In Re Hydrogen Peroxide Antitrust Litigation

OPINION OF THE COURT

At issue in this antitrust action are the standards a district court applies when deciding whether to certify a class. We will vacate the order certifying the class in this case and remand for proceedings consistent with this opinion.

In deciding whether to certify a class under Fed.R.Civ.P. 23, the district court must make whatever factual and legal inquiries are necessary and must consider all relevant evidence and arguments presented by the parties. See Newton v. Merrill Lynch, Pierce, Fenner & Smith, Inc., 259 F.3d 154, 166, 167 (3d Cir.2001) (citing Szabo v. Bridgeport Machs., Inc., 249 F.3d 672, 676 (7th Cir.2001); Manual for Complex Litigation (Third) § 30.1 (1995)). In this appeal, we clarify three key aspects of class certification procedure. First, the decision to certify a class calls for findings by the court, not merely a “threshold showing” by a party, that each requirement of Rule 23 is met. Factual determinations supporting Rule 23 findings must be made by a preponderance of the evidence. Second, the court must resolve all factual or legal disputes relevant to class certification, even if they overlap with the merits — including disputes touching on elements of the cause of action. Third, the court’s obligation to consider all relevant evidence and arguments extends to expert testimony, whether offered by a party seeking class certification or by a party opposing it.

I.

Purchasers of hydrogen peroxide and related chemical products brought this antitrust conspiracy action against chemical manufacturers. 1 An inorganic liquid, hydrogen peroxide is used most prominently as a bleach in the pulp and paper industry with smaller amounts appearing in chemicals and laundry products, environmental applications, textiles, and electronics. Hydrogen peroxide is available in solutions of different concentrations and grades depending on its intended use. Major concentrations are 35, 50, and 70 percent. The grades, roughly in order from least- to most-expensive, are: standard, food/cosmetic (which must meet FDA standards), electronic, and propulsion. All defendants sold the standard grade, but not all defendants sold all other grades. Defendants sold different amounts of each of the grades. Each grade has different supply and demand conditions because the grades are sold to end-users in a variety of industries with different economic characteristics. According to defendants, the differ *308 ent grades are not economic substitutes for each other, but plaintiffs disagree. Prices diverge dramatically among grades; electronic or propulsion grade can be as much as five times more expensive than standard grade.

The other two products at issue are sodium percarbonate and sodium perbo-rate, together known as persalts, which are granular solids containing hydrogen peroxide used primarily as detergents. Among the defendants, only Solvay produced and sold sodium percarbonate in the United States during the class period. Solvay Chemicals, Degussa Corp., and FMC sold sodium perborate in the United States during the class period. Akzo, Arkema, and Kemira did not sell or produce sodium perborate in the United States during the class period.

After the United States Department of Justice and the European Commission began investigating possible violations of the antitrust laws in the hydrogen peroxide industry, 2 several plaintiffs filed class action complaints against producers of hydrogen peroxide and persalts under § 4 of the Clayton Act, 15 U.S.C. § 15, alleging a conspiracy in restraint of trade violating § 1 of the Sherman Act, 15 U.S.C. § 1. The Judicial Panel on Multidistrict Litigation transferred all cognate federal actions to the United States District Court for the Eastern District of Pennsylvania, which consolidated the cases. See In re Hydrogen Peroxide Antitrust Litig., 374 F.Supp.2d 1345 (J.P.M.L.2005). The consolidated amended complaint alleged that during an eleven-year class period (January 1, 1994-January 5, 2005) defendants (1) communicated about prices they would charge, (2) agreed to charge prices at certain levels, (3) exchanged information on prices and sales volume, (4) allocated markets and customers, (5) agreed to reduce production capacity, (6) monitored each other, and (7) sold hydrogen peroxide at agreed prices.

The District Court denied defendants’ motion to dismiss the complaint for failure to state a claim. Following extensive discovery, 3 plaintiffs moved to certify a class of direct purchasers of hydrogen peroxide, sodium perborate, and sodium percarbo-nate, over an eleven-year class period. In support of class certification, plaintiffs offered the opinion of an economist. Defendants, opposing class certification, offered the opinion of a different economist. Defendants separately moved to exclude the opinion of plaintiffs’ economist as unreliable under Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993). Concluding plaintiffs’ expert’s opinion was admissible and supported plaintiffs’ motion for class certification, the District Court certified a class of direct purchasers of hydrogen peroxide, sodium perborate, and sodium per-carbonate under Fed.R.Civ.P. 23(b)(3). See In re Hydrogen Peroxide Antitrust Litig., 240 F.R.D. 163 (E.D.Pa.2007). The District Court identified seven issues to be tried on a class-wide basis: (1) whether defendants and others engaged in a combi *309 nation and conspiracy to fix, raise, maintain, or stabilize prices; allocate customers and markets; or control and restrict output of hydrogen peroxide, sodium perbo-rate, and sodium percarbonate sold in the United States; (2) the identity of the participants in the alleged conspiracy; (3) the duration of the alleged conspiracy and the nature and character of defendants’ acts performed in furtherance of it; (4) the effect of the alleged conspiracy on the prices of hydrogen peroxide and persalts during the class period; (5) whether the alleged conspiracy violated the Sherman Act; (6) whether the activities alleged in furtherance of the conspiracy or their effect on the prices of hydrogen peroxide and persalts during the class period injured named plaintiffs and the other members of the class; and (7) the proper means of calculating and distributing damages. The class was defined as:

All persons or entities, including state, local and municipal government entities (but excluding defendants, their parents, predecessors, successors, subsidiaries, and affiliates as well as federal government entities) who purchased hydrogen peroxide, sodium perborate, or sodium percarbonate in the United States, its territories, or possessions, or from a facility located in the United States, its territories, or possessions, directly from any of the defendants, or from any of their parents, predecessors, successors, subsidiaries, or affiliates, at any time during the period from September 14, 1994 to January 5, 2005.

We granted defendants’ petition for an interlocutory appeal under Fed.R.Civ.P. 23(f). 4

II.

