PRECEDENTIAL
UNITED STATES COURT OF APPEALS FOR THE THIRD CIRCUIT
______________
No. 22-1950
______________
IN RE WAWA, INC. DATA SECURITY LITIGATION
THEODORE H. FRANK,
Appellant
______________
On Appeal from the United States District Court for the Eastern District of Pennsylvania (D.C. Civil No. 2:19-cv-06019)
District Judge: Honorable Gene E. K. Pratter
______________
Argued March 30, 2023
Before: MATEY, FREEMAN, and FUENTES, Circuit Judges.
(Filed: November 2, 2023)
______________
Theodore H. Frank Adam E. Schulman [ARGUED] Hamilton Lincoln Law Institute 1629 K Street, N.W. Suite 300 Washington, DC 20006 Counsel for Appellant
Donald E. Haviland, Jr. Haviland Hughes 201 South Maple Street Suite 110 Ambler, PA 19002
Gerard A. Dever Roberta D. Liebenberg Fine Kaplan & Black One South Broad Street Suite 2300 Philadelphia, PA 19107
Samantha E. Holbrook Benjamin F. Johns [ARGUED] Jonathan Shub Shub & Johns 200 Barr Harbor Drive Four Tower Bridge, Suite 400 West Conshohocken, PA 19428 Counsel for Plaintiffs-Appellees
3
Kristin M. Hadgis Gregory T. Parks [ARGUED] Morgan, Lewis & Bockius 2222 Market Street Philadelphia, PA 19103
Michael E. Kenneally Morgan, Lewis & Bockius 1111 Pennsylvania Avenue, N.W. Suite 800 North Washington, DC 20004 Counsel for Defendants-Appellees
Melissa Holyoak Office of Attorney General of Utah 350 North State Street Suite 230 Salt Lake City, UT 84114 Counsel for Amicus Appellant
______________
OPINION OF THE COURT
______________
MATEY, Circuit Judge.
Convenience is king at Wawa, Inc., where guests are invited to gas up, chow down, and swipe, tap, or click to pay before heading on their way. Throughout 2019, uninvited guests stopped by too. Hackers, who infiltrated Wawa’s payment systems and helped themselves to the credit and bank card data of some twenty-two million customers. Wawa announced the breach on December 19, 2019; by the next day,
4
attorneys had rounded up plaintiffs and filed the first of many class action suits seeking damages for the disclosures. A brisk nine months later, Wawa and plaintiffs’ class counsel shook hands on a settlement making $9 million in gift cards and some other compensation available to customers (of which $2.9 million was claimed) and giving $3.2 million to class counsel for fees and expenses (the “Settlement Agreement”). Objections arrived, prompting modifications to the proposal. But the changes are not enough to ensure class counsel receives only a reasonable fee award, and we clarify two considerations that loom large in that calculation: the ratio between the fee award and amount recovered by the class members, and side agreements between class counsel and the defendant. Because the District Court lacked the benefit of our fresh guidance, we will vacate the fee award and remand for further consideration.
I.
When Wawa announced that malware had been stealing payment information for nearly a year, litigation erupted overnight. Moving to order a ballooning docket, the District Court consolidated the multiplying lawsuits into one class action with three tracks: financial institutions, employees, and consumers. The resulting master complaint asserts claims against Wawa for negligence, negligence per se, breach of implied contract, unjust enrichment, and violations of multiple states’ consumer protection and data privacy laws. Our focus is the consumer track plaintiffs who reached a proposed settlement in September 2020 (the “Proposed Settlement Class”).
5
The Proposed Settlement Class includes around 22 million people 1 who used electronic payments (be it credit, debit, or something else) at a Wawa between March 4, 2019, and December 12, 2019. The Settlement Agreement provided three tiers of relief:
Tier 1 customers who attest that they spent at least some time monitoring their credit can get a $5 Wawa gift card. Total Tier 1 compensation is subject to a $6 million cap and a $1 million floor.
Tier 2 customers who saw a fraudulent charge that required some effort to sort out can receive a $15 Wawa gift card for their trouble. Total Tier 2 compensation is subject to a $2 million cap with no floor.
Tier 3 customers who show certain out-ofpocket losses caused by the breach can receive $500 (in currency, not Wawa gift cards). Total Tier 3 compensation is subject to a $1 million cap without a floor.
