6
Va. J.L. & Tech. 7 (2001), at http://www.vjolt.net
1522-1687 / 2001
Virginia Journal of Law and Technology Association
VIRGINIA JOURNAL of LAW and TECHNOLOGY
|
UNIVERSITY OF VIRGINIA |
SPRING 2001 |
6 VA. J.L. & TECH. 7 |
Microsoft
III[1] and the Metes and Bounds
of Software Design and Technological Tying Doctrine[2]
By Charles M. Gastle and Susan Boughs[3]
[T]he most enlightened judicial policy is to let people manage their own
business in their own way, unless the ground for interference is very clear...[4]
I. Introduction
II. The Plausible Claim vs. Consumer Demand Technological Tying Test
A. Court of Appeals Plausible Claim Test of Technological Tying
B. Judge Jacksons Rejection of the Plausible Claim Test
C. The Distinction between Functional Integration and
Market/Functional Linkage
E.
The
Consumer Demand Test and Tying Doctrine Generally, Are Ripe for Review
F.
But will
the Microsoft III Court of Appeals Accept the Challenge?
III. Experience & Logic: Operating Systems Evolve
A. Experience: The Evolution of Minix to Linux through the Addition
of Functionality
B. Logic: Charting an Evolutionary Trajectory through an
Exponentially Expanding Functional Universe
C. Experience & Logic Does Not Support Per Se Treatment
IV. The Enigma of Weighing Dynamic Efficiencies
A. The Fragile Nature of Increasing Returns Economics
B. The Remarkable Mixture of Stilted Antitrust Theory with Avant Garde
Economic Theory
C. Coase & The Nature of the Operating System: A Surrogate Theory
for Software Design
D. A Cosian Explanation of Operating System Design in the Microsoft
Case
V. Conclusion
1.
In
Microsoft III, Judge
Jackson applied the consumer demand test to determine the issue of product
separability. Remarkably, he rejected guidance from the Court of Appeals which
indicated that tying could not be established if there was a plausible claim to
efficiency resulting from the integration of Internet Explorer into Windows.
The consumer demand test is consistent with tying doctrine, but the plausible
claim test is justified in technological tying cases involving software. A
static snapshot of the pattern of consumer demand is not suitable for
determining product separability because it changes as the operating system
evolves over time by adding functionality. Some form of efficiency analysis
must be undertaken and the alternatives were provided in the Court of Appeals
decision: the plausible claim test proposed by the majority and the efficiency
calculus proposed by the minority. Courts do not appear to have the necessary
institutional competence to weigh efficiencies against possible anticompetitive
aspects of the integration. Increasing returns economics is a positive theory
and it is not capable of predicting the evolutionary trajectory of complex
innovation processes. The bias of antitrust law and the impossibility of
designing an objective measure that can weigh efficiency claims against
antitrust values- such as buyer freedom- suggests that a deferential test to
software design decisions is justified.
2.
The holding that
Microsoft engaged in technological tying when Internet Explorer was integrated
into the Windows 98 operating system is a disturbing aspect of the judgment
handed down by Judge Thomas Penfield Jackson in Microsoft III. Jackson characterized browsers and operating systems
as two separate products on the basis that a separate consumer demand for
browsers existed during the relevant time period. The Microsoft Judgment
includes definitions of an operating system,[5]
middleware,[6]
and, by default, applications. These definitions will help define the software
market for the duration of the judgment - a period of ten years. To the extent
that the Microsoft Judgment limits the design decisions that can be made by
Microsoft programmers, the question arises whether the courts should engage in
this kind of software engineering.
3.
The central issue in a
technological tying determination is whether the operating system and browser
represent one product or two. If the combination constitutes one product,
Microsoft should have the freedom to defend its core business asset. The
potential costs to the Internet economy may be significant if the case is
decided in error, due to the manner in which innovation processes may be
compromised not only in this case but in others that will be decided on the
principles established herein. Surprisingly, the implications are such that the
open source software movement may be vulnerable to an antitrust complaint,
which seems a somewhat bizarre result, but serves an appropriate litmus test as
to the rationality of Judge Jacksons decision. The open source Apache server
software[7]
is made vulnerable due the nature of competition in the computer industry, such
that monopolies in various market segments are likely. Instead of recognizing
the dynamics of this process and being sensitive to them, the bias of antitrust
principles will tend to interfere with the industry.[8]
4.
Microsoft III
represents a crossroads in both law and economic theory. From the legal standpoint,
tying doctrine is one of the few remaining per
se antitrust offenses, where anticompetitive intent and effect are presumed
once certain prerequisites are established. Tying doctrine has resisted the
kind of rationalization that the balance of antitrust law has undergone, where per se offences have been replaced by a
rule of reason analysis. The erosion of per
se liability offenses has occurred in part through rigorous analysis on the
basis of economic efficiency, as championed by the Chicago School. Judge
Jackson adopted the traditional consumer demand test to determine the issue of
market separability, using a static snapshot of the pattern of consumer demand
for operating system functionality, even though such demand is evolutionary in
nature and subject to change. He did so despite adoption by the Court of
Appeals of a test in Microsoft II
which was deferential to software design decisions as long as there was a
plausible claim to technical value.[9] The
appellate decision was released shortly after the Microsoft III Complaint was issued in May, 1998, and the majority
of the Court of Appeals indicated specifically that the plausible claim test
was intended to provide guidance in subsequent litigation.[10]
Remarkably, Judge Jackson distinguished his own Court of Appeals and then
attempted to bypass it by certifying a direct reference to the Supreme Court.[11]
The rejection of the case by the Supreme Court,[12] sets
the stage for the Court of Appeals argument that occurred on February 26th-27th,
and the reasons for judgment that are expected to be released in the next month
or two.[13]
5.
This article starts
with an analysis of the Court of Appeals decision in Microsoft II, with
an explanation of the majoritys plausible claim test and the efficiency calculus
proposed by the minority. Judge Jacksons rejection of the guidance of the Court
of Appeals decision is then reviewed, along with his acceptance of the consumer
demand test of product separability. The consumer demand test has been seen as
a way in which to admit a rule of reason analysis through the backdoor.[14]
In Microsoft III, it is used in a per se manner and one which avoids any
examination of efficiency claims. If efficiency claims cannot be ignored, the
consumer demand test is inadequate to determine the issue of product
separability in the software market. We suggest that logic and experience[15]
do not support the application of the customer demand tests per se reasoning with respect to the issue of product
separability. To the contrary, they suggest that the efficiencies involved in
software integration must be considered in some manner. An analysis of the
evolutionary pressure on operating system design is then undertaken to show
that the pattern of consumer demand must and will change over time. In light of
this, the consumer demand test is not appropriate and cannot act as a proxy for
an efficiency analysis. Measuring consumer demand during a base period results
in a static analysis that is out of step with reality.
6.
The evolutionary
pressure on operating systems is illustrated by Andrew Tanenbaums Minix
program, developed as a teaching aid for university courses on operating system
design. Minix enjoys a rather special place in software lore, as Tanenbaums
textbook and code was the inspiration for Linus Torvalds to write the Linux
kernel, which, at least initially, was a Minix clone. Tanenbaum restricted the
development of Minix to keep it within a student’s budget, and he ignored
pressure to add functionality to the program. His decision to do so made
possible the Linux phenomena, because Torvalds had no such reservations in
adding functionality. The limited functionality that the initial Minix
operating system contained fits well within the narrow technical definition of
an operating system included within the Microsoft Judgment. The definition is
out of touch with the design of existing operating systems, as it excludes much
of what has already become an accepted part of the various operating systems
(i.e., the user interface). The evolution from Minix to Linux through the
adoption of new functionality that falls outside the narrow definition, as well
as the fact that all operating systems contain a functionality that does
not fall within the Judgment definition, suggests that it is too restrictive
and unsuitable to guide the development of the software industry for the next
ten years.
7.
The article then turns
to the economic theory underlying Judge Jackson’s decision. He is forced to
adopt an avant garde theory - increasing returns economics - that suggests that
it is possible for the marketplace to lock-in to an inferior technology. Judge
Jacksons acceptance of this theory appears to result in part from the fact that
Microsoft does not act as a traditional monopolist by restricting output to
increase price. In the absence of this traditional antitrust evil, inhibition
of innovation based on concepts such as network effects, path dependency, and
lock in, supplies a policy justification for intervention - a newly discovered
flaw in the market system.[16] The
possibility that inferior technologies might achieve lock-in, and/or maintain
their position through active management, allegedly gives rise to an antitrust
concern and provides a reason to intervene. This theory has made the transition
from theory to policy in antitrust innovation market analysis[17]
and the Microsoft decisions. In both
instances, the objective is to measure dynamic efficiencies in recognition that
traditional static price theory is not well suited to do so. However,
increasing returns economics is fragile and its assumptions are not robust.
Innovation processes are complex in nature, and the relationship between
concentration and innovation intensity is ambivalent. The question then arises
whether an intrusive structural remedy can be justified by an archaic antitrust
theory, dependent upon the utilization of a new economic paradigm which is
incapable of predicting the evolutionary trajectory of innovation processes in
the affected market.
8.
If the efficiencies of
the software integration cannot be ignored, either the Court of Appeals
majority plausible claim test or the minority efficiency calculus analysis
should be utilized. The choice will depend on the institutional competence of
the courts to undertake the full blown efficiency analysis implicit in the
minoritys calculus. Their ability to do so is hampered by the weakness of
increasing returns economics, which is positive in nature and only able to
explain the manner in which a particular market evolved; it cannot predict the
course of innovation. It is incapable of identifying the manner in which
intervention should be structured to improve innovation and carries with it a
danger of picking winners and losers, a fact that the Department of Justice
recognized at the time of the 1995 Consent Decree in Microsoft I.[18]
9.
The article notes that
there is a gap in economic theory created by the lack of an explanation from
the standpoint of software design as to what functionality is properly included
within an operating system or application. This is similar to the gap in
neoclassical economics, which lacks a rigorous theory explaining the existence
of firms and the economic activities that they undertake. The absence of such a
theory in both instances feeds the bias implicit in antitrust doctrine, such
that any ununderstandable market practice is sought to be explained in terms of
monopoly abuse.[19]
This article looks for a surrogate theory of software design, akin to Ronald
Coases theory of the firm,[20] that
attempts to explain the existence and role of firms in terms of transaction
costs, which may be defined more broadly as risk management. Approaching
software design from the standpoint of risk management, a rational basis for
incorporating new functionality into the operating system is provided. In
particular, browsing functionality is an obvious candidate due to the
importance of the Internet, which represents an inflection point in the
information economy. This challenges Judge Jacksons conclusion that Microsoft
engaged in predatory innovation.[21]
10. The article concludes with the
recommendation that the difficulty in utilizing an efficiency calculus is such
that the plausible claim test is the only principled alternative. To carry out
a full blown efficiency analysis, the court must measure the intrinsic value of
dynamic efficiencies of an innovation process, and subsequently balance them in
some manner against possible anticompetitive effects. This is difficult to do
in an objective manner. Increasing returns economics has difficulty identifying
the value of particular integrations and the effect they will have on the
course of innovation. This difficulty results from the complexity of different
sources of evolutionary pressure acting in combination on operating system
design. The complexity is such that the course of innovation in the software
market can exhibit a sensitive dependence upon initial conditions, which is a
hallmark of chaos theory (although the software market does not descend into
the instability associated with chaotic processes).[22]
11. Even if innovation processes could
be predicted and measured against one another, a question arises as to what
value antitrust principles such as buyer freedom should be given by courts.
When market power is found to exist, antitrust doctrine views market practices
with suspicion, a questionable predisposition in a market segment in which
serial monopolies are to be expected as the result of normal market processes.[23]
To a certain extent, weighing technical efficiencies against entrenched
antitrust values involves measuring apples and oranges, and cannot be done in
an objective manner. The efficiency analysis, along with the application of
increasing returns economics, provides an opportunity for the re-emergence of
political values in antitrust - such as the preservation of rivalry at the
expense of economic efficiency - that had been eclipsed with the rise in the
influence of the Chicago School during the past thirty years.
12. The Court of Appeals plausible claim
test, therefore, should be used in cases where technological tying is alleged,
at least in the Software market. The majority of the Court of Appeals appears
to have been correct in reminding Judge Jackson of the limits of the trial
courts institutional competence in this regard.
13. At the time of writing, the oral
argument in the Microsoft III appeal has been completed with the
decision under reserve. It appears that the Court of Appeals is poised to set
aside Judge Jacksons finding that Microsoft engaged in technological tying and
to do so on the basis that no separate browser market exists. It appears that
the determination will be made on the basis of an application of the consumer
demand test and likely not on an acceptance of the plausible claim test. This
is unfortunate, as the determination will be made once again on the basis of a
static snapshot of market characteristics and not on the unique attributes of
the dynamic market process of the software market segment.
II. The Plausible Claim vs. Consumer Demand
Technological Tying Test
14. Microsoft III involved three main issues:
maintenance of a monopoly, attempting to monopolize, and technological tying.
The issue of whether the operating system and browser represents one market or
two underlies at least two of these issues. One cannot attempt to monopolize a
market that it already monopolized and one cannot be involved in tying two
products if only one product actually exists. It also might underlie certain
aspects of the claim that Microsoft illegally maintained its monopoly, at least
the extent to which this allegation relates to the technical integration at
issue.[24]
A distinction might be made between the persistence of illegitimate monopoly
power[25]
and a monopoly lawfully acquired through internal growth and greater business
acumen[26]
reacting to clearly discernable evolutionary pressures through the integration
of new functionality.
15. The integration of Internet Explorer
into Windows was the factor that gave rise to the current round of Microsoft
antitrust proceedings. In Microsoft I
in 1995, the Department of Justice (DOJ) and Microsoft entered into a Consent
Decree that provided certain restrictions on Microsofts per processor licensing
practices, but expressly recognized Microsofts ability to integrate new
functionality into its operating system.[27] Microsoft II commenced in October, 1997,
when the DOJ began contempt and injunctive proceedings against Microsoft on the
basis that it breached the Consent Decree by requiring computer manufacturers
to install Internet Explorer as a condition to licensing Windows 95. Judge
Jackson found that Microsoft had not violated the Consent Decree because it was
too vague to be enforceable, but, in any event, he found that there was a high
probability that Microsoft had violated the Sherman
Act by conditioning its license of Windows 95 on the license of Internet
Explorer.[28]
The injunction was set aside by the Court of Appeals on June 23rd, 1998, but it
had been overtaken by that point by a new and broader proceeding arising from
Microsofts intended release of Windows98. The majority decision attempted to
give Judge Jackson guidance in the new proceeding on the meaning of the term
integration and whether Windows and Internet Explorer represent one product or
two.[29]
Judge Jackson refused to follow this guidance, and instead relied upon the more
traditional consumer demand antitrust test of product separability, thus giving
rise to one of the most important issues on appeal.
16. In overturning Judge Jacksons injunction, the majority of the Court of Appeals found that the United States had failed to present evidence suggesting that Windows 98 was not an integrated product and thus exempt from the prohibitions of the Consent Decree. The Department of Justice also did not dispute Microsofts claim that Windows95 and Internet Explorer were technologically integrated and commented that a simple way to harmonize the parties desires was to read the integration provision of the Consent Decree as permitting genuine technological integrations, regardless whether elements of the integrated package are marketed separately.
This reading requires us, of course, to give substantive content to the concept of integration. We think that an integrated product is most reasonably understood as a product that combines functionalities (which may also be marketed separately and operated altogether) in a way that offers advantages unavailable if the functionalities are bought separately and combined by the purchaser.[30]
And further:
So the combination offered by the manufacturer must be different from what the purchaser could create from the separate products on his own. The second point is that it must also be better in some respect; there should be some technological value to integration .... The concept of integration should exclude a case where the manufacturer has done nothing than to metaphorically bolt two products together, ....[31]
17. The court was cognizant of the
institutional limits to which it was subject that put the legal process in a
relatively weak position to design a remedy to improve the process of innovation.
