IN RE: BROILER CHICKEN ANTITRUST LITIGATION (2025)

In this successive appeal, one class member again objects to the district court's award of attorney's fees. The court's analysis was thorough, and we commend the detailed steps it took to calculate the award. Its sole slip was to include a handful of skewed fee awards when recreating the ex ante market for legal services. Those awards can be easily removed from the calculation, so we affirm the award as modified.

I

Because this appeal is successive, we need not recount much of the procedural history. See In re Broiler Chicken Antitrust Litig. (Broiler I), 80 F.4th 797 (7th Cir. 2023). As a brief refresher, a class often does not agree before litigation on a fee with its attorneys. When awarding fees after the case concludes, then, district courts must do so according to a “hypothetical ex ante bargain”—giving “counsel the market price for legal services, in light of the risk of nonpayment and the normal rate of compensation in the market” at the litigation's outset. Id. at 801–02 (italicization added) (first quoting Williams v. Rohm & Haas Pension Plan, 658 F.3d 629, 636 (7th Cir. 2011); and then quoting In re Synthroid Mktg. Litig. (Synthroid I), 264 F.3d 712, 718 (7th Cir. 2001)).

We previously agreed with objector John Andren and held that, when reconstructing the ex ante market for attorney's fees, the district court erred in two ways. First, the court improperly discounted auction bids made by class counsel in other litigation because the bids incorporated a “declining fee scale” format. Id. at 803. Regardless of the fee structure, auction bids made at the outset of litigation are typically probative of the rate counsel is willing to accept while the risk of loss still exists—the goal of this circuit's ex ante calculus. Id. at 801–02. We caveated our discussion, though, by saying it was an error only for the district court to “categorically” discount these bids. Id. at 803–04. The court retained discretion to “accord” them “appropriate weight,” as “[o]ther aspects of the cases in which the bids were made may show the bids to be poor indicators of what bargain would have been struck ex ante.” Id.

Second, the district court abused its discretion in excluding fee awards from the Ninth Circuit. Id. at 804. The court had done so because that circuit, unlike ours, imposes a megafund rule, which caps fees once a recovery reaches a certain amount. Synthroid I, 264 F.3d at 717–18. We said that although fees in this circuit are not similarly limited, class counsel, “as rational actors,” “assess the risk of being awarded fees below the market rate of their legal services when they seek to represent plaintiffs in the Ninth Circuit.” Broiler I, 80 F.4th at 804. The district court should have accordingly “assigned appropriate weight” to fees awarded in the Ninth Circuit. Id.

On remand, after thoroughly analyzing the parties' arguments, the district court awarded a new fee. A few aspects of its award are relevant here. The court first disregarded three bids class counsel had made to represent plaintiffs in different antitrust suits in the six years before filing this suit. All those bids entitled class counsel to between 13.5% and 20% of the recovery amount. The district court, though, thought these three bids were poor indicators of the ex ante market, as each case was preceded by a government antitrust investigation. Given that the government already laid the groundwork for later private litigation, class counsel had a lighter workload and higher chance of success compared to suits beginning from scratch. As a result, the court did not consider these bids when recreating the ex ante market.

Class counsel also agreed to represent a pension fund in an antitrust action against large financial institutions, in a case called In re Interest Rate Swaps Antitrust Litigation (IRS). That agreement used a declining fee approach taken from another antitrust case. See In re Payment Card Interchange Fee & Merch. Disc. Antitrust Litig., 991 F. Supp. 2d 437, 445 (E.D.N.Y. 2014). Nobody disputes that if the Payment Card schedule were adopted in this case, class counsel would be entitled to a $47.2 million award, representing 26.6% of the net common fund.

The district court thought IRS and this case were “good comparators.” Yet it did not award 26.6% here due to a material difference: Because the defendants in IRS were large financial institutions with significant assets, the court concluded that the cases' potential settlement values were not comparable. Even though the court did not incorporate the 26.6% award here wholesale, it still gave IRS an outsized weight in its final calculation, as discussed below.

Next up for the district court's consideration were awards from the Ninth Circuit. It said fees capped under that circuit's megafund rule are not evidence of the market rate, else there would be no need for an artificial limit. So, although counsel make the informed decision to continue litigating there, all other circuits maintain a higher “market” rate for class-action awards.

