In November 1976 the United States instituted this action under Section 4 of the Sherman Act, 15 U.S.C.A. § 4, to enjoin the continuing violation of Section 1 of the Sherman Act, 15 U.S.C.A. § 1, by three vaW bureaus. These defendants, Southern Motor Carriers Rate Conference, Inc. (SMCRC), North Carolina Motor Carriers Association, Inc. (NCMCA), and Motor Carriers Traffic Association, Inc. (MCTA), represent common carriers before the regulatory commissions of the states of Alabama, Georgia, Mississippi, North Carolina, and Tennessee. The rate bureaus perform three basic functions: (1) they provide a forum for competing member carriers to discuss and agree on rates for intrastate transportation of general commodities to be proposed to state public service commissions for approval; (2) they publish tariffs and supplements containing the rates on which the carriers agree; and (3) they provide counsel, staff experts, and facilities for the preparation of cost studies and other exhibits and testimony for use in support of proposed rates at hearings held by the regulatory commissions.1 The government challenged the first of these functions as price fixing in violation of Section 1 of the Sherman Act.
The district court granted the government’s motion for summary judgment and held that defendants were in violation of Section 1. The court rejected arguments2 that defendants’ activities were immune under the state action doctrine or under the Noerr-Pennington doctrine. Defendants SMCRC and NCMCA and intervenor NA-RUC appealed, and a panel of the former Fifth Circuit affirmed. United States v. Southern Motor Carriers Rate Conference, Inc., 672 F.2d 469 (5th Cir. Unit B 1982). The Court en banc voted to rehear the case to consider two issues: first, whether a private party may avail itself of the state action exception to federal antitrust laws only if the state compels it to perform the disputed actions; and second, whether, if state compulsion is not necessary, the state policies here at issue were sufficiently clearly articulated and affirmatively expressed for the state action exception to apply. We hold that compulsion is required of private parties and thus do not reach the second issue. Accordingly, we affirm the judgment of the district court.
*535The Supreme Court first clearly articulated the “state action” exception to the antitrust laws in Parker v. Brown, 317 U.S. 341, 63 S.Ct. 307, 87 L.Ed. 315 (1943). At issue in Parker, a suit against state officials, was the applicability of the Sherman Act to an agricultural proration system established by a California statute authorizing a commission to impose marketing programs for raisins upon the petition of raisin growers. Once the program was approved and instituted by the commission, participation by the growers was mandatory; under the statute, growers who refused to follow a program once instituted were subject to fines and imprisonment. Id. at 347, 63 S.Ct. at 311-312.3 Analyzing the intent of Congress in enacting the Sherman Act, the Court concluded that, insofar as the Act applied only to “persons,” Congress intended application of the Act to the “business combinations” of “individuals and corporations.” Id. at 351, 63 S.Ct. at 313-314. The Act was not, therefore, intended to “restrain a state or its officers or agents from activities directed by its legislature.” Id. at 350-51, 63 S.Ct. at 313. Congress made no mention of states, and “[i]n a dual system of government in which, under the Constitution, the states are sovereign, save only as Congress may constitutionally subtract from their authority, an unexpressed purpose to nullify a state’s control over its officers and agents is not lightly to be attributed to Congress.” Id. at 351, 63 S.Ct. at 313.4 Applying this analysis to the California program, the Court reasoned that the Sherman Act was not intended to apply to the state of California in its establishment of a program operating under “state command.” “The state in adopting and enforcing the prorate program,” the Court concluded, “made no contract or agreement and entered into no conspiracy in restraint of trade or to establish monopoly but, as sovereign, imposed the restraint as an act of government which the Sherman Act did not undertake to prohibit.” Id. at 352, 63 S.Ct. at 314. The Court’s opinion in Parker therefore established the state action issue as a question of whether the challenged acts could be attributed to the state as sovereign.
