In Re Microsoft Corp. Antitrust Litigation

185 F. Supp. 2d 519, 2002 U.S. Dist. LEXIS 728, 2002 WL 99709

• Court: District Court, D. Maryland
• Decided: January 11, 2002
• Docket: MDL 1332
• Status: Published

Opinion

OPINION

MOTZ, District Judge.

The MDL plaintiffs and Microsoft have filed a motion for preliminary approval of a proposed class settlement agreement into which they have entered. I have concluded, as a procedural matter, that I cannot presently determine the adequacy of the proposed settlement because the record has not been sufficiently developed on the question of the underlying value of the class claims. I have also concluded, as a substantive matter, that the record as it now exists demonstrates that the charitable foundation contemplated by the agreement is not sufficiently funded both to fulfill the eleemosynary purposes justifying a cy pres remedy and to assure that effectuation of the agreement would not have anti-competitive effects. Therefore, I will deny the motion for preliminary approval.

I.

A.

Under the proposed settlement, all state and federal claims of the members of a nationwide settlement class (defined generally as persons and entities who have acquired licenses for Microsoft operating system or applications software in the United States since January 1,1985) would be released. In return, Microsoft would make certain contributions to a national “eLearning Foundation” (“the Foundation”) that would be established for the purpose of providing computer technology through a grants program to the nation’s most economically impoverished schools. Those schools are defined in the agreement as public schools, kindergarten through high school, in the United States and its territories at which at least 70 percent of the attending students are eligible to receive benefits under the National School Lunch Program. Class members would not personally receive any payments under the settlement. They would have a right to opt out, and, if there were a certain number of opt-outs (defined in a confidential provision of the agreement), Microsoft would have the right to withdraw from the settlement.

The provisions of the agreement relating to the Foundation may be briefly described as follows:

1.The Foundation would be governed by an independent board of directors selected by the court. It would be intended to have a perpetual existence, but during the five-year “settlement period” its primary responsibility would be to administer the grants program established by the settlement agreement.

2. Microsoft would be committed to contribute $400 million to the Foundation to be distributed to eligible schools over five years. Of this sum, $150 million would be for technology acquisitions, $160 million would be for technical support, and $90 million would be for professional development. In addition, Microsoft would contribute up to an additional $100 million to match (on a $1 to $2 basis) funds donated to the Foundation by other sources. The costs of running the Foundation would be paid out of contributed funds.

3. Microsoft would make available to all eligible schools a standard subscription to its TechNet technical support program.

4. Microsoft would guarantee at least 200,000 refurbished computers would be available through Microsoft Authorized Refurbishers to the eligible schools during each of the five years of the settlement period. The computers would consist of Macintosh computers or Pentium-class personal computers or better. The Foundation would establish specifications intended to provide the best technology reasonably likely to be available. Microsoft would make up any annual shortfall in the number of available machines by donating computers that met the specifications. The Foundation would pay the greater of one-half of the cost of the refurbished computers or the actual cost less $50. This means that in the present market eligible schools would pay $50 for each refurbished computer since the current estimated cost of refurbished computers meeting the projected quality standards is $130 to $150.

5. Subject to certain caps, Microsoft would provide free of charge a wide range of its own software for all computers owned by eligible schools or acquired by them through the Foundation. Donated software would include Microsoft software designed for use on Macintosh computers.

6. The Foundation’s funds earmarked for technology acquisition would be used for purchasing refurbished computers, new hardware and other technology equipment, and non-Microsoft software. As indicated above, the Foundation would pay the greater of one-half the cost of refurbished computers or the actual cost less $50. It would pay one-third the cost of new hardware and other technology equipment. It would pay the entire cost of non-Microsoft software.

7. When seeking grants, eligible schools would be free to package their requests in any way they deemed proper, e.g., one school might seek to purchase only refurbished computers with donated Microsoft software, another school might seek to purchase a combination of refurbished and new computers with different kinds of software, or another school might seek to purchase only Macintoshes with donated Mac Office. In awarding grants, the Foundation could not discriminate in favor of schools requesting refurbished computers or Microsoft software.

