POSNER, Circuit Judge.

This is a class action suit under section 1 of the Sherman Act, 15 U.S.C. § 1, that after certification of the class was dismissed on the merits when shortly before the trial was scheduled to begin the district judge ruled that the case could not go to trial on a theory of per se liability. The plaintiffs could have gone to trial on a theory of liability under the rule of reason, but preferred to appeal the dismissal, hoping we would order the reinstatement of their per se case. The dismissal is final because the plaintiffs have made clear that the case is over if they are not allowed to try it as a per se case.

The class consists of chemical companies that purchase sulfuric acid as one of the inputs into their production of chemicals. The defendants own smelters that process nonferrous minerals such as nickel and copper. They also produce sulfuric acid and sell or sold it to the members of the class.

The defendants cross-appeal, asking that if (but only if) we reverse the dismissal of the suit, we decertify the class. The purpose of the "only if" qualification is to make the judgment bind the entire class if we affirm the dismissal, which it would not do if the class were decertified. See Smith v. Bayer Corp., ___ U.S. ___, 131 S.Ct. 2368, 2380, 180 L.Ed.2d 341 (2011). If we reverse, and allow the class members to press their theory of per se liability, the defendants would prefer to fight the class members one by one, which would be the result of decertification of the class, rather than have to face all of them in a single trial that could produce a monstrous judgment. It is such threats of ruin that force most defendants in class action suits to settle if a class is certified. In re Rhone-Poulenc Rorer, Inc., 51 F.3d 1293, 1299-1300 (7th Cir.1995).

The abiding puzzle of the plaintiffs' appeal is why the lawyers for the class, having spent almost nine years litigating the case in the district court, refused to go to trial. Though the trial would have been governed by the rule of reason, probably all that this would have meant in a case such as this is that the defendants would have had greater latitude for offering justifications for what the plaintiffs claim is a price-fixing conspiracy than if the standard governing the trial had been the per se rule, which treats price fixing by competitors as illegal regardless of consequences or possible justifications. Texaco Inc. v. Dagher, 547 U.S. 1, 5, 126 S.Ct. 1276, 164 L.Ed.2d 1 (2006). The plaintiffs do not concede that the conduct they challenge was reasonable and therefore lawful; but their refusal to go to trial under the rule of reason suggests that they expected a jury to find that it was.

From remarks by their lawyer at the oral argument we infer that they think that in a trial governed by the rule of reason they would have had to prepare a radically different case in chief, proving not only that the defendants fixed prices (all they'd have to prove, besides damages, in a per se case), but also that the defendants had market power (that is, the power to raise price above the competitive level without losing so much business to other sellers that the price would quickly fall back to that level) and that their collusive activity was indeed anticompetitive. Doubtless in most cases the prima facie case under the rule of reason requires proof "that the defendant has sufficient market power to restrain competition substantially," as we said in General Leaseways, Inc. v. National Truck Leasing Association, 744 F.2d 588, 596 (7th Cir.1984). But a plaintiff who proves that the defendants got together and agreed to raise the price (whether directly or by restricting output, which would have the same effect) that he pays them for their products — which is what the plaintiffs in this case would have had to prove under the per se rule to establish liability and obtain damages — has made a prima facie case that the defendants' behavior was unreasonable. He need not prove market power; even though by definition without it a firm or group of firms can't harm competition, it is not a part of the prima facie case of illegal per se price fixing. E.g., National Collegiate Athletic Association v. Board of Regents, 468 U.S. 85, 109-10, 104 S.Ct. 2948, 82 L.Ed.2d 70 (1984). But even if a challenged practice doesn't quite rise to the level of per se illegality, it may be close enough to shift to the defendant the burden of showing that appearances are deceptive and really the behavior that the plaintiffs have challenged is not anticompetitive. Of course there would be more work for the plaintiffs if the defendants in this case were able to create a triable issue of justification, but, as we have just explained, probably less than they think.

