United States: Class Actions

In summary

Even though there was ‘little dispute’ that tuna producers fixed prices in Olean Wholesale Grocery Coop, Inc v Bumble Bee Foods LLC, a panel of Ninth Circuit judges let them off the hook. Private antitrust enforcement depends on joining claims in a single class action, so the panel’s decision to reverse the district court’s order granting class certification effectively dismissed the case. But the full Ninth Circuit breathed new life into the case when it voted to rehear the case en banc. This article explains the panel’s decision and previews the issues the Ninth Circuit will address en banc.

Discussion points

• In Olean Wholesale Grocery Coop, Inc v Bumble Bee Foods LLC, a three-judge panel of Ninth Circuit judges erected new barriers to class certification.
• Those barriers could make antitrust enforcement less likely, even in cases such as Olean Wholesale where there is ‘little dispute’ that the defendants conspired to fix prices.
• A majority of judges on the Ninth Circuit voted to rehear the case, giving the victims of a price-fixing conspiracy another shot at holding the defendants accountable.

Referenced in this article

• Olean Wholesale Grocery Coop, Inc v Bumble Bee Foods LLC, 993 F.3d 774 (9th Cir. 2021), reh'g en banc granted, 5 F.4th 950 (9th Cir. 2021).
• Tyson Foods, Inc v Bouaphakeo, 577 US 442 (2016)
• Anderson v Mt Clemens Pottery Co, 328 US 680 (1946)
• Bigelow v RKO Radio Pictures, 327 US 251 (1946)
• In re Asacol Antitrust Litig, 907 F.3d 42 (1st Cir. 2018)

‘There is little dispute over the existence of [the] price-fixing scheme.’^[1] That is how a divided panel of Ninth Circuit judges described the conspiracy among the three largest tuna producers, Tri-Union Seafoods (doing business as Chicken of the Sea), Bumble Bee Foods, and Starkist, to fix packaged seafood prices in Olean Wholesale Grocery Coop, Inc v Bumble Bee Foods LLC.

After all, who could seriously deny a conspiracy that had produced guilty pleas, convictions, lengthy prison sentences and criminal fines? Starkist and its senior vice president pled guilty to ‘a conspiracy among major packaged-seafood-producing firms, the primary purpose of which was to fix, raise, and maintain the prices of packaged seafood in the United States’.^[2] Bumble Bee Foods also pled guilty to that conspiracy,^[3] while a jury convicted the company’s CEO for conspiring to fix canned tuna prices, and the court sentenced him to 40 months in prison.^[4] Chicken of the Sea admitted fixing prices and agreed to cooperate with federal investigators.^[5] All told, the conspirators received over US$100 million in criminal fines.^[6]

One might think that holding them accountable in a civil antitrust action would be rather straightforward. And yet, in the same opinion that acknowledged the tuna producers’ price-fixing conspiracy, the panel let them off the hook for private antitrust liability.

Private litigants play a key role in enforcing antitrust laws.^[7] The government cannot prosecute every violation of federal antitrust law. Nor does the government see its role compensative of the victims of antitrust violations. Private enforcement fills these gaps.^[8]

The class-action mechanism makes that enforcement possible.^[9] Antitrust cases are expensive to litigate, and an individual plaintiff’s recovery is usually too small go it alone. So, victims of an anticompetitive conspiracy join together to resolve their claims in a single class action. If a court denies them the ability to do so, however, they’re effectively out of luck, regardless of the defendants’ liability.

That’s what happened to the victims of the tuna producers’ price-fixing conspiracy did in Olean Wholesale. A Ninth Circuit panel vacated the district court’s decision to grant class status and erected new barriers that will make it more difficult to hold antitrust violators accountable in future cases.

But the full Ninth Circuit may knock those barriers down. A majority of Ninth Circuit judges voted to vacate that decision and rehear the case en banc.^[10] This article explains the panel’s decision, the issues that the judges will want to address en banc, and the impact the case could have going forward.

