1 JCOMLE 303                                                    Journal of Competition Law & Economics                      CLASS CERTIFICATION IN THE MICROSOFT INDIRECT PURCHASER LITIGATION                                                          June, 2005 (Approx. 35 pages) 1 J. Competition L. & Econ. 303 Journal of Competition Law & Economics June, 2005 *303 CLASS CERTIFICATION IN THE MICROSOFT INDIRECT PURCHASER LITIGATION William H. Page [FNa1] Copyright © 2005 by Oxford University Press; William H. Page ABSTRACT The indirect purchaser cases brought against Microsoft under state law in the wake of the government's victory in United States v. Microsoft have been certified as class actions much more frequently than other indirect purchaser cases. In this study, I explore possible explanations for this disparity, including lenient standards of certification; factual characteristics of the Microsoft cases that might be favorable to certification; and the supposed preclusive effect of findings in the government case. In the conclusion, I consider whether the certification of these cases under lenient standards has been in the public interest. I. BACKGROUND The government's monopolization case against Microsoft has spawned a myriad of private lawsuits seeking treble damages. [FN1] This study examines an important subset of those lawsuits, the class actions brought under state law on behalf of "indirect purchasers" of Microsoft's software, mostly consumers and businesses who bought the software along with new computers from Original Equipment Manufacturers (OEMs) like Compaq, Gateway, and Dell. Although the Illinois Brick [FN2] rule denies indirect purchasers the right to sue under the federal antitrust laws, many states have authorized these suits under state antitrust or consumer protection law. [FN3] Before the Microsoft cases, however, courts in the "Illinois Brick repealer" states had refused, more often than not, to certify indirect purchaser suits as class actions on the grounds that issues specific to the individual plaintiffs, particularly the issue of impact, "predominated" over the issues common to the class. [FN4] In these courts' "skeptical" view, the complexities of proving whether and by how much the direct purchasers had passed on the overcharge to the plaintiffs would likely overwhelm any common *304 issues, like questions of liability, and thus make the cases unsuitable for trial as class actions. The courts in the Microsoft indirect purchaser cases, however, have certified them as class actions much more frequently. Of the fourteen opinions in these cases, three (two in Michigan and one in Maine) denied certification, [FN5] while eleven granted it. [FN6] In this study, I explore possible explanations for this disparity in the rate of certification, and consider tentatively whether the experience in these cases helps us decide whether indirect purchaser class actions are in the consumer interest. Antitrust law, at both state and federal levels, is ostensibly about protecting the consumer interest from private conduct that reduces competition. [FN7] Courts, for example, interpret the substantive law to prohibit conduct that harms consumers rather than conduct that harms only competitors or small businesses. [FN8] Paradoxically, however, consumers often cannot sue for the harms they suffer as a result of federal antitrust violations. [FN9] Although they have the right to sue under federal law for their antitrust injuries, [FN10] most are foreclosed from doing so by the rule of Illinois Brick, which denies indirect purchasers the right to sue for damages from monopolistic overcharges. Consumers are usually indirect purchasers, because they do not often buy monopolized or cartelized products directly from the offenders. But the absence of a remedy for consumers is not necessarily inconsistent with the goal of protecting the consumer interest. Whether a statute protects *305 a class is a separate question from whether it would efficient for that class to have a private right of action. Many federal statutes, such as the Federal Trade Commission Act, are designed to protect consumers, but do not accord them a private right to sue, presumably because to do so would not be the most efficient mechanism for securing compliance with substantive standards. [FN11] Similarly, the Supreme Court in Illinois Brick reasoned that concentrating the right to recovery in direct purchasers would ensure more effective deterrence. [FN12] It would, for example, be very difficult to apportion the overcharge among the various levels of a chain of distribution. Although the theoretical incidence of an overcharge is clear, the theory depends on facts that are usually not available and assumptions that are rarely met. Nevertheless, to many readers of Illinois Brick, it seemed anomalous that direct purchasers who have passed on most of the overcharge could recover treble damages for all of it, while consumers, to whom at least a share of the overcharge was passed on, could recover nothing. This anomaly was sufficiently disturbing to cause a political backlash. Many states, under the banner of compensating consumers, have enacted so-called Illinois Brick repealer statutes that allow indirect purchasers to sue in state courts for passed-on overcharges. [FN13] Other states' courts have interpreted pre-existing antitrust or consumer protection laws to authorize indirect purchaser standing. [FN14] The Supreme Court in ARC America [FN15] upheld these statutes. Although a third group of states has followed the federal courts in barring indirect purchaser suits, [FN16] they are now in the minority. Indirect purchaser statutes have certainly been effective in generating litigation on behalf of consumers. [FN17] Most antitrust violations of national scope are followed by class action lawsuits in the repealer states behalf of indirect purchasers. The lawsuits are, by and large, in addition to parallel lawsuits in the federal courts on behalf of direct purchasers. [FN18] (The Microsoft cases are a partial exception to this generalization; I will return to this point later in this study.) The question remains, however, whether indirect purchaser class actions have been *306 effective in the more meaningful sense of benefiting the consumers who actually paid overcharges, either by providing compensation or more effective deterrence. Before the present study, I have examined this issue in two contexts, and have discovered that the states that repealed Illinois Brick have not managed to repeal the difficulties that the Supreme Court predicted would plague indirect purchaser cases. Those difficulties remain persistent obstacles to compensation of consumers and to effective antitrust enforcement. First, I have studied class certification decisions in those states that authorize indirect purchaser suits, focusing whether class issues "predominate" over individual issues, as required by class action rules. In an article published in 1999, [FN19] I reported striking differences in approach to the certification inquiry. [FN20] Some courts freely certified indirect purchaser suits as class actions with minimal scrutiny of the evidence on which the plaintiffs expected to prove classwide impact. I characterized the approach of these courts, which I traced to the dissent in Illinois Brick, as the "sanguine view" of the problems of proving passing on. [FN21] Other courts, however, have scrutinized the proposed theories of classwide proof, and have identified a multitude of impediments to providing relief to the consumers who actually paid an overcharge. I characterized this approach as the "skeptical view." [FN22] Obviously, the states with the most skeptical view of these issues have not repealed Illinois Brick at all. But even in some states that have repealed Illinois Brick, courts take seriously the difficulties the Supreme Court identified in that case, and insist that representatives of the plaintiff class account for them in their proposals for certification. In most instances, the courts taking the skeptical view have denied class certification. The sanguine and skeptical views amount to different standards of certification, even if the courts do not acknowledge them as such. There is a well-recognized tension in law governing certification. [FN23] On the one hand, Eisen laid down the principle that courts should not "conduct a preliminary inquiry into the merits of the suit in order to determine if it may be maintained as a class action." [FN24] On the other hand, Falcon requires that courts conduct a "rigorous analysis" of the requirements for certification to assure "actual, not *307 presumed" compliance. [FN25] Some courts have interpreted the Falcon standard to require consideration of the plaintiffs' evidence, not merely their allegations; and that inquiry can involve "entanglement with the merits." [FN26] As one of the certification cases I studied in 1999 stated the dilemma, the court must "walk the fine line between a rigorous analysis of the basic claims and methods of proof presented by the plaintiffs and the inappropriate delving into an assessment of the merits of those claims." [FN27] The stakes in this balance are extraordinarily high. [FN28] Courts emphasize that denying certification of classes with small individual damages will likely mean denying relief. [FN29] They do sometimes recognize that granting certification will place great pressure on the defendant to settle for economic reasons, [FN30] although they mitigate this concern by holding out the (mostly theoretical) possibility of decertification at a later stage. [FN31] Sanguine and skeptical courts balance these considerations in different ways. Sanguine courts suggest that "any doubts as to whether the class should be granted certification should be resolved in favor of certification." [FN32] More important, they examine proposed expert testimony only to determine if meets some minimal standard of validity, [FN33] in many cases requiring only that the expert propose a methodology that is "not so insubstantial as to amount to no method at all." [FN34] Moreover, they refuse to resolve a battle of experts, because to do so would implicate the merits. [FN35] Skeptical courts, by contrast, insist that the experts' methods "bridge the gap between theory and reality" [FN36] by *308 testing, at least preliminarily, whether the evidence will permit proof of harm by classwide proof. One might fairly characterize this sort of inquiry as going to the merits. Skeptical courts sometimes justify their inquiry as a substantive requirement of their state's statute. [FN37] The issue of predominance requires courts to anticipate "what substantive issues will be important and how much of the litigation those issues will consume." [FN38] To the extent courts examine evidence in support of those issues, they inquire into the merits. Predicting whether the plaintiffs' evidence will be able to establish harm by classwide proof involves an estimate of the likelihood of success on a substantive issue. [FN39] I argued in 1999 that whether a state applied one or the other standard was the single most important determinant of whether a class would be certified. [FN40] Apart from these standards of certification, however, I identified factual characteristics of classes that make proving passing on more difficult. [FN41] One characteristic is simply the availability of evidence. It counts against certification if the class consists of millions of consumers who purchased the product in frequent, undocumented transactions from a variety of sellers. [FN42] Equally important are factors that complicate the task of calculating either the amount of the initial overcharge or the rate at which it was passed on to the class members. Proof is more complex, for example, if the initial overcharge varies over time; if firms at the intermediate levels of distribution have market power on the buying side or on the selling side (because variations in market power make buying prices and markup policies less consistent); if there is more than one intermediate level through which the overcharge is passed; or if the putative class members themselves could have passed on the overcharge to others. [FN43] Certification is also less likely if the monopolized product comes in many varieties, or if it is transformed or incorporated into other products in the chain of distribution. [FN44] I argued that these obstacles implied only a small subset of injured consumers will ever be members of an indirect purchaser class. The most difficult "ingredient" cases, for example, will never even be brought. Notice that these limitations do not mean there was no overcharge passed on to consumers; they only mean it would be impossible to prove it, at least by classwide proof. The second context in which I have examined the effects of indirect purchaser litigation on consumers is in the settlements indirect purchaser class actions. In an article published in 2003, John Lopatka and I found that indirect *309 purchaser class actions provided very little recovery to consumers who demonstrably paid overcharges. [FN45] The difficulties that the court identified in Illinois Brick were invoked as justifications for cy pres payments to charities and other beneficiaries. Worthy as these beneficiaries may be, they are not consumers who paid overcharges. [FN46] We concluded