A & M SUPPLY CO. v. Microsoft Corp.

Whitbeck, C.J. In this action under the Michigan Antitrust Reform Act (MARA), MCL 445.771 et seq., plaintiff A&M Supply Company alleges that defendant Microsoft Corporation illegally monopolized the personal computer software market in Michigan, harming A&M and many other residents of this state who purchased certain Microsoft products. In August 2001, the trial court entered an opinion and order certifying this matter as a class action. Microsoft appeals that order by leave granted. We reverse and remand. I. legal overview Generally speaking, there are two types of private antitrust lawsuits. The first type of action is a direct purchaser action in which the plaintiff purchases a product directly from the purported monopolist and *583 sues for the injuries sustained. 1 The second type of action is an indirect, or “pass-on,” action in which the plaintiff is one or more steps removed from the direct purchaser, essentially purchasing the product after it has changed hands through middlemen. 2 34This case is a pass-on lawsuit. That state law forms the basis for this suit is no coincidence. In Hanover Shoe, Inc v United Shoe Machinery Corp 3 and Illinois Brick Co v Illinois, 4 the United States Supreme Court barred indirect purchaser suits in federal courts. However, the Court did not foreclose states from allowing indirect purchaser actions. Accordingly, in 1984, 5 6 the Michigan Legislature passed an Illinois Brick repealer law when it enacted MARA subsection 8(2), which states: Any other person threatened with injury or injured directly or indirectly in his or her business or property by a violation of this act may bring an action for appropriate injunctive or other equitable relief against immediate irreparable harm, actual damages sustained by reason of a violation of this act, and, as determined by the court, interest on the damages from the date of the complaint, taxable costs, and reasonable attorney’s fees. If the trier of fact finds that the violation is flagrant, it may increase recovery to an amount not in excess of 3 times the actual damages sustained by reason of a violation of this act.[ 61 *584 The dispositive issue in this case revolves around injury, which MARA subsection 8(2) makes so critical in providing a basis for recovering “actual damages.” Microsoft argues that, in an indirect purchaser case, a plaintiff who asserts that each putative class member actually sustained the “fact of injury” must demonstrate two factors with common proof. Under this analytical construct, the plaintiff must first prove that the prices the defendant charged the direct purchasers were consistently higher than the prices it would have charged in a competitive environment. We label this an “overcharge” requirement. Second, the plaintiff must prove that the overcharge, or some portion thereof, passed through the chain of distribution to indirect purchasers. We label this a “pass-on” requirement. We agree with Microsoft that proving overcharge and pass-on are essential to succeeding in an indirect purchaser suit under mara, and therefore adopt this construct. Without using this terminology, the trial court, when it certified the class action, concluded that A&M had met its burden with respect to the overcharge and pass-on requirements. As we explain in more detail, the trial court’s approach to the overcharge requirement may have been correct. However, that is an issue we need not reach because the trial court erred in concluding that A&M satisfied the pass-on requirement; this element of the trial court’s reasoning requires us to reverse and remand. R. BASIC FACTS AND PROCEDURAL HISTORY James Barnard originally filed this case in December 1999 in Livingston County. Barnard sought to hold Microsoft liable to consumers who purchased Win *585 dows operating systems and Internet Explorer software for alleged damages attributable to Microsoft’s monopolistic posture and antitrust violations in Michigan. The complaint consisted of a single count alleging that Microsoft had violated mara. The case was first assigned to Livingston Circuit Judge Daniel Burress. In an opinion and order entered on May 4, 2000, Judge Burress certified a class comprised of all individuals who purchased, leased, or licensed a copy of Windows 95 or Windows 98 from an entity other than Microsoft. 7 Microsoft then filed a motion in the Michigan Supreme Court for superintending control. This motion prompted the Michigan Supreme Court to issue Administrative Order No. 2000-5, 8 which transferred all cases brought in this state’s trial courts against Microsoft for violating MARA to the Wayne Circuit Court to coordinate both pretrial and trial proceedings. As a result, this case moved from the Livingston Circuit Court to the Wayne Circuit Court, which is the trial court that made the decision at issue in this appeal. In the trial court, Microsoft moved to set aside Judge Burress’ certification order, arguing that there had been no discovery before that judge decided the motion and because, in Microsoft’s view, the decision to certify the class was itself erroneous. The trial court granted the motion, vacating the previous certification order. The trial court then gave the parties *586 time to conduct further discovery and granted Barnard leave to amend his complaint. When Barnard filed his amended complaint, he added A&M as the class representative because A&M had allegedly purchased more Microsoft products than Barnard, including Microsoft NT, Microsoft Windows 95, Microsoft Windows 98, Microsoft Word, Microsoft Excel, Microsoft Office, Microsoft Outlook, and Internet Explorer. Barnard had purchased only Internet Explorer and one version of the Windows operating system. In the amended complaint, A&M claimed that Microsoft had violated mara in three ways. The first two counts in the complaint implicated MARA § 3, which makes unlawful “[t]he establishment, maintenance, or use of a monopoly, or any attempt to establish a monopoly, of trade or commerce in a relevant market by any person, for the purpose of excluding or limiting competition or controlling, fixing, or maintaining prices.” 9 The third count relied on mara § 2, which simply declares that “[a] contract, combination, or conspiracy between 2 or more persons in restraint of, or to monopolize, trade or commerce in a relevant market is unlawful.” 10 Underlying these three counts was A&M’s theory that, since the mid-1980s, Microsoft had used monopolistic and anticompetitive means to dominate production of operating systems and software for personal computers. According to A&M, Microsoft, through its predatory conduct, was able to eliminate its competition in the operating systems market. As a *587 result, A&M claimed, consumers have had no viable choices for an alternative operating system for their personal computers since 1994. A&M also alleged that Microsoft engaged in the same conduct regarding its word processing, spreadsheet, and office suite software. A&M asserted that Microsoft’s behavior allowed it to raise its prices for its products above competitive pricing levels in Michigan, creating artificially high prices consumers cannot escape. As support for these allegations, A&M cited findings and conclusions made against Microsoft in other federal and state antitrust cases. In March 2001, A&M became the sole named plaintiff. A&M subsequently filed a motion asking the trial court to certify this matter as a class action. In its motion to certify the class, A&M asked the trial court to certify a single class for consumers who acquired operating systems software produced by Microsoft (MS-DOS, Windows 95, and Windows 98), dropping its previous request to certify subclasses for consumers who purchased Microsoft’s word processing, spreadsheet, and office suite software. Therefore, under the language A&M proposed, the class would consist of [a]ll persons and entities who acquired for their own use, and not for further selling, leasing or licensing, a license in Michigan from Microsoft, an agent of Microsoft or an entity under Microsoft’s control, for an Intel-compatible PC version of MS-DOS, Windows 95, upgrades to higher MS-DOS versions, upgrades to or of Windows 95, Windows 98, upgrades to or of Windows 98, or other software products in which MS-DOS or Windows have been incorporated in full or part (“Microsoft operating systems software”) at any time during the class period. *588 On August 21, 2001, about a month after oral arguments, the trial court issued a written opinion and order certifying the class. In explaining its decision, the trial court focused on the disputed issue, which it determined to be whether A&M could show a common method or means of proving injury and damages to the class members. The trial court concluded that the three methods A&M advanced for demonstrating the injury Microsoft’s actions allegedly inflicted on the class as a whole were sufficient to certify the class because A&M was required only to demonstrate this factor through common proof that may have affected the class. The trial court indicated that any lack of common proof of damages for class members was not fatal to certification. Finally, the trial court concluded that handling this matter as a class action was superior to other available methods. EL STANDARD OF REVIEW This Court reviews a trial court’s decision to certify a class to determine whether the trial court clearly erred. 11 “Generally speaking, factual findings are clearly erroneous if there is no evidence to support them or there is evidence to support them but this Court is left with a definite and firm conviction that a mistake has been made.” 12 IV. UNITED STATES v MICROSOFT CORP As is well known, the federal government and several states brought antitrust proceedings against Mic *589 rosoft in the United States District Court for the District of Columbia in the 1990s, eventually consolidating most, but not all, 13 suits under the name United States v Microsoft Corp. 14 The federal district court’s 412 individual factual findings in that case are too numerous to recount, even summarily. 15 However, with respect to the legal questions involved in the case, the federal district court determined that Microsoft violated §§ 1 and 2 of the Sherman Anti-Trust Act, 16 and numerous analogous state antitrust laws, including MARA §§ 2 and 3. 17 In June 2000, after Microsoft and the many plaintiffs failed to negotiate a mutually agreeable settlement to remedy Microsoft’s alleged antitrust violations, 18 the federal district court ordered sweeping changes in Microsoft, including a multiphase process of dividing the corporation into two entities, as well as other efforts to ensure Microsoft’s compliance with antitrust laws. 19 Microsoft then asked the United States Supreme Court to hear its appeal, but the Supreme Court denied certiorari, 20 leaving Microsoft to appeal the federal district court’s decision to the United States *590 Court of Appeals for the District of Columbia. 21 The federal appeals court reversed the federal district court’s finding that Microsoft had violated the Sherman Act by attempting to monopolize the Internet browser market. 22 The federal appeals court also remanded for determination by the federal district court the issue whether Microsoft had illegally tied its browser and operating systems. 23 The federal appeals court, however, affirmed the federal district court’s findings that Microsoft had violated the Sherman Act by possessing monopoly power and maintaining its monopoly in the operating systems software market through anticompetitive means, 24 noting that its “judgment extends to the District Court’s findings with respect to the state law counterparts of the plaintiffs’ Sherman Act claims.” 25 Thus, the federal appeals court affirmed the federal district court’s findings that Microsoft had violated MARA with respect to the operating systems software market, overturned the federal district court’s remedy, and assigned the task of determining the tying issue and new remedy to a different federal district judge. 26 After Microsoft’s second attempt to bring this case to the United State Supreme Court failed, 27 the case returned to the federal district court, with a different judge presiding. The United States decided not to pur *591 sue the tying claim on remand, leaving only the proper remedy in dispute. 28 In November 2001, Microsoft reached a settlement with the federal government and nine states, including Michigan. 