Freeland v. AT & T Corp.

238 F.R.D. 130, 2006 U.S. Dist. LEXIS 57394, 2006 WL 2380410

• Court: District Court, S.D. New York
• Decided: August 17, 2006
• Docket: No. 04 CIV. 8653(DLC)
• Status: Published

Opinion

OPINION & ORDER

COTE, District Judge.

The plaintiffs move for class certification in this antitrust action pursuant to Rules 23(b)(2) and (b)(3), Fed.R.Civ.P. That motion is denied.

Background

The plaintiffs are purchasers of cellular telephones and wireless telephone service from the defendants, the four largest providers of wireless telephone services in the United States.¹ They allege that the defendants have violated Sections 1 and 2 of the Sherman Antitrust Act, 15 U.S.C. §§ 1 & 2, by conspiring to restrain trade, conspiring to monopolize, tying, and making contracts in restraint of trade.

A. Procedural History

1. Brook Litigation

This motion comes before the Court in a somewhat unusual procedural posture. The full procedural history has been recited elsewhere, see In re Wireless Tel. Seros. Antitrust Litig., 385 F.Supp.2d 403, 406-07 & 406 n. 2 (S.D.N.Y.2005) (explaining history and collecting cases), and those discussions are incorporated herein by reference. Only the procedural history most relevant to the present motion is described here.

On April 5, 2002, the plaintiffs in Brook v. AT & T Cellular Servs., Inc., No. 02 Civ. 2637(DLC), filed suit in this District, alleging that several wireless service providers had violated federal antitrust law. At a conference on November 19, 2002, it was agreed that the Brook plaintiffs could amend their complaint on the understanding that no further amendments would be made. The amended complaint, filed January 19, 2003, alleged that each defendant individually tied the sale of handsets (telephones) to the sale of wireless services and that each defendant monopolized the market for wireless handsets compatible with its wireless services. The defendants moved to dismiss the amended complaint on February 21, 2003.

Shortly thereafter, the Judicial Panel on Multidistrict Litigation transferred four putative class actions raising similar claims to this Court. Through an Order on August 11, 2003, and with consent of the parties, the four transferred actions were consolidated with Brook for pretrial purposes. The August 11 Order further provided that the amended Brook complaint would serve as the consolidated class action complaint for any action “alleging antitrust claims against national wireless services carriers and assigned to the undersigned which is subsequently filed in or transferred to this Court.” Brook v. AT & T Cellular Servs., Inc., Nos. 02 Civ. 2637(DLC), 03 Civ. 1712(DLC), 03 Civ. 1713(DLC), 03 Civ. 1714(DLC), 03 Civ. 1715(DLC), 2003 WL 21911123, at *1 (S.D.N.Y. Aug. 11, 2003). Following the issuance of the August 11 Order, the consolidated actions were collectively referred to as In re Wireless Telephone Services Antitrust Litigation. Id.

The next day, August 12, an Opinion and Order was issued addressing the defendants’ motion to dismiss the amended complaint. See In re Wireless Tel. Servs. Antitrust Li-tig., No. 02 Civ. 2637(DLC), 2003 WL 21912603 (S.D.N.Y. Aug.12, 2003) (“Motion to Dismiss Opinion”). The Motion to Dismiss Opinion dismissed all of the plaintiffs’ monopolization claims and held that the plaintiffs had failed to state a per se tying claim. Id. at *9-10, *7. The Opinion held, however, that the plaintiffs had sufficiently stated a tying claim under the rule of reason doctrine. Id. at *7. The Motion to Dismiss Opinion went on to explain that “at trial, the plaintiffs will have the burden to show that each [defendant's market power and tying arrangement had an anticompetitive impact on the handset market.” Id. at *8.

On July 30, 2004, two months away from the end of fact discovery, the plaintiffs sought leave to amend their amended complaint. The proposed second amended complaint included an allegation that the defendants, both collectively and individually, have significant market power in both the tying and the tied product markets. An Opinion and Order issued on October 6, 2004, denied the plaintiffs’ motion for leave to amend. In re Wireless Tel. Servs. Antitrust Litig., No. 02 Civ. 2637(DLC), 2004 WL 2244502, at *1 (S.D.N.Y. Oct.6, 2004). The October 6 Opinion noted that the proposed second amended complaint would “transform the lawsuit from one asserting five tying claims against each of the defendants individually to a lawsuit alleging collective action on the tying claim.” Id. at *6. Plaintiffs had voluntarily dropped conspiracy allegations included in the original complaint when the amended complaint was filed. The subsequent decision to “reassert the conspiracy allegations initially abandoned” was found to not constitute the good cause required for “a substantial and untimely amendment” that would significantly delay the resolution of the litigation. Id.

2. Summary Judgment

Following the completion of fact and expert discovery in early 2005, the defendants moved for summary judgment. That motion was granted in an Opinion and Order issued on August 29, 2005, and corrected on September 2, 2005. In re Wireless, 385 F.Supp.2d 403 (S.D.N.Y.2005) (“Summary Judgment Opinion”).²

Summary judgment was granted on two grounds. First, the plaintiffs were unable to show that any one of the defendants had market power in the wireless services industry. To prevail on a tying claim, a plaintiff must prove that the defendant has “sufficient economic power in the tying product market to coerce purchaser acceptance of the tied product.” Id. at 414 (quoting Hack v. President & Fellows of Yale Coll., 237 F.3d 81, 86 (2d Cir.2000)). Yet “[n]one of the defendants enjoys a market share that would, standing alone, permit an inference of market power to be drawn, and the plaintiffs ha[d] not shown that questions of fact exist with respect to any other issue which, when combined with a defendant’s market share, would allow a finding of market power.” Id. at 417.

Second, the plaintiffs had not shown that each defendant’s tying arrangement had an anticompetitive effect on the handset market. Although not required for a per se claim, a successful tying claim brought under the rule of reason requires proof that “the defendants’ challenged behavior had an actual adverse effect on competition as a whole in the tied product market.” Id. at 415 (quoting Geneva Pharms. Tech. Corp. v. Barr Labs., Inc., 386 F.3d 485, 506-07 (2d Cir. 2004)).

The Opinion noted a fundamental problem with the plaintiffs’ tying claim. “The plaintiffs acknowledge that vigorous competition exists in the wireless services market.” Id. at 417. Yet “in a tying product market characterized by competition, even if all market participants engage in packaged sales, such conduct cannot have actual anticompeti-tive effects.” Id. at 424. Furthermore, it was “undisputed that the handset manufacturers compete with one another to offer the highest quality, maximum spectral efficiency, and lowest prices to the wireless service providers who purchase their handsets.” Id. at 409. Given this level of competition in both markets, it was “not surprising that plaintiffs [were] ... unable to show that the tying practice in which each defendant is alleged to have engaged has resulted in any anticom-petitive effects in the handset market.” Id. at 424; see also Jefferson Parish Hosp. Dist. No. 2 v. Hyde, 466 U.S. 2, 38,104 S.Ct. 1551, 80 L.Ed.2d 2 (1984) (O’Connor, J., concurring) (noting that a tying arrangement is economically harmful only if there is “a substantial threat that the tying seller will acquire market power in the tied market” and that “[n]o such threat exists if the tied product market is occupied by many stable sellers who are not likely to be driven out by tying”).

3. The Freeland Action

Meanwhile, on August 17, 2004 (during the briefing on the motion for leave to amend), plaintiffs’ counsel brought this action in the Northern District of California. The action was transferred to this Court in November 2004. An amended complaint was filed on October 18, 2005.

While the named plaintiffs are different, Freeland and the consolidated Brook actions involve all of the same major players. Plaintiffs’ counsel, defendants’ counsel, and defendants are the same in both. The named plaintiffs in each case brought their action “on behalf of persons who have purchased cellular or PCS telephone products from the five major carriers within the United States” and sought certification of a class consisting of all such purchasers.³

There is also considerable overlap in the substance of the actions. The Freeland complaint alleges collective action with respect to the defendants’ tying practices, but the theory of how the tie was accomplished and its anticompetitive effects on the handset market remain the same. The complaint also includes new charges that the defendants have entered into a conspiracy to restrain trade, contracts in restraint of trade, and a conspiracy to monopolize, but the gravamen of the complaint remains that the defendants have illegally locked handsets so that they cannot be used on another service provider’s network and tied the sale of handsets to the sale of their cellular services. See In re Wireless, 385 F.Supp.2d at 430 (“The allegation that each of the defendants locks handsets sold for use on its respective network is perhaps the heart of the plaintiffs’ case.”).

For all intents and purposes, then, this separately filed action is a continuation of the consolidated Wireless litigation that was decided on summary judgment in 2005. Discovery in this action was accordingly limited so as to not duplicate the lengthy and detailed discovery that was undertaken with respect to the Brook phase of the litigation. Given this limitation, the motion for class certification has been made following what is essentially complete fact and expert discovery in this case. Consideration of the motion was delayed four months further by a request from the parties for an extension of time and the submission of an additional round of briefing papers prompted by the submission of an unexpected expert report, as described below. Thus, while normally “[cjlass certification is to be decided ‘at an early practicable time’ after the commencement of a suit,” Jones v. Ford Motor Credit Co., 358 F.3d 205, 215 (2d Cir.2004) (quoting Fed.R.Civ.P. 23(c)(1)), this motion for class certification comes close to the end of litigation.^{3 4}

When plaintiffs submitted their brief in support of this motion, they did not file any additional substantive expert reports, choosing instead to support their motion with expert reports submitted as part of the Brook litigation. With their reply brief, however, plaintiffs submitted a declaration by one of their experts updating the analysis included in his previous reports. Even though plaintiffs had failed to demonstrate why it was appropriate to submit a new report with their reply, to which defendants would have no opportunity to respond, the defendants’ motion to strike the updated analysis was denied. Instead, the defendants were given an opportunity to write a surreply, to which the plaintiffs could respond in a sur-sur-reply. When the defendants submitted two rebuttal expert reports of their own with the sur-reply, plaintiffs moved to strike the additional reports.

That motion to strike is denied. Plaintiffs were given an opportunity to respond to the sur-reply and the accompanying reports in their sur-sur-reply. Moreover, defendants would not have felt the need, nor had the opportunity, to submit these additional reports had the plaintiffs not submitted their own supplemental report with their reply. Any inconvenience plaintiffs might have suffered in the course of responding to the defendants’ reports was of their own creation.

