This antitrust case presents questions concerning issue preclusion, horizontal price fixing with a union’s assistance, and the Noerr-Pennington doctrine. The questions arise from an agreement in 1976 between the National Electrical Contractors Association, Inc. (the Association), which comprises firms doing 50-60% of the nation’s electrical contracting work, and the International Brotherhood of Electrical Workers, AFL-CIO (the Union), which represents employees of these and many other contractors. The agreement established the National Electrical Industry Fund (the Fund), and members of the Association pay 1% of their gross payroll to the Fund to finance its activities. The Fund helps to defray the costs of the Association’s bargaining with the Union on behalf of its members and administering their collective bargaining agreements. The Fund also pays for some research and educational programs. The 1976 agreement calls for the Union to obtain, as part of any collective bargaining agreement with a firm that is not a member of the Association, a requirement that the firm contribute 1% of its gross payroll to the Fund.
I
Firms that were outside the Association objected to the Union’s effort to divert 1% of each employer’s payroll to the Fund. Characterizing the contribution requirement as a cartel, they filed an antitrust suit in the federal court in Maryland in 1977. Two of the plaintiffs in the Maryland case asked that it be certified as a class action on behalf of all electrical contractors that did not belong to the Association. Despite the requirement of Fed.R.Civ.P. 23(c)(1) that the court determine “[a]s soon as practicable” whether a suit may be maintained as a class action, the district court allowed that issue to slide for three years. It then simultaneously decided the merits and the application for certification of the class. National Constructors Ass’n v. National Electrical Contractors Ass’n, Inc., 498 F.Supp. 510 (D.Md.1980). The court held the contribution requirement unlawful per *360se under § 1 of the Sherman Act, 15 U.S.C. § 1, see 498 F.Supp. at 529-43. It also certified a class of electrical contractors that do not belong to the Association and have signed agreements with the Union requiring them to make the 1% contribution to the Fund. Id. at 544-51. The court denominated this as a class under Rule 23(b)(3), id. at 548, which required that each member be given notice and offered an opportunity to opt out. Certification under Rule 23(b)(3) was appropriate, the judge concluded, because all members of the class had identical claims for damages, which could be computed in a mechanical fashion. Id. at 549-50. The district court then deferred the required notice while the Association, Union, and Fund took an appeal under 28 U.S.C. § 1292(a)(1) from the injunction against their demand that firms pay the 1% fee.
On September 16, 1980, seven days after the Maryland court filed its opinion, Premier Electrical Construction Co. — a member of the class that had been certified in Maryland — filed this suit in Chicago. Premier’s complaint tracked that of the plaintiffs in Maryland, with one exception. The plaintiffs in the Maryland case wanted an injunction and damages of treble the amounts they had paid into the Fund. Although Premier had signed contracts containing the 1% requirement, it had refused to pay and therefore could not obtain the kind of damages the class representatives had requested. Because Premier had refused to pay, the Fund had filed three suits (one for each year of the contract) in state courts of Illinois. Premier defended these suits, arguing among other things that the contribution requirement violated the Sherman Act. These state suits proceeded for a while as the Maryland litigation was ongoing. They were stayed after a time, and the Fund dismissed them voluntarily when it lost the Maryland case. Premier’s complaint in federal court asked for treble damages based on the expense, including attorneys’ fees, of defending the contract actions in Illinois.
Two weeks after filing the suit in Chicago, Premier informed the Maryland court that it would move to consolidate the cases. It did not file the necessary motion, however, until February 1982, when the court in Chicago demanded that Premier show cause for its inaction. Premier quickly filed with the Judicial Panel on Multidistrict Litigation a motion to transfer the case under 28 U.S.C. § 1407(c)(ii). This was a peculiar motion, because § 1407 allows transfer and consolidation only for pretrial proceedings, and the Maryland case was already on appeal from a dispositive judgment. The Panel denied the motion in light of “the more advanced stage of the Maryland action” and the “minimal number of actions involved in this litigation”. Premier did nothing further concerning the Maryland litigation until March 1983.
Meanwhile the Fourth Circuit affirmed the Maryland court’s decision. National Electrical Contractors Ass’n, Inc. v. National Constructors Ass’n, 678 F.2d 492 (4th Cir.1982). The court of appeals agreed with the district court that the 1% requirement was unlawful per se as price fixing. Premier did not participate in the appeal, and apparently no one asked the Fourth Circuit to address the propriety of the class certification. The Association, Union, and Fund asked the Supreme Court to review the case. While the petition was pending, the case was settled. The defendants consented to the entry of an injunction against collecting the 1% contribution from firms that did not belong to the Association. They also offered to create a fund of $6 million, on which class members could draw in proportion to their contributions to the Fund since 1977.
On learning of the terms of the proposed settlement, Premier objected that only payments into the Fund could be a basis of access to the $6 million. In April 1983 it asked the Maryland court to defer still further the giving of notice under Rule 23 and to allow it to add and litigate claims on behalf of a new subclass that had incurred expenses of litigation. The subclass Premier proposed would include contractors sued, threatened with suit, or subject to any other efforts to collect the fee. Premier apparently wanted payments to this sub*361class to come on top of the $6 million. The Maryland court concluded that Premier had tarried far too long in proposing a new subclass and rebuffed its requests. It stated that the case had “progressed to the point that ... it would be totally unfair to interrupt it by permitting intervention along the lines requested and staying the distribution of the class notice and redefining the class. It seems to me that Premier, being a member of the class in this suit, should, if it so desires, object to the settlement, either that or seek to opt out of the settlement and pursue any further claims it may have elsewhere.” The notices to the class were issued in April 1983 and told class members that they could accept the benefits of the settlement, object to the settlement (see Rule 23(e)), or opt out of the class. Premier opted out. The district court approved the settlement, and in August 1983 the defendants dismissed their petition for certiorari. 463 U.S. 1234, 104 S.Ct. 26, 77 L.E.2d 1449 (1983).
The Maryland case was over, but the Chicago case had just begun. There had been only limited discovery, and Premier wanted to keep things that way. It asked the Chicago court to hold that the defendants are bound by the Fourth Circuit’s decision that they violated the Sherman Act. The defendants opposed this request and added the contention that the sort of damages Premier requested are unavailable under the Sherman Act unless the state litigation was a “sham”. Both sides moved for summary judgment.
The district court held that the defendants are bound by the Maryland decision under principles of issue preclusion (collateral estoppel, here the offensive, non-mutual variety). Premier Electrical Construction Co. v. IBEW, 627 F.Supp. 957, 960-63 (N.D.Ill.1985). The court concluded that class members should be entitled to the benefit of preclusion even when they opt out, because preclusion will reduce claims on judicial resources.1 This did not carry the day for Premier, however, because the court also held that the Noerr-Pennington doctrine precludes the recovery as damages of the costs of defending the Fund's suits. Id. at 964-66. Premier had not alleged that the suits were shams in the sense that the Fund did not care whether it won, or that the Fund brought them for the sake of the costs they would impose rather than because of the prospect of winning.
The court did not grant the defendants’ motions for summary judgment, however, because of the possibility that Premier could show some other kind of compensable injury. Premier asked the district court to certify its order for an interlocutory appeal under 28 U.S.C. § 1292(b), arguing that the court’s interpretation of the Noerr-Pennington doctrine warranted immediate review. In the course of requesting the certification, Premier conceded that it did not expect to be able to show any other injury — at least not enough other injury to make further litigation worthwhile. The district court took this as a concession that the case is over and entered judgment for the defendants.
II
We start with the district court’s conclusion that Premier is entitled to the benefit of the Maryland court’s holding that the defendants violated the Sherman Act. If the question of preclusion had arisen in 1967, there would have been a ready answer. Most federal courts would apply estoppel only among the parties to the original suit. This is the mutuality requirement — the rule that unless a party would have obtained the full benefit of any victory against the person relying on preclusion, it does not bear the risk of loss. See *362Bigelow v. Old Dominion Copper Co., 225 U.S. 111, 127, 32 S.Ct. 641, 642, 56 L.Ed. 1009 (1912); Restatement of Judgments § 93 (1942). Had the settlement broken down and the defendants won this case in the Supreme Court, Premier would not have been bound. Until recently, that would have prevented Premier from taking advantage of a loss by the defendants in the Maryland case.
