intercollegiate athletics as an integral part of the educational program and the athlete as an integral part of the student body and, by so doing, retain a clear line of demarcation between college athletics and professional sports.” Constitution and Interpretations of the NCAA, Art. II, §2(a), reprinted in App. 216. See 546 F. Supp. 1276, 1309 (WD Okla.1982).

The NCAA, in short, “exist[s] primarily to enhance the contribution made by amateur athletic competition to the process of higher education as distinguished from realizing maximum return on it as an entertainment commodity.” Association for Intercollegiate Athletics for Women v. NCAA, 558 F. Supp. 487, 494 (DC 1983), aff’d, 236 U. S. App. D. C. 311, 735 F. 2d 577 (1984). In pursuing this goal, the organization and its members seek to provide a public good — a viable system of amateur athletics — that most likely could not be provided in a perfectly competitive market. See Hennessey v. NCAA, 564 F. 2d 1136, 1153 (CA5 1977). “Without regulation, the desire of member institutions to remain athletically competitive would lead, them to engage in activities that deny amateurism to the public. No single institution could confidently enforce its own standards since it could not trust its competitors to do the same.” Note, Antitrust and Nonprofit Entities, 94 Harv. L. Rev. 802, 817-818 (1981). The history of intercollegiate athletics prior to the advent of the NCAA provides ample support for this conclusion. By mitigating what appears to be a clear failure of the free market to serve the ends and goals of higher education, the NCAA ensures the continued availability of a unique and valuable product, the very existence of which might well be threatened by unbridled competition in the economic sphere.

In pursuit of its fundamental goal and others related to it, the NCAA imposes numerous controls on intercollegiate athletic competition among its members, many of which “are similar to those which are summarily condemned when *123undertaken in a more traditional business setting.” • Weistart & Lowell, supra, §5.12.b. Thus, the NCAA has promulgated and enforced rules limiting both the compensation of student-athletes, see, e. g., Justice v. NCAA, 577 F. Supp. 356 (Ariz. 1983), and the number of coaches a school may hire for its football and basketball programs, see, e. g., Hennessey v. NCAA, supra; it also has prohibited athletes who formerly have been compensated for playing from participating in intercollegiate competition, see, e. g., Jones v. NCAA, 392 F. Supp. 295 (Mass. 1975), restricted the number of athletic scholarships its members may award, and established minimum academic standards for recipients of those scholarships; and it has pervasively regulated the recruitment process, student eligibility, practice schedules, squad size, the number of games played, and many other aspects of intercollegiate athletics. See 707 F. 2d 1147, 1153 (CA10 1983); 546 F. Supp., at 1309. One clear effect of most, if not all, of these regulations is to prevent institutions with competitively and economically successful programs from taking advantage of their success by expanding their programs, improving the quality of the product they offer, and increasing their sports revenues. Yet each of these regulations represents a desirable and legitimate attempt “to keep university athletics from becoming professionalized to the extent that profit making objectives would overshadow educational objectives.” Kupec v. Atlantic Coast Conference, 399 F. Supp. 1377, 1380 (MDNC 1975). Significantly, neither the Court of Appeals nor this Court questions the validity of these regulations under the Rule of Reason. See ante, at 100-102, 117; 707 F. 2d, at 1153.

Notwithstanding the contrary conclusion of the District Court, 546 F. Supp., at 1316, and the majority, ante, at 117,1 do not believe that the restraint under consideration in this case — the NCAA’s television plan — differs fundamentally for antitrust purposes from the other seemingly anticompet-itive aspects of the organization’s broader program of self-*124regulation. The television plan, like many of the NCAA’s actions, furthers several complementary ends. Specifically, the plan is designed

“to reduce, insofar as possible, the adverse effects of live television . . . upon football game attendance and, in turn, upon the athletic and related educational programs dependent upon the proceeds therefrom; to spread football television participation among as many colleges as practicable; to reflect properly the image of universities as educational institutions; to promote college football through the use of television, to advance the overall interests of intercollegiate athletics, and to provide college football television to the public to the extent compatible with these other objectives.” App. 35.

See also id., at 244, 323, 640, 651, 672. More generally, in my view, the television plan reflects the NCAA’s fundamental policy of preserving amateurism and integrating athletics and education. Nor does the District Court’s finding that the plan is intended to maximize television revenues, 546 F. Supp., at 1288-1289, 1315-1316, warrant any implication that the NCAA and its member institutions pursue this goal without regard to the organization’s stated policies.

Before addressing the infirmities in the Court’s opinion, I should state my understanding of what the Court holds. To do so, it is necessary first to restate the essentials of the NCAA’s television plan and to refer to the course of this case in the lower courts. Under the plan at issue, 4-year contracts were entered into with the American Broadcasting Cos. (ABC), Columbia Broadcasting System (CBS), and Turner Broadcasting System (Turner) after competitive bidding. Every fall, ABC and CBS were to present 14 exposures of college football and Turner would show 19 evening games. The overall price for each network was stated in the contracts. The networks select the games to be telecast and pay directly to the colleges involved what has developed to be *125a uniform fee for each game telecast. Unless within one of the exceptions, only the designated number of games may be broadcast, and no NCAA member may arrange for televising its games other than pursuant to the plan. Under this scheme, of course, NCAA members must compete against one another for television appearances, although this competition is limited somewhat by the fact that no college may appear on television more than six times in any 2-year period. In 1983, 242 games were televised, 89 network games and 153 under the exceptions provided in the television plan. In 1983, 173 schools appeared on television, 89 on network games and an additional 84 teams under the exceptions. Report of the 1983 NCAA Football Television Committee to the 78th Annual Convention of the NCAA 61-65 (1984).1

The District Court held that the plan constituted price fixing and output limitation illegal per se under § 1 of the Sherman Act; it also held that the scheme was an illegal group boycott, was monopolization forbidden by § 2, and was in any event an unreasonable restraint of trade. It then entered an injunction that for all practical purposes excluded the NCAA from interfering with or regulating its members’ arrangements for televising their football games. The Court of Appeals, while disagreeing with the boycott and monopolization holdings, otherwise upheld the District Court’s judgment that the television plan violated the Sherman Act, focusing almost entirely on the price-fixing and output-limiting aspects of the television plan. The Court of Appeals, however, differed with the District Court with respect to the injunction. After noting that the injunction vested exclusive control of television rights in the individual schools, the court stated that, “[w]hile we hold that the NCAA cannot *126lawfully maintain exclusive control of the rights, how far such rights may be commonly regulated involves speculation that should not be made on the record of the instant case.” 707 F. 2d, at 1162. The court expressly stated, for example, that the NCAA could prevent its members from telecasting games on Friday night in competition with high school games, ibid., emphasized that the disparity in revenue between schools could be reduced by “[a] properly drawn system of pass-over payments to ensure adequate athletic funding for schools that do not earn substantial television revenues,” id., at 1159, and indicated that it was not outlawing “membership-wide contracts] with opt-out and pass-over payment provisions, or blackout rules.” Id., at 1162. It nevertheless left the District Court’s injunction in full force and remanded the case for further proceedings in light of its opinion. Anticipating that the Court would grant certiorari, I stayed the judgment of the Court of Appeals. 463 U. S. 1311 (1983).

In affirming the Court of Appeals, the Court first holds that the television plan has sufficient redeeming virtues to escape condemnation as a per se violation of the Sherman Act, this because of the inherent characteristics of competitive athletics and the justifiable role of the NCAA in regulating college athletics. It nevertheless affirms the Court of Appeals’ judgment that the NCAA plan is an unreasonable restraint of trade because of what it deems to be the plan’s price-fixing and output-limiting aspects. As I shall explain, in reaching this result, the Court traps itself in commercial antitrust rhetoric and ideology and ignores the context in which the restraints have been imposed. But it is essential at this point to emphasize that neither the Court of Appeals nor this Court purports to hold that the NCAA may not (1) require its members who televise their games to pool and share the compensation received among themselves, with other schools, and with the NCAA; (2) limit the number of times any member may arrange to have its games shown on *127television; or (3) enforce reasonable blackout rules to avoid head-to-head competition for television audiences. As I shall demonstrate, the Court wisely and correctly does not condemn such regulations. What the Court does affirm is the Court of Appeals’ judgment that the NCAA may not limit the number of games that are broadcast on television and that it may not contract for an overall price that has the effect of setting the price for individual game broadcast rights.2 I disagree with the Court in these respects.

II

“In a competitive market,” the District Court observed, “each football-playing institution would be an independent seller of the right to telecast its football games. Each seller would be free to sell that right to any entity it chose,” and “for whatever price it could get.” 546 F. Supp., at 1318. Under the NCAA’s television plan, member institutions’ competitive freedom is restrained because, for the most part, television rights are bought and sold, not on a per-game basis, but as a package deal. With limited exceptions not particularly relevant to antitrust scrutiny of the plan, broadcasters wishing to televise college football must be willing and able to purchase a package of television rights without knowing in advance the particular games to which those rights apply. The real negotiations over price and terms take place between the broadcasters and the NCAA rather *128than between the broadcasters and individual schools. Knowing that some games will be worth more to them than others, the networks undoubtedly exercise whatever bargaining power they possess to ensure that the minimum aggregate compensation they agree to provide for the package bears some relation to the average value to them of the games they anticipate televising. Because some schools’ games contribute disproportionately to the total value of the package, see id., at 1293, the manner in which the minimum aggregate compensation is distributed among schools whose games are televised has given rise to a situation under which less prominent schools receive more in rights fees than they would receive in a competitive market and football powers like respondents receive less. Id., at 1315.

As I have said, the Court does not hold, nor did the Court of Appeals hold, that this redistributive effect alone would be sufficient to subject the television plan to condemnation under § 1 of the Sherman Act. Nor should it, for an agreement to share football revenues to a certain extent is an essential aspect of maintaining some balance of strength among competing colleges and of minimizing the tendency to professionalism in the dominant schools. Sharing with the NCAA itself is also a price legitimately exacted in exchange for the numerous benefits of membership in the NCAA, including its many-faceted efforts to maintain a system of competitive, amateur athletics. For the same reasons, limiting the number of television appearances by any college is an essential attribute of a balanced amateur athletic system. Even with shared television revenues, unlimited appearances by a few schools would inevitably give them an insuperable advantage over all others and in the end defeat any efforts to maintain a system of athletic competition among amateurs who measure up to college scholastic requirements.

The Court relies instead primarily on the District Court’s findings that (1) the television plan restricts output; and (2) the plan creates a noncompetitive price structure that is unresponsive to viewer demand. Ante, at 104-106. See, *129 e. g., 546 F. Supp., at 1318-1319. These findings notwithstanding, I am unconvinced that the television plan has a substantial anticompetitive effect.

First, it is not clear to me that the District Court employed the proper measure of output. I am not prepared to say that the District Court’s finding that “many more college football games would be televised” in the absence of the NCAA controls, id., at 1294, is clearly erroneous. To the extent that output is measured solely in terms of the number of televised games, I need not deny that it is reduced by the NCAA’s television plan. But this measure of output is not the proper one. The District Court found that eliminating the plan would reduce the number of games on network television and increase the number of games shown locally and regionally. Id., at 1307. It made no finding concerning the effect of the plan on total viewership, which is the more appropriate measure of output or, at least, of the claimed anticompetitive effects of the NCAA plan. This is the NCAA’s position, and it seems likely to me that the television plan, by increasing network coverage at the expense of local broadcasts, actually expands the total television audience for NCAA football. The NCAA would surely be an irrational “profit maximizer” if this were not the case. In the absence of a contrary finding by the District Court, I cannot conclude that respondents carried their burden of showing that the television plan has an adverse effect on output and is therefore anticompetitive.

Second, and even more important, I am unconvinced that respondents have proved that any reduction in the number of televised college football games brought about by the NCAA’s television plan has resulted in an anticompetitive increase in the price of television rights. The District Court found, of course, that “the networks are actually paying the large fees because the NCAA agrees to limit production. If the NCAA would not agree to limit production, the networks would not pay so large a fee.” Id., at 1294. Undoubtedly, this is true. But the market for television rights to college football competitions should not be equated to the markets *130for wheat or widgets. Reductions in output by monopolists in most product markets enable producers to exact a higher price for the same product. By restricting the number of games that can be televised, however, the NCAA creates a new product — exclusive television rights — that are more valuable to networks than the products that its individual members could market independently.

The television plan makes a certain number of games available for purchase by television networks and limits the incidence of head-to-head competition between football telecasts for the available viewers. Because competition is limited, the purchasing network can count on a larger share of the audience, which translates into greater advertising revenues and, accordingly, into larger payments per game to the televised teams. There is thus a relationship between the size of the rights payments and the value of the product being purchased by the networks; a network purchasing a series of games under the plan is willing to pay more than would one purchasing the same games in the absence of the plan since the plan enables the network to deliver a larger share of the available audience to advertisers and thus to increase its own revenues. In short, by focusing only on the price paid by the networks for television rights rather than on the nature and quality of the product delivered by the NCAA and its member institutions, the District Court, and this Court as well, may well have deemed anticompetitive a rise in price that more properly should be attributed to an increase in output, measured in terms of viewership.

Third, the District Court’s emphasis on the prices paid for particular games seems misdirected and erroneous as a matter of law. The distribution of the minimum aggregate fees among participants in the television plan is, of course, not’ wholly based on a competitive price structure that is responsive to viewer demand and is only partially related to the value those schools contribute to the total package the networks agree to buy. But as I have already indicated, see *131 supra, at 128, this “redistribution” of total television revenues is a wholly justifiable, even necessary, aspect of maintaining a system of truly competitive college teams. As long as the NCAA cannot artificially fix the price of the entire package and demand supercompetitive prices, this aspect of the plan should be of little concern.' And I find little, if anything, in the record to support the notion that the NCAA has power to extract from the television networks more than the broadcasting rights are worth in the marketplace.