The majority commented that [a]ntitrust scholars have long recognized the
undesirability of having courts oversee product design, and any dampening of
technological innovation would be at cross-purposes with antitrust law.[32]
As a result, the Court of Appeals was prepared to defer to the software
designers, expressing concern that any interpretation of the Consent Decree
that prohibited the distribution of Windows 98 would put judges and juries in
the unwelcome position of designing computers.[33]
18. In its majority opinion, the court stated:
The short answer is thus that integration may be considered genuine if it is beneficial when compared to a purchaser combination. But we do not propose that in making this inquiry the court should embark on product design assessment. In antitrust law, from which this whole proceeding springs, the courts have recognized the limits of their institutional competence and rejected theories of technological tying. A courts evaluation of a claim of integration must be narrow and deferential. As the Fifth Circuit put it, [S]uch a violation must be limited to those instances where the technological factor tying the hardware to the software has been designed for the purpose of tying the products, rather than to achieve some technologically beneficial result. Any other conclusion would enmesh the courts in a technical inquiry into the justifiability of product innovations.[34]
19. The measure of discretion granted to software designers is reflected in the test posed by the majority, in which [t]he question is not whether the integration is a net plus but merely whether there is a plausible claim that it brings some advantage.[35] Microsoft had clearly met the burden of ascribing facially plausible benefits to its integrated design as compared to an operating system combined with a stand-alone browser such as Netscapes Navigator. Incorporating browsing functionality into the operating system allows applications to avail themselves of that functionality without starting up a separate browser application.[36] The majority also noted that Internet Explorer provides system services not directly related to web browsing, enhancing the functionality of a wide variety of applications and also upgrades some aspects of the operating system unrelated to web browsing. The majority stated that the inquiry does not end at this point because [i]t is also necessary that there be some reason Microsoft, rather than the OEMs or end users, must bring functionalities together.[37] It noted that if Microsoft gave OEMs a separate browser and operating system, OEMs could not knit the two of them together and that [t]his reprogramming would be absurdly inefficient and [c]onsequently, it seems clear that there is a reason why the integration must take place at Microsofts level.[38] The majority concluded that the Windows 95/IE package is a genuine integration. [39] The majority also stated that their interpretation of the integration provision is consistent with tying law.[40]
20. In the minority opinion, Judge Wald indicated that the majoritys plausible claims test was too safe a harbour with too easily navigable an entrance.[41] The minority offered a different test, allowing Microsoft to offer an integrated product to OEMs under one license only if the integrated product achieves synergies great enough to justify Microsofts extension of its monopoly to an otherwise distinct market.[42] An efficiency calculus is proposed which balances two factors: first, whether there are real benefits to the consumer associated with integrating two software products, and second, whether there is independent evidence that a genuine market exists for the two products provided separately. The greater the evidence of distinct markets, the more of a showing of synergy Microsoft must make in order to justify incorporating what would otherwise be an other product into an integrated whole.[43] Judge Wald concludes that this test is more consonant with the .... weight of antitrust law.[44]
21. The majority responded that the efficiency calculus test requires Microsoft to counter evidence of historically separate markets, with evidence of synergistic efficiency gains that courts are not equipped to evaluate.[45] The majority stated:
Apart from the lack of textual support, we think that a balancing test that requires courts to weigh the synergies of an integrated product against the evidence of distinct markets, is not feasible in any predictable or useful way. Courts are ill-equipped to evaluate the benefits of hightech product design, and even could they place such an evaluation on one side of the balance, the strength of the evidence of distinct markets, proposed for the other side of the scale, seems quite incommensurable. .... If, as the record suggests, Microsoft proposed modification of the integration proviso because of concern about vague or subjective criteria, an interpretation requiring courts to weigh evidence that establishes distinctness (or does not) against a sliding scale of net synergistic value looks like the most total transvaluation one can imagine.[46]
22. In the Conclusions of Law issued on April 4th, 2000, Judge Jackson listed the four requirements of liability for tying under Section 1 of the Sherman Act,[47] and then dealt with the product separability issue and the plausible claim test established by the majority of the Court of Appeals. Judge Jackson distinguished his own Court of Appeals, claiming to rely upon the controlling authority of Supreme Court precedent.[48] He stated that he did not believe that the D.C. Circuit intended [their decision] to state a controlling rule of law for purposes of this case, although noting that his finding of liability would arguably be at variance with it and that the majority decision sought to guide this Court.[49] In his view, the departure was justified because the Court of Appeals was interpreting a single provision of a Consent Decree which primarily was a matter of determining contractual intent. Therefore, its observations on the institutional competence of courts to review software design determinations, were in the strictest sense obiter dicta and are thus not formally binding.[50]
23. Judge Jackson believed that the majority opinion showed an extraordinary degree of respect for software design changes and, read literally, would immunize any product design from antitrust scrutiny. [51] He stated that it is inconsistent with Supreme Court precedent[52] and, inter alia, ignored reality because the claim of advantage needed only be plausible and not proved, and it dispensed with balancing the hypothetical advantages against any anticompetitive effects.[53] He noted that in both Jefferson Parish and Eastman Kodak, the Supreme Court considered theoretical valid business reasons for the combination in question and rejected them. He drew the conclusion that the Court of Appeals rejection of a product design assessment as to whether the integration represents a net plus, is at odds with the Supreme Courts own approach.[54] In his view, the two Supreme Court cases established that the correct test of product separability involved the character of the demand for the constituent components and not their functional relationship. The test was met in Microsoft III because consumers perceived that web browsers and operating systems were distinguishable and that a separate demand existed for each independent of the other.[55]
24. In his Conclusions of Law, Judge Jackson determined the product separability issue on the character of consumer demand, and did not engage in evaluating the efficiencies implicit in the integration of Internet Explorer into Windows. However, in the Findings of Fact released five months earlier in November, 1999, Judge Jackson found that the integration - without providing a mechanism for removing Internet Explorer functionality - harmed consumers and provided no benefit at all.[56]
25. Once Judge Jackson made the finding that operating systems and browsers were separate markets, the remaining liability requirements were quickly met. The integration of Windows and Internet Explorer were found to constitute coercion, as buyers were forced into the purchase of a tied product they might not want or might have purchased elsewhere on different terms. He noted that the purpose of the forcing inquiry was to expose those product bundles that raised the cost or difficulty of doing business for would-be competitors thereby depriving consumers of the opportunity to evaluate a competing product.[57] Thus, competitive foreclosure was a factor as well, notwithstanding Judge Jacksons dismissal of exclusive dealing allegations under Section 1 of the Sherman Act on the grounds that Microsofts arrangements with various firms did not foreclose enough of the relevant market to constitute a 1 violation.[58]
26. Judge Jackson concluded his finding of technological tying by commenting on the risk of error and intervention in circumstances of a genuine integration:
The Court is fully mindful of the reasons for the admonition of the D.C. Circuit in Microsoft II of the perils associated with a rigid application of the traditional separate products test to computer software design. Given the virtually infinite malleability of software code, software upgrades and new application features, such as Web browsers, would virtually always be configured so as to be capable of separate and subsequent installation by an immediate licensee or end users. A court mechanically applying a strict separate demand test could improvidently wind up condemning integrations that represent genuine improvements to software that are benign from the standpoint of consumer welfare and a competitive market. Clearly, this is not a desirable outcome.[59]
Nevertheless, he stated that he was not at liberty to extrapolate a new rule governing the tying of software products, in light of Supreme Court decisions that had spoken authoritatively on the issue of technological tying.[60] As will be seen, however, a new rule that embraces either the majority or minority Court of Appeals opinions is precisely what is necessary.
27. Jefferson Parish is the leading Supreme
Court case establishing consumer demand as the product separability test. The
case dealt with the question whether hospital surgery and anesthesiological
services represented one product or two. The Supreme Court unanimously
determined that the combination of services represented one product for antitrust
purposes. The majority accepted the traditional treatment of tying as a per se offence - one which logic and
experience dictates is anticompetitive when certain prerequisites are met.[61]
The antitrust evil noted with respect to tying doctrine is the sellers
exploitation of its control over the tying product to force the buyer into the
purchase of tied product that the buyer either did not want at all, or might
have preferred to purchase elsewhere on different terms.[62]
The use of market power becomes actionable when it impairs competition by
allowing a potentially inferior product to be insulated from competitive
pressures.[63]
The majority stated with respect to the issue of product separability that in
this case no tying arrangement can exist unless there is a sufficient demand
for the purchase of anesthesiological services separate from hospital services
to identify a distinct product market in which it is efficient to offer
anesthesiological services separately from hospital services.[64]
28. The Jefferson Parish precedent became firmly entrenched with the
subsequent Supreme Courts decision in Eastman
Kodak,[65]
which dealt with the issue of whether the delivery of repair services and parts
represented one product or two. The majority[66] held
that sufficient evidence of product separability had been provided to defeat
Kodaks motion for summary judgment. For service and parts to be considered two
distinct products, there must be sufficient consumer demand so that it is
efficient for a firm to provide service separately from parts.[67]
Sufficient evidence was found on the basis that service and parts continued to
be sold separately to self-service equipment owners, with the development of
the entire high-technology service industry providing evidence of the efficiency
of a separate market for service.[68] The
majority rejected Kodaks insistence that there could not be separate markets
for service and parts because there is no demand for parts separate from
service,[69]
stating that:
[b]y that logic, we would be forced
to conclude that there can never be separate markets, for example, for cameras
and film, computers and software, or automobiles and tires. That is an
assumption we are unwilling to make. We have often found arrangements involving
functionally linked products at least one of which is useless without the other
to be prohibited tying devices.[70]
29. As a result, Judge Jackson did find
support in the governing Supreme Court decisions regarding the test of product
separability, but the majority of the Court of Appeals that attempted to
provide guidance to him also dealt with these cases, albeit briefly. The
majority commented that, in Eastman
Kodak, the Supreme Court found parts and service to be separate products
because sufficient consumer demand existed to make separate provision
efficient. The majority expressed doubt as to whether the Supreme Court would
have subjected a self-repairing copier to the same analysis, in which
circumstances the separate markets for parts and service would not suggest that
such an innovation was really a tie-in.[71] The
point appears to be that parts and service are different in nature: one is a
physical object and the other is a pure service involving the installation of
the replacement parts. The antitrust concern is that the goods and attendant
services - separate in nature even if they are functionally linked - are joined
contractually.
30. The example of a self-repairing
copier suggests that the analysis should change if the difference is eliminated,
such that the underlying economic activity giving rise to the antitrust
concern, is unitary - precisely and exactly the same object - but exhibiting
different characteristics in a manner suggesting participation in two separate
markets.[72]
Here, the repair functionality is added to the copier which previously had been
provided separately by a repair person. The self-repairing copier example
suggests that two economic activities which are provided through separate
agencies (the copier and the repair person) which are functionally linked can
give rise to two markets, but different functionality provided through the same
economic agent - the self-repairing copier - should not give rise to a finding
of two separate markets. The majority imposes a condition on this proposition,
in that the integration must have a plausible efficiency claim and not simply
represent two products bolted together.[73]
31. Analysing the self-repairing copier
example in this manner is not without its problems. From a reductionist
viewpoint, it is possible to suggest that the instrumentality generating the
new functionality is different in some manner from that generating the
traditional functionality in which the monopoly was established. The
distinction may be as simple as the fact that different parts within the copier
give rise to the different functionality.
32. Whether the distinction between
functional integration and market/functional linkage is justified, the self-repairing copier example can
be extended by assuming that scanning functionality and laser printing
capability are integrated. The new self-repairing/scanning/laser printing
copier now includes functionality from a number of separate product categories.
To make the analogy work further, certain assumptions should be made: the
copier company has an 80% market share;[74] it
takes its scanning and laser printing products off the market;[75]
and, the new copier is sold below the price that the printer, scanner and laser
printer could be sold separately.[76] If
the consumer demand test is taken to its logical conclusion, the copier must be
taken off the market and consumers forced to buy the functionality separately.
This is due to the separate consumer demand for the scanning and laser printing
products at the time of integration, assuming that the suppliers of the
separate products continue to offer them for sale. It does not matter that
consumers could substantially benefit by having one office appliance that can
simplify the imaging of documents to be included in data bases, as well as
collating the images with various other kinds of documents or data, and then,
printing the collated group of documents.[77]
33. The argument can be made that the
efficiencies are so striking in this circumstance that even Judge Jackson might
agree that the integration should be permitted. The problem, if the consumer
demand test is determinative, is that an analysis of the nature of the
functional relationship is irrelevant, as is any consideration of the
efficiencies produced thereby.[78] If
the efficiencies are substantial, the presumption appears to be that there
would no longer be sufficient consumer demand for the products to be marketed
separately.[79]
The existence of a separate consumer demand indicates that the integration does
not yield efficiencies that outweigh the anticompetitive consequences that
exist when market power is present. Consumer demand thereby becomes a per se proxy for an efficiency analysis,[80]
the justification for which must be found in the basis of per se rules, in that logic and experience support them so
profoundly that no further judicial inquiry is necessary.[81]
34. If it is conceded that an efficiency
analysis is required, the current state of the law determining product
separability must be viewed as inadequate. If the efficiencies have to be
considered in some manner, the debate between the majority and the minority of
the Court of Appeals becomes apposite; should plausible efficiency claims be
sufficient, or must they be weighed against possible anti-competitive effects?[82]
A new rule must be recognized, as the consumer demand test is insufficient on
its own to deal with cases of technological tying.
35. The need to consider the
efficiencies in circumstances of software design is suggested by the
implications of the consumer demand test when it is taken to an extreme case.
The per se rule of technological
tying is applicable whenever a technology becomes established and enjoys
substantial market power. The nature of network effects is that particular
technologies will grow to dominate their respective software segments.
Inevitably, there will be leaders in most software categories with substantial
market power.[83]
A point will occur in the development of each one of these segment leaders
where the addition of new functionality will be vulnerable to claims of
technological tying.
36. The nature of the precedent can be
tested by its applicability to the open source movement, which has produced,
among other software products, Linux and Apache server software. The strength
of this movement is reflected in the adoption of the model by more traditional
software companies in trying to harness the power of distributed programming.[84]
The model involves placing the source code on the Internet and making it
available to programmers all over the world. If the program achieves the
requisite degree of mindshare, programmers will begin joining development teams
on a voluntary basis and developing the functionality that is the focus of the
particular development team. The reward is recognition in the development
records of the particular program as one of the contributing hackers. Eric
Raymond - the philosopher of the open source movement - has suggested that the
open source movement is a gift culture.[85] None
of the leaders receive a salary, although they become legends that are able to
parlay their fame into careers through such means as the publication of books
regarding open source programs or positions in more traditional software
companies.
37. The organizational structure is one
of a recognized leader or group that determines which patches of code developed
by the hackers are integrated into the source code of the program.[86]
Leaders exist for each sub-group, with overall control being vested in the head
of the program itself. Linus Torvalds and his team have the final decision
regarding the Linux kernel. The patches that are not integrated are still made
available to the user community through posting on open source web sites (such
as Metalabs website, http://www.ibiblio.org/pub/Linux/!INDEX.html).
38. The business model of the open
source movement is unique. The source code is distributed on the basis of an
open licensing model (the General Public License (GPL)).[87]
Internet distribution of the software program involves posting the latest
stable code, including those patches that have been incorporated into the
program, as well as posting a cutting edge buggy version for hackers.
Commercial distribution is encouraged, while Red Hat, Mandrake, Caldera
and Corel, among others, have adopted a secondary services model, in which the
vendor distributing open source software charges a fee for sales, support and
integration, but not the code itself. If the Caldera license is taken as an
example, open source code is bundled with Caldera and third party proprietary
software. The license includes the GPL with respect to the open source Linux, but
imposes terms preventing the modification or redistribution of the proprietary
software.