We choose 1967 as the benchmark because the preceding year the Supreme Court rewrote Fed.R.Civ.P. 23. One of the complaints about the old Rule 23 was that it allowed courts to entertain what were called “spurious class actions” — actions for damages in which a decision for or against one member of the class did not inevitably entail the same result for all. One party could style the case a “class action”, but the missing parties would not be bound. A victory by the plaintiff would be followed by an opportunity for other members of the class to intervene and claim the spoils; a loss by the plaintiff would not bind the other members of the class. (It would not be in their interest to intervene in a lost cause, and they could not be bound by a judgment to which they were not parties. Hansberry v. Lee, 311 U.S. 32, 40, 61 S.Ct. 115, 117, 85 L.Ed. 22 (1940).) So the defendant could win only against the named plaintiff and might face additional suits by other members of the class, but it could lose against all members of the class. This came to be known as “one-way intervention”, which had few supporters. A principal purpose of the 1966 revision of Rule 23 was to end “one-way intervention”. See the Advisory Committee’s note to new Rule 23(c)(3), and, e.g., C. Wright, A. Miller & M. Kane, 7B Federal Practice and Procedure § 1789 at 266-67 (2d ed. 1986). See also H. Kalven & M. Rosenfield, The Contemporary Function of the Class Suit, 8 U.Chi.L.Rev. 684 (1941).
The drafters of new Rule 23 assumed that only parties could take advantage of a favorable judgment. Given that assumption, it was a simple matter to end one-way intervention. First, new Rule 23(b)(3) eliminated the “spurious” class suit and allowed the prosecution of damages actions as class suits with preclusive effects. Second, new Rule 23(c)(3) required the judgment in a Rule 23(b)(3) class action to define all members of the class. These members of the class were to be treated as full-fledged parties to the case, with full advantage of a favorable judgment and the full detriments of an unfavorable judgment. Third, new Rule 23(c)(1) required the district courts to decide whether a case could proceed as a class action “as soon as practicable” after it was filed. The prompt decision on certification would both fix the identities of the parties to the suit and prevent the absent class members from waiting to see how things turned out before deciding what to do. Finally, new Rule 23(c)(2) allowed members of a 23(b)(3) class action to opt out immediately after the certification in accordance with 23(c)(1). So a person’s decision whether to be bound by the judgment — like the court's decision whether to certify the class — would come well in advance of the decision on the merits. Under the scheme of the revised Rule 23, a member of the class must cast his lot at the beginning of the suit and all parties are bound, for good or ill, by the results. Someone who opted out could take his chances separately, but the separate suit would proceed as if the class action had never been filed. As the Advisory Committee put it: “Under proposed subdivision (c)(3), one-way intervention is excluded; the action will have been early determined to be a class or a nonclass action, and in the former case the judgment, whether or not favorable, will include the class”.
The drafters of new Rule 23 did not anticipate that courts would give preclusive effect to judgments in the absence of mutuality. Blonder-Tongue Laboratories, Inc. v. University of Illinois Foundation, 402 U.S. 313, 91 S.Ct. 1434, 28 L.Ed.2d 788 (1971), and Parklane Hosiery Co. v. Shore, 439 U.S. 322, 99 S.Ct. 645, 58 L.Ed.2d 552 (1979), which severely curtailed the mutuality doctrine in federal litigation, washed away the foundation on which the edifice of Rule 23 had been built. A rule requiring each person’s decision whether to be bound by the judgment to precede the decision on *363the merits works only when the choice is conclusive. The curtailment of the mutuality requirement meant that a decision to opt out might not be conclusive. The drafters also did not anticipate the degree to which district judges would disregard Rule 23(c)(1). Rules 23(c)(1) and (2) together force class members to choose the binding effect of the judgment in advance of decision on the merits. (If they can choose later, it’s one-way intervention all over again.) But district courts frequently postpone deciding whether a case may be maintained as a class action until the case has been settled or a decision has been rendered on the merits. That is what happened in the Maryland litigation. The district judge decided the merits and certified the case as a class action simultaneously, more than three years after it had been filed. The judge then did not give the Rule 23(c)(2) notice for another 2lk years, by which time the judgment had been affirmed and the case had been settled. So by the time Premier and the other members of the class were asked to choose, they knew how the case had come out. Delay in certifying the class is regrettably frequent, even though appellate courts continually condemn the practice. E.g., Glidden v. Chromalloy American Corp., 808 F.2d 621 (7th Cir.1986); Watkins v. Blinzinger, 789 F.2d 474, 475 n. 3 (7th Cir.1986) (collecting other cases, which in turn collect still more).
The district court in Chicago concluded that these two unanticipated developments make all the difference. (Actually the district court did not refer to Rule 23, but its holding logically entails a belief that the decision made in 1966 to do away with one-way intervention is no longer binding.) It relied principally on Parklane, which allowed a private plaintiff in a securities case to obtain the benefit of a judgment against the defendant in an action brought by the Securities and Exchange Commission. The Court criticized the mutuality doctrine for “failing to recognize the obvious difference in position between a party who has never litigated an issue and one who has fully litigated and lost”, 439 U.S. at 327, 99 S.Ct. at 649, and for wasting the time of courts (and parties) on the way to a duplication of the initial result. Even if the second case differed from the result of the first, this might mean that the second decision was the error. If the first litigation is complete, and the parties have good reason to present their cases fully, the decision is as likely to be accurate as any later outcome, and there is accordingly no reason to endure a series of similar cases. The application of the first outcome to all parties is as accurate as a sequence of cases with varying outcomes. Still, the Court recognized, the use of non-mutual, offensive issue preclusion might leave the defendant being pecked to death by ducks. One plaintiff could sue and lose; another could sue and lose; and another and another until one finally prevailed; then everyone else would ride on that single success. This sort of sequence, too, would waste resources; it also could make the minority (and therefore presumptively inaccurate) result the binding one. Parklane therefore gave the district courts discretion not to use offensive issue preclusion. It stated: “The general rule should be that in cases where a plaintiff could easily have joined in the earlier action or where ... the application of offensive estoppel would be unfair to a defendant, a trial judge should not allow the use of offensive collateral estoppel.” Id. at 331, 99 S.Ct. at 651-52.
The parties concentrate their attention on this phrase, from which each draws comfort. Premier points out that both the Panel on Multidistrict Litigation and the Maryland court declined to allow it to press its theory of damages in Maryland. The defendants observe that Premier was a party to the Maryland case for six years before voluntarily opting out, and that it could have raised its damages theory there had it not slept through most of that period. The district judge in Chicago thought Premier closer to the mark. We doubt, however, that Parklane contains the answer to our problem. The Supreme Court did not address the extent to which class members may use issue preclusion after opting out of the class; indeed it did not mention Rule 23. The rules that govern *364the extent to which one judgment in a federal case precludes litigation in a second federal case are part of the federal common law. Cf. University of Tennessee v. Elliott, — U.S.-, 106 S.Ct. 3220, 3225-27, 92 L.Ed.2d 635 (1986) (even the effect of state judgments in federal litigation is a matter of common law when 28 U.S.C. § 1738 is inapplicable). Issue preclusion is made available when it is sound to do so in light of the effects on the rate of error, the cost of litigation, and other instrumental considerations. When there are good reasons to allow relitigation of a category of disputes, preclusion does not apply. So the Court held in United States v. Mendoza, 464 U.S. 154, 104 S.Ct. 568, 78 L.Ed.2d 379 (1984), that offensive non-mutual issue preclusion may not be invoked against the federal government.