Ill

Even if I were convinced that the District Court did not err in failing to look to total viewership, as opposed to the number of televised games, when measuring output and anti-competitive effect and in failing fully to consider whether the NCAA possesses power to fix the package price, as opposed to the distribution of that package price among participating teams, I would nevertheless hold that the television plan passes muster under the Rule of Reason. The NCAA argues strenuously that the plan and the network contracts “are part of a joint venture among many of the nation’s universities to create a product—high-quality college football— and offer that product in a way attractive to both fans in the stadiums and viewers on [television]. The cooperation in producing the product makes it more competitive against other [television] (and live) attractions.” Brief for Petitioner 15. The Court recognizes that, “[i]f the NCAA faced ‘interbrand’ competition from available substitutes, then certain forms of collective action might be appropriate in order to enhance its ability to compete.” Ante, at 115, n. 55. See Continental T. V., Inc. v. GTE Sylvania Inc., 433 U. S. 36, 54-57 (1977). It rejects the NCAA’s proffered pro-competitive justification, however, on the ground that college football is a unique product for which there are no available substitutes and “there is no need for collective action in *132order to enable the product to compete against its nonexistent competitors.” Ante, at 115 (footnote omitted). This proposition is singularly unpersuasive.

It is one thing to say that “NCAA football is a unique product,” 546 P. Supp., at 1299, that “intercollegiate football telecasts generate an audience uniquely attractive to advertisers and that competitors are unable to offer programming that can attract a similar audience.” Ante, at 111 (footnote omitted). See 707 F. 2d, at 1158-1159; 546 F. Supp., at 1298-1300. It is quite another, in my view, to say that maintenance or enhancement of the quality of NCAA football telecasts is unnecessary to enable those telecasts to compete effectively against other forms of entertainment. The NCAA has no monopoly power when competing against other types of entertainment. Should the quality of the NCAA’s product “deteriorate to any perceptible degree or should the cost of ‘using’ its product rise, some fans undoubtedly would turn to another form of entertainment.... Because of the broad possibilities for alternative forms of entertainment,” the NCAA “properly belongs in the broader ‘entertainment’ market rather than in . . . [a] narrower marke[t]” like sports or football. Grauer, Recognition of the National Football League as a Single Entity Under Section 1 of the Sherman Act: Implications of the Consumer Welfare Model, 82 Mich. L. Rev. 1, 34, n. 156 (1983). See National Football League v. North American Soccer League, 459 U. S. 1074, 1077 (1982) (Rehnquist, J., dissenting from the denial of certio-rari); R. Atwell, B. Grimes, & D. Lopiano, The Money Game 32-33 (1980); Hanford, at 67; J. Michener, Sports in America 208-209 (1976); Note, 87 Yale L. J., at 661, and n. 31.

The NCAA has suggested a number of plausible ways in which its television plan might enhance the ability of college football telecasts to compete against other forms of entertainment. Brief for Petitioner 22-25. Although the District Court did conclude that the plan is “not necessary for effective marketing of the product,” 546 F. Supp., at 1307, its *133finding was directed only at the question whether college football telecasts would continue in the absence of the plan. It made no explicit findings concerning the effect of the plan on viewership and thus did not reject the factual premise of the NCAA’s argument that the plan might enhance competition by increasing the market penetration of NCAA football. See also 707 F. 2d, at 1154-1156, 1160. The District Court’s finding that network coverage of NCAA football would likely decrease if the plan were struck down, 546 F. Supp., at 1307, in fact, strongly suggests the validity of the NCAA’s position. On the record now before the Court, therefore, I am not prepared to conclude that the restraints imposed by the NCAA’s television plan are “such as may suppress or even destroy competition” rather than “such as merely regulat[e] and perhaps thereby promot[e] competition.” Chicago Board of Trade v. United States, 246 U. S. 231, 238 (1918).

<1

Finally, I return to the point with which I began — the essentially noneconomic nature of the NCAA’s program of self-regulation. Like Judge Barrett, who dissented in the Court of Appeals, I believe that the lower courts “erred by subjugating the NCAA’s educational goals (and, incidentally, those which Oklahoma and Georgia insist must be maintained in any event) to the purely competitive commercialism of [an] ‘every school for itself’ approach to television contract bargaining.” 707 F. 2d, at 1168. Although the NCAA does not enjoy blanket immunity from the antitrust laws, cf. Goldfarb v. Virginia State Bar, 421 U. S. 773 (1975), it is important to remember that the Sherman Act “is aimed primarily at combinations having commercial objectives and is applied only to a very limited extent to organizations . . . which normally have other objectives.” Klor’s, Inc. v. Broadway-Hale Stores, Inc., 359 U. S. 207, 213, n. 7 (1959).

The fact that a restraint operates on nonprofit educational institutions as distinguished from business entities is as “rele*134vant in determining whether that particular restraint violates the Sherman Act” as is the fact that a restraint affects a profession rather than a business. Goldfarb v. Virginia State Bar, supra, at 788, n. 17. Cf. Community Communications Co. v. Boulder, 455 U. S. 40, 56, n. 20 (1982). The legitimate noneconomic goals of colleges and universities should not be ignored in analyzing restraints imposed by associations of such institutions on their members, and these noneconomic goals “may require that a particular practice, which could properly be viewed as a violation of the Sherman Act in another context, be treated differently.” Goldfarb v. Virginia State Bar, supra, at 788, n. 17. The Court of Appeals, like the District Court, flatly refused to consider what it termed “noneconomic” justifications advanced by the NCAA in support of the television plan. It was of the view that our decision in National Society of Professional Engineers v. United States, 435 U. S. 679 (1978), precludes reliance on noneconomic factors in assessing the reasonableness of the television plan. 707 F. 2d, at 1154; see Tr. of Oral Arg. 24-25. This view was mistaken, and I note that the Court does not in so many words repeat this error.

Professional Engineers did make clear that antitrust analysis usually turns on “competitive conditions” and “economic conceptions.” 435 U. S., at 690, and n. 16. Ordinarily, “the inquiry mandated by the Rule of Reason is whether the challenged agreement is one that promotes competition or one that suppresses competition.” Id., at 691. The purpose of antitrust analysis, the Court emphasized, “is to form a judgment about the competitive significance of the restraint; it is not to decide whether a policy favoring competition is in the public interest, or in the interest of the members of an industry.” Id., at 692. Broadly read, these statements suggest that noneconomic values like the promotion of amateurism and fundamental educational objectives could not save the television plan from condemnation under the Sherman Act. *135But these statements were made in response to “public interest” justifications proffered in defense of a ban on competitive bidding imposed by practitioners engaged in standard, profit-motivated commercial activities. The primarily non-economic values pursued by educational institutions differ fundamentally from the “overriding commercial purpose of [the] day-to-day activities” of engineers, lawyers, doctors, and businessmen, Gulland, Byrne, & Steinbach, Intercollegiate Athletics and Television Contracts: Beyond Economic Justifications in Antitrust Analysis of Agreements Among Colleges, 52 Ford. L. Rev. 717, 728 (1984), and neither Professional Engineers nor any other decision of this Court suggests that associations of nonprofit educational institutions must defend their self-regulatory restraints solely in terms of their competitive impact, without regard for the legitimate noneconomic values they promote.

When these values are factored into the balance, the NCAA’s television plan seems eminently reasonable. Most fundamentally, the plan fosters the goal of amateurism by spreading revenues among various schools and reducing the financial incentives toward professionalism. As the Court observes, the NCAA imposes a variety of restrictions perhaps better suited than the television plan for the preservation of amateurism. Ante, at 119. Although the NCAA does attempt vigorously to enforce these restrictions, the vast potential for abuse suggests that measures, like the television plan, designed to limit the rewards of professionalism are fully consistent with, and essential to the attainment of, the NCAA’s objectives. In short, “[t]he restraints upon Oklahoma and Georgia and other colleges and universities with excellent football programs insure that they confine those programs within the principles of amateurism so that intercollegiate athletics supplement, rather than inhibit, educational achievement.” 707 F. 2d, at 1167 (Barrett, J., dissenting). The collateral consequences of the spreading of *136regional and national appearances among a number of schools are many: the television plan, like the ban on compensating student-athletes, may well encourage students to choose their schools, at least in part, on the basis of educational quality by reducing the perceived economic element of the choice, see Note, 87 Yale L. J., at 676, n. 106; it helps ensure the economic viability of athletic programs at a wide variety of schools with weaker football teams; and it “promot[es] competitive football among many and varied amateur teams nationwide.” Gulland, Byrne, & Steinbach, supra, at 722 (footnote omitted). These important contributions, I believe, are sufficient to offset any minimal anticompetitive effects of the television plan.

For all of these reasons, I would reverse the judgment of the Court of Appeals. At the very least, the Court of Appeals should be directed to vacate the injunction of the District Court pending the further proceedings that will be necessary to amend the outstanding injunction to accommodate the substantial remaining authority of the NCAA to regulate the telecasting of its members’ football games.

Mr. Justice Clark

delivered the opinion of the Court.

This action for treble damages and injunctive relief, brought under § 4 of the Clayton Act,1 tests the application of the antitrust laws to the business of professional football. Petitioner Radovich, an all-pro guard formerly with the Detroit Lions, contends that the respondents 2 *447entered into a conspiracy to monopolize and control organized professional football in the United States, in violation of §§ 1 and 2 of the Sherman Act; 3 that part of the conspiracy was to destroy the All-America Conference, a competitive professional football league in which Radovich once played; and that pursuant to agreement, respondents boycotted Radovich and prevented him from becoming a player-coach in the Pacific Coast League. Petitioner alleges that respondents’ illegal conduct damaged him in the sum of $35,000, to be trebled as provided by the Act. The trial court, on respondents’ motion, dismissed the cause for lack of jurisdiction and failure to state a claim on which relief could be granted. The Court of Appeals affirmed, 231 F. 2d 620, on the basis of Federal Baseball Club v. National League, 259 U. S. 200 (1922), and Toolson v. New York Yankees, Inc., 346 U. S. 356 (1953), applying the baseball rule to all “team sports.” It further found that even if such application was erroneous and that United States v. International Boxing Club, 348 U. S. 236 (1955), applied, Radovich had not grounded his claim on conduct of respondents which was “calculated to prejudice the public or unreasonably restrain interstate commerce.” 231 F. 2d, at 623. We granted certiorari, 352 U. S. 818, in order to clarify the application of the Toolson doctrine and determine whether the business of football comes within the scope of the Sherman Act. For the reasons hereafter stated we conclude that Toolson and Federal Baseball do not con*448trol; that the respondents’ activities as alleged are within the coverage of the antitrust laws; and that the complaint states a cause of action thereunder.

I.

Since the complaint was dismissed its allegations must be taken by us as true. It is, therefore, important for us to consider what Radovich alleged. Concisely the complaint states that:

1. Radovich began his professional football career in 1938 when he signed with the Detroit Lions, a National League club. After four seasons of play he entered the Navy, returning to the Lions for the 1945 season. In 1946 he asked for a transfer to a National League club in Los Angeles because of the illness of his father. The Lions refused the transfer and Radovich broke his player contract by signing with and playing the 1946 and 1947 seasons for the Los Angeles Dons, a member of the All-America Conference.4 In 1948 the San Francisco Clippers, a member of the Pacific Coast League which was affiliated with but not a competitor of the National League, offered to employ Radovich as a player-coach. However, the National League advised that Radovich was black-listed and any affiliated club signing him would suffer severe penalties. The Clippers then refused to sign him in any position. This black-listing effectively prevented his employment in organized professional football in the United States.

2. The black-listing was the result of a conspiracy among the respondents to monopolize commerce in professional football among the States. The purpose of the conspiracy was to “control, regulate and dictate the terms upon which organized professional football shall be played throughout the United States” in violation of §§ 1 and 2 of the *449Sherman Act. It was part of the conspiracy to boycott the All-America Conference and its players with a view to its destruction and thus strengthen the monopolistic position of the National Football League.

3. As part of its football business, the respondent league and its member teams schedule football games in various metropolitan centers, including New York, Chicago, Philadelphia, and Los Angeles. Each team uses a standard player contract which prohibits a player from signing with another club without the consent of the club holding the player’s contract. These contracts are enforced by agreement of the clubs to black-list any player violating them and to visit severe penalties on recalcitrant member clubs. As a further “part of the business of professional football itself” and “directly tied in and connected” with its football exhibitions is the transmission of the games over radio and television into nearly every State of the Union. This is accomplished by contracts which produce a “significant portion of the gross receipts” and without which “the business of operating a professional football club would not be profitable.” The playing of the exhibitions themselves “is essential to the interstate transmission by broadcasting and television” and the actions of the respondents against Radovich were necessarily related to these interstate activities.

In the light of these allegations respondents raise two issues: They say the business of organized professional football was not intended by Congress to be included within the scope of the antitrust laws; and, if wrong in this contention, that the complaint does not state a cause of action upon which relief can be granted.

II.

Respondents’ contention, boiled down, is that agreements similar to those complained of here, which have for many years been used in organized baseball, have *450been held by this Court to be outside the scope of the antitrust laws.5 They point to Federal Baseball and Toolson, supra, both involving the business of professional baseball, asserting that professional football has embraced the same techniques which existed in baseball at the time of the former decision.6 They contend that stare decisis compels the same result here. True, the umbrella under which respondents hope to stand is not so large as that contended for in United States v. International Boxing Club, supra, nor in United States v. Shubert, 348 U. S. 222 (1955). There we were asked to extend Federal Baseball to boxing and the theater. Here respondents say that the contracts and sanctions which baseball and football find it necessary to impose have no counterpart in other businesses and that, therefore, they alone are outside the ambit of the Sherman Act. In Toolson we continued to hold the umbrella over baseball that was placed there some 31 years earlier by Federal Baseball. The Court did this because it was concluded that more harm would be done in overruling Federal Baseball than in upholding a ruling which at best was of dubious validity. Vast efforts had gone into the development and organization of baseball since that decision and enormous capital had been invested in reliance on its permanence. Congress had chosen to make no change.7 All this, combined with the flood of litigation that would follow its repudiation, the harass*451ment that would ensue, and the retroactive effect of such a decision, led the Court to the practical result that it should sustain the unequivocal line of authority reaching over many years.

