In Re Glumetza Antitrust Litigation

Opinion

NORTHERN DISTRICT OF CALIFORNIA

In re No. C 19-05822 WHA GLUMETZA ANTITRUST No. C 19-05831 WHA LITIGATION. No. C 19-06138 WHA No. C 19-06156 WHA No. C 19-06839 WHA This Document Relates to: No. C 19-07843 WHA

No. C 19-05822 WHA (Consolidated) No. C 19-06138 WHA No. C 19-06839 WHA ORDER RE FINAL APPROVAL OF CLASS SETTLEMENTS AND MOTION FOR ATTORNEY’S FEES

In this antitrust action, defendant brand and generic manufacturers of a type 2 diabetes drug engaged in a reverse-payment scheme and agreed not to compete with each other by delaying the introduction of a generic version of the drug for several years. On the precipice of trial, the direct-purchaser class reached settlement agreements with all defendants. To the extent stated, the motion for final approval of the settlements is GRANTED. Class counsel’s motion for attorney’s fees and costs is GRANTED IN PART. The order denying summary judgment described the facts of this action in detail (Dkt. No. 537). Briefly, in 2009, pharmaceutical company Depomed, Inc. sued Lupin Pharmaceuticals, Inc. in its various corporate forms for infringing Depomed’s patents on blood sugar levels for individuals suffering from type 2 diabetes. Lupin had entered Depomed’s crosshairs because it had filed an Abbreviated New Drug Application (ANDA) with the FDA, seeking to manufacture and market a generic version of Glumetza. The parties litigated before the Honorable Phyllis J. Hamilton of our district until a claim construction order prompted the parties to settle in February 2012. But this settlement hid an underlying conspiracy from Judge Hamilton — a quid pro quo where Lupin would delay market entry of its generic for four years (until February 1, 2016) and, in return, Depomed paid Lupin $3 million and promised it would not launch an authorized generic to compete with Lupin for one year (until February 2017, atop an FDA-granted 180- day period of market exclusivity) and would also protect Lupin from other generic competition. First, if any other manufacturer succeeded in marketing a generic Glumetza before February 2016, Lupin could market immediately. Second, Depomed covenanted not to license any other generic Glumetza manufacturers until at least 180 days following Lupin’s market entry. The agreement resulted in a windfall for defendants. Depomed had already sold its marketing rights in Glumetza to defendant Santarus, Inc., but then sold its royalty rights to PDL Biopharma in October 2013 (Dkt. No. 464-1). The juiced Glumetza portfolio aided Santarus’ sale to defendant Salix Pharmaceuticals for $2.6 billion in November 2013, and Salix’s sale to defendant Valeant Pharmaceuticals for $14.5 billion in April 2015. Then, in the summer of 2015, Valeant hiked the price of Glumetza on the order of 800%. The wholesale acquisition cost of a 500 mg pill spiked from $5.72 in May to $51.48 by the end of July. At the same time, the cost of a 1000 mg pill rose from $12.37 to $113.36. When Lupin joined the market in February, the supposed-affordable generic did so at around $45 for each 500 mg tablet. Both direct and indirect purchasers started suing in September 2019. The cases were consolidated before the undersigned, the indirect purchasers eventually exited the case, and a direct-purchaser class was certified in August 2020. An order dated May 6, 2021, denied all late July 2021, under the supervision of Judge Donna M. Ryu, the class reached a settlement agreement with Bausch. In August, an order ruled on the parties’ ten Daubert motions, and the final pre-trial conference then occurred over two days, on September 8 and 9. All of the parties’ motions in limine were decided therein. By that point, the class had also reached an agreement with Assertio. Shortly after the hearing on September 8 concluded, the class filed a motion for preliminary approval of the Bausch settlement. Class counsel filed an amended motion for the Bausch settlement on September 14, and a preliminary approval motion regarding the Assertio agreement followed on September 16. On September 20, the class filed a motion for preliminary settlement approval for Lupin, the last to fall. A September 22 order granted preliminary approval to all three settlements. The final approval hearing took place telephonically due to the ongoing COVID-19 pandemic. Class counsel notified all class members of the telephonic hearing via the class- action website and via an email from the claims administrator. The docket and the Court calendar also explained how to attend the hearing remotely. The three national wholesalers that had filed a written objection to class counsel’s motion for attorney’s fees participated in the hearing. No other class members joined the call or appeared in the courtroom, which remained open for class members wishing to take part in the hearing. “The class action device, while capable of the fair and efficient adjudication of a large number of claims, is also susceptible to abuse and carries with it certain inherent structural risks.” Officers for Just. v. Civ. Serv. Comm’n of City & Cty. of San Francisco, 688 F.2d 615, 623 (9th Cir. 1982). A settlement purporting to bind absent class members must be fair, reasonable, and adequate. See FRCP 23(e). Rule 23(e)(2) requires district courts to employ a two-step process: First, the parties must show the district court will likely be able to approve the proposed settlement. Second, the district court must hold a hearing to make a final determination of whether the settlement is fair, reasonable, and adequate. We have arrived at step two. Our court of appeals recently explained that the final fairness assessment must analyze the eight Churchill factors: (1) the strength of the plaintiff’s case; (2) the suit’s risk, expense, complexity, and the likely duration of further litigation; (3) the risk of maintaining class action status throughout the trial; (4) the amount offered in settlement; (5) the extent of discovery and the stage of the proceedings; (6) the experience and views of counsel; (7) the presence of a governmental participant (if any); and (8) the reaction of the class members to the proposed settlement. Kim v. Allison, 8 F.4th 1170, 1178–79 (9th Cir. 2021) (quoting In re Bluetooth Headset