The fact that the Court in Parker concluded that Congress did not intend to restrain anticompetitive acts attributable to states of course did not mean that only states could assert a state action defense. Thus, while Parker's holding applied to state party defendants, the Supreme Court in Goldfarb v. Virginia State Bar, 421 U.S. 773, 95 S.Ct. 2004, 44 L.Ed.2d 572 (1975), suggested that a defense of state action may also be available to private parties under certain limited circumstances. Gold-farb was a class action against the Virginia State Bar and a Virginia county bar association in which the Court considered the validity under the Sherman Act of a minimum fee schedule for lawyers published by the county bar association. After determining that the county bar association was not a state agency,5 the Court addressed the “threshold inquiry” of whether the bar association's “activity is required by the State acting as sovereign.” Id. at 790, 95 S.Ct. at 2015. The county bar association had ar*536gued that its action was “prompted” by the state, but a unanimous Court held that this was not enough for a state action defense. Rather, the action must be “compelled by direction of the State acting as a sovereign.” Id. at 791, 95 S.Ct. at 2015. Deciding that the state of Virginia did not require the publication of minimum fee schedules, the Court stated that it “need not inquire further into the state-action question,” id. at 790, 95 S.Ct. at 2015, and held that the challenged action was not state action for Sherman Act purposes.
After Goldfarb, the question remained as to what would be required in addition to the threshold compulsion requirement for a finding of state action for private parties. This question was partially answered the next term in Cantor v. Detroit Edison Co., 428 U.S. 579, 96 S.Ct. 3110, 49 L.Ed.2d 1141 (1976). At issue in Cantor was a program in which Detroit Edison — a private defendant — provided light bulbs to its electricity customers without a separate charge. Detroit Edison could not discontinue its program, the Court stated, without the approval of a Michigan state commission, and in this sense the threshold compulsion requirement was at least formally satisfied. See id. at 592, 96 S.Ct. at 3118 (“In this case we are asked to hold that private conduct required by state law is exempt from the Sherman Act.” (emphasis added)); id. at 594, 96 S.Ct. at 3119.6 As Goldfarb had hinted, however, compulsion alone was not enough, particularly in Cantor, where “there can be no doubt that the option to have, or not to have, such a program is primarily [Detroit Edison’s], not the [State’s],” id. at 594, 96 S.Ct. at 3119, and where the state was neutral as to the desirability of providing the light bulbs. Noting that it would not find an exception to federal antitrust law unless one was required for the state regulatory program to function, a majority of the Court concluded that Michigan’s policy interest in regulating electricity would not be affected by the inapplicability of the state action doctrine.7 Thus, although Detroit Edison’s action was formally commanded by the state, the Court declined to grant the company state action immunity where no significant state interest would be undermined by enforcement of federal antitrust law. This discussion as to the factors necessary for private party invocation of the state action doctrine is, however, for the most part not at issue in the present case. It is enough for our purposes that the Court premised its discussion on the fact that Detroit Edison’s action was “required” by the state. Id. at 592, 96 S.Ct. at 3118. In fact, although there were four separate opinions in Cantor, none of the Justices questioned the continued validity of Goldfarb’s requirement — which had been imposed by a unanimous Court — that a private party must first meet the compulsion requirement as a threshold matter.8
*537In our view, the analytical basis for Goldfarb’s compulsion requirement for private parties traces its roots directly to the analysis in Parker. The lesson of Parker is that the legislative history and wording of the Sherman Act point to the conclusion that Congress wished to restrain the anticompetitive actions of “persons,” that is, “individuals and corporations.” Thus, an anticompetitive action undertaken by a private party necessarily is proscribed by the Sherman Act unless for some reason that action cannot fairly be attributed to the private party. Under Goldfarb, when a state compels the private party to act, it deprives him of his freedom of choice; his action, being mere obedience, cannot be described as individual action and must therefore be attributed to the state. The Court elaborated on this freedom of choice analysis in Cantor after reviewing Supreme Court cases involving state action immunity:
In each of these cases the initiation and enforcement of the program under attack involved a mixture of private and public decisionmaking. In each case, notwithstanding the state participation in the decision, the private party exercised sufficient freedom of choice to enable the Court to conclude that he should be held responsible for the consequences of his decision.