29 This prompted the federal district court to “deconsolidate” the action, separating the settling governmental entities from the nine states that, in addition to the District of Columbia, continued to litigate the remedy phase of the case. 30 The settlement, as it concerns the federal government, has passed several preliminary thresholds, including the federal district court’s determination that the Tunney Act, 31 which governs settlements between antitrust violators and the United States government, applies in this case; however, the federal district court has yet to decide whether the settlement agreement satisfies the Tunney Act’s requirements. 32 Regardless of the current procedural posture of Michigan’s lawsuit in federal court, only two aspects of the federal litigation are relevant to this case: the federal district court’s factual findings and legal con- *592 elusions. 33 Those factual findings and legal conclusions, especially as they relate to mara, 34 are the legs on which this case stands. Indeed, because the federal district court held a bench trial to determine the facts in this case, A&M may rely on those facts not affected by the federal appeals court’s decision. As mara § 10 states: A final judgment or decree determining that a person has violated this act in an action brought by the state under section 7, 8(1), or 9 [MCL 445.777, 445.778(1), or 445.779] other than a consent judgment or decree entered before any testimony has been taken, is prima facie evidence against the person in any other action against the person under section 8 [MCL 445.778] as to all matters with respect to which the judgment or decree would be an estoppel between the parties to the action. This section does not affect the application of collateral estoppel or issue preclusion.1 35 1 The net effect of mara § 10 is that the federal district court’s findings that Microsoft is, essentially, an illegal monopoly establishes that matter as a fact in this case. V. LEGAL FRAMEWORK FOR INDIRECT PURCHASER SUITS Any discussion of indirect purchaser suits must start with a discussion of Hanover Shoe because it “played a landmark role in defining the antitrust standing of private plaintiffs,” eliminated the pass-on defense for antitrust defendants, and “prescribed a *593 damage calculation.” 36 In broad terms, in Hanover Shoe the United States Supreme Court announced a prudential federal rule that antitrust defendants could not try to show that direct purchaser plaintiffs were not injured (and therefore had no damages claim) because they had passed on the price increase they sustained to indirect purchasers. Rather, irrespective of whether the direct purchasers pass on the price increase, they could sue for the entire overcharge. The Court reasoned that it should not further complicate already complex antitrust actions with attempts to trace a price increase down the distribution chain[, stating]: “A wide range of factors influences a company’s pricing policies. Normally the impact of a single change in the relevant conditions cannot be measured after the fact; indeed a businessman may be unable to state whether, had one fact been different (a single supply less expensive, general economic conditions more buoyant, or the labor market tighter, for example), he would have chosen a different price. Equally difficult to determine in the real economic world rather than an economist’s economic model, is what effect a change in a company’s price will have on its total sales. Finally, costs per unit for a different volume of total sales are hard to estimate .... Since establishing the applicability of the passing-on defense would require a convincing showing of each of these virtually unascertainable figures, the task would normally prove insurmountable.’ 37 In Illinois Brick, the Supreme Court amplified its perspective on indirect purchaser suits, explaining that [t]he principal basis for the decision in Hanover Shoe was the Court’s perception of the uncertainties and difficulties *594 in analyzing price and output decisions “in the real economic world rather than an economist’s hypothetical model,” and of the costs to the judicial system and the efficient enforcement of the antitrust laws of attempting to reconstruct those decisions in the courtroom. This perception that the attempt to trace the complex economic adjustments to a change in the cost of a particular factor of production would greatly complicate and reduce the effectiveness of already protracted treble-damages proceedings applies with no less force to the assertion of pass-on theories by plaintiffs than it does to the assertion by defendants. However “long and complicated” the proceedings would be when defendants sought to prove pass-on, they would be equally so when the same evidence was introduced by plaintiffs. Indeed, the evidentiary complexities and uncertainties involved in the defensive use of pass-on against a direct purchaser are multiplied in the offensive use of pass-on by a plaintiff several steps removed from the defendant in the chain of distribution. The demonstration of how much of the overcharge was passed on by the first purchaser must be repeated at each point at which the price-fixed goods changed hands before they reached the plaintiff.! 38 1 In light of the Court’s statement on these issues, other commentators have described the Supreme Court’s decision in Illinois Brick as a reaction to three factors: the possibility of multiple liability for defendants, the complexity and uncertainty of apportioning damages among participants in the supply chain, and the advantage of enforcing antitrust laws more effectively through direct purchaser suits. 39 Notwithstanding this broad policy against federal pass-on suits, Illinois Brick did not foreclose states *595 from allowing indirect purchaser actions. 40 As a result, following the Elinois Brick decision, a number of states passed Elinois Brick repealer statutes to allow indirect purchaser suits. 41 Thus, in enacting this state’s repealer law, the Michigan Legislature incorporated language in MARA subsection 8(2) allowing suit by a direct or indirect purchaser, such as an end user or licensee. Though states have this freedom to allow indirect purchaser suits, some state courts still grapple with the factors that concerned the United States Supreme Court in Hanover Shoe and Elinois Brick. Consequently, state courts facing the indirect purchaser issue have either tended toward a “skeptical” or a “sanguine” view. The skeptical view, typified by the [Illinois Brick\ majority, values deterrence over compensation, efficiency over equity, and accuracy over approximation. In this view, the purpose of antitrust remedies is to impose a deterrent penalty on the wrongdoer, and only incidentally to provide compensation to those harmed. The overcharge is initially paid by the direct purchaser, who has most frequent dealings with the offenders and thus is most likely to be aware of the illegal activity. While a portion of the overcharge may be passed on to others the amount of the overcharge and hence the appropriate total damage award remains the same. Allowing indirect purchasers to sue creates costly satellite litigation over how to divide the common fund. It also creates a risk that the right to sue will be divided among so many claimants that collective action problems will prevent any suit from being filed. And it creates the risk of duplicative damages if the sum of the awards to the *596 various direct and indirect purchasers exceeds the amount of the overcharge.1 42 1 In contrast to the concern for deterrence and the unmanageable nature of pass-on suits reflected in the skeptical perspective, the sanguine view, typified by the [Illinois Brick] dissent, values compensation over deterrence; equity over efficiency, and approximation over accuracy. The primary consideration in this view is the assumption that the consumer will bear the major part of the overcharge and should be compensated for it. According to this view, because the initial overcharge to the direct purchaser is almost always passed on, allowing direct purchasers to recover the full overcharge would give them an unacceptable windfall. The tax incidence theory and other generalizations about passing on provide a reasonable basis for judicial approximation of actual harms suffered by the various levels of distribution. “In essence, estimating the amount of damages passed on to an indirect purchaser is no different from and no more complicated than estimating what the middleman’s selling price would have been, absent the violation.” Moreover, because “reasoned estimation” is necessary in all antitrust damage cases, “lack of precision in apportioning damages between direct and indirect purchasers is . . . plainly not a convincing reason for denying indirect purchasers an opportunity to prove their injuries and damages.”1 43 1 Each of these theories has its own merit, particularly the dueling questions of efficiency and equity. However, we need not adopt either a strictly sanguine or *597 purely skeptical perspective on the basis of theory alone. Instead, we take our cues about the necessary proof of damages from our state court rules governing certification of a class action, MARA, and other relevant cases. VI. CLASS CERTIFICATION A. THE FACTORS IN THE COURT RULES The court rules define the factors relevant to a trial court’s decision to certify a class. Under MCR 3.501(A)(1), one or more members of a specific class may bring suit on behalf of other members of the class only if: (a) the class is so numerous that joinder of all members is impracticable; (b) there are questions of law or fact common to the members of the class that predominate over questions affecting only individual members; (c) the claims or defenses of the representative parties are typical of the claims or defenses of the class; (d) the representative parties will fairly and adequately assert and protect the interests of the class; and (e) the maintenance of the action as a class action will be superior to other available methods of adjudication in promoting the convenient administration of justice. The word “and” following the semicolon in subsection d is critical because it reveals that the action must meet all the requirements in MCR 3.501(A)(1); a case cannot proceed as a class action when it satisfies only some, or even most, of these factors. Further, the plaintiff, when the plaintiff is seeking to certify *598 the class, has the burden of proving that the action satisfies every one of the factors in MCR 3.501. 44 B. CLASS MEMBERSHIP The threshold question in any proposed class action is whether the proposed class representative is a member of the class. “A plaintiff who cannot maintain the cause of action as an individual is not qualified to represent the proposed class.” 45 A&M has alleged facts sufficient to show that it purchased multiple Microsoft operating systems involved in Microsoft’s documented antitrust violations. Thus, assuming solely for the sake of analysis that there are no other defects in A&M pleadings and proofs, A&M is an appropriate class representative because, having purchased numerous Microsoft products, it fits the class definition of an indirect purchaser that Microsoft’s anticompetitive activity allegedly harmed. C. NUMEROSITY According to MCR 3.501(A)(1)(a), the first specific class action factor is numerosity. A&M alleged in its complaint that Microsoft distributed 300 million copies of its operating systems in the United States, several million of which were distributed in Michigan. By the time A&M moved for class certification, it had lowered this estimate to hundreds of thousands. Even so, the class members are so numerous that their joinder would be impracticable, satisfying this aspect of the court rale. Indeed, attempting to join hundreds *599 of thousands of plaintiffs in a single suit would border on the impossible. D. COMMONALITY MCR 3.501(A)(1)(b) prescribes that, to certify a class action, there must be common questions of law or fact that predominate over individual questions. Microsoft challenges the trial court’s finding that A&M satisfied this inquiry. This Court has explained that [t]he common question factor is concerned with whether there “is a common issue the resolution of which will advance the litigation.” Sprague v General Motors Corp, 133 F3d 388, 397 (CA 6, 1998), cert den 524 US 923; 118 S Ct 2312; 141 L Ed 2d 170 (1998). It requires that “the issues in the class action that are subject to generalized proof, and thus applicable to the class as a whole, must predominate over those issues that are subject only to individualized proof.” Kerr v West Palm Beach, 875 F2d 1546, 1557-1558 (CA 11, 1989).