B. Factual Background

The detailed descriptions of the evolution of the wireless services industry, recent developments in handsets, and the distribution and sales of handsets included in the Summary Judgment Opinion are incorporated here by reference. See In re Wireless, 385 F.Supp.2d at 408-12. A brief sketch of the relevant facts are included below for the reader’s convenience. The facts included here are drawn principally from the Free-land complaint and the Summary Judgment Opinion.

The alleged conspiracy began when the defendants, acting as members of various trade groups, agreed to the adoption of industry standards for programming locks on all cellular telephone handsets. These locks prevent the unauthorized reprogramming of handsets, making it difficult for subscribers (and competing carriers) to reprogram the handsets to operate on another network. Through these groups, the defendants proceeded in a conspiracy to “lock[] handsets, t[ie] the sale of handsets to the sale of eellu-lar[] services, eliminat[e] alternate sources of handsets from the market, and horizontally divid[e] the market for handsets, with the intent and expectation of causing and maintaining artificially elevated prices for handsets and cellular[ ] services.”

None of the defendants manufactures handsets. Instead, each purchases handsets from manufacturers to whom they have transmitted specifications that each handset must meet in order to be activated on the carrier’s network. Because defendants control which handsets will be activated, they are able effectively to control the trajectory of handset development. Thus, it is ultimately the defendants who decide which new features will be included on handsets that reach the market.

The plaintiffs contend that the conspiracy to tie the purchase of a handset to the provision of service has allowed the defendants to sell handsets with features that many consumers do not want or use that add to the average cost of handsets, such as internet browsers. While the added features themselves may not be overpriced, because these features are unused, the plaintiffs contend the consumer has been injured by being forced to purchase an unnecessarily expensive handset to obtain wireless service.⁵

Manufacturers sell handsets in bulk to carriers and sales agents of the carriers, but do not sell directly to consumers in the United States. It is undisputed, though, that the market for handsets is competitive, as manufacturers compete to offer the highest quality and lowest prices to the service providers who purchase handsets. In re Wireless, 385 F.Supp.2d at 409. It is also undisputed that defendants use their offers of handsets at the lowest possible price to compete with each other for subscribers. Id. at 430 n. 40.

As part of the implementation of the handset locking scheme, the defendants have adopted a uniform tying policy such that each defendant markets handsets and services in essentially the same way. Typically, the handsets are bundled with a service contract (usually 12 or 24 months) that has substantial penalties for early termination. It is undisputed that wireless service providers have consistently sold their respective service and handsets as a package, and that in doing so, the carriers have subsidized the cost of handsets to make initial entry into the wireless services market more palatable for the consumer. Id. at 410. The subscriber agreements that govern the terms of wireless services provision do not themselves require the subscribers to purchase handsets from the defendants. But commercial realities limit alternatives — manufacturers do not sell directly to consumers, and handsets purchased from a competing carrier are programmed and locked for use on that carrier’s network. Thus, plaintiffs contend that subscribers have little choice but to purchase a handset from the carrier that provides their wireless service.

Discussion

The plaintiffs seek certification pursuant to Federal Rule of Civil Procedure 23 of a class consisting of “[a]ll persons who have purchased wireless telephone services or handsets in the United States from any defendant [after] January 1, 1999.” They further seek certification of subclasses consisting of class members who have purchased wireless services or handsets from each of the defendants.

A. Requirements for Class Certification

To qualify for class certification, a putative class must meet the four threshold requirements of Federal Rule of Civil Procedure 23(a). In re Visa Check/MasterMoney Antitrust Litig., 280 F.3d 124, 132 (2d Cir. 2001). Thus, the plaintiffs will be able to sue defendants as representatives of the class

only if (1) the class is so numerous that joinder of all members is impracticable, (2) there are questions of law or fact common to the class, (3) the claims or defenses of the representatives are typical of the claims or defenses of the class, and (4) the representative parties will fairly and adequately protect the interests of the class.

Rule 23(a), Fed.R.Civ.P.

Once these criteria are satisfied, an action may be maintained as a class action only if it qualifies under at least one of the categories provided in Rule 23(b). Visa Check, 280 F.3d at 133. In this case, plaintiffs seek to certify a class under Rules 23(b)(2) and (b)(3). Rule 23(b)(2) permits certification where “the party opposing the class has acted or refused to act on grounds generally applicable to the class, thereby making appropriate final injunctive relief or declaratory relief with respect to the class as a whole.” Fed.R.Civ.P. Rule 23(b)(2). Rule 23(b)(3) permits certification where “the court finds that the questions of law or fact common to the members of the class predominate over any questions affecting only individual members, and that a class action is superior to other available methods for the fair and efficient adjudication of the controversy.” Fed.R.Civ.P. Rule 23(b)(3). The burden of proving compliance with all of the requirements of Rule 23 rests with the party moving for certification. Visa Check, 280 F.3d at 132.

“A motion for class certification is not an occasion for examination of the merits of the ease,” Caridad v. Metro-North Commuter R.R., 191 F.3d 283, 291 (2d Cir.1999), but a trial court must nonetheless “conduct a rigorous analysis to ensure that the prerequisites of Rule 23 have been satisfied before certifying a class.” Visa Check 280 F.3d at 135 (citation omitted). This limitation applies equally to the court’s evaluation of expert evidence submitted in support of a motion for class certification. The court “must ensure that the basis of the expert opinion is not so flawed that it would be inadmissible as a matter of law.” Id. But the question at this stage is “whether plaintiffs’ expert evidence is sufficient to demonstrate common questions of fact warranting certification of the proposed class, not whether the evidence will ultimately be persuasive.” Id. As a result, a court must avoid weighing conflicting expert testimony which is otherwise admissible. Id.

B. Rule 23(a) Threshold Requirements

1. Numerosity

Rule 23(a) requires a finding that the numerosity of the putative class makes joinder of all class members “impracticable.” Robidoux v. Celani, 987 F.2d 931, 935 (2d Cir.1993). Numerosity is presumed when a class consists of forty or more members. Consol. Rail Corp. v. Town of Hyde Park, 47 F.3d 473, 483 (2d Cir.1995); In re World-Com, Inc. Sec. Litig., 219 F.R.D. 267, 279 (S.D.N.Y.2003).

Defendants do not contest the numerosity of the putative class or plaintiffs’ claim that defendants have more than 100 million subscribers. A class of this size exceeds the threshold for numerosity by several orders of magnitude.

2. Commonality

“The commonality requirement is met if plaintiffs’ grievances share a common question of law or fact.” Robinson v. Metro-North Commuter R.R., 267 F.3d 147, 155 (2d Cir.2001) (citation omitted). Even a single common legal or factual question will suffice. See In re Agent Orange Prod. Liab., 818 F.2d 145, 166-67 (2d Cir.1987).

There are several common questions of law or fact shared by members of the putative class. Among these are whether the defendants conspired to, inter alia, lock handsets and tie the sale of handsets to the sale of wireless services. Other common questions include whether the defendants entered into contracts in restraint of trade, whether the defendants exercise market power in the market for wireless services, and whether their marketing practices have had anticompetitive effects in the market for handsets.

Defendants do not argue that there is not a common question of law or fact that would support a finding of commonality. Instead, they focus on the predominance requirement of Rule 23(b)(3), noting that it is more stringent than the 23(a) commonality requirement and therefore more often the focus of contention when there is a motion to certify a damages class. Defendants are correct that “the predominance criterion is far more demanding” than the commonality requirement, Amchem Products, Inc. v. Windsor, 521 U.S. 591, 624, 117 S.Ct. 2231, 138 L.Ed.2d 689 (1997), but this is not relevant to the analysis of the commonality requirement, which is clearly satisfied in this case.

3. Typicality

The claims of the class representatives are typical of those of the class when “each member’s claim arises from the same course of events, and each class member makes similar legal arguments to prove the defendant’s liability.” Robinson, 267 F.3d at 155 (citation omitted). The typicality and commonality requirements tend to merge into one another as both “serve as guideposts for determining whether the named plaintiffs claim and the class claims are so interrelated that the interests of the class members will be fairly and adequately protected in their absence.” Caridad, 191 F.3d at 291 (quoting Gen. Tel. Co. v. Falcon, 457 U.S. 147, 157, n. 13, 102 S.Ct. 2364, 72 L.Ed.2d 740).

Here, each class member’s claim arises from the same course of events, namely the alleged conspiracy to lock handsets and tie the sale of handsets to the sale of wireless services, and the subsequent anti-competitive actions taken by each defendant to implement the conspiracy. See Robidoux, 987 F.2d at 936-37 (“When it is alleged that the same unlawful conduct was directed at or affected both the named plaintiff and the class sought to be represented, the typicality requirement is usually met irrespective of minor variations in the fact patterns underlying individual claims.”). All members’ claims are also based on the same legal theories of conspiring to tie and take other actions that violate federal antitrust laws. See In re Drexel Burnham Lambert Group, Inc., 960 F.2d 285, 291 (2d Cir.1992) (finding typicality where “each class member’s claim is based on [defendant’s] purported violations of securities laws”). Consequently, the typicality requirement is satisfied in this case. Defendants do not contend otherwise.

4. Adequacy of Representation

“[A]dequacy of representation entails inquiry as to whether: 1) plaintiffs interests are antagonistic to the interest of other members of the class and 2) plaintiffs attorneys are qualified, experienced and able to conduct the litigation.” Baffa v. Donaldson, Lufkin & Jenrette Sec. Corp. 222 F.3d 52, 60 (2d Cir.2000). The adequacy inquiry “serves to uncover conflicts of interest between named parties and the class they seek to represent.” Amchem, 521 U.S. at 625, 117 S.Ct. 2231. Not every potential conflict will preclude a finding of adequacy, however. Visa Check, 280 F.3d at 145. “The conflict that will prevent a plaintiff from meeting the Rule 23(a)(4) prerequisite must be fundamental, and speculative conflict should be disregarded at the class certification stage.” Id. (citation omitted).