Prods. Liab. Litig., 654 F.3d 935, 946 (9th Cir. 2011)); Churchill Vill. v. Gen. Elec., 361 F.3d 566 (9th Cir. 2004). Additionally, Rule 23(e)(2) requires the district court to consider an overlapping set of factors, including the adequacy of the notice procedure, “the terms of any proposed award of attorney’s fees,” to scrutinize the settlement for evidence of collusion or conflicts of interest, and to review other, relevant factors before deeming the settlement fair. Kim, 8 F.4th at 1179; Briseño v. Henderson, 998 F.3d 1014, 1023–26 (9th Cir. 2021). Among the other relevant factors that will be considered are those listed by this Court in its notice regarding factors to be evaluated for any proposed class settlement, filed herein November 4, 2019 (Dkt. No. 39). In short, in consideration for the dismissal of this action with prejudice and a release of all claims, Bausch, Assertio, and Lupin agree to make settlement payments of $300 million, $3.85 million, and $150 million, respectively, to be allocated to the direct-purchaser class members on a pro rata basis. Defendants’ settlements can be considered jointly for many of the factors this order considers, others require individualized analyses. 1. THE CHURCHILL FACTORS. We begin with the eight Churchill factors. They support settlement with the defendants. First, the strength of the direct purchaser’s case favors settlement. In this pharmaceutical reverse-payment action, plaintiffs would have had the burden of establishing defendants’ violation of the rule of reason then proving a causal antitrust injury. See FTC v. Actavis, 570 U.S. 136 (2013). As a threshold issue, the rule of reason analysis would have required plaintiffs to demonstrate that defendants wielded market power. To that end, the summary judgment order denied plaintiffs’ motion for summary judgment on that issue: True, a reasonable trier of fact could find that defendants wielded market power during the relevant time period. . . . But a reasonable trier could reject these considerations in light of defendants’ evidence of the cross-elasticity of demand between Glumetza and the many, and more prevalent, variants of metformin hydrochloride available to consumers (Summary Judgment Order 10, Dkt. No. 537). Beyond market power, as part of the rule of reason analysis, defendants would have also been able to offer to our jury procompetitive justifications for the reverse payment, such as defendants’ subjective beliefs that Depomed would win the patent litigation (Br. 9). Moreover, although rejected at the summary judgment stage, defendants had several arguments why plaintiffs could not demonstrate causation. At trial, plaintiffs would have had to argue that either Lupin and the brand defendants would have settled their patent litigation with an earlier agreed-upon entry date, or that Lupin would have ultimately won the patent litigation. Both paths had noteworthy pitfalls. The Daubert order, for example, declined to exclude either sides patent-law expert, or the technical experts which they relied upon, paving the way for a battle of the experts at trial (Dkt. No. 593). As stated in summary judgment, material issues abounded, many of which weighed against plaintiffs. And even if plaintiffs achieved victory at trial, the almost inevitable appeal would have required further litigation of these issues before our court of appeals. Second, the risk, expense, complexity, and likely duration of the case supports settlement. As explained for the first factor, plaintiffs faced notable risks in obtaining a positive result in this litigation. Moreover, a reverse payment antitrust action such as this one requires evaluation of the underlying patent litigation, which in turn involves still more legal frameworks, such as the Hatch-Waxman Act. This action also had the additional wrinkle of a strategic, selective waiver of privilege by generic defendant Lupin, without corresponding waivers by brand defendants Bausch and Assertio. This waiver of privilege by a single players in the pharmaceutical industry comes large law firms and legal bills. Plaintiffs faced a long trial ahead of them. Third, the risk of maintaining class action status throughout trial is neutral. The direct- purchaser class was certified in an August 2020 (Dkt. No. 347). Fourth, the amount of the award supports the settlement agreement for Bausch and Lupin, and it is neutral for Assertio. The combined settlement amount of $453,850,000 constitutes the third-largest absolute dollar recovery for a direct-purchaser class in a generic delay pharmaceutical antitrust action (Dkt. No. 681 App’x A). Counsel contend this amounts to 50% to 90% of the class’s claimed aggregate single damages of $500 to $900 million, which represents the highest percentage recovery of any settlement of a generic delay pharmaceutical antitrust action (Barnes Class Counsel Decl. ¶ 107, Dkt. No. 682). While this order does not contest counsel’s statements, the percentage of recovery depends on how you look at it. If plaintiffs had prevailed at trial, they would have been awarded treble damages rather than single damages. 15 U.S.C. § 15(a). The class’s estimated aggregate treble damages range from $1.5 to $2.7 billion, so the percentage of recovery could be correctly deemed to range from 17% to 30%. With this qualification in mind, we consider each individual settlement amount in turn using the class’s single damages statistics. Bausch has agreed to a settlement payment of $300 million. According to Dr. Jeffrey Leitzinger, the class’s economics and damages experts, the $300 million recovery amounts to between 33% and 60% of the class’s total aggregate single damages, which equates to between 57.5% and 103% of the claimed aggregate single damages attributable to purchases from Bausch (Leitzinger Decl. Exh. 3; Dkt. No. 643 at 2). The amount offered by Bausch in settlement supports approval. Lupin’s settlement of $150 million amounts to between 16.7% and 30% of the class’s claimed aggregate single damages and between 39.7% and 71% of their claimed single damages attributable to purchases from Lupin (Dkt. No. 657 at 2). These figures are lower than those of the Bausch settlement, but still represent a satisfactory level of return to the class Analysis of the $3.85 million