B.

It is undisputed that Microsoft would be committed to contributing $400 million in cash to the Foundation and that it would be required to contribute up to an additional $100 million on a $1 to $2 matching basis for each dollar the Foundation raised from other sources. The other components of the settlement are valued differently by various parties. Professor Keith Leffler, an economist retained by the MDL plaintiffs, values the settlement at approximately $1.6 billion; Professor Jeffrey MacKie-Mason, an economist retained by counsel for the California plaintiffs who object to preliminary approval, values the settlement at approximately $705 million; and Professor Robert Hall, an economist retained by Microsoft, values the settlement at between approximately $1 billion and $1.6 billion. 1

II.

I will first address three objections lodged against the proposed settlement that would not cause me to deny preliminary approval: the alleged unavailability of a cy pres remedy as a matter of law; alleged irreconcilable conflicts among the members of the proposed settlement class; and considerations of comity.

The California plaintiffs and other objectors argue that the Fourth Circuit’s decision in Windham v. American Brands, Inc., 565 F.2d 59 (4th Cir.1977), cert. denied, 435 U.S. 968, 98 S.Ct. 1605, 56 L.Ed.2d 58 (1978), forecloses a cy pres recovery here. I do not agree. All that the court held in Windham is that a “fluid recovery” theory cannot be used as a mechanism for certifying a litigation class. Id. at 72; see also Eisen v. Carlisle & Jacquelin, 479 F.2d 1005, 1012 (2d Cir.1973), vacated on other grounds, 417 U.S. 156, 94 S.Ct. 2140, 40 L.Ed.2d 732 (1974). That is an entirely different question than the one presented here.

Although the cy -pres approach is most frequently used for the purpose of distributing the residue of a class settlement fund, see, e.g., Powell v. Georgia-Pacific Corp., 119 F.3d 703 (8th Cir.1997); In re Motorsports Merck. Antitrust Litig., 160 F.Supp.2d 1392 (N.D.Ga.2001); Jones v. Nat’l Distillers, 56 F.Supp.2d 355 (S.D.N.Y.1999), it has also been utilized as a means for distributing the entirety of a class fund where the proceeds cannot be economically distributed to the class members. See, e.g., In re Toys “R” Us Antitrust Litig., 191 F.R.D. 347 (E.D.N.Y.2000); N ew York v. Reebok Int’l Ltd,., 903 F.Supp. 532 (S.D.N.Y.1995), aff'd, 96 F.3d 44 (2d Cir.1996); 2 Herbert B. Newberg & Alba Conte, Newberg on Class Actions § 11-20 (3d ed.1992); 7B Charles Alan Wright, Arthur R. Miller & Mary K. Kane, Federal Practice and Procedure § 1784, at 84 (2d ed.1986). As Professor Miller succinctly stated during the course of his presentation at the preliminary approval hearing, “[C]ourts are saying, the game isn’t worth the candle.”

In accordance with these authorities, if I ultimately were to find that the proceeds of any class recovery could not be economically distributed to class members — and if I were to find (as I presently do not) that the Foundation contemplated by the agreement was sufficiently funded to fulfill its intended purpose without potential adverse effects upon competition — I would approve the settlement. 2

Objectors to the proposed agreement argue that there are so many conflicts among the members of both the litigating and the proposed settlement classes that the agreement cannot withstand scrutiny under Amchem Prods., Inc. v. Windsor, 521 U.S. 591, 117 S.Ct. 2231, 138 L.Ed.2d 689 (1997). In an earlier opinion I ruled that Illinois Brick v. Illinois, 431 U.S. 720, 97 S.Ct. 2061, 52 L.Ed.2d 707 (1977), bars any suit for monetary damages under the federal antitrust laws by persons who did not purchase their licenses for operating system or applications software directly from Microsoft. Some states have statutorily repealed the Illinois Brick rule under their own antitrust laws; courts in several other states have held that the rule does not apply to the licensing of computer software. Therefore, the internal class conflict most strongly emphasized by counsel for the California plaintiffs and other objectors is between “repealer” and “non-repealer” states.