The Olean Wholesale Grocery Coop, Inc v Bumble Bee Foods LLC litigation

The plaintiffs in Olean Wholesale Grocery Coop, Inc v Bumble Bee Foods LLC are grocery stores, commercial food preparers, and individual consumers. They filed a class action lawsuit against the three largest domestic packaged tuna producers, Tri Union Seafoods (doing business as Chicken of the Sea), Bumble Bee Foods and Starkist. The plaintiffs alleged that the tuna producers violated state and federal antitrust law by conspiring to fix packaged seafood prices.^[11]

The plaintiffs primarily relied on two statistical models to show the class-wide impact of defendants’ price-fixing conspiracy.^[12] The first ‘calculated what the price for wholesale tuna would have been “but for” the alleged price fixing’.^[13] This model compared prices during the price-fixing conspiracy to prices either before or after the impacted period, a clean benchmark where there was no anticompetitive activity.^[14] That comparison revealed that each class member suffered an average overcharge of 10.28 per cent.^[15] Second, the plaintiffs ran a regression against that benchmark to control for ‘explanatory variables’ like ‘product characteristics, supply and demand factors, and the period of alleged conspiracy’.^[16] The regression model concluded that ‘1,111 out of 1,176 direct purchasers (or 94.5%) were injured by Defendants’ actions.’^[17]

The defendants took issue with the regression model. They argued that the model should have considered certain differences between individual class members. And if it had, the model would have showed that only ‘72% paid an inflated price, meaning 28% of the class members suffered no injury at all’.^[18]

The district court left it to the jury to determine whose ‘theory [was] right or wrong’.^[19] At class certification, it was ‘only concerned with whether the method itself is capable of showing [impact] to all, or nearly all of the Class members – not that it does in fact show that the injury occurred’.^[20] Once the plaintiffs met that burden, the court certified the plaintiff classes.^[21] The defendants appealed that decision to the Ninth Circuit.

The panel’s decision on appeal

A three-judge panel vacated the district court’s order granting class certification.^[22] The panel made three rulings on appeal. First, it adopted a preponderance of the evidence standard at class certification. Going forward, antitrust plaintiffs would have to show they met each element of Rule 23, including predominance, by a preponderance of the evidence. Second, it declined to prohibit plaintiffs from using statistical or representative evidence to show predominance. Third, the court held that predominance is defeated when a proposed class contains more than a de minimus number of uninjured class members, so district courts must resolve competing claims as to the number of uninjured class members before it certifies a class.

A preponderance-of-the evidence standard during class certification

The court held that plaintiffs must show that their proposed class meets Rule 23(b)’s predominance requirement by a preponderance of the evidence.^[23] The court reasoned that this evidentiary burden best accords with the text of the rule and Supreme Court precedent. The court noted that rule provides that ‘courts can certify a class “only if . . . the court finds that the questions of law or fact common to class members predominate” over individual ones’.^[24] The court further noted that the Supreme Court instructed courts to certify a class only after conducting a ‘rigorous analysis’ to ensure that the plaintiff has met Rule 23.^[25] This means that courts must do more than find that the plaintiffs have provided admissible class-wide evidence that could establish predominance; courts must also ‘“judg[e] the persuasiveness of the evidence presented” for and against certification’.^[26]

Rejecting a categorical rule against statistical evidence during class certification

The panel rejected the invitation from defendants and their amici to issue a broad rule prohibiting plaintiffs from using statistical evidence to show predominance.^[27] Instead, the it chose a flexible standard: ‘whether a representative sample can “establish classwide liability” at the certification stage “will depend on the purpose for which the sample is being introduced and on the underlying causes of action”.’^[28]

Having found statistical evidence permissible in the abstract, the panel turned to whether the plaintiffs could use their statistical evidence to prove antitrust impact on a class-wide basis. To do so, plaintiffs must show that class members could use the evidence to prove liability in an individual suit, a nexus between their evidence and their theory of liability, and that their evidence is reliable and does not mask individualised differences.^[29]

The plaintiffs met the first two requirements easily. Individual antitrust plaintiffs establish the impact of the defendants’ price-fixing conspiracy by ‘comparing the actual world with a “hypothetical” world that would have existed “but for” the defendant's unlawful activities’.^[30] Like plaintiffs in other price-fixing cases, the plaintiffs’ expert constructed a clean ‘benchmark’ period with representative evidence, and ‘compar[ed] it to [the] market price before and after the benchmark . . . to isolate the “but-for” effect of the price-fixing conspiracy’.^[31] Plaintiffs also ran a regression analysis to calculate their injury that was ‘consistent with . . . regression models [that] prove price-fixing impact in other cases’.^[32] And since the plaintiffs had asserted that defendants had fixed prices, they had sufficiently linked their statistical models with their theory of liability.^[33]

