These consolidated appeals derive from the plaintiff-appellant’s injunctive suit1 to halt alleged antitrust violations by the defendants-appellees, two foundations for medical care (FMCs),2 and the county medical society associated with one of them. The challenged conduct is the setting by majority vote of maximum fees that physician members may claim in full payment for health services they provide to policyholders of FMC-approved insurance plans. The appellant State of Arizona appeals from the vacation of a temporary restraining order that had already been extended without a hearing beyond the usual period. It also appeals from denial of summary judgment on the issue of liability. Pursuant to 28 U.S.C. § 1292(b) the district court certified the question whether the FMC membership agreements, which contain the promise to abide by maximum fee schedules, are illegal per se under section 1 of the Sherman Act, 15 U.S.C. § 1. We have jurisdiction to answer the certified question, but do not reach the jurisdictional issue raised by the other appeal. We affirm.
I.
Factual Background.
The FMCs in this case exemplify a type of organization that is beginning to play a significant part in the health services market. See C. Steinwald, An Introduction to Foundations for Medical Care (1971); Havighurst, Professional Restraints on Innovation In Health Care Financing, 1978 Duke L.J. 303, 315-16. The two challenged FMCs are not-for-profit corporations licensed as “insurance administrators” by the State of Arizona. Their activities include polling their members from time to time to set upper limits on fees they may charge patients covered by insurance plans the FMCs *555approve. This “price fixing,” the FMCs claim, subserves a general purpose of setting minimum standards and performing peer review and administrative tasks for health insurance plans. Under the heading of peer review, the foundations evaluate the medical necessity and appropriateness of treatment given and, in some instances, the use of hospital services. They also serve as agents for the underwriters, drawing funds directly from insurers’ bank accounts to pay doctors’ bills. Participation in the foundations is open to all physicians licensed in Arizona.
The State of Arizona charges that the foundations’ fee schedules have raised members’ fees above the average and median fees charged by Arizona doctors. The foundations dispute the appropriateness of the statewide figures used for the comparison, but concede that eighty-five to ninety-five percent of physicians in Maricopa County bill at or above the maximum reimbursement levels set by the county FMC.3 The foundations point out that the State’s own maximum reimbursement levels for its worker’s compensation insurance program and Comprehensive Medical/Dental Program for Foster Children are higher than the foundations’. Although the State primarily argues that the FMC membership agreements are contracts to fix prices, it hints that the FMCs serve a separate anti-competitive end by aiding the exchange of price information among doctors in the counties concerned. The State contends that the FMC fee schedules are not inseparable from the professional standards review of FMC-approved insurance plans, since the foundations offer peer review and administrative services for at least one health program — the foster child program — in which prices paid the doctors are not set by themselves but by the third party payor. Although the foundations say their purpose is to provide an alternative to “closed panel prepaid health insurance plans,” which are sometimes called “health maintenance organizations” (HMOs), see Havighurst, Health Maintenance Organizations and the Market for Health Services, 35 Law & Contemp. Prob. 716 (1970), the record does not show that HMOs exist in Maricopa or Pima County or have been kept from entering the health service market in those counties by the operation of the FMCs.
II.
The No. 79-3612 Appeal.
Our jurisdiction to consider the appeal in No. 79-3427 is based on 28 U.S.C. § 1292(b). As for the appeal in No. 79-3612, we note that the State of Arizona would be entitled to reinstatement of an originally proper preliminary injunction only if the court below, in dissolving it, had committed an abuse of discretion. The controlling standard is a sliding scale correlation of the “balance of hardships” with the likelihood that a permanent injunction will be granted. Benda v. Grand Lodge of International Association of Machinists, 584 F.2d 308, 314-15 (9th Cir. 1978), cert. dismissed, 441 U.S. 937, 99 S.Ct. 2065, 60 L.Ed.2d 667 (1979); William Inglis & Sons Baking Co. v. ITT Continental Baking Co., 526 F.2d 86, 88 (9th Cir. 1975). The State argued that without preliminary relief, its citizens would suffer pecuniary loss, and prospective patients who could not afford a physician’s care at supracompetitive prices would suffer physical injury and mental distress. On the other hand, a preliminary injunction threatened the very existence of the FMCs. Given this balance of hardships and the uncertainties attending proof of the antitrust violation, which we consider below, we cannot sáy that the district court would have erred in dissolving a preliminary injunction.