Defendants argue that there are three fundamental conflicts that make class certification inappropriate. First, some class members welcome the inclusion of the additional features that the named plaintiffs argue are unwanted by consumers. Second, many, and perhaps most, class members actually benefit from the subsidization that accompanies the bundling that the named plaintiffs claim is illegal. And third, both owners and other subscribers alike benefit from the deployment of newer, more efficient handsets, which is encouraged by defendants’ current locking and tying practices. Defendants argue that if these practices are held illegal or enjoined, then all of the consumers who have benefitted from them will be harmed as a result of the named plaintiffs’ legal action. This, according to the defendants, creates fundamental conflicts within the class that preclude a finding of adequacy of representation.

The adequacy requirement of Rule 23(a)(4), however, “is designed to ferret out potential conflicts between representatives and other class members.” Id. (emphasis added). The focus, then, should be on the ways in which the interest of the named plaintiffs in this case is antagonistic to the interests of other class members. With respect to the first conflict, defendants have not identified a divergence of interests. Indeed, the defendants point specifically to four named plaintiffs who appear to value the inclusion of additional features in their handsets. Defendants do not explain why these individual plaintiffs could not fairly and adequately represent the interests of others like them in this litigation.

As for the second and third conflicts, the defendants’ arguments reduce to a claim that most plaintiffs actually benefit from the defendants’ marketing practices. The defendants present no basis for a conclusion, though, that the named plaintiffs do not also benefit from these practices. Thus, defendants provide no reason to believe that the named plaintiffs’ interests are any different from those of the class members they represent.

The fundamental problem with defendants’ arguments regarding adequacy is that they do not actually concern the adequacy of representation of the class by the named plaintiffs. Rather, these arguments are little more than a repackaging of the defendants’ claim that their tying and locking practices have not caused an antitrust injury on a classwide basis. Their assertion that “[t]he differences among the purported class members that make it impossible to satisfy the predominance inquiry likewise underlie the ‘fundamental conflicts’” is telling in this regard.

At best, the asserted conflicts threaten the cohesion of the class as whole, but there is no reason to believe that the named plaintiffs would not adequately represent the interests of the various subgroups within the class to the extent those interests diverge. Moreover, the Court remains free to “define, redefine, subclass, and decertify as appropriate in response to the progression of the case from assertion to facts.” Richardson v. Byrd, 709 F.2d 1016, 1019 (5th Cir.1983), quoted with approval in Boucher v. Syracuse Univ., 164 F.3d 113, 118 (2d Cir. 1999); cf. Visa Check, 280 F.3d at 141 (finding no abuse of discretion in the district court’s manageability finding when the court “specifically recognized its ability to modify its class certification order ... or even decertify the class if such an action ultimately became necessary’).

The plaintiffs have carried their burden of demonstrating that they “possess the same interest and suffer the same injury as the class members,” Amchem, 521 U.S. at 625-26, 117 S.Ct. 2231 (citation omitted), assuming (as the Court must at this stage) that such an injury has occurred. Any conflict between the named plaintiffs and the rest of the class is entirely speculative. Moreover, there is no reason to believe that plaintiffs’ counsel, who has led the Wireless litigation for the past several years, lacks the competency to handle the class action. Named plaintiffs can be trusted to represent fairly and adequately the interests of the class in this action.

C. Rule 23(b)(3): Predominance of Common Questions

“In order to meet the predominance requirement of Rule 23(b)(3), a plaintiff must establish that the issues in the class action that are subject to generalized proof, and thus applicable to the class as a whole, predominate over those issues that are subject only to individualized proof.” Visa Check, 280 F.3d at 136 (citation omitted). Predominance is established if the legal or factual issues that can be resolved through generalized proof are “more substantial than the issues subject only to individualized proof.” Moore v. PaineWebber, Inc., 306 F.3d 1247, 1252 (2d Cir.2002). In making this determination, a court considers whether the putative class members “could establish each of the ... required elements of [their] claim[s] ... using common evidence.” Visa Check, 280 F.3d at 136. The essential inquiry for predominance is whether the proposed classes are “sufficiently cohesive to warrant adjudication by representation.” Amchem, 521 U.S. at 623,117 S.Ct. 2231.

There are two elements of the plaintiffs’ claims that will require individualized proof: antitrust injury and coercion. Because these issues are so substantial and individualized proof of each so burdensome, plaintiffs cannot carry their burden of proving that certification of a(b)(3) class is appropriate.

1. Antitrust Injury

To prevail on a claim under the Sherman Act, a plaintiff must, in addition to proving a violation of antitrust law, also establish “antitrust injury, which is to say injury of the type the antitrust laws were intended to prevent and that flows from that which makes defendants’ acts unlawful.” Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 489, 97 S.Ct. 690, 50 L.Ed.2d 701 (1977). Antitrust injury is a required element of all of plaintiffs’ claims in this action. See Daniel v. Am. Bd. of Emergency Med., 428 F.3d 408, 438 (2d Cir.2005); Visa Check 280 F.3d at 134 n. 5.

The plaintiffs propose to establish antitrust injury on a classwide basis at trial by proving that the defendants’ locking and tying practices have inflated handset prices above competitive levels.⁶ Antitrust injury based on payment of an illegal overcharge may be subject to proof on a class-wide basis. Bogosian v. Gulf Oil Corp., 561 F.2d 434, 455 (3d Cir.1977). If the plaintiffs could establish that “the defendants engaged in an unlawful national conspiracy which had the effect of stabilizing prices above competitive levels, and further establish that the [plaintiffs] were consumers of that product, ... the jury could reasonably conclude that [defendants’] conduct caused injury to each [class member].” In re Master Key Antitrust Litigation, 528 F.2d 5, 12 n. 11 (2d Cir.1975) (dicta). Injury would thereby be established by common evidence and further individualized inquiry would be unnecessary. Plaintiffs seeking class certification may carry their burden of establishing that injury is subject to generalized proof by submitting an expert report that “posits class-wide injury resulting from every single class member’s overpaying for [a tied product] as a direct result of the tie.” Visa Check, 280 F.3d at 137 (citation omitted); see also id. at 138 (concluding that district court’s resolution of this issue did not constitute an abuse of discretion).

a. Admissibility of Economides’s Regression Analysis

The plaintiffs rely on the benchmark and regression analyses of Professor Economides to demonstrate class-wide injury.⁷ Econom-ides argues that average wholesale price of handsets became inflated after 1999, when the defendants had fully implemented their locking and tying agreements. He contends that the inflation is attributable to a number of factors. One is that the defendants have required manufacturers to produce handsets with features, such as internet browsers, that add to the cost of handsets but are in fact never used, and presumably unwanted, by the large majority of subscribers. Other factors that he identifies as contributing to the increase in handset price include the elimination of direct sales from manufacturers, conflation of the price of the handset with the price of the service plan, and the elimination of many low-cost handset models from the market.⁸

According to Economides, his regression analysis is intended to answer the following question: “What has been the effect, if any, of the tying and locking of handsets and services on the price of handsets sold in the United States from January 1, 1999 to the present?” Regression analysis is a commonly accepted statistical tool used to examine “the effect of independent variables on a dependent variable.” Bickerstaff v. Vassar Coll., 196 F.3d 435, 448 (2d Cir.1999). As reflected in the question above, Economides’s analysis is intended to reveal the effect of the defendants’ tying practices (an independent, or explanatory variable) on the price of handsets (the dependent variable).

While the analysis is intended to explain effects on the retail price of handsets, Eco-nomides has chosen the average wholesale price of handsets as the dependent variable. Economides claims that the defendants’ marketing practices obscure the true economic price of the handset in the retail transaction, making accurate determinations of retail price difficult. Assuming that 100% of the wholesale price would be passed on to consumers, Economides theorizes that any changes in wholesale price attributable to the effects of defendants’ tying practices are ultimately borne by consumers. He therefore uses wholesale price as a substitute for retail price in his analysis. Because most handsets are manufactured abroad, Economides adopts as a measure of wholesale price data from the United States International Trade Commission (“USITC”) on the declared value of handsets that are imported into the United States.

In selecting his independent variable, Eco-nomides begins with the assumption that the prices of handset components and of other electronic goods (such as digital cameras) built from similar components should serve as rehable proxies for the factors that affect the wholesale price of handsets. He identifies three price indices published by the Bureau of Labor Statistics (“BLS”) that reflect the average prices of semiconductors, computers, and electronic goods, respectively, and performs a separate regression on each of the three indices for the years 1994 to 1998. He then uses the “most conservative” of these regressions on the BLS indices for the years before the alleged conspiracy was fully implemented to predict what handset prices would have been for the years 1999 to 2003 but for the defendants’ tying practices.⁹ The difference between the predicted price and the actual observed price for those years represents Economides’s answer to the question posed above. Economides’s analysis does not permit attribution of the difference between his predicted handset price and actual handset price to any cause other than the defendants’ tying and locking practices. He explained that he “could not think of another explanation” for why prices might have changed after 1999. In re Wireless, 385 F.Supp.2d at 428 (quoting Economides’s testimony).

The regression described thus far was originally submitted to the Court to support plaintiffs’ summary judgment papers in the Brook phase of this litigation. See id. at 427-30 (discussing Economides’s expert report). The Summary Judgment Opinion held that Economides’s regression analysis was methodologically unsound and therefore could not be admitted as expert testimony under Rule 702, Fed.R.Evid. See id. at 427.

The primary defect identified by the Court was Economides’s “failure to test for [the] obvious and significant alternative explanations” identified by the defendants to account for the discrepancy between the projected and actual prices. Id. at 428. Defendants had pointed to two major developments that might have affected wholesale price after 1999: the shift from analog to digital technology and advances in handset quality. Eco-nomides’s failure to explore these alternative explanations rendered his analysis “essentially worthless.” Id. (citation omitted).

In support of the present motion for class certification, plaintiffs offer a supplemental declaration from Economides explaining updates he has made to his analysis of handset price, including extending his forecast from 2003 to 2005 using newly available USITC data. Economides also included a new series of regressions (one for each BLS index variable) that included a dummy variable to represent the presence or absence of the alleged conspiracy.¹⁰ The use of a dummy variable, suggested originally by defendants’ expert Dennis Carlton, enabled the incorporation of index data for all relevant years (1994-2005) into the regression, instead of limiting the regression analysis to pre-conspiracy years and then using results from that limited regression analysis to forecast prices for the conspiracy years. Economides suggests that the dummy-variable model provides for less precise measures of damages, but nonetheless offers a useful cross-check on his original prediction model. Significantly, he never attempted to rectify the shortcomings identified by the Court or introduce any other explanatory variables into the regression model with the dummy variable. Instead, plaintiffs (and Economides) renew their previously made arguments that the omissions of the two identified variables are justified.