But this is a detail; the question is whether the judge was right to think this a rule of reason case. Before turning to that question, we address briefly the plaintiffs' argument that the district court's ruling on the question was not only substantively unsound but procedurally irregular. The district judge who had handled the lengthy pretrial proceedings in this case had retired and the case had been reassigned. The original judge, in denying summary judgment (except on one issue) for the defendants, had, rather oddly, refused to decide whether the case should proceed as a per se case or a rule of reason case. After the case was reassigned and a trial date set, the defendants became concerned that they didn't know what kind of trial to prepare for, so they asked the judge to decide, and he said rule of reason. His ruling was abrupt and not explained, but whether it was correct is a question of law that we can decide without benefit of an analysis by the district judge. See Deutscher Tennis Bund v. ATP Tour, Inc., 610 F.3d 820, 829 n. 7 (3d Cir.2010).

So let's decide it; concretely, let's decide whether the challenged practices are the sort that fall within the scope of the per se rule against price fixing, or fall outside it in which event the judge was right to rule that a trial would be governed by the rule of reason.

The principal defendants are Noranda, Inc. and Falconbridge Ltd., Canadian mining companies that in 2005, after the period of the alleged antitrust violations (1988-2002), merged to form a single company named Xstrata Canada. During the relevant period Noranda owned between 46 and 60 percent of Falconbridge's common stock, and as a result controlled Falconbridge. They thus were affiliated rather than independent producers, and in fact pooled and jointly sold their sulfuric acid.

The smelting of nonferrous minerals generates sulfur dioxide as a byproduct, and sulfur dioxide reacts with the water vapor in the atmosphere to create sulfuric acid, which is the acid in acid rain. For environmental reasons the Canadian government requires the mining companies to process the sulfur dioxide into sulfuric acid, which unlike sulfur dioxide does not enter the atmosphere and so does not contribute to the formation of acid rain. Although there is a market for sulfuric acid — it is used in manufacturing fertilizer, paper, and other products — Noranda and Falconbridge didn't want to produce the acid because the Canadian market for it is limited and what is not sold is costly to store or — because of further restrictions imposed by the Canadian government to protect the environment — to dispose of other than by sale. When in the mid-1980s the government increased the amount of sulfur dioxide that smelters were required to capture, Noranda had to build a large new sulfuric acid plant at one of its smelter sites in order to be in compliance with the new requirement.

Thus at the same time that Noranda was involuntarily contributing to the solution of the acid rain problem, it was compounding its own problem (its "personal" problem as it were) of excess production of sulfuric acid — excess because as we said it is a product costly to store or dispose of and difficult to find a market for in Canada. And so in the 1980s Noranda and Falconbridge began looking at the large U.S. market for sulfuric acid. Having virtually no capability for distributing their acid in the United States — no distributors, no customer relationships — they had to create a U.S. distribution network if they wanted to sell sulfuric acid in this country. The logical candidates for such a network were the U.S. producers of sulfuric acid. Although sulfuric acid was an unwanted byproduct of the smelting operations of Canadian mining companies, chemical companies in the United States manufactured it from sulfur and sold it to firms that used it in their own manufacturing. The domestic U.S. production of sulfuric acid (called "virgin acid") was not very profitable, however, so the Canadian companies saw an opportunity to persuade the producers to distribute the Canadian companies' sulfuric acid ("smelter acid") in lieu of producing their own. The U.S. companies were also the distributors of the sulfuric acid they produced and so had the customer relationships necessary for distribution of the Canadian companies' acid in the United States.

The effort at persuasion was successful. The U.S. producers were willing to curtail production and devote their distribution facilities to the Noranda-Falconbridge acid and be compensated by the difference between what the Canadian companies would charge them for sulfuric acid and what they could resell it for to the U.S. consumers of the acid. The U.S. producers were afraid that unless they agreed to become distributors for the Canadian companies the latter would create their own U.S. distribution network and underprice the U.S. producers, thereby driving them out of the sulfuric acid market rather than keeping them in as distributors of the Canadian acid. As one producer put it, "they [Noranda and Falconbridge] are leaving a path of destruction in their wake. They have not picked up any business without decreasing the pricing at least $10-$15 per ton to the customer, and the threat of their participation [in the U.S. market] is causing [other U.S. producers] to significantly reduce pricing in an effort to maintain" sales.