The Settlement Agreement also specified injunctive relief, including upgraded security and processing systems, which
6
class counsel and Wawa valued at about $35 million. For this work, class counsel sought a lump-sum award of $3.2 million, comprised of $3,040,060 in attorney’s fees, $45,940 in litigation expenses, some $100,000 in settlement administration fees, and $14,000 in class representative awards. The parties added those fees, expenses, and awards to the $9 million offered to the class to create what they call a “constructive common fund” of $12.2 million. App. 19–20, 22. That combination of attorney and class recovery into a single amount is at center stage in this appeal. 2
7
Class member Theodore H. Frank objected to the settlement and the request for attorney’s fees. Frank argued the constructive common fund was miscalculated and the settlement unfair because, stripped of the labels, class counsel would receive a disproportionate share of the amount Wawa would pay in gift cards or cash. And he pointed to other perks class counsel secured in the deal, including a “clear sailing” clause, under which Wawa agreed not to contest class counsel’s fee petition. 3 He also objected to the “fee reversion,”
8
a provision that returned any reductions in the fee award to Wawa, and not to the class. Frank urged a different approach: cap attorney’s fees at 25% of the actual claims made and paid, rather than funds and gift cards offered but never used.
Amendments followed Frank’s objections. A Second Amended Settlement clarified that the gift cards would not expire and granted automatic eligibility for Tier 1 gift cards to Wawa app users with valid email addresses. And a Third Amended Settlement eliminated the fee reversion so any reduction in fees awarded would be redistributed to Tier 1 and Tier 2 gift card holders. Finally, a claims administrator would email the 575,162 eligible Wawa app users, explaining that they will receive $5 electronic gift cards once the settlement is finalized. The administrator also plans to remind unused gift card holders to use their credit by sending an email nine months after distribution. These adjustments boosted the estimated redemption rate from about 0.035% (about 8,000 claims of the 22 million class members) to as much as 2.6% (around 564,000 claims). This brought the total projected distribution amount to $2,905,195, including $2,815,075 for Tier 1 (up from $33,720 before the amendment), $10,290 for Tier 2, and $79,830 for Tier 3.
Frank then withdrew his objection to the settlement. But he maintained his objection to the attorney’s fees because they
9
were still based on the constructive common fund, not the amounts paid to the class, a several million-dollar difference. He also pointed to the never-deleted clear sailing clause as evidence of collusion between class counsel and Wawa.
The District Court disagreed, endorsing the $12.2 million calculation for the constructive common fund and finding that class counsel’s requested $3,040,060 fee award— totaling just shy of 25% of that fund—was not unreasonable. Analyzing the fee award under the factors outlined in Gunter v. Ridgewood Energy Corp., 223 F.3d 190, 195 n.1 (3d Cir. 2000), the District Court found that class counsel’s blended $653 hourly rate was reasonable; the litigation was complex; there was a substantial risk of nonpayment (since class counsel worked on contingency); and the total payout fell below other data breach settlements. Cross-checking those conclusions, the District Court ran a lodestar analysis—resulting in an award of roughly $3.8 million. Class counsel’s requested fees of $3,040,060 is less than that number.
Finally, the District Court found that the clear sailing clause was typical of class action settlements. And the District Court was satisfied there was no collusion because an independent mediator attested that the fee agreement was discussed only after the terms of the class settlement were already set. So the District Court approved the settlement and the fee award. Having withdrawn his objection to the settlement’s approval under Rule 23(e), Frank now appeals the fee award as unreasonable under Rule 23(h). 4
10
II.
Frank contests the District Court’s $3,040,060 attorney’s fee award to class counsel. 5 Attorney’s fee awards are governed by Rule 23(h) of the Federal Rules of Civil Procedure, 6 which demands that any awarded fees be
11
“reasonable”—a capacious phrase refined by reference to the history that surrounds it. All with the goal of “giv[ing] effect to the rule maker’s aim.” Epsilon Energy USA, Inc. v. Chesapeake Appalachia, LLC, 80 F.4th 223, 230 (3d Cir. 2023) (citing Brown v. Barry, 3 U.S. (3 Dall.) 365, 367 (1797)). We lay out that analysis below but start with the takeaway.
12
Two considerations must play central roles in the assessment of a fee award under Rule 23(h): 1) how the amount awarded stacks up against the benefit given to the class, using either the amounts paid or the sums promised; 7 and 2) whether side agreements between class counsel and the defendant suggest an unreasonable attorney’s fee award. We will vacate the District Court’s order approving the fee award and remand for a hard look at these “red flags.” 8
13
A.