39. The mechanistic nature of tying
doctrine is such that leading open source solutions, such as Apache server
software, are vulnerable to a finding of technological tying, notwithstanding
the gift culture and lack of
traditional managerial control. Apache was developed by a group of webmasters
who pooled their resources to avoid undertaking the cost of developing server
software themselves.[88] As
an open source program, it achieved a 61.66% market share by August, 2000,[89]
and by September, 2000, had been growing at a faster rate than Microsoft’s
Internet‑Information‑Server. Apache is vulnerable to a finding of market
power due to its dominant position. Additionally, the question arises whether a
technological tying determination could be made. The first requirement for such
a finding is product separability, which turns on the character of demand for
the functionality in question.[90]
Critics of the open source movement suggest that it has been quite adept at
developing functionality to mimic other operating systems such as Unix, OS2 and
Windows. It is suggested that the movement may lose momentum when it has to develop
new concepts on its own in a leadership, versus follower, position. Eric
Raymond suggests that it will easily rise to the task, due to the sheer number
of hackers that can be concentrated on a particular project. Whether this
criticism is warranted or not, it can be expected that Apache programmers will
find at least some of their inspiration from existing programs on the market
and will be vulnerable to integrating functionality that previously had been
marketed separately.[91]
40. The second technological tying
requirement is that there must be evidence of actual coercion by the seller
that in fact forced the buyer to accept the tied product.[92]
The mere fact of integration of the browser into Windows provided evidence of
coercion in the Microsoft case, with no method to remove the code (such as
through the add/remove utility). With respect to Apache, as soon as a decision
is made to integrate the code functionality in question directly into the
underlying program and the stable version distributed on the Internet, the
coercion element should be satisfied, assuming for the moment that commercial
distributions automatically adopt the stable version incorporating the impugned
functionality.
41. The next requirement is that the
seller must possess sufficient economic power in the tying product market to
coerce purchaser acceptance of the tied product.[93] As
this is a strictly market share test, the Apache share of 61.66% should be
sufficient to meet the requirement, or at least confirm that Apache has the potential
to achieve the threshold market share that would satisfy this element.
42. A further requirement is that the
Court must find that there are anticompetitive effects in the tied markets.[94]
This would depend on the characteristics of the functionality that were
integrated into the Apache kernel. Taking the fate of Netscape as an archetype
example of an anticompetitive effect, it is possible that some new
revolutionary functionality might be invented by a small company and marketed
it as an application or middleware, instantly building a strong market share at
a relatively modest cost. The open source community takes significant pride in
its belief that its model can provide cutting edge functionality and can be
expected to follow evolutionary trends to incorporate new concepts. If the new
functionality is mimicked and included in the Apache Code, the anticompetitive
effects might be found in the elimination or truncation of the small company
that first developed the functionality.
43. Finally, there must be involvement
of a not insubstantial amount of commerce in the tied product market. Apaches
market position alone would satisfy this requirement, as a not insubstantial
amount can be found in market positions falling well short of monopoly.[95]
44. Critics might suggest that the open
source software movement would be immune from antitrust scrutiny because the
software code is distributed for free. However, one of the remarkable aspects
of the open source software movement is that companies are encouraged to
distribute the software at a charge, by virtue of the secondary services model
referred to above. Judge Jackson rejected the argument that Microsoft did not
sell Internet Explorer, but simply distributed it without charge. He found that
a value could be ascribed to the code, as it was incorporated into a product
that was sold, and, in the Microsoft Judgment, he instituted a discount scheme
to provide a price reduction to those who want the Windows operating system
without Internet Explorer.[96] The secondary
services model would similarly offer a commercial nexus upon which antitrust
oversight might be grounded.
45. In the end, the analysis included in
the Microsoft case can yield some seemingly perverse results. A dominant
participant in computer software market segments is not only likely but
expected. Once the requisite market share is present, the per se liability kicks in as soon as the prerequisites can be
established.
46. The
consumer demand test, and possibly the per
se treatment of tying, appear ripe for review. This is suggested in part by
the strong dissents in each of the Jefferson
Parish and Eastman Kodak cases.
In writing the minority opinion in Jefferson Parish, Justice O’Connor
stated that it is time to abandon the per
se label and refocus the inquiry on the potential benefits and adverse
economic effects that the particular tie may have.[97]
She noted that the existence of a tied product normally will not increase the
profit that a seller with market power can extract from sales of the tying product.[98]
She stated that extension of market power is unlikely to pose a threat of
economic harm unless three threshold criteria are established: the seller must
have power in the tying-product market; there must be a substantial threat that
the tying seller will acquire market power in the tied-product market; and
there must be a coherent economic basis for treating the tying and tied
products as distinct.[99] With
respect to the third criterion, there may be economic justifications for
selling the combined product even if the tied product has a use separate from
that of the tying product.[100]
However, Justice O’Connor noted that tie-ins may prove acceptable, even if all
three criteria are met, as tie-ins may have economic benefits as well as
economic harms. Therefore, according to O’Connor, these benefits should be
considered in the rule-of-reason analysis.[101]
47. In Eastman Kodak, Justice Scalia wrote the minority opinion and, while
not directly calling for the elimination of per
se treatment of tying allegations, he strongly criticized the effect of
such treatment, reflected in the majority’s application of tying doctrine to
after-market parts and service in the inter-brand market for Kodak products.[102]
The issue in the case was the tying of after sales repair services with parts.
Justice Scalia stated that summary judgment in favor of Kodak should be upheld
in circumstances of strong intra-brand competition. The tie-in question did not
permit the seller to project power over a class of consumers distinct from that
which it is already able to exploit prior to the tie-in under review.[103]
He noted that there are a number of legitimate purposes that the tie might
serve,[104]
and thus called for a consideration of the efficiencies.
48. It has been suggested that the two
product test admits a rule of reason analysis in an indirect manner.[105]
This highlights the static nature of tying analysis, in that the consumer
demand test analyses the past or present pattern of consumer demand and draws
an inference that it will remain stable in the future. We suggest that the
consumer demand test acts as a proxy for an efficiency analysis in that, if a separate
consumer demand is found to exist, the inference is drawn that distribution of
an integrated product adds no cognisable advantage over the separate
distribution of the functionalities in question.[106] The
consumer demand test acts in a per se
manner, providing the basis for an inference that the risk of injury from
distribution of an integrated product is so pronounced that no further judicial
inquiry is necessary with respect to the issue of product separability.[107]
As a result, no consideration of efficiencies is necessary. Logic and
experience do not support the inference that the efficiencies are so irrelevant
that the analysis can just stop dead in its tracks. In fact, they argue that
the efficiencies must be considered because the pattern of consumer demand is
not stable and is subject to evolutionary pressures over time.
F.
But will the Microsoft III Court
of Appeals Accept the Challenge?
49. At the time of writing, the oral
argument in the Microsoft III appeal has been completed and the judgment is
under reserve. It appears from the oral argument that the Court of Appeals is
poised to set aside the technological tying determination. While we agree with
this outcome, it appears unfortunate that the Court of Appeals may not do so on
the basis of the plausible claim test. The Court may apply the consumer demand
test, finding that there is no separate browser market. During argument, the
Court commented: [i]ts a highly questionable proposition for anyone to now suggest
that there is a separate market for a browserless operating system in part
because all operating systems include browsers and for lots of other reasons.
Its a fairly absurd proposition. The Court also stated [b]ut the basic theme,
it seems to me of Jefferson Parish and looking at separate demand is to
ask the question, and this is a reading of the test are competitive sellers
also bundling the item? And if they are, the belief is that the competition is
taking care of this and the fact that people find themselves in possession of
one makers product is therefore innocent or at least not covered by per se.[108]
50. The DOJ attempted to respond by
stating that Microsoft was the only company that enforced the tie by refusing
to allow the browser to be removed through Windows add/remove utility. The
Court of Appeals does not appear receptive to this argument, holding to the
traditional view that forcing/tying most occur at the time of sale. The Court
suggested that the DOJ was a pioneer by moving [tying doctrine] into new
territory of a peculiar situation where the defendant has said, in a sense, you
cant throw it away or were going to make it slightly hard for you to throw it
away.[109]
51. The Court of Appeals will not have
to address the plausible claim test if it does set aside the tying
determination on the ground that no separate browser market exists due to the
pattern of supply by Microsoft’s competitors. This is unfortunate in part
because the Court will make the determination on a static factor, even though
it acknowledged the markets dynamic nature and its propensity to develop serial
monopolies.[110]
The pattern of supply is subject to change in a dynamic market and so the
perpetuation of the consumer demand test, leaves the law in its current muddled
state.
53. The finding that IE and Windows98
are two different products is unsettling because it appears to adopt a static
view of the software market. The history
of software development, and particularly of operating system design, is marked
by rapid evolution and the addition of new functionality. James Allchin
testified at trial that integration of new functionality is the engine of
innovation within the software market.[111]
Often, the functionalities to be integrated are marketed separately. Judge
Jackson noted Microsoft’s position in December 1997 that nearly every new
feature incorporated into its operating systems over the last sixteen years was
once available separately.[112]
Judge Jackson and the Court of Appeals refer to the virtually infinite
malleability of software code.[113] The
software market is subject to dynamic evolution and the pattern of consumer
demand will change. In such circumstances, is it justifiable to measure
consumer demand at a single point in time?
54. The Microsoft judgment adopts a
definition of an operating system that includes software that controls the
allocation and usage of hardware resources.[114]
This is consistent with a narrow, technical definition which is not viable in
the marketplace. Andrew Tanenbaum developed the Minix operating system in the
mid-1980s as an aid for teaching software design.[115]
Minix provides a good proxy for the kind of operating system that fits the
judgment definition. It is designed to handle process management, input/output
device management, memory management and file management.[116]
Tanenbaum indicated that the operating system provided for an orderly and
controlled allocation of the processors, memories, and input/output devices
among the various programs competing for them. As an example, multiprocessing
appears to involve the simultaneous operation of two or more programs. In fact,
the processor is only able to handle one program at a time and multiprocessing
actually involves the processor switching between programs so quickly that it
gives the appearance of multiprocessing. Obviously, the operating system
regulates the manner in which this switching occurs,[117]
and does so by converting the user instructions or system calls upon the
operating system into the machine language necessary to regulate the various
components.[118]
The narrowness of this technical definition is demonstrated by the fact that it
does not include the command interpreter, or shell.
On top of the operating system is
the rest of the systems software. Here we find the command interpreter (shell),
window systems, compilers, editors and similar application-independent
programs. It is important to realize that these programs are not part of the
operating system even though they are typically supplied by the computer
manufacturer. This is a crucial, but subtle, point. The operating system is
that portion of the software that runs in kernel
mode or supervisor mode. It is
protected from user tampering by the hardware (ignoring for the moment some of
the older microprocessors that do not have hardware protection at all).
Compilers and editors run in user mode.
If a user does not like a particular compiler, he is free to write his own if
he so chooses; he is not free to write his own disk interrupt handler, which is
part of the operating system and is normally protected by hardware against
attempts by users to modify it.[119]
55. Tanenbaum excludes the Minix shell
from the definition of an operating system,[120]
even though it is technically simplistic when compared to the Windows user
interface.
56. Judge Jacksons narrow concept of an
operating system would have to exclude key elements of the Windows operating
system which have been accepted by the Court of Appeals as integral operating
system elements. The exclusion of the Windows user interface within the context
of the Microsoft Judgment is supported by the definition of middleware[121]
which includes software that operates, directly or through other software,
between an Operating System and another type of software. The Windows user
interface should be characterized as middleware under the Microsoft Judgment,
as it operates between applications and the operating system which simply
provides the basic control over the computers components. The narrow definition
included in the Microsoft Judgment would tend to prevent the kind of evolution
in operating system design that has marked the software market, assuming the
prerequisites of a tying claim are met. The question arises, for instance,
whether it makes sense that browsers will be defined as middleware for the next
ten years while voice recognition software remain an application, even though
it might soon be called upon by numerous other applications.[122]
57. It is clear that operating systems must
provide more than just the basic device drivers. This is illustrated by the
story of Minix, which occupies a special role in the history of the open source
movement in the software industry. Historically, universities used Unix for
teaching purposes because it was the open source code at the time. However,
AT&T began asserting copyright with Unix Version 7, due to its obvious
commercial potential, as well as a 1983 antitrust structural remedy that broke
up AT&T but relaxed certain business restrictions that had limited the
business segments in which it could compete. A number of universities
thereafter taught operating system theory and reduced the emphasis on the
practical aspects of operating system design through the use of hands on code.
Minix was written to fill this vacuum. Tanenbaum designed it to remain a
relatively limited program and he studiously avoided the inclusion of more
exotic functionality. It was designed, in Tanenbaums words, to fit within a
students budget and thus was a program that could be run off floppy disks.
58. After Minix was released, a Usenet
group was created that grew quickly to a 40,000 user community that soon
clamoured for the addition of new functionality. Users supplied snippets of
code to Tanenbaum, who continually refused to add new functionality
notwithstanding the interest therein. Now we reach the historical importance of
the story. Linus Torvalds used Tanenbaums textbook and the source code for
Minix, included therein, to develop Linux. He then announced the posting of the
kernel in a Minix newsgroup on August 25th, 1991.[123]
Torvalds was open to adding new functionality and quickly took advantage of the
latent interest in a full feature Minix that had built since the release of the
Minix source code. There is some dispute as to the extent to which Linux
mimicked Minix. At the time it was launched in 1991, Torvalds claimed that
Linux was free from Minix code, while Tanenbaum calls Linux a Minix clone.[124]
Nevertheless, the opportunity for Linux directly resulted from the evolutionary
pressure on Minix to develop more complex functionality and Tanenbaums stalwart
refusal to do so in order to maintain a small, compact teaching vehicle. It is
quite possible that, had Tanenbaum allowed the addition of new functionality to
the Minix code, Tanenbaum would occupy the position in popular culture that
Torvalds and Linux now enjoy.
59. A comparison of the functionality of
Linux today (which Tanenbaum describes as a feature heavy production system),[125]
and the initial release of Minix, indicates the degree of pressure on operating
systems to evolve, and the willingness of users to look elsewhere to find the
promise of this kind of evolutionary trajectory. It is interesting to note that
even Minix has now yielded to this evolutionary pressure, as it is grown to
thirty megabytes and contains more complex features, including TCP/IP
networking functionality.[126]
60. The narrow definition of an
operating system included in the Microsoft Judgment is out of step with the
evolution of Windows and Linux, and even Minix to a certain extent, which have
added a broad range of functionality beyond a simple allocation of computer
resources.
B. Logic: Charting an Evolutionary Trajectory through an Exponentially Expanding Functional Universe
61. The addition of new functionality to
existing programs represents the essence of competition within the software
market. From one standpoint, the existing user base will have no incentive to
upgrade from the legacy system unless it is given a rather substantial reason
to do so. From another standpoint, it may be argued that it is the promise of
continual upgrading through the addition of such functionality that gives a
program lasting value. Software programs must promise an evolutionary
trajectory or their user base will erode over time.[127]
62. Evolutionary pressure on software
design arises from the rapid evolution in computer components, which may be
considered supply side in nature. The most important advance in componentry is
the exponential increase in processing capacity which has held true to Moores
law, which postulated in the 1960s that chip capacity would double every 24
months.[128]
This rapid advancement in processing capacity makes new functionality possible
that prior thereto could not have existed, at least in the personal computer
space. New functionality is also made possible by advances in other components,
including the availability of such characteristics as larger RAM and hard disk
capacity, higher modem and clock speeds, and greater broadband connectivity,
amongst a myriad of other advances in componentry. As the capacity and speed of
the computer increases, new and exciting functionality becomes possible in
personal computers that used to be available only in workstation (and higher)
computers.
63. If we consider all of the
functionality that is technically available to be included in software at any
given point during the evolution of the personal computer, as occupying a kind
of spatial dimension, this functional universe[129] has
been expanding constantly since the introduction of the IBM PC. It can be
considered a multi-dimensional universe, if each computer component that
contributes to functionality represents a separate dimension. An analogy can be
found in super string theory, which postulates that the physical universe
contains more dimensions than the four which are instantly understandable.
Super string theory indicates that there are many additional dimensions curled
up in strange, non-intuitive ways, but necessary to make sense of quantum
theory.[130]
64. If we adopt this rather strange
model, we begin to glimpse the pressure on operating system designers. Any one
of these dimensions might expand, with a technological advance, such as a
breakthrough in broadband technology. We can think of this model in another
way. Evolutionary concepts have also been explained in terms of a fitness
landscape, with the topography defined by the combination of characteristics
that promote survival. In evolution and economic models, a dancing landscape
problem has been recognized when the fitness surface has a constantly changing
topography.[131]
This occurs because the factors that define the landscape are constantly
fluctuating. Whether we think of the model in terms of a constantly expanding
functional universe or a dancing landscape, software designers must identify
the functionality made possible by the technical developments and which aspects
might become core functionality that adds significant value to the operating
system platform.