Class certification is proper only “if the trial court is satisfied, after a rigorous analysis, that the prerequisites” of Rule 23 are met. 5 Gen. Tel. Co. of Sw. v. Falcon, 457 U.S. 147, 161, 102 S.Ct. 2364, 72 L.Ed.2d 740 (1982); see Beck v. Maximus, Inc., 457 F.3d 291, 297 (3d Cir.2006); see also Amchem Prods., Inc. v. Windsor, 521 U.S. 591, 615, 117 S.Ct. 2231, 138 L.Ed.2d 689 (1997) (Rule 23(b)(3) requirements demand a “close look”). “A class certification decision requires a thorough examination of the factual and legal allegations.” Newton, 259 F.3d at 166. 6

*310 The trial court, well-positioned to decide which facts and legal arguments are most important to each Rule 23 requirement, possesses broad discretion to control proceedings and frame issues for consideration under Rule 23. See Amchem, 521 U.S. at 630, 117 S.Ct. 2231 (Breyer, J., concurring in part and dissenting in part) (recognizing that the decision on class certification may implicate “highly fact-based, complex, and difficult matters”). But proper discretion does not soften the rule: a class may not be certified without a finding that each Rule 23 requirement is met. Careful application of Rule 23 accords with the pivotal status of class certification in large-scale litigation, because

denying or granting class certification is often the defining moment in class actions (for it may sound the “death knell” of the litigation on the part of plaintiffs, or create unwarranted pressure to settle nonmeritorious claims on the part of defendants)....

Newton, 259 F.3d at 162; see id. at 167 (“Irrespective of the merits, certification decisions may have a decisive effect on litigation.”); see also Coopers & Lybrand v. Livesay, 437 U.S. 463, 476, 98 S.Ct. 2454, 57 L.Ed.2d 351 (1978). In some cases, class certification “may force a defendant to settle rather than incur the costs of defending a class action and run the risk of potentially ruinous liability.” Fed.R.Civ.P. 23 advisory committee’s note, 1998 Amendments. Accordingly, the potential for unwarranted settlement pressure “is a factor we weigh in our certification calculus.” Newton, 259 F.3d at 168 n. 8. The Supreme Court recently cautioned that certain antitrust class actions may present prime opportunities for plaintiffs to exert pressure upon defendants to settle weak claims. See Bell Atl. Corp. v. Twombly, 550 U.S. 544, 127 S.Ct. 1955, 1967, 167 L.Ed.2d 929 (2007).

III.

Here, the District Court found the Rule 23(a) requirements were met, a determination defendants do not now challenge. Plaintiffs sought certification under Rule 23(b)(3), which is permissible when the court “finds that the questions of law or fact common to class members predominate over any questions affecting only individual members, and that a class action is superior to other available methods for fairly and efficiently adjudicating the controversy.” 7 Fed.R.Civ.P. 23(b)(3). The twin requirements of Rule 23(b)(3) are known as predominance and superiority.

Only the predominance requirement is disputed in this appeal. Predomi *311 nance “tests whether proposed classes are sufficiently cohesive to warrant adjudication by representation,” Amchem, 521 U.S. at 623, 117 S.Ct. 2231, a standard “far more demanding” than the commonality requirement of Rule 23(a), id. at 623-24, 117 S.Ct. 2231, “requiring more than a common claim,” Newton, 259 F.3d at 187. “Issues common to the class must predominate over individual issues.... ” In re Prudential Ins. Co. Am. Sales Practice Litig., 148 F.3d 283, 313-14 (3d Cir.1998). Because the “nature of the evidence that will suffice to resolve a question determines whether the question is common or individual,” Blades v. Monsanto Co., 400 F.3d 562, 566 (8th Cir.2005), “‘a district court must formulate some prediction as to how specific issues will play out in order to determine whether common or individual issues predominate in a given case,’ ” In re New Motor Vehicles Can. Exp. Antitrust Litig., 522 F.3d 6, 20 (1st Cir.2008) [hereinafter New Motor Vehicles ] (quoting Waste Mgmt. Holdings, Inc. v. Mowbray, 208 F.3d 288, 298 (1st Cir.2000)). 8 “If proof of the essential elements of the cause of action requires individual treatment, then class certification is unsuitable.” Newton, 259 F.3d at 172. Accordingly, we examine the elements of plaintiffs’ claim “through the prism” of Rule 23 to determine whether the District Court properly certified the class. Id. at 181.

A.

The elements of plaintiffs’ claim are (1) a violation of the antitrust laws — here, § 1 of the Sherman Act, (2) individual injury resulting from that violation, and (3) measurable damages. 15 U.S.C. § 15; Am. Bearing Co. v. Litton Indus., Inc., 729 F.2d 943, 948 (3d Cir.1984); Blades, 400 F.3d at 566. Importantly, individual injury (also known as antitrust impact) is an element of the cause of action; to prevail on the merits, every class member must prove at least some antitrust impact resulting from the alleged violation. Bogosian v. Gulf Oil Corp., 561 F.2d 434, 454 (3d Cir.1977); see Newton, 259 F.3d at 188 (In antitrust and securities fraud class actions, “[p]roof of injury (whether or not an injury occurred at all) must be distinguished from calculation of damages (which determines the actual value of the injury)”).

In antitrust cases, impact often is critically important for the purpose of evaluating Rule 23(b)(3)’s predominance requirement because it is an element of the claim that may call for individual, as opposed to common, proof. See New Motor Vehicles, 522 F.3d at 20 (“In antitrust class actions, common issues do not predominate if the fact of antitrust violation and the fact of antitrust impact cannot be established through common proof.”); Bell Atl. Corp. v. AT & T Corp., 339 F.3d 294, 302 (5th Cir.2003) (“[W]here fact of damage cannot be established for every class member through proof common to the class, the need to establish antitrust liability for individual class members defeats Rule 23(b)(3) predominance.”); see also Blades, 400 F.3d at 572 (“[P]roof of conspiracy is not proof of common injury.”).