The Court was careful to restrict Toolson’s coverage to baseball, following the judgment of Federal Baseball only so far as it “determines that Congress had no intention of including the business of baseball within the scope of the federal antitrust laws.” 346 U. S., at 357. The Court reiterated this in United States v. Shubert, supra, at 230, where it said, “In short, Toolson was a narrow application of the rule of stare decisis.” And again, in International Boxing Club, it added, “Toolson neither overruled Federal Baseball nor necessarily reaffirmed all that was said in Federal Baseball. . . . Toolson is not authority for exempting other businesses merely because of the circumstance that they are also based on the performance of local exhibitions.” 348 U. S., at 242. Furthermore, in discussing the impact of the Federal Baseball decision, the Court made the observation that that decision “could not be relied upon as a basis of exemption for other segments of the entertainment business, athletic or otherwise. . . . The controlling consideration in Federal Baseball . . . was . . . the degree of interstate activity involved in the particular business under review.” Id., at 242-243. It seems that this language would have made it clear that the Court intended to isolate these cases by limiting them to baseball, but since Toolson and Federal Baseball are still cited as controlling authority in antitrust actions involving other fields of business, we now specifically limit the rule there established to the facts there involved, i. e., the business of organized professional baseball. As long as the Congress continues to acquiesce we should adhere to — but not extend — the interpretation of the Act made in those cases. We did not extend them to boxing or the theater because we believed *452that the volume of interstate business in each — the rationale of Federal Baseball — was such that both activities were within the Act. Likewise, the volume of interstate business involved in organized professional football places it within the provisions of the Act.

If this ruling is unrealistic, inconsistent, or illogical, it is sufficient to answer, aside from the distinctions between the businesses,8 that were we considering the question of baseball for the first time upon a clean slate we would have no doubts. But Federal Baseball held the business of baseball outside the scope of the Act. No other business claiming the coverage of those cases has such an adjudication. We, therefore, conclude that the orderly way to eliminate error or discrimination, if any there be, is by legislation and not by court decision. Congressional processes are more accommodative, affording the whole industry hearings and an opportunity to assist in the formulation of new legislation. The resulting product is therefore more likely to protect the industry and the public alike. The whole scope of congressional action would be known long in advance and effective dates for the legislation could be set in the future without the injustices of retroactivity and surprise which might follow court action. Of course, the doctrine of Toolson and Federal Baseball must yield to any congressional action and continues only at its sufferance. This is not a new approach. See Davis v. Department of Labor, 317 U. S. 249, 255 (1942); 9 Compare Rutkin v. United States, 343 U. S. 130 (1952).

*453III.

We now turn to the sufficiency of the complaint. At the outset the allegations of the nature and extent of interstate commerce seem to be sufficient. In addition to the standard allegations, a specific claim is made that radio and television transmission is a significant, integral part of the respondents’ business, even to the extent of being the difference between a profit and a loss. Unlike International Boxing, the complaint alleges no definite percentage in this regard. However, the amount must be substantial and can easily be brought out in the proof. If substantial, as alleged, it alone is sufficient to meet the commerce requirements of the Act. See International Boxing, supra, at 241.

Likewise, we find the technical objections to the pleading without merit. The test as to sufficiency laid down by Mr. Justice Holmes in Hart v. B. F. Keith Vaudeville Exchange, 262 U. S. 271, 274 (1923), is whether “the claim is wholly frivolous.” While the complaint might have been more precise in its allegations concerning the purpose and effect of the conspiracy, “we are not prepared to say that nothing can be extracted from this bill that falls under the act of Congress . . . .” Id., at 274. See also United States v. Employing Plasterers Assn., 347 U. S. 186 (1954).

Petitioner’s claim need only be “tested under the Sherman Act’s general prohibition on unreasonable restraints of trade,” Times-Picayune Publishing Co. v. United States, 345 U. S. 594, 614 (1953), and meet the requirement that petitioner has thereby suffered injury. Congress has, by legislative fiat, determined that such prohibited activities are injurious to the public10 and has *454provided sanctions allowing private enforcement of the antitrust laws by an aggrieved party. These laws protect the victims of the forbidden practices as well as the public. Mandeville Island Farms, Inc. v. American Crystal Sugar Co., 334 U. S. 219, 236 (1948). Furthermore, Congress itself has placed the private antitrust litigant in a most favorable position through the enactment of § 5 of the Clayton Act.11 Emich Motors Corp. v. General Motors Corp., 340 U. S. 558 (1951). In the face of such a policy this Court should not add requirements to burden the private litigant beyond what is specifically set forth by Congress in those laws.

Respondents’ remaining contentions we believe to be lacking in merit.

We think that Radovich is entitled to an opportunity to prove his charges. Of course, we express no opinion as to whether or not respondents have, in fact, violated the antitrust laws, leaving that determination to the trial court after all the facts are in.

Reversed.

*455Mr. Justice Frankfurter,

dissenting.

The difficult problem in this case derives for me not out of the Sherman Law but in relation to the appropriate compulsion of stare decisis. It does not derive from the Sherman Law because the most conscientious probing of the text and the interstices of the Sherman Law fails to disclose that Congress, whose will we are enforcing, excluded baseball — the conditions under which that sport is carried on — from the scope of the Sherman Law but included football. I say this, fully aware that the Sherman Law’s applicability turns on the particular circumstances of activities pursued in trade and commerce among the several States. But whether the conduct of an enterprise is within or without the limits of the Sherman Law is, after all, a question for judicial determination, and conscious as I am of my limited competence in matters athletic, I have yet to hear of any consideration that led this Court to hold that “the business of providing public baseball games for profit between clubs of professional baseball players was not within the scope of the federal antitrust laws,” Toolson v. New York Yankees, 346 U. S. 356, 357, that is not equally applicable to football.

But considerations pertaining to stare decisis do raise a serious question for me. That principle is a vital ingredient of law, for it “embodies an important social policy.” Helvering v. Hallock, 309 U. S. 106, 119. It would disregard the principle for a judge stubbornly to persist in his views on a particular issue after the contrary had become part of the tissue of the law. Until then, full respect for stare decisis does not require a judge to forego his own convictions promptly after his brethren have rejected them.

The considerations that governed me two years ago in United States v. International Boxing Club, 348 U. S. 236, *456have not lost their force by reason of the authority that time gives to a single decision. And so I am confronted with the Toolson case, supra, which guides me to find the present situation within its scope, and the Boxing case, supra, which, while it looks the other way, left Toolson as a living authority. Respect for the doctrine of stare decisis does not yet require me to disrespect the views I expressed in the Boxing case.

I would affirm.

Mr. Justice Harlan,

with whom Mr. Justice Brennan joins,

dissenting.

What was foreshadowed by United States v. International Boxing Club, 348 U. S. 236, has now come to pass. The Court, in holding that professional football is subject to the antitrust laws, now says in effect that professional baseball is sui generis so far as those laws are concerned, and that therefore Federal Baseball Club v. National League, 259 U. S. 200, and Toolson v. New York Yankees, Inc., 346 U. S. 356, do not control football by reason of stare decisis. Since I am unable to distinguish football from baseball under the rationale of Federal Baseball and Toolson, and can find no basis for attributing to Congress a purpose to put baseball in a class by itself, I would adhere to the rule of stare decisis and affirm the judgment below.

If the situation resulting from the baseball decisions is to be changed-, I think it far better to leave it to be dealt with by Congress than for this Court to becloud the situation further, either by making untenable distinctions between baseball and other professional sports, or by discriminatory fiat in favor of baseball.

Justice Thomas

delivered the opinion of the Court.

From 1998 until 2002, petitioners Texaco Inc. and Shell Oil Co. collaborated in a joint venture, Equilon Enterprises, to refine and sell gasoline in the western United States under the original Texaco and Shell Oil brand names. Respondents, a class of Texaco and Shell Oil service station owners, allege that petitioners engaged in unlawful price fixing when Equilon set a single price for both Texaco and Shell Oil brand gasoline. We granted certiorari to determine whether it is per se illegal under § 1 of the Sherman Act, 15 U. S. C. § 1, for a lawful, economically integrated joint venture to set the prices at which the joint venture sells its products. We conclude that it is not, and accordingly we reverse the contrary judgment of the Court of Appeals.

I

Historically, Texaco and Shell Oil have competed with one another in the national and international oil and gasoline *4markets. Their business activities include refining crude oil into gasoline, as well as marketing gasoline to downstream purchasers, such as the service stations represented in respondents’ class action.

In 1998, Texaco and Shell Oil formed- a joint venture, Equilon, to consolidate their operations in the western United States, thereby ending competition between the two companies in the domestic refining and marketing of gasoline. Under the joint venture agreement, Texaco and Shell Oil agreed to pool their resources and share the risks of and profits from Equilon’s activities. Equilon’s board of directors would comprise representatives of Texaco and Shell Oil, and Equilon gasoline would be sold to downstream purchasers under the original Texaco and Shell Oil brand names. The formation of Equilon was approved by consent decree, subject to certain divestments and other modifications, by the Federal Trade Commission, see In re Shell Oil Co., 125 F. T. C. 769 (1998), as well as by the state attorneys general of California, Hawaii, Oregon, and Washington. Notably, the decrees imposed no restrictions on the pricing of Equilon gasoline.

After the joint venture began to operate, respondents brought suit in District Court, alleging that, by unifying gasoline prices under the two brands, petitioners had violated the per se rule against price fixing that this Court has long recognized under § 1 of the Sherman Act, ch. 647, 26 Stat. 209, as amended, 15 U. S. C. § 1. See, e. g., Catalano, Inc. v. Target Sales, Inc., 446 U. S. 643, 647 (1980) (per curiam). The District Court awarded summary judgment to Texaco and Shell Oil. It determined that the rule of reason, rather than a per se rule or the quick look doctrine, governs respondents’ claim, and that, by eschewing rule of reason analysis, respondents had failed to raise a triable issue of fact. The Ninth Circuit reversed, characterizing petitioners’ position as a request for an “exception to the per se prohibition on price-fixing,” and rejecting that request. Dagher v. *5 Saudi Refining, Inc., 369 F. 3d 1108, 1116 (2004). We consolidated Texaco’s and Shell Oil’s separate petitions and granted certiorari to determine the extent to which the per se rule against price fixing applies to an important and increasingly popular form of business organization, the joint venture. 545 U. S. 1138 (2005).

II

Section 1 of the Sherman Act prohibits “[e]very contract, combination in the form of trust or otherwise, or conspiracy, in restraint of trade or commerce among the several States.” 15 U. S. C. § 1. This Court has not taken a literal approach to this language, however. See, e. g., State Oil Co. v. Khan, 522 U. S. 3, 10 (1997) (“[T]his Court has long recognized that Congress intended to outlaw only unreasonable restraints” (emphasis added)). Instead, this Court presumptively applies rule of reason analysis, under which antitrust plaintiffs must demonstrate that a particular contract or combination is in fact unreasonable and anticompetitive before it will be found unlawful. See, e. g., id., at 10-19. Per se liability is reserved for only those agreements that are “so plainly anti-competitive that no elaborate study of the industry is needed to establish their illegality.” National Soc. of Professional Engineers v. United States, 435 U. S. 679, 692 (1978). Accordingly, “we have expressed reluctance to adopt per se rules . . . ‘where the economic impact of certain practices is not immediately obvious.’ ” State Oil, supra, at 10 (quoting FTC v. Indiana Federation of Dentists, 476 U. S. 447, 458-459 (1986)).

Price-fixing agreements between two or more competitors, otherwise known as horizontal price-fixing agreements, fall into the category of arrangements that are per se unlawful. See, e. g., Catalano, supra, at 647. These cases do not present such an agreement, however, because Texaco and Shell Oil did not compete with one another in the relevant market — namely, the sale of gasoline to service stations in the western United States — but instead participated in that *6market jointly through their investments in Equilon.1 In other words, the pricing policy challenged here amounts to little more than price setting by a single entity — albeit within the context of a joint venture — and not a pricing agreement between competing entities with respect to their competing products. Throughout Equilon’s existence, Texaco and Shell Oil shared in the profits of Equilon’s activities in their role as investors, not competitors. When “persons who would otherwise be competitors pool their capital and share the risks of loss as well as the opportunities for profit.. . such joint ventures [are] regarded as a single firm competing with other sellers in the market.” Arizona v. Maricopa County Medical Soc., 457 U. S. 332, 356 (1982). As such, though Equilon’s pricing policy may be price fixing in a literal sense, it is not price fixing in the antitrust sense. See Broadcast Music, Inc. v. Columbia Broadcasting System, Inc., 441 U. S. 1, 9 (1979) (“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”).