If the scope of issue preclusion is a matter of federal common law, then Parklane is not a sufficient reason to upset the balance struck in Rule 23. Under the Rules Enabling Act, 28 U.S.C. § 2072, the Rules of Civil Procedure have the effect of statutes. A development in the common law of judgments is not a reason to undo a statute, to treat a thorough rethinking of the law as so much fluff. The revision of Rule 23 in 1966 does away with one-way intervention in class actions. It should stay done-away-with until the Supreme Court adopts a new version. Whether class members should get the benefit of a favorable judgment, despite not being bound by an unfavorable judgment, was considered and decided in 1966. That decision binds us still.
One could reply that the effect of the 1966 revision on one-way intervention was just a supposition of the drafters. They enacted the rule, not its effects, and the translation from rule to effects depended on mutuality of estoppel. When the mutuality requirement died, the effects of Rule 23 changed. It is not judicial legislation to recognize this alteration, so long as all of the actual commands of Rule 23 are obeyed. Compare Illinois Brick Co. v. Illinois, 431 U.S. 720, 733 n. 14, 97 S.Ct. 2061, 2068 n. 14, 52 L.Ed.2d 707 (1977) (distinguishing between a supposition that underlies a statute and the rule of law contained in the statute), with Califano v. Sanders, 430 U.S. 99, 97 S.Ct. 980, 51 L.Ed.2d 192 (1977) (treating a fundamental supposition, without which the statute would have made no sense, as a rule of law). Maintaining the line between the expected consequences of a rule and the contents of the rule itself is one of the most delicate tasks of the judicial system. Doubtless the Framers of the Constitution expected that a consequence of their structure of government would be a federal government small in relation to the states, and they predicted that political processes would protect property from regulation or expropriation. They did not anticipate the effects of cheap communications and transportation on the size of the national government, but that does not make the governmental structure unconstitutional; they did not predict how the commerce clause could be used to regulate property, see Wickard v. Filbum, 317 U.S. 111, 63 S.Ct. 82, 87 L.Ed. 122 (1942), but that does not make the regulation invalid. Courts have declined to treat the suppositions underlying the Constitution’s structure as part of our governing law— but there are exceptions. The fourth amendment rests on some suppositions about privacy that have been the basis of rules regulating wiretapping, as the supposition underlying the first amendment that willing listeners would be allowed to hear the speech that the amendment protects has been turned into a rule of positive law. All of this is related to the difficult question whether the text, structure, history, or intentions underlying a rule should be given the primary role. That is, when is the rule of law to be found in the words (plus their structure and history), and when are the words just evidence of the “real” rule, which is to be teased out of the purposes of those who wrote and approved the rule and their expectations about the rule’s consequences? Questions of this sort continue to divide the Supreme Court. E.g., Tashjian v. Republican Party of Connecticut,—U.S.-, 107 S.Ct. 544, 555-56, 93 L.Ed.2d 514 (1986), in which the majority treated the unqualified language of Art. I *365§ 2 as simply evidence of the real rule, which would be located in the reasons that text was adopted, while Justices Stevens and Scalia, in dissent, id. at 557-59, treated the language as definitive.
The Supreme Court has not decided when predictions about the effects of a rule should be treated as part of the rule. If suppositions are themselves law, then judges must rummage the minds of the drafters, and what they find there may have more in common with the judges’ beliefs than with the authors’. Worse, a transmutation from “intent” to law bypasses the processes that form the heart of the constitutional method of legislation. Cf. INS v. Chadha, 462 U.S. 919, 103 S.Ct. 2764, 77 L.Ed.2d 317 (1983). Yet laws are designed to produce effects by the application of prescribed means. When the court can both perceive the planned effect and follow the means to the letter, it can fulfill the whole plan, and almost always should. Perhaps it is not possible to reduce to a verbal formula the complex process by which the domain of a rule is fixed. It is unlikely we shall make much headway in this opinion. We need not. Even if a court sometimes may allow the effects of the 1966 revision of Rule 23 to diverge from the course planned by the Advisory Committee, such a modification would not be appropriate when the very problem at hand was anticipated and resolved. See Polk v. Montgomery County, 782 F.2d 1196, 1202 (4th Cir.1986) (holding that a class member who opts out may not take advantage of a judgment favoring the class); Sarasota Oil Co. v. Greyhound Leasing & Financial Corp., 483 F.2d 450, 452 (10th Cir.1973) (same, although before Parklane). But see Saunders v. Naval Air Rework Facility, 608 F.2d 1308, 1312 (9th Cir.1979) (stating, without much discussion, that a class member who opts out of the damages portion of a class suit seeking an injunction may use the preclusive force of the decision granting an injunction as the jumping-off point in pursuit of damages).
We also lack a sound reason to deviate from the plan of 1966. The district court concluded that application of issue preclusion would produce judicial economy. 627 F.Supp. at 962-63. See also In re Transocean Tender Offer Securities Litigation, 455 F.Supp. 999 (N.D.Ill.1978). “Judicial economy” sounds like a sure-fire Good Thing — especially to the ears of judges. Conservation of resources is the principal reason the Supreme Court gave for its decision in Parklane. And it has been used as a reason for preclusion when there are two actions, one for injunction and one for damages. For example, we held in Crowder v. Lash, 687 F.2d 996, 1011-12 (7th Cir.1982), that a member of the class in an injunctive case may use the findings underlying the grant of injunctive relief to preclude defendants from relitigating certain issues in a separate suit for damages. Cf. Cooper v. Federal Reserve Bank of Richmond, 467 U.S. 867, 104 S.Ct. 2794, 81 L.Ed.2d 718 (1984) (a judgment that an employer did not discriminate against an entire class does not necessarily preclude a member of the class from showing, in a separate action, that the employer discriminated against her, because the issues in the pattern-or-practice class suit may be different from the contentions in the individual action). When there are bound to be two separate actions, as we assumed in Crowder, a plaintiff who does not opt out of the injunctive action is entitled to keep his victory, as he will be bound by defeat. (Crowder is a case of mutual estoppel.) It serves the interest of economy to have as few issues litigated in the second suit as possible. The district court took our case as a similar situation. Yet this assumes that there will be two suits against the Association, Union, and Fund. Parklane was a case of this type. The SEC’s suit and the private suit were bound to go forward independently; nothing the court could do would reduce the number of cases that had to be litigated. If there are bound to be two suits, why not cut down on the number of issues? The district court treated this case as similar to Parklane, but it is not. Here there were not bound to be two suits. There were two suits in fact, but there need not have been.
An approach that asks how to hold down the costs of litigation given the existence of *366multiple suits is an ex post perspective on judicial economy. It is the wrong perspective when inquiring about the consequences of a legal rule. A decision to make preclusion available to those who opt out of a class influences whether there will be multiple suits. The more class members who opt out may benefit from preclusion, the more class members will opt out. Preclusion thus may increase the number of suits, undermining the economy the district court hoped to achieve. The effect of the legal rule may be the opposite of the effect of applying preclusion to a given case. To determine whether a rule is beneficial, a court must examine how that rule influences future behavior. The influence of a rule of preclusion cannot be known for sure, but we are not confident that there would be net benefits.
This case makes the point. The class action in Maryland lasted six years. If Premier had known that it must start from scratch in Chicago, it might well have stayed in the Maryland litigation. It had the right under Rule 23(e) to protest the settlement and obtain a decision about its adequacy. Instead of finishing the litigation in Maryland, Premier walked away to try again, fortified by its belief that it could win in Chicago but not lose. If others behave similarly, the application of issue preclusion would multiply the number of suits and undermine the benefit of class actions in centralizing litigation. Reducing the cost to a class member of carrying the litigation to another forum also would complicate and reduce the number of settlements. The defendants in Maryland thought that for $6 million and an injunction they would purchase peace. Premier wants to deny defendants that boon, but not to refund any of the $6 million. If enough other class members had opted out, the settlement would have collapsed, and the Maryland litigation would have dragged on — perhaps before the Supreme Court, perhaps in the district court as the defendants contested damages. If defendants anticipate significant opting out, they also will reduce the amounts they offer in settlement, which may in turn make it worthwhile for more parties to opt out. The more attractive it is to opt out — and giving the parties who opt out the benefit of preclusion makes it very attractive — the fewer settlements there will be, the less the settlements will produce for the class, and the more cases courts must adjudicate. This is not judicial economy at work!