As to the third factor, the court found that the plaintiffs’ statistical evidence did not defeat predominance.^[34] The court saw ‘no issue’ with the plaintiffs’ regression models because they merely ‘averaged the overcharge calculation using Defendants’ own data, and then used that average in a regression model to calculate what percentage of the class was impacted’.^[35] These models did not mask individualised differences in ‘purchasing power [or] retail price strategy’ because the defendants’ ‘scheme could have raised the baseline price at the start of negotiations and could have affected the range of prices that resulted from negotiation’.^[36] And, even if class members suffered individualised damages, ‘the presence of individualized damages cannot, by itself, defeat class certification under Rule 23(b)(3).’^[37]

District courts must determine the number of uninjured class members to determine whether predominance has been met

The panel’s final holding was its most significant. It held that district courts must resolve the parties’ competing claims as to the number of uninjured plaintiffs before certifying a class.^[38] It reiterated that district courts must resolve factual disputes necessary to decide whether common issues predominate over individual ones,^[39] and it identified the number of uninjured class members as ‘an essential component of predominance’.^[40] The panel reasoned that ‘[i]f a substantial number of class members “in fact suffered no injury,” the “need to identify those individuals will predominate”’.^[41]

Turning to the parties’ dispute, the panel noted that the plaintiffs’ statistical model showed an injury for 94.5 per cent of the class.^[42] The defendants, however, argued that the plaintiffs could only show that the conspiracy injured 72 per cent of class members.^[43] The panel held that the district should abused its discretion by not resolving that conflict.^[44] The court reasoned that if the defendants were right that 28 per cent of the class was uninjured, then the impact of defendants’ conspiracy would not be common to the class, which would raise concerns that that the plaintiffs’ models ‘mask[ed] individual differences among the class members, such as bargaining power, negotiation positions, and marketing strategies’.^[45]

The panel declined to set the percentage of uninjured class members large enough to defeat predominance.^[46] The court noted that classes ‘inevitably’ have uninjured class members, but their number must be ‘de minimus’.^[47] A statistical model that could not show antitrust impact for a quarter of the class could not prove predominance ‘under any rubric’.^[48]

Issues that the full Ninth Circuit must resolve en banc

A majority of Ninth Circuit judges voted to vacate the panel’s decision and will rehear the case en banc. On rehearing, the Ninth Circuit will have to address two errors that the panel made that undermine each of its holdings.

Tyson Foods applies to antitrust cases

The panel’s first error concerns its new preponderance-of-the-evidence standard. That holding rests entirely on a single footnote that distinguishes the Supreme Court’s opinion in Tyson Foods v Bouaphakeo.^[49] In that case, the Court held that ‘[o]nce a district court finds evidence to be admissible, its persuasiveness is . . . the near-exclusive province of the jury’.^[50] If Tyson Foods applied to antitrust claims, district courts would not have the discretion to weigh the evidence ‘for and against certification’ to determine which party a preponderance of the evidence favoured.^[51] Instead, they could deny class certification only if the evidence was so unpersuasive that ‘no reasonable juror’ could believe it persuasive.^[52] Before it could announce a new evidentiary standard for proving predominance, the panel had to first ‘cabin’ the case ‘to wage-and-hour suits’.^[53]

The panel distinguished the wage-and-hour class action in Tyson Foods from antitrust class actions. It explained that courts apply a relaxed evidentiary standard in wage cases. ‘[B]ecause defendants often fail to keep proper records of hours worked by employees,’ courts allow the factfinder to ‘infer harm in individual wage-and-hour suits’ with representative evidence.^[54] This inference makes representative evidence ‘presumptively usable at the class certification stage’, but it was ‘cabined to wage-and-hour suits and doesn’t apply [to antitrust cases]’.^[55] Because that inference applied only to wage-and-hour cases, so did the Supreme Court’s instructions that ‘once a district court finds representative evidence “admissible, its persuasiveness is . . . a matter for the jury,” and class certification should only be denied if “no reasonable juror” could have found the plaintiffs’ representative evidence persuasive.’^[56]

But the Supreme Court did not limit the use of representative evidence in Tyson Foods to wage cases. The Court affirmed its use in cases where the defendants caused the uncertainty in the plaintiffs’ damages, which is invariably true in antitrust actions.