that deterrence should remain the paramount goal of antitrust enforcement. Since my first article on class certification appeared, there have been many new decisions involving indirect purchaser class actions, [FN47] including some in federal court certifying multiple class actions under state law. [FN48] More than half of the new indirect purchaser certification decisions have been the cases brought against Microsoft. In this article, I consider why courts have so overwhelmingly certified these cases as class actions. [FN49] In the following Part, I place the Microsoft indirect purchaser cases in the context of indirect purchaser litigation generally, analyzing how the rate of certification has been affected by the state in which the case was filed, the time when the case was filed, and the product that the case involved. That inquiry reveals that, while the high rate of certification may be partly attributable to the cases have been filed in "sanguine" states, or by a possible general trend in recent years toward greater leniency in certification, it also seems partly attributable to the specific character of the cases as Microsoft cases. Consequently, in Part III, I place the Microsoft indirect purchaser cases in the context of the much larger world of Microsoft litigation, pointing out that the government case does not support the claim of an overcharge. In Part IV, I examine the reasoning of the Microsoft certification cases in an effort to explain their unusually high rate of certification. I find that most of the difference appears to be attributable to the standard of certification, but I also identify characteristics of the case that courts identified as warranting certification. I separately consider one of these characteristics, the fact that these cases followed the government's victory in the Microsoft monopolization case. I suggest that some of the courts granting certification may have overestimated the relevance of the government case. Finally, in *310 the Conclusion, I consider whether the overwhelming certification and subsequent settlement of these cases was in the public interest. II. THE PLACE OF THE MICROSOFT INDIRECT PURCHASER CASES IN INDIRECT PURCHASER LITIGATION In this Part I try to explain why the Microsoft cases have granted certification so frequently by looking for possible correlations between the results in indirect purchasers cases and a few simple factors. I classified the cases listed in the Appendix and their results according to year, product, state, and result, and calculated the frequency of certification according to each variable. For clarity, I have limited these statistics to cases in which the court considered classes limited to a single state; at the end of the section, I discuss separately three cases in which federal courts certified indirect purchaser classes in multiple states. To summarize the results at the outset, I found that indirect purchaser classes are being certified more frequently since my last study, but that most of that change is attributable to the Microsoft cases. Whether a Microsoft indirect purchaser class is certified appears to depend in large part on whether it is filed in a skeptical or sanguine jurisdiction. Most of the states certifying Microsoft classes never failed to certify an indirect purchaser class; Michigan, which denied certification, has never certified one. But some courts in previously skeptical jurisdictions also certified Microsoft classes. Table 1 shows that the high frequency of certification of the Microsoft cases does not simply reflect a general increase in frequency of certification of indirect purchaser cases since my last examination of the issue. Table 1. Rate of Certification by Year of Decision       --------------------------------------------------------                 Decisions          Granted  % Granted   --------------------------------------------------------                                                         Pre-1999         27                      10  37          1999-2004                                                 Total          27                      18  67           Non-Microsoft  13                       7  54           Microsoft      14                      11  79          Total            54                      28  52          -------------------------------------------------------- Courts have certified about half of all of the proposed classes to date, but the frequency has changed in recent years. Courts certified just over a third of the proposed classes before 1999, the period I considered in my last study. Since then, the courts have certified classes almost twice as frequently. But it turns out that the increase in frequency of certification is largely attributable to the Microsoft cases, [FN50] which account for more than half of more recent decisions and have been certified at a rate of almost 80%. It is true that half of the more recent non-Microsoft cases have been certified, an increase in frequency over *311 the older set of cases, but far less than the rate in the Microsoft cases. This increase may be attributable to class counsel having learned from experience to file better cases in more favorable jurisdictions. Table 2 shows the frequency of certification of class actions involving the same product. Table 2. Rate of Certification by Product                -------------------------------------------------------- Product                    Decisions  Granted  % Granted --------------------------------------------------------                                                         Caustic Soda/chlorine              1        0          0 HFCS and citric acid               1        0          0 Lysine                             1        0          0 Thermal fax paper                  2        0          0 Tissue                             1        0          0 Sugar                              1        0          0 Milk                               1        0          0 Cigarettes                         5        1         20 Contact lenses                     2        1          0 Infant formula                     7        3         43 Methionine                         4        2         50 Prescription drugs [FN50]          9        5         56 Microsoft software                14       11         79 Compact disks                      1        1        100 Glass containers                   1        1        100 Vitamins                           3        3        100 Total                             53       27         51 -------------------------------------------------------- There have been far more certification decisions involving Microsoft software than any other product, and the vast majority of those classes have been certified. Classes involving other products have been certified at higher and lower rates. Although the samples are too small to draw any strong conclusions, they do suggest that there are characteristics of products and how they are distributed that may influence whether a class alleging an overcharge on the product will be certified. Table 3 shows the frequency with which each state has certified indirect purchaser classes, both generally and in Microsoft cases. Table 3. Rate of Certification by State                                ---------------------------------------------------------------------- State           Decisions  Granted  % Granted  Microsoft cases/granted ----------------------------------------------------------------------                                                                       Alabama                 5        0          0  0/0                     Mississippi             1        0          0  0/0                     Michigan                6        0          0  2/0                     Minnesota              10        2         20  2/2                     Maine                   3        1         33  1/0                     Florida                 2        1         50  1/1                     Wisconsin               7        4         57  1/1                     California              6        6        100  1/1                     DC                      1        1        100  0/0                     Iowa                    1        1        100  1/1                     Kansas                  4        4        100  1/1                     New Mexico              1        1        100  1/1                     North Carolina          1        1        100  0/0                     North Dakota            1        1        100  1/1                     South Dakota            2        2        100  1/1                     Tennessee               3        3        100  1/1                     Total                  54       28         51  14/11                   ---------------------------------------------------------------------- This Table shows why I suggest that states have applied different standards of class certification. Of the sixteen states that have issued written certification decisions, nine have granted certification in every case, while three states have denied certification in every case. [FN51] Michigan has always denied certification, and did so in its two Microsoft cases. Six other states have always granted *312 certification, and did so likewise in their Microsoft cases. Notice that the Microsoft case was the only indirect purchaser class action in some of these states; in fact, every state that had no track record in the earlier period certified their Microsoft case. Four states are less consistent in their application of certification standards. Of these, all but Maine granted certification in their Microsoft cases. Maine is thus of interest, because it is the only state to deny certification in a Microsoft case while granting certification in another case. Minnesota is also interesting, because it has only granted certification twice in ten decisions, both times in the same Microsoft case--one of those decisions was the initial grant of certification, [FN52] and one was the refusal to decertify the case after discovery, even though plaintiff's proposed methods for proof of classwide harm had changed. As we will see in Part IV, the court in that case took some pains to explain why the characteristics of the Microsoft case justified a different result from all of the other Minnesota certification cases. In the forgoing tables, I have only considered cases, almost all in state courts, involving proposed classes that focused on a single state. In 2004, *313 however, three federal district courts certified multiple state classes of indirect purchasers in cases alleging harms from monopolistic conduct involving drug patents. [FN53] These cases were sufficiently different from the single-state cases to warrant separate mention. The courts managed to find federal subject matter jurisdiction. [FN54] Two of these cases certified proposed classes in seventeen states that permitted indirect purchaser suits. These courts applied lenient federal standards of certification, without even acknowledging differences in state standards for class actions. They both, for example, certified classes for the very skeptical states (such as Michigan and Alabama) and even certified classes for Mississippi, which does not even authorize class actions in its state courts. One might argue that Hanna v. Plumer [FN55] requires federal courts to apply Federal Rule 23 in certification of class actions alleging violations of state law. Interestingly, however, one of the federal cases refused to certify exemplar classes in Florida, Maine, Michigan, and Minnesota because courts in those states "have interpreted their respective indirect purchaser statutes with a more 'skeptical view' of antitrust policies and remedies" [FN56] and that their statutes "require a somewhat stronger and more precise showing of individual impact." [FN57] Evidently, this court viewed the distinction between skeptical and sanguine views as a substantive matter of state law. [FN58] With the enactment of the Class Action Fairness Act, [FN59] more indirect purchaser classes will likely be heard in federal court because of that statute's relaxed requirements for diversity jurisdiction and for removal. On the evidence of these cases, it appears that those federal courts that treat certification as a federal issue will apply a relatively sanguine standard. [FN60] Those that look to state standards will presumably replicate the split between sanguine and skeptical states in their decisions. *314 III. THE PLACE OF THE MICROSOFT INDIRECT PURCHASER CASES IN THEWORLD OF MICROSOFT ANTITRUST LITIGATION The foregoing discussion leaves open the possibility that one reason for the higher rate of certification in the Microsoft cases may lie in the characteristics of the Microsoft antitrust litigation. Consequently, it may be helpful to describe briefly the relationship between the Microsoft indirect purchaser cases and the government litigation that preceded them. In 1999, Judge Thomas Penfield Jackson issued findings of fact accepting the government's contention that Microsoft had bundled its Internet Explorer browser with its dominant Windows operating system, and imposed other contractual and technological measures, all with the goal of limiting usage of Netscape's Navigator browser and Sun's Java programming language, and thus preventing them from evolving, together or separately, into