Plaintiffs’ burden at the class certification stage is not to prove the element of antitrust impact, although in order to prevail on the merits each class member must do so. Instead, the task for plaintiffs at class certification is to demonstrate that the element of antitrust impact is capable of proof at trial through evidence that is common to the class rather than individual *312 to its members. Deciding this issue calls for the district court’s rigorous assessment of the available evidence and the method or methods by which plaintiffs propose to use the evidence to prove impact at trial. See Fed.R.Civ.P. 23 advisory committee’s note, 2003 Amendments (“A critical need is to determine how the case will be tried.”); see, e.g., In re Linerboard Antitrust Litig., 305 F.3d 145, 155 (3d Cir.2002) (“rejecting] the contention that plaintiffs did not demonstrate that sufficient proof was available, for use at trial, to prove antitrust impact common to all the members of the class”).

Here, the District Court found the predominance requirement was met because plaintiffs would be able to use common, as opposed to individualized, evidence to prove antitrust impact at trial. On appeal, defendants contend the District Court erred in three principal respects in finding plaintiffs satisfied the predominance requirement: (1) by applying too lenient a standard of proof for class certification, (2) by failing meaningfully to consider the views of defendants’ expert while crediting plaintiffs’ expert, and (3) by erroneously applying presumption of antitrust impact under Bogosian, 561 F.2d at 454-55.

We review a class certification order for abuse of discretion, which occurs if the district court’s decision “rests upon a clearly erroneous finding of fact, an errant conclusion of law or an improper application of law to fact.” Newton, 259 F.3d at 165. “[W]hether an incorrect legal standard has been used is an issue of law to be reviewed de novo.” In re Initial Pub. Offering Sec. Litig., 471 F.3d 24, 32 (2d Cir.2006) [hereinafter IPO ] (citation omitted). 9

B.

We summarize briefly the evidence and arguments offered to the District Court. As noted, both plaintiffs and defendants presented the opinions of expert economists. Importantly, the experts disagreed on the key disputed predominance issue— whether antitrust impact was capable of proof at trial through evidence common to the class, as opposed to individualized evidence.

Plaintiffs’ expert, John C. Beyer, Ph.D., offered an opinion purporting to show that “there is common proof that can be used to demonstrate that the alleged conspiracy to raise prices, restrict output and allocate customers would have impacted all purchasers of hydrogen peroxide, sodium perborate, and sodium percarbonate.” Beyer’s “market analysis” suggested that conditions in the hydrogen peroxide industry favored a conspiracy that would have impacted the entire class. First, hydrogen peroxide and persalts are fungible, undifferentiated commodity products, which means producers compete on price, not quality or other features. Second, production is heavily concentrated in a small group of manufacturers. 10 Third, *313 there are high barriers to entry in the industry and no close economic substitutes, preventing any competitors from entering the market and undercutting prices. Fourth, defendants’ geographic markets overlapped, so that purchasers would have benefited from price competition if not for the alleged conspiracy.

Beyer also observed a “pricing structure” in the hydrogen peroxide industry which, he contended, showed prices across producers, grades and concentrations of hydrogen peroxide, and end uses moved similarly overtime. 11 This, according to Beyer, suggested a conspiracy would have impacted all class members:

My analysis of the similarity in price movements over time indicates that hydrogen peroxide prices charged by different manufacturers are affected by the same market forces of supply and demand. ... These similarities in movement are sometimes referred to as “pricing structure” or “structure to prices.” This analysis confirms that prices would have behaved similarly, in a consistent and generalized manner[,] to a conspiracy to fix prices at artificially high levels [and] to restrict output or to allocate customers.

Beyer also pointed to coordinated increases in list prices by defendants as evidence of common impact.

Beyer identified two “potential approaches” to estimating damages on a class-wide basis: (1) benchmark analysis, which would compare actual prices during the alleged conspiracy with prices that existed before the class period; and (2) regression analysis, through which it “may be possible ... to estimate the relationship between price of hydrogen peroxide, sodium perborate, and sodium percarbonate and the various market forces that influence prices, including demand and supply variables.” These methods, according to Beyer, could be used to estimate the prices plaintiffs would have faced but for the conspiracy. Beyer stated that “sufficient reliable data” exist to allow him to employ one or both of the potential approaches.

Defendants offered the opinion of their own expert economist, Janusz A. Ordover, Ph.D., to “provide an independent expert assessment of whether certification of the proposed class of Plaintiffs is appropriate in this matter.” Specifically, Ordover set out to address “whether, assuming a conspiracy of the kind described in the Complaint, the Plaintiffs will be able to show, through common proof, that all or virtually all of the members of the proposed class suffered economic injury caused by the alleged conspiracy.” Ordover also “opine[d] on whether a formulaic approach exists by which impact could be demonstrated and damages to the class could be reasonably calculated.” Ordover responded to and disputed many of Beyer’s opinions.

First, Ordover disputed Beyer’s finding that hydrogen peroxide and persalts are fungible, contending that the “various grades of hydrogen peroxide ... [and per-salts] have different supply characteristics and face different demand conditions. The existence of supply and demand characteristics that are specific to the various grades and uses requires individualized assessment of the impact of the alleged conspiracy at least across these different grades and uses. Consequently, a finding of class-wide impact from the alleged conspiracy cannot be inferred from the mere fact of the conspiracy and from common *314 evidence.” Second, Ordover alleged that, over the eleven-year proposed class period, “the industry experienced prolonged periods of increasing capacity, increasing production, and an overall trend of declining real and nominal prices in the face of stable or increasing costs.” Ordover disputed Beyer’s pricing structure analysis, contending “there is no tendency for prices charged to individual customers to move together, which indicates that the alleged conspiracy cannot be shown to have had class-wide impact,” necessitating individualized inquiries to determine whether a customer incurred impact.

Ordover also found some of defendants’ price-increase announcements were ineffective- — -actual prices did not follow the purported announcements — suggesting list prices could not be used to measure antitrust impact on a basis common to the class. Ordover observed that a number of contracts for the sale of hydrogen peroxide were individually negotiated, with a variety of contract terms. And deposition testimony from named plaintiffs indicated list prices were sometimes disregarded. Ord-over opined that the statistical methods by which Beyer proposed to demonstrate common impact and damages were not feasible. Given the record of prices and output in the industry and the apparent influence of individualized factors on pricing, “class-wide assessment of impact based on aggregate price information [was] impossible,” "and any formulaic approach to determine a set of “but-for prices” would have to incorporate a multitude of different “variables,” defeating any reasonable notion of proof common to the class.