The burden of proving the admissibility of expert evidence, as with all evidence, rests with its proponent. See Fed R. Evid. 702 Advisory Committee Notes (“[T]he admissibility of all expert testimony is governed by the principles of Rule 104(a). Under that Rule, the proponent has the burden of establishing that the pertinent admissibility requirements are met by a preponderance of the evidence.”); Bourjaily v. United States, 483 U.S. 171, 175-76, 107 S.Ct. 2775, 97 L.Ed.2d 144 (1987). The proponent of expert evidence must establish that the evidence is reliable and will assist the trier of fact to understand or determine a fact in issue. Rule 702, Fed.R.Evid. Expert evidence must also meet the relevancy requirement of Rule 401. Bickerstaff, 196 F.3d at 449. As it determines whether these requirements have been met, a court is mindful of its obligation to act as a gatekeeper to ensure the “reliability and relevancy of expert testimony.” Kumho Tire Co., Ltd. v. Carmichael, 526 U.S. 137, 152, 119 S.Ct. 1167, 143 L.Ed.2d 238 (1999).

Ordinarily, the failure to include a variable in a regression analysis will affect the probative value of the analysis and not its admissibility. Bazemore v. Friday, 478 U.S. 385, 400, 106 S.Ct. 3000, 92 L.Ed.2d 315 (1986). Where a study accounts for the “major factors” but not “all measurable variables,” it is admissible. Id. (citation omitted). Where significant variables that are quantifiable are omitted from a regression analysis, however, the study may become so incomplete that it is inadmissible as irrelevant. Bickerstaff, 196 F.3d at 449. Because the burden of proving helpfulness and relevance rests on the proponent of a regression analysis, it is the proponent who must establish that the major factors have been accounted for in a regression analysis.

The plaintiffs attempt to shift that burden to the defendants. Specifically, the plaintiffs contend that a party opposing the admission of a regression analysis as incomplete for the failure to include a given variable is required by the Second Circuit’s opinion in Sobel v. Yeshiva University, 839 F.2d 18 (2d Cir. 1988), to make certain showings with respect to the omitted variable.

The plaintiffs’ reliance on Sobel is misplaced. Sobel does require “a defendant challenging the validity of a multiple regression analysis” to “make a showing of relevance for each particular variable it contends [the proponent of the analysis] ought to include.” Sobel, 839 F.2d at 34. And it further suggests that the opponent “[i]deally ... would seek to do so by offering its own regression that includes the variable it contends improperly was omitted.” Id. But the Court of Appeals was describing the showing to be made by a party challenging the weight during a bench trial of an admissible regression analysis. See id. (observing that the objections to the regressions went “to the weight rather than their admissibility”). So-bel did not purport to, and cannot be read to, alter the burdens imposed by the Federal Rules of Evidence and Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993), on the proponent of scientific evidence. See Daubert, 509 U.S. 579, 592 & n. 10, 113 S.Ct. 2786,125 L.Ed.2d 469 (1993).¹¹

Plaintiffs attempt to carry their burden by offering a series of arguments to explain why the missing variables identified by the Summary Judgment Opinion should not be included in the regression analysis. The reasons given by the plaintiffs for excluding a percent-digital variable are that the variable is tainted by the alleged anticompetitive conduct; that the shift to digital would not affect handset prices; and that a percent-digital variable suffers from multicollinearity with (i.e., it is highly correlated with) the single explanatory variable included in each regression. The reasons for excluding a quality variable are that handsets improved in quality from 1994 to 1998 while prices continually declined; that the comparator products underwent similar improvements in quality while prices continually declined; and that quality is unquantifiable. None of these reasons is persuasive.

i. The percent-digital variable

The first reason for excluding the percent-digital variable is utterly without merit. Plaintiffs argue in passing that because the defendants’ illegal conspiracy allowed them to dictate handset specifications, a percent-digital variable is tainted by their conduct. The plaintiffs also contend, however, that the transition from analog to digital handsets was virtually complete by 1999, while the conspiracy to tie and lock handsets was not fully implemented until that year. Accepting the plaintiffs’ view of the evidence, the conspiracy can hardly account for the transition. Even more significantly, the shift from analog to digital technologies occurred across the consumer electronics industry, a point plaintiffs rely on in making their second argument for exclusion of this variable. The suggestion that this widespread technological shift was influenced by the defendants’ decision to lock and tie handsets is nonsense.

With respect to the second reason, Eco-nomides’s unsupported opinion that the shift to digital technology would not affect handset price is not a sufficient reason for its exclusion from the regression analysis. Indeed, the entire objective of undertaking the regression analysis was to discover which independent variables affected handset price. Completely irrelevant variables, to be sure, should be omitted, and marginally relevant variables may be omitted, but Economides offered no empirical analysis to support his conclusion that the percentage of handsets employing digital technology was irrelevant to average price.¹² Defendants’ expert, Dennis Carlton, on the other hand, did offer empirical evidence and analysis to demonstrate why a shift to digital handsets would affect average handset price. Carlton further performed his own regression analysis including a percent-digital variable and observed that this modification resulted in a reduced estimate of the effect of the tying practices.

As for multicollinearity, plaintiffs rely on Sobel for the proposition that when a party challenges a regression analysis for failure to include a specific variable, the omitted variable “must be shown not to be multicollinear with those variables already included.” So-bel, 839 F.2d at 35. As described above, Sobel does not shift any part of the burden of proving admissibility to the opponent of a regression analysis. Nor can Sobel be read to justify the admission of a regression analysis that lacks a major explanatory variable if that variable is multicollinear with another explanatory variable already included in the analysis. The issue in Sobel was not the admissibility of a regression analysis but its probative value. The Court of Appeals was considering the criteria for adding an additional variable to a regression analysis that was already admissible, which is to say that the regression was already sufficiently complete as to be relevant to the issue at hand. Furthermore, the court had just observed that the defendants had failed to demonstrate that the excluded variables would have made a difference to the plaintiffs’ regression analysis when it referred as well to the need to consider multicollinearity. Although the Second Circuit did not explain the rationale behind its no-multicollinearity requirement, it seems likely that the court was concerned with protecting the integrity of a possibly incomplete, but nonetheless probative, regression analysis.¹³

In this case, however, the regression performed by Economides is “so incomplete as to be inadmissible as irrelevant.” Bazemore, 478 U.S. at 400, 106 S.Ct. 3000 (1986). In contrast to the situation addressed in Sobel, the inclusion of a demonstrably (and logically) relevant but multicollinear variable into such an incomplete regression analysis as Economides’s could only increase its proba-five value. For as Judge Posner has noted, “there are statistical methods for solving this problem (the problem of multicollinearity, as it is called by statisticians).” In re High Fructose Corn Syrup Antitrust Litig., 295 F.3d 651, 660 (7th Cir.2002); see also Federal Judicial Center, Reference Manual on Scientific Evidence 197 (2d ed. 2000) (“Reference Manual”) (“Fortunately, the reported regression statistics take into account any multicollinearity that might be present.”).¹⁴ In sum, the shift from analog to digital technology in handsets is “too significant not to be accounted for in the regression analysis in this ease,” Bickerstajf, 196 F.3d at 449, even if it is multicollinear with another variable in the regression.

ii. The quality variable

Plaintiffs’ arguments regarding a quality variable fare no better. They first argue that quality changes should not be expected to have affected prices after 1998 because handsets improved in quality from 1994 to 1998 while decreasing in average price. It is undoubtedly true that handsets improved in quality from 1994 to 1998, but that improvements in quality did not cause prices to rise in certain years does not mean that they never could, particularly if the improvements were dramatic. Two examples come readily to mind: the introduction of camera phones and “smart” phones. See, e.g., New York v. Microsoft Corp., 224 F.Supp.2d 76, 125 (D.D.C.2002) (defining “smart phones” as “devices that combine the PC-like capabilities of handheld devices and mobile telephone technology”). The integration of digital cameras and “computing devices that perform many of the functions of an ordinary personal computer,” id., into a cellular telephone are remarkable improvements in quality. These changes, moreover, are arguably more dramatic than the changes in size and weight referred to by plaintiffs for the earlier time period, and might be reasonably expected to have a correspondingly more dramatic effect on average handset price. An unsupported assumption that changes in quality would never affect handset price is insufficient to justify its exclusion from the regression analysis, particularly in light of evidence to the contrary, discussed below.

For largely similar reasons, the comparison to price trends for the comparator products fails to justify the omission of a quality variable from the regression analysis. Again, it is undoubtedly true that other consumer electronic products improved in quality, but Economides does not contend that the improvements were identical, and has not established that the improvements were even similar. As defendants point out, handsets that include digital cameras have become common, but digital cameras that include cellular telephones have not. It cannot simply be assumed, as Economides does, that improvements in handset quality did not affect handset price because the prices of other products behaved differently over the same time period.

The primary argument advanced by Eco-nomides and plaintiffs to explain why a quality variable should not be included in the regression is that “quality” is not measurable in such a way that it can be quantified and controlled. It is true that variables that cannot be measured cannot for that reason be included in a regression analysis. See Reference Manual at 188. But this does not mean that a more easily quantified proxy for quality could not be included, such as the percentage of handsets that include a novel feature. Carlton suggested, for example, the percentage of handsets that include cameras or data functionality (i.e., smart phones). Economides did not explain why a proxy could not be identified and included in his regression.

Carlton goes further, though, than merely suggesting how a quality variable might be included in the regression analysis. Carlton performed his own regressions including a variable that reflected the percentage of handsets that include either a camera or data functionality. He found that the inclusion of a variable that captures these changes in handset characteristics had profound effects on the results of the analysis, eliminating the statistical significance and drastically reducing the estimated impact of the alleged conspiracy.