This conclusion is confirmed by respondents’ apparent concession that there would be no per se liability had Equilon simply chosen to sell its gasoline under a single brand. See Tr. of Oral Arg. 34. We see no reason to treat Equilon differently just because it chose to sell gasoline under two dis*7tinct brands at a single price. As a single entity, a joint venture, like any other firm, must have the discretion to determine the prices of the products that it sells, including the discretion to sell a product under two different brands at a single, unified price. If Equilon’s price unification policy is anticompetitive, then respondents should have challenged it pursuant to the rule of reason.2 But it would be inconsistent with this Court’s antitrust precedents to condemn the internal pricing decisions of a legitimate joint venture as per se unlawful.3

The court below reached the opposite conclusion by invoking the ancillary restraints doctrine. 369 F. 3d, at 1118-1124. That doctrine governs the validity of restrictions imposed by a legitimate business collaboration, such as a business association or joint venture, on nonventure activities. See, e. g., National Collegiate Athletic Assn. v. Board of Regents of Univ. of Okla., 468 U. S. 85, 113-115 (1984); Citizen Publishing Co. v. United States, 394 U. S. 131, 135-136 (1969). Under the doctrine, courts must determine whether the nonventure restriction is a naked restraint on trade, and thus invalid, or one that is ancillary to the legitimate and competitive purposes of the business association, and thus valid. We agree with petitioners that the ancillary restraints doctrine has no application here, where the business practice being challenged involves the core activity of the joint venture itself — namely, the pricing of the very *8goods produced and sold by Equilon. And even if we were to invoke the doctrine in these cases, Equilon’s pricing policy is clearly ancillary to the sale of its own products. Judge Fernandez, dissenting from the ruling of the court below, put it well:

“In this case, nothing more radical is afoot than the fact that an entity, which now owns all of the production, transportation, research, storage, sales and distribution facilities for engaging in the gasoline business, also prices its own products. It decided to price them the same, as any other entity could. What could be more integral to the running of a business than setting a price for its goods and services?” 369 F. 3d, at 1127.

See also Broadcast Music, supra, at 23 (“Joint ventures and other cooperative arrangements are ... not usually unlawful, at least not as price-fixing schemes, where the agreement on price is necessary to market the product at all”).

* * *

Because the pricing decisions of a legitimate joint venture do not fall within the narrow category of activity that is per se unlawful under § 1 of the Sherman Act, respondents’ antitrust claim cannot prevail. Accordingly, the judgment of the Court of Appeals is reversed.

It is so ordered.

Justice Alito took no part in the consideration or decision of these cases.

Justice Breyer

delivered the opinion of the Court.

The question in this ease arises at the intersection of the Nation’s labor and antitrust laws. A group of professional *234football players brought this antitrust suit against football club owners. The club owners had bargained with the players’ union over a wage issue until they reached impasse. The owners then had agreed among themselves (but not with the union) to implement the terms of their own last best bargaining offer. The question before us is whether federal labor laws shield such an agreement from antitrust attack. We believe that they do. This Court has previously found in the labor laws an implicit antitrust exemption that applies where needed to make the collective-bargaining process work. Like the Court of Appeals, we conclude that this need makes the exemption applicable in this case.

I

We can state the relevant facts briefly. In 1987, a collective-bargaining agreement between the National Football League (NFL or League), a group of football clubs, and the NFL Players Association, a labor union, expired. The NFL and the Players Association began to negotiate a new contract. In March 1989, during the negotiations, the NFL adopted Resolution G-2, a plan that would permit each club to establish a “developmental squad” of up to six rookie or “first-year” players who, as free agents, had failed to secure a position on a regular player roster. See App. 42. Squad members would play in practice games and sometimes in regular games as substitutes for injured players. Resolution G-2 provided that the club owners would pay all squad members the same weekly salary.

The next month, April, the NFL presented the developmental squad plan to the Players Association. The NFL proposed a squad player salary of $1,000 per week. The Players Association disagreed. It insisted that the club owners give developmental squad players benefits and protections similar to those provided regular players, and that they leave individual squad members free to negotiate their own salaries.

*235Two months later, in June, negotiations on the issue of developmental squad salaries reached an impasse. The NFL then unilaterally implemented the developmental squad program by distributing to the clubs a uniform contract that embodied the terms of Resolution G-2 and the $1,000 proposed weekly salary. The League advised club owners that paying developmental squad players more or less than $1,000 per week would result in disciplinary action, including the loss of draft choices.

In May 1990, 235 developmental squad players brought this antitrust suit against the League and its member clubs. The players claimed that their employers’ agreement to pay them a $1,000 weekly salary violated the Sherman Act. See 15 U. S. C. § 1 (forbidding agreements in restraint of trade). The Federal District Court denied the employers’ claim of exemption from the antitrust laws; it permitted the case to reach the jury; and it subsequently entered judgment on a jury treble-damages award that exceeded $30 million. The NFL and its member clubs appealed.

The Court of Appeals (by a split 2-to-l vote) reversed. The majority interpreted the labor laws as “waiving] antitrust liability for restraints on competition imposed through the collective-bargaining process, so long as such restraints operate primarily in a labor market characterized by collective bargaining.” 50 F. 3d 1041, 1056 (CADC 1995). The court held, consequently, that the club owners were immune from antitrust liability. We granted certiorari to review that determination. Although we do not interpret the exemption as broadly as did the Appeals Court, we nonetheless find the exemption applicable, and we affirm that court’s immunity conclusion.

II

The immunity before us rests upon what this Court has called the “nonstatutory” labor exemption from the antitrust laws. Connell Constr. Co. v. Plumbers, 421 U. S. 616, 622 (1975); see also Meat Cutters v. Jewel Tea Co., 381 U. S. 676 *236(1965); Mine Workers v. Pennington, 381 U. S. 657 (1965). The Court has implied this exemption from federal labor statutes, which set forth a national labor policy favoring free and private collective bargaining, see 29 U. S. C. § 151; Teamsters v. Oliver, 358 U. S. 283, 295 (1959); which require good-faith bargaining over wages, hours, and working conditions, see 29 U. S. C. §§ 158(a)(5), 158(d); NLRB v. Wooster Div. of Borg-Warner Corp., 356 U. S. 342, 348-349 (1958); and which delegate related rulemaking and interpretive authority to the National Labor Relations Board (Board), see 29 U. S. C. § 153; San Diego Building Trades Council v. Garmon, 359 U. S. 236, 242-245 (1959).

This implicit exemption reflects both history and logic. As a matter of history, Congress intended the labor statutes (from which the Court has implied the exemption) in part to adopt the views of dissenting Justices in Duplex Printing Press Co. v. Deering, 254 U. S. 443 (1921), which Justices had urged the Court to interpret broadly a different explicit “statutory” labor exemption that Congress earlier (in 1914) had written directly into the antitrust laws. Id., at 483-488 (Brandeis, J., joined by Holmes and Clarke, JJ., dissenting) (interpreting § 20 of the Clayton Act, 38 Stat. 738, 29 U. S. C. § 52); see also United States v. Hutcheson, 312 U. S. 219, 230-236 (1941) (discussing congressional reaction to Duplex). In the 1930’s, when it subsequently enacted the labor statutes, Congress, as in 1914, hoped to prevent judicial use of antitrust law to resolve labor disputes — a kind of dispute normally inappropriate for antitrust law resolution. See Jewel Tea, supra, at 700-709 (opinion of Goldberg, J.); Marine Cooks v. Panama S. S. Co., 362 U. S. 365, 370, n. 7 (1960); A. Cox, Law and the National Labor Policy 3-8 (1960); cf. Duplex, supra, at 485 (Brandeis, J., dissenting) (explicit “statutory” labor exemption reflected view that “Congress, not the judges, was the body which should declare what public policy in regard to the industrial struggle demands”). The implicit (“nonstatutory”) exemption interprets the labor statutes in *237accordance with this intent, namely, as limiting an antitrust court’s authority to determine, in the area of industrial conflict, what is or is not a “reasonable” practice. It thereby substitutes legislative and administrative labor-related determinations for judicial antitrust-related determinations as to the appropriate legál limits of industrial conflict. See Jewel Tea, supra, at 709-710.

As a matter of logic, it would be difficult, if not impossible, to require groups of employers and employees to bargain together, but at the same time to forbid them to make among themselves or with each other any of the competition-restricting agreements potentially necessary to make the process work or its results mutually acceptable. Thus, the implicit exemption recognizes that, to give effect to federal labor laws and policies and to allow meaningful collective bargaining to take place, some restraints on' competition imposed through the bargaining process must be shielded from antitrust sanctions. See Connell, supra, at 622 (federal labor law’s “goals” could “never” be achieved if ordinary anti-competitive effects of collective bargaining were held to violate the antitrust laws); Jewel Tea, supra, at 711 (national labor law scheme would be “virtually destroyed” by the routine imposition of antitrust penalties upon parties engaged in collective bargaining); Pennington, supra, at 665 (implicit exemption necessary to harmonize Sherman Act with “national policy ... of promoting ‘the peaceful settlement of industrial disputes by subjecting labor-management controversies to the mediatory influence of negotiation’ ”) (quoting Fibreboard Paper Products Corp. v. NLRB, 379 U. S. 203, 211 (1964)).

The petitioners and their supporters concede, as they must, the legal existence of the exemption we have described. They also concede that, where its application is necessary to make the statutorily authorized collective-bargaining process work as Congress intended, the exemption must apply both to employers and to employees. *238Accord, Volkswagenwerk Aktiengesellschaft v. Federal Maritime Comm’n, 390 U. S. 261, 287, n. 5 (1968) (Harlan, J., concurring); Jewel Tea, supra, at 729-782, 735 (opinion of Goldberg, J.); Brief for AFL-CIO as Amicus Curiae in Associated Gen. Contractors of Cal., Inc. v. Carpenters, O. T. 1981, No. 81-334, pp. 16-17; see also P. Areeda & H. Hovenkamp, Antitrust Law ¶ 229’d (1995 Supp.) (collecting recent Court of Appeals cases); cf. H. A. Artists & Associates, Inc. v. Actors’ Equity Assn., 451 U. S. 704, 717, n. 20 (1981) (explicit “statutory” exemption applies only to “bona fide labor organization^]”). Nor does the dissent take issue with these basic principles. See post, at 253-254. Consequently, the question before us is one of determining the exemption’s scope: Does it apply to an agreement among several employers bargaining together to implement after impasse the terms of their last best good-faith wage offer? We assume that such conduct, as practiced in this case, is unobjectionable as a matter of labor law and policy. On that assumption, we conclude that the exemption applies.

Labor law itself regulates directly, and considerably, the kind of behavior here at issue — the postimpasse imposition of a proposed employment term concerning a mandatory subject of bargaining. Both the Board and the courts have held that, after impasse, labor law permits employers unilaterally to implement changes in pre-existing conditions, but only insofar as the new terms meet carefully circumscribed conditions. For example, the new terms must be “reasonably comprehended” within the employer’s preimpasse proposals (typically the last rejected proposals), lest by imposing more or less favorable terms, the employer unfairly undermined the union’s status. Storer Communications, Inc., 294 N. L. R. B. 1056, 1090 (1989); Taft Broadcasting Co., 163 N. L. R. B. 475, 478 (1967), enf’d, 395 F. 2d 622 (CADC 1968); see also NLRB v. Katz, 369 U. S. 736, 745, and n. 12 (1962). The collective-bargaining proceeding itself must be free of *239any unfair labor practice, such as an employer’s failure to have bargained in good faith. See Akron Novelty Mfg. Co., 224 N. L. R. B. 998, 1002 (1976) (where employer has not bargained in good faith, it may not implement a term of employment); 1 R Hardin, The Developing Labor Law 697 (3d ed. 1992) (same). These regulations reflect the fact that impasse and an accompanying implementation of proposals constitute an integral part of the bargaining process. See Bonanno Linen Serv., Inc., 243 N. L. R. B. 1093, 1094 (1979) (describing use of impasse as a bargaining tactic), enf’d, 630 F. 2d 25 (CA1 1980), aff’d, 454 U. S. 404 (1982); Colorado-Ute Elec. Assn., 295 N. L. R. B. 607, 609 (1989), enf. denied on other grounds, 939 F. 2d 1392 (CA10 1991), cert. denied, 504 U. S. 955 (1992).

Although the case law we have cited focuses upon bargaining by a single employer, no one here has argued that labor law does, or should, treat multiemployer bargaining differently in this respect. Indeed, Board and court decisions suggest that the joint implementation of proposed terms after impasse is a familiar practice in the context of multiem-ployer bargaining. See, e. g., El Cerrito Mill & Lumber Co., 316 N. L. R. B. 1005 (1995); Paramount Liquor Co., 307 N. L. R. B. 676, 686 (1992); NKS Distributors, Inc., 304 N. L. R. B. 338, 340-341 (1991), rev’d, 50 F. 3d 18 (CA9 1995); Sage Development Co., 301 N. L. R. B. 1173, 1175 (1991); Walker Constr. Co., 297 N. L. R. B. 746, 748 (1990), enf’d, 928 F. 2d 695 (CA5 1991); Food Employers Council, Inc., 293 N. L. R. B. 333, 334, 345-346 (1989); Tile, Terazzo & Marble Contractors Assn., 287 N. L. R. B. 769, 772 (1987), enf’d, 935 F. 2d 1249 (CA11 1991), cert. denied, 502 U. S. 1031 (1992); Salinas Valley Ford Sales, Inc., 279 N. L. R. B. 679, 686, 690 (1986); Carlsen Porsche Audi, Inc., 266 N. L. R. B. 141, 152-153 (1983); Typographic Service Co., 238 N. L. R. B. 1565 (1978); United Fire Proof Warehouse Co. v. NLRB, 356 F. 2d 494, 498-499 (CA7 1966); Cuyamaca Meats, Inc. v. Butchers’ *240 and Food Employers’ Pension Trust Fund, 638 F. Supp. 885, 887 (SD Cal. 1986), aff'd, 827 F. 2d 491 (CA9 1987), cert. denied, 485 U. S. 1008 (1988). We proceed on that assumption.

Multiemployer bargaining itself is a well-established, important, pervasive method of collective bargaining, offering advantages to both management and labor. See Appendix, infra, p; 251 (multiemployer bargaining accounts for more than 40% of major collective-bargaining agreements, and is used in such industries as construction, transportation, retail trade, clothing manufacture, and real estate, as well as professional sports); NLRB v. Truck Drivers, 353 U. S. 87, 95 (1957) (Buffalo Linen) (Congress saw multiemployer bargaining as “a vital factor in the effectuation of the national policy of promoting labor peace through strengthened collective bargaining”); Charles D. Bonanno Linen Service, Inc. v. NLRB, 454 U. S. 404, 409, n. 3 (1982) (Bonanno Linen) (multiemployer bargaining benefits both management and labor, by saving bargaining resources, by encouraging development of industry-wide worker benefits programs that smaller employers could not otherwise afford, and by inhibiting employer competition at the workers’ expense); Brief for Respondent NLRB in Bonanno Linen, O. T. 1981, No. 80-931, p. 10, n. 7 (same); General Subcommittee on Labor, House Committee on Education and Labor, Multiemployer Association Bargaining and its Impact on the Collective Bargaining Process, 88th Cong., 2d Sess., 10-19, 32-33 (Comm. Print 1964) (same); see also C. Bonnett, Employers’ Associations in the United States: A Study of Typical Associations (1922) (history). The upshot is that the practice at issue here plays a significant role in a collective-bargaining process that itself constitutes.an important part of the Nation’s industrial relations system.