[ 46 J Still, there “is no requirement in the rule that all questions necessary for ultimate resolution be common to the members of the class.” 47 A&M clearly showed that there are common issues and facts involved in its claim that Microsoft engaged in anticompetitive and monopolistic activities, as the suit in the federal district court proved with respect to mara itself. Consequently, the pertinent question is whether A&M can demonstrate with common proof that the members of the class have suffered a com *600 mon injury as indirect purchasers of Microsoft operating systems. 48 Microsoft contends that A&M lacks this proof. Microsoft provides no authority suggesting that A&M had to prove actual damages to any particular degree of certainty at this preliminary class certification stage. A&M nevertheless had to provide some basis for the trial court to conclude that all members of the class had a common injuiy that could be demonstrated with generalized proof, rather than evidence unique to each class member. 49 To be clear, the question is not whether each member of the class has sustained an identical amount of damage because of a defendant’s anticompetitive behavior, which can be estimated, 50 but, rather, whether “the common issues [that] determine liability predominate.” 51 Federal precedent applying Michigan law suggests that when a statutory scheme requires proof of actual damages, as MARA subsection 8(2) does, this language necessarily entails the individualized inquiry that is inconsistent with the commonality requirement, thus preventing a class action. 52 Nevertheless, contrary federal precedent indicates that if a plaintiff proposes a method or *601 formula by which the court could determine that the defendant’s conduct caused actual damages or injury to each class member, even if the amount of damages due each party is later computed at a separate trial, then a class action is possible. 53 We address this difficult issue in detail in a following section. E. TYPICALITY MCR 3.501(A)(1)(c) requires that the claims or defenses of representative parties be typical of the claims or defenses of the class. The parties do not dispute this factor. F. ADEQUACY MCR 3.501(A)(1)(d), the fourth class action factor, requires the trial court to scrutinize whether the representative parties will fairly and adequately assert and protect the interests of the class. Similar to the typicality question, there is no dispute concerning this factor. G. SUPERIORITY MCR 3.501(A)(1)(e), the fifth and final class action factor, asks whether a class action, rather than individual suits, will be the most convenient way to decide the legal questions presented, making a class action a superior form of action. In deciding this factor, the court may consider the practical problems that can arise if the class action is allowed to pro *602 ceed. 54 “The relevant concern ... is whether the issues are so disparate” that a class action would be unmanageable. 55 To determine whether a class action is a superior form of action, a trial court must consider: (a) whether the prosecution of separate actions by or against individual members of the class would create a risk of (i) inconsistent or varying adjudications with respect to individual members of the class that would confront the party opposing the class with incompatible standards of conduct; or (ii) adjudication with respect to individual members of the class that would as a practical matter be dispositive of the interests of other members not parties to the adjudications or substantially impair or impede their ability to protect their interests; (b) whether final equitable or declaratory relief might be appropriate with respect to the class; (c) whether the action will be manageable as a class action; (d) whether in view of the complexity of the issues or the expense of litigation the separate claims of individual class members are insufficient in amount to support separate actions; (e) whether it is probable that the amount which may be recovered by individual class members will be large enough in relation to the expense and effort of administering the action to justify a class action; and (f) whether members of the class have a significant interest in controlling the prosecution or defense of separate actions.1 56 1 *603 Because this factor is intertwined with the class action factor concerning commonality, 57 we also address it in the following section. vn. THE BATTLE OF THE EXPERTS A. LEFFLER I A&M fired the opening salvo in the battle of the experts in this case by submitting Dr. Keith B. Leffler’s April 13, 2001, affidavit (Leffler I) to support its motion for class certification. Dr. Leffler is an Associate Professor of Economics at the University of Washington. In his first affidavit, Dr. Leffler indicated that A&M had asked him to describe the type of economic analysis, evidence, and data he would likely use to determine the scope of the effect, if any, Microsoft’s alleged monopolization had on “Intel-PC compatible operating systems on Michigan end-users.” Additionally, A&M asked him to describe methods by which “any damages to Michigan end-users can be calculated on a classwide basis.” Clearly, this description focused on the pass-on requirement, the second of the two steps in the analytical construct we set out at the beginning of this opinion. Nevertheless, most of Dr. Leffler’s analysis in Leffler I concerned the first step in this analytical construct, the overcharge requirement. Dr. Leffler described three alternative “yardstick” approaches that could be used to estimate the overcharges. First is a “comparable market yardstick,” which is a comparison price from a comparable market that was not affected by anticompetitive activity. Second is a “com *604 petitive market yardstick,” which relies on margins earned by sellers in a comparable market free from antitrust violations. Third is a “violation-free-period yardstick,” which is a price in the market at issue during a period free of the anticompetitive conduct. This is not to say, however, that in Leffler I Dr. Lef-fler ignored the pass-on requirement. Indeed, he referred to it numerous times, saying that basic economic principles show that the proposed class members paid higher prices during the class period for Microsoft operating system software whether purchased as a software package or preinstalled on a PC because of Microsoft’s monopolization. Basic economic principles, therefore, can be used to show that all proposed class members have been injured. Damages to end user class members can be estimated using a class wide statistical determination of the relationship between the costs faced by Microsoft’s customers and the prices paid by end users. The intense competition in the distribution of Microsoft operating systems implies that a Microsoft overcharge will affect a particular end user group the same regardless of the details of how the operating system reaches the end-users through the distribution chain. It also suggests a 100 percent pass-on is likely. Therefore, a class wide damage methodology should be viable given the practical realities of the Microsoft operating system market. As a consequence, as in any indirect purchaser case, the issue arises of the extent to which Microsoft’s overcharges to OEMs [original equipment manufacturers], distributors or retailers are reflected in the prices paid by the end-users of the operating system. *605 There is an [sic] extensive economics literature analyzing the relationship between cost changes at one level of distribution and price changes downstream. For the purposes of class certification, I am assuming, consistent with Judge Jackson’s finding, that consumers do not have access to close substitutes for Microsoft’s operating systems. The economics literature therefore demonstrates that any overcharge by Microsoft is expected to result in all end-users (i.e. class members) paying higher prices regardless of the specific point in the distribution chain from which they purchased. In my opinion, the data should be readily available to apply sound, accepted economic and statistical methodologies to estimate the impact of Microsoft’s overcharges on the prices to end-users. The analysis of the relationship between end user price and Microsoft customer costs requires a relatively simple statistical estimation. The data for such estimation should be readily available. The prices charged end users are available from public sources and, likely, from third party discovery. The impact of such actual cost reductions [the reduced costs of OEMs and system builders for many components such as disk drives and hard drives] can be statistically related through regression analysis to actual end users’ prices. The end-user price impact of different overcharges to Microsoft customers can then be inferred from such regression analysis. [T]he end-user prices of products of varying costs to the retailer or distributor can be compared to infer statistically *606 how such cost differences are reflected in end-user prices. Again the end-user price impact of different Microsoft overcharges to its customers can then be inferred from such statistical analysis. . . . Modem computer systems and statistical tools can readily handle such data analysis. I would expect that separate end-user overcharges and damages could be estimated for each appropriate Microsoft customer channel through use of a single statistical model. [T]he economic expectation in competitive markets is that cost changes will be fully reflected in the price to the end users. Recent economic literature on the effect of taxes on prices also supports the expectation that there will be a full pass on of a cost change as a consequence of the elimination of a Microsoft operating system overcharge. Industry experts also believe that cost changes in the PC industry are fully and rapidly reflected in end user prices. Microsoft’s expert in the government trial recognized that changes in operating system costs will lead to congment changes in the prices paid by end users. Therefore, it is likely that only relatively simple analysis will be required to confirm the strong expectation that there will be a full pass on to end users of the Microsoft operating system overcharges.[ 58 ] Dr. Leffler was also careful to note that “merits discovery” had not yet commenced and that he had “not at this time actually performed the detailed careful economic and statistical work that will be required to accurately determine class damages.” Most specialized fields—and some not so specialized—have their own esoteric terms and particular *607 ized methods; the fields of law and economics come quickly to mind. Recognizing the risk of oversimplification, we believe it fair to say that in Leffler I, Dr. Leffler adopted the “tax incidence” theory as a method of determining whether Microsoft’s alleged overcharge was passed-on to indirect purchasers. 59 As one commentator explains the tax incidence theory: An overcharge to all of the firms at an intermediate level of distribution is conventionally analogized to an excise tax imposed on each unit sold by the overcharged firms. The “tax” results in a parallel upward shift in the direct purchaser’s marginal costs. Under the conventional assumptions of economics, such a cost would lead the intermediate purchaser to increase its price to its customers, and thus to transfer some part of the overcharge to them. Under those assumptions, the relative share of the overcharge borne by the direct and the indirect purchaser would depend upon the relative elasticities of supply and demand in the resale market. [T]he only circumstances in which the intermediate purchaser would not pass on any of the overcharge would be when demand is perfectly elastic, or horizontal, or when supply is perfectly inelastic, or vertical. In the former case, the seller is unable to raise prices without losing all of its sales; in the latter case, the seller is unable to alter its output in response to changes in the price of goods that it purchases, so the price at retail must remain the same. Consequently, in both cases the intermediate purchaser must absorb the entire price increase. Thus, under the assumptions of the incidence model, the indirect purchaser will *608 bear some of the overcharge, except in extreme circumstances.! 60 1 This analysis assumes that “the retail market is perfectly competitive.” 61 B. NICHOLS AFFIDAVIT Microsoft’s responsive salvo came, basically, in the form of two affidavits. 