428 U.S. at 593, 96 S.Ct. at 3119. This reasoning also explains the Court’s statement in Parker that a state may “not give immunity to those who violate the Sherman Act by authorizing them to violate it, or by declaring that their action is lawful,” 317 U.S. at 351, 63 S.Ct. at 314, and its statement in Cantor that “state authorization, approval, encouragement, or participation in restrictive private conduct confers no antitrust immunity,” 428 U.S. at 592-93, 96 S.Ct. at 3118-3119 (footnotes omitted). So long as the state merely allows, but does not compel, anticompetitive conduct, the decision to adopt such conduct remains that of the individual and is within the scope of the Sherman Act. Regardless of how strongly a state may wish to remove private action from the Act’s jurisdiction, it will not succeed unless it structures a program so as to make individual actions the actions of the state itself — by means of the mechanism of compulsion.9
A necessary corollary of the rationale of Goldfarb and Cantor is that compulsion should not be required of state defendants. Regardless of their motives for undertaking anticompetitive behavior, states are nevertheless acting as states.10 Since Goldfarb, the Court has consistently pointed out that the analysis of state action differs substantially depending on whether the defendant is a private party or a public institution. See, e.g., Bates, 433 U.S. at 361, 97 S.Ct. at 2697-2698 (Cantor would have been an “entirely different case” had the action been against a public official or public agency rather than against a private party); Cantor, 428 U.S. at 600-601, 96 S.Ct. at 3122-3123 (plurality opinion) (noting that Parker might have been different had there been a private defendant); id. at 585-92, 96 S.Ct. at 3115-3118 (distinguishing Parker as case against public officials). The same analysis applies to municipalities, at least when they act as agents of the state. Thus, the Supreme Court has not required compulsion for municipalities, but has held that when a state indicates its intention that the municipality may act as an instrumentality of the state — i.e., when the state “sanctions” anticompetitive behavior of the municipality — a state action defense may be *538available to the municipality. Community Communications Co. v. City of Boulder, 455 U.S. 40, 51, 102 S.Ct. 835, 840, 70 L.Ed.2d 810 (1982); City of Lafayette v. Louisiana Power & Light Co., 435 U.S. 389, 413, 98 S.Ct. 1123,1136-1137, 55 L.Ed.2d 364 (1978) (plurality opinion).
Under the foregoing analysis of the state action doctrine, we have little difficulty in holding that the activities of appellants are not immune from Sherman Act coverage. Specifically, a defense of state action fails because this litigation was brought against private defendants and none of the states in question compels the ratemaking bureaus to set rates collectively.11 Because the anticompetitive practices challenged by the government are undertaken at the individual election of the carriers and the rate bureaus, appellants fail to pass the threshold test established in Goldfarb. Thus, appellants’ action cannot be said to constitute state action.
Appellants, however, urge that state action analysis has been changed in their favor by the recent Supreme Court decision in California Retail Liquor Dealers Association v. Midcal Aluminum, Inc., 445 U.S. 97, 100 S.Ct. 937, 63 L.Ed.2d 233 (1980). Midcal involved a challenge under the Sherman Act to a California statute that required wine producers and wholesalers to file fair trade contracts or price schedules with the state. Under the statute, wholesalers could not resell wine to retailers at prices below those on the contracts or schedules. In order to evaluate whether the state defendant California Department of Alcoholic Beverage Control was state action immune, the Court briefly reviewed several of the state action cases discussed above and then stated that “[t]hese decisions establish two standards for antitrust immunity under Parker. First, the challenged restraint must be ‘one clearly articulated and affirmatively expressed as state policy’; second, the policy must be ‘actively supervised’ by the state itself. City of Lafayette v. Louisiana Power & Light Co., 435 U.S. 389, 410 [98 S.Ct. 1123, 1135, 55 L.Ed.2d 364] (1978) (opinion of Brennan, J.).” 445 U.S. at 105, *539100 S.Ct. at 943. California’s regulatory system, the Court ruled, met the first standard but not the second; it therefore refused to hold that the state defendant’s activity was outside the coverage of the Sherman Act. Applying Midcal to the facts in the present case, the government and appellants agree that each state actively supervises the collective ratemaking process,12 so the second Midcal standard is satisfied. As to the first Midcal standard, the parties disagree concerning the importance of the fact that none of the five states compels the challenged anticompetitive activity. Appellants assert that, despite the lack of compulsion, their actions meet the first standard. This is because the Supreme Court allegedly abandoned the practice established in Goldfarb of requiring compulsion of private parties, as evidenced by the fact that the first standard did not specifically refer to compulsion. We disagree with appellants’ argument for several reasons.