Significantly, Ordover presented empirical analysis of the data on individual sales transactions and found that different customers purchasing the two most common grades and three most common concentrations from the same hydrogen peroxide producer in a given year were as likely to experience a decline in actual prices over the year as an increase, while other similarly situated customers experienced no change in price. Defendants contend this disparity goes to the core of the predominance issue — plaintiffs and their expert, Beyer, failed to “explain ... how or which common proof could be used to determine that the alleged conspiracy impacted customers whose prices declined, as well as customers whose prices increased or stayed the same, over the same time period.” 12 Br. of Appellant at 5. Beyer, according to defendants, only “promised” to come up with a method to overcome this obstacle, without showing or even suggesting how it might be done. Defendants contend the market analysis is “generic” and note it would apply equally to a large number of industries. With- respect to the pricing structure analysis, they contend Beyer’s use of average prices, rather than those of individual transactions, to show pricing structure, was erroneous because it glossed over differences in actual prices. The theme of defendants’ argument is that the data, which Ordover analyzed, rebut Beyer’s “theory” that common proof was feasible. Beyer’s and Ordover’s analyses are irreconcilable.

In addition to presenting Ordover’s testimony, defendants moved to exclude Beyer’s testimony as unreliable, citing Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 *315 (1993). 13 The District Court denied the Daubert motion in its memorandum and order certifying the class.

C.

The District Court concluded the predominance requirement was met. It held that “[ejither [Beyer’s] market analysis or the pricing structure analysis would likely be independently sufficient at this stage. Plaintiffs and Dr. Beyer have provided us with both. Despite defendants’ claims to the contrary, we should require no more of plaintiffs in a motion for class certification.” Because hydrogen peroxide is fungible, the court found, “purchasing decisions [are] made primarily on the basis of price rather than quality or specific properties,” and “price is by far the most significant means of competition among producers and an agreement to control prices will seriously hinder competition.” The court rejected defendants’ objection that different grades and concentrations of hydrogen peroxide called into question its fungibility. The prices of the grades and concentrations were related to each other, so in the view of the court, the differences would not preclude common proof of antitrust impact. Defendants’ high combined market share meant that “no competitor who was not a member of the conspiracy would be able to take up the slack and keep prices stable.” The high barriers to entry and lack of economic substitutes implied “a conspiracy such as the one alleged here [could] continue indefinitely with limited risk that a new competitor would enter the market and undercut the agreed-upon prices.” Also, the court accepted Beyer’s opinion that “prices in the hydrogen peroxide industry moved similarly over time and the industry exhibited structure in pricing.” The court added that it believed “plaintiffs would be able to show antitrust impact on all purchasers merely by showing that defendants kept list prices that were artificially high because of their conspiracy.”

The District Court held that it was sufficient that Beyer proposed reliable methods for proving impact and damages; it did not matter that Beyer had not completed any benchmark or regression analyses, and the court would not require plaintiffs to show at the certification stage that either method would work.

IV.

A.

Defendants contend the District Court applied too lenient a standard of proof with respect to the Rule 23 requirements by (1) accepting only a “threshold showing” by plaintiffs rather than making its own determination, (2) requiring only that plaintiffs demonstrate their “intention” to prove impact on a class-wide basis, and (3) singling out antitrust actions as appropriate for class treatment even when compliance with Rule 23 is “in doubt.”

Although it is clear that the party seeking certification must convince the district *316 court that the requirements of Rule 23 are met, little guidance is available on the subject of the proper standard of “proof’ for class certification. 14 The Supreme Court has described the inquiry as a “rigorous analysis,” Falcon, 457 U.S. at 161, 102 S.Ct. 2364, and a “close look,” Amchem, 521 U.S. at 615, 117 S.Ct. 2231, but it has elaborated no further.

1.

The following principles guide a district court’s class certification analysis. First, the requirements set out in Rule 23 are not mere pleading rules. Szabo, 249 F.3d at 675-77. The court may “ ‘delve beyond the pleadings to determine whether the requirements for class certification are satisfied.’ ” Newton, 259 F.3d at 167 (quoting 5 James Wm. Moore et al., Moore’s Federal Practice § 23.61[5]); see Beck, 457 F.3d at 297 (same); see also Johnston v. HBO Film Mgmt., Inc., 265 F.3d 178, 189 (3d Cir.2001) (district court properly “examine[d] the factual record underlying plaintiffs’ allegations in making its certification decision”). 15

An overlap between a class certification requirement and the merits of a claim is no reason to decline to resolve relevant disputes when necessary to determine whether a class certification requirement is met. Some uncertainty ensued when the Supreme Court declared in Eisen v. Carlisle & Jacquelin, 417 U.S. 156, 177, 94 S.Ct. 2140, 40 L.Ed.2d 732 (1974), that there is “nothing in either the language or history of Rule 23 that gives a court any authority to conduct a preliminary inquiry into the merits of a suit in *317 order to determine whether it may be maintained as a class action.” Only a few years later, in addressing whether a party may bring an interlocutory appeal when a district court denies class certification, 16 the Supreme Court pointed out that “the class determination generally involves considerations that are ‘enmeshed in the factual and legal issues comprising the plaintiffs cause of action.’ ” Livesay, 437 U.S. at 469, 98 S.Ct. 2454 (quoting Mercantile Nat’l Bank v. Langdeau, 371 U.S. 555, 558, 83 S.Ct. 520, 9 L.Ed.2d 523 (1963)). As we explained in Newton, 259 F.3d at 166-69, Eisen is best understood to preclude only a merits inquiry that is not necessary to determine a Rule 23 requirement. Other courts of appeals have agreed. 17 Because the decision whether to certify a class “requires a thorough examination of the factual and legal allegations,” id. at 166, the court’s rigorous analysis may include a “preliminary inquiry into the merits,” id. at 168, and the court may “consider the substantive elements of the plaintiffs’ case in order to envision the form that a trial on those issues would take,” id. at 166 (quot *318 ing 5 Moore’s Federal Practice § 23.46[4]) (quotation marks omitted). See id. at 168 (“In reviewing a motion for class certification, a preliminary inquiry into the merits is sometimes necessary to determine whether the alleged claims can be properly resolved as a class action.”). 18 A contested requirement is not forfeited in favor of the party seeking, certification merely because it is similar or even identical to one normally decided by a trier of fact. Although the district court’s findings for the purpose of class certification are conclusive on that topic, they do not bind the fact-finder on the merits. 19