62 The first affidavit was from Dr. Donald A. Nichols, dated June 8, 2001 (Nichols Affidavit). Dr. Nichols is a Professor of Economics and Public Affairs at the University of Wisconsin-Madison. Dr. Nichols indicated that he had been asked to do a survey of the prices in Michigan of personal computers with the Windows operating system preinstalled and the prices of boxed Windows operating systems offered for retail sale. In his affidavit, Dr. Nichols described his data sources, survey ranges, and survey methods. He concluded that, in general, there was substantial variation across Michigan retailers in the prices quoted for the same computer model on any given date, with greater price variation for only some models. He also concluded that it was clear from the data related to Michigan retailers’ prices for the Compaq Presario 12XL325 that some Michigan buyers paid different prices for computers that had identical components while some Michigan buyers paid identical prices for computers that retailers promoted as having different components. From other data related to Internet *609 prices for the Compaq Presado 12XL325, Dr. Nichols observed an even wider variation of prices than in the retailer survey. Dr. Nichols also discussed “price points,” specifically prices that end with $49.99 or $99.99. He found that a high percentage of Michigan retailers’ prices, after taking into account rebates, are within $1.00 of these price points. C. HAUSMAN I Microsoft’s second volley was an affidavit from Dr. Jerry A. Hausman (Hausman I), submitted in opposition to the motion for class certification. 63 Dr. Haus-man is a Professor of Economics at the Massachusetts Institute of Technology. In Hausman I, he indicated that he had been asked to provide an expert opinion on whether, when using common proof, it is possible to determine with any reasonable degree of certainty “whether Michigan end-users of Microsoft PCs were in fact injured by the conduct of Microsoft” alleged in A&M’s complaint. Microsoft also asked him to determine whether “the damages sought for the members of the proposed classes could be estimated with reasonable accuracy using common proof.” Although using different terminology, Dr. Hausman pointed out early in Hausman I that A&M must demonstrate that Microsoft raised its prices as a result of illegal anticompetitive actions (the overcharge requirement) and that the price increases were passed on through intermediaries to the final *610 purchasers (the pass-on requirement). Dr. Hausman’s analysis was wide-ranging, but, when concentrating on these two requirements, he concluded that economic theory does not provide any straightforward way to analyze pass-through without an individualized analysis that takes account of the diversity among the thousands of affected manufacturers, distributors, and retailers. The conclusion ... is that the pass-through, if any, of an alleged overcharge to the next level of distribution depends on the demand, costs, and competition-related factors of the selling firm and the specific product subject to the overcharge. In this case, one would need to know the demand, cost, and competition-related factors for the thousands of firms in the distribution chain as well as all of the Microsoft operating systems at issue. [Tjhis undertaking would be infeasible using common proof in this case because the large number of final purchasers, firms, and products, the highly diverse business strategies used by firms, and the length of the proposed class period produce numerous variations in the possibilities of any seller passing through any portion of an overcharge. Plaintiffs’ expert Professor Leffler proposes several “yardsticks” against which he claims the alleged overcharge on Microsoft operating systems could be measured. ... It is important to note that these proposals concern estimating the overcharge to direct purchasers. None of these proposals addresses the essential issue for the plaintiff: the more difficult problem of estimating the amount of the overcharge that was passed through to final purchasers. Moreover, all of Professor Leffler’s proposed yardsticks are fundamentally flawed, they are speculative, and they do not account for the economic realities of this case. Professor Leffler claims that to estimate pass-through, “there is no need to model how the overcharge is passed on at each stage, rather the interest is only in how an over *611 charge by Microsoft to its customers impacts the final price.” ... In other words, Professor Leffler suggests that he need only econometrically analyze the relationship between end user prices and Microsoft’s prices. The economic conditions in the intermediate stages of distribution are deemed “irrelevant” for his analysis . . . resulting in “a relatively simple statistical estimation”. . . . However, Professor Leffler’s assertions are inconsistent with generally accepted econometric theory and practice in several ways. Plainly, Dr. Hausman raised a number of objections to Dr. Leffler’s “yardstick” methods for calculating the alleged overcharge, but he concentrated his fire on Dr. Leffler’s approach to the pass-on requirement. There, Dr. Hausman asserted that “[t]he factors that enter into determining whether a specific member of the proposed class was in fact harmed cannot be analyzed on a class-wide basis” and that “determining with any degree of certainty who in fact was injured would require detailed, individualized factual inquiries.” D. LEFFLER II A&M’s countervolley consisted primarily of a second affidavit from Dr. Leffler, dated July 6, 2001 (Leffler II). At that time, Dr. Leffler noted that “data have not been produced in discovery from which” he could “actually calculate damages.” Consequently, he conceded, he could “only evaluate in a conceptual rather than empirical way the likely class-wide methods of damage calculations.” For instance, Dr. Leffler explained in a footnote that [licensing agreements for some OEMs [original equipment manufacturers] and retailers are available. However, these agreements do not necessarily provide information as to the *612 actual prices paid by the OEMs or the retailers. The actual cost to the OEMs and retailers that contract with Microsoft will be provided in Microsoft databases that are not yet available. Some list price information for distributors is also available, however, again, without the Microsoft database I cannot know what the actual transactions prices were. Information as to the prices charged end-users by OEMs and retailers is being obtained from third party vendors and through third party discovery but these data are not yet available in any systematic way. Despite the absence of hard numbers to support A&M’s claims, Dr. Leffler summarized his position with respect to the overcharge requirement and the pass-on requirement: Professor Hausman mischaracterizes my proposed yardstick damage methodologies. He offers no conceptual methodological reasons that such yardsticks cannot be implemented. Therefore, determination of Microsoft’s overcharge is amenable to standard economic yardsticks. The damages plaintiff seeks to recover in this case are those suffered by end-users. These damages can be estimated by a simple statistical determination of the relationship between prices paid by Microsoft’s customers and the prices paid by end-users. Professor Hausman himself relies upon such a simple approach. There is no economic reason to estimate a complete supply and demand model for each seller at each level of the distribution chain. Professor Hausman’s extensive detailed description of myriad distribution channels is of no relevance to proper estimation of the damages to end-users. While Professor Hausman recognizes the intense competition in distribution of Microsoft software, he ignores the implications of this intense competition for the damage estimation. Intense competition in distribution implies that a Microsoft overcharge will affect a particular end-user group the same regardless of the details of how the operating system reaches the end-users through the distribution chain. It also suggests a 100 percent pass-on is likely. Therefore, a *613 class-wide damage methodology should be viable given “the practical realities of the case.” Dr. Leffler also responded to several of Dr. Haus-man’s specific points on the pass-on requirement: I understand that plaintiff, having established the adverse impact on all members of the class, is required to reasonably estimate damages to the class. This is distinct from the estimation of the damages to each individual member of the class. Professor Hausman’s concern seems to focus on the allocation of the class-wide damages among individuals. I believe it is premature to address this at this time. Indeed, the answer may be manifest from the ultimate class-wide methodology. For example, if a proper damage analysis demonstrates that there was a 100% percent [sic] pass on of' the Microsoft overcharge and the overcharge was 50 percent of the Microsoft price, it would be a simple matter to determine a specific individual’s damage. . . . However, to calculate the damages to end-users there is no need to model how the overcharge is passed on at each stage. Rather, the goal is only to determine how an overcharge by Microsoft to its customers impacts the price to end-users. The supply and demand elasticities at each distribution stage are not needed to predict how end-users are impacted. Rather, analysis of how the actual prices charged end-users were impacted by actual cost changes to Microsoft’s customers will show how the overcharge damaged end-users. The analysis of relationships between end-user price and Microsoft customer cost requires relatively simple statistical estimation. . . . Therefore, it is likely that relatively simple analysis will be required to confirm the strong expectation that there will be a full pass on to end-users of the Microsoft operating system overcharges. *614 B. HAUSMAN H The battle of the experts concluded with Dr. Haus-man’s second affidavit, dated July 17, 2001 (Hausman II). In Hausman II, Dr. Hausman summarized his critique of Leffler II: Professor Leffler proposes an approach that “assumes away” the fundamental problems with using recognized empirical economic methods to establish antitrust impact on a classwide basis. For instance, Professor Leffler’s conclusions rely heavily on his assumption that the PC industry is essentially perfectly competitive. In particular, his conclusion that 100% of the alleged overcharge will be passed through to final purchasers rests on this assumption .... However, his assumption of perfect competition is not based on any analysis and, indeed, is inconsistent with the basic economic characteristics of the PC industry, where the products are highly differentiated, not homogenous [sic]; brand names are an important source of competitive advantage; and firms face significant fixed costs. Dr. Hausman also responded to several of Dr. Lef-fler’s specific points on the pass-on requirement: In saying that computer retailing is “extremely competitive,” Professor Leffler appears to be claiming that computer retailing is essentially “perfectly competitive,” which provides the justification for his assumption that pass-through would be 100%. Perfect competition is a textbook economic construct that is rarely, if ever, encountered in the real world. Because OEMs compete with each other in the context of differentiated products, not “perfect competition,” Professor *615 Leffler’s assumption that the PC industry is essentially perfectly competitive (and the implied assumption that pass through would be 100%) is not valid. While Professor Leffler wants the situation to be simple so that he could apply “very basic propositionfs] of introductory economics” . . . the economic realities of the industry are much more complex and thus a much more complex economic analyses [sic] must be performed. VIH. PRECEDENT REGARDING CLASS CERTIFICATION IN INDIRECT PURCHASER SUITS A. OVERVIEW There are no reported appellate cases in Michigan interpreting the indirect purchaser provisions of MARA in the context of certifying a class action. Consequently, the proof necessary to demonstrate commonality and superiority, the first and fifth class action factors, in an indirect purchaser action presents an issue of first impression in our state at the appellate level. In considering this matter, however, we examine trial court decisions in Michigan involving class actions by indirect purchasers. Our search of foreign case law has been slightly more fruitful, revealing some decisions from appellate courts as well as trial courts in other states addressing class certification in indirect purchaser suits against Microsoft. 