65. The dynamic nature arises not only
from supply-side factors, but also from demand-side developments. A new
software program involving a relatively limited code base, can mark an
expansion of a market niche to functionality that is soon common to a large
number of applications. It may give rise to a cluster of new functionalities as
well as spur the development of complimentary hardware, the advanced
characteristics of which make possible further enhancements. Similarly, the
addition of new functionality in a competing product will likely compel its
competitors to add the functionally to remain competitive.
66. An example of a demand side pressure
is that the development of the Internet to its present level of importance was
not due to the increasing sophistication of computer components. The Internet
was first launched in 1969,[132] but
usage did not increase dramatically until after the introduction of the World
Wide Web by Tim Berners Lee while employed at CERN at Berne, Switzerland, on
May 17th, 1991.[133]
While the Web was immediately successful and the growth remarkable, the Web was
not launched on its exponential trajectory until January 23rd, 1993, when Marc
Andreesen announced the posting of the Mosaic browser on the Internet.[134]
The Mosaic browser was developed in three months by a group of five students. A
browser had been developed prior thereto by Tim Berners-Lee, but it did not
catch fire as did the Mosaic browser. One of Mosaics significant advantages was
that it added an <IMG> file capability that allowed the browser to access
image files.[135]
Netscape was developed by the same student group - with financial support
provided by Jim Clark, the founder of SGI - in a matter of a few months, and
was immediately successful, achieving six million downloads in the first four
months.[136]
The growth was exponential:
The numbers are startling. As of
April 1998, use of the Internet by Americans was doubling once every 100 days,
according to a US. Commerce Department report. In 1994, 3 million people
worldwide used the Internet. By the close of 1998, eMarketer predicts there
will be 130 million active Internet users worldwide, and 350 million by 2003.
It took radio 38 years and television 13 years to gain 50 million domestic
users. The Internet matched that figure in less than four years.[137]
67. It is not surprising that Microsoft
had to react to the development of the Web, due to its rapid growth as well as
its adoption of a different strategy - the development of the MSN proprietary
network, similar to America On-Line. The strategy made sense when connectivity
was clearly seen as the future of personal computing, but the unique and
revolutionary business strategies made possible by the Web were unknown
territory.[138]
The emergence of this latent functionality affected the evolutionary trajectory
of the computer industry - both hardware and software and expanded the
functional universe exponentially. The Internet explosion was a code
development resulting from the introduction of the browser, which became a
killer application.
68. Logic and experience does not
support per se treatment when dealing
with the question of product separability in respect of software functionality.
To the contrary, it suggests that operating systems must evolve and change such
that the pattern of demand for functionality in the future might be radically
different than the present. The complexity of innovation processes within the
software market arises in part from the recognition that there are different
sources of evolutionary pressure,[139]
from both supply and demand side factors. These different sources can interact
in complex, non-intuitive ways such that the course of innovation can reflect a
sensitive dependence on initial conditions, which is the hallmark of a
quasi-chaotic process. The efficiencies of a quasi-chaotic process cannot be
reduced to a per se inference such
that considerations of efficiency are irrelevant. Kenneth Arrow captured the
lack of predictability of complex systems in a paper contained in the inaugural
publication of the Santa Fe Institute:
The tension between chaotic
behaviour and perfect foresight was observed. Start with an equilibrium
dynamics of a standard type derived from the hypothesis that future prices are
predicted perfectly. Suppose that the solution to the difference equations
characterizing the solution exhibits chaotic behaviour. Is it realistic to
assume that the future, even though deterministic, is in fact predictable?
Clearly, part of the lessons drawn by natural scientists, especially
meteorologists, from nonlinear dynamics is precisely the opposite; chaotic
behaviour implies that small errors of observation in the starting position may
lead to virtually total unpredictability after some period of time. This
creates no difficulties of consistency when the predictor is not part of the
system being predicted. But when the predictors are the economic agents being
examined, there is a fundamental inconsistency ... The problem was dubbed the
cloudy crystal ball....[140]
69. In 1995, the only evidence that the Department of Justice relied upon in Microsoft I was an affidavit from Kenneth Arrow to support its position that no restriction upon Microsofts ability to innovate was warranted. He stated that there was no principled basis upon which the government could intervene because the difficulty, if not impossibility, in predicting the course of innovation in a complex self-organizing dynamic system.
We
are dealing with a complex system where the outcome is not easily predictable.
Indeed, predictions in the modern history of the information business have been
very poor. AT&T did not realize
the consequences of the development of the transistor, which eventually
destroyed its monopoly. IBM was
hesitant about entering the electronic computer industry altogether and failed
to understand the potential of PCs; otherwise, it would have made a very
different contract with Microsoft. Xerox
developed the basic ideas which developed into Apple and took no economic
advantage of them. This unpredictability is precisely what would be expected of
a complex self‑organizing dynamic system. But it also means that the
government is not in a position to predict, and interference to pick the winner
of this dynamic process is likely to be counterproductive.[141]
70. Professor Arrow then rejected the concept of monopoly tipping, which is part of the economic theory that underlies Judge Jacksons trial decision. He rejected the position of an amici curiae that closely resembles Judge Jacksons determination:
The amici curiae brief notes that, "once a market is 'tipped' in
favor of a particular competitor, it would take truly massive forces to return the market to a state of equilibrium
(i.e., competition)" ... There are two remarks to be made here. (1)
Clearly, competition is not a state of equilibrium or at any rate not of stable
equilibrium, as a preceding quotation on the same page makes clear. (2) "Truly
massive" forces are very likely to impose their own truly massive costs,
which have to be weighed against the gain from competition, which, under
increasing returns, is sure to be inefficient, or from "tipping" the
equilibrium in the right direction, which is usually unknowable.
To be more concrete, in this situation any set of remedies is likely to be of
the form of penalizing whatever firm happens to be leading; Microsoft in this
instance. This may take the form of disintegrating the firm horizontally or
vertically, or of imposing constraints on its ability to enter certain markets.
A rule of penalizing market successes that are not the result of
anti-competitive practices will, among other consequences, have the effect of
taxing technological improvements and is unlikely to improve welfare in the
long run. [142]
71. A
comparison of the Departments position in 1995 and 1998 marks a distinct change
in outlook. The Department appears to have rejected its earlier position that
it is difficult if not impossible to predict the dynamic process which
innovation entails. It now embraces the interventionist position which had been
espoused by the amici curiae in 1995,
including the adoption of the concept of monopoly tipping.[143]
72. The challenge for software developers
is to chart a course through a constantly changing functional universe, and
they must do so through a cloudy crystal ball. To make the process even more
difficult, it is an evolutionary process springing from different sources and
with a memory, as backwards compatibility is a highly-prized commodity. The
dynamic, quasi-chaotic nature of innovation processes renders it unsuitable to
use the per se reasoning implicit in the consumer demand test. This
evolutionary pressure suggests quite strongly that some further inquiry is
necessary.
IV. The Enigma of Weighing Dynamic
Efficiencies
73. If the issue of product integration
can only be determined on the basis of a consideration of the efficiencies in
some manner, the question arises whether the economic theory adopted by Judge
Jackson is up to the task. If the economic theory can robustly evaluate the
efficiencies implicit in competing software models and can easily weigh them
against the anticompetitive aspects of the alternative designs, the minority
efficiency calculus technological tying test may be justified. However, if the
economic theory chosen by Judge Jackson is ambivalent about the weight that
should be accorded to relative efficiencies of the computer models involved,
deference to software designers is warranted, with the remaining question being
the degree of deference that should be accorded.
74. A question arises as to what
efficiencies need be measured. The Chicago School, which has been influential in
antitrust theory, takes the position that maximizing allocative and productive
efficiencies[144]
should be the policy objective. The problem is that allocative and productive
efficiencies are largely static concepts, based on neoclassical price theory,
which does not easily deal with evolutionary pressures in a given market. Price
theory is also problematic in circumstances where Microsoft does not act like a
traditional monopolist in restricting output to raise prices and thereby enjoy
the worst of all anti-competitive evils- the quiet life.[145]
The position of the Chicago School, as explained by Richard Posner, is that the
proper purpose of the antitrust laws is to promote competition, as that term is
understood in economics.[146] He
is of the opinion that economic theory provides a firm basis for the belief
that monopoly pricing, which results when firms create an artificial scarcity
of their product and thereby drive price above its level under competition, is
inefficient.[147]
What happens when this traditional antitrust evil is not present and the goal
it pursues - static measures of efficiency - may not be appropriate in a
dynamic market segment marked by persistent evolutionary pressures? Instead of
maximizing the static efficiency at a given point in time, the antitrust goal
must be to maximize dynamic efficiency, which necessarily involves a time
element and one that requires a prescient ability to weigh the relative merits
of various market trajectories that might result through government or court
interference in the existing market process.
75. Judge Jackson adopts increasing
returns economics and its attendant concept of path-dependency to provide a
replacement antitrust evil - domination of the course of innovation, such that
the economy may be locked-in to an inferior technological standard that would
otherwise be replaced in a competitive market. The question arises whether
increasing returns economics is sufficiently robust to justify the intrusive
remedies that can result from a finding of an antitrust violation.
76. The evolutionary nature of the
software market is due to its dynamic characteristics, which make it difficult
to model from the standpoint of economic theory. Neoclassical economics, in the
form of price theory, finds its origin in the 19th Century.[148]
It is a static model in which the principle of diminishing returns plays an
important role, as it is the primary mechanism which lead supply and demand to
converge at an equilibrium point. It makes a number of simplifying assumptions
- markets are in perfect competition with perfect factor mobility, and economic
agents are perfectly rational with perfect information and complete insight
into market processes. These assumptions make it a mathematically robust
discipline, as the conduct of economic agents becomes predictable. It also is a
normative discipline and capable of developing policy prescriptions to guide
regulation or intervention.
77. The traditional model of
neoclassical economics has significant difficulty in explaining key elements
evident in the economy and common in everyday experience. For instance, it has
difficulty in explaining the way in which technological development occurs, and
until 20 years ago, the notion was that technologies came at random out of the
blue and fell from heaven in celestial blocks.[149]
Technological development was considered exogenous and governed by non-economic
forces. Another problem with neo-classical economics is its lack of ability to
explain spatial relationships within the economy. It explains how equilibria
are reached, but not which industries thrive and grow in a particular region.
Neoclassical economics also fails to explain the existence of firms, or provide
a robust explanation as to why firms grow to a particular size or vertically
integrate to include some activities but not others. Firms are considered
simply as production functions.
78. There is increasing dissatisfaction
with the state of economic thought and attempts are being made to explain some
of its shortcomings. The New Institutional Economics, which deals with the
existence of firms, markets and other economic institutions, finds its origin
in 1937 with Ronald Coase’s seminal article, The Nature of the Firm.[150] Attempts
are also being made to challenge the 19th Century mechanistic paradigm by
choosing an evolutionary metaphor, to more closely model emergent economic
processes. Another example is provided by the revival of increasing returns
economics as a method to explain economic developments and the dynamic
conditions within the information economy, particularly the software market.[151]
This initiative is related to what has become known as chaos theory - a
sensitive dependence on initial conditions in which prediction becomes
impossible because of dynamic properties.
79. The Microsoft case is at a
crossroads in economic thought because it is one instance in which these new
theories intending to revitalize economic thought, have migrated from the realm
of economic theory to policy.[152] Judge Jackson found that Microsoft
had a single-minded determination to protect an applications barrier to entry
that is the source of its monopoly power and its ability to coerce other
segments of the computer industry to do its bidding. This concept postulates
that in the software industry, network effects abound and the value of a
particular program increases with each new user thereof. This is an application
of increasing returns economics where the incremental value increases with each
successive copy of the program sold, (otherwise known in economics as an
increasing marginal utility). As sales increase, the particular technology may
be continually reinforced in a process that has been termed a positive feedback
loop and the standards underlying the technology are locked-in with the market
segment becoming path dependent thereupon. As the positive feedback loop
progresses, competing operating systems become less attractive platforms for
independent software vendors, due to the cost of writing programs or porting
(transferring) code written from Windows to the competing platform. The absence
of applications available for a competing platform to Windows - so the theory
goes - dooms them to a marginal presence in the marketplace.
80. Increasing returns economics was a
relatively obscure economic theory[153]
that has now found a new popularity and is being used to challenge the
deterministic predictability of neoclassical economics. The Santa Fe Institute,
which was founded in New Mexico in the mid-1980s, has become one of the leading
inter-disciplinary academic institutions studying this form of economics. The
self-styled Santa Fe approach considers the economy to be an emergent and
evolutionary process in which the self-organization of economic agents is
marked by limited-rationality and learning, such that the rules governing the
economy continually change. The economy is bifurcated into a traditional,
mature, low-technology sector in which neo-classical economic principles
(including the principle of diminishing returns) continue to apply, and a
high-technology sector governed by a sensitive dependence upon initial
conditions and the principle of increasing returns.[154]
By active management, it is possible to maintain technological leadership
through successive waves of new technology. Instead of one equilibrium point
which is the most efficient allocation of resources, the economy can lock into
any number of equilibria and efficiency is not assured. The quintessential
example of lock-in is the QWERTY keyboard, which was designed in the 19th
Century for mechanical keyboards and is said to be inferior to other keyboard
configurations.[155] It
remains the dominant keyboard design because it has been locked-in. In a manner
underscoring his adoption of this economic theory, Judge Jackson adopts the
concept of inflection points which represent phase transitions in the
technological landscape such that an emergent organizing principle creates a
new cluster of industrial opportunities.[156]
81. Increasing return economics gained
popularity in the 1980s as a branch of strategic trade theory, which attempted
to provide policy guidance to mature economies under pressure from
international trade. It provides encouragement for governments to take a more
interventionist role in the management of key industrial sectors and also
encourages them to be vigilant in defending these sectors from foreign
competition.[157]
Increasing returns economics is a is a positive theory that explains how a
particular market segment evolved, but it provides a relatively poor basis upon
which to develop policy rules or legal remedies. Paul Krugman states:
The rise of the economics of QWERTY
felt like an intellectual revolution to those who participated in it; phrases
like paradigm shift were used routinely. Yet
when it came to actual policy applications, the professors were cautious. There were at least three reasons for that
caution. One is that while an acknowledgement of the importance of QWERTY
refutes the near-religious faith of conservatives in free markets, it is not at
all easy to decide which direction the government should pursue. Weve already
seen how subtle the issue of strategic trade policy becomes once one tries to
deal with real-world complications. So unlike, say, the rational expectations
school, the new economic theorists did not find that their theory translated
readily into simple policy recommendations. That does not devalue the
significance of the theory: it is unreasonable to expect each intellectual
advance to be ready for immediate policy consumption. Nonetheless, the failure
of QWERTY to yield policy conclusions has been a real disappointment.[158]
82. Apart from its status as a positive
discipline, the archetypical example of increasing returns is placed in jeopardy
by an interesting analysis undertaken by S. J. Liebowitz and Stephen Margolis.[159]
They note that Paul Krugman, ... speaks glowingly of this entire literature and
that [t]he significance of the keyboard example to this literature cannot be
overstated.[160]
They note that the argument depends upon an assertion that another keyboard
design, the Dvorak keyboard, was superior to QWERTY but:
[i]n reality, research shows that
the QWERTY keyboard is about as good a design as the Dvorak keyboard and was
better than most competing designs existing in the late 1800s. Ignored in these
stories of Dvoraks superiority is a carefully controlled experiment conducted
under the auspices of the General Service Administration (GSA) in the 1950s
comparing QWERTY with Dvorak. That experiment contradicted the claims made by
Dvorak advocates and concluded that it made no sense to retrain typists on the
Dvorak keyboard. This influential study ended any serious efforts to shift from
QWERTY to Dvorak.[161]
If the archetypal example does not support the
theory of lock-in to an inferior technology, how robust is increasing returns
economics as a normative discipline? The question arises whether this form of
analysis provides a principled foundation for international strategic trade theory
or domestic antitrust policy.
83. Liebowtiz and Margolis identify other evidence that tends to undercut the concepts of monopoly tipping and pernicious lock-in. They argue that should monopoly tipping in a network market exist, the build in market share should be slow in the initial stages and then begin to accelerate more rapidly beyond the tipping point when the incumbents advantage is reversed. This was not what they found when they studied the market share trends in different software segments.