Furthermore, Carlton provides data on handset price broken down by categories corresponding to handset characteristics to demonstrate how focus on overall handset price could yield a misleading picture of price change over time. To do this, Carlton provided data on wholesale handset price and number of units sold by class of handset. He demonstrated that the dramatic shift in sales from lower- to higher-priced (and higher quality) categories could lead to an increase in overall average price even though the average price within each class of handset declined over the same time period. Carlton’s analysis convincingly demonstrates how changes in handset quality could affect average handset price. The plaintiffs have not carried their burden of proving the relevance and helpfulness of a regression that does not account for this effect.

The impact from the exclusion of a relevant variable can be significant. “Failure to include a major explanatory variable that is correlated with the variable of interest in a regression model may cause an included variable to be credited with an effect that actually is caused by the excluded variable.” Reference Manual at 188. If this occurred, and the effect on price attributable to the omitted variables was erroneously attributed Eco-nomides’s sole explanatory variable in each regression, then the forecasts of prices in the years following the implementation of the conspiracy would be of questionable value. These forecasts are the basis for Econom-ides’s claim that handset prices would have been lower but for the alleged conspiracy and the primary means by which plaintiffs intend to prove injury on a classwide basis.

The lack of any attempt to include an independent variable for quality in an investigation of price changes is especially remarkable. An analysis of price change in a product must account for changes in the product over time “so that only real price changes will be measured.” Bureau of Labor Statistics, U.S. Dep’t of Labor, BLS Handbook of Methods, Ch. 14 at 3, available at http://www.bls.gov/ opub/hom/pdf/homchl4.pdf. Indeed, this is why the BLS indices, including the indices used by Economides in his regression, incorporate adjustments for changes in the characteristics of the products over time.¹⁵ By failing to include an independent variable for quality in his regression analysis, Econom-ides cannot be sure that he is measuring true price changes.

At least two significant variables have been omitted from Economides’s regression analysis. The plaintiffs’ arguments for exclusion of these variables are unpersuasive and they have failed to carry their burden of proving that the incomplete analysis is sufficiently relevant and helpful to justify its admission into evidence in this case. The largely unmodified regression analysis remains “essentially worthless” and is inadmissible under Federal Rules of Evidence 401 and 702.

Without the regression analysis, plaintiffs are left with only their benchmark model comparing market prices for handsets since the start of the alleged conspiracy with market prices for comparable consumer electronics. Economides treats the benchmark model as a “check on the reasonableness” of his regression analysis, and neither he nor the plaintiffs explain how this model makes it “clear that all members of the class suffered some damage.” Bogosian, 561 F.2d at 455. Defendants for their part essentially ignore the benchmark model in their briefs and do not appear to object to its admissibility. For reasons described below, however, plaintiffs could not carry their burden of showing that antitrust injury could be established by generalized proof, even if all of their statistical evidence were admissible.

b. The Plaintiffs’ Methodology

Beyond admissibility issues there are more fundamental problems with plaintiffs’ approach to proving antitrust injury. To establish antitrust injury, “a plaintiff must show (1) an injury-in-fact; (2) that has been caused by the violation; and (3) that is the type of injury contemplated by the statute.” Blue Tree Hotels Inv., Ltd. v. Starwood Hotels & Resorts Worldwide, Inc., 369 F.3d 212, 220 (2d Cir.2004). Plaintiffs fail to identify generalized proof that could establish the first prong of this element on a classwide basis.

i. Injury-in-fact

There are two basic methods for assessing harm in tying eases.¹⁶ See Visa Check, 280 F.3d at 142. -Under the “tied product” approach, the harm to the plaintiff is “the difference between the price actually paid for the tied product and the price for which the item could have been purchased on the open market.” Id. at 143. Under the “package” approach, the plaintiff is harmed “only to the extent that the plaintiff overpaid for the combination of the tied and tying products.” Id. Although other circuits have split on the issue, the question of which approach more accurately captures the extent of antitrust injury is an open question in this Circuit. Id. at 144.

Because “the focus of ‘antitrust injury’ is on whether the challenged conduct has actually caused harm to the plaintiff,” Blue Tree Hotels, 369 F.3d at 220, the package approach is more consistent with the antitrust injury requirement. See Visa Check, 280 F.3d at 155 (Jacobs, J., dissenting) (concluding that “the statutory text compels th[e] rule” that “the correct measure of injury in fact is the package price of the tying and tied products”). Under the tied product method, a plaintiff is deemed injured, and may recover, to the extent that the price of the tied product has increased, regardless of whether that increase was completely offset, or even exceeded, by a reduction in price of the tying product. Yet such a consumer has suffered no economic harm as a result of the tying; any damages awarded to such a plaintiff are pure windfall. See id. (“[T]he tied buyer is injured only when he is forced to pay above-market prices for the sum of the tying and the tied products.” (citation omitted)). This is a particularly valid concern in this case where it is undisputed that the defendants each subsidize the price of handsets so that the price of handsets does not discourage consumers from becoming wireless customers.

To prevent recovery of windfall damages, “injury resulting from a tie-in must be shown by establishing that payments for both the tied and tying products exceeded their combined market value.” Kypta v. McDonald’s Corp., 671 F.2d 1282, 1285 (11th Cir.1982); see also 2 Phillip E. Areeda, Herbert Hovenkamp & Roger D. Blair, Antitrust Law: An Analysis of Antitrust Principles and Their Application ¶358b, at 464 (2d ed. 2000) (“Areeda”) (same). Plaintiffs have offered no methodology that takes into account the prices paid for the combination of service and handset sold by the defendants.¹⁷ And there is no way to know from Economides’s report whether the increases in handset price, assuming they are real, have not been offset by a reduction in the price of service plans offered by defendants.

Defendants, meanwhile, have argued that the sale of newer handsets has obviated the need for capital investment in the network. See Wireless, 385 F.Supp.2d at 408 (“[Digital service enabled wireless service producers to accommodate more users within any given amount of spectrum, thereby reducing their need to set up additional transmitters or cell sites.”). Newer models of handsets also facilitate the provision of higher quality service. See id. at 409 (“[T]he use of ‘outmoded’ handsets not only affects the quality of that subscriber’s service, but also diminishes the quality of service to other subscribers.”). Because the market for wireless services is competitive, see In re Wireless, 385 F.Supp.2d at 412 (noting repeated determinations by the FCC that the wireless services market is competitive), these savings are presumably passed on to customers in the form of lower service prices. There is thus reason to believe that any increase in handset price has been offset, at least in part, by a concomitant decrease in service price. Indeed, it is generally true that “if a tie causes a buyer to pay more than the market price for the tied product, the buyer is most likely paying less than the price that the seller could profitably charge for the tying product if sold separately.” Visa Check, 280 F.3d at 155 (Jacobs, J., dissenting). Plaintiffs’ inability to identify a methodology to demonstrate that this has not happened precludes a showing of injury on a classwide basis. Cf. Visa Check, 280 F.3d at 143 (noting the district court’s determination that “plaintiffs had made a showing, sufficient to support class certification, that the price of the tied package would decline without the tie”).

Even if the tied product approach were a potential avenue for showing the existence of injury-in-fact, the plaintiffs have not identified a method of demonstrating that injury-on a classwide basis. The specific injury identified by plaintiffs is an increase in the average price of handsets. Using averages that include prices for different products “can lead to serious analytical problems.” ABA Section of Antitrust Law, Econometrics: Legal, Practical, and Technical Issues 220 (2005) (“Econometrics”). For example, the increase in average handset price might reflect the fact that more customers are freely choosing to purchase more sophisticated (and therefore more expensive) types of phones rather than an across-the-board increase in prices. Indeed, defendants’ expert Carlton provides market research suggesting this to be the case.

Plaintiffs have offered no methodology that could prove that the price of a particular model of handset has increased as a result of defendants’ tying and locking practices. Plaintiffs attempt to get around the problems inherent in basing their claim of injury upon average price by invoking a “presumption of impact.” In the appropriate case, a presumption of impact enables proof of injury for each individual class member to be made on a classwide basis. Where the alleged antitrust injury is an overcharge, for example, the presumption of impact reheves plaintiffs of the burden of proving that every individual in fact paid an inflated price.

This presumption does not, however, relieve plaintiffs of proving that the overcharge operated throughout the market for the affected product. Thus, the Third Circuit has observed that “it would be clear that all members of the class suffered some damage” “[i]f the price structure in the industry is such that ... prices at the wholesale level fluctuated within a range which, though different in different regions, was higher in all regions than the range which would have existed in all regions under competitive conditions.” Bogosian, 561 F.2d at 455. Plaintiffs in this action must similarly demonstrate that the alleged overcharge affected prices for all handset models, or at least all categories of handset models. Their offer of generalized proof fails to do this, though, as it looks only at the overall average price of handsets.

Nor have plaintiffs alleged the type of “deliberate interference with market forces” that “[t]he economic laws of supply and demand ... in tandem with the tenets of logic” would predict to have a direct marketwide impact on prices. In re Linerboard Antitrust Litig., 305 F.3d 145, 152 (3d Cir.2002). There is no offer of evidence, for example, that defendants “deliberately cut ... supply” of handsets in order to boost sagging prices. Id. To be sure, plaintiffs have alleged a conspiracy that they claim has interfered with the market in handsets. But the “tenets of logic” do not explain why a conspiracy to add programming locks to handsets, would directly lead to an increase in the wholesale price of handsets (unless, of course, the locks added significantly to production costs, which plaintiffs do not allege). And it is not immediately clear why a conspiracy to eliminate direct sales from manufacturers to consumers would affect the wholesale prices paid by carriers. Because plaintiffs have failed to present generalized proof, based on either evidence or basic logic, that an overcharge operated throughout the market for handsets, it cannot be presumed that every member of the putative class in fact paid an inflated price for a handset.

To the extent that the plaintiffs’ theory rests on their assertion that the conspiracy has forced consumers to purchase handsets with features the consumers do not use, it is worth observing that the plaintiffs do not contend that there was any overcharge for those features. They do not claim that handset manufacturers overcharge their customers for those (or any) features, or that consumers’ handsets are not competitively priced in terms of their quality and features.¹⁸ They instead contend that the average handset price has increased because the defendants have demanded that handset manufacturers create and provide more sophisticated and therefore more expensive handsets than many consumers need or want. As a result, some percentage, indeed perhaps a large percentage, of consumers do not use many of the features that add to the cost of their handsets.