In these circumstances, to subject the practice to antitrust law is to require antitrust courts to answer a host of important practical questions about how collective bargaining over *241wages, hours, and working conditions is to proceed — the very result that the implicit labor exemption seeks to avoid. And it is to place in jeopardy some of the potentially beneficial labor-related effects that multiemployer bargaining can achieve. That is because unlike labor law, which sometimes welcomes anticompetitive agreements conducive to industrial harmony, antitrust law forbids all agreements among competitors (such as competing employers) that unreasonably lessen competition among or between them in virtually any respect whatsoever. See, e. g., Paramount Famous Lasky Corp. v. United States, 282 U. S. 30 (1930) (agreement to insert arbitration provisions in motion picture licensing contracts). Antitrust law also sometimes permits judges or juries to premise antitrust liability upon little more than uniform behavior among competitors, preceded by conversations implying that later uniformity might prove desirable, see, e. g., United States v. General Motors Corp., 384 U. S. 127, 142-143 (1966); United States v. Foley, 598 F. 2d 1323, 1331-1332 (CA4 1979), cert. denied, 444 U. S. 1043 (1980), or accompanied by other conduct that in context suggests that each competitor failed to make an independent decision, see, e. g., American Tobacco Co. v. United States, 328 U. S. 781, 809-810 (1946); United States v. Masonite Corp., 316 U. S. 265, 275 (1942); Interstate Circuit, Inc. v. United States, 306 U. S. 208, 226-227 (1939). See generally 6 P. Areeda, Antitrust Law ¶¶ 1416-1427 (1986); Turner, The Definition of Agreement Under the Sherman Act: Conscious Parallelism and Refusals to Deal, 75 Harv. L. Rev. 655 (1962).

If the antitrust laws apply, what are employers to do once impasse is reached? If all impose terms similar to their last joint offer, they invite an antitrust action premised upon identical behavior (along with prior or accompanying conversations) as tending to show a common understanding or agreement. If any, or all, of them individually impose terms that differ significantly from that offer, they invite an unfair *242labor practice charge. Indeed, how can employers safely discuss their offers together even before a bargaining impasse occurs? A preimpasse discussion about, say, the practical advantages or disadvantages of a particular proposal invites a later antitrust claim that they agreed to limit the kinds of action each would later take should an impasse occur. The same is true of postimpasse discussions aimed at renewed negotiations with the union. Nor would adherence to the terms of an expired collective-bargaining agreement eliminate a potentially plausible antitrust claim charging that they had “conspired’' or tacitly “agreed” to do so, particularly if maintaining the status quo were not in the immediate economic self-interest of some. Cf. Interstate Circuit, supra, at 222-223; 6 Areeda, supra, ¶ 1425. All this is to say that to permit antitrust liability here threatens to introduce instability and uncertainty into the collective-bargaining process, for antitrust law often forbids or discourages the kinds of joint discussions and behavior that the collective-bargaining process invites or requires.

We do not see any obvious answer to this problem. We recognize, as the Government suggests, that, in principle, antitrust courts might themselves try to evaluate particular kinds of employer understandings, finding them “reasonable” (hence lawful) where justified by collective-bargaining necessity. But any such evaluation means a web of detailed rules spun by many different nonexpert antitrust judges and juries, not a set of labor rules enforced by a single expert administrative body, namely the Board. The labor laws give the Board, not antitrust courts, primary responsibility for policing the collective-bargaining process. And one of their objectives was to take from antitrust courts the authority to determine, through application of the antitrust laws, what is socially or economically desirable collective-bargaining policy. See supra, at 236-237; see also Jewel Tea, 381 U. S., at 716-719 (opinion of Goldberg, J.).

*243III

Both petitioners and their supporters advance several suggestions for drawing the exemption boundary line short of this case. We shall explain why we find them unsatisfactory.

A

Petitioners claim that the implicit exemption applies only to labor-management agreements — a limitation that they deduce from case law language, see, e. g., Connell, 421 U. S., at 622 (exemption for “some union-employer agreements”) (emphasis added), and from a proposed principle — that the exemption must rest upon labor-management consent. The language, however, reflects only the fact that the cases previously before the Court involved collective-bargaining agreements, see id., at 619—620; Pennington, 381 U. S., at 660; Jewel Tea, supra, at 679-680; the language does not reflect the exemption’s rationale, see 50 F. 3d, at 1050.

Nor do we see how an exemption limited by petitioners’ principle of labor-management consent could work. One cannot mean the principle literally — that the exemption applies only to understandings embodied in a collective-bargaining agreement — for the collective-bargaining process may take place before the making of any agreement or after an agreement has expired. Yet a multiemployer bargaining process itself necessarily involves many procedural and substantive understandings among participating employers as well as with the union. Petitioners cannot rescue their principle by claiming that the exemption applies only insofar as both labor and management consent to those understandings. Often labor will not (and should not) consent to certain common bargaining positions that employers intend to maintain. Cf. Areeda & Hovenkamp, Antitrust Law ¶ 229’d, at 277 (“[J]oint employer preparation and bargaining in the context of a formal multi-employer bargaining unit is clearly exempt”). Similarly, labor need not consent to certain tactics that this Court has approved as part of the multiemployer *244bargaining process, such as unit-wide lockouts and the use of temporary replacements. See NLRB v. Brown, 380 U. S. 278, 284 (1965); Buffalo Linen, 353 U. S., at 97.

Petitioners cannot save their consent principle by weakening it, as by requiring union consent only to the multi-employer bargaining process itself. This general consent is automatically present whenever multiemployer bargaining takes place. See Hi-Way Billboards, Inc., 206 N. L. R. B. 22 (1973) (multiemployer unit “based on consent” and “established by an unequivocal agreement by the parties”), enf. denied on other grounds, 500 F. 2d 181 (CA5 1974); Weyerhaeuser Co., 166 N. L. R. B. 299, 299-300 (1967). As so weakened, the principle cannot help decide which related practices are, or are not, subject to antitrust immunity.

B

The Government argues that the exemption should terminate at the point of impasse. After impasse, it says, “employers no longer have a duty under the labor laws to maintain the status quo,” and “are free as a matter of labor law to negotiate individual arrangements on an interim basis with the union.” Brief for United States et al. as Amici Curiae 17.

Employers, however, are not completely free at impasse to act independently. The multiemployer bargaining unit ordinarily remains intact; individual employers cannot withdraw. Bonanno Linen, 454 U. S., at 410-413. The duty to bargain survives; employers must stand ready to resume collective bargaining. See, e.g., Worldwide Detective Bureau, 296 N. L. R. B. 148, 155 (1989); Hi-Way Billboards, Inc., supra, at 23. And individual employers can negotiate individual interim agreements with the union only insofar as those agreements are consistent with “the duty to abide by the results of group bargaining.” Bonanno Linen, supra, at 416. Regardless, the absence of a legal “duty” to act jointly is not determinative. This Court has implied antitrust immunities *245that extend beyond statutorily required joint action to joint action that a statute “expressly or impliedly allows or assumes must also be immune.” 1 P. Areeda & D. Turner, Antitrust Law ¶ 224, p. 145 (1978); see, e. g., Gordon v. New York Stock Exchange, Inc., 422 U. S. 659, 682-691 (1975) (immunizing application of joint rule that securities law permitted, but did not require); United States v. National Assn. of Securities Dealers, Inc., 422 U. S. 694, 720-730 (1975) (same).

More importantly, the simple “impasse” line would not solve the basic problem we have described above. Supra, at 241-242. Labor law permits employers, after impasse, to engage in considerable joint behavior, including joint lockouts and replacement hiring. See, e. g., Brown, supra, at 289 (hiring of temporary replacement workers after lockout was “reasonably adapted to the achievement of a legitimate end — preserving the integrity of the multiemployer bargaining unit”). Indeed, as a general matter, labor law often limits employers to four options at impasse: (1) maintain the status quo, (2) implement their last offer, (3) lock out their workers (and either shut down or hire temporary replacements), or (4) negotiate separate interim agreements with the union. See generally 1 Hardin, The Developing Labor Law, at 516-520, 696-699. What is to happen if the parties cannot reach an interim agreement? The other alternatives are limited. Uniform employer conduct is likely. Uniformity — at least when accompanied by discussion of the matter— invites antitrust attack. And such attack would ask antitrust courts to decide the lawfulness of activities intimately related to the bargaining process.

The problem is aggravated by the fact that “impasse” is often temporary, see Bonanno Linen, supra, at 412 (approving Board’s view of impasse as “a recurring feature in the bargaining process, ... a temporary deadlock or hiatus in negotiations which in almost all cases is eventually broken, through either a change of mind or the application of economic force”) (internal quotation marks omitted); W. Sim*246kin & N. Fidandis, Mediation and the Dynamics of Collective Bargaining 139-140 (2d ed. 1986); it may differ from bargaining only in degree, see 1 Hardin, supra, at 691-696; Taft Broadcasting Co., 163 N. L. R. B., at 478; it may be manipulated by the parties for bargaining purposes, see Bonanno Linen, supra, at 413, n. 8 (parties might, for strategic purposes, “precipitate an impasse”); and it may occur several times during the course of a single labor dispute, since the bargaining process is not over when the first impasse is reached, cf. J. Bartlett, Familiar Quotations 754:8 (16th ed. 1992). How are employers to discuss future bargaining positions during a temporary impasse? Consider, too, the adverse consequences that flow from failing to guess how an antitrust court would later draw the impasse line. Employers who erroneously concluded that impasse had not been reached would risk antitrust liability were they collectively to maintain the status quo, while employers who erroneously concluded that impasse had occurred would risk unfair labor practice charges for prematurely suspending multiemployer negotiations.

The United States responds with suggestions for softening an “impasse” rule by extending the exemption after impasse “for such time as would be reasonable in the circumstances” for employers to consult with counsel, confirm that impasse has occurred, and adjust their business operations, Brief for United States et al. as Amici Curiae 24; by reestablishing the exemption once there is a “resumption of good-faith bargaining,” id., at 18, n. 5; and by looking to antitrust law’s “rule of reason” to shield — “in some circumstances”— such joint actions as the unit-wide lockout or the concerted maintenance of previously established joint benefit or retirement plans, ibid. But even as so modified, the impasse-related rule creates an exemption that can evaporate in the middle of the bargaining process, leaving later antitrust courts free to second-guess the parties’ bargaining decisions *247and consequently forcing them to choose their collective-bargaining responses in light of what they predict or fear that antitrust courts, not labor law administrators, will eventually decide. Cf. Dallas General Drivers, Warehousemen and Helpers, Local Union No. 74 v. NLRB, 355 F. 2d 842, 844-845 (CADC 1966) (“The problem of deciding when further bargaining ... is futile is often difficult for the bargainers and is necessarily so for the Board. But in the whole complex of industrial relations few issues are less suited to appellate judicial appraisal... or better suited to the expert experience of a board which deals constantly with such problems”).

C

Petitioners and their supporters argue in the alternative for a rule that would exempt postimpasse agreement about bargaining “tactics,” but not postimpasse agreement about substantive “terms,” from the reach of antitrust. See 50 F. 3d, at 1066-1069 (Wald, J., dissenting). They recognize, however, that both the Board and the courts have said that employers can, and often do, employ the imposition of “terms” as a bargaining “tactic.” See, e. g., American Ship Building Co. v. NLRB, 380 U. S. 300, 316 (1965); Colorado-Ute Elec. Assn., Inc. v. NLRB, 939 F. 2d 1392, 1404 (CA10 1991), cert. denied, 504 U. S. 955 (1992); Circuit-Wise, Inc., 309 N. L. R. B. 905, 921 (1992); Hi-Way Billboards, 206 N. L. R. B., at 23; Bonanno Linen, 243 N. L. R. B., at 1094. This concession as to joint “tactical” implementation would turn the presence of an antitrust exemption upon a determination of the employers’ primary purpose or motive. See, e. g., 50 F. 3d, at 1069 (Wald, J., dissenting). But to ask antitrust courts, insulated from the bargaining process, to investigate an employer group’s subjective motive is to ask them to conduct an inquiry often more amorphous than those we have previously discussed. And, in our view, a labor/ antitrust line drawn on such a basis would too often raise *248the same related (previously discussed) problems. See supra, at 237, 241-242; Jewel Tea, 381 U. S., at 716 (opinion of Goldberg, J.) (expressing concern about antitrust judges “roaming at large” through the bargaining process).

D

Petitioners make several other arguments. They point, for example, to cases holding applicable, in collective-bargaining contexts, general “backdrop” statutes, such as a state statute requiring a plant-closing employer to make employee severance payments, Fort Halifax Packing Co. v. Coyne, 482 U. S. 1 (1987), and a state statute mandating certain minimum health benefits, Metropolitan Life Ins. Co. v. Massachusetts, 471 U. S. 724 (1985). Those statutes, however, “ ‘neither encourage[d] nor discourage[d] the collective-bargaining processes that are the subject of the [federal labor laws].’” Fort Halifax, supra, at 21 (quoting Metropolitan Life, supra, at 755). Neither did those statutes come accompanied with antitrust’s labor-related history. Cf. Oliver, 358 U. S., at 295-297 (state antitrust law interferes with collective bargaining and is not applicable to labor-management agreement).