a rival platform for applications software. [FN61] The following year, he held those actions unlawful under Sections 1 and 2 of the Sherman Act, [FN62] and ordered sweeping structural and injunctive relief. [FN63] The D.C. Circuit reversed Judge Jackson's remedial order and his holdings of liability for attempted monopolization of the browser market and for tying Windows and Internet Explorer, but affirmed many of the holdings that Microsoft had illegally maintained its monopoly in the market for Intel-compatible PC operating systems. [FN64] Microsoft later agreed with the Department of Justice and with most of the states to a consent decree, which was upheld by the District Court and the D.C. Circuit. [FN65] In follow-on litigation, [FN66] virtually all of Microsoft's significant rivals sued for damages from Microsoft's exclusionary practices, [FN67] and most of these cases have settled. [FN68] The picture for purchasers is more complicated. Significantly, *315 some of the largest direct purchasers, OEMs, have not sued, at least as long as they have remained in business. [FN69] Other purchasers of Microsoft software, however, have filed numerous cases, many of them class actions, in both state and federal court seeking both damages and injunctive relief under both state and federal law. On various jurisdictional grounds, Microsoft removed to federal court a number of the cases filed initially in state court. The Judicial Panel on Multidistrict Litigation then transferred to the District of Maryland over one hundred of the actions pending in federal court, including some consumer actions based on state law. [FN70] The district court in Maryland subsequently dismissed the federal indirect purchaser claims and some of the state claims on a variety of grounds, [FN71] and declined to certify some classes. [FN72] Consumers who purchased software directly from Microsoft successfully formed a class [FN73] and later settled. [FN74] The class actions that are the subject of this study are part of a larger group of follow-on actions by consumers who bought Microsoft products from intermediaries. In at least one case, consumers alleged that they were actually direct purchasers, and therefore not barred by Illinois Brick, because the OEMs from whom they purchased had acted in concert with Microsoft, *316 but the court rejected this argument because the plaintiffs had failed to allege that OEMs violated the antitrust laws [FN75] or were co-conspirators with Microsoft. [FN76] Numerous state-law indirect purchaser claims, however, remained pending in state courts. Some of these cases have been dismissed either on jurisdictional grounds [FN77] or on the grounds that state law did not provide a right of recovery to indirect purchasers. [FN78] Other courts, however, have allowed indirect purchaser cases to proceed under state law. [FN79] Of these, fourteen have issued written opinions on the question of class certification; [FN80] this set of cases is the subject of the present study. Many of the indirect purchaser cases have now been settled in amounts that, as of Jan. 2005, totaled $1.87 billion. [FN81] Follow-on cases are common because of the "widespread belief that if the Antitrust Division acts, there must be substantial reason for its action." [FN82] If the government case results in a favorable judgment, it also may collaterally estop the defendant on some issue of liability, even in a lawsuit in state court under state antitrust law. [FN83] But it is important to recognize the limited significance of *317 collateral estoppel, particularly in the Microsoft litigation. Both federal [FN84] and state [FN85] courts have rejected broad readings of the preclusive effect of the government case, instead giving collateral estoppel effect only to rulings that were essential to support the liability determinations that the D.C. Circuit affirmed. [FN86] Very few of the findings and conclusions meet this standard. The only state court to address the issue in a Microsoft indirect purchaser suit gave collateral estoppel effect only to the findings on relevant market and monopoly power; the court of appeals' statements of Microsoft's specific acts of monopolization; and the conclusion that Microsoft maintained its monopoly illegally. [FN87] More important, even if all of Judge Jackson's findings were given collateral estoppel effect, they would be mainly of value to Microsoft's rivals that sued for damages from unlawful exclusion; they would be of limited relevance to claims by purchasers for overcharge damages. Judge Jackson held, and the court of appeals affirmed, that Microsoft had monopoly power in the market for Intel-compatible operating systems, as evidenced by its market share and the existence of the applications barrier to entry, [FN88] and that it had illegally maintained that power by its exclusionary actions against Netscape and Java. [FN89] But he did not find, nor did the government ever allege, that Microsoft acquired its monopoly power unlawfully. Indeed, in earlier litigation, the government had argued that Microsoft gained its initial lead in the OS market mainly by luck, and that the positive feedback associated with network effects cause the market to tip in its direction. [FN90] Nor did Judge Jackson hold that the continuation of Microsoft's *318 monopoly power after the introduction of Internet Explorer was solely a function of Microsoft's illegal acts. [FN91] In addition, Judge Jackson held that the evidence did not support a finding that Microsoft was exercising its monopoly power to charge a profit-maximizing monopoly price for Windows. [FN92] Judge Jackson did hold that Microsoft could have charged a supracompetitive price, [FN93] but that it may have chosen to "keep the price of Windows low today" in order to spur the growth of the OS market; [FN94] or it may have used some of its monopoly power by offering lower prices to OEMs who agreed to restrictive terms. [FN95] Moreover, it is a fair reading of Judge Jackson's opinion that Microsoft charged either nothing [FN96] or a negative price [FN97] for the Internet Explorer browser, which it *319 bundled with Windows. The trial court did hold that Microsoft acted consistently with monopoly power in some of its pricing decisions. [FN98] For example, Microsoft raised the price of an outdated version of Windows upon introducing Windows 98; [FN99] it priced an upgrade of Windows 98 at $89 even though an internal study had found that $49 would have been profitable; [FN100] and it charged different prices to different OEMs. [FN101] But these findings only related to whether Microsoft had monopoly power, not to whether the actions themselves were unlawful, or constituted any sort of overcharge. Thus, far from holding that Microsoft imposed an overcharge, United States v. Microsoft actually suggests that prices of the operating system and the browser were lower, at least in the short run, because of the illegal conduct alleged in the government's case. As the government's expert witness testified "Microsoft has used its power to protect its operating system's monopoly from a threat that might not have materialized by this time anyway. And, in doing that, it has given away a lot of things." [FN102] This conclusion does mean that consumers were necessarily better off because of Microsoft's actions during the browser wars. Judge Jackson suggested they may have suffered a loss in system performance in being forced to take Internet Explorer, [FN103] and may have been hurt by reduced innovation in platform software. But these harms, if they occurred, are not evidence of an overcharge. The Microsoft indirect purchaser cases allege a conventional overcharge theory of damages. Recognizing the limited support that the government case lends to this theory, the classes have alleged that Microsoft's illegal conduct began in the 1980s, long before the browser wars that were focus of the government case. Thus, the classes contended that Microsoft did acquire its monopoly power in the OS market illegally. Some of the classes alleged that this illegality extended to the Microsoft Office suite and other products that were not at issue in the government litigation. As we will see, the courts certifying some of the Microsoft classes sometimes failed to recognize this disjuncture between the government case and the indirect purchaser class actions. *320 IV. THE REASONING OF THE MICROSOFT CERTIFICATION DECISIONS In the Microsoft certification decisions, as in others, the most important question was whether class issues predominated over individual issues. Issues of the defendant's liability were common to the class; thus, whether the common issues predominated over individual issues depended upon whether the plaintiffs offered a reasonable basis for proving the fact and amount of damages on a classwide basis. In most of the Microsoft certification cases, plaintiffs relied on expert testimony of Dr Keith Leffler of the University of Washington to show how they planned to prove the effect of Microsoft's monopolistic conduct on the state's consumers on a classwide basis. [FN104] Dr Leffler apparently made substantially the same initial submission in support of certification in all of the cases, outlining methodologies that he could use to prove damages, but without actually implementing any of the methods, or examining data specific to the state. [FN105] Microsoft, in each case, offered objections to the methodologies, pointing to complexities in the relevant state market that would make Dr Leffler's arguments unworkable. Dr Leffler then responded to these objections. In his initial submissions, Dr Leffler proposed to prove, first, Microsoft's overcharge to its direct customers and, second, how that overcharge affected end users. First, assuming that there would be proof that Microsoft engaged in illegal actions in operating system markets prior to the damage period (for example, involving MS-DOS and early versions of Windows) an overcharge would be "embedded" in prices at every level of distribution. [FN106] To prove the overcharge, he identified three possible yardstick methods: the "comparable market yardstick," which would compare Microsoft's price with the prices in an otherwise similar market in which there was no illegal activity; the "competitive market yardstick," which would compare Microsoft's profit margins to those in a comparable competitive market; and the "violation-free period yardstick," which would compare Microsoft's margins with those charged in the same market in a period free from violations. [FN107] To prove how much of this overcharge the intermediate purchasers passed on to end users, he *321 proposed to estimate the relationship between Microsoft's prices to its customers and the prices paid by end users using "basic economic principles," standard statistical methods, and data that "should be readily available." [FN108] Relying on the tax incidence model, he stated that, because distribution of Microsoft's products is competitive, the intermediate purchasers would pass on 100% of the overcharge to end users. Because end users have no good substitutes for Microsoft's product, all would pay higher prices regardless of where they purchased the product. Estimating the effect of an upstream overcharge on end users would be a matter of using regression analysis to control for the various other costs of the computer systems in which Microsoft's operating system is installed. According to Dr Leffler, the overcharge for all consumers, regardless of where they purchased their computers, could be estimated with one model. [FN109] Microsoft challenged these assertions through its own experts, claiming that Dr Leffler's account did not take account of numerous demand, cost, and competitive factors in different locations and at different times that would affect the rate of passing on. [FN110] Microsoft also argued that Dr Leffler's assumption that the distribution level was perfectly competitive ignored the role of brand names, product differentiation, and high fixed costs in computer manufacture. [FN111] As we will see, courts have differed in their view of the sufficiency of Dr Leffler's proposed methods. I begin by examining the three decisions that denied certification, then turn to those that granted certification. In the latter category, I give special attention to the decisions in Gordon v. Microsoft, which certified a Microsoft class even though it was the only indirect purchaser class ever certified in the state. A. Reasons for Denying Certification Courts in only two states have refused to certify Microsoft classes. In both instances, the most important reason for denial was the application of an exacting standard for certification. The courts required proof of damage to each member of the class rather than an average harm to the class as a whole; expressed skepticism that tax incidence and regression analysis could provide a basis for estimating actual damages to each class member on a classwide basis; and therefore insisted that the proponents of certification implement their methodologies using real-world evidence. In A&M Supply, [FN112] the Michigan court of appeals reversed the certification of a class of indirect purchasers of Microsoft operating system software, holding that plaintiff had not "set forth a viable method for proving actual damages on a class-wide basis" as required by *322 