The evidence and arguments a district court considers in the class certification decision call for rigorous analysis. A party’s assurance to the court that it intends or plans to meet the requirements is insufficient. See id. at 191 (“[W]here the court finds, on the basis of substantial evidence as here, that there are serious problems now appearing, it should not certify the class merely on the assurance of counsel that some solution will be found.”) (quoting Windham v. Am. Brands, Inc., 565 F.2d 59, 70 (4th Cir.1977)) (quotation marks omitted); Wachtel v. Guardian Life Ins. Co., 453 F.3d 179, 184, 186 (3d Cir.2006) (the requirement that a district court include in its class certification order “a clear and complete summary of those claims, issues, or defenses subject to class treatment” provides for the “full and clear articulation of the litigation’s contours at the time of class certification”).

Support for our analysis is drawn from amendments to Rule 23 that took effect in 2003. First, amended Rule 23(c)(1)(A) altered the timing requirement for the class certification decision. The amended rule calls for a decision on class certification “[a]t an early practicable time after a person sues or is sued as a class representative,” while the prior version had required that decision be made “as soon as practicable after commencement of an action.” We recognized in Weiss v. Regal Collections, 385 F.3d 337, 347 (3d Cir.2004), that this change in language, though subtle, reflects the need for a thorough evaluation of the Rule 23 factors — for this reason the rule does not “require or encourage premature certification determinations.” We explained:

Fed.R.Civ.P. 23 directs that certification decisions be made “at an early practicable time.” Fed.R.Civ.P. 23(c)(1)(a). This recent amendment replaced the language of the old rule: The former “ ‘as soon as practicable’ exaction neither reflected] prevailing practice nor *319 eapture[ed] the many valid reasons that may justify deferring the initial certification decision.” See Fed.R.Civ.P. 23(c)(1)(a) Advisory Committee Notes....

Allowing time for limited discovery supporting certification motions may ... be necessary for sound judicial administration. See [Newton, 259 F.3d at 166] (“[I]t may be necessary for the Court to probe behind the pleadings before coming to rest on the certification question.”) (quoting [Falcon, 457 U.S. at 160, 102 S.Ct. 2364])....

Id. at 347-48 n. 17; see Gariety, 368 F.3d at 365 (noting the change). 20 Relatedly, in introducing the concept of a “trial plan,” the Advisory Committee’s 2003 note focuses attention on a rigorous evaluation of the likely shape of a trial on the issues:

A critical need is to determine how the case will be tried. An increasing number of courts require a party requesting class certification to present a “trial plan” that describes the issues likely to be presented at trial and tests whether they are susceptible of class-wide proof.

Fed.R.Civ.P. 23 advisory committee’s note, 2003 Amendments.

Additionally, the 2003 amendments eliminated the language that had appeared in Rule 23(c)(1) providing that a class certification “may be conditional.” 21 The Advisory Committee’s note explains: “A court that is not satisfied that the requirements of Rule 23 have been met should refuse certification until they have been met.” The Standing Committee on Rules of Practice and Procedure advised:

The provision for conditional class certification is deleted to avoid the unintended suggestion, which some courts have adopted, that class certification may be granted on a tentative basis, even if it is unclear that the rule requirements are satisfied.

*320 Committee Report, supra, at 12; see 5 Moore’s Federal Practice § 23.80[2] (“The 2003 amendment clarifies that courts should not grant certification except after searching inquiry, and that courts should not rely on later developments to determine whether certification is appropriate.”)-

While these amendments do not alter the substantive standards for class certification, they guide the trial court in its proper task — to consider carefully all relevant evidence and make a definitive determination that the requirements of Rule 23 have been met before certifying a class. See IPO, 471 F.3d at 39 (2003 amendments “arguably combine to permit a more extensive inquiry into whether Rule 23 requirements are met than was previously appropriate”); Oscar, 487 F.3d at 267 (noting that these “subtle changes” reflect that “a district court’s certification order often bestows upon plaintiffs extraordinary leverage, and its bite should dictate the process that precedes it”).

To summarize: because each requirement of Rule 23 must be met, a district court errs as a matter of law when it fails to resolve a genuine legal or factual dispute relevant to determining the requirements.

2.

Class certification requires a finding that each of the requirements of Rule 23 has been met. 22 See Unger, 401 F.3d at 321 (“The plain text of Rule 23 requires the court to ‘find,’ not merely assume, the facts favoring class certification.”); Gariety, 368 F.3d at 365 (Rule 23(b)(3) requires the court to find predominance). Factual determinations necessary to make Rule 23 findings must be made by a preponderance of the evidence. In other words, to certify a class the district court must find that the evidence more likely than not establishes each fact necessary to meet the requirements of Rule 23. See Teamsters Local 445 Freight Div. Pension Fund v. Bombardier Inc., 546 F.3d 196, 202 (2d Cir.2008).

In reviewing a district court’s judgment on class certification, we apply the abuse of discretion standard. A district court abuses its discretion in deciding whether to certify a class action if its “decision rests upon a clearly erroneous finding of fact, an errant conclusion of law or an improper application of law to fact.” In re Gen. Motors Corp. Pick-Up Truck Fuel Tank Prods. Liab. Litig., 55 F.3d 768, 783 (3d Cir.1995) (citation and quotation marks omitted); see Newton, 259 F.3d at 165. Under these Rule 23 standards, a district court exercising proper discretion in deciding whether to certify a class will resolve factual disputes by a preponderance of the evidence and make findings that each Rule 23 requirement is met or is not met, having considered all relevant evidence and arguments presented by the parties. The abuse of discretion standard requires the judge to exercise sound discretion — failing that, the judge’s decision is not entitled to the deference attendant to discretionary rulings.