The plaintiffs in Tyson Foods sued their employer under the Fair Labor Standards Act for unpaid overtime wages for the time they spent putting on and taking off their protective gear.^[57] Their employer failed to keep records of that time, despite a statutory obligation to do so.^[58] And so, the plaintiffs had to ‘fill an evidentiary gap created by the employer’s failure to keep adequate records’.^[59] Rather than punish the plaintiffs for their employer’s poor recordkeeping, the Court allowed the plaintiffs to introduce a representative sample^[60] to prove the number of hours they worked ‘as a matter of just and reasonable inference’.^[61]

The Supreme Court relied on an earlier case, Anderson v Mt Clemens Pottery Co, in applying the ‘just and reasonable inference’ standard.^[62] The plaintiffs in Mt Clemens worked in a pottery plant, and they brought a class action against their employer under the FLSA for unpaid overtime.^[63] The plaintiffs alleged that their employer did not credit plaintiffs’ wages for the time spent clocking in and walking to and from their workstations.^[64] The employer kept no records of this time, in violation of its statutory duty.^[65] After the district court found for the plaintiffs, the appellate court dismissed the case because it was ‘insufficient for them merely to offer an estimated average of overtime worked’.^[66] The Supreme Court reversed.^[67]

The Court began by noting that employees can ‘easily discharge [their] burden’ to prove the amount of hours they worked when their employers keep ‘proper and accurate’ records.^[68] But when employers fail to do so, a ‘more difficult problem arises’, because the employees cannot ‘prove the precise extent of uncompensated work’.^[69]

The solution to this problem, however, was not ‘penaliz[ing] the employee[s] by denying [them] any recovery’.^[70] Doing so would ‘place a premium on an employer’s failure to keep proper records’ and ‘allow the employer to keep the benefits of an employee’s labors without paying due compensation’.^[71] Such a result would ‘perver[t] of fundamental principles of justice . . . and . . . relieve the wrongdoer from making any amend for his acts’.^[72]

Instead, the Court held that employees could carry their burden if they showed ‘the amount and extent of that work as a matter of just and reasonable inference . . . even though the result be only approximate’.^[73] The Court reasoned that ‘[t]he employer cannot . . . complain that the damages lack the exactness and precision of measurement that would be possible had he kept records in accordance with . . . the [law].’^[74] Because ‘[t]he uncertainty lies only in the amount of damages’ arising from the employers’ statutory violation, its employees need only show ‘a reasonable inference as to the extent of the damages’.^[75]

The Court did not fabricate this standard out of whole cloth for Mt Clemens. The Court borrowed it from three antitrust cases, each of which applied the ‘just and reasonable inference’ standard.^[76] None of those cases addressed the FLSA or an employer’s statutory recordkeeping duty. They each turned on how a defendant’s conduct inherently makes a plaintiff’s damages uncertain in antitrust cases.

The Supreme Court articulated this reasoning clearest in a case it had decided four months Mt Clemens called Bigelow v RKO Radio Pictures. The plaintiffs in Bigelow owned a movie theater in Chicago.^[77] The defendants distributed movies and, like the plaintiffs, also owned their own movie theaters in Chicago.^[78] The plaintiffs alleged that the defendants entered into an anticompetitive agreement that allowed them to show the movies exclusively in their own theatres, which prevented plaintiffs from securing early, desirable movie runs at their theater.^[79] By comparing their receipts before and after the conspiracy, they calculated that defendants’ conduct caused them US$120,000 in damages.^[80]

The jury returned a verdict of US$120,000 for the plaintiffs.^[81] The circuit court sustained the jury’s finding of an unlawful conspiracy, but it reversed the damages award because the comparison could not prove what the plaintiffs’ earnings would have been in the absence of the conspiracy.^[82]