the Michigan Antitrust Reform Act (MARA). [FN113] The court recognized the distinction between skeptical and sanguine views of the practicalities of indirect purchaser class actions, [FN114] and, after reviewing Michigan trial court decisions denying certification, interpreted MARA to require the "rigorous analysis" of the "skeptical" view. [FN115] The court took the lesson of the Michigan cases to be that certification would only be appropriate if there were a "minimum number of variables involved" and the plaintiff does "more than promise that [the necessary calculation of damages] will be available in the future." [FN116] The proposed methods for proving the overcharge might have been sufficient under this standard, but the methods for proving passing on were not. The court rejected Dr Leffler's proposed use of tax incidence theory to prove passing on, because he had not taken account of the complexities of the market identified by Microsoft's experts. [FN117] The court characterized as "slogans" his proposal to use "relatively simple statistical estimations" and "basic economic principles." [FN118] Although Dr Leffler had not had access to data yet, his "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," [FN119] and it was the plaintiff's choice to move for class certification at such an early stage. The proposed use of regression analysis to isolate the effect of the overcharge on retail prices was likewise insufficient, because it was a bare assertion. [FN120] The A&S court did not characterize the Microsoft indirect purchaser class actions as involving complexities substantially greater than other indirect purchaser class actions. The court found that "some of the concerns" earlier Michigan courts had about proposed classes were present in the Microsoft class. It noted that the class was enormous and covered six products over several years. Moreover, Microsoft's expert's study had shown that computer distributors were not perfectly competitive, and charged different prices and relied on price points, and so would likely have had different rates of passing on at different times. [FN121] *323 The other state to have denied certification was Maine in Melnick. [FN122] Of the only two earlier Maine indirect purchaser decisions cases, one had certified a class and the other had not. [FN123] The court in Melnick, however, applied a stringent approach typical of the skeptical view. The court required the plaintiffs to offer a method of proving damages to each class member by common proof, not only aggregate damage to the class, [FN124] and refused to recognize a presumption of injury in the indirect purchaser context. [FN125] It rejected Dr Leffler's proposed methods for accomplishing this task, emphasizing that he relied only on general theoretical assertions, despite extensive discovery, while Microsoft's experts had offered specific studies of competitive conditions in the Maine computer market, and the various actors' actual pricing practices. [FN126] Although the court agreed that it should not resolve a battle of experts at the certification stage, it insisted that the expert offer some specific evidence that theoretical claims are borne out by real world facts. [FN127] B. Reasons for Granting Certification 1. The Sanguine View Generally A number of states have granted certification in every case for which courts have issued a written opinion. In some of these states, the Microsoft cases are the only indirect purchaser cases in which certification has been proposed. In some instances, the courts offered no reasoning in support of their decisions, presumably adopting the extreme sanguine position that certification was so obviously warranted that it required no justification. [FN128] The cases that did offer reasoning to support certification did so on similar grounds. Some suggested that the adoption of an indirect purchaser statute created a presumption that *324 consumer classes should be certified [FN129] and that to deny certification would deny consumers a remedy. [FN130] In the cases granting certification, the plaintiffs proposed to prove both the fact and the amount of damages by offering essentially the same submissions by Dr Leffler that the courts in Maine and Michigan rejected. [FN131] In these sanguine cases, the courts accepted the expert's proposed yardstick measures of proving the overcharge, and the statistical methods of proving passing on. [FN132] The fact that the submission described only abstract methodologies did not disqualify it. [FN133] The courts also were willing to accept the prospect of many individualized issues on the amount of damage to members of the case. [FN134] As in Michigan and Maine, Microsoft's experts challenged Dr Leffler's methods, and proposed various ways in which the overcharge might vary among individual members of the class. [FN135] The sanguine courts, however, characterized these differences as a "battle of experts" that could only be resolved by an examination of the merits, which was precluded at the certification stage. [FN136] Many courts applied the peculiarly deferential standard that the expert need only offer a method that was not so "insubstantial as to amount to no method at all." [FN137] Some noted also that if the evidence ultimately did not establish impact by *325 common proof, the class could be decertified at a later stage. [FN138] There also appeared to be a kind of snowball effect--later cases cited earlier ones as precedent, and characterized the contrary decisions in A&M and Melnick as aberrations. [FN139] The Florida Microsoft [FN140] case deserves special mention, because Florida courts had denied certification in Execu-Tech, [FN141] one of cases that figured most prominently in my 1999 discussion of the skeptical view. The court in Execu-Tech had denigrated the plaintiff's proposal to use tax incidence theory to establish passing on as "inherently speculative." [FN142] Thus, one might have thought that the Florida Microsoft case, in granting certification, might have identified the characteristics of the case that warranted a different result. But, while the court cited Execu-Tech, it made no effort to distinguish it, offering instead essentially the sanguine rationale for certification. [FN143] 2. Minnesota The foregoing cases are distinguishable from A&M and Melnick mainly in their far more lenient standards of certification; few of them discussed characteristics of the Microsoft cases that might have made them either more or less appropriate for certification than other indirect purchaser cases. More illuminating in this respect are the court's opinions in Gordon, the Minnesota Microsoft case. This case is unique in that it certified a class despite unanimously skeptical state precedents [FN144] and actually went to trial before finally settling. *326 Before Gordon, all of Minnesota's nine indirect purchaser decisions rejected certification on skeptical grounds. [FN145] In spite of this record of unrelenting skepticism, the court in Gordon granted certification. In its initial grant of certification, [FN146] it noted that, if there were common proof of liability and impact, individual issues and complexity in the proof of actual damages would not necessarily prevent class treatment. [FN147] It recognized that no Minnesota court had yet certified an indirect purchaser class, but found all of the prior cases distinguishable. [FN148] In one, there were many defendants and multitude of different transactions; [FN149] in another, there were many defendants and many different drugs at issue; [FN150] in a third, the products at issue gained value through repackaging at intermediate levels; [FN151] in a fourth, buyers of had no records of the amount of the product they bought or the price they paid; [FN152] and in a fifth, the expert could not show that anyone had ever used his methods. [FN153] In Gordon, in contrast, there was a single seller of only two primary products; there were relatively infrequent transactions for each class member, and there *327 were likely to be records of those transactions; and the expert had offered a plausible method of proving damages. [FN154] The court reviewed Dr Leffler's proposed methods, and Microsoft's objections, [FN155] and found the methods sufficient, declining to resolve the battle of experts at the certification stage. [FN156] More than 2 years later, after discovery had occurred, and on the eve of trial, Microsoft moved to decertify the class. A number of courts, in certifying Microsoft indirect purchaser classes, had mentioned the possibility of decertification after discovery, but Gordon was the only one actually to confront the issue. [FN157] Although the plaintiff, during discovery, had switched experts and theories of classwide injury, the court declined to decertify the class, saying that the battle of experts should now be resolved at a trial. [FN158] Microsoft again challenged the proposed methods of proving an "embedded" overcharge with respect to early purchasers of Microsoft products; the court responded that, even if true, this criticism would only shorten the damage period. The court also rejected Microsoft's claim that, to prove passing on, it would be necessary for all of the intermediate purchasers to testify to how they responded to price changes. Even under Microsoft's view of the evidence, all of the distributors passed on some of the costs, and proof of the amount of damages is subject to a lesser standard of proof. [FN159] *328 C. The Dubious Role of United States v. Microsoft The Microsoft indirect purchaser plaintiffs undoubtedly expected to benefit from the rulings in United States v. Microsoft. [FN160] As we have seen, however, the government case offered little support for them. [FN161] First, traditional principles of collateral estoppel limited the preclusive effect of the government case to its irreducible core of findings and conclusions. [FN162] Second, the government case suggested that Microsoft was pursuing a low-price strategy, while the indirect purchasers alleged it was imposing an overcharge. [FN163] Third, the government case established only that Microsoft illegally maintained its monopoly by actions beginning in 1995, while the indirect purchasers alleged that Microsoft illegally acquired its monopoly by actions against entirely different firms beginning the late 1980s. [FN164] The courts in the Microsoft indirect purchaser cases at times either failed to recognize either the existence of these kinds of differences or their significance. [FN165] The usual sticking point in certification of indirect purchaser class actions is the issue of passing on; but uncertainty in the amount of the overcharge to direct purchasers can weigh against certification. [FN166] The experience of the Minnesota Microsoft court is instructive. In initially granting certification, the court wrote that "Judge Jackson not only described an overcharge to direct purchasers but specified an amount," citing the finding *329 that Microsoft charged $89 for an upgrade of Windows when it could have charged $49. [FN167] There are a number of problems with using this finding in this way. First, the finding did not characterize the difference between $89 and $49 as an overcharge attributable to illegal conduct; it was intended only to show that Microsoft had monopoly power. An overcharge, for antitrust purposes, is the difference between the actual price and the price that would have prevailed absent illegal conduct; it is not the difference between the actual price and the seller's cost. [FN168] Even if the indirect purchaser plaintiffs were able to prove that Microsoft acquired its monopoly illegally by its actions before 1994, [FN169] it would not follow that a supracompetitive price reflected an illegal overcharge. It is not clear that, absent any illegal conduct, the market for operating systems would have produced a competitive price, much less a price of $49 for a comparable product. The theory of network effects, on which the government's theory of liability rested, implies that the market for operating systems would be likely to tip toward a single dominant producer even without efforts by Microsoft to influence the process. Consequently, it is likely that a dominant producer would have emerged in any but for world. [FN170] Second, the finding did not have binding effect as evidence of supracompetitive pricing. The Minnesota court itself later determined that the only essential findings on monopoly power were that that Microsoft could have charged a supracompetitive price for Windows because of its high market share, the applications barrier to entry, and the lack of available substitutes. [FN171] *330 Thus, the findings did not establish that Microsoft charged a supracompetitive price at all; indeed there was evidence that Microsoft was charging low prices for strategic reasons. This instance shows the wisdom of limiting collateral estoppel effect to essential rulings. It seems that Judge Jackson intended the episode of the $89 price to illustrate Microsoft's power to charge a supracompetitive price, not to suggest that it consistently charged supracompetitive prices. A number of the certification decisions noted Judge Jackson's finding that Microsoft's actions "have harmed consumers in ways that are immediate and easily discernible." [FN172] But this holding is also of little relevance to the