If a class is certified, “the text of the order or an incorporated opinion must include (1) a readily discernible, clear, and precise statement of the parameters defin *321 ing the class or classes to be certified, and (2) a readily discernible, clear, and complete list of the claims, issues or defenses to be treated on a class basis.” Wachtel, 453 F.3d at 187; see Fed.R.Civ.P. 23(c)(1)(B).

B.

Although the District Court properly described the class certification decision as requiring “rigorous analysis,” some statements in its opinion depart from the standards we have articulated. The District Court stated, “So long as plaintiffs demonstrate their intention to prove a significant portion of their case through factual evidence and legal arguments common to all class members, that will now suffice. It will not do here to make judgments about whether plaintiffs have adduced enough evidence or whether their evidence is more or less credible than defendants’.” With respect to predominance, the District Court stated that “[pjlaintiffs need only make a threshold showing that the element of impact will predominantly involve generalized issues of proof, rather than questions which are particular to each member of the plaintiff class.” (quoting Lumco Indus., Inc. v. Jeld-Wen, Inc., 171 F.R.D. 168, 174 (E.D.Pa.1997)). As we have explained, proper analysis under Rule 23 requires rigorous consideration of all the evidence and arguments offered by the parties. It is incorrect to state that a plaintiff need only demonstrate an “intention” to try the case in a manner that satisfies the predominance requirement. Similarly, invoking the phrase “threshold showing” risks misapplying Rule 23. A “threshold showing” could signify, incorrectly, that the burden on the party seeking certification is a lenient one (such as a prima facie showing or a burden of production) or that the party seeking certification receives deference or a presumption in its favor. So defined, “threshold showing” is an inadequate and improper standard. “[T]he requirements of Rule 23 must be met, not just supported by some evidence.” IPO, 471 F.3d at 33; see e.g., id. at 40, 42 (rejecting the view that a party seeking certification need only make “some showing” with respect to the Rule 23 requirements).

Citing Cumberland Farms, Inc. v. Browning-Ferris Industries, 120 F.R.D. 642, 645 (E.D.Pa.1988), the District Court reasoned, “[i]t is well recognized that private enforcement of [antitrust] laws is a necessary supplement to government action. With that in mind, in an alleged horizontal price-fixing conspiracy case when a court is in doubt as to whether or not to certify a class action, the court should err in favor of allowing the class.” See also Eisenberg v. Gagnon, 766 F.2d 770, 785 (3d Cir.1985) (citing Kahan v. Rosenstiel, 424 F.2d 161,169 (3d Cir.1970)) (advising that in a “doubtful” case when presented with a putative securities class action, court should err, if at all, in favor of certification). These statements invite error. Although the trial court has discretion to grant or deny class certification, the court should not suppress “doubt” as to whether a Rule 23 requirement is met— no matter the area of substantive law. Accordingly, Eisenberg should not be understood to encourage certification in the face of doubt as to whether a Rule 23 requirement has been met. Eisenberg predates the recent amendments to Rule 23 which, as noted, reject tentative decisions on certification and encourage development of a record sufficient for informed analysis. See Fed.R.Civ.P. 23 advisory committee’s note, 2003 Amendments (“A court that is not satisfied that the requirements of Rule 23 have been met should refuse certification until they have been met.”). We recognize the Supreme Court has observed that “[pjredominance is a *322 test readily met in certain cases alleging consumer or securities fraud or violations of the antitrust laws.” Amchem, 521 U.S. at 625, 117 S.Ct. 2231. But it does not follow that a court should relax its certification analysis, or presume a requirement for certification is met, merely because a plaintiffs claims fall within one of those substantive categories. See Fed.R.Civ.P. 28(b)(3) advisory committee’s note, 1966 Amendment (“Private damage claims by numerous individuals arising out of concerted antitrust violations may or may not involve predominating common questions.”); Robinson v. Tex. Auto. Dealers Ass’n, 387 F.3d 416, 420-21 (5th Cir.2004) (“There are no hard and fast rules ... regarding the suitability of a particular type of antitrust case for class action treatment. Rather, the unique facts of each case will generally be the determining factor governing certification.” (citations and quotation marks omitted)). “[A]ctual, not presumed, conformance” with the Rule 23 requirements remains necessary. Newton, 259 F.3d at 167 (quoting Falcon, 457 U.S. at 160, 102 S.Ct. 2364); see, e.g., E. Tex. Motor Freight Sys., Inc. v. Rodriguez, 431 U.S. 395, 405, 97 S.Ct. 1891, 52 L.Ed.2d 453 (1977) (although a putative class may bring a type of claim that “typically” involves common questions of law or fact, “careful attention to the requirements of [Rule 23] remains ... indispensable”).

To the extent that the District Court’s analysis reflects application of incorrect standards, remand is appropriate. We recognize that the able District Court did not have the benefit of the standards we have articulated. Faced with complex, fact-intensive disputes, trial courts have expended considerable effort to interpret and apply faithfully the requirements of Rule 23. One important reason for granting interlocutory appeals under Fed. R.Civ.P. 23(f) is to address “novel or unsettled questions of law” like those presented here. Newton, 259 F.3d at 164; see Fed.R.Civ.P. 23 advisory committee’s note, 1998 Amendments (cases in which “the certification decision turns on a novel or unsettled question of law” are among the best candidates for interlocutory appeal).

C.

Defendants contend the District Court erred as a matter of law in failing to consider the expert testimony of defendants’ expert, Ordover, instead deferring to the opinion of plaintiffs’ expert, Beyer. Plaintiffs do not dispute that a district court may properly consider expert opinion with respect to Rule 23 requirements at the class certification stage, but maintain that in this case the District Court considered and rejected Ordover’s opinion and defendants’ arguments based on it.