Within those two broad subclasses, however, there are a multitude of other potential conflicts. For example, lawyers for plaintiffs in states where classes have already been certified contend that their claims should be valued substantially higher than claims in states where class certification has been denied. Counsel for plaintiffs in some non-repealer states assert that, because of the alleged viability of other statutory or common law claims under the laws of their states, their claims are stronger than claims in other non-repealer states and are just as strong as the claims in repealer states. Moreover, within individual states, lawyers argue that the claims of the particular plaintiffs they represent are stronger than the claims of other plaintiffs. For example, it is said that, because of one of the findings made by Judge Jackson in the government’s case against Microsoft, purchasers of licenses to Windows 98 have stronger claims than other plaintiffs.

Several state attorneys general have likewise lodged objections on a variety of grounds, some of which relate to matters of general antitrust policy while others are premised upon more parochial considerations. Thus, the attorney general for Minnesota complains about a settlement designed to help the nation’s poorest schools because, inter alia, it allegedly would unfairly discriminate against his state because of its comparative wealth. The attorney general for Massachusetts likewise objects, arguing that each Massachusetts consumer should recover a $25 penalty under Mass. Gen. Laws Ch. 93A, § 9.

Conflicts among class members thus abound. Amchem, however, is distinguishable in one critical respect. Involving, as it did, an attempted nationwide class settlement of all asbestos claims, present and future, the proposed agreement in Am-chem would have been binding on persons who did not yet know what injuries, if any, they would eventually suffer as a result of their having been exposed to asbestos. Potential plaintiffs, therefore, were entirely without adequate information to know whether or not they should opt out of the class. Amchem, 521 U.S. at 628, 117 S.Ct. 2231. In contrast, the present case involves conduct and alleged injury that has already occurred. All potential plaintiffs know whether they purchased Microsoft licenses during the class period, and each one of them, upon being notified of the proposed settlement, could make a rational decision whether or not to remain in the settling class.

This is not to say that in every case a realistic opportunity to opt out, standing alone, provides a solution to the problems caused by internal class conflicts. It is, however, a factor to be considered. That is especially true in a case such as this where (as the plaintiffs themselves emphasize in opposing a cy pres recovery) many members of the class have large claims and the incentive to scrutinize the terms of the proposed settlement carefully when deciding whether to exercise their opt-out right.

In any event, there is a more fundamental point I find to be dispositive of the issue. If a cy pres class recovery is appropriate (which, subject to resolution of factual questions regarding the economic feasibility of distributing a class recovery to class members, I find it to be), any disagreements among class members as to the value of their respective claims is irrelevant. The fact that different class members have claims of different potential value is material only if the class members themselves are directly receiving the benefits of the settlement. It would, for exam- pie, in that event be significant that the value of the claim of A, a class member who purchased a license for Windows 98 at full retail price in a repealer state where a class has been certified, was x, whereas the value of the claim of B, a class member who purchased at a discounted price in a non-repealer state a license for another software product with a lower alleged overcharge, was y. But if neither A nor B himself receives any money personally from the settlement, it does not matter that their claims are different from one another. At least that is true if A and B stand in equal relation to C, the third party who is being benefited by the cy pres recovery. In this case, the latter condition would be met since, although A and B, by definition, purchased their software in different states, they would, as residents of the United States, benefit equally from a bridging of the “digital divide,” a national problem the proposed settlement is designed to address. 3

C.

Objectors to the proposed settlement make the related argument that since I have held that Illinois Brick bars any monetary recovery by indirect purchasers under the federal antitrust laws, plaintiffs’ claims should be resolved solely on a state by state basis. 4 In other words, according to the objectors, considerations of comity foreclose any national settlement.

I am mindful of the institutional damage that can be caused by federal courts’ unwarranted intrusion into state judicial proceedings. I also am cognizant that the approval of a national class settlement on the terms that have been proposed would result in the release of all state, as well as any viable federal, claims. But while the considerations require that I carefully consider the interests of the various states in assessing the adequacy and propriety of the proposed settlement, I am not persuaded that they prevent the very possibility of a national settlement.