Initially, we question the very premise upon which appellants’ argument is based— that the specific choice of wording in Midcal implies a rejection of a compulsion requirement for private parties. In the context of private parties, the first Midcal standard makes sense only if it includes a compulsion requirement. This is because we do not see how a private party can carry out a clearly articulated and affirmatively expressed state policy when it is left to the private party to carry out that policy or not as he sees fit. Our interpretation of the disputed passage in Midcal is that it should be read as a broad, general restatement of the Court’s earlier holdings; the absence of any mention of compulsion in one specific application of the standard — that is, to private parties — therefore does not strike us as unusual. One additional factor points to the conclusion that the Court’s choice of wording in Midcal does not imply a rejection of its prior holdings. The Court introduced the two standards with the statement that the prior decisions “establish” the standards, 445 U.S. at 105,100 S.Ct. at 943, thus indicating that it did not view them as anything new. Furthermore, the Court used similar or identical language in earlier cases. See New Motor Vehicle Bd., 439 U.S. at 109, 99 S.Ct. at 411-412; City of Lafayette, 435 U.S. at 410, 98 S.Ct. at 1135 (plurality opinion); Bates, 433 U.S. at 362, 97 S.Ct. at 2698. If these cases did not alter the standard for state action immunity for private parties, we see no reason to believe that Midcal did.
Our interpretation of Midcal is supported by the consideration that that case had nothing to do with the applicability of a state action defense to a private party. Midcal involved an action by a wholesale distributor of wine for a writ of mandate for an injunction against the California Department of Alcoholic Beverage Control — a state agency.13 Because, as we said earlier, Parker immunity for state defendants does not require compulsion, it makes sense that the Court did not require compulsion of the state Department.14 Moreover, because the Court’s holding against state action immunity in Midcal was grounded on the fact that the second standard, requiring active state supervision, was lacking, it is not surprising that the Court did not articulate a standard designed to encompass every conceivable state action case. The Court’s citation to City of Lafayette — a non-private party case — immediately after listing the two standards further evidences that the Court’s focus was only on cases involving public defendants. The Supreme Court recently agreed with our reading of this passage, referring to Midcal as having adopted the principle “that anticompetitive restraints engaged in by state municipalities or subdivisions must be ‘clearly articu*540lated and affirmatively expressed as state policy’ in order to gain an antitrust exemption. Midcal, 435 U.S. at 105 [100 S.Ct. at 943].” Community Communications Co., 455 U.S. at 51 n. 14, 102 S.Ct. at 841 n. 14 (emphasis added). The fact remains that the only two state action cases in the Supreme Court to date involving private defendants have applied the threshold compulsion requirement.15
Appellants suggest that their view of the state action doctrine does not really constitute a significant change from prior law, arguing that Goldfarb would have come out the same way under their reading of Midcal because in Goldfarb the minimum fee schedule lacked even state authorization. Such speculation, however, ignores the fact that the Supreme Court in Goldfarb could not possibly have been more clear in imposing a compulsion requirement on private parties. The very basis of the Court’s holding was that compulsion was lacking; having made that determination, the Court reached no other issues, including whether the state had authorized the bar association’s action. More importantly, reading a compulsion requirement out of the private party state action doctHne would require a complete reformulation of the theoretical foundations for state action first established in Parker. Under the view of federalism espoused in Parker and followed in Goldfarb and Cantor, the Court’s inquiry focuses on the wording and intent of the Sherman Act in determining whether private action is mere obedience to a state command and therefore effectively the action of the state itself. Under appellants’ view of federalism, on the other hand, Congress’ explicit legislative determination to reach the conduct of “individuals and corporations” may be overridden by a state’s decision to implement a voluntary supervision scheme that permits but does not compel