In addressing defendants’ Dauber l motion to exclude Beyer’s opinion, the court discussed whether it should consider Ordover’s opinion in deciding whether Beyer’s opinion was admissible. The court stated it would be improper to “weigh the relative credibility of the parties’ experts” — in other words, to weigh Ordover’s opinion against Beyer’s — for the purpose of deciding whether to admit or exclude Beyer’s opinion. Concluding Beyer’s opinion was admissible, the court denied the Daubert motion. But in addressing the Rule 23 requirements, the court did not confront Ordover’s analysis or his substantive rebuttal of Beyer’s points. Nor did the court address Ordover’s finding of substantial price disparities among similarly situated purchasers of hydrogen peroxide. The court appears to have assumed it was barred from weighing Ordover’s opinion against Beyer’s for the purpose of deciding whether the requirements of Rule 23 had been met. This was erroneous.

*323 l.

Expert opinion with respect to class certification, like any matter relevant to a Rule 23 requirement, calls for rigorous analysis. See West v. Prudential Sec., Inc., 282 F.3d 935, 938 (7th Cir.2002). 23 It follows that opinion testimony should not be uncritically accepted as establishing a Rule 23 requirement merely because the court holds the testimony should not be excluded, under Daubert or for any other reason. See IPO, 471 F.3d at 42 (rejecting the view that “an expert’s testimony may establish a component of a Rule 23 requirement simply by being not fatally flawed” and instructing that “[a] district judge is to assess all of the relevant evidence admitted at the class certification stage and determine whether each Rule 23 requirement has been met, just as the judge would resolve a dispute about any other threshold prerequisite for continuing a lawsuit”); 24 Blades, 400 F.3d at 569-70, 575 (affirming denial of class certification where the district court denied defendants’ Daubert motion and “considered all expert testimony offered by both sides in support of or in opposition to class certification” and “afforded that testimony such weight as [it] deemed appropriate”). Under Rule 23 the district court must be “satisfied,” Falcon, 457 U.S. at 161, 102 S.Ct. 2364, or “persuaded,” IPO, 471 F.3d at 41, that each requirement is met before certifying a class. Like any evidence, admissible expert opinion may persuade its audience, or it may not. This point is especially important to bear in mind when a party opposing certification offers expert opinion. The district court may be persuaded by the testimony of either (or neither) party’s expert with respect to whether a certification requirement is met. Weighing conflicting expert testimony at the certification stage is not only permissible; it may be integral to the rigorous analysis Rule 23 demands. See Blades, 400 F.3d at 575 (“[I]n ruling on class certification, a court may be required to resolve disputes concerning the factual setting of the case,” including “the resolution of expert disputes concerning the import of evidence concerning the factual setting — such as economic evidence as to business operations or market transactions”); West, 282 F.3d at 938 (cautioning that neglecting to resolve disputes between experts “amounts to a delegation of judicial power to the plaintiffs, who can obtain class certification just by hiring a competent expert”); see also Cordes & Co. Fin. Servs, v. A.G. Edwards & Sons, Inc., 502 F.3d 91, 106-07 (2d Cir.2007) (analyzing the opinions of plaintiffs’ and defendants’ experts); In re PolyMedica Corp. Sec. Litig., 432 F.3d 1, 5-6, 19 (1st Cir.2005) (vacating class certification order but finding no error in the “level of inquiry” the district court applied under Rule 23 when it “went well beyond the four corners of the pleadings, considering both parties’ expert reports and literally hundreds of pages of exhibits focused on market efficiency”). 25

*324 Resolving expert disputes in order to determine whether a class certification requirement has been met is always a task for the court — no matter whether a dispute might appear to implicate the “credibility” of one or more experts, a matter resembling those usually reserved for a trier of fact. Rigorous analysis need not be hampered by a concern for avoiding credibility issues; as noted, findings with respect to class certification do not bind the ultimate fact-finder on the merits. A court’s determination that an expert’s opinion is persuasive or unpersuasive on a Rule 23 requirement does not preclude a different view at the merits stage of the case.

That weighing expert opinions is proper does not make it necessary in every case or unlimited in scope. As the Court of Appeals for the Second Circuit instructed,

To avoid the risk that a Rule 23 hearing will extend into a protracted mini-trial of substantial portions of the underlying litigation, a district judge must be accorded considerable discretion to limit both discovery and the extent of the hearing on Rule 23 requirements. But even with some limits on discovery and the extent of the hearing, the district judge must receive enough evidence, by affidavits, documents, or testimony, to be satisfied that each Rule 23 requirement has been met.

IPO, 471 F.3d at 41. In its sound discretion, a district court may find it unnecessary to consider certain expert opinion with respect to a certification requirement, but it may not decline to resolve a genuine legal or factual dispute because of concern for an overlap with the merits. Genuine disputes with respect to the Rule 23 requirements must be resolved, after considering all relevant evidence submitted by the parties. See West, 282 F.3d at 938 (“Tough questions must be faced and squarely decided, if necessary by holding evidentiary hearings and choosing between competing perspectives.”); Szabo, 249 F.3d at 676 (district court must “resolve the disputes before deciding whether to certify the class”); IPO, 471 F.3d at 41 (Rule 23 calls for “definitive assessment” of its requirements); id. at 42 (rejecting the view that “a district judge may not weigh conflicting evidence and determine the existence of a Rule 23 requirement just because that requirement is identical to an issue on the merits”).

2.

Plaintiffs contend the District Court’s acceptance of their expert’s opinion was consistent with In re Linerboard Antitrust Litigation, 305 F.3d 145 (3d Cir.2002), an antitrust conspiracy action in which we affirmed class certification. There are a number of surface similarities between this case and Linerboard. Plaintiffs’ expert, Beyer, also appeared as an expert for the plaintiffs in Linerboard, and in both cases he presented an analysis of the industry and a “structure in pricing” analysis. Id. at 153. As in Linerboard, Beyer here proposed to demonstrate antitrust impact through the use of “benchmarks” and “multiple regression analysis.” Id. at 153-54. In affirming the district court’s grant of class certification in Linerboard, we concluded that Beyer, along with another *325 expert for the plaintiffs, “effectively utilized supporting data, including charts and exhibits, to authenticate their professional opinions that all class members would incur” antitrust impact. Id. at 155.