indirect purchaser plaintiffs' cases. First, it was probably not entitled to collateral estoppel effect, because it was not essential to the finding of liability, which was based on a theory of incipient harm. The court in Gordon gave collateral estoppel effect to the court of appeals' statement that "Microsoft maintained its monopoly power by ... conduct which caused 'harm [to] the competitive process and thereby harm [to] consumers."' [FN173] The court specifically refused to give preclusive effect to any of the specific harms Judge Jackson enumerated. [FN174] In any event, the findings would offer little aid to the indirect purchasers, because none of the harms identified by Judge Jackson was an overcharge. Judge Jackson suggested that consumers were harmed by an unspecified reduction in innovation--a type of harm the indirect purchaser plaintiffs did not allege. [FN175] In the Florida Microsoft case, Microsoft pointed to holdings in the government case that some of Microsoft's conduct, such as providing a free web browser, benefited some consumers. [FN176] Of course, as nonparties to the government case, the plaintiffs were not bound by this finding. Nevertheless, it makes a self-evident point that is apparently relevant to the amount of the overcharge. Moreover, it suggests that the harm to computer users was not uniform: some consumers who did not want a web browser might have been injured by receiving a free browser, while those who did want one might have benefited. Nevertheless, the Florida court accepted Dr Leffler's response that *331 there is no economic basis to offset operating system overcharges because of predatory actions with respect to browsers. The predatory browser action may benefit consumers in the browser market in the short run until Microsoft establishes its monopoly in the browser market but this has no implications as to the injury and the amount of injury suffered by consumers in the separate operating system market. Moreover, the competitive browser price may have been zero, given that Netscape and Microsoft have given away their browsers for several years. If so, there is no consumer benefit to including IE with Windows. Finally, as emphasized in Judge Jackson's findings, the competitive harm from the browser tie-in by Microsoft included computer performance degradation and costs to consumers that were independent of the price charged for the browser. [FN177] This passage fails to refute Microsoft's point. Judge Jackson did not characterize the actions of Microsoft that benefited consumers as predatory. The court of appeals opinion makes clear that even providing the browser integrated with the operating system was not harmful; only the failure to provide means of removing IE was harmful, and was so only to consumers who would have wanted to remove the browser. And even if the browser is a separate product, providing it free is a benefit, comparable to a coupon, that reduces the effective price of Windows at least for consumers that wanted the product. The point that the competitive price for the browser may have been zero fails to recognize that the price of Netscape's browser was driven to zero only by Microsoft's introduction of its browser. Interestingly, none of the courts in the indirect purchaser cases seems to have addressed another complicating factor in the United States v. Microsoft: Judge Jackson's suggestion that Microsoft charged different prices for Windows to different OEMs. [FN178] Those that agreed to restrictive terms paid *332 less than those who did not. Consequently, the size of any overcharge might have depended upon whether the OEM cooperated with Microsoft with respect to particular arrangements. As I noted in my 1999 study, a number of courts have noted the variability of the initial overcharge as a reason for denying certification. [FN179] For whatever reason, this concern never presented an obstacle to certification of the Microsoft class actions. V. CONCLUSION: THE CONSUMER INTEREST The dust has largely settled in the Microsoft indirect purchaser litigation. It now seems that many of the certifications in these cases were based on the more lenient standard I have called the sanguine view of the problems of proving passing on of an overcharge. Certainly, some aspects of the case made it better suited to certification than some earlier cases. These included the existence of a single seller and a relatively small number of intermediate levels of distribution that were generally, though far from perfectly, competitive. The classes were enormous, but the purchases of Microsoft software by each class member were relatively few and probably documented better than most consumer transactions. [FN180] On the other hand, some aspects of the case seemed to count against certification. The OS was incorporated into another product, and accounted for only a small part of the final price of the computer. Moreover, there were significant differences in prices to OEMs that might have affected any passed on overcharge, and the evidence apparently showed significant differences in the OEMs' policies relating to passing on. Nevertheless, only a very few of the decisions closely analyzed these differences. In general, the decisions were based on bare proposals to use statistical methodologies, not on application of those methodologies to the specific markets in the state. Microsoft's objections based on the specifics of the market were generally met with the response that class certification was not the time to resolve a battle of experts. Some of the courts may have overestimated the significance of United States v. Microsoft for the issue of whether an overcharge to class members could be shown by classwide proof. It is difficult to evaluate whether the overwhelming use of the sanguine view in the Microsoft cases has been in the public interest. All but one of the cases have settled before trial stage, and the one that went to trial was settled without *333 reaching the jury. Consequently, we cannot tell if the sanguine courts' deferential approach to certification was warranted in these cases. In the abstract, a lenient standard makes erroneous grants of certification more likely, while a strict standard makes erroneous denials more likely. [FN181] Which is preferable depends on the ex ante estimate of the magnitude of the costs associated with the two types of error. Sanguine courts estimate that the costs of an erroneous denial are greater, first, because a denial ends the action, while a grant is only provisional, and, second, because an erroneous denial thwarts the substantive law's policy, while an erroneous grant only defers to a later time the proper termination of the action. Consequently, these courts suggest that the lenient certification standard is appropriate. Robert Bone and David Evans have objected to this reasoning. They point out, first, that virtually all cases settle soon after certification because of economic pressures on the defendant. [FN182] Where the risk of an erroneous finding of liability is significant, a defendant has strong incentives to settle even an unmeritorious lawsuit. Courts moreover only very rarely decertify a class action. [FN183] Thus, an erroneous grant is likely to be final, and to produce a costly settlement. [FN184] Bone and Evans also dispute the premise that an erroneous denial of certification necessarily thwarts social policies. In some instances, individual suits are possible in place of a class action; in cases in which they are not possible, the class action would not likely have produced compensation, so the primary consideration for social policy is deterrence, which may be achieved by other means, including public enforcement. [FN185] They also argue that a lenient certification standard encourages more frivolous lawsuits to be filed. [FN186] On balance, they suggest that the social costs of the lenient rule outweigh the costs of a more stringent rule. [FN187] With this analysis in mind, can we find anything in the experience in the Microsoft indirect purchaser class actions that would help us decide whether the more stringent skeptical rule of certification is preferable to the more lenient sanguine rule? Many of Bone and Evans' empirical observations are consistent with the Microsoft experience. All of the cases settled without an adjudication on the merits. Only one ever reached the stage of a motion for decertification, and that motion was denied, albeit on different grounds than *334 the original grant. Moreover, like virtually all of the indirect purchaser class action settlements, the Microsoft settlements have provided little compensation to consumers. [FN188] The vouchers were for small amounts, and only a relatively small percentage of consumers ever filed claims to receive the payments. For example, in the Minnesota case, 5 days before the deadline, 87 000 of an estimated 1 million possible claimants had filed claims for vouchers worth $15 for operating system software, $9 for Word, and $23 for Office and Excel. [FN189] The lion's share of the unclaimed amounts generally went to the state or to schools; in some of the settlements, a portion reverted to Microsoft. Thus, even if the three denials of certification in Microsoft indirect purchaser cases were erroneous, they probably did not thwart a social policy in favor of compensation. If, however, there was an overcharge, the actions might be justified on a deterrence theory. As Bone and Evans point out, the primary goal of class actions involving small individual stakes is deterrence. In fact, deterrence is preferable to compensation as an antitrust goal, because, if it is effective, it renders compensation unnecessary. The rationale of Illinois Brick was to concentrate the full right to recover in the direct purchasers, to assure that they had an adequate incentive to sue for the overcharge. But that rationale assumed that, if direct purchasers were given the full overcharge, the direct purchasers would actually sue, and the indirect purchasers would not. Since the rise of indirect purchaser suits in state court, of course, the issue has changed, because both direct and indirect purchasers can sue. This change has largely eliminated the concern about adequate deterrence, but created much greater concerns about overdeterrence. If multiple levels of purchasers could sue for the full overcharge, then defendants could be subject to duplicative recoveries. [FN190] Penalties should not be arbitrarily large; they should reasonably approximate an optimal penalty. It is difficult to say with any confidence, of course, what an optimal penalty is in a particular case. [FN191] But if we take as a benchmark *335 the federal measure of the treble damages for those who have suffered antitrust injury and have standing, it is doubtful that indirect purchaser class actions are typically necessary to achieve deterrence. Consequently, a more stringent standard for certification is likely to have minimal social costs. The complicating factor in the Microsoft litigation is that some of the direct purchasers have not sued, as one of the courts in the indirect purchaser cases noted. [FN192] Thus, one might argue that indirect purchaser suits are necessary to impose an adequate deterrent penalty. But this argument fails to recognize that Microsoft, unlike most overcharge cases, was about exclusionary practices, which imposed harm most directly on competitors rather than purchasers. [FN193] Predictably, all of the excluded rivals have sued and most have already received substantial settlements. [FN194] Consequently, it is doubtful that the indirect purchaser overcharge litigation was necessary to impose an adequate deterrent penalty. APPENDIX Class Certification Decisions Alabama: Pre-1999: Ashley v. Archer-Daniels-Midland Co., No. CV 95-336-R (Ala. Cir. Ct. DeKalb Co. Mar. 13, 1998) (lysine) (denied); Durden v. Abbott Labs., No. CV 93- 663 (Ala. Cir. Ct. Calhoun Co. Jan. 16, 1996) (infant formula) (denied); Harbin v. Johnson & Johnson Vision Prods., No. CV 94-2872 (Ala. Cir. Ct. Mobile Co. Sept. 12, 1995) (contact lenses) (denied); In re Brand Name Prescription Drugs Antitrust Litig. (Price v. American Home Prods.), No. 94 C 897, 1994 WL 663590 (N.D. Ill. Nov. 18, 1994) (prescription drugs) (denied); McCarter v. Abbott Labs., No. CV 91-050 (Ala. Cir. Ct. Shelby Co. Apr. 9, 1993) (infant formula) (denied). Arizona: 1999-2004: Friedman v. Microsoft Corp., CV2000-000722 (Ariz. Super. Ct. Maricopa Co. Nov. 15, 2000) (MS Software) (granted). California: Pre-1999: Lethbridge v. Johnson & Johnson, No. B105754 (Cal. Ct. Ap. Nov. 10, 1997) (contact lenses) (granted); Feitelberg v. Abbott Labs., No. 953865 (Cal. Super. Ct. San Francisco Co. Jun. 23, 1995) (prescription drugs) (granted); Preciado v. Abbott Labs., No. 962294 (Cal. Super. Ct., San Francisco Co., Aug. 16, 1995) (prescription drugs) (granted); B.W.I. Custom Kitchen v. Owens-Illinois, Inc., 235 Cal. Rptr. 228 (Ct. App. 1987) (glass containers) (granted). *336 1999-2004: In re Cipro Cases I & II, 121 Cal. App. 4th 402 (2004) (prescription drugs) (granted); Coordination Proceedings, Microsoft I-V Cases, No. J.C.C.P. 4106 (Cal. Super. Ct. San Francisco Co. Aug. 29, 2000) (MS Software) (granted). District of Columbia: Pre-1999: Goda v. Abbott Labs., No. CIV. A. 01445-96, 1997 WL 156541 (D.C. Super. Ct. Feb. 3, 1997) (prescription drugs) (granted). Florida: Pre-1999: Execu-Tech Business Systems, Inc., v. Appleton Papers, Inc., No. 96-9639 CACE 05 (Fla. Cir. Ct. Broward Co. Dec. 15, 1997) (thermal fax paper) (denied), aff'd, 743 So. 2d 19 (Fla. Dist. Ct. App. 1999). 