settlement payment from Assertio poses different issues. The class admits the settlement reflects only 0.43% to 0.77% of the class’s claimed aggregate single damages. Acknowledging that this recovery is “small,” the class contends the number reflects the reality of Assertio’s rocky financial circumstances (Dkt. No. 654 at 8–13). To support the payment, the class enlisted its expert Dr. Mark L. Frigo, who concluded: Assertio Therapeutics would not be able to survive a significant adverse financial event, such as an adverse judgment arising from one of the legal claims currently pending against it. I am further of the opinion that, based on Assertio Therapeutics current position and recent performance trends as a stand-alone entity, the proposed class settlement . . . is at the high end of the range of payments it can afford while still meeting its other obligations (Frigo Decl. ¶ 21, Dkt. No. 654-11). Dr. Frigo reviewed Assertio’s problematic cash position, working capital position, accounts receivable, and the 10-K of Assertio’s parent company, Assertio Holdings, Inc. The class also notes that Assertio Holdings’ recent 10-Q contains several serious risk factors related to ongoing opioid litigation and our litigation. To bolster this analysis at final approval, the class retained another expert, financial valuation expert Michael Atkinson, who reached a similar conclusion as Dr. Frigo: “[Assertio] Therapeutics faces current and future significant operational and financial challenges that hinder its ability to afford a settlement amount greater than the agreed upon $3.85 million” (Atkinson Decl. ¶ 12). In addition, Mr. Atkinson reviewed a $55 million intercompany transfer from (named defendant) Assertio Therapeutics to (non-defendant) Assertio Holdings to determine whether, in class counsel’s words, “it was a legitimate transaction” (Hrg. Tr. 8–9). After reviewing the specifics of the transaction, Mr. Atkinson concluded that “the promissory note did not negatively impact the book value of Therapeutics” (Atkinson Decl. ¶¶ 37–40). Upon review, this order accepts this analysis. “It is well-settled law that a cash settlement amounting to only a fraction of the potential recovery does not per se render the settlement inadequate or unfair.” Officers for Justice v. Civil Serv. Comm., 688 F.2d 615, 628 (9th Cir. 1982). Our court of appeals has also held that a defendant’s financial condition can be considered, can indeed predominate the analysis, when recovery. See Torrisi v. Tuscon Elec. Power Co., 8 F.3d 1370, 1376 (9th Cir. 1993); Linney v. Cellular Ak. P’ship, 151 F.3d 1234, 1242 (9th Cir. 1998). The class has demonstrated that the Assertio settlement payment qualifies as adequate given Assertio’s problematic financials. Fifth, the stage of the proceedings supports settlement. The parties had almost run the gamut by the time they agreed to settle. Bausch settled in late July. Assertio and Lupin settled in mid-September mere weeks from trial, scheduled for the first week of October. As instructed, settlement negotiations did not begin until after class certification. The parties also had the benefit of months of settlement discussions under the guidance of Judge Ryu. At the time of settlement, the parties had engaged in extensive discovery and briefed motions to dismiss, a motion for class certification, and motions for summary judgment. Lupin, the last defendant standing, settled after most of the motions in limine were decided at the final pre- trial conference. Sixth, the abilities and views of counsel support settlement. Class counsel have extensive experience in pharmaceutical antitrust class actions and have handled reverse payment cases previously (Dkt. Nos. 61-2; 247-1 at 17–18). As explained in more detail below, this order finds class counsel have zealously litigated against competent defense counsel and approached settlement negotiations with the same ardor. Seventh, while there was no governmental participant in this action, CAFA requires notice of a settlement to be given to the Department of Justice. The DOJ has not raised any issues with any of the three settlements. This factor moderately favors settlement. Eighth, no class member has objected to the proposed settlements, although, as will be discussed, three members objected to class counsel’s motion for attorney’s fees. This factor moderately supports settlement. In sum, the Churchill factors support final approval of the class’s settlements with Bausch, Assertio, and Lupin. 2. RULE 23 AND THIS COURT’S SETTLEMENT FACTORS. Consideration of the Churchill factors alone is not enough. Kim, 8 F.4th at 1179. This 1 adequate under Rule 23(e)(2), as well as this Court’s notice regarding factors to be evaluated 2 for any proposed class settlement (Dkt. No. 39), to the extent not considered in the Churchill 3 analysis above. 4 First, recovery of 17% of the Class’s aggregate treble damages — the low end of 5 counsel’s estimate accounting for the treble damages available at trial — represents an 6 adequate cost-benefit analysis for absent class members given the risks discussed above. The 7 combined settlement, as far as this order is aware, represents the third-largest absolute dollar 8 recovery for a generic delay pharmaceutical antitrust case (Dkt. No. 681 App’x A; Winick 9 Decl. Exh. 1). In addition, the parties have included the requisite “most-favored nation” clause 10 to protect the class: 11 It is the intention of the parties to this Settlement Agreement that the recovery by the Direct Purchaser Class, in proportion to units 12 purchased during the period May 2012 to November 2019, be at least as favorable, if not more favorable, on a pro rata basis than 13 any settlement by either (a) the Retailer Plaintiffs or (b) Humana

v 14 (Bausch Settlement 4 6(b), Dkt. No. 643-2; Assertio Settlement 4 6(b), Dkt. No. 654-2; Lupin

15 Settlement 4 6(b), Dkt. No. 657-2). To that end, for purposes of the clause, the parties agreed Q 16 to a baseline based on the total amount of units purchased from May 2012 to November 2019:

Plaintiff Units Purchased 17 8 Direct Purchaser Class (Excluding Retailer 73,810,441,704 mg 18 Plaintiffs and Humana) 19 Retailer Plaintiffs 20 21 (Ubid.). The parties’ most-favored nation provision sufficiently protects the class and serves as 22 an effective stopgap measure for absent class members. 23 Second, the plan of allocation is fair and reasonable, and the class members are treated 24 equitably relevant to each other (Allocation Plan, Dkt. No. 643-7). The settlement agreements 25 provide that plaintiffs will calculate each class members pro rata share of the settlement based 26 on net units purchased of brand and generic Glumetza during the period May 6, 2012, through 27 August 15, 2020. Class Counsel’s expert Dr. Leitzinger will calculate each claimant’s 28 percentage share of the net settlement fund using transactional sales data for brand and generic

Glumetza, supplemented with IQVIA data as necessary. Dr. Leitzinger has determined that the per-unit dollar overcharge is roughly equivalent for brand and generic Glumetza over the class period, which avoids the need for class members to determine the proportion of generic to brand Glumetza they purchased (Leitzinger Decl. ¶ 6 n.9, Dkt. No. 643-8). Third, the claim procedure for class members is appropriate given the complexity of the action and amount of money involved. Each class member receives a claim form with class counsel’s calculation of its pro rata share of the settlement. The form requests each class member to verify the accuracy of the information and provides instructions for challenging the figures. Each claimant must execute and return the claim form to receive any distribution. The settlement administrator will follow up if any class member does not timely return the form (Allocation Plan ¶ 1.4). The claims administrator is not simply cutting checks, but this approach permits each class member to verify its payment against its own records, acceptable given the sophistication of each class member. The notice, moreover, is straightforward and comprehensible. Fourth, the release of claims against Bausch, Assertio, and Lupin have a permissible scope. The settlements release defendants from claims related to claims under Sections 1 and 2 of the Sherman Antitrust Act arising from or relating to the purchase of brand or generic Glumetza from defendants. The settlements explicitly limit the scope of the release: “For the avoidance of doubt, the scope of Released Claims will be interpreted to be no broader than that required by ¶ 4 of the Court’s November 4, 2019 Notice and Order on Putative Class Actions and Factors to Be Evaluated for Any Proposed Settlement (ECF No. 39)” (Bausch Settlement ¶ 9; Lupin Settlement ¶ 8; Assertio Settlement ¶ 9). The parties have explicitly limited the release to a permissible scope. Fifth, a class-expansion issue arose here, but was resolved satisfactorily. “The class must be notified of a proposed settlement in a manner that does not systematically leave any group without notice.” Officers for Justice v. Civil Serv. Comm’n of City & Cty. of San Francisco, 688 F.2d 615, 624 (9th Cir. 1982). The claims administrator began disseminating the class before the final approval hearing, class counsel acknowledged that they had recently identified three additional class members that had not received notice of the class settlements: Healthsource Distributors, LLC; NDC Distributors LLC; and Republic Pharmaceuticals. Retracing their steps, class counsel explained how they received updated transaction data from Bausch in September 2021 (covering the period between December 2019 and August 2020) for use in settlement allocation planning. In late December 2021, class counsel recognized that the updated information contained details on three class members that had not received notice of the impending settlement. They accordingly sent out notices and followed up several times via email, telephone, and (for two class members) in person (Dkt. No. 688). At a status conference held on the notice issue on January 12, 2022, class counsel were given one week to acquire sworn declarations from the three class members regarding notice detailing whether they would opt out of the class, object to the settlement, or object the motion for attorney’s fees. On January 18, class counsel filed the necessary declarations. All three class members stated that they would not opt of the class and that they did not object to the settlement or the request attorney’s fees (Dkt. No. 699). The class and counsel’s efforts to rectify this issue support settlement. Sixth, the remaining factors for consideration all support settlement. As discussed, the class has adequately demonstrated that Assertio has minimal assets and that a reduction of the settlement amount was warranted. The request for attorney’s fees was left to the judge’s discretion and there was no tentative agreement to any division between the parties. Relatedly, the terms and timing of payment of attorney’s fees, as discussed in detail below, are permissible. This order finds, as explained and further considered below, class counsel and the class representatives adequately represented the class. The class representatives have also not requested an incentive payment, and there is no reversion provision in the settlement agreements. Lastly, this order finds no evidence of collusion or conflicts of interest. As noted, the settlements were all negotiated with active involvement of Judge Ryu and appear to be the result of arms’ length negotiations. In sum, this Court’s factors as well as the factors listed in Rule 23 support settlement. Final approval of the settlement is GRANTED. 