Apparently Noranda and Falconbridge were not content to wait for the economics of the sulfuric acid market to convince the U.S. producers to stop producing. For the two Canadian companies entered into contracts with several U.S. producers that provided that in exchange for becoming a distributor of the Canadian companies' sulfuric acid (and with an exclusive territory in which to distribute it), each producer would curtail its own production and be compensated for this by the Canadian companies' selling sulfuric acid to it (for resale) cheaply enough to make distribution more profitable than production. These "shutdown agreements" are the main focus of this case. The plaintiffs argue that by reducing total sales of acid in the United States the agreements raised the market price, and that an agreement to restrict output and therefore raise price is the per se illegal offense of price fixing.

This is a possible interpretation of the shutdown agreements, and if it were the only plausible one this would indeed be a per se price-fixing case. But it is not the only plausible interpretation. If you are Noranda and Falconbridge, gazing into the American market for sulfuric acid, you see opportunity but also risk. The opportunity is to make money from your unwanted byproduct of mining. For this you need distribution. The U.S. producers of the acid are the firms that can undertake distribution in the United States. But as they are also producers, you have to worry about the effect of their production on the profitability of your venturing into the U.S. market.

We do not know much about the cost structure of the Canadian companies' acid; the plaintiffs haven't told us what we would need to know in order to be persuaded by them that the shutdown agreements are garden-variety price-fixing agreements. What we do know is that the Canadian companies incur costs both in converting sulfur dioxide into sulfuric acid and in transporting it to market, that the acid is costly to transport because it is extremely corrosive and special equipment and training are therefore required for its safe transportation, and that most buyers of sulfuric acid in the United States are located far from the Canadian smelters. Suppose the U.S. distributors of sulfuric acid, being themselves producers, decided to continue producing. Supply, being augmented by the shipments of the Canadian companies into the United States, would now greatly exceed demand, and prices might plummet to a level at which it was no longer profitable for the Canadian companies to incur the costs of trying to sell their sulfuric acid in the United States.

They might be willing to sell their acid at a loss, because they might lose even more money if as a result of ceasing to export the acid they had to close down some of their smelters or build new sulfuric acid plants in order to comply with Canadian environmental regulations. But if therefore they sold their acid in the United States at a loss, the U.S. producers might bring antidumping proceedings against them, arguing that the Canadian companies were selling below their cost in Canada. See generally Harvey Kaye & Christopher A. Dunn, International Trade Practice §§ 15:1, 19:2 (2012). We don't know whether the Canadian companies could defend successfully by proving that they would lose even more money by not selling below cost, because of the losses they would incur by closing down some of their smelters or building new sulfuric acid plants, in which event their loss selling in the United States would be in a relative sense profitable.

The Canadian companies might also be troubled by the prospect of distributing their sulfuric acid through companies that are also competitors by virtue of producing their own sulfuric acid. It is not per se illegal to insist that a distributor agree not to carry a competing line of goods to the supplier's, Roland Machinery Co. v. Dresser Industries, Inc., 749 F.2d 380, 394 (7th Cir.1984); 11 Herbert Hovenkamp, Antitrust Law ¶ 1820, pp. 174-79 (3d ed. 2011); what difference should it make that the competing line is produced by the distributor himself? And so the shutdown agreements might be found to be an innocent species of exclusive dealing.

Our analysis suggests that had it been the rule, when the Canadian companies were contemplating entry into the U.S. market, that shutdown agreements (or some equivalent, like requirements contracts) would be per se violations of U.S. antitrust law, the companies might have been deterred from entering — and the price of sulfuric acid in the United States would be higher. Moreover, given the cost advantage of the Canadian companies and the fact that a number of U.S. producers got out of the business of producing sulfuric acid because they knew they couldn't compete with those companies and so couldn't remain in the sulfuric acid business without signing shutdown agreements, the effect of the agreements on price and output may have been merely to accelerate slightly an inevitable trend created by the Canadian companies' entry into the U.S. market.