Petitioners also say that irrespective of how the labor exemption applies elsewhere to multiemployer collective bargaining, professional sports is “special.” We can understand how professional sports may be special in terms of, say, interest, excitement, or concern. But we do not understand how they are special in respect to labor law’s antitrust exemption. We concede that the clubs that make up a professional sports league are not completely independent economic competitors, as they depend upon a degree of cooperation for economic survival. National Collegiate Athletic Assn. v. Board of Regents of Univ. of Okla., 468 U. S. 85, 101-102 (1984); App. 110-115 (declaration of NFL Commissioner). In the present context, however, that circumstance *249makes the league more like a single bargaining employer, which analogy seems irrelevant to the legal issue before us.

We also concede that football players often have special individual talents, and, unlike many unionized workers, they often negotiate their pay individually with their employers. See 'post, at 255 (Stevens, J., dissenting). But this characteristic seems simply a feature, like so many others, that might give employees (or employers) more (or less) bargaining power, that might lead some (or all) of them to favor a particular kind of bargaining, or that might lead to certain demands at the bargaining table. We do not see how it could make a critical legal difference in determining the underlying framework in which bargaining is to take place. See generally Jacobs & Winter, Antitrust Principles and Collective Bargaining by Athletes: Of Superstars in Peonage, 81 Yale L. J. 1 (1971). Indeed, it would be odd to fashion an antitrust exemption that gave additional advantages to professional football players (by virtue of their superior bargaining power) that transport workers, coal miners, or meat packers would not enjoy.

The dissent points to other “unique features” of the parties’ collective-bargaining relationship, which, in the dissent’s view, make the case “atypical.” Post, at 255. It says, for example, that the employers imposed the restraint simply to enforce compliance with league-wide rules, and that the bargaining consisted of nothing more than the sending of a “notice,” and therefore amounted only to “so-called” bargaining. Post, at 256-257. Insofar as these features underlie an argument for looking to the employers’ true purpose, we have already discussed them. See supra, at 247-248. Insofar as they suggest that there was not a genuine impasse, they fight the basic assumption upon which the District Court, the Court of Appeals, petitioners, and this Court rest the case. See 782 F. Supp. 125, 134 (DC 1991); 50 F. 3d, at 1056-1057; Pet. for Cert. i. Ultimately, we cannot find a satisfactory basis for distinguishing football players from *250other organized workers. We therefore conclude that all must abide by the same legal rules.

* * *

For these reasons, we hold that the implicit (“nonstatu-tory”) antitrust exemption applies to the employer conduct at issue here. That conduct took place during and immediately after a collective-bargaining negotiation. It grew out of, and was directly related to, the lawful operation of the bargaining process. It involved a matter that the parties were required to negotiate collectively. And it concerned only the parties to the collective-bargaining relationship.

Our holding is not intended to insulate from antitrust review every joint imposition of terms by employers, for an agreement among employers could be sufficiently distant in time and in circumstances from the collective-bargaining process that a rule permitting antitrust intervention would not significantly interfere with that process. See, e. g., 50 F. 3d, at 1057 (suggesting that exemption lasts until collapse of the collective-bargaining relationship, as evidenced by decer-tification of the union); El Cerrito Mill & Lumber Co., 316 N. L. R. B., at 1006-1007 (suggesting that “extremely long” impasse, accompanied by “instability” or “defunctness” of multiemployer unit, might justify union withdrawal from group bargaining). We need not decide in this case whether, or where, within these extreme outer boundaries to draw that line. Nor would it be appropriate for us to do so without the detailed views of the Board, to whose “specialized judgment” Congress “intended to leave” many of the “inevitable questions concerning multiemployer bargaining bound to arise in the future.” Buffalo Linen, 353 U. S., at 96 (internal quotation marks omitted); see also Jewel Tea, 381 U. S., at 710, n. 18.

The judgment of the Court of Appeals is affirmed.

It is so ordered.

*251APPENDIX TO OPINION OF THE COURT

*252Justice Stevens,

dissenting.

In his classic dissent in Lochner v. New York, 198 U. S. 45, 75 (1905), Justice Holmes reminded us that our disagreement with the economic theory embodied in legislation should not affect our judgment about its constitutionality. It is equally important, of course, to be faithful to the economic theory underlying broad statutory mandates when we are construing their impact on areas of the economy not specifically addressed by their texts. The unique features of this case lead me to conclude that the Court has reached a decision that conflicts with the basic purpose of both the antitrust laws and the national labor policy expressed in a series of congressional enactments.

I

The basic premise underlying the Sherman Act is the assumption that free competition among business entities will produce the best price levels. National Soc. of Professional Engineers v. United States, 435 U. S. 679, 695 (1978). Collusion among competitors, it is believed, may produce prices that harm consumers. United States v. Socony-Vacuum Oil Co., 310 U. S. 150, 226, n. 59 (1940). Similarly, the Court has held, a marketwide agreement among employers setting wages at levels that would not prevail in a free market may violate the Sherman Act. Anderson v. Shipowners Assn. of Pacific Coast, 272 U. S. 359 (1926).

The jury’s verdict in this case has determined that the marketwide agreement among these employers fixed the salaries of the replacement players at a dramatically lower level than would obtain in a free market. While the special characteristics of this industry may provide a justification for the agreement under the rule of reason, see National Collegiate Athletic Assn. v. Board of Regents of Univ. of Okla., 468 U. S. 85, 100-104 (1984), at this stage of the proceeding our analysis of the exemption issue must accept the premise that the agreement is unlawful unless it is exempt.

*253The basic premise underlying our national labor policy is that unregulated competition among employees and applicants for employment produces wage levels that are lower than they should be.1 Whether or not the premise is true in fact, it is surely the basis for the statutes that encourage and protect the collective-bargaining process, including the express statutory exemptions from the antitrust laws that Congress enacted in order to protect union activities.2 Those statutes were enacted to enable collective action by union members to achieve wage levels that are higher than would be available in a free market. See Trainmen v. Chicago R. & I. R. Co., 353 U. S. 30, 40 (1957).

The statutory labor exemption protects the right of workers to act collectively to seek better wages, but does not *254“exempt concerted action or agreements between unions and nonlabor parties.” Connell Constr. Co. v. Plumbers, 421 U. S. 616, 621-622 (1975). It is the judicially crafted, non-statutory labor exemption that serves to accommodate the conflicting policies of the antitrust and labor statutes in the context of action between employers and unions. Ibid.

The limited judicial exemption complements its statutory counterpart by ensuring that unions which engage in collective bargaining to enhance employees’ wages may enjoy the benefits of the resulting agreements. The purpose of the labor laws would be frustrated if it were illegal for employers to enter into industrywide agreements providing supra-competitive wages for employees. For that reason, we have explained that “a proper accommodation between the congressional policy favoring collective bargaining under the NLRA and the congressional policy favoring free competition in business markets requires that some union-employer agreements be accorded a limited nonstatutory exemption from antitrust sanctions.” Id., at 622.

Consistent with basic labor law policies, I agree with the Court that the judicially crafted labor exemption must also cover some collective action that employers take in response to a collective-bargaining agent’s demands for higher wages. Immunizing such action from antitrust scrutiny may facilitate collective bargaining over labor demands. So, too, may immunizing concerted employer action designed to maintain the integrity of the multiemployer bargaining unit, such as lockouts that are imposed in response to “a union strike tactic which threatens the destruction of the employers’ interest in bargaining on a group basis.” NLRB v. Truck Drivers, 358 U. S. 87, 93 (1957).

In my view, however, neither the policies underlying the two separate statutory schemes, nor the narrower focus on the purpose of the nonstatutory exemption, provides a justification for exempting from antitrust scrutiny collective action initiated by employers to depress wages below the level *255that would be produced in a free market. Nor do those policies support a rule that would allow employers to suppress wages by implementing noncompetitive agreements among themselves on matters that have not previously been the subject of either an agreement with labor or even a demand by labor for inclusion in the bargaining process. That, however, is what is at stake in this litigation.

II

In light of the accommodation that has been struck between antitrust and labor law policy, it would be most ironic to extend an exemption crafted to protect collective action by employees to protect employers acting jointly to deny employees the opportunity to negotiate their salaries individually in a competitive market. Perhaps aware of the irony, the Court chooses to analyze this case as though it represented a typical impasse in an unexceptional multiem-ployer bargaining process. In so doing, it glosses over three unique features of the case that are critical to the inquiry into whether the policies of the labor laws require extension of the nonstatutory labor exemption to this atypical case.

First, in this market, unlike any other area of labor law implicated in the cases cited by the Court, player salaries are individually negotiated. The practice of individually negotiating player salaries prevailed even prior to collective bargaining.3 The players did not challenge the prevailing *256practice because, unlike employees in most industries, they want their compensation to be determined by the forces of the free market rather than by the process of collective bargaining. Thus, although the majority professes an inability to understand anything special about professional sports that should affect the framework of labor negotiations, ante, at 248-249, in this business it is the employers, not the employees, who seek to impose a noncompetitive uniform wage on a segment of the market and to put an end to competitive wage negotiations.

Second, respondents concede that the employers imposed the wage restraint to force owners to comply with league-wide rules that limit the number of players that may serve on a team, not to facilitate a stalled bargaining process, or to revisit any issue previously subjected to bargaining. Brief for Respondents 4. The employers co'üld have confronted the culprits directly by stepping up enforcement of roster limits. They instead chose to address the problem by unilaterally preventing players from individually competing in the labor market.

Third, although the majority asserts that the “club owners had bargained with the players’ union over a wage issue until they reached impasse,” ante, at 234, that hardly constitutes a complete description of what transpired. When the employers’ representative advised the union that they proposed to pay the players a uniform wage determined by the owners, the union promptly and unequivocally responded that their proposal was inconsistent with the “principle” of individual salary negotiation that had been accepted in the past and that predated collective bargaining.4 The so-called “bar*257gaining” that followed amounted to nothing more than the employers’ notice to the union that they had decided to implement a decision to replace individual salary negotiations with a uniform wage level for a specific group of players.5

Given these features of the case, I do not see why the employers should be entitled to a judicially crafted exemption from antitrust liability. We have explained that “[t]he nonstatutory exemption has its source in the strong labor policy favoring the association- of employees to eliminate competition over wages and working conditions.” Connell Constr. Co., 421 U. S., at 622. I know of no similarly strong labor policy that favors the association of employers to eliminate a competitive method of negotiating wages that predates collective bargaining and that labor would prefer to preserve.

Even if some collective action by employers may justify an exemption because it is necessary to maintain the “integrity of the multiemployer bargaining unit,” NLRB v. Brown, 380 U. S. 278, 289 (1965), no such justification exists here. The employers imposed a fixed wage even though there was no dispute over the pre-existing principle that player salaries should be individually negotiated. They sought only to prevent certain owners from evading roster limits and thereby gaining an unfair advantage. Because “the employer’s interest is a competitive interest rather than an interest in regulating its own labor relations,” Mine Workers v. Pennington, 381 U. S. 657, 667 (1965), there would seem to be no *258more reason to exempt this concerted, anticompetitive employer action from the antitrust laws than the action held unlawful in Radovich v. National Football League, 352 U. S. 445 (1957).

The point of identifying the unique features of this case is not, as the Court suggests, to make the case that professional football players, alone among workers, should be entitled to enforce the antitrust laws against anticompetitive collective employer action. Ante, at 249. Other employees, no less than well-paid athletes, are entitled to the protections of the antitrust laws when their employers unite to undertake anti-competitive action that causes them direct harm and alters the state of employer-employee relations that existed prior to unionization. Here that alteration occurred because the wage terms that the employers unilaterally imposed directly conflict with a pre-existing principle of agreement between the bargaining parties. In other contexts, the alteration may take other similarly anticompetitive and unjustifiable forms.

III

Although exemptions should be construed narrowly, and judicially crafted exemptions more narrowly still, the Court provides a sweeping justification for the exemption that it creates today. The consequence is a newly minted exemption that, as I shall explain, the Court crafts only by ignoring the reasoning of one of our prior decisions in favor of the views of the dissenting Justice in that case. Of course, the Court actually holds only that this new exemption applies in cases such as the present in which the parties to the bargaining process are affected by the challenged anticompetitive conduct. Ante, at 250. But that welcome limitation on its opinion fails to make the Court’s explanation of its result in this case any more persuasive.

The Court explains that the nonstatutory labor exemption serves to ensure that “antitrust courts” will not end up substituting their views of labor policy for those of either the *259Labor Board or the bargaining parties. Ante, at 236-237. The Court concludes, therefore, that almost any concerted action by employers that touches on a mandatory subject of collective bargaining, no matter how obviously offensive to the policies underlying the Nation’s antitrust statutes, should be immune from scrutiny so long as a collective-bargaining process is in place. It notes that a contrary conclusion would require “antitrust courts, insulated from the bargaining process, to investigate an employer group’s subjective motive,” a task that it believes too “amorphous” to be permissible. Ante, at 247.

The argument that “antitrust courts” should be kept out of the collective-bargaining process has a venerable lineage. See Duplex Printing Press Co. v. Deering, 254 U. S. 443, 483-488 (1921) (Brandeis, J., joined by Holmes and Clarke, JJ., dissenting). Our prior precedents subscribing to its basic point, however, do not justify the conclusion that employees have no recourse other than the Labor Board when employers collectively undertake anticompetitive action. In fact, they contradict it.

We have previously considered the scope of the nonstatu-tory labor exemption only in cases involving challenges to anticompetitive agreements between unions and employers brought by other employers not parties to those agreements. Ante, at 243. Even then, we have concluded that the exemption does not always apply. See Mine Workers v. Pennington, 381 U. S., at 663.

As Pennington explained, the mere fact that an antitrust challenge touches on an issue, such as wages, that is subject to mandatory bargaining does not suffice to trigger the judicially fashioned exemption. Id., at 664. Moreover, we concluded that the exemption should not obtain in Pennington itself only after we examined the motives of one of the parties to the bargaining process. Id., at 667.