This is not a theory of price inflation, however, that is subject to classwide proof. Those consumers who do use (or desire even if they do not use) the more expensive features have not suffered any injury-in-faet. Indeed, under the plaintiffs’ theory, these consumers in all likelihood benefitted since mass production of the more sophisticated handsets has probably lowered their prices. Moreover, while injury may be proven on a classwide basis when a defendant has used its market power to charge a supracompeti-tive price for every individual tied product, see, e.g., Key, 528 F.2d at 12 n. 11, the plaintiffs’ unwanted feature theory would find antitrust injury even if each individual handset were priced appropriately in terms of its quality and features as long as there has been an increase in the average price. Such proof, even though made on a classwide basis, cannot establish injury to each individual class member.

ii. Causation

An antitrust plaintiff need only show that the defendant’s “violation was at least a material cause of the plaintiffs injury.” Blue Tree Hotels, 369 F.3d at 220 (quoting Bohack Corp. v. Iowa Beef Processors, Inc., 715 F.2d 703, 710-11 (2d Cir.1983)). But “a purchaser may recover only for the price increment that ‘flows from’ the distortion of the market caused by the monopolist’s anticompetitive conduct.” Berkey Photo, Inc. v. Eastman Kodak Co., 603 F.2d 263, 297 (2d Cir.1979). Thus, when there are multiple causes for a plaintiffs injury, “the plaintiff must provide the jury with a reasonable basis for separating the impact of the violation from other factors.” Areeda, ¶338, at 319.

The expert reports offered by plaintiffs to support their motion for class certification do not seem to provide a methodology for disaggregating the effect on prices due to defendants’ agreement to tie and lock handsets from any effects due to other causes, such as changes in quality or customer preference for a feature-rich handset. It is therefore unclear at this point how the plaintiffs will be able to establish causation at trial. See Allegheny Pepsi-Cola Bottling Co. v. Mid-Atlantic Coca-Cola Bottling Co., Inc., 690 F.2d 411, 415 (4th Cir.1982). Further consideration of this element of antitrust injury is unnecessary at this point, however. To the extent that plaintiffs are able to prove causation, they will do so on a classwide basis with generalized proof. That is all that the plaintiffs need to show at this point in the litigation.

iii. Type of injury

The third feature of antitrust injury is that it be an “injury of the type that antitrust laws were intended to prevent.” Atl. Richfield Co. v. USA Petroleum Co., 495 U.S. 328, 334, 110 S.Ct. 1884, 109 L.Ed.2d 333 (1990) (quoting Brunswick, 429 U.S. at 489, 97 S.Ct. 690). This requirement “ensures that a plaintiff can recover only if the loss stems from a competition -¿reducing aspect or effect of the defendant’s behavior.” Id. at 344, 110 S.Ct. 1884.

Plaintiffs face significant problems with this element of antitrust injury. They do not dispute that handset manufacturers compete with each other in order to offer defendants the highest-quality handsets for the lowest prices. Nor do they dispute that defendants vigorously compete with each other for customers of their wireless services and that one of the strategies they use is to offer handsets at the lowest possible price. Given this level of competition in both the wholesale and retail markets for handsets and the retail market for wireless service, it could be difficult for plaintiffs to establish that any loss they have suffered stems from a competition-reducing aspect of defendants’ behavior. Indeed, summary judgment was granted to the defendants in Brook because plaintiffs were unable to show that any defendant’s individual practices had anticompetitive effects in the handset market. See In re Wireless, 385 F.Supp.2d at 424-31.

Nevertheless, whether the conspiracy alleged in the Freeland complaint has had such anticompetitive effects is a separate question, and one that is subject to proof by classwide evidence. Moreover, the existence of such effects goes to the merits of plaintiffs’ claims and is not properly resolved on a motion for class certification.

iv. Burden of proof for antitrust injury

Plaintiffs’ final argument, relied on heavily in their sur sur reply, is that Economides’s report meets the standard set out by the Second Circuit in New York v. Julius Nasso Concrete Corp., 202 F.3d 82 (2d Cir.2000). Nasso addressed the burden imposed on the government plaintiff to show the amount of damages proximately caused by the defendants’ violation. Id. at 88. The Court of Appeals recognized that, “if causation is established, ... the State’s burden of proving antitrust damages is not as rigorous as in other types of cases.” Id. It observed, “[djamage issues in antitrust cases are rarely susceptible of the kind of concrete, detailed proof of injury which is available in other contexts” and more particularly that establishing market prices in the “but for” world free of anticompetitive conduct could be made difficult by the “dearth of market information unaffected by the collusive actions of the defendants.” Id.

Nasso is of no assistance to the plaintiffs in this case, though. Nasso does not affect the burden to prove antitrust injury, but only the quality of evidence required to show the amount of damages. Even under that relaxed standard, a plaintiff must be able to offer “some relevant data from which the [factfinder] can make a reasonable estimated calculation of the harm suffered.” Id. at 89. As described above, plaintiffs’ generalized proof provides neither a method of establishing that any economic harm was suffered at all nor a method for preventing recovery for price increases not attributable to defendants’ tying practices. These limitations would prevent a factfinder from making even a reasonable estimation of harm suffered. Moreover, the difficulties with plaintiffs’ theory of injury do not principally stem from a dearth of market information due to the collusive activities of the defendants. The problems here owe more to conceptual failings than to any lack of specific evidence. Finally, and perhaps most fundamentally, Nasso concerns the standard that applies to a decision on the merits of an antitrust case; it in no way affects the “rigorous analysis” a court must undertake to decide whether plaintiffs have made a proper showing of predominance pursuant to Rule 23(b)(3). Plaintiffs’ generalized proof simply does not establish that injury could be proven on a classwide basis, and Nasso does not relieve them of that requirement.

2. Conspiracy and Restraint of Trade Counts

Plaintiffs allege four separate counts of antitrust violation in their complaint: conspiracy to restrain trade, conspiracy to monopolize, tying, and contracts in restraint of trade. As indicated above, antitrust injury must be proven with respect to each one of these claims. Thus, if plaintiffs cannot establish that they will be able to prove antitrust injury through common evidence, then individual questions will predominate with respect to all four of their claims.

The remaining elements of the non-tying counts may all be proven on a class-wide basis. To prove a conspiracy (or contract) in restraint of trade in violation of Section 1 of the Sherman Act, a plaintiff must prove two elements: “(1) a combination or some form of concerted action between at least two legally distinct economic entities that (2) unreasonably restrains trade.” Geneva Pharms., 386 F.3d at 506. A Section 2 conspiracy to monopolize claim requires “(1) proof of a concerted action deliberately entered into with the specific intent to achieve an unlawful monopoly, and (2) the commission of an overt act in furtherance of the conspiracy.” AD/SAT, Div. of Skylight, Inc. v. Associated Press, 181 F.3d 216, 233 (2d Cir.1999). All of these elements focus on either the behavior of the defendants or the conditions of the markets in which they operated; none requires individualized proof. See, e.g., 6 Alba Conte & Herbert Newberg, Newberg on Class Actions §§ 18:28 & 18:29 (4th ed.2002) (noting that allegations of antitrust conspiracies generally establish predominance of common questions). Defendants do not dispute that these issues may all be proven on a classwide basis.

3. Tying

The Second Circuit requires proof of five elements to establish an illegal tie:

first, a tying and a tied product; second, evidence of actual coercion by the seller that forced the buyer to accept the tied product; third, sufficient economic power in the tying product market to coerce purchaser acceptance of the tied product; fourth, anticompetitive effects in the tied market; and fifth, the involvement of a “not insubstantial” amount of interstate commerce in the “tied” market.

Hack, 237 F.3d at 86. Plaintiffs argue, and defendants do not contend otherwise, that the first, third, fourth, and fifth elements are susceptible to proof by generalized evidence. These four focus on conditions of the markets in which the defendants operate and do not require individualized proof. The sole dispute is whether coercion, required by the second element, must be proven on an individual basis or may be established by common evidence.

As Justice Marshall explained, the Second Circuit has “consistently made clear [that] ‘unless the buyer can prove that it was the unwilling purchaser of the allegedly tied products, actual coercion has not been established and a tying agreement cannot be found to exist.’ ” Trans Sport, Inc. v. Starter Sportswear, Inc., 964 F.2d 186, 192 (2d Cir.1992) (Marshall, J.) (quoting Unijax, Inc. v. Champion International, Inc., 683 F.2d 678, 686 (2d Cir.1982)); see also Am. Mfrs. Mut. Ins. Co. v. Am. Broad.-Paramount Theatres, Inc., 446 F.2d 1131, 1137 (2d Cir. 1971) (“[T]here can be no illegal tie unless unlawful coercion by the seller influences the buyer’s choice.”). The essence of an illegal tying arrangement is the “seller’s exploitation of its control over the tying product to force the buyer into the purchase of a tied product that the buyer either did not want at all, or might have preferred to purchase elsewhere on different terms.” Jefferson Parish, 466 U.S. at 12, 104 S.Ct. 1551. Evidence of aggressive salesmanship, even “strong persuasion, encouragement, or cajolery to the point of obnoxiousness,” will not suffice to establish this element of a tying claim. Unijax, 683 F.2d at 685 (citation omitted). “[A] reasonable trier of fact must be able to find actual coercion by the seller that in fact forced the buyer to purchase the tied product,” and such a finding is proper only where the defendant “goes beyond persuasion and conditions [the plaintiffs] purchase of one product on the purchase of another product.” Trans Sport, 964 F.2d at 192 (citation omitted).