We start with the text of Rule 23(h): “In a certified class action, the court may award reasonable attorney’s fees and nontaxable costs that are authorized by law or by the parties’ agreement.” Fed. R. Civ. P. 23(h) (emphasis added). 9 Frank claims the District Court erred when it found the attorney’s fees reasonable because the District Court: 1) considered only “the funds made available to class members rather than the amount actually claimed during the claims process,” App. 19; and 2) inadequately scrutinized any side agreements between class counsel and Wawa. The text of Rule 23(h)—that the award must be “reasonable”—when read in history and context, explains why those objections are correct.
1.
Arguments about the reasonableness of attorney’s fees arrived relatively recently in our profession’s history. 10 In the early
14
American colonies, fees paid to lawyers largely resembled those paid in England, where the prevailing party “recovered attorney fees as part of the costs, and the right to recovery was grounded on statute.” John Leubsdorf, Toward a History of the American Rule on Attorney Fee Recovery, 47 L. Contemp. Probs. 9, 12 (1984) (citing 3 Blackstone, Commentaries *399–401); see also Alyeska Pipeline Serv. Co. v. Wilderness Soc’y, 421 U.S. 240, 247 & n.18 (1975). Juries usually included these costs in their calculations when determining damages, and in practice the recovered costs were often low. See 3 Blackstone, Commentaries *399; Leubsdorf, supra, at 11–12, 14. Lawyers in the colonies and early Republic regularly recovered less than they wanted because of statutory limits on attorney’s fees and cost awards. 11 The reasonableness of fees only became an issue when legislatures began repealing these statutory limits and American courts started applying a new principle requiring each litigant to pay their own attorney’s fees, “win or lose, unless a statute or contract provides otherwise.” Hardt v. Reliance Standard Life Ins. Co., 560 U.S. 242, 252–53 (2010); see also Leubsdorf, supra, at
15
13–14. 12 This “American Rule” has since been called a “bedrock principle,” Baker Botts L.L.P. v. ASARCO LLC, 576 U.S. 121, 126 (2015), applied broadly with only a few longrunning exceptions.
One of the “well-recognized” exceptions to the American Rule is the common fund. Boeing Co. v. Van Gemert, 444 U.S. 472, 478 (1980). The theory is rooted in the equitable principle that a “lawyer who recovers a common fund for the benefit of persons other than himself or his client is entitled to a reasonable attorney’s fee from the fund as a whole.” Id.; see also Samuel R. Berger, Court Awarded Attorneys’ Fees: What is “Reasonable”?, 126 U. Pa. L. Rev. 281, 281–82 (1977) (noting “the historic equity power of the federal courts to compel all of the beneficiaries of a ‘common fund’ recovered or preserved by the plaintiff to pay, out of the fund, their proportionate share of the compensation to which plaintiff’s attorneys are entitled”).
Reasonableness has always been the measurement for fees in a common fund, beginning with Trustees v. Greenough, which adopted the equitable practice of paying fees “where one of many parties having a common interest in a trust fund, at his own expense takes proper proceedings to save it from
16
destruction and to restore it to the purposes of the trust.” 105 U.S. 527, 532–33 (1881). Subsequent cases granting attorney’s fees reinforced this equitable practice and evaluated awards against “the standard of reasonableness.” United States v. Equitable Tr. Co. of New York, 283 U.S. 738, 744, 746 (1931) (“It is a general rule in courts of equity that a trust fund which has been recovered or preserved [by an advocate may be] charged with the costs and expenses, including reasonable attorney’s fees, incurred in that behalf.”). Cases dealing with common funds emphasized that reasonableness is tied to the benefit rendered to the class. See, e.g., Cent. R.R. Banking Co. of Georgia v. Pettus, 113 U.S. 116, 124–26 (1885) (approving “reasonable compensation” for attorneys’ “professional services . . . and that such compensation should be made with reference to the amount of all claims filed in the cause”). The adoption of Rule 23 in 1938 did not change this wellestablished practice, and courts continued to evaluate attorney’s fees for reasonableness. See, e.g., Powell v. Pennsylvania R.R. Co., 267 F.2d 241, 245–46 (3d Cir. 1959) (determining that an attorney’s fee award was “reasonable” by “considering all the facts”). Nor did the significant amendments to Rule 23 in 1966, which omitted any mention of attorney’s fees and continued to commit reasonableness to judicial discretion. 13