The absence of a monetary remedy for indirect purchasers under the Sherman Act does not mean that the resolution of the widespread consumer antitrust litigation pending against Microsoft both in the state and federal courts is not a matter of appropriate federal concern. At least in theory, plaintiffs still have a claim for far-reaching injunctive relief that is unaffected by Illinois Brick. 5 Moreover, the orderly and efficient operation of the national economy self-evidently is a federal interest, and that interest dictates that where (as unquestionably is the case here) a federal court’s jurisdiction has been properly invoked, the federal court is an appropriate forum for forging a constructive solution to a multi-jurisdictional problem of national scope. The ultimate fate of any settlement would lie in the hands of the class members who, if they found it unacceptable, could undo it by exercise of their right to opt out. But if a settlement meeting the requirements of federal law was reached, it could be effected under the aegis of this court and through the exercise of its power. Comity does not require tolerance of economic balkanization and litigation chaos.

III.

I will now address the first of the issues that prevent my granting preliminary approval to the settlement agreement. Assessment of the adequacy of a proposed settlement requires evaluation both of the value of the settlement and the value of the claims being settled. Although, as I have indicated in part I.B., the proponents and the opponents of the proposed settlement disagree about its worth, the record contains sufficient information by which to evaluate the parties’ respective contentions and reach a conclusion as to a reasonable range of value. However, as to the second factor in the equation, the parties have widely divergent views which, on the present record, are largely theoretical and have not been sufficiently tested.

Professor MacKie-Mason, the expert economist retained by counsel for the objecting California plaintiffs, has estimated that the national class can prove single, i.e., non-trebled, damages in the range of $10.3 billion to $18.9 billion. The upper level of overcharges estimated by the MDL plaintiffs’ economic expert, Professor Leffler, falls within Professor MacKie-Ma-son’s range. However, Professor Leffler states various reasons why he believes his upper level estimate might not be recoverable. Of course, as both Professors MacK-ie-Mason and Leffler recognize, their damage estimates do not themselves establish the value of the class recovery because they must be adjusted for litigation risks — a legal, not an economic, judgment.

Microsoft has not proffered specific damage estimates, but has challenged various assumptions made by Professors MacKie-Mason and Leffler. Its economist, Professor Hall, has submitted testimony in which he stresses the necessity of tailoring damage claims to specific alleged antitrust violations — something which, according to Professor Hall, has not been done by plaintiffs’ experts. He also critiques the methodologies employed by Professors MacKie-Mason and Leffler, and he opines that even under the approaches they have adopted, if different assumptions are made, the damages could be as little as zero to $200 million. Microsoft also cites an independent study conducted by McKinsey Global Institute, which concluded that the quality-adjusted price of Microsoft’s operating systems, from DOS in 1988 to Windows in 1998, declined at a compound rate of 11.7 percent. While ac knowledging that declining prices are not theoretically inconsistent with overcharges, Microsoft points to the McKinsey report as evidence that Professor MacKie-Mason’s damage estimates are unrealistic and would not be appealing to a jury.

The preliminary approval process has, perhaps, had the salutary effect of causing the parties to focus more closely upon the viability of various damage theories. Were settlement negotiations to begin now, the record that has been created over the past weeks would require supplementation in several respects, but would provide a substantial basis for evaluating the strengths and weakness of the parties’ positions. However, prior to the start of the preliminary approval process, the parties had not made expert disclosures about damages in the MDL proceedings or (to the best of my knowledge) in any of the state cases. Moreover, while I permitted some informal cross-examination of the experts during the preliminary approval hearing, time was limited and the questioning was far less extensive than it would be on deposition. Further, additional questions were raised by the parties’ arguments that perhaps need to be explored. For example, although the MDL and related state cases have been characterized as “consumer actions,” the record established during the approval process clarifies that business organizations (many of them very large), not individual consumers, hold approximately 85 percent of the claims. This fact may have an impact on various issues, including the propriety of using retail prices as a premise for damage calculations, the determination of appropriate “pass through” rates, 6 and the value of Microsoft’s products as perceived by the business organizations themselves.