3. MOTION FOR ATTORNEY’S FEES AND COSTS. A. LODESTAR OR PERCENTAGE METHOD FOR CALCULATING ATTORNEY’S FEES? We turn now to class counsel’s motion for attorney’s fees. Class counsel request the benchmark attorney’s fee award of 25% of the settlement fund, or $112,817,643.00, plus any interest. This equates to a 4.99 multiplier on counsel’s lodestar of $22,586,167. McKesson Corporation, Cardinal Health, Inc., and AmerisourceBergen Drug Corporation (hereinafter “objectors”) are national wholesalers and the three largest class members in this case who likely represent over 90% of the subject purchases. They oppose class counsel’s motion, arguing an award of the requested size would constitute a windfall for counsel, and recommend instead a multiplier of 2 for counsel’s lodestar. A district court must ensure that attorney’s fees awarded in a class action settlement are reasonable. Staton v. Boeing Co., 327 F.3d 938, 963–64 (9th Cir. 2003). “Because the relationship between class counsel and class members turns adversarial at the fee-setting stage, district courts assume a fiduciary role that requires close scrutiny of class counsel’s requests for fees and expenses from the common fund.” In re Optical Disk Drive Prods. Antitrust Litig., 959 F.3d 922, 930 (9th Cir. 2020). In common-fund cases, where the settlement or award creates a large fund for distribution to the class, the district court has discretion to use either a lodestar or percentage method for awarding attorney’s fees. Staton, 327 F.3d at 968. The lodestar method multiples the number of hours reasonably expended on the litigation (as supported by adequate documentation) by the reasonable hourly rate for the region and experience of the lawyer. Bluetooth, 654 F.3d at 941. Though the lodestar figure is presumptively reasonable, the district court may adjust the lodestar figure upward or downward to reflect a host of reasonableness factors, including: (1) the quality of the representation; (2) the benefit obtained for the class; (3) the complexity and obtained for the class remains the foremost consideration when adjusting the lodestar. Id. at 941–42; Hensley v. Eckerhart, 461 U.S. 424, 434–36 (1983). The percentage-of-recovery method awards a portion of the common fund as attorney’s fees. Our court of appeals has recognized 25% as the benchmark award under this method. Similar to the lodestar analysis, upward or downward departures from the benchmark require analysis of a set of factors that measure the reasonableness of the fee award and the benefit obtained for the class. See Vizcaino v. Microsoft Corp., 290 F.3d 1043, 1048–52 (9th Cir. 2002); Hanlon v. Chrysler Corp., 150 F.3d 1011, 1029 (9th Cir. 1998). Under either method, our court of appeals has “encouraged courts to guard against an unreasonable result by cross- checking their calculations against a second method.” Bluetooth, 654 F.3d at 944. This order employs the lodestar method for three reasons. First, this is a megafund action (a fund larger than $100 million), where the 25% benchmark is “of little assistance.” Optical Disk Drive, 959 F.3d at 931. Empirical data on class-action settlements suggests for funds north of $175.5 million the average for attorney’s fees awards was 12%. A second study calculated that the average award when a fund was over $72.5 million was 18.4%. See William B. Rubenstein, Newberg on Class Actions § 15:81 (2021 ed.) (citing studies); see also In re High-Tech Employee Antitrust Litig., No. C 11-02509 LHK, 2015 WL 5158730, at *13 (N.D. Cal. Sept. 2, 2015) (Judge Lucy H. Koh) (median award for megafund actions 10.2%). Moreover, the pertinent direct purchaser generic delay pharmaceutical antitrust settlements, discussed below, have fairly divergent percentage awards. The variance of these percentages from the benchmark indicate that the 25% marker provides limited insight with which to make an appropriate attorney’s fees award here. Second, without the benefit of a sturdy percentage benchmark, this order will necessarily rely on the lodestar cross-check for some sort of foundation. Trying to “craft an appropriate percentage for this megafund case without first reviewing counsel’s actual timesheets would likely result in ‘picking a number out of the air.’” Gutierrez v. Wells Fargo Bank, N.A., No. C 07-05923 WHA, 2015 WL 2438274, at *5 (N.D. Cal. May 21, 2015) (Judge William Alsup) (quoting In re Washington Public Power Supply System Sec. Litig., 19 F.3d 1291, 1297, 1303 (9th Cir.1994)). Third, objectors assert that use of the benchmark would result in class counsel enjoying windfall profits. Indeed, the 25% award translates into a blended hourly rate of $3,269 per hour for every one of class counsel’s timekeepers, regardless of whether they are a partner, paralegal, or summer associate (Opp. 9, Dkt. No. 686). Where awarding “25% of a ‘megafund’ would yield windfall profits for class counsel in light of the hours spent on the case, courts should adjust the benchmark percentage or employ the lodestar method instead.” Bluetooth, 654 F.3d at 942. After considering all of the circumstances, and given the risk of a windfall, this order finds using the lodestar method with a percentage cross-check will achieve the fairest and most reasonable result. B. CALCULATING THE LODESTAR: ATTORNEY BILLING RATES AND HOURS. This order proceeds to review class counsel’s lodestar in this action. As an initial matter, objectors do not flag any concerns with counsel’s lodestar, acknowledging: “As to the reasonableness of class counsel’s lodestar, the National Wholesalers have little to say” (Opp. 8). Class counsel’s billing rates, although on the high end in some respects, qualify as reasonable. Our class is represented by seven different firms: (1) Hagens Berman Sobol Shaprio LLP; (2) Sperling & Slater P.C.; (3) Hilliard & Shadowen LLP; (4) Taus, Cebulash & Landau, LLP; (5) Roberts Law Firm; (6) Tadler Law LLP; and (7) Frank LLP. For brevity, and because the various rates are similar across firms, this order considers the attorneys from all seven firms jointly. For the thirty-two partners who billed for this action, the hourly rate ranged from Brian Morrison ($525) to Paul Slater ($1,105). The three of counsel involved in this action billed from James Nicklaus ($650) to Jeffrey Bergman ($725). For the twelve associates, the rate ranged from Charles Goulding ($300) to Tina Miranda ($750). One summary associate’s billing was included: Matthew Simkovits ($275). Next, two staff attorneys participated herein: Lloyd Donders ($250) and Rachel Downey ($400). The hourly rate for the one case manager involved, Jason Joseph, was $425. Eleven interns participated, with their hourly rate ranging from Payton Renfroe ($40) to Sylvia Krakauer ($350). Seventeen paralegals assisted in this matter, with their billing rate ranging from Deirdre Mulligan ($150) to Ed Notargiacomo ($350) (Barnes HBSS Decl. Exh. A, Dkt. No. 682-1; Vanek Decl. Exh. A; Shadowen Decl. Exh. A; Taus Decl. Exh. A; Roberts Decl. Exh. A; Bartolomeo Decl. Exh. A; Frank Decl. Exh. A). Upon review, this order will not reduce the billing