Perhaps this is misleading. There may be economies in adjudicating liability in a single action while having separate actions to determine damages. See Note, Offensive Assertion of Collateral Estoppel by Persons Opting Out of a Class Action, 31 Hastings L.Rev. 1189 (1980) (arguing that issue preclusion should be available to those who opt out if and only if their theory of recovery is sufficiently distinctive from that pressed by the representative plaintiffs). The potential saving from breaking one complex case down into simpler ones accounts for decisions such as Crowder, which, allow the plaintiff in a damages case to take advantage of the outcome of an injunctive case. But see Cooper, which, in declining to bind the single plaintiff by the adverse result of a class action seeking an injunction on a different theory of liability, establishes that a sequence of suits with preclusive effect is not always preferable.
We doubt that there are benefits in separating questions of liability and damages when the first class action also seeks damages. Different members of the class may suffer different kinds of damages, but this is a reason to establish subclasses (or to appeal the approval of a one-sided settlement) rather than to increase the number of separate suits. If the result of the first class action binds only the parties, then class members who have different theories of damages will be induced to present them as early as possible in order to have a subclass certified. If the differences in the kinds of damages suffered are indeed substantial, perhaps it is mistaken to certify any class. See Rule 23(b)(3) (class action appropriate only if common questions predominate). If, as Premier contends, a class member dissatisfied by the measure of damages selected by the representative party may pick up his portfolio and start *367again, there will be an incentive to stand on the sidelines and see how things turn out. If the first case proceeds well, the bystander will take the benefit, and if not it will try again. The benefits of presenting common claims in a single forum will be lost. Premier should have presented its theory of damages in the Maryland suit as quickly as possible, because it cast doubt on the district court’s belief (498 F.Supp. at 545, 549-50) that common questions predominated and that damages could be computed mechanically. It would be an unwelcome development to induce class members to keep to themselves reasons to doubt the propriety of class certifications.
One more consideration. The district court believed that issue preclusion would ensure equal treatment of members of the class. Equality, like judicial economy, is desirable; the judicial system uses a variety of devices, including stare decisis and rules of preclusion, to treat likes alike. But the threat to equal treatment of class members came from Premier itself. The class members who stayed in the Maryland litigation were treated equally. Premier wanted to be treated differently; it wanted a unique measure of damages. This measure might well be appropriate (we discuss this below); it might even produce “equality” from a different standpoint. “Equality” is an open term. See P. Westen, The Empty Idea of Equality, 95 Harv.L.Rev. 537 (1982). Equal with respect to what is the essential question, the answer to which must come from some independent source. Maybe only a unique measure of damages will lead to equal treatment with respect to injury suffered (that is, to recompense for an equal percentage of the damage incurred). Still, Premier’s decision to opt out is the best indicator that it wanted special treatment. It wanted a greater recovery without risk. It could have had equal treatment (though not necessarily with identical effect) by staying in the Maryland case.
We conclude that class members who opt out may not claim the benefits of the class’s victory. Like the Supreme Court’s decision in Mendoza, this is a categorical rule. The Court forbade application of nonmutual estoppel of the federal government, avoiding any need to determine the scope of the district court’s discretion. Here, too, the rule affects a category of cases — the one settled by the 1966 revision of Rule 23.
Ill
To say that issue preclusion does not apply is not to say that this case must follow the dreary course of many antitrust matters and bury the court under mountains of documents, to be followed by an interminable trial. Preclusion is not an all-or-nothing matter; there are degrees. The doctrine of stare decisis supplies some of the lesser degrees. A decision by the Supreme Court that the Agreement violates the Sherman Act would be authoritative, precluding further contention in Chicago. A decision by the Fourth Circuit is not authoritative in district courts of the Seventh Circuit, but it is entitled to respect, both for its persuasive power and because it involves the same facts. Colby v. J.C. Penney Co., 811 F.2d 1119, 1123-24 (7th Cir.1987). The application of stare decisis will produce most of the judicial economy the district court sought to achieve.
The value of the first decision as a prediction of how other courts will act produces part of the savings. If the class wins, the opting-out plaintiff should expect to win for the same reasons that persuaded the first court. If the class loses, the opting-out plaintiff should expect to meet the same fate. He may drop the suit; if he presses the suit, the earlier decision again shortens the path to disposition of the second case. Stare decisis should be particularly potent in cases suitable for class treatment. The conclusion of the first court that certification of a class is proper, if correct, means that there will be few significant differences between the class suit and the opt-out suit. The second litigation may safely concentrate on those differences, whether or not issue preclusion comes into play.
When as here the defendant’s activities span more than one court of appeals, only *368the gravest reasons should lead the court in the opt-out suit to come to a conclusion that departs from that in the class suit. This circuit pays respectful attention to the decisions of others, to the point of suppressing doubts in order to prevent the creation of a conflict, e.g., FDIC v. Elefant, 790 F.2d 661, 665-66 (7th Cir.1986), or overruling existing cases to eliminate a conflict, e.g., Bauzo v. Bowen, 803 F.2d 917, 920-22 (7th Cir.1986). The benefits of following the decision of another circuit are particularly apparent when the two courts are dealing with the same set of facts. A conflict among the circuits about a single collective bargaining agreement or other common endeavor may compel the Supreme Court to hear the case even though the dispute lacks independent significance. E. g., CPSC v. GTE Sylvania, Inc., 447 U.S. 102, 100 S.Ct. 2051, 64 L.Ed.2d 766 (1980) (resolving a conundrum in which one court of appeals had directed the CPSC to disclose some documents that another court of appeals had forbidden it to reveal). We therefore approach the merits of this case with a strong presumption in favor of the Fourth Circuit’s disposition. The presumption does not eliminate the need for independent analysis, but it does mean that doubts should be resolved in favor of the Fourth Circuit’s disposition.
The Fourth Circuit held that the 1% contribution requirement is price-fixing, a per se violation of the antitrust laws. 678 F.2d at 500-01. The Association enlisted the Union in collecting the 1% from nonmembers, the Maryland district court concluded, 498 F.Supp. at 533-35, when the nonmembers began underbidding members. The nonmembers, free of the 1% contribution, had lower costs of doing business and therefore could make money at lower prices. The Association used the Union to increase its rivals’ costs of doing business, the better to eliminate a source of competition. The result was higher prices to purchasers of electrical work and higher profits for members of the Association — both because there is more in the Fund, for the Association’s use, and because the reduction in competition enabled the members to capture more of the market. The extra profits could be shared with the Union, compensating it for the assistance. The principal purpose of the antitrust laws is to prevent overcharges to consumers, see Hospital Corp. of America v. FTC, 807 F.2d 1381, 1386 (7th Cir.1986), and Illinois Corporate Travel, Inc. v. American Airlines, Inc., 806 F.2d 722, 727-28 (7th Cir.1986) (both collecting cases); cf. Cargill, Inc. v. Monfort of Colorado, Inc.,—U.S.-, 107 S.Ct. 484, 489-92, 93 L.Ed.2d 427 (1986) (stressing that the antitrust laws are designed to protect consumers rather than competitors). The Association raised its rivals’ costs, and thereby raised the market price to its own advantage. See T. Krattenmaker & S. Salop, Anticompetitive Exclusion: Raising Rivals’ Costs to Achieve Power Over Price, 96 Yale L.J. 209 (1986).