anticompetitive behavior.16 Even if *541such a view of federal-state relations could be supported, its adoption would require at least some explanation or discussion. In Mideal there was none. Additionally, appellants’ view of state action for private parties would raise extremely difficult practical questions of interpretation and application. For example, appellants presumably do not dispute the continued validity of the Court’s statement in Cantor in 1976 that state authorization, approval, encouragement, or participation in restrictive private conduct does not confer antitrust immunity. See 428 U.S. at 592-93, 96 S.Ct. at 3118-3119. Yet appellants’ interpretation of Mideal would have us find antitrust immunity where a state grants private parties the option of participating in a scheme of state supervision. If there is a significant difference between these two schemes, it eludes us. Thus, in light of the significant theoretical and practical difficulties that a reformulation in state action doctrine would create, we are not inclined to infer such a reformulation from a Supreme Court opinion which is completely silent on the subject.
The final reason why we conclude that Mideal does not represent a rejection of a compulsion requirement for private parties is found in the language of the opinion itself. The Court chose to cite Goldfarb with approval, and even quoted from it to the effect that “[i]t is not enough that ... anticompetitive conduct is ‘prompted’ by state action; rather, anticompetitive activities must be compelled by direction of the State acting as a sovereign.” 445 U.S. at 104, 100 S.Ct. at 943 (quoting 421 U.S. at 791, 95 S.Ct. at 2015) (emphasis added). Moreover, in a footnote immediately after setting out the two standards for state action immunity applicable to that case, the Court cited á student note in which the only argument contained at the cited page was that private action must be compelled by the state in order for the private party to invoke state action immunity.17
In addition to arguing that Mideal represents a rejection of a compulsion requirement for private parties, appellants also rely on various policy arguments against affirming the judgment of the district court. Even if we were not bound by the Supreme Court’s decisions in Goldfarb and Cantor, we would find these arguments unpersuasive.
Appellants argue first that competition is undermined by a state action requirement that allows states to eliminate competition completely but not to maintain a policy that gives parties the choice of whether or not to compete. But the rationale of the state action doctrine is federalism, not the maximization of competition. As we stated earlier, it is the source of the decision not to compete — not the extent of competition — which determines whether state action exists.18 Unless it is the state *542rather than the individual or corporation that decides to act anticompetitively, a state action defense would be inconsistent with the wording and intent of the Sherman Act.
Appellants also urge that if we uphold the injunction against collective ratemaking and require competition among carriers, carriers will face the burden of additional expense in preparing individual rate filings. They also assert that the ability of state regulators to perform their task will be undermined. Even if these assertions were supported in the record,19 they would be irrelevant to proper antitrust analysis. Appellants’ objections amount to the argument that competition does not work well, so the antitrust laws should not apply. This argument has consistently been rejected by the Supreme Court. See National Society of Professional Engineers v. United States, 435 U.S. 679, 98 S.Ct. 1355, 55 L.Ed.2d 637 (1978); United States v. Socony-Vacuum Oil Co., 310 U.S. 150, 220-22, 60 S.Ct. 811, 842-843, 84 L.Ed. 1129 (1940). Moreover, if collective ratemaking is as desirable as appellants insist, then it should be but a small step for the states to compel it of all carriers. This latter point underscores a critical element of the state action issue. Congress has instructed courts not to treat lightly overt collusion by private parties in setting prices. Permitting states to confer antitrust immunity on private parties who voluntarily elect state regulation would, in our view, open the way for the creation of large loopholes in federal antitrust law, loopholes whose contours would be difficult to contain. If a state wishes to displace federal antitrust laws, it should be permitted to do so only if it speaks in the clearest and strongest possible terms — by compelling private parties to submit to its own regulatory system.