IV.

For the reasons just stated, I am not satisfied that the record has been sufficiently developed on various damages issues for me to assess reasonably the value of the class claims. If, however, that were the only difficulty I perceived in the proposed settlement, I might — instead of now denying preliminary approval — defer ruling until a more complete record had been made. The reason I am not following that course is that I have determined that the damages valuation problem aside, the present record establishes that the Foundation contemplated by the agreement is critically underfunded. Therefore, however the potential class recovery is valued, I do not believe that the proposed settlement falls into “the range of possible approval.” Manual for Complex Litigation (Third) § 30.41, at 265 (2000); see also In re Mid-Atlantic Toyota Antitrust Litig., 564 F.Supp. 1379, 1383 (D.Md.1983).

As proffered and testified to by various persons during the preliminary approval hearing, school boards and administrators typically budget $3 of support and training for every $1 of technology acquisition. If a $1 billion value is placed upon the proposed settlement, that ratio is converted into a ratio approximating $1 of support for every $3 of technology acquisition. This is so because under the proposed agreement Microsoft would contribute between $650 and $750 million for technology acquisition (assuming a $500 million value for its donated software), as compared to $250 million for technical support and professional development. 7

Dramatic though this ratio reversal may appear, it would not necessarily be fatal to the proposed settlement. As the settlement’s proponents point out, some of the technology acquisition funds would be used to purchase replacement, rather than new, hardware and software. The support and training funds for the hardware and software being replaced presumably are already in the schools’ budgets. Moreover, the Foundation is not intended to cure the problems of the digital divide all by itself. What it would do would be to provide substantial seed money and technology to the boards and administrators responsible for the operation of the eligible schools and serve as a catalyst for further funding and innovation. In that regard, the proposed settlement agreement does not contemplate providing permanent funding for salaried staff positions in the schools. Rather, the agreement provides that in considering grants for technical support funds, the Foundation would give preference to clinical and student-based programs, such as IT Academy Clinical Programs and the Gen-SCI program, and encourage volunteer organizations, like TechCorps.

Nevertheless, even as designed, the proposed settlement is thinly funded. Somewhat troublesome in itself, this fact becomes more acutely problematic when viewed in the context of the proposed settlement’s potential adverse impact upon competition. If more fully funded, the Foundation would be able to award grants which would reduce the danger of anti-competitive effects.

On its face the agreement is entirely “platform neutral.” It would empower eligible schools to formulate their grant requests in any way they deem proper, and it specifically would prohibit the Foundation from discriminating in favor of Microsoft (or any other manufacturer or supplier) in making decisions on grant applications. The agreement also would create a source of funding for eligible schools to purchase non-Microsoft software and Macintoshes (“Macs”), as well as personal computer (“PC”) hardware. However, the agreement raises legitimate questions since it appears to provide a means for flooding a part of the kindergarten through high school market, in which Microsoft has not traditionally been the strongest player (particularly in relation to Apple), with Microsoft software and refurbished PCs.

The first provision requiring scrutiny is the one under which Microsoft would guarantee that during each of the five years of the settlement period, 200,000 refurbished computers, meeting objectively defined quality and performance standards, would be made available to the eligible schools through Microsoft Authorized Refurbish-ers. The difficulty with this provision is that businesses, governmental agencies, and other organizations are the primary suppliers of used computers to refurbish-ers, and 95 percent of them use PCs rather than Macs. Moreover, although eligible schools are free to request funding from the Foundation for new computers (including Macs) rather than refurbished computers, if the refurbished computer program is successful (as the proponents of the settlement assert it will be), it would be more rational for the schools to seek to purchase refurbished computers instead of new computers since that would maximize their use of resources. A reasonably foreseeable effect of these combined circumstances would be to increase the number of PCs, both absolutely and in comparison to Macs, being used in the eligible schools. This, in turn, might depress the development of software designed for use in Macs, further reducing the attractiveness of Apple products over time. 8