64 Those cases from foreign jurisdic *616 tions relying on evidence similar to the evidence likely to be presented in this case are of particular interest to us. Thus, we also examine these foreign cases to determine whether they might have components persuasive to our decision in this case. B. MICHIGAN TRIAL COURT DECISIONS (1) HOLMES v ABBOTT LABORATORIES In Holmes v Abbott Laboratories, 65 a case Microsoft submits is persuasive, 66 the plaintiff moved to certify a class action on behalf of consumers who purchased infant formula between 1980 and 1992, and for which the defendants had allegedly fixed the prices pursuant to a conspiracy that implicated the way health care professionals helped advertise and market the formula. 67 At the hearing on the motion to certify the class, the plaintiff’s attorney argued that, in demonstrating that there was common proof of impact, it was unnecessary to prove that *617 each purchaser bought [the infant formula] in the same manner from the same seller or in the same quantity. Proof of impact for class certification does not require that each purchase have the same size claim or even paid the same amount. Proof of impact is whether or not there is a common basis to judge at a preliminary stage that all purchases of the effective [sic] product were impacted! 68 1 This argument prompted the trial court to ask the plaintiff’s counsel to address a defense expert’s testimony concerning the infant formula prices, including delays and plateaus, in particular the probability that “a person could get in and out of the market and never have experienced any kind of pass through price increase.” 69 The plaintiff’s counsel responded that the defense would have the court believe that the conspiracy was to fix prices over time, in which case comparing price increases over time would reveal an overcharge as a step before determining whether consumers eventually absorbed the overcharge. 70 However, from the plaintiff’s perspective, examining a change in price alone over time was too simplistic because it failed to answer the critical question whether, absent the alleged conspiracy, the prices consumers paid would have been lower. 71 In other words, the plaintiff’s argument suggested that, while prices for many products may naturally increase over time because of a variety of factors, not eveiy price will increase, nor will it necessarily increase at the rate that consumers experienced. Thus, it was possible that even products that had a modest increase in *618 price over a certain period might be passing an overcharge to consumers that would not have existed in a competitive market. 72 Turning to the evidence, the plaintiffs attorney argued that it was possible to calculate a common basis for overcharge and pass-on. Indeed, he contended, Dr. Beyer, the plaintiff’s expert, had pointed to “benchmarks” to make the appropriate calculations. 73 Though the defendants challenged “the validity of those benchmarks,” the plaintiff’s attorney countered that whether a benchmark was valid was a “merits determination,” not appropriate for the preliminary questions addressed in the process of certifying a class. 74 The plaintiff’s attorney conceded that the trial court need not accept at face value the representation that the method of calculating effect was appropriate, but that the plaintiff had to demonstrate by a preponderance of the evidence 75 on the record a method “that is a viable, rational method or that the method is not so insubstantial that it amounts to no method at all.” 76 The plaintiff’s attorney then proceeded to outline Dr. Beyer’s methodology, highlighting the areas in which Dr. Beyer and the defense expert agreed. In short, as the plaintiff’s attorney portrayed Dr. Beyer’s earlier testimony, the expert was able to demonstrate, partly by using pricing related to the infant formula used in the Women Infants and Children (wic) program, 77 that the wholesale price the *619 defendants charged to retailers affected the price retailers then charged to consumers, connecting the initial overcharge with the pass-on. 78 Critically, by calculating the aggregate difference between what the class actually paid and what it would have paid without the conspiracy, Dr. Beyer’s method avoided any need to calculate how much overcharge was passed-on to each individual class member. 79 Despite these arguments, the trial court remained unconvinced that the plaintiff had satisfied her burden under MCR 3.501. Analyzing each class action certification factor, the trial court found that the class was so numerous that joinder would be impracticable. 80 Skipping to the third factor, the trial court found that the evidence met the typicality requirement. 81 The plaintiff had also proved the fourth and fifth factors to the trial court’s satisfaction because the trial court had no doubt that the plaintiff would be an appropriate class representative and because a class action would be the most efficient way to handle the legal interests of the tens of thousands of individuals who purchased infant formula in the years at issue. However, returning to the second factor concerning the common proofs of impact, the trial court noted the inconsistency in Dr. Beyer’s testimony, in which he once claimed that the overcharge had harmed the retailers, which the trial court interpreted to exclude the possibility that the overcharge had been passed-on to the proposed class. 82 The trial *620 court remarked that Dr. Beyer noted that retailers did not pass-on overcharges consistently over time or when compared to other retailers. 83 Dr. Beyer had also conceded that he did not interview any Michigan retailers and was unaware of any way to adjust his calculations for this inconsistent pass-on rate. 84 Further, acknowledging the actual damages requirement in MARA subsection 8(2), the trial court observed that Dr. Beyer had discussed his methodology in terms of the “insight” it offered: When Dr. Beyer says that the wic data points provide the best insight available in terms of what the price competition would have looked like and so on and later says the wic open market bid prices did offer among the various benchmarks not only the best insight but in my judgment a good insight into what price competition and so on, that falls short of clearly establishing class-wide a means of determining damages. Then we get to the point made by [the defense expert] that when you have 47 different types of products sold and more than a million consumers, we get to [defense counsel’s] point in his argument as it relates to manageability.[ 85 1 Consequently, though the trial court found Dr. Beyer credible, 86 it determined that the plaintiff had “failed to establish by a preponderance of the evidence that there are questions of law and fact common to the members of the class that predominate over questions of fact affecting only individual [class] members.” 87 *621 Therefore, the trial court denied the motion to certify the class. 88 (2) WILCOX v ARCHER-DANIELS-MIDLAND CO In Wilcox v Archer-Daniels-Midland Co, 89 the plaintiff alleged that the defendants had conspired to fix the price of high fructose com syrup and citric acid. 90 Com syrup and citric acid are ingredients in an enormous number of commercially prepared food products, such as “soft drinks and breakfast cereals.” 91 Under the plaintiff’s theory, the defendants, who controlled “the vast majority of the market” for producing these ingredients, conspired to raise the prices for these products charged to the food manufacturers. 92 In turn, the food manufacturers passed on this overcharge to the consumers who bought the wide array of food products containing these ingredients. 93 Thus, as the trial court put it, the class the plaintiff proposed for certification “would, in essence, consist of everyone residing in the state during” a two-year period starting January 1, 1992, and ending December 31, 1994. 94 The trial court, as was proper, then analyzed each of the factors relevant to class certification under *622 MCR 3.501. The trial court found that a class consisting of “nearly every person in this state” would be too numerous to make joinder the more practical approach in the case. 95 Similarly, the trial court found that the plaintiff would make an appropriate class representative because he had the same interests in the suit as the other members of the proposed class and would vigorously pursue it, as his choice of experienced and vigorous attorneys had demonstrated up to that point. 96 In its analysis of the factor in MCR 3.501(A)(1)(b), the trial court squarely addressed the defendants’ contention “that it would be extremely difficult to demonstrate any consistent damages suffered by any one individual in the proposed class, and therefore, the class is impracticable.” 97 The trial court rejected this argument, stating that “[t]he fact that there are some questions that are difficult or not common to the group will not necessarily defeat the Plaintiff’s request to certify a class action.” 98 In the trial court’s opinion, despite some likely differences between members of the plaintiff’s proposed class, critical questions, such as whether a conspiracy actually existed, would be uniform for the class. 99 The trial court did not find proof problematic under this factor. The trial court’s analysis thus had proceeded smoothly in a direction favoring class certification through the first four certification factors. 100 Yet, the final class certification factor, which synthesized the *623 other factors in asking whether a class action would be superior to other forms of lawsuits, troubled the trial court. The trial court concluded that the likelihood that many of the class members would have sustained such small damages that suing individually would not be feasible militated in favor of a class action. 101 However, the trial court returned to the question of proof, explaining that a class of the size proposed “creates a vast problem of proof that in this Court’s opinion tolls the death knell for the request to certify a class in this case.” 102 First, the trial court found that the proposed class was a “fluid group,” making it virtually impossible to determine who actually was in the class. 103 For example, though the plaintiff claimed that proving and quantifying residency during the period defined for the class would be possible, the trial court questioned how the parties could account for people who had lived in the state but had moved, people who worked in this state but lived elsewhere, and the myriad other situations a proposed class of that size posed. 104 As the trial court put it, dealing with the paperwork would be “staggering.” 105 Second, the trial court noted that the prevalence of citric acid and com syrup made it highly likely that the members of the class would not have known when they consumed the ingredients, much less that they would have kept any records. 106 This plainly *624 posed a “serious and substantial problem . . . [in] ascertaining the amount of damages suffered by any of the claimants.” 107 The trial court acknowledged that consumer goods had been the subject of class actions in the past, but noted that this case was relatively unusual in the sense that it related to ingredients rather than finished products. 108 The plaintiff suggested avoiding this problem by assuming that “everyone [in the class] ate or drank the same amount of the various products during the applicable period.” 109 The trial court rejected this effort to “generalize” the proof of damages, stating that [t]he purpose of this litigation would be to compensate those people who were damaged by the alleged illegal activities . . . not just to fix some punishment for antitrust violations. While the people of this state deserve protection from illegal price fixing, there are other mechanisms available to punish such illegal actions.1 110 1 Consequently, the trial court denied the motion for class certification, resting its decision on the difficulty of defining the class and proving damages. (3) WOOD v ABBOTT LABORATORIES In the 1990s, Shirley Wood filed an action in Oakland Circuit Court accusing Abbott Laboratories of engaging in price fixing in one segment of its pharma *625 ceutical market. 