The Court’s only attempt to square its decision with Pennington occurs at the close of its opinion. It concludes that *260the exemption applies because the employers’ action “grew out of, and was directly related to, the lawful operation of the bargaining process,” “[i]t involved a matter that the parties were required to negotiate collectively,” and that “concerned only the parties to the collective-bargaining relationship.” Ante, at 250.

As to the first two qualifiers, the same could be said of Pennington. Indeed, the same was said and rejected in Pennington. “This is not to say that an agreement resulting from union-employer negotiations is automatically exempt from Sherman Act scrutiny simply because the negotiations involve a compulsory subject of bargaining, regardless of the subject or the form and content of the agreement.” 381 U. S., at 664-665.

The final qualifier does distinguish Pennington, but only partially so. To determine whether the exemption applied in Pennington, we undertook a detailed examination into whether the policies of labor law so strongly supported the agreement struck by the bargaining parties that it should be immune from antitrust scrutiny. We concluded that because the agreement affected employers not parties to the bargaining process, labor law policies could not be understood to require the exemption.

Here, however, the Court does not undertake a review of labor law policy to determine whether it would support an exemption for the unilateral imposition of anticompetitive wage term's by employers on a union. The Court appears to conclude instead that the exemption should apply merely because the employers’ action was implemented during a lawful negotiating process concerning a mandatory subject of bargaining. Thus, the Court’s analysis would seem to constitute both an unprecedented expansion of a heretofore limited exemption, and an unexplained repudiation of the reasoning in a prior, nonconstitutional decision that Congress itself has not seen fit to override.

*261The Court nevertheless contends that the “rationale” of our prior cases supports its approach. Ante, at 243. As support for that contention, it relies heavily on the views espoused in Justice Goldberg’s separate opinion in Meat Cutters v. Jewel Tea Co., 381 U. S. 676 (1966). At five critical junctures in its opinion, see ante, at 236, 237-238, 242, 247-248, the Court invokes that separate concurrence to explain why, for purposes of applying the nonstatutory labor exemption, labor law policy admits of no distinction between collective employer action taken in response to labor demands and collective employer action of the kind we consider here.

It should be remembered that Jewel Tea concerned only the question whether an agreement between employers and a union may be exempt, and that even then the Court did not accept the broad antitrust exemption that Justice Goldberg advocated. Instead, Justice White, the author of Pennington, writing for Chief Justice Warren and Justice Brennan, explained that even in disputes over the lawfulness of agreements about terms that are subject to mandatory bargaining, courts must examine the bargaining process to determine whether antitrust scrutiny should obtain. Jewel Tea, 381 U. S., at 688-697. “The crucial determinant is not the form of the agreement — e. g., prices or wages — but its relative impact on the product market and the interests of union members.” Id., at 690, n. 5 (emphasis added). Moreover, the three dissenters, Justices Douglas, Clark, and Black, concluded that the union was entitled to no immunity at all. Id., at 736-738.

It should also be remembered that Justice Goldberg used his separate opinion in Jewel Tea to explain his reasons for dissenting from the Court’s opinion in Pennington. He explained that the Court’s approach in Pennington was unjustifiable precisely because it permitted “antitrust courts” to reexamine the bargaining process. The Court fails to explain its apparent substitution in this case of Justice Gold*262berg’s understanding of the exemption, an understanding previously endorsed by only two other Justices, for the one adopted by the Court in Pennington.

The Court’s silence is all the more remarkable in light of the patent factual distinctions between Jewel Tea and the present case. It is not at all clear that Justice Goldberg himself understood his expansive rationale to require application of the exemption in circumstances such as those before us here. Indeed, the main theme of his opinion was that the antitrust laws should not be used to circumscribe bargaining over union demands. Jewel Tea, 381 U. S., at 723-725. Moreover, Justice Goldberg proved himself to be a most unreliable advocate for the sweeping position that the Court attributes to him.

Not long after leaving the Court, Justice Goldberg served as counsel for Curt Flood, a professional baseball player who contended that major league baseball’s reserve clause violated the antitrust laws. Flood v. Kuhn, 407 U. S. 258 (1972). Although the Flood case primarily concerned whether professional baseball should be exempt from antitrust law altogether, see Federal Baseball Club of Baltimore, Inc. v. National League of Professional Baseball Clubs, 259 U. S. 200 (1922); Toolson v. New York Yankees, Inc., 346 U. S. 356 (1953), the labor law dimensions of the case did not go unnoticed.

The article that first advanced the expansive view of the nonstatutory labor exemption that the Court appears now to endorse was written shortly after this Court granted certio-rari in Flood, see Jacobs & Winter, Antitrust Principles and Collective Bargaining by Athletes: Of Superstars in Peonage, 81 Yale L. J. 1 (1971), and the parties to the case addressed the very questions now before us. Aware of both this commentary, and, of course, his own prior opinion in Jewel Tea, Justice Goldberg explained in his brief to this Court why baseball’s reserve clause should not be protected from antitrust review by the nonstatutory labor exemption.

*263“This Court has held that even a labor organization, the principal intended beneficiary of the so-called labor exemption, may not escape antitrust liability when it acts, not unilaterally and in the sole interest of its own members, but in concert with employers ‘to prescribe labor standards outside the bargaining unit[.]’ And this is so even when the issue is so central to bargaining as wages. Mine Workers v. Pennington, 381 U. S. at 668. Compare Meat Cutters v. Jewel Tea Co., 381 U. S. 676 (1965). See Ramsey v. Mine Workers, 401 U. S. 302, 307 (1971). . . .

“The separate opinion on which respondents focus did express the view that ‘collective bargaining activity on mandatory subjects of bargaining’ is exempt from antitrust regulation, without regard to whether the union conduct involved is ‘unilateral.’ Meat Cutters v. Jewel Tea Co., 381 U. S. at 732 (concurring opinion). But the author of that opinion agreed with the majority that agreements between unions and nonlabor groups on hard-core restraints like ‘price-fixing and market allocation’ were not exempt. 381 U. S. at 733. And there is no support in any of the opinions filed in Meat Cutters for Baseball’s essential, if tacit, contention that unilateral, hard-core anticompetitive activity by employers acting alone — the present case — is somehow exempt from antitrust regulation.” Reply Brief for Petitioner in Flood v. Kuhn, O. T. 1971, No. 71-32, pp. 13-14.

Moreover, Justice Goldberg explained that the extension of antitrust immunity to unilateral, anticompetitive employer action would be particularly inappropriate because baseball’s reserve clause predated collective bargaining.

“This case is in fact much clearer than Pennington, Meat Cutters, or Ramsey, for petitioner does not challenge the fruits of collective bargaining activity. He seeks relief from a scheme — the reserve system — which

*264Baseball admits has been in existence for nearly a century, and which the trial court expressly found was 'created and imposed by the club owners long before the arrival of collective bargaining.’” Id., at 14..

I would add only that this case is in fact much clearer than Flood, for there the owners sought only to preserve a restraint on competition to which the union had not agreed, while here they seek to create one.

Adoption of Justice Goldberg’s views would mean, of course, that in some instances “antitrust courts” would have to displace the authority of the Labor Board. The labor laws do not exist, however, to ensure the perpetuation of the Board’s authority. That is why we have not previously adopted the Court’s position. That is also why in other contexts we have not thought the mere existence of a collective-bargaining agreement sufficient to immunize employers from background laws that are similar to the Sherman Act. See Fort Halifax Packing Co. v. Coyne, 482 U. S. 1 (1987); Metropolitan Life Ins. Co. v. Massachusetts, 471 U. S. 724 (1985).6

*265IV

Congress is free to act to exempt the anticompetitive employer conduct that we review today. In the absence of such action, I do not believe it is for us to stretch the limited exemption that we have fashioned to facilitate the express statutory exemption created for labor’s benefit so that unions must strike in order to restore a prior practice of individually negotiating salaries. I therefore agree with the position that the District Court adopted below.

“Because the developmental squad salary provisions were a new concept and not a change in terms of the expired collective bargaining agreement, the policy behind continuing the nonstatutory labor exemption for the terms of a collective bargaining agreement after expiration (to foster an atmosphere conducive to the negotiation of a new collective bargaining agreement) does not apply. To hold that the nonstatutory labor exemption extends to shield the NFL from antitrust liability for imposing restraints never before agreed to by the union would not only infringe on the union’s freedom to contract, H. K. Porter Co. v. NLRB, 397 U. S. at 108 . . . (one of fundamental policies of NLRA is freedom of contract), but would also contradict the very purpose of the antitrust exemption by not promoting execution of a collective bargaining agreement with terms mutu*266ally acceptable to employer and labor union alike. Labor unions would be unlikely to sign collective bargaining agreements with employers if they believed that they would be forced to accept terms to which they never agreed.” 782 F. Supp. 125, 139 (DC 1991) (footnote omitted).

Accordingly, I respectfully dissent.

WINTER, Circuit Judge:

O. Leon Wood, an accomplished point-guard from California State University at Fullerton and a member of the gold medal-winning 1984 United States Olympic basketball team, appeals from Judge Carter’s dismissal of his antitrust action challenging certain provisions of a collective bargaining agreement between the National Basketball Association (“NBA”), its member-teams, and the National Basketball Players Association (“NBPA”). Wood contends that the “salary cap,” college draft, and prohibition of player corporations violate Section 1 of the Sherman Act, 15 U.S.C. § 1 (1982), and are not exempt from the Sher*957man Act by reason of the non-statutory “labor exemption.” We disagree and affirm.

The challenged provisions are in part the result of the settlement of an earlier antitrust action brought by players against the NBA. Robertson v. National Basketball Ass’n, 12 F.R.D. 64 (S.D.N.Y.1976), aff'd, 556 F.2d 682 (2d Cir.1977). In that case, a class consisting of all NBA players challenged both the merger of the NBA with the now-defunct American Basketball Association and certain NBA employment practices, including the college draft system by which teams obtain the exclusive right to negotiate with particular college players. Following extensive pre-trial proceedings, see Robertson v. National Basketball Ass’n, 413 F.Supp. 88 (S.D.N.Y.1976); 67 F.R.D. 691 (S.D.N.Y.1975); 389 F.Supp. 867 (S.D.N.Y.1975), the parties settled the case on April 29, 1976. The Settlement Agreement provided for the payment of $4.3 million to the class and for substantial modification of the practices attacked by the plaintiffs.

The Settlement Agreement is effective through the 1986-87 season. It modified the college draft system by limiting to one year the period during which a team has exclusive rights to negotiate with and sign its draftees. If a draftee remains unsigned at the time of the next year’s draft, he may re-enter the draft. Most important, the Settlement Agreement instituted a system of free agency allowing veteran players to sell their services to the highest bidder subject only to their current team’s right of first refusal that allows it to match the best offer.

On October 10, 1980, the NBA and NBPA signed a collective bargaining agreement that incorporated the provisions of the Settlement Agreement pertinent to this action. The 1980 collective agreement expired on June 1, 1982, however, and the 1982-83 season began before a new agreement had been reached. Negotiations between the NBA and the NBPA continued and centered on the league’s insistence upon controls on the growth in players’ salaries. The NBA claimed that increases in players’ salaries resulting from free agency were in part responsible for mounting losses and that a number of its teams might face bankruptcy absent some stabilization of expenses. The NBPA, after reviewing the teams’ financial data, reached an agreement in principle with the NBA on March 31, 1983, some 48 hours before a strike deadline set by the players. This agreement was memorialized in writing on April 18, 1983, in a Memorandum of Understanding (the “Memorandum”).

The Memorandum continued the college draft and free agency/first refusal provisions of the earlier agreements and, like those agreements, included provisions for fringe benefits such as pensions and medical and life insurance. However, the Memorandum also established a minimum for individual salaries and a minimum and maximum for aggregate team salaries. The latter are styled the salary cap provisions, even though they establish a floor as well as a ceiling. Under the salary cap, a team that has reached its maximum allowable team salary may sign a first-round draft choice like Wood only to a one-year contract for $75,000. An integral part of the method by which the floor and ceiling on aggregate team salaries were to be determined was a guarantee that the players would receive 53 percent of the NBA’s gross revenues, including new revenues, in salaries and benefits. This combination of fringe benefits, draft, free agency, a floor and a ceiling on aggregate team salaries, and guaranteed revenue sharing was unique in professional sports negotiations.

The Memorandum also prohibited the use of “player corporations,” which had been formed by players to enter into contracts with teams. According to the NBPA’s general counsel, the players agreed to this prohibition in light of changes in the tax laws that “virtually eliminated” the tax advantages of incorporation. The teams sought the prohibition to eliminate administrative and accounting difficulties and potential withholding tax liability that arose *958from their entering into contracts with player corporations.

Because the Memorandum altered certain terms and conditions of the Settlement Agreement, the NBA and NBPA jointly requested district court approval of a modification of the Settlement Agreement. Fed.R.Civ.P. 23(e). After a hearing at which class members were invited to address the fairness and adequacy of the modification, Judge Carter approved it on June 13, 1983.

Against this background, the Philadelphia 76ers drafted Wood in the first round of the 1984 college draft. At the time of the draft, the 76ers’ team payroll exceeded the amount permitted under the salary cap. The 76ers therefore tendered to Wood a one-year $75,000 contract, the amount stipulated under the salary cap. This offer was a formality, however, necessary to preserve its exclusive rights to sign him. In fact, the team informed Wood’s agent of its intention to adjust its roster so as to enable it to negotiate a long-term contract with Wood for substantially more money. Wood understandably did not sign the proffered contract.

On September 13, 1984, he turned from the basketball court to the district court and sought a preliminary injunction restraining enforcement of the agreement between the NBA and NBPA and compelling teams other than the 76ers to cease their refusal to deal with him except on the terms set out in the collective bargaining agreement and Memorandum.