The agreement’s provisions relating to Microsoft’s donation of free software likewise require thoughtful analysis. Self-evidently, these provisions raise antitrust concerns from the perspective of other software manufacturers. To put it bluntly, in the words of the opponents of the proposed settlement, the donation of free software could be viewed as constituting “court approved predatory pricing.” 9 Again, although eligible schools would be free to apply for grants for non-Microsoft software, and although the Foundation would be bound by its charter not to discriminate against schools making such grant requests, it would be a fact known both to the schools and to the Foundation that, under the agreement, Microsoft software would genuinely be free, whereas the purchase of non-Microsoft software through the Foundation would decrease the amount of funds available for other purposes.

Real though these problems are, they could perhaps be overcome by a settlement in which the Foundation was more substantially funded. Although the opponents of the proposed settlement vigorously contest the amount Microsoft has agreed to pay (in cash and in kind), they suggest that a “simple and elegant” solution would be the creation of an all-cash fund. Having donated the monies to create the fund, Microsoft could then compete with other software manufacturers to sell licenses for its products to the eligible schools through the grants program.

This approach would require Microsoft to infuse more cash into the settlement on the front end, but if successful in the ensuing competition created by the Foundation’s grants, Microsoft would recoup a substantial portion of its donation. Alternatively, though it would require a careful balancing, a settlement might be structured under which Microsoft would continue to provide its own software free of charge, but would contribute a greater amount of cash to the Foundation to be made available for the purchase of non-Microsoft software requested by the schools. 10 Likewise, if a restructured settlement were to continue to include a refurbished computer component, in order to address anti-competitive concerns, greater funding would have to be provided for the purchase of new computers and the percentage of the cost of new computers paid by the Foundation might have to be increased. 11

I have described some of the adjustments that might be made to the settlement proposal for the purpose of demonstrating the interrelationship between the underfunding of the Foundation and competitive concerns. I recognize that those adjustments themselves might be properly perceived as inadequate, either in enhancing the ability of the Foundation to fulfill its intended mission or in reducing adverse competitive effects. Obviously, I pass no judgment upon objections not yet made to a settlement agreement not yet reached. The time to consider such objections would come if and when renewed negotiations resulted in the submission of a revised agreement for approval. For the present, however commendable the efforts of counsel and however laudable the goals of the proposed Foundation may be, I deny the motion for preliminary approval of the settlement that has been presented.

1 . Each of the listed values is from the perspective of the benefits received by the Foundation and the eligible schools and their students. As the opponents of the proposed settlement point out, the cost to Microsoft of the proposed settlement would be much less since it could deduct its contributions for tax purposes and since its marginal cost of producing the donated software would be nominal. Moreover, the settlement would generate goodwill that Microsoft might be able to convert to immediate marketing advantage in the education market. Another incidental benefit to Microsoft would appear to be that by effectively assuming control over the refurbished computer market, it could decrease the incidents of software piracy inherent in that market.

2 . Although the record is not fully developed on the question of whether a class recovery could be economically distributed to class members, I might have been content to leave the development of a supplemental record to confirmatory discovery if I had been persuaded that preliminary approval should otherwise be granted. The sheer number of potential class members (it has been estimated that there are as many as 100 million possible claimants), the transient nature of the U.S. population, the high rate of software piracy, and the fact that many individual consumers would not have retained proof of purchase documents all point to the unfeasibility of economic distribution of class proceeds. Moreover, Dr. Harvey Rosen, an expert retained by the MDL plaintiffs, has testified that the cost of processing claims in other cases has ranged from $32 to $292 per claim. Although Dr. Rosen concedes that in this case there would be certain economies of scale and although the objectors challenge his figures, their own expert estimates that the total processing cost per claimant would be $7.52 to $9.00. While other factual questions would need to be explored, including (1) the ease and cost of identifying the business organizations who the parties agree hold the bulk of the potential claims, and (2) the possibility of using a “product key” written into every item of Microsoft’s software as a means for verifying a claim, the practical difficulties of distributing any class recovery seem immense.