1999-2004: In re Florida Microsoft Antitrust Litig., No. 99-27340 CA 11 (Fla. Cir. Ct. Miami-Dade Co. Aug. 26, 2002) (MS software) (granted). Iowa: 1999-2004: Comes v. Microsoft Corp., No. CL82311 (Iowa Dist. Ct. Polk Co. Sept. 16, 2003) (MS Software) (granted). Kansas: Pre-1999: Donelan v. Abbott Labs., Inc., No. 94 C 709 (Kan. Dist. Ct. Sedgwick Co. May 3, 1995) (infant formula) (granted). 1999-2004: Bellinder v. Microsoft Corp., No. 00-C-0855, 00-C-00092, 99CV17089, 2001 WL 1397995 (Kan. Dist. Ct. Johnson Co. Sept. 7, 2001) (MS software) (granted); Premier Pork, Inc., v. Phone-Poulenc, S.A., No. 00 C 3 (Kan. Dist. Ct. Scott Co. May 1, 2004) (Methionine) (granted); Smith v. Philip Morris Cos., Inc., No. 00-CV-26 (Kan. Dist. Ct. Seward Co. Nov. 15, 2001) (cigarettes) (granted). Maine: Pre-1999: Karofsky v. Abbott Labs., No. CV-95-1009 (Me. Super. Ct. Cumberland Co. Oct. 15, 1997) (prescription drugs) (denied). 1999-2004: Melnick v. Microsoft Corp., No. CV-99-709, CV-99-752, 2001 WL 1012261 (Me. Super. Ct. Cumberland Co. Aug. 24, 2001) (MS software) (denied); Semba v. Lonza AG, No. CV-99-50 (Me. Super. Ct. Hancock Co. Jul. 28, 1999) (vitamins) (granted). Michigan: Pre-1999: Wilcox v. Archer-Daniels-Midland Co., No. 96-82473-CP (Mich. Cir. Ct. Ingham Co. Sept. 29, 1997) (high-fructose corn syrup and citric acid) (denied); Wood v. Abbott Labs., No. 96-512561-CZ, 1997 WL 824019 (Mich. Cir. Ct. Oakland Co. Sept. 11, 1997) (prescription drugs) (denied); Holmes v. Abbott Labs., No. 94-744 CP (Mich. Cir. Ct. Calhoun Co. May 22, 1995) (infant formula) (denied). See also In re Cardizem CD Antitrust Litig., 200 F.R.D. 326 (E.D. Mich. 2001) (drugs) (granted). 1999-2004: A&M Supply Co. v. Microsoft Corp., 654 N.W.2d 572 (Mich. Ct. App. 2002) (MS Software) (denied); Fish v. Microsoft Corp., No. 00- 031126-NZ (Mich. Cir. Ct. Wayne Co. Apr. 8, 2004) (MS Software) (denied); Ren v. Philip Morris, Inc., No. 00-004035-CZ, 2002 WL 1839983 (Mich. Cir. Ct. Wayne Co. Jun. 11, 2002). *337 Minnesota: Pre-1999: Fischenich v. Abbott Labs., Inc., No. MC-94-6868 (Minn. Dist. Ct. Hennepin Co. May 26, 1995) (infant formula) (denied); Kerr v. Abbott Labs., No. 96-002837, 1997 WL 314419 (Minn. Dist. Ct. Hennepin Co. Feb. 19, 1997) (prescription drugs) (denied); Keating v. Philip Morris, Inc., 417 N.W.2d 132 (Minn. Ct. App. 1987); Humphrey v. Philip Morris Inc., 551 N.W.2d 490 (Minn. 1996) (cigarettes) (denied); City of St. Paul v. FMC Corp., Civ. No. 3-89-0466, 1990 WL 259683 (D. Minn. Nov. 14, 1990) (caustic soda and chlorine) (denied). 1999-2004: Gordon v. Microsoft Corp., No. 00-5994, 2001 WL 366432 (Minn. Dist. Ct. Hennepin Co. Mar. 30, 2001) (MS OS software) (granted); Gordon v. Microsoft Corp., NO. MC 00-5994, 2003 WL 23105552 (Minn. Dist. Ct. Hennepin Co. Mar. 14, 2003) (MS applications software) (granted); [FN195] Gordon v. Microsoft Corp., No. MC00-5994 (Minn. Dist. Ct. Hennepin Co. Dec. 15, 2003) (MS software) (decertification denied); Godden v. Marigold Foods, Inc., No. C8- 97-932 (Minn. Dist Ct. Ramsey Co. Mar. 20, 2000) (milk) (denied); Ludke v. Philip Morris Cos., Inc., No. MC 00-1954, 2001 WL 1673791 (Minn. Dist. Ct. Hennepin Co. Nov. 21, 2001) (cigarettes) (denied); Peridot, Inc., v. Kimberly-Clark Corp., No. MC 98-012686, 2000 WL 673933 (Minn. Dist. Ct. Hennepin Co. Feb. 2, 2000) (tissue) (denied). Mississippi: Pre-1999: Borden, Inc., v. Universal Indus. Corp., 88 F.R.D. 708, 710 (N.D. Miss. 1981) (sugar) (denied--alternate ground). New Mexico: 1999-2004: In re New Mexico Indirect Purchasers Microsoft Antitrust Litig., No. D-0101-CV-2000-1697 (N.M. Dist. Ct. Santa Fe Co. Oct. 2, 2002) (MS Software) (granted). North Carolina: 1999-2004: Picos v. Lonza AG, No. 99 CVS 5035 (N.C. Super. Ct. Cumberland Co. Jul. 20, 1999) (vitamins) (granted). North Dakota: 1999-2004: Howe v. Microsoft Corp., 656 N.W.2d 285 (N.D. 2003) (MS Software) (granted). South Dakota: Pre-1999: In re South Dakota Microsoft Antitrust Litig., 657 N.W.2d 668 (S.D. 2003) (MS Software) (granted). 1999-2004: Hagemann v. Abbott Labs., Inc., No. 94-221 (S.D. Cir. Ct. Hughes Co. Nov. 21, 1995) (infant formula) (granted). Tennessee: Pre-1999: Robinson v. EMI Music Distrib., No. CIV. A. No. L-10462. 1996 WL 495551 (Tenn. Cir. Ct. Blount Co. Jul. 8, 1996) (compact disks) (granted). 1999-2004: Panitz v. Lonza AG, No. 302829-8 TD (Tenn. Cir. Ct. Shelby Co. Jul. 28, 1999) (vitamins) (granted); Sherwood v. Microsoft Corp., No. 99C-3562 (Tenn. Cir. Ct. Davidson Co. Dec. 20, 2002) (MS Software) (granted). *338 Wisconsin: Pre-1999: Derzon v. Appleton Papers, Inc., No. 96-CV-3678, 1998 WL 1031504 (Wis. Cir. Ct. Milwaukee Co. Jul. 7, 1998) (thermal fax paper) (denied); Carlson v. Abbott Labs., Inc., No. 94-CV-002608 (Wis. Cir. Ct. Milwaukee Co. Mar. 23, 1995) (infant formula) (granted); K-S Pharmacies, Inc., v. Abbott Labs., No. 94CV-002384 1996 WL 33323859 (Wis. Cir. Ct. Dane Co. May 17, 1996) (prescription drugs) (granted). 1999-2004: Capp v. Microsoft Corporation, No. 00-CV-0637 (Wis. Cir. Ct. Dade Co. Sep. 12, 2001) (MS Software) (granted); In re Methionine Antitrust Litig., 204 F.R.D. 161 (N.D. Cal. 2001) (Methionine) (denied); In re Methionine Antitrust Litig., MDL No. 00-1311 (N.D. Cal. Sept. 24, 2002) (Methionine) (granted); In re Methionine Antitrust Litig., No. 00-1311, 2003 WL 22048232 (N.D. Cal. Aug. 26, 2003) (Methionine) (decertification granted). [FNa1]. Marshall M. Criser Eminent Scholar, University of Florida Levin College of Law. Email: page@law.ufl.edu. I am grateful to John Lopatka, Ronald Cass, Chuck Casper and the participants in a workshop at the University of Florida for helpful comments on earlier drafts. I also benefited from the observations of Spencer Waller and Samuel Issacharoff, who commented on this paper at a conference on Litigating Conspiracy at the University of Western Ontario School of Law. Finally, I thank Diego Puig for research assistance. [FN1]. See infra, Part III. [FN2]. Ill. Brick Co. v. Illinois, 431 U.S. 720 (1977). [FN3]. See infra, n 13, 14. [FN4]. See generally William H. Page, The Limits of State Indirect Purchaser Suits: Class Certification in the Shadow of Illinois Brick, 67 ANTITRUST L.J. 1 (1999) [hereinafter Limits]. [FN5]. A&M Supply Co. v. Microsoft Corp., 654 N.W.2d 572 (Mich. Ct. App. 2002); Fish v. Microsoft Corp., No. 00-031126-NZ (Mich. Cir. Ct. Wayne Co. Apr. 8, 2004); Melnick v. Microsoft Corp., No. CV-99-709 2001 WL 1012261 (Me. Super. Ct. Cumberland Co. Aug. 24, 2001). [FN6]. See Bellinder v. Microsoft Corp., No. 00-C-0855, 00-C-00092, 99CV17089, 2001 WL 1397995 (Kan. Dist. Ct. Johnson Co. Sept. 7, 2001); Capp v. Microsoft Corporation, No. 00-CV-0637 (Wis. Cir. Ct. Dade Co. Sep. 12, 2001); Comes v. Microsoft Corp., No. CL82311 (Iowa Dist. Ct. Polk Co. Sept. 16, 2003); Coordination Proceedings, Microsoft I-V Cases, No. J.C.C.P. 4106 (Cal. Super. Ct. San Francisco Co. Aug. 29, 2000); In re Florida Microsoft Antitrust Litig., No. 99-27340 CA 11 (Fla. Cir. Ct. Miami-Dade Co. Aug. 26, 2002); Gordon v. Microsoft Corp., No. 00-5994, 2001 WL 366432 (Minn. Dist. Ct. Hennepin Co. Mar. 30, 2001); Gordon v. Microsoft Corp., NO. MC 00-5994, 2003 WL 23105552 (Minn. Dist. Ct. Hennepin Co. Mar. 14, 2003) (applications software); Gordon v. Microsoft Corp., No. MC00-5994 (Minn. Dist. Ct. Hennepin Co. Dec. 15, 2003) (refusing to decertify); Howe v. Microsoft Corp., 656 N.W.2d 285 (N.D. 2003); In re New Mexico Indirect Purchasers Microsoft Antitrust Litig., No. D-0101-CV-2000-1697 (N.M. Dist. Ct. Santa Fe Co. Oct. 2, 2002); Sherwood v. Microsoft Corp., No. 99C-3562 (Cal. Cir. Ct. Davidson Co. Dec. 20, 2002); In re South Dakota Microsoft Antitrust Litig., 657 N.W.2d 668 (S.D. 2003); Friedman v. Microsoft Corp., CV2000-000722 (Ariz. Super. Ct. Maricopa Co. Nov. 15, 2000). Final settlements have been approved in California, Montana, North Dakota, South Dakota, Tennessee, West Virginia, Florida and Kansas. [FN7]. See, e.g., Seacoast Motors, Inc., v. DaimlerChrysler Motors Corp., 271 F.3d 6, 9 (1st Cir. 2001) ("The central aim of the antitrust laws is to protect consumers against certain abusive business practices--especially price-fixing and monopoly."). [FN8]. See John E. Lopatka & William H. Page, Monopolization, Innovation, and Consumer Welfare, 69 GEO. WASH. L. REV. 367, 373-76 (2001). [FN9]. See, e.g., Leider v. Ralfe, NO. 01 CIV. 3137 (HB), 2003 WL 22339305 (S.D.N.Y. Oct. 10, 2003) (denying federal antitrust standing to consumers as indirect purchasers of diamonds). [FN10]. Reiter v. Sonotone Corp., 442 U.S. 330, 340 (1979). [FN11]. Alfred Dunhill, Ltd. v. Interstate Cigar Co., 499 F.2d 232, 237 (2d Cir. 1974) (holding that "the Federal Trade Commission Act may be enforced only by the Federal Trade Commission" and not by competitors or consumers (citations omitted)). [FN12]. Ill. Brick Co. v. Illinois, 431 U.S. 720, 734 (1977). [FN13]. See, e.g., Edward D. Cavanagh, Illinois Brick: A Look Back and a Look Ahead, 17 LOYOLA CONSUMER L. REV. 1, 2, n.4 (2004) (citing "[t]wenty-five states and the District of Columbia [that] have enacted Illinois Brick repealer statutes," six of which require that the suit be brought by the attorney general on behalf of consumers). [FN14]. Id. (citing four such states). [FN15]. California v. ARC Am. Corp., 490 U.S. 93 (1989). [FN16]. See infra, n 78. [FN17]. For a comprehensive study of issues surrounding these cases, see ABA SECTION OF ANTITRUST LAW, INDIRECT PURCHASER HANDBOOK (forthcoming 2005), especially Chapter 7 on class certification issues. [FN18]. See John E. Lopatka & William H. Page, Indirect Purchaser Suits and the Consumer Interest, 48 ANTITRUST BULL. 531, 561, n.114 (2003) (hereinafter Consumer Interest). [FN19]. See Page, Limits, supra n 4. See also Chris S. Coutroulis & D. Matthew Allen, The Pass-on Problem in Indirect Purchaser Class Litigation, 44 ANTITRUST BULL. 179, 184-86 (Spring 1999). [FN20]. For a study finding similar divisions in the federal courts' approaches to certification of antitrust damage class actions over 30 years, see Robert H. Klonoff, Antitrust Class Actions: Chaos in the Courts, in LITIGATING CONSPIRACY (forthcoming 2005). [FN21]. Id., at 18-19. [FN22]. Id., at 17-18. [FN23]. Robert G. Bone & David S. Evans, Class Certification and the Substantive Merits, 51 DUKE L.J. 1251, 1268 (2002) ("Ever since Falcon, the federal courts have struggled to chart a middle course between Falcon's requirement of a rigorous analysis and Eisen's prohibition of a preliminary inquiry into the merits."). See also Geoffrey C. Hazard, Jr, Review of the Merits in Class Action Certification, 33 HOFSTRA. L. REV. 51 (2004); Bartlett H. McGuire, The Death Knell for Eisen: Why the Class Action Analysis Should Include an Assessment of the Merits, 168 F.R.D. 366 (1996). [FN24]. Eisen v. Carlisle & Jaquelyn, 417 U.S. 156, 177 (1974). [FN25]. Gen. Telephone Co. v. Falcon, 457 U.S. 147, 161 (1982). Until 2003, the federal class action rule required that the certification decision be made "as soon as practicable"; that provision was amended in 2003 to read "at an early practicable time," in order to allow adequate discovery for the certification decision. Fed. R. Civ. P. 23(c)(1)(a). [FN26]. Coopers & Lybrand v. Livesay, 437 U.S. 463, 469, n.12 (1978). See also Earnest v. Amoco Oil Co., 859 So.2d 1255, 1258 (Fla. App. 2003) ("In conducting its 'rigorous analysis,' the trial court may look beyond the pleadings and, without resolving disputed issues, determine how disputed issues might be addressed on a classwide basis."). [FN27]. Karofsky v. Abbott Labs., No. CV-95-1009, slip op. at 6 (Me. Super. Ct. Cumberland Co. Oct. 15, 1997). [FN28]. Hazard, supra n 23, at 78 (observing that the certification decision has "potentially devastating consequences" for the losing side). [FN29]. In re South Dakota Microsoft Antitrust Litig., 657 N.W.2d 668, 675 (S.D. 2003). [FN30]. Id. [FN31]. Id. [FN32]. Id. [FN33]. See, e.g., Howe v. Microsoft Corp., 656 N.W.2d 285, 295-96 (N.D. 2003); S.D. Microsoft, 657 N.W.2d at 675. [FN34]. In re Potash Antitrust Litig., 159 F.R.D. 682, 697 (D.Minn.1995). See, e.g., Gordon v. Microsoft Corp., No. 00-5994, 2001 WL 366432, at 5 (Minn. Dist. Ct. Hennepin Co. Mar. 30, 2001); Bellinder v. Microsoft Corp., No. 00-C-0855, 00-C-00092, 99CV17089, 2001 WL 1397995, at 7 (Kan. Dist. Ct. Johnson Co. Sept. 7, 2001). But cf. S.D. Microsoft, 657 N.W.2d at 673 (recognizing that "such a standard is little better than no standard," but nevertheless approving certification). [FN35]. S.D. Microsoft, 657 N.W.2d at 678-79. Cf. Caridad v. Metro North Commuter R.R., 191 F.3d 283, 292 (2d Cir.1999) ("statistical dueling" between experts is not relevant to the certification issue). [FN36]. A&M Supply Co. v. Microsoft Corp., 654 N.W.2d 572, 628 (Mich. Ct. App. 2002). [FN37]. Id. (observing that "there is a distance between theory and reality, ... and the actual damages requirement in [the Michigan antitrust statute] mandates proof of pass-on's reality"). [FN38]. Bone & Evans, supra n 23, at 1268. [FN39]. Id., at 1269. [FN40]. Page, Limits, supra n 4, at 21. [FN41]. Id., at 27-34. [FN42]. Id., at 31-33. [FN43]. Id., at 27-34. [FN44]. Id., at 31. [FN45]. Lopatka & Page, supra n 18. But cf. Stephen Calkins, Perspectives on State