Judge Carter denied Wood’s motion. Wood v. National Basketball Ass’n, 602 F.Supp. 525 (S.D.N.Y.1984). He found that both the salary cap and college draft provisions

affect only the parties to the collective bargaining agreement — the NBA and the players — involve mandatory subjects of bargaining as defined by federal labor laws, and are the result of bona fide arms-length negotiations. Both are proper subjects of concern by the Players Association. As such these provisions come under the protective shield of our national labor policy and are exempt from the reach of the Sherman Act.

Id. at 528.

In addition, he rejected Wood’s claim with respect to the prohibition of player corporations, noting that such a provision was probably a mandatory subject of bargaining, and that even if it was not, it raised no antitrust problems. Id. at 529. Meanwhile, Wood signed a contract with the 76ers that provided for $1.02 million in total compensation over a four-year period, including a $135,000 signing bonus. Wood has since been traded.

In January 1986, the parties made an evidentiary submission to Judge Carter for a decision on the merits. This consisted of papers submitted with the motion for a preliminary injunction and a stipulation of additional facts. On February 5, 1986, Judge Carter granted judgment to the defendants. This appeal followed.

DISCUSSION

Plaintiff views the salary cap, college draft and prohibition of player corporations as an agreement among horizontal competitors, the NBA teams, to eliminate competition for the services of college basketball players. As such, he claims, they constitute per se violations of Section 1 of the Sherman Act.1

Because the college draft and salary cap are the centerpieces of this litigation, we put the player corporation issue momentarily to one side in order to simplify our discussion. We may assume for purposes of this decision that the individual NBA teams and not the league are the relevant employers and that Wood would obtain considerably more favorable employment terms were the draft and salary cap eliminated so as to allow him to offer his services to the highest bidder among NBA *959teams. We may further assume that were these arrangements agreed upon by the NBA teams in the absence of a collective bargaining relationship with a union representing the players, they would be illegal and plaintiff would be entitled to relief.

The draft and salary cap are not, however, the product solely of an agreement among horizontal competitors but are embodied in a collective agreement between an employer or employers and a labor organization reached through procedures mandated by federal labor legislation. Their legality therefore, cannot be assessed without reference to that legislation. The interaction of the Sherman Act and federal labor legislation is an area of law marked more by controversy than by clarity. See R. Gorman, Labor Law 631-35 (1976) (“Gorman"). We need not enter this debate or probe the exact contours of the so-called statutory or non-statutory “labor exemptions,” however, because no one seriously contends that the antitrust laws may be used to subvert fundamental principles of our federal labor policy as set out in the National Labor Relations Act. 29 U.S.C. §§ 151-169 (1982). Wood’s claim is just such a wholesale subversion of that policy, and it must be rejected out of hand. As a result, whether the draft and salary cap are per se violations of the antitrust laws or subject to rule of reason analysis need not be decided.

Although the combination of the college draft and salary cap may seem unique in collective bargaining (as are the team salary floor and 53 percent revenue sharing agreement), the uniqueness is strictly a matter of appearance. The nature of professional sports as a business and professional sports teams as employers calls for contractual arrangements suited to that unusual commercial context. However, these arrangements result from the same federally mandated processes as do collective agreements in the more familiar industrial context. Moreover, examination of the particular arrangements arrived at by the NBA and NBPA discloses that they have functionally identical, and identically anti-competitive, counterparts that are routinely included in industrial collective agreements.

Among the fundamental principles of federal labor policy is the legal rule that employees may eliminate competition among themselves through a governmentally supervised majority vote selecting an exclusive bargaining representative. Section 9(a) of the National Labor Relations Act explicitly provides that “[rjepresentatives ... selected ... by the majority of the employees in a unit ... shall be the exclusive representatives of all the employees in such unit for the purposes of collective bargaining.” 29 U.S.C. § 159(a). Federal labor policy thus allows employees to seek the best deal for the greatest number by the exercise of collective rather than individual bargaining power. Once an exclusive representative has been selected, the individual employee is forbidden by federal law from negotiating directly with the employer absent the representative’s consent, NLRB v. Allis-Chalmers Mfg. Co., 388 U.S. 175, 180, 87 S.Ct. 2001, 2006, 18 L.Ed.2d 1123 (1967), even though that employee may actually receive less compensation under the collective bargain than he or she would through individual negotiations. J.I. Case Co. v. NLRB, 321 U.S. 332, 338-39, 64 S.Ct. 576, 580-81, 88 L.Ed. 762 (1944).

The gravamen of Wood’s complaint, namely that the NBA-NBPA collective agreement is illegal because it prevents him from achieving his full free market value, is therefore at odds with, and destructive of, federal labor policy.2 It is true that the diversity of talent and specialization among professional athletes and the widespread exposure and discussions of *960their “work” in the media make the differences in value among them as “workers” more visible than the differences in efficiency and in value among industrial workers. High public visibility, however, is no reason to ignore federal legislation that explicitly prevents employees, whether in or out of a bargaining unit, from seeking a better deal where that deal is inconsistent with the terms of a collective agreement.

Indeed, examination of the criteria that Wood advances as the basis for striking down the draft and salary cap reveals that there is hardly a collective agreement in the nation that would survive his legal theory. For example, Wood emphasizes his superior abilities as a point-guard and his selection in the first round of the college draft as grounds for enabling him to bargain individually for a higher salary. However, collective agreements routinely set standard wages for employees with differing responsibilities, skills, and levels of efficiency. Wood’s theory would allow any employee dissatisfied with his salary relative to those of other workers to insist upon individual bargaining, contrary to explicit federal labor policy. As one commentator has noted, “Congress gave to the majority representative the task of harmonizing and adjusting the conflicting interests of employees within the bargaining unit, no matter how diverse their skills, experience, age, race or economic level.” Gorman at 379. And the Supreme Court has observed, “The complete satisfaction of all who are represented is hardly to be expected.” Ford Motor Co. v. Huffman, 345 U.S. 330, 338, 73 S.Ct. 681, 686, 97 L.Ed. 1048 (1953).

Wood also attacks the draft and salary cap because they assign him to work for a particular employer at a diminished wage. However, collective agreements in a number of industries provide for the exclusive referral of workers by a hiring hall to particular employers at a specified wage. See Local 357, International Brotherhood of Teamsters v. NLRB, 365 U.S. 667, 81 S.Ct. 835, 6 L.Ed.2d 11 (1961); Associated General Contractors of America, Houston Chapter, 143 N.L.R.B. 409, enforced, 349 F.2d 449 (5th Cir.1965), cert. denied, 382 U.S. 1026, 86 S.Ct. 648, 15 L.Ed.2d 540 (1966). The choice of employer is governed by the rules of the hiring hall, not the preference of the individual worker. There is nothing that prevents such agreements from providing that the employee either work for the designated employer at the stipulated wage or not be referred at that time. Otherwise, a union might find it difficult to provide the requisite number of workers to employers. Such an arrangement is functionally indistinguishable from the college draft.

Wood further attacks the draft and salary cap as disadvantaging new employees. However, newcomers in the industrial context routinely find themselves disadvantaged vis-a-vis those already hired. A collective agreement may thus provide that salaries, layoffs, and promotions be governed by seniority, Ford Motor Co. v. Huffman, 345 U.S. at 337-39, 73 S.Ct. at 685-87, even though some individuals with less seniority would fare better if allowed to negotiate individually.

Finally, Wood argues that the draft and salary cap are illegal because they affect employees outside the bargaining unit. However, that is also a commonplace consequence of collective agreements. Seniority clauses may thus prevent outsiders from bidding for particular jobs, and other provisions may regulate the allocation or subcontracting of work to other groups of workers. See Fibreboard Paper Products Corp. v. NLRB, 379 U.S. 203, 210-15, 85 S.Ct. 398, 402-05, 13 L.Ed.2d 233 (1964). Indeed, the National Labor Relations Act explicitly defines “employee” in a way that includes workers outside the bargaining unit. 29 U.S.C. § 152(3).3

*961If Wood’s antitrust claim were to succeed, all of these commonplace arrangements would be subject to similar challenges, and federal labor policy would essentially collapse unless a wholly unprincipled, judge-made exception were created for professional athletes. Employers would have no assurance that they could enter into any collective agreement without exposing themselves to an action for treble damages. Moreover, recognition of a right to individual bargaining without the consent of the exclusive representative would undermine the status and effectiveness of the exclusive representative, and result in individual contracts that reduce the amount of wages or other benefits available for other workers. Wood’s assertion that he would be paid more in the absence of the draft and salary cap also implies that others would receive less if he were successful. It can hardly be denied that the NBA teams would be more resistant to benefits guaranteed to all, such as pensions, minimum salaries, and medical and insurance benefits. ,In fact, the salary cap challenged by Wood is one part of a complex formula including minimum team salaries and guaranteed revenue sharing.

The policy claim that one can do better through individual bargaining is nothing but the flip side of the policy claim that other employees need unions to protect their interests. Congress has accepted the latter position, and we are bound by that legislative choice.

The fact that one cannot alter important provisions of a collective agreement without undermining other provisions demonstrates that Wood’s antitrust claim fundamentally conflicts in yet another way with national labor policy. That policy attaches prime importance to freedom of contract between the parties to a collective agreement. See H. Wellington, Labor and the Legal Process 49-90 (1968). Freedom of contract is an important cornerstone of national labor policy for two reasons. First, it allows an employer and a union to agree upon those arrangements that best suit their particular interests. Id. at 28-29. Courts cannot hope to fashion contract terms more efficient than those arrived at by the parties who are to be governed by them. Second, freedom of contract furthers the goal of labor peace. Id. at 45. To the extent that courts prohibit particular solutions for particular problems, they reduce the number and quality of compromises available to unions and employers for resolving their differences.

Freedom of contract is particularly important in the context of collective bargaining between professional athletes and their leagues. Such bargaining relationships raise numerous problems with little or no precedent in standard industrial relations. As a result, leagues and player unions may reach seemingly unfamiliar or strange agreements. If courts were to intrude and to outlaw such solutions, leagues and their player unions would have to arrange their affairs in a less efficient way. It would also increase the chances of strikes by reducing the number and quality of possible compromises.

The issues of free agency and entry draft are at the center of collective bargaining in much of the professional sports industry. It is to be expected that the parties will arrive at unique solutions to these problems in the different sports both because sports generally differ from the industrial model and because each sport has its own peculiar economic imperatives. The NBA/NBPA agreement is just such a unique bundle of compromises. The draft and the salary cap reflect the interests of the employers in stabilizing salary costs and spreading talent among the various teams. Minimum individual salaries, fringe benefits, minimum aggregate team salaries, and guaranteed revenue sharing reflect the interests of the union in enhancing standard benefits applicable to all players. The free agency/first refusal provisions in turn allow individual players to exercise a degree of individual bargaining *962power.4 Were a court to intervene and strike down the draft and salary cap, the entire agreement would unravel.5 This would force the NBA and NBPA to search for other avenues of compromise that would be less satisfactory to them than the agreement struck down. It would also measurably increase the chances of a strike. We decline to take that step.6

We also agree with the district court that all of the above matters are mandatory subjects of bargaining under 29 U.S.C. § 158(d). Each of them clearly is intimately related to “wages, hours, and other terms and conditions of employment.” Indeed, it is precisely because of their direct relationship to wages and conditions of employment that such matters are so controversial and so much the focus of bargaining in professional sports. Wood’s claim for damages, for example, is based on an allegation of lost wages. For similar reasons, the prohibition on player corporations cannot be challenged on antitrust grounds. Use of player corporations may (or currently may not) increase the true salary received by players because of tax advantages, but also may increase the true wages paid by the NBA and its teams because of enhanced administrative costs. As such, it is a paradigmatic subject for bargaining.

It is true that the combination of the draft and salary cap places new players coming out of college ranks at a disadvantage. However, as noted earlier, that is hardly an unusual feature of collective agreements. In the industrial context salaries, promotions, and layoffs are routinely governed by seniority, with the benefits going to the older employees, the burdens to the newer. Wood has offered us no’ reason whatsoever to fashion a rule based on antitrust grounds prohibiting agreements between employers and players that use seniority as a criterion for certain employment decisions. Even if some such arrangements might be illegal because of discrimination against new employees (players), the proper action would be one for breach of the duty of fair representation. See Ford Motor Co. v. Huffman, supra; Ferro v. Railway Express Agency, Inc., 296 F.2d 847 (2d Cir.1961); cf. Britt v. Trailmobile Co., 179 F.2d 569 (6th Cir.), cert. denied, 340 U.S. 820, 71 S.Ct. 52, 95 L.Ed. 603 (1950).

Wood relies for legal support primarily upon the Supreme Court’s decisions in Connell Construction Co. v. Plumbers and Steamfitters Local Union No. 100, 421 U.S. 616, 95 S.Ct. 1830, 44 L.Ed.2d 418 (1975); Local Union 189, Amalgamated Meat Cutters v. Jewel Tea Co., 381 U.S. 676, 85 S.Ct. 1596, 14 L.Ed.2d 640 (1965); *963and United Mine Workers v. Pennington, 381 U.S. 657, 85 S.Ct. 1585, 14 L.Ed.2d 626 (1965). He reads those decisions as holding generally that a person outside the bargaining unit, in his case an unsigned first-round draft choice, who is injured in an anticompetitive fashion by a collective agreement may challenge that agreement on antitrust grounds. However, these cases are so clearly distinguishable that they need not detain us. Each of the decisions involved injuries to employers who asserted that they were being excluded from competition in the product market. Wood cites no case in which an employee or potential employee was able to invalidate a collective agreement on antitrust grounds because he or she might have been able to extract more favorable terms through individual bargaining. We need not determine the precise limits of the rules laid down by the cases cited or consider fine distinctions going to whether product- or labor-market activities are in issue. Wood’s claim is beyond peradventure one that implicates the labor market and subverts federal labor policy. It must, therefore, be rejected.

Affirmed.