Because the ultimate question is whether the purchase of the tying product was in fact conditioned on the purchase of the unwanted tied product, “[a]n unremitting policy of tie-in” will constitute coercion. Hill v. A-T-O, Inc., 535 F.2d 1349, 1355 (2d Cir. 1976). The Second Circuit has found such a policy, and therefore coercion, subject to generalized proof in two circumstances. The first is when the policy of conditioned sales is admitted to by the defendant. See id. (noting that “defendants admit to a policy of never offering the [tying product] separately from the [tied product]”); see also Hardy v. City Optical Inc., 39 F.3d 765, 770-71 (7th Cir.1994) (finding “a blanket policy” where the defendants “made it clear from the start of this ease — and their counsel repeated at oral argument — that they ... interpret [a state] regulation as forbidding them” from providing the tying product without the tied product). The second is when a contractual provision applicable to all members of the class requires the tie. See Visa Check, 280 F.3d at 136 (affirming the district court’s conclusion that “coercion was ... amenable to proof on a class-wide basis because the contractual provision to which all class members were subject ... would establish the requisite coercion”); see also Response of Carolina, Inc. v. Leasco Response, Inc., 537 F.2d 1307, 1327 (5th Cir.1976) (“[W]here one product is sold on the express contractual condition that the purchaser purchase supplies for that product in the future only from the seller, the purchaser’s future course of action is limited by that contract, and coercion is found in the agreement itself.”). “[W]hen a class action tie-in claim is made, without basis in an express agreement,” or a defendant’s admission to a tying arrangement (which effectively dissolves plaintiffs’ burden of proof on this element), “proof of actual coercion on an individual basis is necessary to prove the existence of a tie.” Bo-gosian, 561 F.2d at 451.

The plaintiffs have not offered to prove the existence of a tie through a common contractual provision to which all class members are subject. To the contrary, plaintiffs concede that “[s]ome carriers, such as Verizon and T-Mobile, include fíne print in their subscriber agreements suggesting that subscribers are not required to purchase handsets from the carrier.”

Instead, the plaintiffs offer three types of evidence to support their assertion that defendants had an unremitting policy of coercing consumers to purchase handsets from them and that such coercion can be proved through generalized proof: first, an admission from defendant Sprint that it has never activated a different provider’s handset on its network and that no other provider has activated a Sprint handset on its own network; second, concessions by all defendants that they do not offer a “service-only” plan; and third, statistics indicating that nearly 100% of customers who purchase service plans also purchase handsets from the same defendant.

This evidence does not establish that coercion could be proven on a classwide basis. Sprint did not admit that it has refused to sell wireless service unless a customer bought a handset from Sprint, as opposed to a third party source.¹⁹ The admission that it in fact had never activated a different provider’s handset is evidence that Sprint customers do not purchase their handsets from other service providers, but it does not establish that all Sprint customers purchased their handsets from Sprint or that those who did purchase handsets from Sprint were forced to do so. And of course Sprint’s admission could not be used to establish coercion for the customers of any other defendant. Similarly, the concessions by all defendants, as summarized by a chart in Carlton’s expert report, that they do not offer “service-only” plans does not mean that they all refuse to sell only wireless service to a prospective customer. Rather, the defendants admit that they charge the same price for the service plan regardless of whether the customer buys a handset at the same time.²⁰

This leaves plaintiffs with only the statistical evidence by which to prove coercion on a classwide basis. But “proof that a high percentage of ... customers” have purchased the allegedly tied product from a defendant in a tying suit is “insufficient, standing alone, to demonstrate coercion.” Cm. Petrolera Caribe, Inc. v. Avis Rental Car Corp., 735 F.2d 636, 638 (1st Cir.1984); see also Ungar v. Dunkin’ Donuts of America, Inc., 531 F.2d 1211, 1225 n. 14 (3d Cir. 1976) (explaining that “evidence of acceptance by a large number of buyers of the same type of offer ... alone would not suffice to establish coercion prima facie”); Areeda ¶1756b, at 299-300 (“A high percentage, even 100 percent, of bundled sales does not itself indicate that two products have been tied together.”). The reason, of course, is that “[b]uyers often find package sales attractive,” Jefferson Parish, 466 U.S. at 12, 104 S.Ct. 1551, and these statistics simply cannot establish that any particular member of the putative class (or even any member at all) was an “unwilling purchaser” of a handset or that a defendant made the purchase of a handset a condition for that member’s purchase of wireless service.

The plaintiffs have not established that coercion, an essential element of a tying claim under Second Circuit law, can be proven on a classwide basis. As a result, individual questions will predominate with respect to their tying claim and class certification is inappropriate. See 5 Moore’s Federal Practice § 23.45[5][ c] (3d ed. 2006) (“In the absence of a common contractual provision, proof of a tie-in is an individual question, and individual questions will predominate.”).

4. Conclusion

The predominance inquiry evaluates whether a proposed class is cohesive enough to justify adjudication through representation. Because antitrust injury, an essential element for each of plaintiffs’ claims, and coercion, a critical component of the tying claims, are not susceptible to generalized proof, certification of the class action is not appropriate.²¹ The issues that are subject to generalized proof are not more substantial than those that are subject to individualized proof.

D. Rule 23(b)(2)

A class may be certified under Rule 23(b)(2) if the defendant has acted or refused to act on grounds generally applicable to the class, “thereby making appropriate final injunctive relief or corresponding declaratory relief with respect to the class as a whole.” Rule 23(b)(2), Fed.R.Civ.P. When presented with a motion for class certification under Rule 23(b)(2) of a claim seeking injunctive relief and non-incidental money damages, a court may allow certification if it finds that “(1) the positive weight or value to the plaintiffs of the injunctive or declaratory relief sought is predominant ... and (2) class treatment would be efficient and manageable, thereby achieving ... judicial economy.” Robinson v. Metro-North Commuter R.R., 267 F.3d 147, 164 (2d Cir.2001) (citation omitted). The assessment of whether injunctive relief predominates requires an ad hoc balancing, but a court should at a minimum, satisfy itself of two things: first, that reasonable plaintiffs would bring the suit to obtain injunctive relief even in the absence of a possible monetary recovery; and second, that the injunctive relief sought would be both reasonably necessary and appropriate were the plaintiffs to succeed on the merits. Id.

Plaintiffs have not demonstrated that the value of injunctive relief is predominant. To be sure, their experts offered conclusory statements that the value of injunctive relief exceeded the value of damages sought, but there is no way to evaluate the accuracy of these statements. Plaintiffs simply do not explain why the injunctive relief would be so valuable to existing customers. It is thus unclear why a reasonable plaintiff would seek injunctive relief if she could not recover the amount she has overpaid for her handset.

There is a further, more fundamental problem. Proof of antitrust injury is required of a plaintiff seeking injunctive relief from a Sherman Act violation. Cargill, Inc. v. Monfort of Colorado, Inc., 479 U.S. 104, 113, 107 S.Ct. 484, 93 L.Ed.2d 427 (1986).²² Yet plaintiffs have provided no methodology by which antitrust injury could be proven on a classwide basis. Antitrust injury will thus have to be proven on an individual basis for a class member to be afforded injunctive relief. Coercion must also be proven individually. As a result, even if the named plaintiffs succeed on the merits of their claims, only they would be entitled to injunctive relief. The claims of the rest of the class members would remain unproven, and “final injunctive relief or corresponding declaratory relief with respect to the class as a ivhole, ” would thus be inappropriate. Fed.R.Civ.P. 23(b)(2)(emphasis added).

The Fifth, Seventh, and Eleventh Circuits have all held that certification under Rule 23(b)(2) is improper where an element of liability must be proven on an individualized basis. See Heffner v. Blue Cross & Blue Shield of Ala., Inc., 443 F.3d 1330, 1344-45 (11th Cir.2006) (concluding that “individual reliance issues weigh against Rule 23(b)(2) certification”); In re Allstate Ins. Co., 400 F.3d 505, 507-08 (7th Cir.2005) (vacating (b)(2) certification where “more than a thousand individual hearings will be necessary in order to determine which members were really forced to quit and which quit voluntarily”); Bolin v. Sears, Roebuck & Co., 231 F.3d 970, 978 (5th Cir.2000) (holding that “the individual findings of reliance necessary to establish RICO liability ... preclude ... (b)(2) certification”). Their reasoning is persuasive and consistent with the plain language of Rule 23(b)(2).

Denial of certification in these circumstances is also consistent with the Second Circuit’s requirement that class treatment be efficient and manageable. There is nothing particularly efficient or manageable about a class action that would require upwards of 100 million individual hearings before a class-wide injunction could issue. “[G]iven all the facts and circumstances of this] case,” especially those highlighted above, certification under Rule 23(b)(2) is not appropriate. Robinson, 267 F.3d at 164 (citation omitted).

Conclusion

Plaintiffs have not carried their burden of establishing that this action meets the requirements of Rule 23. Although the four threshold requirements of Rule 23(a) are satisfied, the action is not maintainable under either 23(b)(2) or 23(b)(3). The motion for class certification is denied.²³ For reasons described above, the plaintiffs’ motion to strike is also denied.

1 . The defendants named in the complaint are AT & T Cellular Services, Inc.; Cellular Telephone Company; AT & T Wireless Services, Inc.; AB Cellular LA, LLC; AT & T Corporation (collectively "AT & T Wireless” or "AT & T defendants"); Cingular Wireless LLC; Cingular Wireless Spectrum Sub B LLC; Pacific Telesis Mobile Services, LLC (collectively "Cingular” or "Cin-gular defendants”); Sprint Corporation; Sprint Spectrum LP; Sprint PCS License LLC (collectively “Sprint” or "Sprint defendants”); Célico Partnership; Los Angeles SMSA Limited Partnership; Los Angeles SMSA Limited Partnership-Airtouch Cellular (collectively "Verizon” or “Verizon Wireless defendants”); T-Mobile USA, Inc.; Voicestream GSM II, LLC; Omnipoint N.Y. MTA License (collectively "T-Mobile” or "T-Mobile defendants”). Although the complaint and plaintiffs’ papers refer to the defendants as the five largest wireless carriers in the United States, two of them (AT & T Wireless and Cingular) merged in 2004. See In re Wireless Tel. Servs. Antitrust Litig., 385 F.Supp.2d 403, 411 (S.D.N.Y.2005).

2 . Hereinafter, all citations and references to In re Wireless shall refer to the Summary Judgment Opinion.

3 . The first motion for class certification (in the Brook phase of this litigation) was submitted at approximately the same time as the defendants' motion for summary judgment. Because sum-maiy judgment was granted, it was unnecessary to rule on the motion for class certification. See In re Wireless, 385 F.Supp.2d at 408 n. 4.

4 . This fact puts the Court in the somewhat unusual position of having to refrain from evaluating the plaintiffs’ case on the merits while deciding the motion for class certification almost a year after it has undertaken a careful examination of much of their evidence, and on the eve of the submission of summary judgment motions in the Freeland action.