For the reasons stated above, we conclude that the district court properly decided that appellants were not entitled to claim state action immunity. On application for rehearing en banc appellants suggest no fault with the panel’s determination that the Sherman Act was violated or that the Noerr-Pennington doctrine does not apply to the challenged activities. Accordingly, we reinstate that part of the panel opinion dealing with these issues, 672 F.2d 469, 476-81 (5th Cir.1982),20 and AFFIRM the judgment of the district court.
. For further elaboration of the facts regarding the role of the rate bureaus, their relationship with the state regulatory commissions, and the general pattern of regulatory procedures in the five subject states, see the district court’s carefully written opinion. United States v. Southern Motor Carriers Rate Conference, Inc., 467 F.Supp. 471, 476-78 (N.D.Ga.1979).
. In the district court, the state attorneys general of the states of Alabama, Georgia, Mississippi, North Carolina, and Tennessee participated as amici curiae. The National Associa- . tion of Regulatory Utility Commissioners (NA-RUC) also was allowed to participate as an intervenor.
. Appellants seek to characterize Parker as a case in which individual growers were not compelled to act anticompetitively, citing a provision of the program which permitted growers to set aside up to 30% of their crops for sale on the free market. The 30% set-aside provision, however, did not affect the ability of a grower to elect not to participate as to the remaining 70% of his crops. In this sense, the state mandated the participation of every grower and compelled the challenged anticompetitive behavior.
. Significantly, this reference to federalism concerns was directed not to any limitation on Congress’ power to act; instead, it was stated merely as a rule of statutory interpretation— that is, whether and to what extent Congress had chosen to override existing state regulation.
. The State Bar was, for “some limited purposes,” a state agency. 421 U.S. at 791, 95 S.Ct. at 2015. But because the fee schedule itself was published by the county bar association, we agree with those who characterize Goldfarb as an action against a private party. Cf. Bates v. State Bar of Arizona, 433 U.S. 350, 97 S.Ct. 2691, 53 L.Ed.2d 810 (1977) (antitrust claims raised only against state bar, which was a state defendant).
. In determining whether compulsion is present, we do not believe that private parties may have no role in proposing mandatory state action. In Parker and New Motor Vehicle Bd. v. Orrin W. Fox Co., 439 U.S. 96, 99 S.Ct. 403, 58 L.Ed.2d 361 (1979), the Supreme Court made it clear that an action still may be compelled by the state even though private parties petitioned the state to consider requiring such action. See also California Motor Transp. Co. v. Trucking Unlimited, 404 U.S. 508, 92 S.Ct. 609, 30 L.Ed.2d 642 (1972); United Mine Workers v. Pennington, 381 U.S. 657, 85 S.Ct. 1585, 14 L.Ed.2d 626 (1965); Eastern R.R. Presidents Conference v. Noerr Motor Freight, Inc., 365 U.S. 127, 81 S.Ct. 523, 5 L.Ed.2d 464 (1961). In Cantor, formal compulsion was present despite the fact that Detroit Edison could with little difficulty have petitioned the state commission for discontinuance of the program which the company itself had proposed. Thus, by compulsion we mean something less than that envisioned by some commentators. See e.g., 1 P. Areeda & D. Turner, Antitrust Law ¶215 (1978). Still, once a state has adopted some policy, whether or not after a petition by a private party, private parties will be exempt from the antitrust laws only if the state requires some action pursuant to that policy.
. In a prior part of the opinion (Part II) Justice Stevens, joined only by three other members of the Court, concluded that Parker should be limited to suits against state officials.
. The disagreement in the Court centered on what in addition to compulsion was required for state action immunity for private parties. In fact, Justice Stewart, joined by Justices Powell and Rehnquist, argued that Michigan’s decision to adopt Detroit Edison’s proposal and then compel compliance with it should have *537been by itself sufficient for state action immunity.