3 . Of course, all plaintiffs also were residents of a state (or, in the case of many organizations, multiple states) at the time they purchased the licenses to their software. However, contrary to the implicit argument of counsel for the objectors and several attorneys general, this fact does not confer any “group rights” upon consumers who are residents of the same state or confer any right of recovery upon states whose laws would permit a monetary recovery by their residents. The rights held by the members of any plaintiff class, state or federal, are entirely individual and belong only to each of the class members. If an appropriate cy pres settlement designed to bridge the digital divide on a nationwide basis were presented and preliminarily approved, the class members would, by virtue of their opt-out right, themselves have the power to decide whether to relinquish their right to monetary relief in favor of helping to solve a problem of national scope. Some might opt out because they reject the concept of a cy pres settlement at all; others might opt out because, while accepting the concept of a cy pres settlement, they believe that the state in which they reside should receive a higher percentage of support for its schools. But the decision is for them — not for the lawyers representing a state class or for the state itself — to make. That said, any proposed settlement would certainly be stronger, and less subject to challenge, if it were supported by all, or a substantial majority of all, counsel who have participated in both the federal and state proceedings.

4 . My Illinois Brick ruling is, of course, subject to challenge on appeal.

5 . I say "at least in theory" because, as a practical matter, the relief obtained in the case brought against Microsoft by the United States and eighteen state attorneys general might render redundant any relief that could be obtained by plaintiffs here. While I mention this issue so as to not overstate the importance of the federal interest embodied in the MDL plaintiffs' claim for injunctive relief, I am certainly not ruling upon it since it has not been even minimally briefed by the parties.

6 . Professor MacKie-Mason assumes that Microsoft’s alleged overcharges were passed through at the rate of 125 percent. Professor Leffler assumes that the pass-through rate was 100 percent. These assumptions seem somewhat glib. See Illinois Brick, 431 U.S. at 732, 97 S.Ct. 2061 (referring to the “complexities and uncertainties” of a pass-on defense that “are multiplied in the offensive use of pass-on by a plaintiff several steps removed from the defendant in the chain of distribution”).

7 . On the technical support side, Microsoft also would provide eligible schools with free subscriptions to TechNet, its online technical service. Valued at the ordinary subscription rate, this contribution would be worth $31 million. It should, however, also be noted that the provision of these subscriptions might have some adverse competitive impact since TechNet covers only Microsoft products.

8 . All of that said, in considering the potential effects of the refurbished computer program, it should be remembered that the computers would be furnished to a market of economically deprived schools that traditionally has been underserved. The threat to the wider education market is collateral and indirect, arising from the danger that school boards, having purchased refurbished PCs through grants from the Foundation, might standardize their technology throughout their district, shifting from Mac-based to PC-based products. For sake of the completeness of presentation, it should perhaps also be noted that, although Apple has traditionally been strong in the education market, its share of new unit sales has been steadily declining in recent years. Presumably, this is due to the fact that PCs are cheaper than Macs. According to Professor Leffler, even if 100 percent of the Foundation grants were used to purchase PCs, the effect on Apple’s overall share of the kindergarten through twelfth grade market would be negligible, declining by only 1.5 percentage points.

9 . As Professor Hall points out, this characterization is somewhat hyperbolic without a credible proffer that Microsoft would be able to recover the "losses” it suffered from its donation of free software after the five-year settlement period has expired. However, the specter of a market being inundated with free software provided by a manufacturer, which in other contexts has been found to have abused monopoly power, understandably is a matter of concern to Microsoft's competitors.

10 . A variation of this alternative might be to set caps, significantly lower than those established in the present agreement, on the amount of software that could be donated by Microsoft, making up the difference between the value of that software and the value of the software projected to be donated under the present agreement by a payment in cash. That approach might enable Microsoft to pay part of the settlement in kind while simultaneously decreasing the risk of anti-competitive effects substantially.

11 . Under the present agreement, the Foundation would pay for only 33 percent of the cost of new computers compared to between 61.5 percent and 66 percent of refurbished computers (based upon their current cost of $130 to $150).