111 According to Wood, Abbott had priced its products on the basis of the identity of the intermediate (direct) purchasers, charging higher prices to the retail pharmacies it disfavored, and lower prices to the institutional pharmacies, such as hospitals, nursing homes, and health maintenance organizations, it favored. Wood asserted that this scheme negatively affected the consumers who purchased their medications from retail pharmacies, which passed on the higher drug prices to them. Thus, Wood asked the trial court to certify a class consisting of consumers who had purchased name-brand drugs from one of the disfavored retail pharmacies. The trial court wasted no time in analyzing the class certification factors, focusing entirely on whether Wood could prove overcharge and pass-on for the entire class with common proofs. Wood’s expert maintained that findings in a similar federal antitrust case, a discount method comparing the prices Abbott charged the favored and disfavored direct purchasers, and the profit as the logical goal of the conspiracy were sufficient proof of overcharge. The expert suggested that regression analysis, a tax incidence theory, and standard economic theories made it “optimal for a pharmacy to pass on some portion of its cost increases to the consumer,” thus satisfying the pass-on requirement. However, the expert conceded that a regression analysis could not calculate the actual pass-on rate, and that the incremental costs that would affect this analysis would vary. The *626 expert also conceded that the pass-on rate might vary according to the drugs at issue, necessitating individual calculations for each drug. The defense expert emphasized the individual calculations that would be necessary, explaining that his experience demonstrated that retailers did not necessarily pass on all the price increases, and were not even consistent regarding that portion that they did pass on to consumers. In light of these variations, the defense expert also concluded that it was impossible to calculate the amount of pass-on using the opposing expert’s simple cost comparison method. The trial court, in considering these conflicting statements regarding how to calculate pass-on, returned to mara’s language in MCL 445.778(2). Emphasizing that MARA required proof of actual damages, the trial court concluded that the plaintiff’s expert had not provided a common way to calculate each class member’s damages. The expert’s method “would require examination of the drugs each class member purchased from which retailer, the discounts applicable to each retailer for each drug at the time of purchase, and other relevant factors, resulting in thousands of mini-trials and rendering the class unmanageable.” Accordingly, the trial court found that “individual questions of fact as to both injury and damages predominate over the one theory common to the class, that being the existence of the alleged conspiracy,” and denied the motion for class certification. (4) CONCLUSIONS AND COMMON FACTORS Clearly, the facts of the individuals cases weighed most heavily on the trial courts in Holmes, Wilcox, and Wood. However, these three opinions reveal sev *627 eral important factors about the way mara’s actual damages requirement and the court rules identify indirect purchaser cases that are inappropriate for class certification. Proving a defendant’s illegal conduct may not be a particularly difficult matter, even when a class is large. In most cases, no matter how large the class, there are usually only a handful of defendants, at most. All members of the class are highly likely to use the same evidence to prove the defendant’s illegal conduct that essentially binds the class as a single entity. In Wilcox, for instance, the plaintiff sued only five defendants that manufactured high fructose com syrup and citric acid. As the trial court in Wilcox suggested, even though the class members may have purchased these two ingredients in any number of products, the defendants allegedly engaged in the same illegal activities. Indeed, with each defendant being part of the same alleged conspiracy, common proof should have been easy to find, if any proof of the conspiracy was available. More importantly, with respect to the nature of the proof itself, there is a distinct advantage in proving not only the illegal conduct, but also the overcharge in comparison to the pass-on. The evidence may be in the form of testimony from eyewitness whistle-blowers, former employees, and corporate accountants and officers, as well as documents, such as minutes of board meetings and invoices showing charges to direct purchasers like retailers. In Holmes, the plaintiff relied heavily on the defendant’s internal documents to prove, at least preliminarily, that the defendant had fixed prices. Courts and juries are comforta *628 ble examining this sort of evidence because it is grounded in the observable world. In contrast, when proving pass-on, plaintiffs typically must present evidence in the form of economic theory and analysis. These theories and analyses organize vast amounts of data and explain trends that would otherwise defy categorization in the ordinary person’s lexicon. Nevertheless, though the real world often validates economic theories and analyses, they represent evidence that courts and juries may have a difficult time accepting as proof. This is not to say that courts and juries are unable to understand economic theories and analyses, but that they can appreciate that there is a distance between theory and reality and that the actual damages requirement in MARA mandates proof of pass-on’s reality. The economic analyses in these three Michigan cases we have examined ultimately failed to bridge the gap between theory and reality, but they are useful in that they demonstrate a range of pitfalls that frustrate making that theory-reality connection. In Wilcox, it seems quite plausible to us that the plaintiff could have found an economic theoiy to suggest that pass-on actually occurred because the circumstances surrounding the case made pass-on so likely. Consider that the products at issue were used widely, which may have allowed all the direct buyers to pass on the overcharge to consumers without drawing any attention to the increased price. These direct buyers who made the consumer goods that contained the ingredients at issue were virtually required to purchase the ingredients from the defendants because of the defendants’ dominance in the market. This not only made it likely that all direct *629 purchasers had this overcharge to pass on to consumers, the large scale and consistent uses for these ingredients may have made the amount of the pass-on more uniform. The food products in which these ingredients were used may have had a low profit margin without the price fixing, preventing direct purchasers from tolerating the overcharge, thereby requiring the pass-on. Though these circumstances seemed to suggest that pass-on actually occurred, the very breadth of the consumer products involved, the length of the class period, the size of the class, and the likely variation in the amount of consumer products each class member purchased made it impossible to propose an economic analysis that would do anything more than “generalize” damages. As the trial court in Wilcox found, generalization is inherently incompatible with proving actual damages. Further, though the information available in the Wilcox trial court’s opinion is limited, as far as we know, the plaintiff’s expert had yet to name a theory or propose the outline of an analysis that would lead to this generalization once the parties completed discovery. In the end, the trial court’s decision in Wilcox suggests that the high likelihood of real world pass-on does not obviate the need for proof of the damages the members of the class would be entitled to receive. While the class certification stage does not require an exacting analysis of damages, it nevertheless requires more than a promise that, when the time comes, a plaintiff will be able to suggest a general figure for damages. Holmes, however, represents a different type of problem in proving pass-on. The plaintiff in Holmes retained an expert who not only identified the eco *630 nomic theories and analyses that he believed would be Useful in calculating the amount of the overcharge passed on to the infant formula purchasers who would comprise the class, the expert actually performed calculations using real data. Thus, unlike the plaintiff in Wilcox, the plaintiff in Holmes presented credible evidence that she would be able to calculate the amount of pass-on, thereby quantifying the damages the class members sustained from the defendant’s allegedly anticompetitive behavior. Nevertheless, the flaw in the plaintiff’s calculations boiled down to the variations in the behavior of the direct purchasers, that is, the infant formula retailers. With so many different retailers, each with individual economic constraints and interests, the plaintiff’s expert learned that there was no consistent pass-on rate, even at the same retailer over time. Even if the expert could have calculated the pass-on rate for the individual retailers, that enormous task would have destroyed the efficiency a class action otherwise provides. Holmes, and to some extent the extraordinarily large proposed class in Wilcox, illustrate the paradox of indirect purchaser class actions concerning consumer goods. If the amount of overcharge passed on to consumers was large, individuals would pursue individual remedies. However, in many instances, the amount of overcharge passed on to purchasers is relatively small, providing little incentive for individual purchasers to sue on their own. 112 In order to make a class action lawsuit economically feasible for the plaintiff, especially when an overcharge is small and *631 the pass-on likely smaller, the class must be very large. Yet, as the class grows larger, so too does the number of variables affecting the rate at which the individual class members were forced to assume the overcharge through pass-on. This makes it difficult, if not impossible, to devise an economic theory and analysis to account for the pass-on to the class as a whole, regardless of how dependable and readily available the data are. Even assuming that an expert’s calculation would accurately determine the actual damages a portion of the class sustained, it must account for the actual damages of the entire class. Wood exemplifies this paradox. The proposed class in Wood, while not as large as in Wilcox, involved pharmaceutical buyers who not only purchased their medications from a variety of allegedly disfavored pharmacies, they also purchased a variety of medications. Though the trial court had not yet certified a class, much less allowed the case to go to trial, the defense put the plaintiff’s economic theories to the test. Choosing just seven of the medications the defendants produced, the defense expert was able to show a significant disparity in the amount of overcharge passed on to consumers, ranging from zero percent to one hundred percent. These real world data had a definite effect on the trial court, which concluded that the plaintiff’s expert had no way to calculate actual damages on a class-wide basis. In summary, when the only figures available to prove damages are general in nature, they fail to prove the actual damages the class members sustained. Economic theories and analyses are most useful, and therefore come closest to satisfying the actual damages and common proof requirements for *632 pass-on, when there is a minimum, of variables involved. However, even assuming that a consistent pass-on rate exists and can be demonstrated with common proof, the plaintiff must do more than promise that this sort of calculation will be available in the future. Moreover, the burden allocated to the party moving for class certification 113 to find a method of common proof gives the party opposing class certification the luxury of reacting to the moving party’s proposed theories and analyses. By conducting studies, some very preliminary and modest in scope, the party opposing class certification has a reasonable prospect of demonstrating significant variations in the market, casting doubt on whether the moving party can prove actual damages with common proof. C. FOREIGN DECISIONS REGARDING INDIRECT PURCHASER SUITS USING SIMILAR EVIDENCE (1) MELNICK v MICROSOFT CORP In Melnick v Microsoft Corp, 114 the plaintiffs indirectly purchased Microsoft’s operating systems software and made allegations virtually identical to allegations in the instant case. In support of their motion to certify a class, the plaintiffs offered an affidavit from the same Dr. Leffler involved in this case. Their hope was that this affidavit would prove class-wide damages and show how to calculate damages relying on general economic principles because Dr. Leffler *633 did not have access to data that would allow him to calculate damages. The Melnick trial court concluded that Dr. Leffler’s analysis was a generic statement of economic theory, not facts, while Microsoft’s experts provided sufficient data and analysis about the markets in Maine to support its argument. In the end, the Melnick trial court denied the motion for certification because the plaintiffs relied only on general economic theories in support of their case and they were unable to show that such theories were workable under the applicable facts. (2) GORDON v MICROSOFT CORP The facts in Gordon v Microsoft Corp, 115 another indirect purchaser case, are virtually identical to the facts of this case. When the plaintiffs moved to certify a class, the trial court in Gordon analyzed Dr. Leffler’s affidavit and the evidence Microsoft offered, including Dr. Hausman’s affidavit. While the Gordon court discussed both sides’ positions and theories, it apparently accepted the plaintiffs’ offer of proof on the issue of both injury and damages for the class without subjecting it to a rigorous analysis on the grounds that such an analysis is best left for trial. In short, the Gordon court may have decided to certify the class because, otherwise, the case could not go forward. Microsoft appealed the trial court’s decision to Minnesota’s intermediate appellate court, which declined to review the case under its discretionary appellate authority. 