. Another rationale for the compulsion requirement is that if a private party is permitted by the state to make the choice whether or not to compete, there is no necessary conflict between state and federal law because the private party may comply with both federal and state law at the same time.
. We acknowledge that Bates and Parker, which involved state defendants, discussed the presence of compulsion, but we view those cases — as appellants would have us do — as instances in which the presence of compulsion provides the best evidence of clear articulation and affirmative expression of state policy to displace competition.
. The district court found that “regardless of the virtual romance between the states and the ratemaking conferences, no state requires that rates be published collectively.” 467 F.Supp. at 483 (emphasis in original). An analysis of each state’s policies reveals the correctness of the district court’s finding. North Carolina General Statute § 62-152.1(b) states that in order to realize and effectuate the policy of uniform rates among carriers, any party to an agreement with other carriers regarding uni-' form rates may apply to a state commission for approval of the agreement; Section 62-152.-1(h) relieves parties to an agreement so approved from the operation of the antitrust laws. The statute does no more than authorize or permit collective ratemaking. There is no compulsion. Tennessee Public Service Commission Rule 1220-2-1.40 adopts Section 5a of the Interstate Commerce Act, 49 U.S.C.A. § 10706(b), which permits carriers to file joint rates and exempts those who do file such rates from the antitrust laws. In addition, Tenn. Code Ann. § 65-1506 permits the Public Service Commission to establish joint ratemaking procedures. Again, there is no state compulsion. Georgia Code Ann. § 68-613, which requires a state commission to “establish rate-making procedures for all motor common carriers, which procedure shall include, but not be limited to, collective ratemaking procedures,” explicitly indicates that Georgia intends to give carriers the option of filing rates jointly or individually. The State of Georgia, participating as amicus in the district court, presented a 1942 order issued by its public service commission stating that “all for hire motor common carriers operating under Class A certificates be, and they are hereby, directed and required to publish and file with this Commission ... a tariff containing all local and joint class and/or commodity rates applicable between points within Georgia.” Pointing to the italicized language, the state argued that the commission “clearly directed all Class ‘A’ carriers to file a single joint tariff with the Commission,” thus indicating its policy in favor of collective action. Id. We are not persuaded that the language of the 1942 order is susceptible of only this interpretation. At best it is ambiguous. Given the statute, only recently amended to insert the provisions outlined above, it would not, in any event, be controlling. A 1942 order of the Alabama Public Service Commission requires carriers to “prepare, publish and file with the Commission an individual tariff or ... participate in an agency issue.” As with Georgia, Alabama explicitly permits individual rate applications. Finally, Mississippi in Miss.Code Ann. §§ 77-7-1 to -341 merely contemplates cooperation among carriers to the limited extent of establishing “joint rates.” This falls short of compelling joint applications for proposed rates.
. The government points out that the district court conducted no fact-finding as to this issue. The record, however, reveals that the state commissions conduct hearings to review the reasonableness of proposed carrier tariffs and routinely suspend their effectiveness.
. In the Supreme Court, petitioner California Retail Liquor Dealers Association had appealed the case as an intervenor.
. In any event, compulsion was in fact present in Midcal. See 445 U.S. at 100, 100 S.Ct. at 940.
. Appellants cite several lower court cases in support of their position, but only two of them actually found that private parties were immune without compulsion by the state. The first, Caribe Trailer Systems, Inc. v. Puerto Rico Maritime Shipping Authority, 475 F.Supp. 711 (D.D.C.1979), aff'd per curiam, No. 79-1658 (D.C.Cir.1980), cert. denied, 450 U.S. 914, 101 S.Ct. 1355, 67 L.Ed.2d 339 (1981), a district court opinion, is of only limited relevance because in that case the court reviewed a governmental restraint involving a private party (and a state defendant) rather than, as here, a private restraint approved by the government. In the second case, Turf Paradise, Inc. v. Arizona Downs, 670 F.2d 813 (9th Cir.1982), cert, denied, -U.S.-, 102 S.Ct. 2308, 73 L.Ed.2d 1308 (1982), in which the Ninth Circuit reviewed an antitrust challenge to a private agreement allocating horse racing dates between two race track operators, the state had by statute specifically reserved the right to perform the allocation in the event that the private parties could not agree. Id. at 816. Thus, although the state did not compel the private parties to enter into the racing date agreement, it had effectively ensured that the anticompetitive conduct would occur one way or another.