Market domination through the action of networks or other increasing returns influences should lead to some other phenomena that are strongly associated with lock-in, tipping, and inertia. But our data show that where a software product was regarded as the best one available, its market share did not build gradually, growing at an increasing rate; but rather it grew rapidly, increasing quite steadily. This is not tipping or inertia.[162]
84. They also suggest that lock-in should be reflected in higher prices once market dominance has been achieved.[163] As indicated below, Judge Jackson did not find sufficient evidence existed, permitting a conclusion that Microsoft had engaged in premium pricing strategies.
85. Another field in which increasing returns economics has been challenged is its application to antitrust innovation market analysis in the review of mergers.[164] Gilbert and Sunshine indicate that traditional merger analysis had been concerned with the effects of the combination on price in a relevant product and geographical market, and typically had not dealt with the non-price aspects of competition.[165] Innovation shared this general neglect and they state that a change in market structure resulting from a merger or acquisition may have adverse consequences for the pace of innovation.[166] Once again, dynamic efficiency is the objective of the introduction of the new analysis.
One of the more fundamental criticisms leveled at antitrust enforcement is its traditionally static orientation. Focusing most of its energy towards ensuring productive and allocative efficiency, it has often neglected dynamic efficiency. In a world of rapid technological advance, it is important that antitrust law pay greater attention to innovation issues. Adoption of the innovation markets approach in merger enforcement will help sharpen the focus on technological advancement so critical to the continued growth of our economy. By assigning innovation an important role in merger analysis, the innovation markets approach will aid antitrust authorities in adopting a more dynamic perspective.[167]
86. Antitrust authorities acknowledge that there is a lack of a deterministic relationship between research and development spending and innovation, making it difficult to link market structure and the pace of technological innovation.[168] The link between market structure and innovation is further weakened by the inability to identify without error the direction of the chain of causation ... [and the] failure to get the direction of causation right undercuts the strength of any conclusion about the relationship between industry structure and innovative efforts.[169] Notwithstanding this, Gilbert and Sunshine proceed to support the introduction of an innovation market analysis in a manner intending to parallel the mode of analysis described in the 1992 Horizontal Merger Guidelines.[170]
87. There
has been significant criticism of this form of innovation market analysis,
alleging that it is a policy without an economic theory.[171]
Critics suggest that there is no robust economic theory that provides a
sufficient connection between the innovation process and market structure. Such
a theory would permit a principled basis for determining the nature of
government intervention. Innovation market theory concentrates upon the degree
of research and development intensity pre-merger and whether any non-transitory
decrease can be predicted to occur if the merger proceeds unaffected. Richard
Rapp suggests that the utility of the innovation market approach depends upon
the validity of two statements: first, an increase in research and development
concentration is likely to reduce the amount of research and development
undertaken; and second, reducing the amount of research and development is
likely to diminish innovation.[172]
Rapp argues that there is little basis in fact or theory for either statement.
Competition can be a powerful incentive to innovate while price competition in
unconcentrated, highly competitive markets can drive innovative activity to
sub-optimal levels.[173]
Further, research and development expenditure is an input to innovation and not
innovation itself.
When we use the
term innovation market, but measure market shares in terms of research and
development (R&D) expenditures or capacity, we are making either an error
or a leap of faith. The error would be to suppose that innovation and R&D
are the same thing. The leap of faith is to believe that there is a positive
functional relationship between the rate of R&D expenditure (or the amount
of R&D capacity) and the quantum of innovation produced by firm.[174]
88. In
some cases greater concentration can reduce inefficiency and permit an
increased scale of effort.[175] In
other cases research and development competition can be an important propulsive
force and the elimination of competition may lessen the incentive to innovate.[176]
The question then arises, where such a malleable relationship exists as in the
case of innovation and market structure, as to how a system of rules can be
devised which allows court intervention in a thoughtful and effective manner.
89. Apart from the inability of
increasing returns and innovation market analysis to establish reliable policy
guidelines, the implications of these theories are that market forces in
certain sectors of the economy can no longer be relied upon to choose the most
efficient equilibrium point. This touches on an important philosophical debate
in economic theory. Adam Smith introduced the resilient concept of an invisible
hand in which enlightened self-interest benefits society at large. He was part
of the Scottish school of thought that took the position that self-organization
in unexpected forms could emerge from complex interactions within society, the
economy, and other fields of study.[177]
This concept represent a form of ontogenetic[178]
evolutionary theory in which market forces (and similar forces in other fields
of study), will inexorably produce improvement in economic (social, biological,
etc.) circumstances.[179]
Increasing returns economics challenges this kind of Panglossian thought that
unfettered market forces will produce the best of all possible worlds,[180]
and represents phylogenetic evolutionary theory,[181] in
which evolution will not necessarily have a positive result.[182]
90. The philosophical issue is thus
squarely put; can one rely on market forces to achieve beneficial results? The
policy implications for antitrust theory are rather striking. As the political
values of antitrust have receded and Chicago School analysis gained pre-eminence,
antitrust has assumed that unconstrained competitive forces will
ontogenetically produce a superior economic result, sometimes even in the face
of market power. The adoption of increasing returns economics undermines
antitrust philosophy to some extent, and opens the door to the kind of active
management recommended by the Santa Fe Institute. It also provides fertile
ground in which political values within antitrust theory may, once again,
thrive.
B. The Remarkable Mixture of Stilted
Antitrust Theory with Avant Garde Economic Theory
91. The acceptance by Judge Jackson of
increasing returns economics and its attendant concepts of path dependency and
lock-in is necessary because of the unusual characteristics of the Microsoft
case. The hallmark of monopoly abuse that justifies intervention, is the
exercise of market power by increasing price and restricting output, in a
manner that reduces consumer welfare. The problem is that no finding can be
made that Microsoft restricted output in an attempt to extort monopoly rents.
Judge Jackson found that Microsoft enjoys so much power in the market for
Intel-compatible PC operating systems that if it wished to exercise this power
solely in terms of price, it could charge a price for Windows substantially
above that which could be charged in a competitive market.[183]
92. It is notable that Microsoft is
found to have so much power in the market place that it could charge a supra-competitive price if it wished, but there is
no finding that Microsoft ever did so. The only reference found to Microsoft s
pricing practices in Judge Jackson’s decision is a 1997 Microsoft study that
indicated it could have charged $49.00 for an upgrade to Windows 98, but that
$89.00 was the revenue maximizing price, with Microsoft opting for the higher
price.[184]
Judge Jackson then adds that it is not possible with the available data to
determine with any level of confidence whether the price that a
profit-maximizing firm with monopoly power would charge for Windows 98 comports
with the price that Microsoft actually charges. Even if it could be determined
that Microsoft charges less than the profit-maximizing monopoly price, though,
that would not be probative of a lack of monopoly power, for Microsoft could be
charging what seems like a low short-term price in order to maximize its
profits in the future for reasons unrelated to underselling any incipient
competitors.[185]
As a result, there is no finding that Microsoft has abused its market power in
the traditional manner - extorting exorbitant prices at the direct expense of
consumers.
93. It is somewhat ironic that Microsoft
has been found to have engaged in anticompetitive conduct through monopoly
abuse, when a historical analysis indicates that Microsoft has lowered prices
in each market it has entered. The sense of irony is deepened when it is
recognized that the price declines increased as Microsoft built market share.
Liebowitz and Margolis conducted an analysis of the pricing and market share
trends in a number of software market segments. Their conclusions were as
follows:
We then grouped the software market
into three categories: markets where Microsoft is a direct competitor; markets
where Microsoft plays no role; and markets for products that compete with
features in Microsoft’s operating system...
The results are striking. Although
software prices in general have fallen over this period, prices for some
software have fallen far more than prices for others. In particular, in those
categories where Microsoft participates, directly or indirectly, prices have
declined by approximately 60 percent, a far more dramatic drop than the 15
percent drop in markets completely devoid of Microsoft’s influence. It is thus
difficult not to accept the conclusion that Microsoft is responsible for the
price declines.[186]
And further,
First, in markets where Microsoft
participated, prices declined faster than in markets where Microsoft did not
participate. Second, in markets where Microsoft became dominant, prices fell after
Microsoft achieved that dominance. Third, in markets where Microsoft was
dominant, it charged lower prices than it did for the same product in markets
where some other firm was dominant.[187]
94. Liebowitz and Margolis find that
Microsoft’s effect on the software markets has been to lower prices and improve
product quality.[188]
This finding underscores the impact that Microsoft had on the cost of browsers.
The Department of Justice took the position that Microsoft engaged in predatory
conduct by giving browsing technology away for free, as part of Windows,
instead of a market price of approximately $40.00 per copy. It is evidence that
cuts two ways, as Microsoft’s conduct in this regard made browsing technology
readily available to the benefit of consumers. Further, the emergence of new
business models by which profits are achievable through the sale of related
products or services, is such that the conclusion of predatory conduct on the
basis of traditional antitrust precepts, is questionable.
95. In the absence of the traditional
pricing evil, increasing returns economics and the maintenance of the
applications barrier to entry become the evil that is necessary to power the
findings of technological tying and justify the imposition of an intrusive remedy.
There is a degree of irony in the fact that, in this case, one of the last
remaining bastions of per se liability - tying doctrine - is dependent for its
validity on an economic theory which strongly condemns neoclassical economics
for its failure to recognize the complexity of the software market. The irony
is deepened by the reliance on this avant garde theory which is positive in
nature, thus allowing it to explain the development of a particular market
segment, but provides a relatively poor basis upon which to develop normative
remedies, such as the restructuring order made by Judge Jackson.
C. Coase & The Nature of the Operating System: A Surrogate Theory for Software Design
96. The apparent inability to measure
dynamic efficiencies is due in part to the absence of an economic theory that
explains what functionality should be included in a particular software
program. Increasing returns economics
does not appear to provide a satisfactory basis for explaining, from the
standpoint of software design, when a particular functionality should logically
move from an application to the operating system. It does explain why
functionality would be included from the standpoint of first mover advantages
and lock-in, but does it largely from the standpoint of game theory. It does
not appear to provide an adequate explanation of the benefits of the
integration from the standpoint of the efficiency of the software design
itself. When does the integration make sense from the standpoint of the
software platform and its user base that has made the investment in the
technology? The absence of this perspective in the economic theory underlying Microsoft III is notable if the
efficiencies of the software design must be considered in some manner.
97. The recognition that there is no
economic theory explaining the integration of operating system functionality is
based on a similar gap in neoclassical economics. In 1937, Ronald Coase
concentrated on the absence of an economic theory explaining the nature and
size of firms.[189]
Coase wanted to bridge what he saw as a gap in economic theory between the
assumption (made for some purposes) that resources are allocated by means of
the price mechanism and the assumption (made for other purposes) that this
allocation is dependent on the entrepreneur-coordinator.[190]
The question arises at what point it makes sense for another transaction to be
integrated into the firm structure, instead of leaving it to be carried out
through a market interface.[191]
Coase stated that the concept of transaction costs provides the answer.[192]
Firms will emerge to organize what would otherwise be market transactions
whenever the cost to the firm is less than the cost of carrying out the
transactions through the market.[193] Coase
continued,
[w]ithout the concept of transaction
costs, which is largely absent from current economic theory, it is my
contention that it is impossible to understand the working of the economic
system, to analyze many of its problems in a useful way, or to have a basis for
determining policy. The existence of transaction costs will lead those who wish
to trade to engage in practices which bring about a reduction of transaction
costs whenever the loss suffered in other ways from the adoption of those
practices is less than the transaction costs saved.[194]
98. His concept of transaction costs is
broader than the financial costs associated with a particular exchange
transaction, and is better described as search and information costs, bargaining
and decision costs, policing and enforcement costs.[195]
In order to carry out a market transaction it is necessary to discover who it
is that one wishes to deal with, to inform people that one wishes to deal and
on what terms, to conduct negotiations leading up to a bargain, to draw up the
contract, to undertake the inspection needed to make sure that the terms of the
contract are being observed and so on.[196]
99. In 1972, Coase lamented what he
considered to be the appalling ignorance that economists have regarding the
forces that determine the organization of industry.[197]
He noted that the study of industrial organization had become dominated by
antitrust aspects and prevented certain questions from being raised or, at any
rate, made it more difficult for them to be raised.[198]
In his view, the opinions of judges became the analytical starting point, with
economists scrambling to make sense of what they said which so tangled the
discussion that most economics were apparently unaware of having failed.[199]
One important result of this
preoccupation with the monopoly problem is that if an economist finds something
- a business practice of one sort or other - that he does not understand he
looks for a monopoly explanation. And as we are very ignorant in this field,
the number of ununderstandable practices tends to be rather large, and the
reliance on a monopoly explanation is frequent.[200]
And further,
What people do not normally do is
inquire whether it may not be the case that the practice in question is a necessary
element in bringing about a competitive situation. If this were done, I suspect
that a good deal of supposed monopoly would disappear, and competitive
conditions would be seen to be more common than is now generally believed.[201]
100. If your favourite tool is a hammer,
all of your problems will look like nails. Judge Jackson is quick to use the
hammer of per se tying doctrine, by
utilizing increasing returns economics that, by its very nature, supports
intervention. This is carried out in the absence of a theory that explains
software design. Coase concluded his 1972 article with a call for a systematic
study that would concentrate on the activities that firms undertake and the
characteristics of the groupings of activities within the firm.[202]
His proposal thirty years ago seems apposite today with respect to the Internet
economy in general and the nature of software design in particular.
101. Coases theory provides some intriguing
clues as to constructing an economic theory to explain the inclusion of
functionality within computer programs and operating systems. Coase suggested
that firms exist to manage transaction costs among the related product and
market processes that constitute the particular economic activities in which
the firm engages. Transaction costs are broadly defined to include any kind of
risk that might give rise to market failure. From this standpoint, the firm
exists to manage transaction costs and the element of risk that might exist
across a market interface.
102. We can adopt the basic elements of
Coase’s theory, and suggest that software design and the functionality subject
to integration is a reflection of risk management in a rapidly expanding functional
universe in which the evolutionary trajectory is influenced by both supply and
demand factors. The management of risk in the context of software design is
marked by the dynamic nature of the market and one that is dominated by network
effects, requiring vigilance in monitoring market trends to chart the next step
in the evolutionary trajectory. An operating system has a dual purpose, in that
it must control the computers resources and provide the base or platform upon
which the application programs can be written.[203] It
must identify functionality that will add value and lead the market by
introducing concepts to software designers and users that will add significant
value. It must also react to those market developments that identify core
functionalities that may be required by different applications. It is possible
that core functionalities are dependent on one another and integration into the
operating system is necessary for co-evolution of standards to occur in the
interfaces between these functionalities.
103. If the operating system fails to
integrate such core functionalities, a market niche is left open for a
middleware program to fill this technological gap, assuming that other programs
soon recognize the need to make system calls upon the middleware through the
applicable APIs. If it involves an emergent functionality, network effects and
path dependency can soon establish a standard which will be reinforced over
time. Co-evolution of related functionality becomes more difficult if someone
else - possibly a hostile competitor - controls the particular standard. The
loss of control over standards can weaken the value of the operating system
over time to its user base as the platform upon which applications are written
begins to fracture.