Relying on the two antitrust cases it would later cite in Mt Clemens – Story Parchment and Eastman Kodak – the Supreme Court reversed the appellate court and affirmed the jury’s verdict.^[83] The Court noted that the ‘defendant by his own wrong has prevented a more precise computation’ of plaintiffs’ damages.^[84] In such cases, juries ‘may make a just and reasonable estimate of the damage based on relevant data, and render its verdict accordingly’.^[85] The court reasoned that ‘the most elementary conceptions of justice and public policy require that the wrongdoer shall bear the risk of the uncertainty which his own wrong has created^[86] . . . Any other rule . . . would be an inducement to make wrongdoing so effective and complete in every case as to preclude any recovery, by rendering the measure of damages uncertain.’^[87] Foreshadowing that it would soon extend this rule to FLSA cases in Mt Clemens, the Court explained that this ‘ancient’ principle ‘is not restricted to proof of damage in antitrust suits, although their character is such as frequently to call for its application’.^[88]

The Ninth Circuit panel in Olean Wholesale had it backwards. Tyson Foods did not prohibit representative evidence outside of the wage-and-hour context.^[89] It affirmed the use of representative evidence to prove the plaintiff’s damages as a matter of just and reasonable inference when the defendants created the uncertainty as to the plaintiff’s damages. And though this rule is not ‘is not restricted to proof of damage in antitrust suits . . . their character is such as frequently to call for its application’.^[90]

Individual class members do not cause individual issues to predominate if they are identifiable

The panel also erred in holding that district courts must ‘resolve the factual disputes as to how many uninjured class members are included in Plaintiffs’ proposed class’ at class certification.^[91] Courts may not ‘engage in free-ranging merits inquiries at the certification stage’.^[92] They may consider merits questions ‘to the extent – but only to the extent – that they are relevant to determining whether Rule 23 prerequisites for class certification are satisfied’.^[93] The number of uninjured class members is not, as the court in Olean Wholesale found, ‘an essential component of predominance’^[94] when those class members are identifiable.

Under Rule 23(b)(3), plaintiffs must show ‘that the questions of law or fact common to class members predominate over any questions affecting only individual members’.^[95] Common questions are those that have the ‘capacity . . . to generate common answers apt to drive the resolution of the litigation’.^[96] At bottom, the predominance inquiry asks ‘what a class trial would look like’.^[97]

Courts make this determination by considering whether common questions of law or fact predominate regarding the key elements of the claims alleged.^[98] In Olean Wholesale, those elements are: ‘(1) the existence of an antitrust conspiracy; (2) the existence of individual injury, also referred to as “antitrust impact,” as a result of the conspiracy; and (3) resultant damages.’^[99] The panel’s decision to vacate class certification focuses on the second element: antitrust impact. To demonstrate that common issues predominate over individual ones with respect to antitrust impact, plaintiffs must show that ‘the existence of individual injury resulting from the alleged antitrust violation . . . [is] capable of proof at trial through evidence that [is] common to the class rather than individual to its members’.^[100]

The panel never considered ‘what a class trial would look like’^[101] if the class contained uninjured members in Olean Wholesale; the court simply assumes that ‘[i]f a substantial number of class members “in fact suffered no injury,” the “need to identify those individuals will predominate”’.^[102] The court further assumes that the court assumes that more uninjured class members means individual issues are more likely to predominate.^[103]

But the number of uninjured class members is not always proportional to the importance of individual issues or whether they would predominate at trial. The key issue is whether ‘questions common to the class predominate, not that those questions will be answered, on the merits, in favor of the class’.^[104] As a result, ‘non-injury to a subset of class members does not necessarily defeat certification of the entire class, particularly [when] the district court is well situated to winnow out those non-injured members at the damages phase of the litigation, or to refine the class definition’.^[105]

Before launching a fact-finding mission at the class-certification stage to determine the exact number of uninjured class members, courts should first ask: does it matter? Does the presence of uninjured class members actually reflect a lack of predominance as to an element of the plaintiffs’ claim, and if so, does an increase in the number of uninjured plaintiffs result in a corresponding increase in the likelihood that individual issues will predominate over common ones? If the answer to these questions is ‘no’, then they are irrelevant to showing predominance, and district courts cannot resolve factual disputes as to the number of uninjured class members at class certification.^[106]

Consider the following hypothetical: the plaintiffs in Olean Wholesale sued defendants for fixing packaged seafood prices.^[107] They introduced defendants’ guilty pleas at class certification as common evidence proving the existence of the conspiracy,^[108] and the district court certified a class of direct purchasers for a five-year period,^[109] in part, based on that evidence.^[110] In their petition to the Ninth Circuit to vacate that decision, however, the defendants pointed out that the ‘guilty pleas covered a much narrower time period than the [conspiracy] alleged in plaintiffs’ complaints and found in the class definitions’.^[111] If a jury determines the conspiracy covers the period that defendants pled guilty to – say, a four-year period – rather than the one in the current class definition, would individual issues suddenly overwhelm the case because 20 per cent of the class was uninjured? No. The district court or claims administrator can easily ‘winnow out those non-injured members’ as a matter of accounting.^[112]