116 The Minnesota Supreme Court then took *634 the case, but purposefully chose not to address the substantive issues surrounding class certification in light of the nature of Microsoft’s petition with the court. 117 Rather the court explored the factors affecting discretionary review in the lower appellate court, ultimately affirming the intermediate appellate court’s decision to deny review. 118 Thus, in June 2002, the class action against Microsoft in Minnesota was proceeding toward trial. (3) BELLINDER v MICROSOFT CORE In Bellinder v Microsoft Corp, 119 a Kansas trial court certified a class action made up of indirect purchasers of Microsoft’s operating systems software. The Bellinder court also analyzed the parties’ evidence, which included Dr. Leffler’s affidavit, and concluded that its role at the preliminary class certification stage did not extend to deciding which of the experts was correct, which was a matter for trial. Instead, the Bellinder trial court seemed to accept Dr. Leffler’s theories on injury and damages without scrutinizing those theories in light of the contrary evidence Microsoft offered. The trial court saw subclasses as the appropriate way in which to handle some of the evidentiary complexities Microsoft identified. In the end, the trial court, which emphasized the useful role private antitrust actions play and the need to protect consumers in an efficient manner, certified the class. *635 (4) CONCLUSIONS AND COMMON FACTORS In a sense, when we look at these cases from other jurisdictions in comparison to the Michigan trial court cases, the certification decisions illustrate Page’s “skeptical” versus “sanguine” paradigm. Those trial courts that have denied class action certification have tended to value accuracy over approximation, while those trial courts that have granted class action certification have taken the opposite approach, allowing “reasoned estimation” to prevail over precision. Plainly, the more rigorous the analysis, the less likely the courts were to certify the class. Though not binding on us, we gather from Holmes, Wilcox, and Wood that Michigan’s trial courts tend to interpret MARA and our state court rules as requiring a rigorous analysis. We, too, think that mara’s actual damages requirement mandates this approach. IX. THE TRIAL COURT’S DECISION A. OVERVIEW As we mentioned earlier in this opinion, the trial court in this case, without question, took considerable care with its opinion granting class action certification. The trial court identified the standards for class action certification in MCR 3.501 and engaged in an extended analysis of the rule’s five requirements. With respect to the commonality requirement, it distinguished between common questions and predominance, concentrating on predominance. The trial court set out general standards employed by federal courts to determine whether a plaintiff can establish the fact of injury (effect) by common proof before *636 reviewing the assertions and conclusions in the various affidavits. The trial court, without using the terminology adopted in this opinion and used extensively in the case law, then turned to the necessary common proof of overcharge and pass-on. B. THE OVERCHARGE REQUIREMENT The trial court explicitly noted that Dr. Leffler had described three “yardstick” methodologies that could be used to establish Microsoft’s purported overcharge to its direct purchasers: the comparative margin method, the competitive margin method, and the violation-free period method. The trial court acknowledged that Dr. Hausman challenged each of these methodologies insofar as “they do not account for the economic realities of this case.” Deciding that it was not necessary to resolve the dispute among the experts, the trial court concluded that A&M had “demonstrated through Professor Leffler that impact might be shown by common proof.” Though it does not serve as the foundation for our decision, we observe that the trial court’s approach on the overcharge requirement may have been correct. Given the findings in the federal Microsoft antitrust litigation and Dr. Leffler’s three “yardstick” methodologies, it is possible that establishing the fact of injury to direct purchasers is “ ‘virtually a mechanical task,’ ” “ ‘capable of mathematical or formula calculation’ ” 120 in this case. These proposed methodologies may be sufficiently probative of a loss suffered *637 by all the class members attributable to the antitrust violation to satisfy this overcharge requirement, 121 as well as the common proof required under MCR 3.501(A)(1)(b). This common proof that each class member suffered injury also tends to support the trial court’s finding that a class action was the superior form of lawsuit. Were the fact of injury the only substantive question relevant to the class certification question, we might affirm the trial court’s decision. However, the pass-on requirement is equally critical. C. THE PASS-ON REQUIREMENT Under a heading entitled “Damages,” again without using the terminology, the trial court attempted to address the pass-on requirement. The trial court, citing federal law, asserted that the plaintiff need not present a precise formula and that it was not necessary that the plaintiff demonstrate to a certainty that the proposed methods will succeed. Further, the trial court stated that its decision whether to certify the class did not depend on whether the proposed method would eventually succeed on the merits, but rather “whether or not the proposed methods are so insubstantial as to amount to [no] method at all.” The trial court’s citations are correct; its conclusions with respect to the pass-on requirement are not. Essentially, there are two approaches to the pass-on requirement in Dr. Leffler’s affidavits. The first is the tax incidence theory. As noted above, this theory may function as an interpretive tool in a market that is perfectly competitive. To the extent that the mar *638 ketplace for Microsoft’s products is less than perfectly competitive, its utility declines drastically. As Judge Posner once observed: A direct purchaser who finds himself paying a higher price for inputs would love to pass on all of the additional cost to his customers in the form of a higher price, but he cannot do so, because a price that much higher will so reduce the demand for his product that his profits would fall unacceptably. (If the higher price were optimal, the firm would have raised its price without waiting for its costs to increase.) The optimal adjustment by an unregulated firm to the increased cost of the input will always be a price increase smaller than the increase in input cost, and this means that the increased cost will be divided between the two tiers, the direct and indirect purchasers—but in what proportions will often be hard to determine, even by sophisticated techniques of economic analysis. This is the central insight of the Illinois Brick decision.1 122 1 The trial court paid no real attention to the parameters, the implications, and the speculative nature of the tax incidence theory. Inexplicably, it asserted that Dr. Leffler had provided an explanation that was “a means by which to show how customers were damaged.” While Dr. Leffler did refer to the “intense competition” in the software market in his affidavits, he did not describe a formula or method demonstrating that indirect purchasers were damaged, other than the assertion that the pass-on might be one hundred percent. Instead, Dr. Leffler repeatedly mentioned *639 “basic economic principles,” estimated end-user damages calculated by “using a class wide statistical determination,” “extensive economics literature,” “sound, accepted economic and statistical methodologies,” “relatively simple statistical estimation[s],” and “[m]odem computer systems and statistical tools.” Bluntly put, the absence of any meaningful details in Dr. Leffler’s affidavits make these statements slogans, not methods of proof. We appreciate that Dr. Leffler did not have all the data necessary to put forth the evidence he might be able to present at a trial following further discovery. We also appreciate that it is unlikely that A&M has the same financial resources as Microsoft with which to secure preliminary studies. Nevertheless, these broad, nonspecific references fail to describe a method or formula by which a court could determine that Microsoft’s conduct caused actual damages or injury to each class member, even when the assertions are read in context. 123 Even when attempting to minimize the effect of Dr. Hausman’s criticisms, Dr. Leffler continued to refer to his methodologies as a viable way to estimate damages for the class. Thus, A&M failed to articulate a method or formula for calculating the pass-on when it relied wholly on Dr. Leffler’s tax incidence theory and generalized slogans. Perhaps recognizing the inherent difficulties with the incidence theory, Dr. Leffler also suggested in Lef-fler I that regression analyses might be used to calculate the rate at which Microsoft’s direct purchasers passed on the overcharge to indirect purchasers. That may be true. However, the affidavits provide no rea *640 sonable basis on which a court might reach the conclusion that a regression analysis would actually compute the amount of the pass-on rate on a class-wide basis or with respect to individual consumers. Again, the early stage of the proceedings likely did not allow Dr. Leffler sufficient data to perform—or even describe in any great detail—a regression analysis or results. However, the timing of a motion for class action certification within the first ninety-one days after the complaint is filed 124 is entirely a matter that rests in the plaintiffs hands. The decision when to move for class certification is tactical, resting in part on the degree to which the lawsuit’s economic feasibility depends on whether it proceeds as a class action and the likely expense of learning, at a later date, that certification is not possible. Nevertheless, these factors do not obviate a plaintiff’s burden with respect to articulating a method or formula by which a court could determine that the defendant’s conduct caused each member of the proposed class actual damages. A&M did not meet that burden through Dr. Leffler’s references to regression analyses. Even when viewed collectively, none of Dr. Leffler’s methods provided the class-wide common proof of actual damages that MCR 3.501(A)(1)(b) and MARA, mandate. In addition to our conclusion that Dr. Leffler’s vague promises for future analysis failed to bridge the gap between economic theory and the reality of economic damages, some of the concerns that surfaced in Holmes, Wilcox, and Wood also concern us when analyzing whether a class action would be the superior form of lawsuit, as MCR 3.501(A)(1)(e) requires. *641 For instance, this case involves no less than six products sold over a number of years through numerous retailers. As the trial court observed in Wood, this proliferation of factors expands the possibility that the overcharge for each product varied over time, varied at a different rate from the other products, and was passed on at different rates to indirect purchasers. Dr. Nichols’ study confirmed this premise from