An alternative to the shutdowns might have been for the Canadian companies to negotiate long-term supply contracts with the U.S. firms, but it is not suggested by the plaintiffs and it is not clear that the effects would be different from the effects of the shutdown agreements. For if the U.S. firms obtain their supply of sulfuric acid from the Canadian companies, they won't be producing acid themselves.

The shutdown agreements are in effect a form of price fixing, but we know from Broadcast Music, Inc. v. Columbia Broadcasting System, Inc., 441 U.S. 1, 24, 99 S.Ct. 1551, 60 L.Ed.2d 1 (1979), that even price fixing by agreement between competitors — and from Polk Bros., Inc. v. Forest City Enterprises, Inc., 776 F.2d 185, 189 (7th Cir.1985), that other agreements that restrict competition, as well — are governed by the rule of reason, rather than being per se illegal, if the challenged practice when adopted could reasonably have been believed to promote "enterprise and productivity." Id. The entry of Noranda and Falconbridge into the U.S. sulfuric acid market was likely to result in an eventual fall in the price of acid in that market, an unequivocally socially beneficial effect from an economic standpoint. If the agreements facilitated that entry, their net effect on economic welfare may well have been positive, especially since the negative effects may have been few because of the higher production costs of the U.S. companies.

The plaintiffs point out that both the BMI and Polk opinions describe competitive restrictions that are defensible as "ancillary" to (that is, supportive of) socially beneficial business endeavors that create a new "product." The Court in BMI upheld what amounted to a price-fixing arrangement among composers on the ground that it created a new "product," namely a blanket copyright license that greatly reduced the cost of license negotiations. Were it not for blanket licensing, every composer of music would have to negotiate a copyright license separately with hundreds or even thousands of radio stations, nightclubs, and other performance venues, and each radio station, etc., would have to negotiate separately with hundreds of composers. But equally the shutdown agreements could be regarded as enabling or assisting in enabling a new product in the U.S. economy, namely Canadian smelter acid. Anyway "product" talk is an unnecessary and distracting embellishment of the rule of reason. The blanket licenses in BMI were not a product, new or old, but a contractual instrument for marketing music, which was the product. The rule of reason directs an assessment of the total economic effects of a restrictive practice that is plausibly argued to increase competition or other economic values on balance.

Pretrial discovery had supplied the plaintiffs' in the present case with all the evidence they needed in order to be able to make a prima facie case of price fixing. So at least they believed and we can assume without deciding that they were right. But, to repeat an earlier point, if they were right, then all that the rule of reason would have done to alter the litigation would have been to allow the defendants to defend at trial by showing that the shutdown agreements, even if they were a form of price fixing or output restriction, were on balance procompetitive because they facilitated the entry of very low-cost producers into the U.S. market. That benefited U.S. chemical companies that use sulfuric acid as an input (and are paradoxically the plaintiffs in this class action suit — biting the hand that fed them), and ultimately the consumers of the products that those companies make.

It is relevant that we have never seen or heard of an antitrust case quite like this, combining such elements as involuntary production and potential antidumping exposure. It is a bad idea to subject a novel way of doing business (or an old way in a new and previously unexamined context, which may be a better description of this case) to per se treatment under antitrust law. Leegin Creative Leather Products, Inc. v. PSKS, Inc., 551 U.S. 877, 886-87, 127 S.Ct. 2705, 168 L.Ed.2d 623 (2007); Broadcast Music, Inc. v. Columbia Broadcasting System, Inc., supra, 441 U.S. at 9-10, 99 S.Ct. 1551. The per se rule is designed for cases in which experience has convinced the judiciary that a particular type of business practice has no (or trivial) redeeming benefits ever.