104. The ability of particular
functionality to remain as middleware depends in part on the degree of
sophistication of the functionality at issue and the complexity of the code
structure. If relatively unsophisticated with a simple code structure, the middleware
is vulnerable to integration into the operating system platform, as it lacks
resiliency. Browsers represented the key breakthrough that invigorated the
Internet and were developed by a handful of students in a few months. More
sophisticated middleware exists that is far more resilient due to its function
and complex code structure. An example is provided by The Information Bus, or
TIB messaging software marketed by Tibco. It provides a software infrastructure
designed to allow companies to become event-driven, a concept that entails a
corporation collecting and distributing information in real time. All aspects
of its operations are integrated in a manner that information is provided to
employees in personalized web pages (by the users), while information is
provided directly into analytical computer programs that automatically provide
performance results in graphical formats. The Information Bus represents
middleware that links all programs together in real time. It eliminates the
need for complex one-to-one application interfaces between the various
programs, by translating the output of each program into a common format. It
also involves a publish/subscribe communication format instead of the usual
request/reply Internet structure. In simple terms, the publish/subscribe format
allows the software to broadcast a particular message once to all those who
have subscribed to the particular information. This replaces the usual Internet
format by which the message would have to be sent to each individual user. This
reduces Internet traffic greatly, saving broadband width. The Information Bus,
represents complex middleware that appears to be more robust than a browser and
not easily duplicated.[204]
105. If we take these concepts and apply
them to the emergence of the Internet, it is understandable that Microsoft did
not initially develop browser functionality when the Internet was first
launched in 1991. The Internet had existed for more than twenty years at that
point as a market niche largely for the academic community. Microsoft followed
an independent strategy of developing MSN, its own proprietary network, a
strategy followed by different connectivity providers, including AOL, and CompuServe,
among others. The release of the Mosaic browser and the success of Netscape
when first launched, soon suggested to Gates that an inflection point in the
software market had occurred and that the Internet represented a substantial
new trend in functionality.
106. The development of the browser was a
demand side development that emerged as middleware, but was quickly recognized
by Microsoft as core functionality, to be used in a wide variety of
circumstances and by numerous applications. Co-evolution of browsing standards
with other aspects of the operating system are important, as the organization
of the user interface within Windows was changed in a manner to integrate it
more tightly with the Internet. The browser code structure was relatively
simple in nature, and the fact that it represents core functionality, provides
an explanation as to why it would be included at no apparent cost into the
operating system. The addition of the functionality provided a strong reason to
upgrade to newer versions of the operating system, and added new elements to
the evolutionary trajectory of the operating system by promising a stable
platform that enables a whole new array of applications to be developed.
107. From the standpoint of software
design, the question whether the functionality is included in the operating
system platform, and not left to be
supplied by middleware or and application (which is the equivalent of a market
process within the Coasian model), is dependent on risk management. When placed
in this light, it is not surprising that Microsoft would make the decision to
integrate browsing functionality into Windows as a logical next step in its
evolution. It is one that added value to the platform.
108. In the software market, it is the
measurement of dynamic efficiency that counts and not static efficiency taken
at one snapshot in time.[205]
Increasing returns economics involves an attempt to inject dynamic processes
into neoclassical economic theory. Innovation market analysis is specifically
designed to address dynamic efficiency concerns. Coase’s theory is also an
attempt to recognize the dynamic nature of the expansion and contraction of
firms over time.[206]
109. There is an inherent difficulty in
fashioning a legal remedy in circumstances dependent upon the measurement of
dynamic efficiencies. As indicated above, this was recognized by Professor
Kenneth Arrow at the time of the negotiation of the Consent Decree. The problem
is the identification of a methodology by which one can weigh efficiency claims
resulting from software design against potential alternative innovation
trajectories, each of which is plausible under the circumstances. It is
difficult to know what the effect will be of the integration of the particular
functionality in the future. It may represent core functionality which provides
a platform for an emergent cluster of applications. Software developers who
make the decision to undertake the integration and dedicate the resources to do
so are in a better position to anticipate the effect that such an integration
may have than is the court.
110. The challenge to develop a proper
methodology is complicated further by difficulty in weighing these dynamic
efficiencies against political antitrust values such as the preservation of
small business opportunities. Alan J Meese puts the issue this way:
Unfortunately, Traditionalists do
not explain how the proponent of a tie can prove that its benefits outweigh the
harm associated with it. Indeed, to the extent that the harm included within
the Traditional calculus includes such values as buyer freedom and
entrepreneurial opportunity, this balancing is impossible.[207]
111.
Meese
refers to a rather elegant metaphor that [t]his balancing is analogous to
deciding whether a particular rose smells as sweet as a specified ripe peach
tastes.[208]
The difficulty of weighing efficiencies against political values of buyer
freedom reflects the bias in antitrust law. In the passage quoted above, Coase
refers to the propensity of economists to assume a monopoly explanation when
they are met incomprehensible practices.[209]
Oliver Williamson refers to the same passage from Coase and continues that
Donald Turners views are representative when he stated I approach customer and
territorial restrictions not hospitably in the common law tradition, but
inhospitably in the tradition of antitrust.[210] It
is notable that even with this general predisposition, Turner has called for a
per se rule of legality in the context of tying cases. He has suggested that,
if courts choose not to adopt a rule of per se legality, they should at least
adopt a rule that the defendant should receive summary judgment if there is any
valid dispute over whether the new product is superior. Thus, under Turners
second-best rule, any credible evidence that the product is superior supports
summary judgment for defendants, even if other and allegedly weightier evidence
points the other way.
112. Commenting on Turners proposal, Commissioner Pitofsky of the Federal Trade Commission has observed that it would almost always be the case that litigation over product innovation could thus be forestalled, since [s]omeone will always write a memo saying, Sure was a terrific idea.[211] Robert Pitofsky is one of a number of proponents who argue that values other than efficiency should be considered in antitrust analysis.[212] Since the 1960s, a vigorous debate has taken place with respect to the proper objectives of antitrust law. The issue is whether antitrust law should promote efficiency/consumer welfare as its sole objective, or whether it should pursue other political values which might sacrifice efficiency. The Chicago School, which includes Bork and Posner, argue that economic efficiency/consumer welfare is the only principled objective for antitrust law. Any other goal would lead the court into a subjective weighing of political values which are vague and uncertain and which are only suitable for a legislature. The alternative school suggests that the legislative history of the antitrust statutes indicates that Congress was concerned with questions of individual liberty and decentralized decision making within the economy, expressed in their concerns for the degree of concentration of American industry along with the possible elimination of small business.[213]
113. Pitofsky, as a member of the
alternative school, has taken the position in the past that it is bad history,
bad policy, and bad law to exclude certain political values in interpreting the
antitrust laws. By political values, he means, first, a fear that excessive
concentration of economic power will breed anti-democratic political pressures,
and second, a desire to enhance individual and business freedom by reducing the
range within which private discretion by a few in the economic sphere controls
the welfare of all. A third and overriding political concern is that if the
free market sector of the economy is allowed to develop under antitrust rules
that are blind to all but economic concerns, the likely result will be an
economy so dominated by a few corporate giants that it will be impossible for
the state not to play a more intrusive role in economic affairs.[214]
114. The highwater mark of the political theories of
antitrust occurred in the 1960s, and had been steadily eroded until the Brooke Group case in 1993.[215]
The influence of the Chicago School resulted in the apparent adoption of
consumer welfare in the form of allocative efficiency as the most important
objective of antitrust analysis. The question arises whether these new theories
(increasing returns economics and innovation market analysis), which are
intended to integrate dynamic innovation concepts into economic and antitrust
analysis, unhinge antitrust theory and provide an opportunity for the
re-introduction of political values. The question also arises whether the Microsoft
case represents a turning point in this regard.
115. One of the most important aspects of Microsoft III is the finding of
technological tying on the basis of traditional per se tying principles. By the 1990s, it was generally accepted
that economic efficiency was to be paramount, resulting in the effective
elimination of predatory pricing allegations, greater acceptance of vertical
restraints and fewer per se offences
where anticompetitive intent and effect is presumed. Tying doctrine has
resisted this rationalization of antitrust theory, remaining one of the few per se offences and one where economic
efficiency arguments are generally ignored, except in indirect ways. The
product separability requirement in tying doctrine was developed in part to
introduce a rule of reason analysis through the backdoor.
116. The application of the consumer demand test, at
least in Microsoft III, is applied in
a per se manner. Once a separate
browser market is found to exist, the remaining tying requirements are quickly
found and liability imposed. The dispute that exists between the consumer
demand and plausible claim tests challenges the application of the consumer
demand test in circumstances of technological tying. Once it is conceded that
some evaluation of the efficiencies must be considered when new functionality
is integrated into the operating systems or any other computer program, the
consumer demand test is inadequate to determine whether the integration should
be considered one product or two.
117. The consumer demand test is not competent to
determine product separability in the software market, which experience and
logic demonstrate is remarkable for its dynamism and the changing pattern of
consumer demand over time. The case study of the reasons why Linux eclipsed
Minix in popular culture indicated that it was precisely because Linux adopted
an open development model, while Minix adopted a fixed and limited development
horizon for the express purpose of keeping it to the kind of device control
features that informs the operating system definition in the Microsoft
Judgment. Where the pattern of demand will inevitably change, reliance cannot
depend on a snapshot of consumer demand at an early stage of software
development. Such a static analysis cannot be considered appropriate when
innovation processes are the focus of the inquiry which must involve a
consideration of dynamic efficiencies.
118. The alternatives for undertaking an efficiency
analysis was provided by the Court of Appeals plausible claim test of the
majority and the efficiency calculus of the minority. The latter raises the
question as to the institutional competence of the court to undertake a review
of the efficiencies involved in the question of software design, which is
precisely the issue debated by the majority and the minority positions. In this
article we have attempted to identify some factors that indicate that the court
should adopt the deferential test proposed by the majority. First, we reviewed
the fragile nature of increasing returns economics that provide a relatively
poor basis on which to develop policy options or intrusive remedies. The
problem is that dynamic efficiencies are hard to model due to the number of
factors that make innovation processes quasi-chaotic in nature, with a
sensitive dependence on initial conditions. The difficulty in measuring dynamic
efficiencies suggests support for the Court of Appeals plausible claim test.
Second, we noted that the inhospitable antitrust tradition by which any
incomprehensible business practice will automatically default the analysis to a
monopoly explanation. This bias, when added to the game theoretic bias of
increasing returns analysis, makes it difficult for any market leader to avoid
an antitrust tying determination when the other prerequisites are met. Third,
we noted that any weighing of efficiency claims against anticompetitive aspects
appears to be a difficult task. How does one establish an objective measure of
the intrinsic value of a particular technological integration, and then use the
same objective measure to establish comparative values for such ephemeral
concepts as buyer freedom?
119. The mixture of archaic antitrust policy with
avant garde economic theory, provides an opportunity for the resurgence of
political values in antitrust. The courts should recognize the limits of their
institutional competence and we suggest that the majority of the Court of
Appeals was correct in limiting the inquiry to the question whether a plausible
claim for efficiency exists. The courts are incapable of undertaking a full
blown efficiency analysis with ineffective tools and a tradition of hostility
in circumstances where it is asked to weigh apples and oranges. A deferential
rule should be adopted in circumstances of technological tying allegations in
respect of software code that is so malleable in nature.
120. Our point of departure was Judge Holmes statement that the most enlightened judicial policy is to let people manage their own business in their own way, unless the ground for interference is very clear...[216] In the circumstances of Microsoft III, the consumer demand test does not provide clear grounds to justify interference, at least with respect to the issue of technological tying.
[1] For our purposes, there are three
iterations to the Microsoft dispute:
Microsoft I: United
States v. Microsoft Corp, 159 F.R.D. 318 (D.D.C. 1995), revd,
United States v. Microsoft Corp., 56 F.3d 1448 (D.C. Cir. 1995) [hereinafter Microsoft
I].
Microsoft II: United States v. Microsoft Corp., 147
F.3d 935 (D.C. Cir.1998) [hereinafter Microsoft II].
Microsoft III: United States v. Microsoft Corp., 84 F.Supp.2d 9 (D.D.C. 1999) [hereinafter Microsoft III Findings of Fact]; United States v. Microsoft Corp., 87 F.Supp.2d 30 (D.D.C. 2000) [hereinafter Microsoft III Conclusions of Law]; United States v. Microsoft Corp., 97 F.Supp.2d 59 (D.D.C. 2000) [hereinafter Microsoft III Remedy Order].
[2] Times-Picayune Publg. Co. v. United States, 345 U.S. 594, 611 (1954) (The market, as most concepts in law or
economics, cannot be measured by metes and bounds.)
[3] Charles M. Gastle, B.Comm., LL.B.,
LL.M., D.Jur., is a partner at Shibley Righton Toronto and is an associate with
The Estey Centre for Law and Economics in International Trade (www.esteycentre.com). Susan Boughs LL.B., is the Manager, Eastern
Division Legal Department, Shell Canada, with responsibility for Canadian
competition law. The views expressed herein are those of the authors alone in
their personal capacities. The authors wish to acknowledge the research and
editorial assistance of Nimali Gamage, LL.B.
[4], Dr. Miles Med. Co. v. John D. Park & Sons Co., 220 U.S. 373, 411-12 (1911)(Holmes,
J., dissenting).
[5] See infra note 118 and
accompanying text.
[6] See infra note 119 and
accompanying text.
[7] Apache Server software had a 61.66%
share of the market in August, 2000. The Netcraft Web Server Survey, at http://www.netcraft.com/survey (last visited Sept. 17,
2000).
[8] R. H Coase, Industrial Organization: A Proposal for Research, reprinted in
R. H. Coase, The Firm, the Market and the Law, 63-67 (Univ. of Chi. Press 1988).
[9] Microsoft II, 147 F.3d at 944.
[10] See id. at
944-45 (Silence at this stage would risk considerable waste of litigative
resources ... When reviewing preliminary injunctions we have generally not been
hesitant to offer interpretation and guidance on the substantive legal issues
.... We thus turn to the interpretation of s IV(E)(i), on which the merits of
the Departments case depend.)
[11] See Microsoft III Remedy Order, 97 F.Supp.2d at
59. See also Microsoft
III Conclusions of Law, 87
F.Supp.2d 30.
[12] Microsoft Corp. v. United States,
121 S.Ct. 25 (2000).
[13] FOOTNOTE TO BE REPLACED BY AUTHOR
[14] Herbert Hovenkamp, Federal
Antitrust Policy, The Law of Competition and its Practice
366 (West Publg Co.
1994).
[15] See, e.g., Eastman Kodak Co. v Image Technical Services
Inc., 504 U.S. 451, 486 (1992) (Justice Scalia explaining the basis for
the existence of per se rules: Per se
rules of antitrust illegality are reserved for those situations where logic and
experience show that the risk of injury to competition from the defendants
behaviour is so pronounced that it is needless and wasteful to conduct the
usual judicial inquiry into the balance between the behaviours pro-competitive
benefits and its anticompetitive costs.).
[16] Jack Hirshleifer, Foreword
to Stan J. Liebowitz & Stephen E. Margolis, Winners, Losers & Microsoft, at IX (The Independent Institute 1999).
[17] See IP Guidelines, reprinted in 4 Trade Reg. Rep. (CCH), para. 13,132, at 20,738 (The Guidelines for the Licensing of Intellectual Property provide that innovation markets consists of the research and development directed to particular new or improved goods or processes, and close substitutes for that research and development. The close substitutes are research and development efforts, technologies, and goods that significantly constrain the exercise of market power with respect to the relevant research and development, for example by limiting the ability and incentive of a hypothetical monopolist to retard the pace of research and development). See also Azam H. Aziz, Defining Technology and Innovation Markets: The DOJs Antitrust Guidelines for the Licensing of Intellectual Property, 24 Hofstra L.R. 475, 477 (1995).
[18] See generally Microsoft I, 56 F.3d 1448 (D.C. Cir. 1995).
[19] See Coase, supra note 8, at 33.
[20] See id. at 67.
[21] Stan J. Liebowitz & Stephen E. Margolis, Winners, Losers & Microsoft, 252 (The Independent Institute 1999)(Antitrust
law defines as predatory those actions that are inconsistent with
profit-maximizing behaviour except when they succeed in driving a competitor
out of business.).
[22] See Brian Arthur, Steven Durlauf & David Lane, The Economy as an Evolving Complex System II, A Proceedings Volume in the Santa Fe Institute Studies in the Sciences of Complexity, 1-14 (Addison-Wesley 1997). The Santa Fe
Institute, the dynamic processes evident in the software market exist on the edge
of chaos. A leading interdisciplinary institution in the concepts associated
with increasing returns economics.
[23] See Liebowitz & Margolis, supra note
21, at 10-11 (Sometimes an industry develops in such a way that monopoly is not
only a likely outcome but also a desirable one. In such industries, what we are
likely to witness is not conventional monopoly, but rather serial monopoly: one
monopoly or near monopoly after another.).