Wood, suggesting that the pass-on rates not only varied widely for Microsoft’s products, but were sensitive to the price points used to market them. Even if the relevant market is perfectly competitive and results in one hundred percent pass-on of the overcharge, 125 which Dr. Hausman disputes, Dr. Leffler’s affidavits fail to assure us that the methodologies he has proposed can account or adjust for this breadth of remaining variation for the class. Additionally, though the proposed class in this case is smaller than the proposed class in Holmes, we cannot ignore the prevalence of Microsoft’s products. The class, if certified, would likely be immense, numbering in the hundreds of thousands. While the size of the class and other factors certainly suggest that join-der would be impracticable and that a number of the factors favoring certification in MCR 3.501(A)(2) exist, the paradox we have discussed makes the many variations between class members problematic when it comes to proving actual damages. In short, this case has all the hallmarks of being unmanageable. Dr. *642 Leffler’s methodologies, even if they were to work with respect to small, well-defined subclasses that group class members by a very few strongly unifying characteristics, will essentially require separate trials to determine the different pass-on rates affecting the class as a whole. MCR 3.501(A)(2)(c) suggests that when a proposed class action is unmanageable, a trial court should deny class certification. Though the trial court was willing to accept the task of handling a case this large and complex, we are left with the definite feeling that it made a mistake when it concluded that the plaintiffs had a satisfactory method of demonstrating that a class action was the superior form of adjudicating this dispute. X. CONCLUSION We hold that A&M has not set forth a viable method for proving actual damages on a class-wide basis that MARA, MCL 445.778(2), and the court rules require for a plaintiff in an indirect purchaser suit brought as a class action. We also hold that A&M has not satisfied its burden of proving that a class action would be a superior way to resolve this dispute. Thus, the trial court clearly erred when it certified the class proposed for this lawsuit. We note, however, that MCR 3.501(B)(3)(e) permits A&M as a named plaintiff to maintain this action against Microsoft, relying on individualized proofs. Reversed and remanded for additional proceedings consistent with this opinion. We do not retain jurisdiction. Footnotes [1] See Coutroulis and Allen, The pass-on problem in indirect purchaser class litigation, 44 Antitrust Bull 179 (March 22, 1999), available in 1999 WL 19535295 without page numbers. [2] Id. [3] Hanover Shoe, Inc v United Shoe Machinery Corp, 392 US 481, 491-493; 88 S Ct 86; 20 L Ed 2d 1231 (1968). [4] Illinois Brick Co v Illinois, 431 US 720, 730-735; 97 S Ct 2061; 52 L Ed 2d 707 (1977). [5] See 1984 PA 274, § 8. [6] MCL 445.778(2) (emphasis supplied). [7] Microsoft previously filed an application for leave to appeal with this Court to challenge Judge Burress’ order denying its motion to stay these proceedings, but we declined to grant leave to appeal. See Barnard v Microsoft Corp, unpublished order of the Court of Appeals, entered March 31, 2000 (Docket No. 225694). [8] In re Microsoft Antitrust Litigation, 463 Mich cli (2000). [9] MCL 445.773. [10] MCL 445.772. [11] Zine v Chrysler Corp, 236 Mich App 261, 270; 600 NW2d 384 (1999). [12] Id. [13] See Gravity, Inc v Microsoft Corp, 127 F Supp 2d 728 (D Md, 2001); In re Microsoft Corp Antitrust Litigation, 127 F Supp 2d 702 (D Md, 2001). [14] See, e.g., United States v Microsoft Corp, 84 F Supp 2d 9 (D DC, 1999) (Microsoft I) (factual findings); United States v Microsoft Corp, 87 F Supp 2d 30 (D DC, 2000) (Microsoft II) (legal conclusions); United States v Microsoft Corp, 97 F Supp 2d 59 (D DC, 2000) (Microsoft III) (remedy). [15] See Microsoft I, supra. [16] 15 USC 1 and 2. [17] Microsoft II, supra at 56. [18] See Microsoft III, supra at 61. [19] See id. at 63-70. [20] Microsoft Corp v United States, 530 US 1301; 121 S Ct 25; 147 L Ed 2d 1048 (2000). [21] See United States v Microsoft Corp, 346 US App DC 330; 253 F3d 34 (2001) (Microsoft TV). [22] See id. at 342, 376-380. [23] Id. at 342. [24] Id. at 347. [25] Id. at 342. [26] Id. at 342, 413-415. [27] Microsoft Corp v United States, 534 US 952; 122 S Ct 350; 151 L Ed 2d 264 (2001). [28] See United States v Microsoft Corp, unpublished memorandum opinion of the United States District Court for the District of Columbia, issued July 2, 2002 (Docket No. 98-1232), slip op at 3, n 2 <http://www.dcd. uscourts.gov/98-1232ci.pdf> (Microsoft V). [29] See United States v Microsoft Corp, unpublished order of the United States District Court for the District of Columbia, entered November 8, 2001 (Docket No. 98-1233) <http://www.dcd.uscourts.gov/98-1232 qq.pdf>; see also Ted Bridis, “Michigan accepts revised Microsoft settlement,” Detroit Free Press, November 7, 2001 <http://www.freep.com/ money/tech/micro7_20011107.htm>. [30] See United States v Microsoft Corp, unpublished order of the United States District Court for the District of Columbia, entered February 1, 2002 (Docket No. 98-1232) <http://www.dcd.uscourts.gov/98-1232ae.pdf>. [31] 15 USC 16(b)-(h). [32] See Microsoft V, supra at 4r6, 36. [33] See Microsoft I & II. [34] See Microsoft II, supra at 54, n 7. [35] MCL 445.780. [36] See Blair and Herndon, A note on Hanover Shoe: (calculating damages due to monopoly), 43 Antitrust Bull 351 (1998), available in 1998 WL 16568463 without page numbers. [37] Coutroulis, supra, quoting Hanover Shoe, supra at 492-493. [38] Minois Brick, supra at 731-733 (internal citations omitted). [39] See Coutroulis, supra. [40] See California v ARC America Corp, 490 US 93, 101-102; 109 S Ct 1661; 104 L Ed 2d 86 (1989). [41] See Page, The limits of state indirect purchaser suits: Class certification in the shadow of Illinois Brick, 67 Antitrust L J 1, 2 (1999). [42] Page, supra at 17 (emphasis supplied). [43] Id. at 18-19 (emphasis added and citations to Justice Brennan’s dissent in Ittinois Brick omitted). But see Coutroulis, supra (noting irony in the fact that since Illinois Brick, some repealer states now bar class actions for indirect purchasers because of the same concerns that led the United States Supreme Court to bar pass-on suits under federal law). [44] See Neal v James, 252 Mich App 12, 16; 651 NW2d 181 (2002). [45] Zine, supra at 287. [46] Zine, supra at 289. [47] Grigg v Michigan Nat’l Bank, 405 Mich 148, 184; 274 NW2d 752 (1979). [48] See In re Polypropylene Carpet Antitrust Litigation, 178 FRD 603, 618 (ND Ga, 1997) (“[A]t the class certification stage, Plaintiffs must show that antitrust impact can be proven with common evidence on a classwide basis; Plaintiffs need not show antitrust impact in fact occurred on a classwide basis.”). [49] See In re Cardizem CD Antitrust Litigation, 200 FRD 326, 331 (ED Mich, 2001). [50] In re Domestic Air Transportation Antitrust Litigation, 137 FRD 677, 692 (ND Ga, 1991). [51] Bogosian v Gulf Oil Corp, 561 F2d 434, 456 (CA 3, 1977). [52] See Peters v Cars to Go, Inc, 184 FRD 270, 278 (WD Mich, 1998); Van Vels v Premier Athletic Center of Plainfield, Inc, 182 FRD 500, 509 (WD Mich, 1998); McCoy v Salem Mortgage Co, 74 FRD 8, 12 (ED Mch, 1976). [53] In re Plywood Anti-Trust Litigation, 76 FRD 570, 583-584 (ED La, 1976). [54] Dix v American Bankers Life Assurance Co of Florida, 429 Mich 410, 414, n 6; 415 NW2d 206 (1987), discussing Grigg, supra. [55] Lee v Grand Rapids Bd of Ed, 184 Mich App 502, 504-505; 459 NW2d 1 (1989). [56] MCR 3.501(A)(2). [57] See Zine, supra at 289, n 14. [58] Footnotes omitted throughout these passages. [59] In Leffler I, Dr. Leffler also referred to regression analyses that might determine the amount of Microsoft’s alleged overcharge passed on to indirect purchasers, although he provided no real detail concerning how a regression analysis might function and what it would actually reveal. [60] Page, supra at 13-14, 16. [61] Id. at 16. [62] There were other affidavits filed in this case, but we believe the ones that we summarize here are the critical ones. [63] The proof of service attached to Hausman I erroneously shows June 14, 2000, as the date for the affidavit. Other documentation in the record and references in the affidavit itself make clear that the proper date was June 14, 2001. Regardless, this date does not change our analysis. [64] There are opinions that decline to certify class actions against Microsoft, but which are of limited value in this case either because they do not come from Illinois Brick repealer jurisdictions or because they involve statutes that protect consumers, but not necessarily on antitrust grounds. See, e.g., Pomerantz v Microsoft Corp, 50 P3d 929 (Colo App, 2002); Vacco v Microsoft Corp, 260 Conn 59; 793 A2d 1048 (2002); Berghausen v Microsoft Corp, 765 NE2d 592 (Ind App, 2002); Davidson v Microsoft Corp, 143 Md App 43; 792 A2d 336 (2002); Arnold v Microsoft *616 Corp, unpublished opinion per curiam of the Kentucky Court of Appeals, issued November 21, 2001 (Docket No. 2000-CA-002144-MR); O’ConneU, v Microsoft Coip, unpublished memorandum opinion and order of the Massachusetts Superior Court, issued June 14, 2001 (Docket Nos. CIV.A 00-01743, CIV.A 00-02456); Daraee v Microsoft Corp, unpublished order of the Oregon Circuit Court, entered June 27, 2000 (Docket No. 0004 03311). Thus, because those cases are not particularly relevant, we do not examine them. [65] Holmes v Abbott Laboratories, Inc, unpublished order of the Calhoun Circuit Court, entered July 10, 1995 (Docket No. 94-744-CP). [66] Microsoft has provided this Court with the transcripts of the class certification hearing in Holmes, which the trial court in that case held on May 22, 1995. Because the trial court’s order is particularly brief, we refer to pages of the transcript. [67] Id. at 6, 75. [68] Id. at 8. [69] Id. [70] Id. at 8-9. [71] Id. at 9. [72] See id. at 12. [73] Id. at 15. [74] Id. [75] Id. at 27. [76] Id. at 16. [77] Id. at 78. [78] Id. at 26. [79] Id. at 70. [80] Id. at 76. [81] Id. at 77. [82] Id. at 82-83. [83] Id. at 83. [84] Id. [85] Id. at 86. [86] Id. at 80. [87] Id. at 86-87. [88] Id. at 87. [89] Wilcox v Archer-Daniels-Midland Co, unpublished opinion and order of the Ingham Circuit Court, issued September 29, 1997 (Pocket No. 96-82473-CP). [90] Id. at 1-2. [91] Id. at 2. [92] Id. [93] Id. [94] Id. [95] Id. at 3. [96] Id. at 4. [97] Id. at 3. [98] Id. at 3-4. [99] Id. at 4. [100] Id. at 5. [101] Id. at 5. [102] Id. at 6. [103] Id. [104] Id. [105] Id. [106] Id. at 6-7. [107] Id. [108] Id. at 7. [109] Id. [110] Id. [111] Wood v Abbott Laboratories, unpublished opinion of the Oakland Circuit Court, issued September 11, 1997 (Docket No. 96-512561-CZ), pagination unavailable. [112] See MCR 3.501(A)(2)(d) and (e). [113] Typically, but not always, the party who moves for class certification in a class action is the plaintiff. [114] Melnick v Microsoft Corp, unpublished opinion of the Maine Superior Court, issued August 24, 2001 (Docket Nos. CV-99-709, CV-99-752), pagination unavailable. [115] See Gordon v Microsoft Corp, unpublished memorandum opinion of the Minnesota District Court, issued March 30, 2001 (Docket No. 00-5994). [116] Gordon v Microsoft Corp, 645 NW2d 393, 396 (Minn, 2002). [117] Id. at 396-397. [118] Id. at 397, 403. [119] Bellinder v Microsoft Corp, unpublished memorandum opinion of the Sedgwick County, Kansas District Court, issued September 7, 2001 (Docket No. 00-C-0855). [120] See Alabama v Blue Bird Body Co, Inc, 573 F2d 309, 326-327 (CA 5, 1978), quoting Windham v American Brands, Inc, 565 F2d 59, 68 (CA 4, 1977). [121] See Blue Bird, supra at 325-326, following Bogosian, supra. [122] Illinois ex rel Hartigan v Panhandle Eastern Pipe Line Co, 852 F2d 891, 894 (CA 7, 1988), overruled on other grounds sub nom Illinois ex rel Burris v Panhandle Eastern Pipe Line Co, 935 F2d 1469 (CA 7, 1991); see also In re Brand Name Prescription Drugs Antitrust Litigation, 123 F3d 599, 605 (CA 7, 1997) (“Tracing a price hike through successive resales ... is famously difficult.”). [123] See In re Plywood Anti-Trust Litigation, supra at 583-584. [124] See MCR 3.501(B)(1)(a). [125] Notably, even the United States Supreme Court is willing to consider the possibility that Illinois Brick would permit an indirect purchaser to sue “when, by hypothesis, the direct purchaser will bear no portion of the overcharge and otherwise suffer no injury.” Kansas v Utilicorp United, Inc, 497 US 199, 218; 110 S Ct 2807; 111 L Ed 2d 169 (1990).