and Federal Antitrust Enforcement, 53 DUKE L.J. 673, 691-93 (2003) cites numerous settlements of state antitrust litigation that provided for payments to consumers, but includes some that involved cy pres payments, and does not specify actual claims rates. In general, distributions are only practical if the mechanism relaxes requirements that consumers document their purchases. [FN46]. Lopatka & Page, supra n 18, at 552-56. [FN47]. The class certification decisions I considered for this study are listed in the Appendix. For a Canadian view, see Chadha v. Bayer Inc., (2003) 63 O.R. (3d) 22 (C.A.) (holding that courts may not simply presume passing-on of a portion of the of the overcharge in certification of an indirect purchaser class). [FN48]. Ferrell v. Wyeth-Ayerst Labs., Inc., No. C-1-01-447 (S.D. Ohio 2004); In re Relafen Antitrust Litig., 221 F.R.D. 260 (D. Mass. 2004); In re Terazosin Hydrochloride Antitrust Litig., 220 F.R.D. 672 (S.D. Fla. 2004). [FN49]. John Lopatka and I made a preliminary examination of this issue in Lopatka & Page, supra n 18, at 549-51. [FN50]. This category includes both older and more recent cases in the drug industry alleging very different antitrust violations. [FN51]. For one of these states, Mississippi, the only case was Borden, Inc., v. Universal Indus. Corp., 88 F.R.D. 708 (N.D. Miss. 1981), a federal case that the court also dismissed on jurisdictional grounds. The extent of the court's certification analysis was to note that "[i]ndividual questions of impact and nature and extent of damages would be expected to predominate over proof regarding the existence vel non of a conspiracy." Id., at 710. [FN52]. The court actually issued two initial certification orders, one for indirect purchasers of operating systems, Gordon v. Microsoft Corp., No. 00- 5994, 2001 WL 366432 (Minn. Dist. Ct. Hennepin Co. Mar. 30, 2001), and one for indirect purchasers of applications. Gordon v. Microsoft Corp., NO. MC 00- 5994, 2003 WL 23105552 (Minn. Dist. Ct. Hennepin Co. Mar. 14, 2003). I have not treated the latter as a separate decision, even though it occurred 2 years later. Three other states certified applications classes in the same opinion as the OS classes; in each of these cases I treated the decision as a single certification. None of the states treated the issues for applications as materially different from the issues for operating systems. [FN53]. Ferrell v. Wyeth-Ayerst Labs., Inc., No. C-1-01-447 (S.D. Ohio 2004); In re Relafen Antitrust Litig., 221 F.R.D. 260 (D. Mass. 2004); In re Terazosin Hydrochloride Antitrust Litig., 220 F.R.D. 672 (S.D. Fla. 2004). See also Collusion and Other Anticompetitive Practices: A Survey of Class Action Lawsuits Against Drug Manufacturers, available at www.familiesusa.org/site/DocServer/Second_edition_lawsuit_surveys.pdf? docID=349, (listing 10 antitrust litigations, most including indirect purchaser classes). In J.P. Morgan & Co., Inc., v. Superior Ct., 113 Cal. App. 4th 195, 216 (2003), a California state court refused to certify a multistate class action. [FN54]. In Relafen, for example, the court retained supplemental jurisdiction, 28 U.S.C. § 1367(a) (2000), over state law claims, even after it denied certification of federal claims for injunctive relief. See Relafen, 221 F.R.D. at 265. [FN55]. 380 U.S. 460 (1965). [FN56]. Relafen, 221 F.R.D. at 282. [FN57]. Id. [FN58]. Cf. ibid, at 284-86 (holding that a New York statute barring class actions for statutory penalties unless specifically authorized did not direct directly conflict with Federal Rule 23, and was outcome determinative). [FN59]. Pub.L. 109-2, § 9, 119 Stat. 14 (2005). [FN60]. See infra, n 157, discussing the Methionine case, which was filed in federal court in California, but asserted a right of action under Wisconsin law. [FN61]. United States v. Microsoft Corp., 84 F. Supp. 2d 9 (D.D.C. 1999). Microsoft's monopoly of operating system software was protected by an "applications barrier to entry," the preference of software developers to write programs for the most popular platforms and the corresponding preference of computer users to use the platform for which the most applications are available. The emergence of "middleware," like the browser and Java, to which applications could be written, threatened to undermine this barrier, and thus provoked Microsoft's competitive and anticompetitive responses. [FN62]. United States v. Microsoft Corp., 87 F. Supp. 2d 30 (D.D.C. 2000). [FN63]. United States v. Microsoft Corp., 97 F. Supp. 2d 59 (D.D.C. 2000). [FN64]. United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001). [FN65]. United States v. Microsoft Corp., 231 F. Supp. 2d 144 (D.D.C. 2002), aff'd, Massachusetts v. Microsoft Corp., 373 F.3d 1199 (D.C. Cir. 2004). [FN66]. See generally Thomas E. Kauper & Edward A. Snyder, An Inquiry into the Efficiency of Private Antitrust Enforcement: Follow-on and Independently Initiated Cases Compared, 74 GEO. L.J. 1163 (1986). [FN67]. See, e.g., In re Microsoft Corp. Antitrust Litig., 333 F.3d 517, 530-31 (4th Cir. 2003) (affirming an injunction against distribution of Microsoft's version of the Java Virtual Machine, but reversing an order that Microsoft incorporate Sun's JVM in Windows). [FN68]. Caron Carlson, Microsoft Moves Closer to Wiping Its Antitrust Slate Clean, EWEEK, Nov. 8, 2004 (describing settlements with Novell, Inc., and the Computer & Communications Industry Association), http:// www.eweek.com/article2/0,1759,1721057,00.asp.; Microsoft to pay AOL $750M; Tech titans settle Netscape lawsuit, set seven-year licensing pact for AOL to use Internet Explorer, http://money.cnn.com/2003/05/29/technology/microsoft/; Microsoft and Sun Microsystems Enter Broad Cooperation Agreement; Settle Outstanding Litigation, http://www.microsoft.com/presspass/press/2004/apr04/04- 02SunAgreementPR.asp. [FN69]. In re Microsoft Corp. Antitrust Litig. 218 F.R.D. 449, 451 (D. Md. 2003) (denying certification of a class that included Large Account Resellers, and noting that "[n]o OEM or other member of the putative class presently in business has instituted an antitrust overcharge suit against Microsoft"). But see "Microsoft and Gateway Lay Foundation for Future Cooperation, Resolve Antitrust Claims," Microsoft Press Release, Apr. 11, 2005 (disclosing that Microsoft would pay Gateway $150 million over 4 years to settle a suit arising out of United States v. Microsoft, in which "Gateway was specifically identified in U.S. District Judge Thomas Penfield Jackson's findings of fact as having been impacted in its business by practices on which he ruled against Microsoft"), available at http:// www.microsoft.com/presspass/press/2005/apr05/04-11GatewayPR.asp. [FN70]. In re Microsoft Corp. Antitrust Litig., 127 F. Supp. 2d 702, 704 (D. Md. 2001). [FN71]. Id. See also In re Microsoft Corp. Antitrust Litig., 241 F. Supp. 2d 563 (D. Md. 2003) (dismissing indirect purchaser class actions based on the law of Connecticut, Kentucky, Maryland, and Oklahoma because state law did not provide remedies for indirect purchasers). [FN72]. In re Microsoft Corp. Antitrust Litig., 214 F.R.D. 371 (D. Md. 2003) (denying certification of classes of purchasers of Microsoft's applications software, "Select and/or Enterprise" customers who purchased software from "Large Account Resellers," and purchasers seeking injunctive relief). [FN73]. In re Microsoft Corp. Antitrust Litig., 214 F.R.D. 371, 377 (D. Md. 2003) (approving a class of "persons who acquired licenses for operating system software through the shop.microsoft.com program," but excluding as atypical large "enterprise" customers who bought software directly from Microsoft after Oct. 19, 2001); In re Microsoft Corp. Antitrust Litigation-Consumer Track, NO. MDL 1332, 2003 WL 21781969 (D. Md. Jul. 28, 2003) (expanding the class to "to include persons who purchased Microsoft operating system software as 'Full Packaged Product' in direct marketing campaigns during the class period"). [FN74]. Microsoft, Multi District Litigation Plaintiffs Announce Settlement Agreement in Federal Class Action Antitrust Case, http:// www.microsoft.com/presspass/press/2003/sep03/09-30MDLpr.asp [FN75]. Dickson v. Microsoft Corp., 309 F.3d 193, 199-213 (4th Cir. 2002). [FN76]. Id., at 213-15. [FN77]. See, e.g., Olstad v. Microsoft Corp., NO. 03-1086, 2004 WL 291267 (Wis. App. Feb. 17, 2004) (certifying to the state supreme court, in an appeal from summary judgment for Microsoft, the question of whether Wisconsin law applies to interstate commerce affecting Wisconsin commerce). [FN78]. See Johnson v. Microsoft Corp., 802 N.E. 2d 712 (Ohio App. 2003); Pomerantz v. Microsoft Corp. 50 P.3d 929 (Colo. Ct. App. 2002); Vacco v. Microsoft Corp. 793 A.2d 1048 (Conn. 2002); Hindman v. Microsoft Corp., No. 00-1-0945 (Haw. Dist. Ct. Sept. 21, 2000); Berghausen v. Microsoft Corp., 765 N.E.2d 592 (Ind. Ct. App. 2002), transfer denied 783 N.E.2d 692 (2002); Arnold v. Microsoft, No. 2000-CA-002144-MR, 2001 WL 1835377 (Ky. Ct. App. Nov. 21, 2001); Davidson v. Microsoft Corp., 143 Md. App. 43, 792 A.2d 336 (Md. Ct. Special App. 2002); O'Connell v. Microsoft Corp., No. CA 00-01743, 2001 WL 893525 (Mass. Super. Ct. Jun. 14, 2001); Ireland v. Microsoft Corp., No. 00CV-201515 (Mo. Cir. Ct. Jan. 24, 2001); Krotz v. Microsoft Corp., No. A416361 (Nev. Dist. Ct. Jun. 23, 2000); Minuteman, LLC v. Microsoft Corp., 795 A.2d 833 (N.H. 2002); Major v. Microsoft Corp., 60 P.3d 511 (Okla. Ct. App 2002); Daraee v. Microsoft Corp., No. 0004-03311 (Or. Cir. Ct. Jun. 27, 2000); Siena v. Microsoft Corp., 796 A.2d 461 (R.I. 2002); Weinberg v. Microsoft Corp., No. D-162,526 (Tex. Dist. Ct. Aug. 30, 2000). [FN79]. Sherwood v. Microsoft Corp., NO. M2000-01850-COA-R9CV, 2003 WL 21780975 (Tenn. Ct. App. Jul. 31, 2003); Comes v. Microsoft Corp., 646 N.W. 2d 440 (Iowa 2002); In re Microsoft Antitrust Litig., CIV.A. CV-99-709, 2001 WL 1711517 (Me. Super. Mar. 26, 2001); Arthur v. Microsoft Corp., 676 N.W.2d 29 (2004) (holding indirect purchasers have standing under a consumer protection statute that predated the state's Illinois Brick repeal statute). [FN80]. See supra, n 4, 5. In at least some of these states, direct purchasers are subject to a passing on defense. See, e.g., Farmers Cooperative Elevator Co. v. Akzo Nobel, Inc., No. ACV35453 (Iowa Dist. Ct. Carroll Co. Mar. 12, 2004); In re Vitamins Antitrust Litig., 259 F. Supp. 2d 1, 3-4, 8 (D.D.C. 2003) (recognizing passing on as an affirmative defense under the Michigan antitrust statute, and noting that most repealer states allowed such a defense). [FN81]. See Laurie J. Flynn, Few Takers for Payments from Microsoft Settlements, N.Y. TIMES, Jan. 5, 2005, at C4. Cf. the California case's settlement web site, http://www.microsoftcalsettlement.com/. [FN82]. Id., at 1165. [FN83]. Herbert Hovenkamp, State Antitrust in the Federal Scheme, 58 IND. L.J. 375, 429 (1983). Plaintiffs may also apparently invoke 15 U.S.C. § 16(a) (2000) to use the federal judgment as prima facie evidence against the defendant in an action under state antitrust law. See 11 JOSEPH P. BAUER, KINTNER FEDERAL ANTITRUST LAW 207 (1998); Kauper & Snyder, supra n 66, at 1169. [FN84]. In re Microsoft Corp. Antitrust Litig., 355 F.3d 322, 326-28 (4th Cir. 2004) (reversing the Judge Motz's decision that Netscape, Sun, Burst.com, Be Inc., and a consumer class may rely on any of the findings that only "supported" the decision in the government case). [FN85]. Gordon v. Microsoft Corp., No. MC 005994, 2003 WL 22281574, at 6 (Minn. Dist. Ct. Hennepin Co. Aug. 20, 2003) (rejecting Judge Motz's interpretation of the scope of offensive collateral estoppel under Minnesota law). [FN86]. In re Microsoft, 355 F.3d at 327 (limiting offensive collateral estoppel to the facts that were "critical and necessary" in the limited sense of "essential" to the liability holdings upheld by the D.C. Circuit); Gordon, 2003 WL 22281574, at 9 (holding that indirect purchaser plaintiffs were only entitled to collateral estoppel on "central facts that were necessary and essential" to the finding of monopolization). [FN87]. Gordon v. Microsoft Corp., No. MC 005994, 2003 WL 22281574, at 10-11 (Minn. Dist. Ct. Hennepin Co. Aug. 20, 2003). [FN88]. United States v. Microsoft Corp., 87 F. Supp. 2d 30, 36-37 (D.D.C. 2000), aff'd in part, 253 F. 3d 34, 51-58 (D.C. Cir. 2001. [FN89]. United States v. Microsoft Corp., 87 F. Supp. 2d 30, 37-46 (D.D.C. 2000), aff'd in part, 253 F. 3d 34, 57-78 (D.C. Cir. 2001). [FN90]. In upholding the decree, the court in United States v. Microsoft Corp., 56 F.3d 1448, 1452 (D.C. Cir. 1995), wrote: The government believes that Microsoft's initial acquisition of monopoly power in the operating systems market was the somewhat fortuitous result of IBM choosing for its PCs the operating system introduced by Microsoft ("MS-DOS"), which, with Microsoft's successful exploitation of that advantage, led Microsoft to obtain an installed base on millions of IBM, and IBM-compatible, PCs. See also John E. Lopatka & William H. Page, Microsoft, Monopolization, and Network Externalities: Some Uses and Abuses of Economic Theory in Antitrust Decision Making, 40 ANTITRUST BULL. 317, 354 (1995). [FN91]. Microsoft, 84 F. Supp.2d at 13 (findings of fact ¶ 8) (finding that "Windows 95 enjoyed unprecedented popularity with consumers"). [FN92]. Id., at 27 (findings of fact ¶ 65). [FN93]. Id., at 19 (findings of fact ¶ 33). See also ibid, at 