The defendants’ response is not so much to deny this characterization as to justify the requirement that every firm in the industry contribute to the Fund. According to the defendants the Fund supports collective bargaining that benefits employers, so there are net savings. This is a problematic argument on both legal and factual grounds. Legal because Catalano, Inc. v. Target Sales, Inc., 446 U.S. 643, 100 S.Ct. 1925, 64 L.Ed.2d 550 (1980), holds that firms cannot justify an agreement fixing the terms of credit (one element of the price of goods) on the argument that other elements of the price will adjust; factual because if the Association were right, then the Union would not have participated. New unions want to be the instruments of their own demise. The defendants’ contention is a variation on the “free riding” argument that has been accepted in cases such as Continental T.V., Inc. v. GTE Sylvania Inc., 433 U.S. 36, 97 S.Ct. 2549, 53 L.Ed.2d 568 (1977). See also Illinois Corporate Travel, 806 F.2d at 728-29. The Fund supplies benefits to the whole industry; firms that do not belong to the Association receive the benefits without paying for them; this is a free ride, and the agreement in question simply requires the free riders to pay for their goodies.
Among the defects in this argument is a misunderstanding of the role of free-riding *369analysis in antitrust law. It is usually used in vertical chains of distribution. A manufacturer decides how much competition to allow among its dealers. If too many dealers sell in the same territory, and one supplies presale services, certification of a product’s quality, or other things for which it cannot charge separately, another may be tempted to skimp on these things. The shirking dealer may hope the customer visits the first dealer, obtains the benefits, and then buys from the shirker for the lower price made possible by the skimping. The situation will degenerate as the first dealer finds the provision of services unprofitable. To prevent everyone from abandoning the service that consumers value, the manufacturer may limit who can sell the goods and where. See Monsanto Co. v. Spray-Rite Service Corp., 465 U.S. 752, 762-64, 104 S.Ct. 1464, 1470-71, 79 L.Ed.2d 775 (1984); L. Telser, Why Should Manufacturers Want Fair Trade?, 3 J.L. & Econ. 86 (1960); V. Goldberg, The Free Rider Problem, Imperfect Pricing, and the Economics of Retailing Services, 79 Nw.U.L.Rev. 736 (1984). The grant of an exclusive territory or group of customers may raise the retailer’s markup (the difference between wholesale and retail price), but non-price competition will continue.
From the manufacturer’s perspective, the higher markup is a cost of doing business. Manufacturers do not particularly like higher costs. Other things equal, they would prefer a markup just large enough to keep their dealers out of bankruptcy. Just as manufacturers want to pay less for steel and labor, so they want to pay less for distribution. A manufacturer therefore adopts a restricted distribution policy only when the higher markup enables the dealers to furnish consumers with something the average consumer values by more than the incremental markup. Otherwise the manufacturer will sell through discount chains that collect the lowest markup. In other words, in selecting a method of distribution, manufacturers act in the consumers’ interest. Neither wants a higher markup unless that is beneficial to purchasers. If the manufacturer adopts a restricted distribution policy in order to serve its own interests, therefore, there is no reason to doubt that the policy is beneficial to consumers and consistent with the purposes of antitrust law. Perhaps a given manufacturer is mistaken, but if so the reaction of consumers (and the lower profits) will send the message faster than courts could. Judges know less than manufacturers about optimal distribution arrangements, so it is unlikely that they could improve the position of consumers by forcing manufacturers to adopt distribution policies manufacturers have not favored voluntarily.
This rationale depends on the alignment of interests between consumers and manufacturers. Destroy that alignment and you destroy the power of the argument. If the dealers meet and agree on the price they will charge or the territories they will occupy, and then induce the manufacturer to go along, that agreement will be illegal per se. Monsanto, 465 U.S. at 765-68, 104 S.Ct. at 1471-72; GTE Sylvania, 433 U.S. at 58 n. 28, 97 S.Ct. at 2561 n. 28. See also United States v. Capitol Service, Inc., 756 F.2d 502 (7th Cir.), cert. denied,—U.S.-, 106 S.Ct. 811, 88 L.Ed.2d 288 (1985) (holding a horizontal market allocation among movie exhibitors unlawful per se despite a claim that the allocation would reduce free riding.) It is so here too. No one involved in establishing and collecting the 1% fee had consumers’ interest at heart. No automatic mechanism corrects blunders. If the value of the services the Fund provides to firms that do not belong to the Association is .01% of their payroll, the collection of a 1% fee produces a supra-competitive profit of 0.99%, and no market force (other than cheating by firms that try to escape the fee) will induce the Association to reduce it. The Fund will act as a profit pool, with the 1% payment a “bonus” paid by successful bidders after the fashion of United States v. Addyston Pipe & Steel Co., 85 F. 271 (6th Cir.1898) (Taft, J.), aff'd, 175 U.S. 211, 20 S.Ct. 96, 44 L.Ed. 136 (1899). The arrangement covers 100% of the market, so that other forms of organization cannot *370readily undercut it.2 So despite the incentive for firms to take free rides on the activities of trade associations, see M. Olson, The Logie of Collective Action (1965), and despite the potential gains from having all affected firms contribute to any benefits the Fund produces, this is not the sort of justification that a court may accept. To do so would abolish the per se rule against horizontal price fixing. See H.A. Artists & Associates v. Actors’ Equity Ass’n, 451 U.S. 704, 722, 101 S.Ct. 2102, 2112, 68 L.Ed.2d 558 (1981) (rejecting a similar justification by a union for collecting fees from non-members).
To put the point slightly differently, the “automatic” force that undercuts price-fixing arrangements is “cheating”. Each seller has an incentive to shave the price a little in order to sell more; to maintain their sales other firms must reduce price too; the cartel desperately wants to halt the price cutting, which if continued will drive all sellers down to the competitive price. From the perspective of the cartel, the price cutting by individual firms is a form of “free riding”. The free rider is taking advantage of the supra-competitive price fixed by agreement; knowing that other firms are charging a monopoly price, it charges a little less and makes a handsome profit. This process spreads and ultimately destroys the cartel. So in price-fixing cases “free riding” is the way the cartel unravels. A group of firms trying to extract a supra-competitive price therefore hardly can turn around and try to squelch lower prices — as the Association may have done — by branding the lower prices “free riding”!
This is not to say that all cooperation among firms at the same level of industry is unlawful. It is not. See Broadcast Music, Inc. v. Columbia Broadcasting System, Inc., 441 U.S. 1, 99 S.Ct. 1551, 60 L.Ed.2d 1 (1979) (BMI); NCAA v. Board of Regents, 468 U.S. 85, 104 S.Ct. 2948, 82 L.Ed.2d 70 (1984); United States Trotting Ass’n v. Chicago Downs Ass’n, 665 F.2d 781 (7th Cir.1981) (en banc). But these cases applied the Rule of Reason to forms of cooperative behavior. The “blanket license” sold by performing rights societies in BMI, for example, was a way of vending music that competed with an unencumbered system of direct licenses; it flourished only to the extent blanket licenses were cheaper and more convenient for customers than were individual, direct licenses. In NCAA the Court applied the Rule of Reason to a sports league, on the ground that sports depends on some organization; so too in United States Trotting, in which one horse league cooperated internally in order to compete with thoroughbred racing, football, and other entertainments. Agreements of this sort are ancillary restraints, and the cooperation underlying the restraint has the potential to create the efficient production that consumers value. See Rothery Storage & Van Co. v. Atlas Van Lines, Inc., 792 F.2d 210 (D.C. Cir.1986); Polk Bros., Inc. v. Forest City Enterprises, Inc., 776 F.2d 185, 188-91 (7th Cir.1985).