To arrive at the fee award, the district court created a detailed spreadsheet, compiling data from three sources. See In re Broiler Chicken Antitrust Litig., No. 16-C-8637, 2024 WL 3292794, at *6 (N.D. Ill. July 3, 2024). The first source included all court-awarded fees to counsel in cases without an ex ante agreement—in other words, ex post awards—within a set timeframe before this litigation. The second consisted of awards listed in one of class counsels' expert declarations filed by Professor Robert Klonoff. His data was titled “examples of antitrust cases” both within and outside of the Seventh Circuit “with [fee] percentages of 33 percent or greater.” The third was the 26.6% award from the IRS case. Because the court found that case to be a good comparator, it counted the award ten times in its spreadsheet, thereby giving it an outsized weight.¹

In its spreadsheet, the district court narrowed the cases it considered to only those with recovery amounts of $100 million to $1 billion—what it viewed as most representative of the $181 million settlement recovery here. This left 52 awards. The district court further excluded three Ninth Circuit “outliers” of 9% and 11%, which—as discussed above—it did not find to be representative of this circuit's ex ante market rate.² The remaining 49 awards yielded a median of 31% and mean of 29%. Splitting the difference, the court awarded class counsel 30% of the net common fund. This resulted in a fee of $51.66 million, a decrease of $5.74 million from the initial 33.3% award. Andren timely appeals.

II

Andren challenges the district court's new fee award in four ways. He first argues that the court erred in relying on any ex post awards to reconstruct the ex ante market. The court improperly disregarded the three ex ante bids class counsel made in other litigation, he asserts, which are more relevant than ex post awards. Second, Andren contends that the court erroneously relied on the cases in Klonoff's declaration, which listed only fee awards of 33% and above. Third, he posits that the court should not have excluded the Ninth Circuit “outliers” from its calculation. Last, he argues that the court was required to consider the stage at which this case settled and reduce the fee accordingly.

We review de novo whether the district court followed the correct methodology in awarding a fee. Broiler I, 80 F.4th at 801. So long as the district court understood its obligation to award fees according to the hypothetical ex ante bargain, as it did here, we review how it arrived at its fee only for abuse of discretion. Id. at 802. The district court has substantial latitude here, as it “is closer to the case than we are,” and reasonable fees “often fall within a broad range.” Id. (quoting In re Stericycle Sec. Litig., 35 F.4th 555, 559 (7th Cir. 2022)).

A

We first address Andren's challenge that the district court erred by relying on ex post awards to reconstruct the ex ante market. True, this court has said that ex ante agreements are “particularly useful guidepost[s] for determining the market rate.” Stericycle, 35 F.4th at 560. Those agreements most accurately shed light on the compensation attorneys are willing to accept “in the shadow of the litigation's uncertainty,” when there is still a risk of loss and when counsel has the highest incentive to offer competitive fees. Synthroid I, 264 F.3d at 719.

Yet this court has already “endorsed” looking to fee awards from other cases to recreate the ex ante market. Williams, 658 F.3d at 636; Taubenfeld v. AON Corp., 415 F.3d 597, 600 (7th Cir. 2005).³ Even if this is so, Andren counters, the district court improperly disregarded more relevant evidence: the three bids class counsel made in other cases. These all allowed for a maximum fee recovery between 13.5% and 20%, which Andren asserts is a more accurate representation of the ex ante market.

Though a district court must treat any prior bids as potentially relevant evidence, those bids do not necessarily set a cap for later cases. See Birchmeier, 896 F.3d at 797 (explaining that courts may award higher fees than in other cases when the present suit was “a riskier undertaking,” making prior cases poor comparators). Instead, because the district court is “closer to the case than we are,” Stericycle, 35 F.4th at 559, it is permitted to distinguish what it believes are inapposite cases.

The district court here persuasively explained why the three bids had little bearing on this award. It noted that all were for civil antitrust suits following governmental criminal investigations of the defendants, which greatly reduced the work required of class counsel. See id. at 564 (Prior litigation can “substantially reduce[ ] lead counsel's risk of nonpayment.”). Under our deferential standard of review, we cannot ask more. There was no abuse of discretion when discounting class counsels' bids entirely.