rate for any of the seventy-nine individuals who participated in this matter. Most of the claimed rates correspond with the going rate for counsel in our geographic region with the same levels of skill and experience (Barnes Class Counsel Decl. ¶ 117). See, e.g., In re Lidoderm Antitrust Litig., No. MD 14- 02521 WHO, 2018 WL 4620695, at *2 (N.D. Cal. Sept. 20, 2018) (Judge William H. Orrick). While some rates appear above market, such as Tadler Law’s summer associate rate, they have little impact on the final lodestar amount. We next consider the reasonableness of billing entries submitted by counsel. Counsel billed a total of 34,503.6 hours to this action, detailed by hundreds of pages of time records (Barnes Class Counsel Decl. ¶ 111, App’x A). Counsel used a taxonomy of 24 categories (with 100 subcategories) to classify the work undertaken on behalf of the class. The individual time entries are sufficiently detailed. Because seven different firms represented the class, a high risk exists of duplicative and excessive entries. To combat this, counsel removed 1,904.8 hours, or $936,795.75 from the cumulative lodestar (id. at ¶ 116). In addition, counsel removed 16 timekeepers with under ten hours in the case and all billing incurred after September 30, 2021, on settlement efforts, and a on a handful of other tasks (id. at ¶¶ 114–15). Upon review, this order is satisfied that counsel have reasonably recorded their billable hours for this matter and that no reduction is necessary to protect the class. Again, objectors have not flagged any particular issue despite access to detailed records. Upon a summary incomprehensible. This order also finds that the timesheets do not clearly indicate significant overstaffing, duplication, inefficiency, or too many teleconferences. See Gutierrez, 2015 WL 2438274, at *6. Class counsel’s lodestar ranks as reasonable. C. LODESTAR MULTIPLIER AND PERCENTAGE CROSS-CHECK. This order turns to whether a multiplier should be applied to class counsel’s lodestar and proceeds to performs a percentage cross-check of that multiplier. Class counsel request $112,817,643, which equates to a lodestar multiplier of 4.99. Objectors argue that such a multiplier is too high and results in a windfall for counsel. This order finds a multiplier of 2.2 reasonable given the circumstances. To begin, the reasonableness factors support a positive multiplier. First, counsel provided strong representation for the class. Class counsel discovered and developed this case without the benefit of a government investigation’s coattails. In total, class counsel reviewed 578,790 documents, deposed 19 fact and opposing-expert witnesses, and consulted with and retained 10 expert witnesses of their own (Barnes Class Counsel Decl. ¶¶ 32, 55, 62). Class counsel also successfully defeated defendants’ motions to dismiss, certified a Rule 23 class, and defeated defendants’ summary judgment motions prior to reaching an agreement with all three defendants to settle this action mere weeks before the trial date (id. at ¶¶ 27, 39–42, 76– 82, 83–95). Class counsel accomplished all this despite vigorous opposition from large multi- national companies with high-quality representation from six national law firms. See In re Apple Inc. Device Performance Litig., No. MD 18-02827, 2021 WL 1022866, at *6 (N.D. Cal. Mar. 17, 2021) (Judge Edward J. Davila). Moreover, the COVID-19 pandemic emerged right as fact-witness depositions were set to begin, effectively forcing counsel to figure out how to litigate an antitrust class action remotely via Zoom (Barnes Class Decl. ¶¶ 36–37). Second, the benefit obtained for the class appears to be one of the stronger in this genus of class-action litigation. As noted, the final settlement amount of $453.85 million represents, before trebling, 50% to 90% of the direct-purchaser class’s claimed aggregate single damages. Counsel assert that this ranks as the highest percentage recovery of any generic delay Counsel Decl. ¶ 107). Moreover, on a pro rata gross basis, class members’ recovery is about 1.9 to 2.4 times larger than that of the retailer opt-outs (Reply Br. 10). Third, as explained, the issues in this litigation qualify as complex. Fourth, this action was litigated purely on a contingency fee basis, so the risk of nonpayment was ever present. Class counsel had no upfront retainer fees or allowance for expenses. In fact, class counsel invested nearly $2.5 million in out-of-pocket costs (Barnes Class Counsel Decl. ¶ 129–31). This supports a multiplier. Cf. Vizcaino, 290 F.3d at 1051. With a multiplier warranted, this order must next determine a reasonable range for the multiplier to be applied, so we look first to class counsel and objectors’ collected data on past settlements of direct purchaser generic delay pharmaceutical antitrust cases (Br. App’x A; Winick Decl. Exh. A). This order agrees with class counsel that these two datasets present 35 previous actions from which to establish a baseline (Reply Br. 11 n.44). But many of these previous settlements, while informative, have minimal relevance to this settlement. Non-megafund settlements, i.e., settlement funds of less than $100 million, reflect a much more direct connection between the efforts of counsel and the ultimate recovery. In other words, due to economies of scale, it isn’t ten times harder to try a $100 million case as it is a $10 million case. See Optical Disk Drive, 959 F.3d at 933. Non-megafund settlements would hence yield little insight on a reasonable range for a multiplier here. In addition, our action was filed, tried, and settled with the benefit of the Supreme Court’s 2013 Actavis decision. Actavis clarified the legal landscape of generic delay antitrust actions and charted a more favorable path for future plaintiffs. Actions filed prior to Actavis ultimately provide stale insight into the proper range of attorney’s fees now. Three cases fit these parameters of actions filed post-Actavis with megafund settlements: Case Settlement Fees Lodestar Multiplier Fees Amount Requested Awarded In re Aggrenox Antitrust $146mm $48,600,000 $19,633,232 1.49 $29,200,000 Litig., No. MD 14-02516 SRU, 2017 WL 11636126 (D. Conn. Dec. 19, 2017) (Judge In re Lidoderm Antitrust $166mm $45,000,000 $42,671,874 1.05 $45,000,000 