. Appellants reason that the Court eliminated the compulsion requirement because the basis for state compulsion is the concern that the economic self interest of private parties might affect the rate setting process and, according to appellants, the second standard in Midcal — the requirement of state supervision — assuages that concern. They assert that a supervision requirement adequately ensures that the state has determined that the challenged activity is in furtherance of the state’s policy and that the state does not abdicate its responsibility to private parties. We disagree with this analysis. If the second Midcal standard accomplished as much as appellants claim, there would have been no need for the first standard. Further, if compulsion and supervision served the same purpose, the Supreme Court would have upheld, rather than denied, state action immunity in Midcal because compulsion was present in that case. See 445 U.S. at 99-100, 100 S.Ct. at 940-941. That the Court did not reach this result is evidence that supervision and compulsion serve different purposes. In our view, both supervision and compulsion are necessary for a finding of state action. A state cannot deprive competitors of their ability to subvert competition merely by compelling them to formulate collective rates. Conversely, active supervision of anticompetitive conduct undertaken at the voluntary election of a private party also is not by itself adequate. For example, a state may by regulation supervise the reasonableness of rates, but the reasonableness of any particular rate may have little to do with whether it was arrived at anticompetitively. A state regulatory system does not even necessarily have the regulatory power to supervise anticompetitive behavior if the resulting rate falls somewhere within the zone of reasonableness. As we stated earlier, compulsion serves an entirely different purpose from supervision: *541as Goldfarb and Parker established, the compulsion inquiry is to determine whether the specific anticompetitive activity at issue was actually that of the state acting as sovereign.
. The mere fact that anticompetitive conduct by a private party effectuates a clearly articulated state regulatory policy is not in itself sufficient to justify a state action defense against injunctive relief. The defense is only available if the state has elected to effectuate its policy by requiring the private conduct under attack.
Note, Parker v. Brown Revisited: The State Action Doctrine After Goldfarb, Cantor and Bates, 77 Colum.L.Rev. 898, 916 (1977) (emphasis in original). The Note, at the end of the relevant section, was even more definite: “To summarize, a state action defense against injunctive relief should be available for private conduct only if the state has specifically required the conduct in question without abdicating final decisionmaking authority to private parties.” Id. at 918.
The same footnote in the Mideal opinion also cited Norman’s on the Waterfront, Inc. v. Wheatley, 444 F.2d 1011, 1018 (3d Cir.1971), and Asheville Tobacco Bd. v. FTC, 263 F.2d 502, 509-10 (4th Cir.1959), in which state action defenses were refused. Although both cases are ambiguous as to the requirements for state action, neither one implies that compulsion is not necessary for private parties.
. Thus, a state may exempt a specific subclass of carriers from regulation or, as in Parker, it may predetermine a set amount or type of activity that may remain in the free market. In both of these cases, the state controls and sets the terms of the individual decision as to whether to compete. But where the option to *542compete or join a system of state regulation at all is up to the private party, state action does not exist.
. We note that the government did not seek to enjoin the rate bureaus from performing joint rate studies or from undertaking joint publication of individual rates and agreements involving two or more parties who provide joint service.
. We also agree with the panel’s determination that there is no implied immunity from the antitrust laws resulting from the antitrust exemption granted motor carriers under 49 U.S. C.A. § 10706, or from the existence of pervasive state regulation permitted by 49 U.S.C.A. § 10521(b). See 672 F.2d at 475 n. 9.