Organization is a precondition to competition; few products are produced by solitary artisans. The appropriate scope of organization may be very hard to establish, and since BMI courts have been appropriately modest about their ability to discern the optimal amount of cooperation in an industry. Cases such as BMI, United States Trotting, Rothery, and Polk Bros, do not eliminate the requirement that there be some productive cooperation as a condition of the application of the Rule of Reason, at least when producers with a high market *371share agree on price. See Arizona v. Maricopa County Medical Society, 457 U.S. 332, 348-52, 102 S.Ct. 2466, 2475-77, 73 L.Ed.2d 48 (1982). The agreement alleged in this case affected 100% of an economically significant market; it was a naked agreement on price; enforcement mechanisms were in place; the free-riding justification given for the agreement does not identify any actor who will protect the consumers’ interests and hold the fee to the level justified by the benefits of the Fund. There is no escape from the conclusion of the Fourth Circuit that the agreement is unlawful per se.3
That is, if there was an agreement. We have assumed so far that Premier will be able to prove its claims, as the class representatives did in Maryland. The Union denies making the agreement. Although Article 6 of the 1976 contract between the Union and the Association provides that “[a]ll construction agreements in the electrical industry shall contain” the 1% fee, the Union contends that this is advisory rather than directory language. Premier responds that the Union acted as if it were mandatory, because the Union crammed the clause down Premier’s throat three years running (resulting in the three suits to collect). Perhaps this dispute will prove amenable to summary judgment in Illinois, as it was in Maryland. But because we have held that principles of issue preclusion do not apply, we must afford the Union the opportunity to make good its offer to show that Art. 6 did not mean what it said. If the district court concludes that Art. 6 should be given its natural meaning, then it must declare the agreement a violation of § 1 of the Sherman Act.
IV
This whole discussion is beside the point unless Premier stands to gain from a declaration of illegality. The district judge dismissed the complaint because, she concluded, Premier cannot establish damages. 627 F.Supp. at 964-66. The barrier, the court held, is the Noerr-Pennington doctrine, named after Eastern Railroad Presidents Conference v. Noerr Motor Freight, Inc., 365 U.S. 127, 81 S.Ct. 523, 5 L.Ed.2d 464 (1961), and United Mine Workers v. Pennington, 381 U.S. 657, 85 S.Ct. 1585, 14 L.Ed.2d 626 (1965). See also California Motor Transport Co. v. Trucking Unlimited, 404 U.S. 508, 92 S.Ct. 609, 30 L.Ed.2d 642 (1972). The Noerr-Pennington doctrine is that the antitrust laws do not apply to efforts to persuade the government to organize or support a cartel. Noerr involved a deceptive lobbying campaign by the railroad industry that led the Governor of Pennsylvania to veto a bill that would have helped the trucking industry. The Court held that cooperative, deceitful acts directed at the political branches of government stand outside antitrust, in part because the original purposes of the Sherman Act did not include regulating political activity and in part because it is questionable whether the first amendment allows such regulation. Pennington extended the doctrine to efforts to influence administrative agencies.
California Motor identified the first amendment as the principal source of the Noerr-Pennington doctrine, see 404 U.S. at 510, 92 S.Ct. at 611, and extended it still further to the conduct of litigation. See D. Fischel, Antitrust Liability for Attempts to Influence Government Action: The Basis and Limits of the Noerr-Pennington Doctrine, 45 U.Chi.L.Rev. 80, 94-104 (1977); J. Hurwitz & D. Neveu, The Noerr Doctrine: Its Significance and Current Interpretation, in The Political Economy of Regulation: Private Interests in the Regulatory Process 33, 47-50 (FTC 1984). See also Otter Tail Power Co. v. United States, 410 U.S. 366, 379-80, 93 S.Ct. 1022, 1030-31, 35 L.Ed.2d 359 (1973); Vendo Co. v. Lektro-Vend Corp., 433 U.S. 623, 97 S.Ct. 2881, 53 L.Ed.2d 1009 (1977); and Bill *372Johnson’s Restaurants, Inc. v. NLRB, 461 U.S. 731, 741-44, 103 S.Ct. 2161, 2168-70, 76 L.Ed.2d 277 (1983), all applying the Noerr-Pennington doctrine to litigation. The Court elaborated on a hint dropped in Noerr that “[t]here may be situations in which a publicity campaign, ostensibly directed toward influencing governmental action, is a mere sham to cover what is actually nothing more than an attempt to interfere directly with the business relationships of a competitor and the application of the Sherman Act would be justified.” Noerr, 365 U.S. at 144, 81 S.Ct. at 533. California Motor concluded that baseless litigation brought in bad faith for the purpose of obstruction, and without reasonable prospect of success, would be a sham within the Noerr hint. We elaborated further in Grip-Pak, Inc. v. Illinois Tool Works, Inc., 694 F.2d 466 (7th Cir.1982), cert. denied, 461 U.S. 958, 103 S.Ct. 2430, 77 L.Ed.2d 1317 (1983), holding that a suit brought only because of the costs litigation imposes on the other party also may fit the “sham” exception to the Noerr-Pennington doctrine. We explained, id. at 472: “Many claims not wholly groundless would never be sued on for their own sake; the stakes, discounted by the probability of winning, would be too low to repay the investment in litigation.” If the expected value of a judgment is $10,000 (say, a 10% chance of recovering $100,000), the case is not “groundless”; yet if it costs $30,000 to litigate, no rational plaintiff will do so unless he anticipates some other source of benefit. If the other benefit is the costs litigation will impose on a rival, allowing an elevation of the market price, it may be treated as a sham. “[I]f the improper purpose [imposing costs on rivals] is to use the litigation as a tool for suppressing competition in its antitrust sense ... it becomes a matter of antitrust concern.” Ibid. See also Neumann v. Reinforced Earth Co., 786 F.2d 424, 427-28 (D.C.Cir.1986).
The district judge looked for evidence that the Fund’s three suits in state court met this definition of sham. She did not find it. It is not there. The Fund brought the suits to win, not because, defense was expensive to Premier. The Fund wanted its 1% and probably thought it could prevail; after all, Premier had signed a contract. True, an antitrust defense to this kind of contract could block enforcement, see Kaiser Steel Corp. v. Mullins, 455 U.S. 72, 102 S.Ct. 851, 70 L.Ed.2d 833 (1982), and Premier interposed antitrust defenses to all three suits. But the Fund was defending the litigation in Maryland; it ultimately persuaded one judge of the Fourth Circuit. See 678 F.2d at 503-04 (Hall, J., dissenting). Premier did not try to persuade the district court that the Fund’s suits are shams within the scope of GripPak. It argued instead that the litigation was not the violation but only the source of its injury. The contracts between the Association and the Union, and between Premier and the Union, were the violation. Premier maintained that the damage caused by the violation may be collected under § 4 of the Clayton Act, 15 U.S.C. § 15, whether or not the acts creating the injury were themselves a proper basis of liability under the Noerr-Pennington doctrine. The district court replied, 627 F.Supp. at 965:
Premier presents no explanation of why this labeling makes any difference. In this court’s view, it does not matter whether the litigation is seen as itself violative of the antitrust laws or as an outgrowth of other acts which are violative of the antitrust laws. In either instance, defendants’ first amendment rights would be burdened by imposing treble damages as a penalty for undertaking litigation. Preservation of the first amendment interests which underlie the Noerr-Pennington doctrine must require a showing that the litigation falls within the sham exception____
The district judge then dismissed a series of cases in which this court had awarded damages on account of litigation that carried out a violation. E.g., Locklin v. DayGlo Color Corp., 429 F.2d 873 (7th Cir.1970), cert. denied, 400 U.S. 1020, 91 S.Ct. 582, 584, 27 L.Ed.2d 632 (1971); Hazeltine Research, Inc. v. Zenith Radio Corp., 388 F.2d 25 (7th Cir.1967), aff'd in part, rev’d in part on other grounds, 395 U.S. 100, 89 *373S.Ct. 1562, 23 L.Ed.2d 129 (1969); Dairy Foods, Inc. v. Dairy Maid Products Cooperative, 297 F.2d 805 (7th Cir.1961). The district court thought these inconsistent with California Motor, which was decided more recently. The district judge did not discuss City of Mishawaka v. American Electric Power Co., 616 F.2d 976, 989-90 (7th Cir.1980), cert. denied, 449 U.S. 1096, 101 S.Ct. 892, 66 L.Ed.2d 824 (1981), which is more recent still and awards as damages the costs of litigation aimed at preventing an extrinsic violation of the antitrust laws. We reaffirm this holding of Mishawaka and accordingly disagree with the district judge.