5, 15- 16 (findings of fact ¶¶ 18, 54, 55). [FN94]. Id., at 27 (findings of fact ¶ 65). It added that "[b]y pricing low relative to the short-run profit-maximizing price, thereby focusing on attracting new users to the Windows platform, Microsoft would also intensify the positive network effects that add to the impenetrability of the applications barrier to entry." Id. [FN95]. Id., at 27 (findings of fact ¶ 66). [FN96]. Id., at 44 (findings of fact ¶ 136) ("Microsoft sought to increase the product's share of browser usage by giving it away for free. In many cases, Microsoft also gave other firms things of value (at substantial cost to Microsoft) in exchange for their commitment to distribute and promote Internet Explorer, sometimes explicitly at Navigator's expense."). See also ibid (findings of fact ¶ 137) ("Microsoft decided not to charge an increment in price when it included Internet Explorer in Windows for the first time, and it has continued this policy ever since. In addition, Microsoft has never charged for an Internet Explorer license when it is distributed separately from Windows."); ibid, at 71 (findings of fact ¶ 250) ("Microsoft licensed the [IE Access Kit] including Internet Explorer, to IAPs at no charge."). But cf. United States v. Microsoft Corp., 253 F.3d 34, 96 (D.C. Cir. 2001) (noting "the tension between Findings of Fact ¶¶ 136-37, which Microsoft interprets as saying that no part of the bundled price of Windows can be attributed to IE, and Conclusions of Law, at 50, which says the opposite"). [FN97]. Microsoft, 84 F. Supp.2d at 45 (findings of fact ¶ 139) ("Not only was Microsoft willing to forego an opportunity to attract substantial revenue while enhancing (albeit temporarily) consumer demand for Windows 98, but the company also paid huge sums of money, and sacrificed many millions more in lost revenue every year, in order to induce firms to take actions that would help increase Internet Explorer's share of browser usage at Navigator's expense."). But cf. Microsoft, 87 F. Supp. 2d at 50 ("The fact that Microsoft ostensibly priced Internet Explorer at zero does not detract from the conclusion that consumers were forced to pay, one way or another, for the browser along with Windows.") Properly read, the Judge Jackson here is not saying there was any increment in price; his point is that tying is illegal regardless of whether there is such an increase. See ibid (arguing that the purpose of prohibiting tying "is not, as Microsoft suggests, simply to punish firms on the basis of an increment in price attributable to the tied product"). In any event, the court of appeals reversed the holding on tying, United States v. Microsoft Corp., 253 F.3d 34, 84-93 (D.C. Cir. 2001) and the government did not pursue that claim on remand. New York v. Microsoft Corp., 297 F.Supp.2d 15, 33 (D.D.C. 2003). [FN98]. Microsoft, 87 F. Supp. 2d at 37. [FN99]. Microsoft, 84 F. Supp.2d at 27 (findings of fact ¶ 62). [FN100]. Id. (findings of fact ¶ 63). [FN101]. Id. (findings of fact ¶ 64). [FN102]. Transcript of Testimony of Franklin Fisher, Jan. 12, 1999 (A.M. session) at 29-30 (emphasis added). [FN103]. Microsoft, 84 F. Supp.2d at 111 (findings of fact ¶ 410). [FN104]. In at least two cases, the plaintiffs relied on Jeffery K. MacKie-Mason. See Gordon v. Microsoft Corp., NO. MC 00-5994, 2003 WL 23105550 (Minn. Dist. Ct. Dec. 15, 2003); Coordination Proceedings Special Title (Rule 1550(b)) Microsoft I-V Cases, NO.JCCP.NO. 4106 (Cal.Super.Ct., San Francisco Co., Aug. 29, 2000). In one, the plaintiffs relied on Dr James Levinsohn. Fish v. Microsoft Corp., Case No. 00-031126-NZ (Mich. Cir. Ct. Wayne Co. Apr. 8, 2004). [FN105]. In re Florida Microsoft Antitrust Litig., No. 99-27340 CA 11, 2002 WL 31423620, at 7 (Fla. Cir. Ct. Miami-Dade Co. Aug. 26, 2002) (noting that Dr Leffler had offered affidavits in Kansas, Maine, Minnesota, North Dakota, Michigan, South Dakota, Tennessee, New Mexico, and Wisconsin). [FN106]. Id., at 7. [FN107]. See., e.g., A&M Supply Co. v. Microsoft Corp., 654 N.W. 2d 572, 584-85 (Mich. Ct. App. 2002); Melnick, at 13; Bellinder v. Microsoft Corp., No. 00-C-0855, 00-C-00092, 99CV17089, 2001 WL 1397995, at 6 (Kan. Dist. Ct. Johnson Co. Sept. 7, 2001). [FN108]. See, e.g., A&M Supply, 654 N.W. 2d at 585-86. [FN109]. Id., at 586. [FN110]. Id., at 588. [FN111]. Id., at 590. [FN112]. Id. [FN113]. Id., at 603. [FN114]. Id., at 579-81. [FN115]. Id., at 600. [FN116]. Id., at 598-99. [FN117]. Id., at 602. One expert found that consumers paid different prices for identical Compaq systems, and the same price for very different systems. Retailers apparently did not always raise prices in response to cost increases if they were at "price points" that are effective for marketing. Id., at 587. [FN118]. Id. [FN119]. Id. [FN120]. Id. [FN121]. Id., at 603. A more recent Michigan trial court decision "reluctantly" followed A&S in refusing to certify a class of indirect purchasers of Windows 98. Fish v. Microsoft Corp., Case No. 00-031126-NZ (Mich. Cir. Ct. Wayne Co. Apr. 8, 2004). The court read A&S to require plaintiffs to offer a method of establishing on a common bases a formula for proving damages to each class member. Id., slip op. at 13. The plaintiff's expert in this case, James Levinsohn, offered actual regressions using data from Michigan retailers, but the court rejected methodology on the ground that the did not adequately capture market realities, and would have proven only average harm to the class. For another recent Michigan case emphasizing the same point, see Ren v. Philip Morris, Inc., No. 00-004035-CZ, 2002 WL 1839983, at 15 (Mich. Cir. Ct. Wayne Co. Jun. 11, 2002). [FN122]. Melnick v. Microsoft Corp., No. CV-99-709, CV-99-752, 2001 WL 1012261 (Me. Super. Ct. Cumberland Co. Aug. 24, 2001). [FN123]. Semba v. Lonza AG, No. CV-99-50 (Me. Super. Ct. Hancock Co. Jul. 28, 1999) (granting certification in a vitamins class action); Karofsky v. Abbott Labs., No. CV-95-1009 (Me. Super. Ct. Cumberland Co. Oct. 15, 1997) (denying certification in a prescription drugs class action). [FN124]. Melnick, at 6, [FN125]. Id., at 7. [FN126]. Id., at 14-15. [FN127]. Id., at 16. [FN128]. Friedman v. Microsoft Corp., No. CV2000-000722 (Ariz. Maricopa Co. Ct. Nov. 15, 2000); Capp v. Microsoft Corporation, No. 00-CV-0637 (Wis. Cir. Ct. Dade Co. Sep. 12, 2001). [FN129]. See, e.g., Coordination Proceedings, Microsoft I-V Cases, No. J.C.C.P. 4106, slip op at 20 (Cal. Super. Ct. San Francisco Co. Aug. 29, 2000) (following California principle that procedural rules should not be interpreted so strictly as to prevent certification of indirect purchaser classes). [FN130]. In re South Dakota Antitrust Litigation, 657 N. W. 2d 668, 673 (S.D. 2003) (noting that "denial of class certification may be the death knell of claims due to the de minimum amounts of each individual claim which may economically preclude attempted redress by plaintiffs in separate law suits"). [FN131]. Exceptions are Comes v. Microsoft Corp., No. CL82311 (Iowa Dist. Ct. Polk Co. Sept. 16, 2003); Coordination Proceedings Special Title, (Rule 1550(b)) Microsoft I-V Cases, NO. J.C.C.P. No. 4106 (Cal. Super. Ct. Aug. 29, 2000), which relied on testimony of Dr McKie-Mason. [FN132]. Bellinder v. Microsoft Corp., No. 00-C-0855, 2001 WL 1397995 at 6 (Kan. Dist. Ct. Johnson Co. Sept. 7, 2001). [FN133]. Howe, 656 N.W. 2d at 298. [FN134]. Howe v. Microsoft Corp., 656 N.W. 2d 285, 288-89 (N.D. 2003); In re New Mexico Indirect Purchasers Microsoft Corporation Antitrust Litig., No. D-0101-CV-2000-1697, slip. op. at 2 (N.M. Dist. Ct. Santa Fe Co. Oct. 2, 2002; Comes, slip. op. at 18 (requiring only proposed methods for a reasonable estimate of the amount of damages); Coordination Proceedings, Microsoft I-V Cases, No. J.C.C.P. 4106, slip op. at 17-19 (Cal. Super. Ct. San Francisco Co. Aug. 29, 2000). [FN135]. Bellinder, 2001 WL 1397995 at 7. [FN136]. N.M. Microsoft, at 4; Howe, 656 N.W. 2d. at 291; S.D. Microsoft, 657 N. W. 2d at 675-77; Bellinder, 2001 WL 1397995 at 7; Coordination Proceedings, Microsoft I-V Cases, No. J.C.C.P. 4106, slip op. at 15 (Cal. Super. Ct. San Francisco Co. Aug. 29, 2000). But cf. Blades v. Monsanto Co., 400 F.3d 562, 575 (8th Cir. 2005) ("[I]n ruling on class certification, a court may be required to resolve disputes concerning the factual setting of the case [including] expert disputes concerning the import of evidence concerning the factual setting--such as economic evidence as to business operations or market transactions."). [FN137]. Comes, slip op. at 17; Bellinder, 2001 WL 1397995 at 7. The phrase is drawn from federal law. In re Potash Antitrust Litig., 159 F.R.D. 682, 687 (D.Minn.1995). [FN138]. N.M. Microsoft, at 4. [FN139]. Comes v. Microsoft Corp., No. CL82311, slip. op. at 26 (Iowa Dist. Ct. Polk Co. Sept. 16, 2003); S.D. Microsoft, 657 N.W. 2d. at 678-80. [FN140]. In re Florida Microsoft Antitrust Litig., No. 99-27340 CA 11, 2002 WL 31423620 (Fla. Cir. Ct. Miami-Dade Co. Aug. 26, 2002). [FN141]. Execu-Tech Bus. Sys., Inc., v. Appleton Papers, Inc., No. 96- 9639 CACE 05 (Fla. Cir. Ct. Broward Co. Dec. 15, 1997) (thermal fax paper) (denied), aff'd, 743 So. 2d 19 (Fla. Dist. Ct. App. 1999). [FN142]. Execu-Tech, slip op. at 4. [FN143]. For example, the court observed that most repealer states had granted certification, ibid, at 14, 16, and failing to do so would deny all relief. Id., at 18. Dr Leffler's methodologies provided a reasonable estimate of damages, for certification purposes, of damages based on accepted economic principles, even if they could not calculate the harm to each class member. Id., at 9-10. Whether the methods were sufficient to support damages was an issue for trial, ibid, at 15, and the court should not attempt to resolve a battle of experts. Id., at 15-17. Indeed, the fact that Dr Leffler had data available after substantial discovery did not impose any requirement that he actually use his methodologies. Id., at 17. The court rejected Microsoft's argument that the market was too complex to allow common proof, citing federal courts' certification of direct purchaser cases and sanguine states' certification of Microsoft classes, ibid, at 14-15, but failing to cite Execu-Tech's refusal to certify a Florida indirect purchaser case on this ground. [FN144]. Godden v. Marigold Foods, Inc., No. C8-97-932 (Minn. Dist Ct. Ramsey Co. Mar. 20, 2000); Ludke v. Philip Morris Cos., Inc., No. MC 00- 1954, 2001 WL 1673791 (Minn. Dist. Ct. Hennepin Co. Nov. 21, 2001); Peridot, Inc., v. Kimberly-Clark Corp., No. MC 98-012686, 2000 WL 673933 (Minn. Dist. Ct. Hennepin Co. Feb. 2, 2000); Fischenich v. Abbott Labs., Inc., No. MC-94- 6868 (Minn. Dist. Ct. Hennepin Co. May 26, 1995; Kerr v. Abbott Labs., No. 96-002837, 1997 WL 314419 (Minn. Dist. Ct. Hennepin Co. Feb. 19, 1997); Keating v. Philip Morris, Inc., 417 N.W.2d 132 (Minn. Ct. App. 1987); Humphrey v. Philip Morris Inc., 551 N.W.2d 490 (Minn. 1996); City of St. Paul v. FMC Corp., Civ. No. 3-89-0466, 1990 WL 259683 (D. Minn. Nov. 14, 1990). [FN145]. Three of the more recent Minnesota cases are illustrative. In Peridot, Inc., v. Kimberly-Clark Corp., No. MC 98-012686, 2000 WL 673933, at 2 (Minn. Dist. Ct. Hennepin Co. Feb. 2, 2000); Godden v. Marigold Foods, Inc., No. C8-97-932 (Minn. Dist. Ct. Ramsey Co. Mar. 20, 2000); Ludke v. Philip Morris., Inc., No. MC 00-1954, 2001 WL 1673791 (Minn. Dist. Ct. Hennepin Co. Nov. 21, 2001). Peridot denied certification of a class of businesses alleging a price fixing conspiracy among tissue manufacturers, pointing out that, up to that time, no Minnesota court had certified a class of indirect purchasers. The court was especially concerned that the expert offered no means of controlling for nonconspiratorial factors in his damage model, and that some businesses purchased tissue from suppliers other than the defendants. Peridot, 2000 WL 673933, at 5. Godden rejected a proposed class of consumers who bought milk at retail stores, emphasign the complexity of the state's milk markets and distribution system, and the absence of consumer records of purchases. Godden, slip op. at 2-4. Ludke disagreed with Smith v. Philip Morris Cos., Inc., No. 00-CV-26 (Kan. Dist. Ct. Seward Co. Nov. 15, 2001), and refused to certify a class of retail purchasers of cigarettes, concluding that, given the frequency of purchases of cigarettes, and the absence of records of the purchases, it would be impossible to determine damages by common proof or to administer a claims process. [FN146]. Gordon v. Microsoft Corp., No. 00-5994, 2001 WL 366432 (Minn. Dist. Ct. Mar. 30, 2001). The Minnesota Supreme Court later affirmed the intermediate appellate court's refusal to review the certification decision on procedural grounds. Gordon v. Microsoft Corp., 645 N.W. 2d 393 (Minn. 2002). [FN147]. Id., at 3. [FN148]. Id., at 4-6. [FN149]. Keating v. Philip Morris, Inc., 417 N.W.2d 132 (Minn. Ct. App. 1987). [FN150]. Kerr v. Abbott Labs., No. 96-002837, 1997 WL 314419 (Minn. Dist. Ct. Hennepin Co. Feb. 19, 1997). [FN151]. City of St. Paul v. FMC Corp., Civ. No. 3-89-0466, 1990 WL 259683 (D. Minn. Nov. 14, 1990). [FN152]. Fischencich v. Abbott Labs, No. MC-94-6868 (Minn. Dist. Ct. Hennepin Co. May 26, 1995). [FN153]. Peridot v. Kimberly-Clark Corp., No. MC 98-012686, 2000 WL 673933 (Minn. Dist. Ct. Hennepin Co. Feb. 2, 2000). [FN154]. Id. Gordon I, 2001 WL 366432 at 6. [FN155]. Id., at 8-9. [FN156]. Id., at 11.