Even if the district court did not err in differentiating the three bids, Andren argues, it underweighted other ex ante evidence: the IRS fee agreement, negotiated with a sophisticated pension fund. The court viewed IRS and this case as “good comparators,” but it did not import the IRS fee structure here wholesale. Rather, it believed the potential settlement value of IRS was greater than in this case, so counsel was willing to accept a correspondingly lower percentage of the higher recovery. Related litigation against the financial-institution defendants in IRS had settled just before class counsel filed a similar suit, with the defendants paying about $2 billion in each case. By contrast, the district court noted that the largest antitrust settlement in history against a food-industry defendant was only one-sixth of that amount. It was eminently reasonable, then, for the court to conclude that IRS, with its high anticipated settlement value, was not a perfect comparator for this case.

Andren fails to show why the district court's thorough reasoning was an abuse of discretion. He points out that the two cases had similar total damages—both between $3.9 and $4.5 billion. But the district court recognized this. It instead focused on the potential settlement value, which is the more relevant metric to counsels' fee award. Nor is it dispositive that the IRS plaintiffs ultimately recovered $71 million, less than half the settlement here. The ex post recovery amount does not change the district court's reasoning for why class counsel may have accepted a lower percentage fee award ex ante.

The district court also did not exceed the bounds of reason when it gave the IRS award ten times the weight of any other case in its calculation. The court's spreadsheet contained 39 ex post awards in the $100 million to $1 billion recovery amount range relevant here. The court then counted the IRS percentage ten times, for a total of 49 awards. Although the IRS award was, in a sense, “diluted” by the volume of ex post figures, it was not an abuse of discretion for the district court to structure its calculation this way when presented only with imperfect ex ante comparators.

In sum, the district court did not “categorically” give “little weight” to class counsels' bids and negotiated fee structures because they used declining fee agreements. Broiler I, 80 F.4th at 803. The court instead soundly exercised its discretion in determining whether those cases served as good “indicators of what bargain would have been struck ex ante.” Id.

B

Andren more persuasively argues that the district court relied on a skewed sample of ex post awards. Recall that the court's calculation incorporated data from three sources. First, as just discussed, was the 26.6% award from IRS. Second was a representative summary of ex post awards class counsel received in other cases around the country. Third, the court cited awards contained in Klonoff's declaration—cases that were limited only to awards around the country of “33 percent or greater.”

District courts should consider all evidence bearing on an ex ante award. Indeed, this court has said that when awarding fees, district courts “need[ ] data rather than cherry-picked examples.” Silverman v. Motorola Sols., Inc., 739 F.3d 956, 959 (7th Cir. 2013). It is difficult to see how the cases in Klonoff's declaration provide a representative sample from which to derive an average award. The district court stated that “the sheer volume of ex post awards, relative to the minimal number of ex ante agreements, has a substantial impact on the expectations of class counsel” when deciding whether to represent a class. We do not disagree with this principle in theory. But when analyzing ex post awards, counsel are aware of all relevant awards—not solely those of 33% or more. So, to the extent a district court relies on a sampling of ex post awards to recreate the hypothetical market, it must ensure that sample is representative.

Class counsels' brief does not engage with this issue. Instead, it “[p]ut[s] aside whether the cited awards were ‘representative,’ ” asserting only that the district court was permitted to rely on public information, such as ex post fee awards. Nobody disputes that courts may rely on public information when fashioning awards. Yet this misses the methodological point: Although we decline to limit the various sources a district court may consult, it may not include data improperly skewing—high or low—the resulting award.

At oral argument before us, class counsel argued that a difference of only 3%—between the district court's revised 30% award and the IRS award of nearly 27%—cannot be an abuse of discretion. Oral Arg. at 18:50–22:00. This is incorrect for two reasons. When reviewing awards, this court does not focus solely on the percentage. Rather, what matters most is the process the district court used to calculate the award. Further, this argument is self-defeating. In Broiler I, we held that a 33% award was an abuse of discretion. Yet class counsel in this second appeal defend the district court's revised 30% award, tacitly admitting that a 3% difference may be reversible in some contexts.