III.
The No.. 79-3427 Appeal.
A.
We must approach this appeal mindful that the Supreme Court has made *556it clear that the determination whether an agreement violates the Sherman Act turns on its “impact on competitive conditions.” National Society of Professional Engineers v. United States, 435 U.S. 679, 688, 98 S.Ct. 1355, 1363, 55 L.Ed.2d 637 (1978). The fact that a restraint may for one or more reasons appear reasonable is not controlling. An unreasonable restraint that contravenes the Sherman Act may be “based either (1) on the nature and 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.” Id. at 690, 98 S.Ct. at 1364 (footnote omitted). The key, to repeat, is the agreement’s impact on competition. If competition is promoted the agreement passes muster; if it suppresses or destroys competition it does not. Some agreements so often lack redeeming virtue, so frequently suppress or destroy competition, as to warrant their classification as per se unreasonable. Broadcast Music, Inc. v. Columbia Broadcasting Co., 441 U.S. 1, 7-8, 99. S.Ct. 1551, 1554, 60 L.Ed.2d 1 (1979). Once so classified a violation of the Sherman Act is made out merely by proving that such an unworthy agreement exists. Actual proof of its competitive impact is unnecessary. It is presumed to be anticompetitive.
The State of Arizona insists that the practice of setting maximum fees by majority vote of the members of the FMC constitutes an arrangement without redeeming virtue that suppresses and destroys competition and is thus unreasonable per se. The difficulty with Arizona’s position is that this record reveals nothing about the actual competitive effects of the challenged arrangement nor do the authorities, primary or secondary, afford assurance concerning its competitive impact. In truth, we know very little about the impact of this and many other arrangements within the health care industry. This alone should make us reluctant to invoke a per se rule with respect to the challenged arrangement.
There are, however, additional reasons. Foremost among them is that we are uncertain about the competitive order that should exist within the health care industry pursuant to the Sherman Act as interpreted by the courts. Only recently have the professions been brought by the Supreme Court within the reach of the Act. Goldfarb v. Virginia State Bar, 421 U.S. 773, 785-88, 95 S.Ct. 2004, 2012-13, 44 L.Ed.2d 572 (1975). The guidelines by which each is governed pursuant to the Act are being written on a case by case basis.
The health care industry, moreover, presents a particularly difficult area. The first step to understanding is to recognize that not only is access to the medical profession very time consuming and expensive both for the applicant and society generally, but also that numerous government subventions of the costs of medical care have created both a demand and supply function for medical services that is artificially high. The present supply and demand functions of medical services in no way approximate those which would exist in a purely private competitive order. An accurate description of those functions moreover is not available. Thus, we lack baselines by which could be measured the distance between the present supply and demand functions and those which would exist under ideal competitive conditions.
Perforce we must take industry as it exists, absent the challenged feature, as our baseline for measuring anticompetitive impact. The relevant inquiry becomes whether fees paid to doctors under that system would be less than those payable under the FMC maximum fee agreement. Put differently, confronted with an industry widely deviant from a reasonably free competitive model, such as agriculture, the proper inquiry is whether the practice enhances the prices charged for the services. In simplified economic terms, the issue is whether the maximum fee arrangement better permits the attainment of the monopolist’s goal, viz., the matching of marginal cost to marginal revenue, or in fact obstructs that end.
Approached in this manner, the weakness of Arizona’s suggestion that a per se rule be *557employed here becomes apparent. To assume that the arrangement in question wrongfully increases fees requires the further assumption that the FMCs are but devices to enable the member doctors to capture a greater share of potential monopoly profit, which their monopoly power makes available, than otherwise would be possible. This is an assumption we are not prepared to make on the basis of the record before us.