The plaintiffs deny that this is a novel case — they say it's a rerun of United States v. Socony-Vacuum Oil Co., 310 U.S. 150, 60 S.Ct. 811, 84 L.Ed. 1129 (1940), the famous "hot oil" case. There was a chronic oversupply of oil during the Great Depression. This was in part because subsurface geological changes had made it difficult (given the then-existing technology) for producers to reactivate wells once they had been abandoned, and so they were reluctant to take them out of service even when demand for oil was low, and in part because many of the smaller independent refiners, lacking storage space, could not hold any of their output off the market but had to dump it. The resulting oversupply depressed prices throughout the market. The large refiners agreed among themselves to buy a portion of the small refiners' output and hold it off the market in order to raise the price for the large refiners' own oil. In effect the big refiners were paying the small ones not to sell, just as — the plaintiffs argue — in our case Noranda and Falconbridge were paying U.S. producers of sulfuric acid not to produce. The difference is that the only aim and effect of the price-fixing agreement in Socony-Vacuum were to raise price; in this case the aim was to facilitate entry into the U.S. market, which would (and eventually did, as we'll see) lower prices and prevent the shutdown of Canadian smelting operations, which would have reduced output and raised the price of sulfuric acid in the United States. The overall effect was thus to lower rather than to raise price. The plaintiffs' claim that the price would have been even lower without the shutdown agreements is doubtful, as we have said, because without the agreements the Canadian companies might not have entered the U.S. market.

The plaintiffs retreat to the general language in the Socony-Vacuum opinion, an opinion 72 years old and showing its age. They quote from the opinion that "any combination which tampers with price structures is engaged in an unlawful activity." Id. at 221, 60 S.Ct. 811. Taken literally the quotation would outlaw resale price maintenance, which is no longer illegal per se, see Leegin Creative Leather Products, Inc. v. PSKS, Inc., supra, 551 U.S. at 894, 127 S.Ct. 2705, as well as the agreements upheld in the BMI and Polk Bros. cases.

The plaintiffs do not rest their case entirely on the shutdown agreements. They point also to Noranda and Falconbridge's grant of exclusive territories to their U.S. distributors, which by preventing competition among the distributors shored up the shutdown agreements. The argument is that the Canadian companies compensated the distributors for giving up their production by cutting them in on the supracompetitive profits that eliminating competition can, and in this case according to the plaintiffs did, enable; and so the distributors' spread — the difference between what they paid their Canadian suppliers and what they charged their customers — was fattened as a way of sharing out the supracompetitive profits between them and the Canadian companies. Without exclusive territories the distributors would have competed with each other and in doing so might have competed away their share of the supracompetitive profits.

But there is another way to look at the exclusive territories. When the Canadian companies went to American producer-distributors and said we're entering the market with our very cheap sulfuric acid and we want you to convert from being producers of acid and distributors to being just distributors of our acid, they were asking the producer-distributors to take a big risk by changing their business model. One way to compensate them for taking that risk was to insulate them from competition at the distribution level. Exclusive territories reduce competition at the distributor level but can increase it at the producer level and in this case may well have done so by facilitating the Canadian producers' entry into the U.S. market. See Continental T.V., Inc. v. GTE Sylvania Inc., 433 U.S. 36, 54-57, 97 S.Ct. 2549, 53 L.Ed.2d 568 (1977).

In 1998 Noranda and Falconbridge formed a joint venture with DuPont to supply sulfuric acid to the U.S. market, and the plaintiffs argue that they did so solely to eliminate competition with DuPont and so the joint venture was another per se violation. The joint venture pooled the acid output of all three companies, and, more important (remember that Falconbridge was an affiliate of Noranda; there is no indication that the two firms had ever competed with each other in the sulfuric acid market), made DuPont's very extensive U.S. distribution network available to the Canadian companies. A further anticipated economy was that DuPont would sell the Canadian companies' sulfuric acid jointly with its other chemicals, providing one-stop shopping to buyers of a variety of chemicals. Noranda and Falconbridge could not do that on their own because they are mining companies, not chemical companies, except (so far as appears) for their involuntary production of sulfuric acid.