The same is true as to the of uninjured class members in Olean Wholesale. The parties disagreed as to the proportion of class members subject to the overcharge.^[113] The plaintiffs argue that almost every class member suffered an injury; the defendants say that many have not. But even if defendants were correct, neither party disputes the statistical models could identify the class members that did not pay the overcharge. Because the model returns a result as to whether each class member has paid an overcharge, the percentage of uninjured class members makes no difference as to the district court’s or the jury’s ability to ‘winnow out’^[114] those uninjured members.

Antitrust enforcement going forward

The panel’s decision follows a worrying trend for antitrust enforcement. It followed the First Circuit and DC Circuit in holding that district courts cannot certify a class if it has more than a de minimi number of antitrust plaintiffs.^[115] Given the class mechanism’s centrality to antitrust enforcement, the upshot is that powerful corporations have free reign to violate antitrust laws, so long as they harm ‘only’ 72 per cent, 87.3 per cent or 90 per cent of their purchasers.^[116]

But the case also radically departs from its sister circuits. A key factor leading to the First and DC Circuit’s decisions was the absence of a manageable process to identify uninjured class members.^[117] In Asacol for example, the plaintiffs alleged that a conspiracy among prescription two name-brand drug manufacturers precluded others from introducing a generic version.^[118] The evidence at class certification showed that 10 per cent of class members were uninjured because they would not have switched to the generic drug, even if it had been available.^[119] But there was no way to identify those class members that fell into that 10 per cent, such that they ‘might be picked off in a manageable, individualized process at or before trial’.^[120] Without a mechanism to separate injured and uninjured class members, ‘[t]he need to identify those individuals will predominate and [will] render an adjudication unmanageable’.^[121]

The panel’s blanket prohibition against certifying classes with more than a de minimis number uninjured plaintiffs has nothing to do with whether individual issues predominate. It makes no difference whether uninjured class members ‘might be picked off in a manageable, individualized process at or before trial’.^[122] Nor does it matter whether their presence renders class-wide adjudication unmanageable. Regardless of the significance of individual issues, the panel prohibited such cases from going to the jury altogether:

Suppose the jury ultimately decides Defendants’ expert is right and Plaintiffs’ model sweeps in 28% uninjured class members. Too late: the damage has been done. By then, Defendants would have possibly weathered years of litigation at untold costs, only to discover that the case never should have reached the merits at all. ^[123]

It is anyone’s guess why the Ninth Circuit voted to rehear the case. It may have found the panel made the same errors identified in this article. Perhaps, it believes that the panel should have ratcheted up the plaintiffs’ burden to certify a class even further. But one thing is certain: a decision to reverse the district court’s order granting class certification could save the tuna producers from ‘weather[ing] years of litigation at untold costs’.^[124]

But those savings would not exist because the case ‘never should have reached the merits’.^[125] Indeed, the Ninth Circuit panel saw ‘little dispute over the existence of [the] price-fixing scheme’.^[126] The grand jury in the Northern District of California indicted multiple defendants^[127] and their employees,^[128] and those indictments have resulted in guilty pleas,^[129] convictions^[130] and more than US$100 million in criminal fines for the same conduct that grocery stores, commercial food preparers and individual consumers alleged.^[131]

Rather, the tuna producers would not have to ‘weather years of litigation at untold costs’^[132] because their victims of their conspiracy cannot afford to bring their claims on their own. Should the Ninth Circuit force the class members to pursue their claims individually, the defendants’ actual liability will stay the same, as the 72 to 94 per cent of class members who were shown to have paid defendants’ overcharge could introduce the statistical models in their individual actions. Defendants’ effective liability, however, would fall precipitously because individual plaintiffs would likely drop their lawsuits as it would no longer be economically feasible for them to pursue their cases. Far from protecting the tuna producers from claims that ‘never should have reached the merits at all’,^[133] such a decision would let criminal price-fixers off the hook for civil antitrust liability.

Notes