Our reluctance also has other roots. We are not, for example, prepared to believe without some evidence that a conspiratorial exploitation of monopoly power by the doctors would be supinely accepted by the insurance carriers. Nor are we prepared to reject out of hand the claim that a ceiling on fees does in fact reduce them. No one suggests that peer review is suspect, although it too purports to reduce the aggregate costs of medical services by curtailing unnecessary treatment. If it is proper to assume a conspiratorial exploitation with respect to maximum fees, why should a different assumption be applicable to peer review? Our belief is that neither assumption is appropriate.4
We are by no means unaware that economic motives frequently lie behind even the best of good works. We are, however, simply not prepared to brand the appellees’ conduct as “price-fixing” and thus a per se violation of the Sherman Act on the basis of an unsupported belief that fee enhancement is the likely consequence of the appellees’ maximum fee arrangement. We say this in full recognition that the appellees’ conduct involves some exchange of price information from which an agreement to raise or maintain prices might be implied if sufficient corroborative facts were shown. See Maple Flooring Assn. v. United States, 268 U.S. 563, 586, 45 S.Ct. 578, 586, 69 L.Ed. 1093 (1925). An agreement of that nature properly proven would violate section 1 of the Sherman Act. To affix the per se label to appellees’ conduct is, however, once more to substitute an unsupported belief for proper proof.
We acknowledge that it is quite possible, perhaps even probable, that the price ceiling serves the purpose of forestalling government price control that could force prices to an even lower level. This does not make the ceiling a violation of the Sherman Act. To eschew profit maximization in order to forestall price control is neither irrational nor, under the facts of this record, in violation of the Act. We observe in passing that only a government lost in an impenetrable legal maze, after having contributed substantially to the creation of monopoly conditions, would threaten price control if full monopoly profits are reaped and enforcement of the antitrust laws if private means are used to prevent the harvest.
B.
Appellant argues that our position is impermissible because under United States v. Socony-Vacuum Oil Co., 310 U.S. 150, 221, 60 S.Ct. 811, 843, 84 L.Ed. 1129 (1940), any collective effort to “tamper with price structures” offends antitrust policy. The argument, while by no means fanciful, overlooks Broadcast Music which held that a traditional form of minimum price fixing in the market for music under copyright was not illegal per se. Broadcast Music *558reminds us that whether to classify something as “ ‘per se price fixing’ will often, but not always, be a simple matter.” 441 U.S. at 9 & n.14, 99 S.Ct. 1557 & n.14 (emphasis added).
The issue whether to so classify the price schedules in this case is by no means “a simple matter.” In addition to the uncertainties already referred to, we do not know how health insurers such as Blue Cross fix their fee schedules in the relevant geographical area or whether the fees they offer exceed the appellees’ maximum fees. We are not informed by the record of the identity of, or the role played by, the various institutional components that compete in the relevant market. One may guess that doctors, both within and without the FMC structure, insurance carriers, hospitals, and perhaps HMOs, operate within the market; nonetheless, the record reveals nothing about the nature and extent of the competition between them. This makes it impossible to evaluate the pro- and anti-competitive aspects of a given feature of the total structure, although these aspects must be weighed together in determining whether a per se rule, or even the Rule of Reason, should brand the questioned feature illegal.5
Our unwillingness to employ a per se rule in this case is not overcome by the suggestion of a well-informed commentator that FMCs could and, perhaps in some instances, have unfairly prevented HMOs from recruiting subscribers and thus gaining entry to the relevant health service market. Havighurst, Health Maintenance Organizations and the Market for Health Services, 35 Law & Contemp. Prob. 716, 767-76 (1971). The thought is that FMCs, far from extracting the maximum profit possible, have set fees sufficiently low to discourage entry by potential competitors such as HMOs. See Knutson v. Daily Review, Inc., 548 F.2d 795, 814 n.21 (9th Cir. 1976) (“If the fixed maximum price is higher than cost but lower than a price that would permit new entrants or smaller scale competitors to operate (i. e., a ‘limit price’), then, although not predatory, it could support other efforts to acquire a monopoly.”) (dictum). Some have questioned the theory on which the thought rests. See R. Posner, Antitrust Law: An Economic Perspective 115 n.50 (1976); for a detailed critique of the limit-price theory, see Markovits, Potential Competition, Limit Price Theory, and the Legality of Horizontal and Conglomerate Mergers Under the American Antitrust Laws, 1975 Wis.L.Rev. 658. Doubt springs from the fact that the returns to those possessing monopoly power from pursuing a policy of fixing prices sufficiently low to discourage the entry of competitors are likely to be less over the long run than those resulting from a pricing policy that permits the capture of monopoly profits. In the latter case, high profits likely will draw competitors whose entry will drive prices down to a point approximating that which would have been set by a policy designed to bar entry. Thus, either policy over time yields approximately the same price level, the difference being that in the latter case available monopoly profits were captured while in the former they were foregone. Knowingly to forego such profits is irrational. This so-called “limit-price” theory, therefore, cannot be accepted as the foundation of a per se rule.