5 . There is no allegation of coordination among defendants in setting specifications for features other than handset locks. Thus, each defendant appears to be able independently to dictate the features of handsets to be activated on its own network.

6 . Plaintiffs' expert Professor Nicholas Econom-ides identifies a litany of anticompetitive effects of the defendants' locking and tying practices that he claims has injured members of the class. The plaintiffs do not attempt in their briefs, though, to establish antitrust injury by reference to any of these other effects. Nor could they. An antitrust plaintiff in an action for damages lacks antitrust standing if she cannot establish an antitrust injury “that translates into reasonably quantifiable damages.” Daniel v. Am. Bd. of Emergency Med., 428 F.3d 408, 438 (2d Cir. 2005). Plaintiffs in this action have not offered any methodology to demonstrate that damages related to the other injuries identified by Eco-nomides are reasonably quantifiable. As a result, they lack antitrust standing to seek redress for any of these other claimed injuries.

Similarly, the complaint alleges that the alleged conspiracy has "cause[d] artificially elevated market prices for cellular and PCS services," but the plaintiffs do not invoke inflated service prices in their briefs as an antitrust injury nor do they offer any methodology by which an increase in service prices could be proven on a classwide basis. See, infra, at 143 n. 7.

7 . A second expert for plaintiffs, Gary French, proposed a method to calculate damages attributable to overpayment for service by customers who continue to subscribe to wireless service after their initial contract period has expired. While plaintiffs have included his estimate of total damages in their own total figure, they did not rely on this part of French's report anywhere in their arguments that injury could be established by generalized proof. This omission might reflect an implied concession that French’s report does not show that injury can be proven on a classwide basis. By French’s own admission, only 7.5% to 21% of class members have suffered this type of damage. French's estimate thus falls far short of "posit[ing] class-wide injury resulting from every single class member’s overpaying” for service.

8 . Neither Economides nor the plaintiffs explain how the elimination of low-cost handset models from the market is substantively any different from the incorporation of advanced (and expensive) features into all new models brought into the market.

9 . Although Economides actually performs three separate regression analyses, each one includes a BLS index as the single explanatory variable. Economides intended for each of these indices to function independently as a proxy for factors affecting the wholesale costs of handsets. Thus, while the estimates generated by the regressions differ somewhat, no one regression was intended to generate substantively more (or different) information than any other. This is also why Eco-nomides bases his forecast on the results of a single regression analysis. For these reason, this Opinion will refer to Economides's report as if it consisted of a single regression analysis.

10 . "Dichotomous variables of this kind are called 'dummy variables' because their values are conventional rather than inherent.” Craik v. Minn. St. Univ. Bd., 731 F.2d 465, 508 (8th Cir.1984).

11 . Thus, when the Court of Appeals upheld the exclusion of a regression analysis in Bickerstaff, it discussed only the proponent's failure. See Bickerstaff, 196 F.3d at 449. The court neither cited Sobel nor implied any burden on the opponent of the evidence. Id.; see also Hollander v. Am. Cyanamid Co., 172 F.3d 192, 202-04 (2d Cir.1999) (upholding trial court's decision that plaintiff's flawed expert report was inadmissible); Raskin v. Wyatt Co., 125 F.3d 55, 66-68 (2d Cir.1997) (same).

12 . Economides similarly failed to investigate his assertion that camcorders had undergone a similar shift from analog to digital without an effect on average price. He offers no evidence that would show, for example, that digital and analog camcorders were the same price, or that the shift from digital to analog occurred at the same rate for camcorders as handsets. Carlton, on the other hand, cited a BLS study suggesting that digital camcorders were in fact more expensive than analog camcorders and that analog camcorders were available for sale after analog handsets had been phased out. If the shift to digital technology in camcorders occurred more gradually, such that the more expensive digital camcorders were replacing cheaper analog models at a slower rate, then the effect of the shift on overall average price would reasonably be expected to be less pronounced.

13 . "When two or more variables are highly, but not perfectly, correlated' — that is, when there is multicollinearity — the regression can be estimated, but some concerns remain. The greater the multicollinearity between two variables, the less precise are the estimates of individual regression parameters.” Federal Judicial Center, Reference Manual on Scientific Evidence 197 (2d ed. 2000) ("Reference Manual”); see also id. at 189 n. 29 ("A more precise estimate of a parameter is an estimate with a smaller standard error.”). The plaintiffs do not contend that there is a perfect correlation between the variable they have selected (the price of component or comparator goods) and the shift to digital technology in handsets.

14 . The ABA Section of Antitrust Law suggests that “if the purpose of the regression exercise is to forecast (e.g., to forecast but-for prices into some 'tainted' period based upon a regression model fitted only using data from the benchmark period), multilcollinearity is of little concern.” ABA Section of Antitrust Law, Econometrics: Legal, Practical, and Technical Issues 22 (2005) (“Econometrics”). Economides's regression (at least those without the dummy variables) were apparently designed for precisely this purpose.

The Section of Antitrust Law further suggests that simply excluding a multicollinear variable, as Economides did, may be an inappropriate method of addressing the problem. "[T]he effect of omitting a statistically significant variable that by definition is correlated with another dependent variable violates the statistical assumption that the error terms are not correlated with the independent variables.” Id. at 68. Neither the defendants nor their experts made either of these arguments, however.

15 . See BLS Handbook, supra, Ch. 14 at 3; id., Ch. 15 at 156, available at http://www.bls.gov/ opub/hom/pdf/homchl5.pdf. The BLS Handbook goes on to describe the use of hedonic regression analysis to capture the impact of changing characteristics on product price change. See BLS Handbook, supra, Ch. 14 at 4 ("Hedonic regressions estimate the functional relationship between the characteristics embodied in the products in a market and the products’ prices. Such regressions yield estimates of 'implicit prices’ for specified product characteristics that may be used to value the quality improvement resulting from changes in the various characteristics embodied in a product. The value of the quality improvement can then be removed from the reported price change to obtain a measure of the pure price change...."). The parties did not discuss whether such a method would be appropriate in this case.

16 . Plaintiffs offer the Economides report to establish antitrust injury in their conspiracy and restraint of trade claims as well as their lying claim. The injury claimed, though, remains the same for all: an increase in price of one product sold as part of a bundle. Because such bundling practices are most regularly challenged in tying claims, it is from the tying case law that guidance must be sought in an attempt to evaluate the injury claimed by plaintiffs here. The principles gleaned from those cases are equally applicable to consideration of plaintiffs’ non-tying claims.

17 . Why they have not done so is unclear, particularly in light of their frequent complaint that it was "extremely difficult” to determine the price of the handset alone.

18 . Economides has explicitly disclaimed any opinion that the defendants' practices have allowed handset manufacturers to enrich themselves at the defendants’ expense.

19 . Plaintiffs do not dispute defendants’ contention that Sprint service utilizes unique network technology and a different software platform than the other carriers. Nor do they allege that Sprint chose its unique network technology and software platform "for the purpose of tying the products, rather than to achieve some technologically beneficial result." Response of Carolina, Inc. v. Leasco Response, Inc., 537 F.2d at 1330; see also In re Wireless, 385 F.Supp.2d at 408 (noting that a number of digital technologies were introduced in the 1990s following the FCC's decision to decline to maintain a technological standard for digital service). Thus, to the extent that a consumer must purchase a Sprint handset (though not necessarily one purchased from Sprint itself as part of a tied transaction) in order to obtain complete Sprint service, this reflects the benign reality that a handset formatted to operate on a different carrier’s network will not be fully functional on the Sprint network. Any company policy requiring the use of Sprint handsets on the network is basically a redundancy.

20 . Defendants admit to subsidizing the purchase price of handsets in an effort to attract new customers by offering low up-front costs for starting wireless service. These subsidies are then recovered, at least in part, through service fees over time. This creates the possibility that, even if the purchase of a handset is not a condition of purchase of wireless service, the payment of a handset subsidy is. To the extent this is a tying arrangement, it is not the tie alleged by the plaintiffs. And even if it were, this does not constitute actual coercion in which consumers are forced to purchase their handsets from their service providers. The practice of subsidization is a silent condition and does not show that all consumers reasonably understood that handsets may not be obtained elsewhere.

21 . Because common questions do not predominate over individual questions, it is unnecessary to determine whether a class action would be a superior means of adjudication of this case. Suffice it for now to observe that the individual hearings on the issues of antitrust injury and coercion would be sufficiently burdensome that, even if common questions did predominate, there would be serious difficulties with the management of the class action. See 5 Moore’s Federal Practice, at § 23.46[2][e] ("If ... there is a need for a significant number of separate trials, the class action is unmanageable.”).

22 . It is true that a plaintiffs may bring an action for injunctive relief pursuant to the Clayton Act for a Sherman Act violation without suffering an actual injury. Kruman v. Christie’s Intern. PLC, 284 F.3d 384, 397 (2d Cir.2002), abrogated on other grounds, F. Hoffmann-La Roche Ltd. v. Em-pagran S.A., 542 U.S. 155, 124 S.Ct. 2359, 159 L.Ed.2d 226 (2004). The plaintiff "need only demonstrate a significant threat of injury from an impending violation of the antitrust laws or from a contemporary violation likely to continue or recur.” Zenith Radio Corp. v. Hazeltine Research, Inc., 395 U.S. 100, 130, 89 S.Ct. 1562, 23 L.Ed.2d 129 (1969). Nevertheless, the Clayton Act does not authorize "an injunction against a threatened injury for which he would not be entitled to compensation if the injury actually occurred," and the plaintiff must therefore still demonstrate that the threatened injury is an antitrust injury. Cargill, 479 U.S. at 112, 107 S.Ct. 484. In this case, the injury complained of has already been suffered, and for reasons described in Subsection C.l.b above, the plaintiffs in this case have failed to demonstrate that they can prove antitrust injury through generalized proof.

23 . The AT & T, Cingular, Sprint, and T-Mobile defendants separately have renewed their arguments made in connection with the Brook motion for class certification that certification of a class including their customers would be improper because of arbitration agreements contained in their service contracts. Because the plaintiffs’ motion for certification is denied in its entirety, it is unnecessary to address this argument.