17
So how was that discretion exercised? By scrutinizing “the size of the fund or the amount of benefit produced for the class” to calculate an award based on a “reasonable percentage” of the amount the class recovered. Court Awarded Attorney Fees: Report of the Third Circuit Task Force, 108 F.R.D. 237, 242 (1986). But the shift to a reasonable percentage raised problems. Though rooted in “reasonableness,” awards using a percentage-of-recovery approach sometimes resulted in “strikingly large” attorney’s fees. Id. That led to complaints from the bar and the public that the money disbursed was disproportionately generous given the limited work of class counsel. Id. In response, this Court led the charge away from the percentage-of-recovery method in Lindy Brothers Builders, Inc. of Philadelphia v. American Radiator & Standard Sanitary Corp., 487 F.2d 161 (3d Cir. 1973). There, we coined the lodestar method of calculating fees, which computes the reasonable hours expended by counsel multiplied by a reasonable hourly rate, then adjusts up or down to account for case-specific variables. Id. at 167–68. Other federal courts soon followed our lead, agreeing that the lodestar approach was more sensible than the percentage-of- recovery method. 14
18
But again, criticism stewed. Some, including this Court, 15 complained the Lindy lodestar analysis replaced old problems with new ones—like a perverse incentive for attorneys to inflate their billing rates, “expend excessive hours,” and “engage in duplicative and unjustified work.” Report of the Third Circuit Task Force, 108 F.R.D. at 248. Others lamented the widespread variation in fee awards. 16 Id.; see also John C. Coffee, Jr., Understanding the Plaintiff’s Attorney: The Implications of Economic Theory for Private Enforcement of Law Through Class and Derivative Actions, 86 Colum. L. Rev. 669, 675–76 (1986). Reasonableness still needed a reasonable standard.
2.
The Supreme Court entered the fray in Boeing Co. v.
Van Gemert, 444 U.S. 472 (1980), and affirmed the use of the percentage-of-the-fund method in common-fund cases. In
19
claims-made settlements, Boeing continued, courts can consider funds offered to the class but never used. 17 444 U.S. at 480. But, at the same time, Boeing created no rule requiring courts to use only the percentage of total funds the defendant made available. Nor did it take a position on the converse: “basing attorneys’ fees on only the amount of the fund claimed by class members.” In re Baby Prods. Antitrust Litig., 708 F.3d 163, 177 (3d Cir. 2013) (emphasis added). That choice remained with the district courts. 18
20
This Court has likewise reserved that power to the trial judge based on the history supporting the American Rule. To guide this analysis, we have said that district courts should “consider the level of direct benefit provided to the class in calculating attorneys’ fee[]” awards, which “needs to be, as much as possible, practical and not abstract,” and “may
21
consider, among other things,” the claims rate. In re Baby Prods., 708 F.3d at 170, 174. That has led courts to “delay a final assessment of the fee award to withhold all or a substantial part of the fee until the distribution process is complete.” In re Baby Prods., 708 F.3d at 179 (quoting Manual for Complex Litigation § 21.71 (4th ed. 2008)). But regardless of whether courts use the amount made available or the amount claimed, we have explained the fees must be analyzed against the benefits to the class case-by-case. See In re Prudential Ins. Co. Am. Sales Prac. Litig. Agent Actions, 148 F.3d 283, 334, 342 (3d Cir. 1998) (“What is important is that the district court evaluate what class counsel actually did and how it benefitted the class.”); see also In re Gen. Motors Corp. Pick-Up Truck Fuel Tank Prods. Liab. Litig., 55 F.3d 768, 822 (3d Cir. 1995) (remanding for “some reasonable assessment of the settlement’s value and determine the precise percentage represented by attorneys’ fees”).