[24] Tying allegations traditionally
were based on a concept of leverage. James B.Speta suggests that the Microsoft
case has little to do with leveraging, but rather that Microsoft acted to
protect the Windows monopoly itself. James B. Speta, Tying, Essential Facilities, and Network Externalities: A Comment on
Piraino, 93 Nw. U.L. Rev 1277, 1282 (1999).
[25] Id.
[26] See Robert Bork, The Antitrust Paradox, A Policy at War
with Itself, 67 (Basic Books 1976).
[27] The Consent Decree in Section IV(E)
provided that:
Microsoft shall not enter into any
License Agreement in which the terms of that agreement are expressly or
impliedly conditioned upon:
1. 1. (i) the licensing of any other Covered Product , Operating System Software product or other product (provided , however, that this provision in and of itself shall not be construed to prohibit Microsoft from developing integrated products);...(Emphasis added)
Microsoft II, 147 F.3d at 939.
[28] See David K. Lam, Revisiting the Separate Products Issue,
108 Yale L.J. 1441, 1442 (1999).
[29] Microsoft II, 147 F.3d at
942 (The Court of Appeals released its decision on June 23, 1998,
providing guidelines for Microsoft III, which had been commenced on May 18th,
1998. It adopted the position that, [w]hen the district courts estimate of the
probability of success depends on an incorrect or mistakenly applied legal
premise, the appellate court furthers the interest of justice by providing a
ruling on the merits to the extent that the matter is ripe, though technically
the case is only at the stage of application for preliminary injunction.).
[30] Id. at 948.
[31] Id.
at 949.
[32] Id. at 948.
[33] Id. at 950.
[34] Id.
at 949-50.
[35] Id.
at 950.
[36] Id. at
950-51.
[37] Id. at 951.
[38] Id.
at 951-52.
[39]
Id. at 951.
[40] Id. at 952.
[41] Id. at 957 . The Minority states:
The majoritys test would seem to permit Microsoft to integrate
word-processing programs, spreadsheets, financial management software, and
virtually any other now-separate software product into its operating system by
identifying some minimal synergy associated with such integration. In effect
the majority has fashioned a broad exemption from the antitrust laws for
operating system design, apparently because an operating system is not like a
peripheral, whose physical existence makes it easier to identify the act of
combination.
Id. at 962.
[42] Id. at 958.
[43] Id. at 958-59.
[44] Id. at 956.
[45] Id.
at 952.
[46] Id. at
952-53. The Majority also stated, [t]he view expressed in the separate
opinion seems sure to thwart Microsofts legitimate desire to continue to integrate
products that had been separate - and hence necessarily would have been
provided in distinct markets. By its very nature integration represents a
change from a state of affairs in which products were separate, to one in which
they are no longer. By focusing on the historical fact of separate provision,
the separate opinion puts a thumb on the scale and requires Microsoft to
counterbalance with evidence courts are not equipped to evaluate. Id. at
953.
[47] See Microsoft III Conclusions of Law, 87 F.Supp.2d at 47. They include: (1), two separate products
are involved; (2), the defendant affords its customers no choice but to take
the tied product in order to obtain the tying product; (3), the arrangement
affects a substantial volume of interstate commerce; and (4), the defendant has
market power in the tying product market. Id.
[48] See Id. Shortly
after the Conclusions of Law were released, he recommended to the Department of
Justice in a case conference that a motion should be brought to certify the
case for direct submission to the Supreme Court.
[49] Id.
[50] Id.
[51]
Id.
[52] The Supreme Court decisions
relied on are Jefferson Parish Hosp.
Dist. No. 2 v. Hyde,
466 U.S. 2 (1984) [hereinafter Jefferson Parish] and Eastman Kodak Co. v. Image Technical Serv.
Inc., 504 U.S. 451 (1992) [hereinafter Eastman Kodak].
[53] See Microsoft III Conclusions of Law, 87 F.Supp.2d. at 47-48.
[54] Id. at 48-49. See also Norman W. Hawker, Consistently
Wrong: The Single Product Issue and the Tying Claims Against Microsoft, 35 Cal.W.
L. Rev. 1, 13 (1998), stating:
Nonetheless, closer examination of the Court of
Appeals integrated product test and the consumer demand test used in antitrust
law reveals that the two methods do indeed suffer from irreconcilable
differences. First, the Court of Appeals said that products may be integration
even if the two items are also marketed separately Thus the Court of Appeals
squarely refused to examine the very issue that proved essential to finding
distinct products in both Jefferson
Parish and Kodak, i.e., whether
sufficient consumer demand existed to make it efficient to market the products
separately.
Hawker indicates that both Jefferson Parish and Kodak functionally integrated products
may still be separate products for tying purposes. Not only did Jefferson
Parish state that the single product issue does not depend on the functional
relation between the items in question, the Court also found that a strong
functional relationship between two items might exacerbate the anticompetitive
effects of bundling. Id. at 15.
[55] See Microsoft III Conclusions of Law, 87 F.Supp.2d. at 49.
[56] See Microsoft III Findings of Fact, 84 F.Supp.2d. 9,
50-56 (D.D.C. 1999).
While browsing technologies may provide benefits, there is no justification for
the integration, as the IE Code could be distributed just as easily,
separately, or in a manner allowing it to be removed from Windows through the
Add/Remove facility. Paragraphs 191-92 of the findings state:
191. Therefore, Microsoft could offer consumers
all the benefits of the current Windows 98 package by distributing the products
separately and allowing OEMs or consumers themselves to combine the products if
they wished...
192 Windows 98 offer some benefits unrelated to browsing that a consumer
cannot obtain by combining Internet Explorer with Windows 95. For example,
Windows 98 includes support for new hardware technologies and data formats that
consumers may desire. While nevertheless preferring to do without Web browsing,
Microsoft has forced Windows users who do not want Internet Explorer to
nevertheless license, install, and use Internet Explorer to obtain the
unrelated benefits. ....
Id. at 56.
[57] See Microsoft III Conclusions of Law, 87 F.Supp.2d. at 50.
[58] Id. at 53.
[59] Id. at 51.
[60]
See id.
[61] The court split 5- 4 on the issue
of maintaining tying as a per se offence.
Jefferson Parish Hosp.
Dist. No. 2 v. Hyde,
466 U.S. 2, 3 (1984). Justice Stevens, writing for the majority, stated: [i]t
is far too late in the history of our antitrust jurisprudence to question the
proposition that certain tying arrangements pose an unacceptable risk of
stifling competition and, therefore, are unreasonable per se. Id. at 9. Justice Brennan, in a concurring decision, stated, [a]s
the opinion for the Court demonstrates, we have long held that tying
arrangements are subject to evaluation for per
se illegality under S. 1 of the Sherman
Act. Whatever merit the policy arguments against this longstanding
construction of the Act might have, Congress, presumably aware of our
decisions, has never changed the rule by amending the Act. In such
circumstances, our practice usually has been to stand by a settled statutory
interpretation and leave the task of modifying the statutes reach to Congress. Id.
at 32.
[62] Id. at 12.
Further, the Court stated that Per se condemnation - condemnation without
inquiry into actual market conditions - is only appropriate if the existence of
forcing is probable. Id. at 15.
[63] Id. at 14
[64] Id.
at 21-22.
[65] Eastman Kodak Co. v. Image Technical Serv. Inc. 504 U.S. 451
(1992).
[66] Id. at 453.
The court split 6-3, with Justice Blackmun writing the majority opinion and
Justice Scalia, joined by Justice OConnor, writing for the dissent.
[67] Id. at 462
(citing Jefferson Parish, 466 U.S. at 21-22).
[68] Id.
[69] Id.
at 463.
[70] Id.
(quoting Jefferson Parish, 466 U.S.
at 19).
[71] See Microsoft II, 147 F.3d 935, 950 (1998).
[72] We have found no direct authority
for this proposition. The only article that even hints at a basis on which the
distinction can be made is Thomas A. Piraino, Jr., An Antitrust Remedy for Monopoly Leveraging by Electronic Networks,
93 Nw. U. L. Rev.1, 6 (1998)(In
Windows 98, the browser has become an indivisible part of the operating system,
and no separate product exists that can be tied to the license of the operating
system. While reference to the Court of Appeals decision is made, no analysis
is provided explaining why they should be considered indivisible and one
product.).
[73] The
majority supports its position, stating that Professor Areeda argues that new
products integrating functionalities in a useful way should be considered
single products regardless of market structure. Microsoft II, 147 F.3d at
950.
[74] This assumption is made to answer the
analysis of the self-repairing copier example by Hawker, supra note 54,
at 11-12 :
The Court of Appeals is correct, of
course, that a self-repairing copier would not constitute illegal tying, but
the reason has nothing to do with whether parts and service are separate
products. The manufacturer in Kodak had
monopoly power in the aftermarket for parts, but not service. In the Court of
Appeals hypothetical, service is not being bundled with parts, but with the
copier equipment. The manufacturer in Kodak,
however, did not have monopoly power in the equipment market. Without monopoly
power in either the service or equipment market, bundling these two products
could not constitute an illegal tie-in.
The assumption of an 80% market share is to give Kodak substantial
market power in the equipment market.
[75] This is to feed the cases, and
academic support that requires monopolists to keep products on the market. The
Microsoft remedy itself has this requirement.
[76] This assumption is made to allow
allegations of predatory conduct.
[77] A motion record would be a good
example.
[78] Hawker argues at length that a
consideration of the functional relationship of the economic activities, is not
required under the consumer demand test. Hawker, supra note 54, at 15-17
(citing Digital Equipment Corp. v.
Systems Indus., Inc, 1990-1
Trade Cas. (CCH) 68,901 (D. Mass. 1990) (tying sale of patented interconnect
products to peripheral products) and Service
& Training, Inc. v. Data General Corp., 963 F.2d 680 (4th Cir. 1992)
(tying repair software to repair services)).
Hawker also argues that the Court of
Appeals wrongly associates product separateness with a lack of synergy from
bundling. He suggests that this proposition was implicitly rejected by the
majority as it was the minority that argued that the economic advantages of
joint packaging are substantial[,] the package is not appropriately viewed as
two products. Id. at 17 (citing Jefferson Parish, 466 U.S. at 40).
[79] See id. at 17 n.123.
[80] Lam rejects the majority position
in the Microsoft Court of Appeals on
the ground that it:
...conflated the definitional question of a
tie-ins existence with the question of whether a tie-in should be permissible
on economic grounds. It thereby unwisely transformed the separate products test
into a rule-of-reason test, which directs courts to examine the efficiencies
and other benefits of a challenged arrangement.
Lam, supra note 28, at 1441.
[81] Eastman Kodak, 405 U.S. at 486 (Scalia, J., dissenting).
[82] It is interesting to note that one
of the Department of Justices six affiants who filed affidavits in support of
the remedy, Carl Shapiro, raised this issue in an article published just before
the commencement of trial, Michael Katz and Carl Shapiro, Antitrust in Software Markets 49-51 (Sept 22, 1998) at http://www.haas.berkley.edu/~shapiro:
The proper statement of the policy problem is
this: Is there a rule that, when applied consistently to actual markets by real
policy enforcers, can weigh these competing claims? Such a rule is extremely
difficult to craft, as are workable remedies. Clearly, there are serious
problems with any policy that freezes the definition of an operating system. Clearly,
there are serious problems with any policy that freezes the definition of an
operating system ... (emphasis added)
Looking forward, the debate will likely turn to the question of the actual integration of code. These issues will be much more troublesome. One approach will be to let Microsoft and other software producers engage in any packaging that they want as long as the different programs are sufficiently entangled. Such a policy would do little to limit packaging. Alternatively, antitrust authorities could pursue a policy of requiring a modular approach to the production and sale of code, with well-defined open interfaces between the modules. While in some ways attractive, such an approach clearly raises a thicket of thorny questions including: Who will define the scope of the individual modules? How will openness be defined and monitoring be conducted to enforce openness on a timely basis? How will we know that important economies of scope are not being lost? And what will happen to the incentives to innovate?
[83] See Liebowitz & Margolis, supra note 23.
[84] It gained additional
momentum when the model was adopted by Netscape (Mozilla browser, Jan98), IBM
(Secure Mailer, June98), and Apple (Darwin: core lawyers of MacOSX Server,
March99), among a number of other initiatives. See Eric S. Raymond, The Cathedral & The Bazar, Musings on Linux and Open Source by an Accidental Revolutionary 75-7,
163-4, 202-3 (OReilly Books 1999).
[85] See id. at
99. Raymond compares it somewhat elegantly to the tradition of the potlach ceremonies
of the Kwakiutl tribe on the British Columbia coast.
[86] See id. at 89-90, 122-6.
[87] The General Public License
(GPL) is the most widely used copyleft license with respect to open source
software and it can be used as a baseline. It permits the use, modification,
and re-distribution of the code in unaltered or modified form and access must
be given to the source code. Any derived programs must be distributed under the
same terms as the original license. One of the only restrictions that is
permissible, is that any modification to the code must be distributed in patch
files, such that the original source code is maintained. Any company is free to
distribute the code in almost any form and to charge a fee to do so but the fee
is not to include a fee for the code itself, but the additional services or
bundling that accompanies it. See id.
at 84.
[88] See
id. at 157.
[89] Netcraft Web Server Survey, supra
note 7.
|
Developer |
July 2000 |
Percent |
August 2000 |
Percent |
Change |
|
Apache |
11412233 |
62.81 |
12222228 |
61.66 |
‑1.15 |
|
Microsoft |
3611020 |
19.87 |
3890905 |
19.63 |
‑0.24 |
|
iPlanet |
1298889 |
7.15 |
1431425 |
7.22 |
+0.07 |
[90] See Microsoft II, 87 F.Supp.2d at 47.
[91] See Raymond, supra
note 84, at 116 (arguing that [d]uplicating the functions of existing closed
software counts as highly as original work if by doing so you break open a
closed protocol or format and make that territory newly available.)
[92]
See Microsoft II, 87 F.Supp.2d at 47.
[93]
See id.
[94]
See id.
[95] See id.
[96] See id. at 53.
[97] Jefferson Parish, 466 U.S. at 35.
[98] Id. at 36.
This reflects acceptance of strong academic criticism of the leverage theory
justification of tying doctrine. This justification suggests that tying occurs
so that the party with market power can extract additional profits in a second
market.
[99] Id. at
37-9. With respect to the third criterion:
All but
the simplest products can be broken down into two or more components that are
tied together in the final sale. Unless it is to be illegal to sell cars with
engines or cameras with lenses, this analysis must be guided by some limiting
principle. For products to be treated as distinct, the tied product must, at a
minimum, be one that some consumers might wish to purchase separately without
also purchasing the tying product.
[100] See id. at
40 (noting that joint packaging may provide one such justification).
[101] See id. at
41 (providing that facilitation of entry into new fields and protection of
reputation are examples. In addition, if the tied and tying products are
functionally related, they may reduce costs through economies of joint
production and distribution, (citing Fortner Enterprises, Inc. v. United States
Steel Corp., 394 U.S. 495, 514 (1969)(White, J., dissenting))).
[102] See Eastman Kodak, 504 U.S. at 489. Justice Scalia argued that:
[t]he holding that market power can
be found on the present record causes these venerable rules of selective
proscription to extend well beyond the point where the reasoning that supports
them leaves off. Moreover, because the sort of power condemned by the Court
today is possessed by every manufacturer of durable goods with distinctive
parts, the Courts opinion threatens to release a torrent of litigation and a
flood of commercial intimidation that will do much more harm than good to
enforcement of the antitrust laws and to genuine competition.
Id.
[103] See id. at
498.
[104] See id. at
502.
[105] See. e.g., Hovenkamp, supra note 14.
[106] Areeda appears to share this
viewpoint. See Phillip E. Areeda et al., Antitrust
Law 207 (1996) (stating that [t]he single product
inquiry does not judge this policy question directly (by assessing cost
efficiencies) but indirectly with more easily obtained evidence regarding
actual market practices). See also Hawker, supra note 54, at 11
n.74 (arguing that neither Jefferson
Parish nor Kodak inquired
directly into the actual costs or quality of the items bundled versus
unbundled. Instead, the Court inferred the nature of consumer demand indirectly
from such more readily observed facts as actual consumer requests and market
practices).
[107] The point is that the analysis
stops with respect to the product separability issue. Of course, the other tying
prerequisites must be found to exist. See supra note 47 and accompanying
text.