The joint venture enabled substantial economies in transportation and marketing and after it was launched the price of sulfuric acid in the United States dropped significantly. Yet the plaintiffs argue that the joint venture was spurious. If two or more competing firms, wanting to fix prices, agreed to form a joint venture to sell their output at a price agreed on by the parties, the designation of the price-fixing agreement as a joint venture would not save it from being adjudged illegal per se. Texaco Inc. v. Dagher, supra, 547 U.S. at 5-6 and n. 1, 126 S.Ct. 1276; Starr v. Sony BMG Music Entertainment, 592 F.3d 314, 326 (2d Cir.2010); Addamax Corp. v. Open Software Foundation, Inc., 152 F.3d 48, 52 (1st Cir.1998); 13 Hovenkamp, supra, ¶ 2132, pp. 187-200. But as also explained in the cases and treatise that we've just cited, if a joint venture has a legitimate business purpose, as the defendants' joint venture with DuPont did, the fact that as part of the venture the prices of the venturers are coordinated does not condemn it out of hand, but instead subjects it to scrutiny under the rule of reason. If the coordination is ancillary to (that is, supportive of) the legitimate business purpose of the venture, it may be permissible — a rule of reason question.

The plaintiffs argue that this venture was illegitimate because by organizing it as a limited liability company rather than as a conventional stock corporation, and also by leaving almost all the assets it needed with Noranda and Falconbridge, the venturers avoided (or at least claimed to be entitled to avoid) registering the proposed venture with the Federal Trade Commission under the Hart-Scott-Rodino Act. See "Premerger Notification: Reporting and Waiting Period Requirements," 64 Fed.Reg. 34804-02 (June 29, 1999); 15 U.S.C. § 18a(a)(2)(B)(ii); 16 C.F.R. § 801.40. But so what? The plaintiffs think that if the joint venture violated that Act, the defendants must be guilty of per se illegal price fixing. That is a non sequitur. If there were no joint venture, there would still be no per se violation for there would still be the legitimate business reasons for the defendants to have cooperated with DuPont — indeed nothing bearing on the economic consequences of the arrangement would be altered.

In summary, we agree with the district court's order rejecting the plaintiffs' request to limit the trial to their claims of per se violation of antitrust law. But we disagree with an alternative ground of affirmance that the defendants urge — that the four-year antitrust statute of limitations expired before the suit was filed — and we think it important to register our disagreement in order to head off such an argument in future cases. The argument depends on an assumption that the statute of limitations in an antitrust case begins to run as soon as the antitrust injury (in this case, the alleged effect of the shutdown agreements, territorial restrictions, and joint venture in preventing the U.S. price of sulfuric acid from falling as low as it would otherwise have fallen) occurs, rather than when it is discovered (which may be later). That is incorrect. In re Copper Antitrust Litigation, 436 F.3d 782, 789 (7th Cir.2006). Statutes of limitations in federal civil cases, unless otherwise specified by Congress, begin to run upon discovery of the injury from the alleged violation. Id. The defendants are incorrect in suggesting that Klehr v. A.O. Smith Corp., 521 U.S. 179, 184, 117 S.Ct. 1984, 138 L.Ed.2d 373 (1997), or any other case, has changed that rule for antitrust. The fact that Zenith Radio Corp. v. Hazeltine Research, Inc., 401 U.S. 321, 338, 91 S.Ct. 795, 28 L.Ed.2d 77 (1971), mentioned only injury as the statute of limitations trigger in an antitrust case in which there was no issue regarding discovery does not imply the inapplicability of the discovery rule to antitrust cases in which, as in this case, the date of discovery might matter.

The defendants argue that because the antitrust laws specify treble damages for violations, prospective plaintiffs should not be allowed to sit on their hands after sustaining antitrust injury, in order to run up their damages. But they aren't allowed to sit on their hands; the discovery rule requires diligence. Merck & Co. v. Reynolds, 559 U.S. 633, 130 S.Ct. 1784, 1793-94, 176 L.Ed.2d 582 (2010); Cathedral of Joy Baptist Church v. Village of Hazel Crest, 22 F.3d 713, 717 (7th Cir.1994); SEC v. Gabelli, 653 F.3d 49, 59 (2d Cir.2011). And punitive damages, whether in the form of trebling compensatory damages or in other forms, are available for violations of a number of different federal statutes, to all of which, as far as we know, the discovery rule applies. We can't think of any reason to treat antitrust statutes differently.

Nevertheless, for the reasons discussed earlier, the judgment of the district court is

AFFIRMED.