Though this inquiry into reasonableness involves discretion, it is not without detailed demands. Boeing highlighted features of class action settlements that inform judicial focus, such as when defendants were liable for a “sum certain,” with each class member entitled to “logically ascertainable shares” of the fund. 444 U.S. at 479–81. In cases where defendants keep any unclaimed funds, making their liabilities “contingent upon the presentation of individual claims,” id. at 479 n.5, courts must place greater weight on the claims rate. 19 And when class members must do more than raise
22
their hands to get their payment, the claims rate offers valuable insight into the “effectiveness” of “the method of processing class-member claims.” Fed. R. Civ. P. 23(e)(2)(C)(ii); see also Fed. R. Civ. P. 23, advisory committee’s note to 2018 amendments (describing the utility of courts reviewing the “contemplated claims process and the anticipated rate of claims by class members”). Finally, class members naturally value cash over gift cards and disfavor coupons or similar “hot button indicators” that “show . . . potential unfairness on their face.” 20 Courts should take special notice when class members are offered discounts and tickets while others—like counsel—get cash. See In re Gen. Motors, 55 F.3d at 803 (“[N]on-cash relief . . . is recognized as a prime indicator of suspect settlements”); see also In re Dry Max Pampers Litig., 724 F.3d 713, 718–21 (6th Cir. 2013) (finding settlement provided “illusory” relief to the class when counsel received cash and class members received the opportunity for a refund). These, and similar considerations, help determine the benefit class counsel provided, and how much should be used to ca
23
lculate a reasonable fee percentage. Cf. Lowery v. Rhapsody Int’l, Inc., 75 F.4th 985, 993 (9th Cir. 2023); Pearson v. NBTY, Inc., 772 F.3d 778, 782 (7th Cir. 2014).
3.
With the background painted, we return to Rule 23(h)
and its instruction that a “court may award reasonable attorney’s fees and nontaxable costs that are authorized by law or by the parties’ agreement.” Fed. R. Civ. P. 23(h). 21 Prior to addin
24
g subdivision (h) in 2003, awards of attorney’s fees were governed by Rule 54, which included no reasonableness requirement. See Fed. R. Civ. P. 23, advisory committee’s note to 2003 amendments; see generally Report of the Judicial Conference of the United States to the Committees on the Judiciary of the Senate and House of Representatives, Class Action Settlements (2006). But Rule 23(h) was added in 2003 to incorporate the reasonableness standard that had long been “customary” in common fund cases and class actions. See Report of the Judicial Conference at 2–4; Linda S. Mullenix, No Exit: Mandatory Class Actions in the New Millennium and the Blurring of Categorical Imperatives, 2003 U. Chi. Legal F. 177, 177 (2003).
25
The advisory committee’s note recognizes the rich discussion on reasonableness—its utility and its limitations— that had been occurring before the adoption of subdivision (h). It notes that determining the reasonableness of an award turns on a “variety of factors,” including the calculation of the award using either the lodestar or percentage-of-recovery method. See Fed. R. Civ. P. 23, advisory committee’s note to 2003 amendments. Whatever the methodology, one focus remained “fundamental”—“the result actually achieved for class members.” Id. If the award were calculated as a percentage of the class’s recovery, “results achieved is the basic starting point.” Id. Though the committee suggested that courts may consider the percentage of an award using the amount “actually paid to the class,” it refrained from imposing that amount as the required denominator in every case. Id. Indeed, courts can evaluate the reasonableness of a percentage-based award by reference to either amounts paid or amounts made available. See Manual for Complex Litigation § 14.121 (4th ed. updated 2023).
Assessing a reasonable fee award also requires courts to take a hard look at side agreements 22 between class counsel and the defendant. See Fed. R. Civ. P. 23, advisory committee’s
26
note to 2003 amendments (“Courts have also given weight to agreements among the parties regarding the fee motion, and to agreements between class counsel and others about the fees claimed by the motion.”). Courts, for instance, must be on the lookout for clear sailing clauses, which amount to “agreement[s] by a settling party not to oppose a fee application up to a certain amount.” Id. So too with fee reversions, which “provide[] that if the judge reduces the amount of fees that the proposed settlement awards to class counsel, the savings shall enure not to the class but to the defendant.” Pearson, 772 F.3d at 786.
B.
Against this framework, we will vacate the District Court’s grant of class counsel’s fee petition and remand to consider whether “the funds made available to class members rather than the amount actually claimed during the claims process” is the best measure of reasonableness, App. 19; and whether the fee award is reasonable in light of any side agreements between class counsel and Wawa.
First, the District Court saw itself as bound to consider only the funds made available to the class. 23 But that limitation is no
27
t required by history or precedent. Rather, we have “recognize[d] the difficulty a district court faces” in calculating attorney’s fees before class relief is given out, In re Baby Prods., 708 F.3d at 179, and for that reason, “[i]t is common to delay a final assessment of the fee award and to withhold all or a substantial part of the fee until the distribution process is complete.” Manual for Complex Litigation § 21.71 (4th ed. updated 2023). And while that practice is not required by Rule 23, it seems a sensible starting line to begin the fee award analysis. So we remand for consideration of the amounts distributed to and expected to be claimed by the class.