Litig., No. MD 14-02521 WHO, 2018 WL 7075881 (N.D. Cal. Sept. 20, 2018) (Judge William H. Orrick) In re Namenda Direct $750mm $157,500,000 $34,769,008 2 $69,538,017 Purchaser Antitrust Litig., No. C 15-07488 CM, 2020 WL 3170586 (S.D.N.Y. June 15, 2020) (Judge Colleen McMahon) These three settlements have multipliers that range from 1.05 to 2. Of these three, only Namenda’s settlement amount is in the same ballpark as our fund.1 Class counsel request a multiplier of 4.99, significantly higher than the multipliers previously granted in the pertinent generic delay antitrust actions. Three reasons justify a smaller multiplier than requested. First, the amount of risk undertaken by counsel does not warrant such a large multiplier. Despite the fact that counsel undertook this litigation on a purely contingent basis, the risk of non-payment was spread out over seven different law firms, ensuring that no one firm would take too big a hit upon an adverse ruling. And, as previously discussed, unlike other reverse payment antitrust actions with larger multipliers, counsel initiated this action nearly six years after the Actavis decision. Prior to Actavis, the law for reverse payment antitrust actions was not as favorable to plaintiffs and the lawyers litigating those actions took on greater risk. This distinguishes, for example, King Drug Co. of Florence, Inc. v. Cephalon, Inc. (Provigil), No. C 06-01797, 2015 WL 12843830 (E.D. Pa. Oct. 15, 2015), which granted an attorney’s fees award of $140.8 million, equating to a multiplier of 4.12 on counsel’s $34,133,958 lodestar. Among other notable factors distinguishing Provigil, such as the national wholesalers endorsing rather than objecting to the requested fees, counsel therein litigated for nine years

1 This order notes In re Loestrin 24 Fe Antitrust Litig., No. MD 13-02472 WES, 2020 WL 5203323 (D.R.I. Sept. 1, 2020), was consolidated shortly after Actavis, but several of the related cases were filed prior to the Supreme Court’s decision. Loestrin settled for $120 million, and prior to settlement. More than seven of those years took place before Actavis. The greater the risk, the greater the justification for a high multiplier. Moreover, from the impetus of this action, counsel had two, dramatic facts weighing in their favor that limited their risk: the concealment of the no-authorized-generic payment from Judge Hamilton in the underlying patent litigation, and Valeant’s 800% price hike of Glumetza in the summer of 2015. Both powerful facts would have likely proved persuasive in front of a jury. And as objectors state, the price hike was a basis for the large size of the settlement (Opp. 6). Class counsel disagree, citing cases with estimated aggregate overcharges similar to those at issue here, but with smaller final settlement amounts (Reply Br. 19–20). These stats are unpersuasive. The price hike unquestionably raised the ceiling of potential damages at stake in this action. This would affect the size of the settlement and limit class counsel’s risk. (This order does agree with class counsel the price hike alone did not dictate that the outcome of settlement negotiations, as objectors seem to contend.) Because the size of the recovery is “not entirely attributable to class counsel’s skill,” a smaller multiplier than requested is warranted. Gutierrez, 2015 WL 2438274, at 5. Second, several of class counsel’s justifications for their requested attorney’s fees were mandated by the Court. Counsel contend that they have ensured that “the class members are receiving far better recovery than those who litigated outside of the class” (Hrg. Tr. 37). But counsel merely complied with instructions put forth in the final pre-trial conference: “I have one other thing to say about settlements. I will not approve the class settlement unless there is a provision in there that says that if these defendants settle for more with somebody else, the class gets the same [or] higher amount” (Sept. 8, 2021 Hrg. Tr. 9). In fact, the first settlement agreement with Bausch was denied subject to reconsideration until a most-favored nation clause was inserted into the agreement (Sept. 9, 2021 Hrg. Tr. 4). Counsel deserve credit for finalizing settlements that provided recovery for class members beyond mere parity with the opt-outs, but do not deserve credit for incorporating an explicit requirement for settlement approval. Additionally, counsel takes credit for getting this case to resolution in two years, result in the class members getting paid that much faster than they typically receive any payment in other cases” (Hrg. Tr. 36–37). But counsel did not set the trial schedule for this litigation, the Court did. These types of justifications fail to persuade that such a high multiplier is justified in this action. Third, the national wholesalers’ objection based on Namenda and economies of scale support a smaller multiplier. Unlike almost all other settlements class counsel cite, three class members — in fact, the three largest class members — have objected to counsel’s motion for attorney’s fees, labelling it an unreasonable windfall (Opp. 1–3). The only other reverse payment antitrust action where the national wholesalers objected was Namenda. In Namenda, objectors did withdraw their objection once class counsel voluntarily reduced their multiplier to 4.5. Nevertheless, the final multiplier awarded was further reduced to 2.0, equating to an award of $69.5 million. 2020 WL 3170586, at *2, *6. Class counsel argue that the national wholesalers’ withdrawal of their objection and affirmative support of a 4.5 multiplier in Namenda negates their objection here (Reply Br. 16). Not so. Objectors will not be held to a position taken in a previous, unrelated action. Determining the reasonableness of an attorney’s fees request is an individualized endeavor. Class counsel also take pains to distinguish Namenda on several other grounds (Br. 16–18; Reply Br. 6). Despite those differences, such as duplicative billing and governmental involvement, it remains a proper point of comparison. Not only is Namenda the most recent generic delay pharmaceutical antitrust class settlement, the size of its settlement is also of similar magnitude as the instant common fund. Moreover, class counsel therein also had to prepare for trial prior to settlement and litigated with the benefit of Actavis. Namenda in part justified its downward departure from the requested 4.5 multiplier by explaining: This Court does not accept class counsel’s argument that an unprecedentedly large settlement warrants a commensurately unprecedentedly large fee award. . . . Where class counsel reaches a settlement for hundreds of millions of dollars, a four-or-five times multiplier strikes me as both unseemly and unnecessary to protect the interests of the class. 