In Linerboard we did not address whether such expert opinion offered by the party opposing class certification would have been properly considered by the district court in the exercise of its discretion. But defendants here presented expert opinion disputing much of the evidence and argument plaintiffs offered for certification — specifically, Ordover disputed Beyer’s characterizations of the market and the alleged pricing structure. Defendants contend Ordover raised substantial doubts, not considered by the District Court, about whether common proof would be available for plaintiffs to demonstrate antitrust impact at trial.

We do not question plaintiffs’ general proposition, which the District Court accepted, that a conspiracy to maintain prices could, in theory, impact the entire class despite a decrease in prices for some customers in parts of the class period, and despite some divergence in the prices different plaintiffs paid. But the question at class certification stage is whether, if such impact is plausible in theory, it is also susceptible to proof at trial through available evidence common to the class. When the latter issue is genuinely disputed, the district court must resolve it after considering all relevant evidence. Here, the District Court apparently believed it was barred from resolving disputes between the plaintiffs’ and defendants’ experts. Rule 23 calls for consideration of all relevant evidence and arguments, including relevant expert testimony of the parties. Accordingly, we will vacate the order eerti-fying the class and remand for proceedings consistent with this opinion. 26

D.

Defendants contend the District Court, by relying on Bogosian v. Gulf Oil Corp., 561 F.2d 434 (3d Cir.1977), erroneously presumed the predominance requirement was met. In Bogosian, also a Clayton Act § 4 case in which plaintiffs sought class certification, the district court had denied class certification on the assumption that the issue of antitrust impact would have to be proven on an individual, as opposed to common, basis. 561 F.2d at 454. Finding that assumption erroneous, we reasoned that “when an antitrust violation impacts upon a class of persons who do have standing, there is no reason in doctrine why proof of the impact cannot be made on a common basis so long as the common proof adequately demonstrates some damage to each individual. Whether or not fact of damage can be proven on a common basis therefore depends upon the circumstances of each case.” Id. Applying the concept, we continued:

If, in this case, a nationwide conspiracy is proven, the result of which was to increase prices to a class of plaintiffs beyond the prices which would obtain in a competitive regime, an individual plaintiff could prove fact of damage simply by proving that the free market prices would be lower than the prices paid and that he made some purchases at the higher price. If the price structure in the industry is such that nationwide the conspiratorially affected prices at the wholesale level fluctuated within a range which, though different in different regions, was higher in all regions *326 than the range which would have existed in all regions under competitive conditions, it would be clear that all members of the class suffered some damage, notwithstanding that there would be variations among all dealers as to the extent of their damage.

Id. at 455; see Newton, 259 F.3d at 179 n. 21 (“In antitrust class actions, injury may be presumed when it is clear the violation results in harm to the entire class.”); Linerboard, 305 F.3d at 151-53.

In Linerboard, we found a “strong argument [could] be made that the Bogosian concept of presumed impact was properly applied” on the facts of that case. Id. at 152. Plaintiffs had alleged a horizontal conspiracy by manufacturers to restrict supply and raise prices of linerboard, the paper lining used in corrugated cardboard boxes and sheets. Id. Over a two-year period, defendants allegedly had agreed to idle their plants to reduce inventories to a twenty-year low and boost prices. Id. at 150-51. During the two-year class period, prices had risen by ninety percent. See id. at 152 (“Coincident with [defendants’] interference with the normal market forces, linerboard prices in the eastern United States rose in six consecutive price increases, from a low of around $270 to $290 per ton in third quarter 1993 to $530 per ton by April 1995”). The record in this case is different. Although the price of hydrogen peroxide rose at some points during the lengthy class period, the price was lower, not higher, at the end of the class period than at the beginning. And the evidence, as interpreted by defendants’ expert, shows that through much of the class period the production of hydrogen peroxide was increasing rather than decreasing. Moreover, there was an active dispute between the experts as to the “price structure in the industry” to which Bogosian refers. Defendants cited, for example, Ordover’s empirical analysis showing substantial price disparities among similarly situated customers. Accordingly, defendants contended, it was far from “clear the violation resulted] in harm to the entire class,” Newton, 259 F.3d at 179 n. 21. It is not apparent that the District Court considered, or believed it had the authority to consider, all the evidence in the record with respect to this dispute.

While the District Court found the Bo-gosian presumption applied, it also relied on Beyer’s analysis. Cf. Linerboard, 305 F.3d at 155 (“[T]his was not a case where plaintiffs relied solely on presumed impact and damages.”). As in Linerboard, plaintiffs here stress that they do not rely merely on Bogosian’s presumption to support class certification, but also on their expert’s analysis. Plaintiffs do not contend a bare allegation of a price-fixing conspiracy, in the absence of supporting evidence and analysis, suffices to support class certification consistent with a proper “rigorous analysis” under Rule 23. We emphasize that “[a]ctual, not presumed, conformance” with the Rule 23 requirements is essential. Newton, 259 F.3d at 167 (quoting Falcon, 457 U.S. at 160, 102 S.Ct. 2364). Applying a presumption of impact based solely on an unadorned allegation of price-fixing would appear to conflict with the 2003 amendments to Rule 23, which emphasize the need for a careful, fact-based approach, informed, if necessary, by discovery. See Fed.R.Civ.P. 23 advisory committee’s note, 2003 Amendments (“[Discovery in aid of the certification decision often includes information required to identify the nature of the issues that actually will be presented at trial.”).

The District Court, upon review of all the evidence consistent with this opinion, may again consider whether the reasoning in Bogosian is compatible with the record of this case. See In re Salomon Analyst *327 Metromedia Litig., 544 F.3d 474, 485-86 (2d Cir.2008) (remanding for opportunity for party opposing class certification to present evidence rebutting the fraud-on-the-market presumption, because Rule 23 requires a “definitive assessment” as to the predominance requirement).

V.

For the foregoing reasons, we will vacate the class certification order and remand for proceedings consistent with this opinion.