Our conclusion in this regard does not conflict with the general disapproval of predatory practices, of which predatory pricing is the most commonly examined species. See, e. g., Utah Pie Co. v. Continental Baking Co., 386 U.S. 685, 87 S.Ct. 1326, 18 L.Ed.2d 406 (1967); R. Posner, supra, at *559184-96 (1976); L. Sullivan, Handbook of the Law of Antitrust 108-13 (1977); Areeda & Turner, Predatory Pricing and Related Practices Under Section 2 of the Sherman Act, 88 Harv.L.Rev. 697 (1975); Cooper, Attempts and Monopolization: A Mildly Expansionary Answer to the Prophylactic Riddle of Section Two, 72 Mich.L.Rev. 373, 435-40 (1974); Williamson, Predatory Pricing: A Strategic and Welfare Analysis, 87 Yale L.J. 284 (1977). As these authorities illustrate, however, there is nothing like agreement concerning the signs by which predation is identifiable.6 Without deciding what the proper identifying signs may be, we are confident that none is revealed by this record. Moreover, the evil of predatory pricing is that it makes possible an ultimate capture of monopoly profits. Were prices maintained at the predatory level indefinitely the case against predatory pricing would more resemble that against suicide than that against monopolies.
The question whether the practice of which the appellant complains is sheltered from antitrust challenge by the McCarranFerguson Act, 15 U.S.C. § 1012, is not before us.7 For the purposes of this appeal we assume that the practice of which the appellant complains was not “the business of insurance.” No more was held with respect to the practice involved in Royal Drug. The issue whether that practice violated the antitrust laws expressly was not decided. Here we confront precisely that *560type of issue and merely hold that the challenged practice is not a per se violation.
We conclude by observing that we draw comfort from the Supreme Court’s reiteration in National Society of Professional Engineers, supra, 435 U.S. at 696 n.22, 98 S.Ct. at 1367 n.22, of its statement in Goldfarb v. Virginia State Bar, 421 U.S. 773, 95 S.Ct. 2004, 44 L.Ed.2d 572 (1975) that marketing restraints that regulate professional competition may pass muster under the Rule of Reason even though similar restraints on ordinary business competition would not. We believe this recognizes that a restraint may serve the public, the transcendent end of all professions, even though its presence in a purely commercial setting would violate the antitrust law. See Professional Engineers, supra, 435 U.S. at 696 & n.22, 98 S.Ct. at 1367 & n.22; Boddicker v. Arizona State Dental Ass’n, 549 F.2d 626 (9th Cir. 1977). There is sufficient probability of the challenged practice in this case being sheltered by this principle to justify our refusal to brand it as a per se violation.
Affirmed.
. The case was brought under both federal and Arizona antitrust laws, but the Arizona statute is interpreted in conformity with the federal ones. Uniform State Antitrust Act, Ariz.Rev. Stat. Ann. § 44-1402.
. The Pima County Medical Society was also originally joined as a party but has been dismissed from the suit under a consent order.
. The record is silent concerning the comparison of Pima County FMC fee schedules with statewide average and median fees.
. Two Supreme Court decisions have held maximum resale price agreements to be illegal per se. Albrecht v. Herald Co., 390 U.S. 145, 88 S.Ct. 869, 19 L.Ed.2d 998 (1968); Kiefer-Stewart Co. v. Joseph E. Seagram & Sons, 340 U.S. 211, 71 S.Ct. 259, 95 L.Ed. 219 (1951); see also Quinn v. Mobil Oil Co., 375 F.2d 273 (1st Cir.), cert. dismissed, 389 U.S. 801, 88 S.Ct. 8, 19 L.Ed.2d 56 (1967). This circuit has not extended those rulings to horizontal agreements that establish maximum prices. For an argument that the extension is proper, see L. Sullivan, Antitrust 210 (1977). A consequence of the extension, however, would be to moot an aspect of the currently lively debate concerning “limit-pricing” and predatory price-cutting. See 643 F.2d at 558-560 infra. It can hardly be said, then, that a per se rule forbidding horizontal maximum price agreements is well settled. It is arguable that this case does not present the issue for our decision. The FMC membership agreements do not set price ceilings for given services across the board but only limit the fees recoverable from certain patients, namely, those covered by FMC-approved insurance policies.