[108] February 26th, 2001 p.m.
Appellate Transcript, at http://www.microsoft.com/presspass/trial/transcripts/default.asp
[109] Ibid.
[110] February 26th, 2001 A.M.
Appellate Transcript, Ibid.
[111] Innovation through integration is
the engine that drives the computer industry, bringing the benefits of
computing to hundreds of millions of people. Allchin direct testimony, para 41.
[112] United States vs. Microsoft Corp., 98 F.Supp. 537, 540 (D.D.C. 1997). The Court
notes that Windows 95 itself integrated the functionality of two products - DOS
6 and Windows 3.1, each of which was available separately - into its single new
operating system. Id.
[113] An example of the malleability of
software code, is provided by the operating system design concept of threads
that are used within software to permit simultaneous calls on various
functionality in multiprocessing. Mosaic and Netscape use threads in order to
place simultaneous calls to the various images on a web site in order to
download them. If threads did not exist, and each image had to be called
sequentially, the time to download a website would increase significantly. It
is a recognized principle in software design that these threads may be placed
solely in an application - user space - or in the operating system itself. See
Andrew
S.
Tanenbaum & Albert
S
Woodhull, Operating
Systems Design and Implementation 55 (Prentice Hall 1997).
[114] The definition in its entirety
provides: Operating System means the software that controls the allocation and
usage of hardware resources (such as memory, central processing unit time, disk
space, and peripheral devices) of a computer, providing a platform by exposing
APIs that applications use to call upon the operating systems underlying
software routines in order to perform functions. See Microsoft III Remedy
Order, 97 F.Supp.2d at 73.
[115] Tanenbaum & Woodhull, supra note 113.
[116] See id. at
44.
[117] See id. at
47-49.
[118] Id.
at 2-4.
[119] See id. at
3.
[120] See id. at
20-21.
[121] Middleware means software that
operates, directly or through other software, between an Operating System and another
type of software (such as an application, a server Operating System, or a
database management system) by offering services via APIs or Communications
Interfaces to such other software, and could, if ported to or interoperable
with multiple Operating Systems, enable software products written for that
Middleware to be run on multiple Operating System Products. Examples of
Middleware within the meaning of this Final Judgment include Internet browsers,
e-mail client software, multimedia viewing software, Office, and the Java
Virtual Machine. Examples of software that are not Middleware within the
meaning of this Final Judgment are disk compression and memory management. See
Microsoft III Remedy Order, 97 F.Supp.2d at 72.
[122] No definition of an application is
included in the Microsoft Judgment. A definition is provided for an Application
Business as including the development, licensing, promotion, and support of
client and server applications and Middleware ... Internet Explorer, ....
Streaming Audio and Video client and server software ... XML servers and
parsers ... Media player, voice recognition software ... See Microsoft III Remedy Order, 97 F.Supp.2d at 71.
[123] John Naughton, A
Brief History
of the Future, The Origins
of the Internet, 201-02 (Weidenfeld & Nicolson
1999).
[124] Tannenbaum & Woodhull,
supra note 113, at 15.
[125] Id.
[126] See id.
[127] In our direct experience, software
quickly loses all but its hardcore user base if the development team is lost.
Often, the development team is lost when a software firm encounters financial
difficulty or it is acquired by a firm that does not share the targets culture.
In these circumstances, the value of the software can drop to de minimis levels, when offered on in
asset sale.
[128] See Naughton, supra note 123, at 222, 224.
[129] This is similar to the concept of
genetic space referred to by Richard Dawkins:
What is the point of thinking in
terms of genetic space? Where does it get us? The answer is that it provides us
with a way to understand evolution as a gradual, cumulative process ... Every
evolutionary history consists of a particular winding trajectory through
genetic space ... If only we could draw in nine dimensions we could make each
dimension correspond to one of the nine genes. The position of a particular
animal, say the scorpion or the bat or the insect is fixed in genetic space by
the numerical value of its nine genes. Evolutionary change consists of a step
by step walk through nine-dimensional space. The amount of genetic difference
between one animal and another, and hence the time taken to evolve, and the
difficulty of evolving from one to the other, is measured as the distance in
nine-dimensional space from one to the other.
Richard Dawkins, The Blind Watchmaker 67 (Penguin
Books 1986).
[130] See Brian Greene, The Elegant Universe, Superstrings, Hidden
Dimensions, and the Quest for the Ultimate
Theory 6 (Norton 1999):
Moreover, building on special and
general relativity, string theory requires its own severe revamping of our
conceptions of space and time. For example, most of us take for granted that
our universe has three spatial dimensions. But this is not so according to
string theory, which claims that our universe has many more dimensions than
meet the eye - dimensions that are tightly curled into the folded fabric of the
cosmos. So central are these remarkable insights into the nature of space and
time that we shall use them as a guiding theme in all that follows. String
theory, in a real sense, is the story of space and time since Einstein.
[131] Geoffrey M. Hodgson, Economics and Evolution, Bringing Life Back into Economics 210 (Univ. of Mich. Press 1996).
[132] The Internet was first launched in
1969 with the establishment of the ARPANET. See Naughton, supra note 123, at 138-39. It was launched before the birth of the
personal computer with the marketing of the Altair computer in 1975. It was the
launch of this personal computer that caused Bill Gates to drop out of Harvard.
See id. at 184.
[133] See id. at
235.
[134] The Mosaic browser was the forerunner
of the Netscape and Internet Explorer browsers. Once netscape was formed by
Andreesen and Jim Clark, it had to settle a copyright dispute with NCSA, as it
was clearly based on the Mosaic browser. The fundamental code structure was
readily available to other companies as well and, in fact, was license from
Spyglass by Microsoft and a number of other companies that were struggling to
build the browser. See id. at
250.
[135] Berners-Lee criticized the addition
of the <IMG> file format when he met Andreesen at one point, suggesting
that it was not consistent with the kind of business and research orientation
that was the object of the Web. Andreesen thought that the image file format
was necessary because it was cool and would spark the growth of the Internet,
as it did. See id. at 244-45.
[136] See id. at
250-51.
[137] Vivek Ranadive, The Power of Now, How Winning Companies Sense & Respond to Change Using Real Time Technology
11 (McGraw Hill 1999).
[138] Steve Balmer reportedly did not
know of the IP/TCP stack until January 1993 when on a sales trip, his customers
were clamouring for the addition of the functionality. See id. at 254-55 Microsoft led evidence at
trial that Microsoft decided to include an IP/TCP stack in 1994 before Netscape
launched its browser.
[139] An example of different sources of
evolutionary pressure, can be found in Susan Blackmores theory of meme
evolution, as an additional source of cultural and evolutionary pressure. A
meme is an idea, behaviour, style or usage that spreads from person to person within
a culture. It is a concept developed by Richard Dawkins in 1976. Professor
Blackmore suggests that meme and biological evolution are concurrent in nature
and the combination has resulted in the advance of mankind. See Susan
Blackmore, The Power of Memes, Sci. Am., Oct. 2000, at 64. See also Susan Blackmore, The Meme Machine (Oxford Univ. Press 1999).
[140] Kenneth J. Arrow, Workshop on the Economy as an Evolving
Complex System: Summary, in Philip Anderson, Kenneth
Arrow, David Pines, The Economy as an Evolving Complex
System 280
(Addison-Wesley 1988).
[141] Affidavit of Kenneth Arrow, January
17, 1995, U.S. v. Microsoft Corp., 56
F.3d 1338 (D.C. Cir., 1995), affidavit available on Department of Justice
website re Microsoft complaint, at 4.
[142] Id.
[143] Id.
[144] Hovenkamp, supra note 14, at 61, indicates that the
Chicago School involves the following:
Economic efficiency, the pursuit of
which should be the exclusive goal of the antitrust laws, consists of two
relevant parts: productive efficiency and allocative efficiency. Productive
efficiency is a fraction in which the value of a firms output is the numerator
and the value of its inputs is the denominator; the higher this ratio, the more
efficient the firm. Gains in productive efficiency come about mainly by
research and development. Allocative efficiency refers to the general
efficiency of markets, generally measured by the Pareto criterion. (Ftn 19: A
situation is Pareto optimal when no person can be made better off without
making someone else worse off.)
[145] Liebowitz & Margolis,
supra note 23, at 15:
Weak as the lock-in argument is as an economic
foundation for an antitrust case, it is a foundation that antitrust enforces
cannot really do without. The usual monopoly concern - low output and high
price - is simply not much in evidence in the computer market...
After studying spreadsheet, word processor and
browser segments, they state:
We studied three different market patterns that
might be viewed as tests of the Microsoft-as-monopolist hypotheses. Heres what
we found. First, in markets where Microsoft participated, prices declined
faster and farther than in markets where Microsoft did not participate. Second,
in markets where Microsoft became dominant, prices fell after Microsoft
achieved that dominance. Third, in markets where Microsoft was dominant, it
charged lower prices than it did for the same product in markets where some
other firm was dominant.
In short, Microsofts effect on the software markets has been to lower prices and improve product quality. Such outcomes benefit consumers - though, of course, they hurt Microsofts competitors.
Id. at 228.
[146] Richard
Posner,
Antitrust Law, An Economic
Perspective ix (Univ.
of Chi. Press 1976).
[147] Id.
at 4.
[148] See Hodgson, supra
note 131, at 99-108. See also M. Mitchell Wardrop, Complexity, The Emerging Science at the Edge of Order and Chaos 44 (Touchstone Books 1992); W. Brian Aurthur, Increasing Returns
and Path Dependence in the Economy 2-3 (Univ. of Mich. Press1994).
[149] Wardrop, supra
note 148, at 118.
[150] See Coase, supra note 19, at 33.
[151] See Hodgson, supra note 131, at 100. See also Wardrop, supra note 148, at 44; Aurthur, supra note 148, at 2-3.
[152] Liebowitz & Margolis, supra note 21, at 285. Liebowitz & Margolis
also argue that
[m]ost notably, this is the first market where path dependence and
lock-in theory has played an explicit role in public policy. We refer, of
course, to the antitrust investigations and charges against Microsoft. Id. at
118.
[153] See, e.g., Hodgson, supra note 131, at 100; Wardrop, supra, note 148, at 44; Aurthur,
supra note 148, at 2-3.
[154] Aurthur, supra note 148, at 2-4. See also W.
Brian Arthur, Increasing Returns and the
New World of Business, 74 Harv. Bus. Rev. 101, 101-02
(1996).
[155] See Aurthur, supra note
148, at 25.
[156] See Microsoft III Findings of
Fact, 84 F.Supp.2d at 49-51.
[157] Paul Krugman, Peddling Prosperity, Economic Sense and Nonsense in the Age of Diminished Expectations 241 (Norton 1994). Paul Krugman has
recognized the potential implications for trade law theory:
The new trade theory as I have described it so far sounds fairly
entomological. It is a description of how the world economy is not a
prescription about what to do about it. ... Nonetheless, the temptation to try
to apply theory to policy is irresistible. Sooner or later, the new trade
theory was going to be given a policy spin. And it was not going to be a spin
that conservatives would like. As I have already hinted in the world of QWERTY,
one cannot trust markets to get it right. So it was inevitable that the new
trade theory would be used to provide a justification for a departure from the
principles of laissez-faire.
Id. at 239.
[158] Id.
at 243-44.
[159] S. J. Liebowitz, Should Technology Choice be a Concern of
Antitrust Policy? 9 Harv. J. L. &
Tech. 283 (1996); Stan J. Liebowitz & Stephen E. Margolis, Network Externality: An Uncommon Tragedy,
8 J. Econ. Persps 133 (1994).
[160] Liebowitz, supra note 159, at
312-13. They were referring to Krugmans comments in his book, Peddling Prosperity, Economic Sense and Nonsense in the Age of Diminished Expectations. See Krugman, supra note 154.
[161] Liebowitz, supra note 159, at 313.
[162] Liebowitz & Margolis,
supra note 21, at 239.
[163] See id. at
139.
[164] Section 7 of the Clayton Act prohibits a merger or
acquisition whose effect may be substantially to lessen competition, or tend to
create a monopoly. See Richard Gilbert & Steven Sunshine, Incorporating Dynamic Efficiency Concerns in
Merger Analysis: The Use of Innovation Markets, 63 Antitrust L. J. 569 (1995) (providing the seminal
analysis of this new approach).
[165] See id. at 571-2.
[166] Id.
at 574.
[167] See id. at 601.
[168] See id. at 579.
[169] Id.
[170] See id. at 594-95.
[171] Daniel J. Gifford, Microsoft Corporation, The Justice Department
and Antitrust Theory, 25 Sw U. L. Rev. 621, 622-23, 628 (1996).
[172] Richard T. Rapp, The Misapplication of the Innovation Market
Approach to Merger Analysis, 64 Antitrust L. J. 19, 26-27 (1995).
[173] See id.
[174] See id.
[175] See id. at 29.
[176] Id.
[177] Hodgson, supra note 131, at 56-57, 160.
[178] Ontogeny involves the development
of a particular organism from a set of given and unchanging genes. Its
environment will also affect this development, but nevertheless the growth of
the organism is the result of genetic instructions. Id. at 40.
[179] An example of the manner in which
ontogenetic theory underlies the free market economy, reference is found in Times-Picayune Publg. Co. v. United States, 345 U.S. 594, 605 (1954):
Tying arrangements, we may readily agree, flout the Sherman Acts policy
that competition rule the marts of trade. Basic to the faith that free economy
best promotes the public weal is that goods must stand the cold test of
competition; that the public, acting through the markets impersonal judgment, shall
allocate the Nations resources and thus direct the course it s economic
development will take.
[180] Hodgson, supra note 131, at 27. Voltaire had
Dr. Pangloss insist in his novel Candide:
Tis demonstrated ... that things cannot be otherwise; for since everything is
made for an end, everything is necessarily for the best end. Id. at 197.
[181] Pyhlogeny is the complete and
ongoing evolution of a population, including changes in its composition and
that of the gene pool. It involves changes in the genetic potentialities of the
population, as well as their individual phenotypic development. The distinction
between ontogenetic and phylogenetic evolution can be used to make some
interesting distinctions between differing conceptions of economic evolution. Id. at 40. Further, Given that
socioeconomic behaviour is purposeful, systems of values, visions of the
future, economic expectations, can all guide and accelerate such a process.
This is still phylogenetic evolution, because the nature of the genetic
material - the institutions and routines - is changing. Yet it is a kind of
phylogenetic evolution where purpose and drift are relatively more important
than selection. Id. at 47-48.
[182] Hodgson is of the point of view that
economics must evolve from the form of mechanistic neoclassical thought now
dominant, to one that adopts a phylogenetic standpoint. See id. generally.
[183] Microsoft III Findings of Facts,
84 F.Supp.2d at 27.
[184] See id. at 53. In his
direct testimony, Richard L. Schmalensee, Dean of the MIT Sloan School of
Management, who gave expert evidence on behalf of Microsoft, calculated the
revenue-maximizing price for Windows at more than 16 times its retail price.
Schmalensee Direct Testimony, Microsoft III Conclusions of Law, 87
F.Supp.2d 30 (D.D.C. 2000) (Trial Transcript Jan. 3, 1999). No similar analysis
was provided by the Department of Justice.
[185] Id.,
at 54.
[186] Leibowitz & Margolis, supra note 21, at 156.
[187] Id.
at 228
[188] See Id.
[189] Coase, supra note
19.
[190] See id. at 37.
[191] See id. at
45.
[192] Coase is known for introducing the
concept of transaction cost economics and although his theories were ignored
for a considerable period of time, they have now become the basis for the New
Industrial Economics. See Alan J. Meese, Tying Meets the New Institutional Economics: Farewell to the Chimera of
Forcing, 146 U. Pa. L. Rev. 1, 12 n.37 (1997).
[193] See Coase, supra note 19, at 45. Coase continues, [o]ther things
being equal, therefore, a firm will tend to be larger:
(a) the less the costs or organizing and
the slower these costs rise with an increase in the transactions organized;
(b) the less likely the entrepreneur is to make mistakes an the smaller the increase in mistakes with an increase in the transactions organized;
(c)
the greater the lowering (or the less the rise) in the supply price of
factors of production to firms of larger size.
2.