The proposition that the first amendment precludes the award of the costs of litigation as damages implies the startling result that fee-shifting rules are unconstitutional. Fee-shifting statutes have become common, and the Sherman Act contains one of the first. They require the loser to pay the winner’s fees. Loser-pays-winner statutes discourage litigation with a low chance of success. See S. Shavell, Suit, Settlement, and Trial: A Theoretical Analysis under Alternative Methods for the Allocation of Legal Costs, 11 J. Legal Studies 55 (1982). Some of the statutes, such as 42 U.S.C. § 1988, have been construed as having a pro-plaintiff bias. These encourage the bringing of low-probability cases, and defendants whose position is vindicated rarely recover their losses. The fee-shifting statute in the Sherman Act is the most biased of all, because only plaintiffs may recover fees. The district court apparently would deem fee-shifting statutes unconstitutional unless the loser’s position was a “sham” within the meaning of the NoerrPennington doctrine. Pro-plaintiff statutes would be even more suspect, because they greatly increase the risks defendants bear in exercising their constitutional right to obtain the court’s decision — they can lose but not win. Other statutes and rules affecting costs in litigation also greatly affect the incentive to present one’s case in court. See G. Miller, An Economic Analysis of Rule 68, 15 J. Legal Studies 93 (1986).
Yet for all this, the proposition that the first amendment, or any other part of the Constitution,4 prohibits or even has anything to say about fee-shifting statutes in litigation seems too farfetched to require extended analysis. Fee shifting requires the party that creates the costs to bear them.5 See In re TCI Ltd., 769 F.2d 441, 445-48 (7th Cir.1985); Coleman v. CIR, 791 F.2d 68, 71-73 (7th Cir.1986). This is no more a violation of the first amendment than is a requirement that a person who wants to publish a newspaper pay for the ink, the paper, and the press. Similarly, whoever wants to read the New York Times must buy a copy. The exercise of rights may be costly, and the first amendment does not prevent the government from requiring a person to pay the costs incurred in exercising a right. See also Harris v. McRae, 448 U.S. 297, 100 S.Ct. 2671, 65 L.Ed.2d 784 (1980); Ross v. Moffitt, 417 U.S. 600, 94 S.Ct. 2437, 41 L.Ed.2d 341 (1974). A discovery request that may cost a million dollars to fulfill does not violate the Constitution as a price tag on the exercise of the right of access to the courts. So long as the violation of the Sherman Act may be established without regard to point of view embodied in the “petitioning” activity, the Constitution does not prevent the assignment as damages of the full injury inflicted.
Premier wants more than its fees and costs; it wants them trebled. This does not make a constitutional difference, however. Multipliers are common in fee shifting. E.g., Ohio-Sealy Mattress Mfg. Co. v. Sealy Inc., 776 F.2d 646, 659-64 (7th Cir.1985) (doubling). The multiplier makes up *374for delay in payment and for the risk of failure.6 Trebling damages in antitrust may serve the same functions. Payment in antitrust litigation comes long after the injury, and in the interim the defendant had the benefit of the money. Moreover, not all violations are caught and prosecuted successfully. This violation was in the open, making trebling less important (although still required by statute), but successful prosecution was not assured. When there is a risk that the violation will succeed, the damages must be increased to achieve sufficient deterrence and to compensate those who were injured. See W. Landes, Optimal Sanctions for Antitrust Violations, 50 U.Chi.L.Rev. 652, 657 (1983). The expected sanction the defendant faces in a regime of trebling is not necessarily more than the full costs of its conduct. Sometimes it will escape any penalty, and sometimes the penalty will come so late that even trebling of damages leaves the defendant with a profit. See the examples given in Fishman v. Estate of Wirtz, 807 F.2d 520, 584 (7th Cir.1986) (separate opinion). There was a risk that this violation would escape punishment. The dissenting opinion in the Fourth Circuit demonstrates that this was not open-and-shut litigation. And there have been six years of delay, during which a dollar invested would have earned a healthy return. Trebling under the antitrust laws is no more a constitutional problem than is a multiplier when attorneys’ fees are awarded under 42 U.S.C. § 1988.
The award as damages of the expenses incurred in resisting litigation has support in the Supreme Court’s decisions. For example, Walker Process Equipment, Inc. v. Food Machinery & Chemical Corp., 382 U.S. 172, 86 S.Ct. 347, 15 L.Ed.2d 247 (1965), said that the enforcement through litigation of a patent obtained by fraud on the Patent Office may violate the Sherman Act. This case was decided the Term after Pennington; the Court evidently saw no contradiction. See also United States v. Singer Manufacturing Co., 374 U.S. 174, 83 S.Ct. 1773, 10 L.Ed.2d 823 (1963) (liability on account of the settlement of private litigation). Cases of this sort furnished a basis of our decision in Mishawaka and our earlier opinions. And although Bill Johnson’s reiterates the application of the Noerr-Pennington doctrine to litigation, the Supreme Court did not doubt the power of the NLRB to impose sanctions on an unsuccessful litigant, once the Board had decided that the litigation was part of a campaign of unfair labor practices.7
There is an important difference, for purposes of the Noerr-Pennington doctrine, between using litigation (or other petitions to the government) as a basis of antitrust liability and awarding damages for efforts to use the courts to carry out private cartel agreements. The Noerr-Pennington doctrine protects private efforts to enlist the government as part of private activity. The Sherman Act expresses one policy; people are free to try to persuade their representatives that monopoly is preferable. It would be bizarre if efforts to repeal the Sherman Act violated the Sherman Act. States, too, may shelter their residents from competitive forces. See Parker v. Brown, 317 U.S. 341, 63 S.Ct. 307, 87 L.Ed. 315 (1943); Southern Motor Carriers Rate Conference v. United States, 471 U.S. 48, 105 S.Ct. 1721, 85 L.Ed.2d 36 (1985). The first amendment protects the right of the people to ask for this boon. The Framers anticipated that interest groups would try to use the government for their own interests. E.g., Federalist No. 10 (Madison). The constitu*375tional structure buffers the effects of this propaganda, but the first amendment protects the process all the same. Cf. C. Sun-stein, Interest Groups in American Public Law, 38 Stan.L.Rev. 29 (1985). Much legislation protects private interest groups at the expense of the larger public. That is why many people have sought to use antitrust law to undo the effects of state regulation. E.g., Exxon Corp. v. Governor of Maryland, 437 U.S. 117, 133-34, 98 S.Ct. 2207, 2217-18, 57 L.Ed.2d 91 (1978) (recognizing that a state regulatory system is anticompetitive and harmful to consumers but holding that the antitrust laws do not apply);8 New Motor Vehicle Board v. Orrin W. Fox Co., 439 U.S. 96, 109-11, 99 S.Ct. 403, 411-12, 58 L.Ed.2d 361 (1978) (same). If interest groups obtain the protection they seek, Parker v. Brown will insulate them from antitrust liability. The Noerr-Pennington doctrine provides to these self-serving groups protection from the application of antitrust laws to their efforts to obtain this insulation.
Favors are sought and obtained from the political branches of government. The judicial branch implements the policies the political branches create. So if an interest group wins in the legislature a system of regulation that stifles new entry and competition from existing firms — a system of entry and rate regulation, for example, under which each firm may protest the price reduction of any other — it is entitled to claim the benefit in the way provided. If the designated way is indiscriminately to protest tariffs filed by rivals, then these protests (in administrative bodies or courts) are free of liability. The court must honor the decision of the political branches to create a way in which private parties can stifle competition. It is churlish to turn on a firm that puts its political victory to advantage.