Next, side agreements between class counsel and Wawa require deeper inquiry to assess whether the fee award is reasonable. Start with the clear sailing provision, where Wawa promised as part of the settlement not to challenge class counsel’s request for an agreed-upon attorney’s fee award. Though not an automatic bar to settlement approval, 24 such terms still “deserve careful scrutiny” when calculating a reasonable fee award. In re Nat’l Football League Players Concussion Inj. Litig., 821 F.3d 410, 447 (3d Cir. 2016), as amended (May 2, 2016). “The concern with a clear sailing provision is collusion,” and class counsel’s desire to maintain its expected fees could tempt it to take money from the class in return for a defendant’s agreement to swiftly settle. Id. So a “district court faced with such a provision in a class action
28
settlement should review the process and substance of the settlement and satisfy itself that the agreement does not indicate collusion or otherwise pose a problem.” Id.
The same concerns apply when assessing fee petitions.
The District Court correctly identified that clear sailing provisions require close attention. But we will remand for closer scrutiny based on our refreshed guidance. The District Court found that the clear sailing provision was not collusive because an independent mediator helped the negotiations and explained the provision arrived after there was agreement on the tiered terms for class relief. That outside oversight, while not irrelevant, is alone insufficient, because “the mere presence of a neutral mediator, though a factor weighing in favor of a finding of non-collusiveness, is not on its own dispositive of whether the end product is a fair, adequate, and reasonable settlement agreement.” In re Bluetooth Headset Prods. Liab. Litig., 654 F.3d 935, 948 (9th Cir. 2011).
Nor does the fact that the agreement came after the parties had settled class compensation end the inquiry. See In re Gen. Motors, 55 F.3d at 803–05, 805 n.24. That is because “[c]lass counsel cannot be unaware that fee negotiations are nigh—that is, after all, how plaintiffs’ lawyers finance their work—and that knowledge simply might cause them to push less hard for the interests of their clients, even if they fail to realize that they are doing so.” Gascho v. Glob. Fitness Holdings, LLC, 822 F.3d 269, 302 (6th Cir. 2016) (Clay, J., dissenting). Defendants, too, are already estimating the impending fee request. “Caring only about his total liability, the defendant will not agree to class benefits so generous that when added to a reasonable attorney’s fee award for class counsel they will render the total cost of settlement
29
unacceptable to the defendant.” Pearson, 772 F.3d at 786. Concerns like these make close and careful review of a clear sailing provision necessary when evaluating the reasonableness of a fee award.
Finally, there is the puzzling fee reversion (also known as a reverter or kicker clause), providing that any court-ordered reduction in the attorney’s fee award would be returned to Wawa—not the class. Ordinarily, as is the case here, class members who suspect class counsel has taken an excessive share of the common fund in attorney’s fees can object and ask the court to reduce that share. But “when parties agree to a ‘kicker,’ a 23(h) challenge cannot increase class recovery because the excessive fees wind up back in the defendant’s pockets.” Briseño v. Henderson, 998 F.3d 1014, 1027 (9th Cir. 2021). It is a bewildering proviso: why would class counsel agree to give part of the common fund they secured to the defendant instead of their clients? The unfortunate conclusion is that class counsel asks for it as a “gimmick for defeating objectors.” Pearson, 772 F.3d at 786. A trick that strips class members of their standing to challenge the fee award, because “any action taken by the court would not redress the class member’s purported injury.” Briseño, 998 F.3d at 1027. And when combined with a clear sailing clause, in which the defendant does not object to the fee award, any action under 23(h) is foreclosed. Id.
The original Settlement Agreement contained a fee reversion eventually removed in the Third Amended Settlement. A welcome change, but not as welcome as if the fee reversion had never existed. That is because a fee reversion need not stay in the final approved settlement to serve its deterrent purpose, so courts should investigate potential
30
collusion by considering the “evidence in the negotiation process or the final terms of the settlement.” In re Nat’l Football League, 821 F.3d at 447. On remand, the District Court should explore how the reversion arrived, what purpose it served, and whether its presence, even temporary, suggests coordinated rather than zealous advocacy, that makes the fee request unreasonable. 25
***
We will vacate and remand the attorney’s fee award for the District Court to take a closer look at the reasonableness of the attorney’s fees in proportion to class benefit and to scrutinize the presence of side agreements.