. When a questioned business practice affects more than one sphere of competition, the Rule of Reason of course recognizes that the enhancement of competition in one sphere may offset the weakening of competition in another. Sometimes, for example, “both interbrand and intrabrand competition must be considered”, First Beverages, Inc. v. Royal Crown Cola Co., 612 F.2d 1164, 1170 n.8 (9th Cir. 1980), although not every case requires extensive analysis of these multiple effects or jury instructions concerning them, id. at 1170-72. Here the novelty of the market or markets and the inadequacy of the record make an inquiry into the affected areas of competition essential.
. There is no disagreement about what constitutes predatory pricing. It is “pricing at a level calculated to exclude from the market an equally or more efficient competitor,” R. Posner, supra, at 188, “conduct which has the purpose and effect of advancing the actor’s competitive position, not by improving the actor’s market performance, but by threatening to injure or injuring [competitors], so as to drive them or keep them out of the market, or force them to compete less effectively,” L. Sullivan, supra, at 108. But while this definition of predatory behavior stresses anticompetitive intent, business tactics are exceptionally resistant to state-of-mind analysis. Antitrust theorists have therefore gravitated toward purely economic definitions. For example, Areeda and Turner propose that the courts regard as predatory pricing, policies that yield returns below average or marginal cost, whichever is lowest. Areeda & Turner, supra. Since marginal and average cost include a normal return on capital, however, the difficulties of determining the normal return for the firm in question make this proposal a demanding one for the courts to implement. See L. Sullivan, supra, at 110; R. Posner, supra, at 191-93. Posner believes that a modified form of the below-cost pricing criterion is appropriate. Id. at 188-91. Although this court has entertained the possibility of proving predation by marginal or average cost statistics, Hanson v. Shell Oil Co., 541 F.2d 1352 (9th Cir. 1976), we have not held that mode of proof to be exclusive, id. at 1358 (“[Predatory pricing] could be shown by evidence” of below cost pricing (emphasis added) ); cf. Janich Bros., Inc. v. American Distilling Co., 570 F.2d 848, 858 (9th Cir. 1977), cert. denied; 439 U.S. 829, 99 S.Ct. 103, 58 L.Ed.2d 122 (1978) (“Thus, as stated in Hanson, average variable cost can be used as evidence of marginal cost.”). More recently, the court has recognized “that refinement of the marginal or average variable cost test will be necessary as future cases arise.” California Computer Products, Inc. v. International Business Machines Corp., 613 F.2d 727, 743 (9th Cir. 1979) (citing Williamson, supra, among other recent commentators on predatory pricing). Since there is no contention before us that predation has occurred, we note only that our skepticism about limit-pricing in no way discredits theories of predation.
. The district court reasoned that since
“the agreements between physicians participating in the foundation-approved plans and the insurers underwriting the plans .. . spread no risk peculiar to the business of insurance, the McCarran-Ferguson Act exemption does not apply to exempt the defendants’ conduct from possible antitrust liability .. .. ”
Arizona v. Maricopa County Medical Society, No. CIV-78-800 PHX WPC (D.Ariz. June 11, 1979) (memorandum and order), Designated Record at 22-24. The parties before us have not attacked that holding. Recently, however, the Fourth Circuit decided that a chiropractors’ association performing peer review and fee supervision on behalf of health insurers was within the exemption. Bartholomew v. Virginia Chiropractors Ass’n, 612 F.2d 812 (4th Cir. 1979). The Bartholomew court explained that “the procurement of drugs,” with which Group Life and Health Insurance Co. v. Royal Drug Co., 440 U.S. 205, 99 S.Ct 1067, 59 L.Ed.2d 261 (1979) dealt, was “admittedly an act only thinly tangential to ... the business of insurance,” whereas the chiropractors’ association played a part in insurance as “a single, composite business.” Bartholomew, supra, at 817. We express no opinion regarding the validity of this distinction.