Although a court may not use antitrust law to reverse a political victory, it must insist that the victors not claim more than they have been granted. Perhaps a statute conditions the filing of a protest or suit against a tariff on a good-faith belief that the tariff is unlawful under the criteria (which may be anticompetitive) that the court or agency will apply. Then a baseless protest or suit has no shelter from the political victory. It is an effort to cheat the political branches, to do what the law forbids. California Motor holds that such efforts, “baseless” because they claim more than the governing substantive rule allows, are unprotected by the Noerr-Pennington doctrine. So too with perjury and fraud in the conduct of litigation; no statute authorizes such tactics, and the first amendment does not protect efforts to defy (as opposed to change) the substantive rule of decision.
In Otter Tail, on which the defendants rely, the Court applied the Noerr-Pennington doctrine to several suits filed by an electric utility seeking to stifle competition. The Court held, 410 U.S. at 379-80, 93 S.Ct. at 1030-31, that unless these suits were baseless, “shams”, they could not be the basis of liability. See also United States v. Otter Tail Power Co., 360 F.Supp. 451 (D.Minn.1973) (imposing liability on remand), aff'd mem., 417 U.S. 901, 94 S.Ct. 2594, 41 L.Ed.2d 207 (1974). In order to determine whether protests and suits in a thoroughly regulated industry are shams, the court first must determine whether the political branches have authorized the use of the administrative and judicial processes to frustrate rival producers. If the answer is yes, there is no liability; Otter Tail remanded the case so that the appropriate inquiry could be carried out. In Mishawaka, by contrast, our court concluded that litigation costs incurred to combat the consequences of acts exceeding the authorization granted by the political branches could be the basis of liability.
All of this harks back to the quotation in Noerr that is the foundation of the “sham” doctrine. “There may be situations in which a publicity campaign, ostensibly directed toward influencing governmental action, is a mere sham to cover what is actu*376ally nothing more than an attempt to interfere directly with the business relationships of a competitor and the application of the Sherman Act would be justified.” 365 U.S. at 144, 81 S.Ct. at 533. In other words, it is important to identify the source of the injury to competition. If the injury is caused by persuading the government, then the antitrust laws do not apply to the squelching (Parker v. Brown) or the persuasion (Noerr-Pennington). If the injury flows directly from the “petitioning” — if the injury occurs no matter how the government responds to the request for aid — then we have an antitrust case. When private parties help themselves to a reduction in competition, the antitrust laws apply. Thus Cantor v. Detroit Edison Co., 428 U.S. 579, 594-98, 96 S.Ct. 3110, 3119-21, 49 L.Ed.2d 1141 (1976), held that a utility may be liable for carrying out a program that it had proposed to the regulator. The Court concluded that the regulator had no policy of its own, that the rule of decision had been supplied by the private party without supervision of the state. That made it liable for the results, if they were otherwise unlawful. See also Continental Ore Co. v. Union Carbide & Carbon Corp., 370 U.S. 690, 706-08, 82 S.Ct. 1404, 1414-15, 8 L.Ed.2d 777 (1962) (same); Osborn v. Pennsylvania-Delaware Service Station Dealers Ass’n, 499 F.Supp. 553 (D.Del.1980) (a boycott with direct economic effects is prohibited by the antitrust laws; the fact that it also serves a role in petitioning the government does not activate the Noerr-Pennington doctrine).
The antitrust laws allow people to ask the government for a monopoly, and they allow them to keep what they get. A request for something that, if granted, is lawful, is also lawful. A request for something that, if granted, is unlawful, is also unlawful. And self-help to an unlawful end is unlawful. There is no such thing as the lawful enforcement of a private cartel. The Fund’s suits in the courts of Illinois did not ask the court to enforce a law or a regulatory victory. The Fund wanted the court to enforce a private contract. The Fund invoked a rule of decision created by the Association and the Union. It was not a petition for a favorable rule of law; it was not an effort to implement an existing rule of law; it was an unvarnished effort to enforce a private price-fixing agreement. The first amendment does not protect efforts to enforce private cartels, in court or out. Cf. National Society of Professional Engineers v. United States, 435 U.S. 679, 697-99, 98 S.Ct. 1355, 1368-69, 55 L.Ed.2d 637 (1978). If the effort inflicted injury, § 4 of the Clayton Act supplies a damages remedy.
If it inflicted injury. The district court dismissed the case on the ground that the Noerr-Pennington doctrine bars relief, so it has not yet held that Premier suffered injury. The Fund maintains that it offered to stay the state cases pending resolution of the Maryland litigation and that Premier spurned this offer for a time, so perhaps some or all of Premier’s injury is self-inflicted. When the Fund voluntarily dismissed the state cases after losing in Maryland, Premier also neglected to ask the state court to impose fees and costs as a condition of the dismissal. Perhaps this has preclusive effects. These and other questions must be left to the district court on remand. Circuit Rule 18 shall not apply-
Reversed and Remanded
. Three of the defendants in this case are local unions that were not parties to the Maryland litigation. The district court granted summary judgment to one because Premier had not produced any evidence showing that the local was involved in the events of which Premier complained. 627 F.Supp. at 966-67. It denied the motion of the other local unions because the evidence was not dispositive either way. Id. at 967-68. The parties have not discussed in this court the status of these two local unions, and Premier’s appeal was dismissed on October 1, 1986, to the extent it concerned these unions. We therefore do not discuss the claims made by and against these unions.
. The scope of the arrangement is made possible by the labor laws’ tolerance of multi-employer bargaining groups and unions that organize the entire craft or trade. Without the assistance of the Union the Association could not have achieved its objective. Several thoughtful scholars have questioned industry-wide collective bargaining. See T. Campbell, Labor Law and Economics, 38 Stan.L.Rev. 991, 1047-58 (1986); R. Lande & R. Zerbe, Reducing Unions' Monopoly Power: Costs and Benefits, 28 J.L. & Econ. 297 (1985). But Premier does not object to the scope of the Union's organization of the industry; the Union, for its part, does not repeat the contention (which both the Maryland district court, 498 F.Supp. at 539-43, and the Fourth Circuit, 678 F.2d at 502, rejected) that the labor laws shelter any agreement they reached with the Association. We therefore do not pursue this matter.
. We express no view on the validity of an agreement among members of the Association to collect a 1% fee measured by their own payrolls. That agreement affects a smaller portion of the industry’s output and does not prevent any firm from dropping its membership in order to compete vigorously. An intra-Association fee does not involuntarily raise any rival’s cost.
. The due process clauses of the fifth and fourteenth amendments also protect access to the courts. See Bounds v. Smith, 430 U.S. 817, 97 S.Ct. 1491, 52 L.Ed.2d 72 (1977); Lindsey v. Normet, 405 U.S. 56, 92 S.Ct. 862, 31 L.Ed.2d 36 (1972); but see United States v. Kras, 409 U.S. 434, 93 S.Ct. 631, 34 L.Ed.2d 626 (1973); Fuller v. Oregon, 417 U.S. 40, 94 S.Ct. 2116, 40 L.Ed.2d 642 (1974).
. It also creates a new third-party effect by foisting on the loser the winner’s fees, see R. Posner, Economic Analysis of Law § 21.9 (3d ed. 1986), but we disregard this complication.
. Seafy itself disallowed a multiplier based on risk, but on the rationale that compensation for risk came through the trebling of damages. Id. at 661-62. Here the costs of litigation are the-damages sought to be trebled.
. See 461 U.S. at 745-47, 103 S.Ct. at 2171-72, concluding that the Board may not order the cessation of a lawsuit unless the suit is fatuous, but that if a colorable suit is pressed and the plaintiff loses, the Board "would be warranted in taking that fact into account in determining whether the suit had been filed in retaliation for the exercise of the employees' § 7 rights. If a violation is found, the Board may order the employer to reimburse the employees whom he had wrongfully sued for their attorney’s fees and other expenses.” Id. at 747, 103 S.Ct. at 2172.
. The Court’s analysis has been confirmed by data. J. Barron & J. Umbeck, The Effects of Different Contractual Arrangements: The Case of Retail Gasoline Markets, 27 J.L. & Econ. 313 (1984).