Because the district court went through the effort of creating a detailed spreadsheet, we can “stick as close as possible to [its] approach” and adjust the fee on appeal. In re Synthroid Mktg. Litig. (Synthroid II), 325 F.3d 974, 980 (7th Cir. 2003); In re Trans Union Corp. Privacy Litig., 629 F.3d 741, 748 (7th Cir. 2011). The sole adjustment we make is to remove the fee awards listed only in Klonoff's skewed declaration. We do not change either the spreadsheet's inclusion of awards from class counsels' other cases, or the district court's decision to weight the IRS fee agreement ten times. We also limit the relevant awards, like the district court did, to those with recoveries between $100 million and $1 billion. After making this single change, we calculate that the average award is 27.1%, with a median of 26.6%. Andren requested a wide range of revised awards, with 26.6% at the top end. We therefore adjust the award to that amount.

C

Andren attempts to poke two more holes in the district court's award, but neither persuades us. First is that the court improperly excluded two Ninth Circuit “outliers” from its calculation. Even if these cases were erroneously left out, though, any error was harmless. Adding these cases back into the calculation changes the average somewhat, yet the median remains 26.6%. Because our adjusted fee award remains the same whether these cases are included or excluded, we need not address the merits of Andren's argument.

Andren also asserts that the district court failed to account for the stage of litigation at which this case settled. True enough, we have reversed a fee award when “the district court did not give sufficient weight to the early stage” of settlement. Stericycle, 35 F.4th at 566. A large award is unwarranted if it is “disproportionate to the amount of work expended by class counsel.” Id. (quoting Camp Drug Store, Inc. v. Cochran Wholesale Pharm., Inc., 897 F.3d 825, 833 (7th Cir. 2018)). But this circuit has also explained that courts should avoid “a second major litigation strictly over attorneys' fees.” Siddiqui v. Nat'l Ass'n of Broad. Emps. & Technicians – Commc'ns Workers, 132 F.4th 530, 533 (7th Cir. 2025) (quoting Vega v. Chi. Park Dist., 12 F.4th 696, 702 (7th Cir. 2021)).

In its initial fee order, the district court soundly considered the amount of work expended by class counsel. It noted the lack of prior government investigation here, requiring counsel to “invest[ ] massive resources of time and money”—over 67,500 attorney hours—in guiding the litigation through extensive discovery. The court further discussed counsels' “exemplary” performance, noting how the plaintiffs survived a “relatively close” motion to dismiss with the assistance of in-depth briefing. Indeed, the court's order denying dismissal tallied 92 pages. And although the case settled before a ruling on class certification, class counsel fully briefed the certification issue. This was not a case where “no real litigation” occurred; counsel put forth significant effort, which the district court properly considered. Camp Drug Store, 897 F.3d at 833.

The district court examined in detail the work counsel devoted to this case. To ask for anything further would be to demand a “second major litigation” over fees. Siddiqui, 132 F.4th at 533 (quoting Vega, 12 F.4th at 702).

* * *

We compliment the district court for creating a detailed fee calculation, spreadsheet and all. After removing the few non-representative data points from its analysis, class counsel are entitled to 26.6% of the net common fund. We AFFIRM the district court's fee award AS MODIFIED and REMAND the case for further proceedings.

FOOTNOTES

1.   Although the district court also considered a fourth source of awards, those were outside of the relevant range, as noted in the following paragraph. We therefore do not discuss them here.

2.   These were three data points on the spreadsheet, but two cases were duplicates.

3.   Andren argues that this court's Taubenfeld decision approved of ex post fee awards only because the objector there forfeited her market-mimicking argument. See 415 F.3d at 599. But this court also addressed this point in Williams, expressly approving looking to other cases' fee awards when reconstructing the ex ante market. 658 F.3d at 636, aff'g, 2010 WL 4723725, at *2 (S.D. Ind., Nov. 12, 2010) (awarding a fee after considering affidavits from class counsel listing, among other things, awards from other cases); see also Birchmeier v. Caribbean Cruise Line, Inc., 896 F.3d 792, 796–97 (7th Cir. 2018), aff'g, Aranda v. Caribbean Cruise Line, Inc., 2017 WL 1369741, at *6 (N.D. Ill. Apr. 10, 2017) (looking to fee awards from other cases).

Brennan, Circuit Judge.