2020 WL 3170586, at *5. A primary rationale for the class action procedure is leverage of economies of scale. This remains true when a class action leads to a megafund settlement, where efficiencies in legal services should be passed on not just to counsel, but to the class members, themselves. See Newberg on Class Actions § 1:7. To account for economies of scale and prevent windfalls for class counsel, district courts consider the size of the fund for motions for attorney’s fees, although our court of appeals has not required sliding-scale fee awards. See Vizcaino, 290 F.3d at 1047–48. This order finds a multiplier of 4.99 gives no credence to economies of scale present in this settlement and would unfairly award class counsel a financial windfall. Class counsel argue that the opt-outs here are likely paying their lawyers on a contingency fee basis in the 25% to 30% range, and that class counsel “should not be disadvantaged by comparison” (Hrg. Tr. 16–17). The ultimate truth of counsel’s contention on the fee arrangements that our opt-outs have agreed to with their lawyers will remain unknown, but class counsel’s argument misses the mark. Those fee agreements did not involve a class of plaintiffs certified under Rule 23. For a Rule 23 class, the Court must act as fiduciary, and a 4.99 multiplier is unnecessary to protect the interests of the class and would inequitably pass on benefits intrinsic to the class action procedure to counsel. In sum, considering all of the facts and circumstances and the current state of the law, this order GRANTS a MULTIPLIER OF 2.2 and AWARDS $49,689,567.40 to counsel, a figure more than adequate to compensate and reward counsel for obtaining a positive result for the class. This award constitutes the second highest amount of attorney’s fees granted in a generic delay antitrust action filed post-Actavis. More than a 2.2 multiplier would risk giving class counsel an unnecessary windfall and cut too deeply into the class’s award. Turning to the percentage cross-check, this fee award equates to 11% of the $453.85 million settlement. This is also reasonable. As explained, “special circumstances” justify a downward departure from the 25% benchmark, especially because the benchmark is of little assistance for a megafund settlement. Bluetooth, 654 F.3d at 942. In particular, collected empirical data on megafund settlements support the award. Again, one study on megafunds demonstrated the mean average return for attorney’s fees was 12% for funds greater than $175.5 million. See Newberg on Class Actions § 15:81. Vizcaino listed with approval twenty-four settlement-percentage awards that ranged between 2.8% to 40%. See 290 F.3d at 1052; see also Gutierrez, 2015 WL 2438274, at *8 (listing actions with percentage awards of e.g., 1.4%, 7%, 11.48%); High-Tech Employee, 2015 WL 5158730, at *11–13. Our percentage award sits comfortably within this range. The 11% award also comports with the three post-Actavis megafund generic delay pharmaceutical antitrust cases listed above: Aggrenox (20%); Lidoderm (27.5%); and Namenda (9.30%). Class counsel seek $2,413,688.60 in expenses for, among other things, costs associated with experts, hosting charges for a discovery document database, filing fees, copying, postage, etc. (Barnes Class Counsel Decl. ¶¶ 119–32). Understandably, the grand majority ($2,010,039.34, prior to reductions) went to expert and consulting fees. Counsel retained an independent accounting firm, Robert A Zagrodny CPA, Inc., to track and review expenses, at no cost to the class. COVID-19 grounded counsel during discovery, and, appropriately, counsel seek no travel or hotel reimbursements. The expenses are reasonable. To the extent stated, final approval of the direct-purchaser class’s settlement agreements with Bausch, Assertio, and Lupin are GRANTED and the settlements shall be consummated in accordance with their terms. All of the direct-purchaser class’s claims against Bausch, Assertio, and Lupin are hereby dismissed with prejudice and without costs as provided in the respective settlement agreements. The class’s request that this order find the notice of settlement adequate and satisfied due process is GRANTED. The class’s further request that the Court retain exclusive jurisdiction over the settlements and the settlement agreements as described therein, including the administration and consummation of the settlements and over this order and final judgment is GRANTED. The class’s further request for final approval of the 1 authorization for lead class counsel and Angeion Group to begin administration and 2 distribution of claim forms and the net proceeds of each settlement in accordance with the plan 3 of allocation is GRANTED. 4 Class counsel is AWARDED attorney’s fees in the amount of $49,689,567.40, plus any 5 interest on that amount that may accrue before distribution. Fifty percent of the total attorney’s 6 fees award shall be distributed to class counsel as of the effective date as defined in the 7 settlements. The remaining fifty percent shall be paid when defendants certify that all funds 8 have been properly distributed and the file can be completely closed. Class counsel is 9 AWARDED expenses in the amount of $2,413,688.60. Further costs and expenses under 10 $25,000 that may be incurred in connection with the settlement administration process may be 11 disbursed from the settlement fund without further application to the Court. The direct- 12 purchaser class’s further request that this order deem costs and attorney’s fees not recoverable 13 under 15 U.S.C. § 15(a) is GRANTED. 14 IT IS SO ORDERED. 16

= 17 Dated: February 3, 2022. 18 Pree 19 { A po WILLIAM ALSUP 20 UNITED STATES DISTRICT JUDGE 21 22 23 24 25 26 27 28