January 1999 Antitrust Remedies In High Technology Industries David

january 1999 antitrust remedies in high technology industries david a. balto & james f. mongoven introduction

January 1999
Antitrust Remedies in High Technology Industries
David A. Balto & James F. Mongoven
Introduction
Increasingly, the workload of the federal antitrust enforcement
agencies has focused on high technology markets. This should not be
surprising. Much of the growth in the U.S. economy during the last
decade has been in high technology, from computers and software to
biotechnology and pharmaceuticals. Beyond the importance of these
industries, antitrust enforcers have recently given greater attention
to the importance of high technology as an influence on growth and
innovation, and how technological questions, such as the relationship
between antitrust and intellectual property, should be evaluated.
Some have questioned the application of the antitrust laws to high
technology industries as inappropriate and ineffective. Some critics
wonder whether laws initially enacted in 1890 can be effectively
applied to industries that did not exist then and to patterns of trade
and commerce that have evolved to serve a vastly different society. At
times, this criticism is centered on the issue of remedies. The
argument posed is that high tech industries are too fast moving to
fashion effective antitrust remedies and that the abuse of market
power will more readily be dissipated by market forces, especially
ease of entry. The specific criticism concerning remedies is that the
antitrust agencies will impose orders that are too broad in an attempt
to sweep up all anticompetitive effects, resulting in a stifling of
innovation incentives. This criticism has a grain of truth to it, but
one that is used to spur a cautious approach to remedy. High
technology industries do indeed exhibit some characteristics that make
appropriate remedies more difficult to fashion, but that does not mean
that the field should be abandoned to the play of private market
forces that may be abused by anticompetitive behavior. The Federal
Trade Commission takes account of the differences in high technology
industries in crafting remedies that preserve competition without
harming unilateral and collaborative efficiencies in research and
development, manufacturing, distribution, or incentives to innovate.
This article reviews the remedies used in several recent Federal Trade
Commission and Department of Justice enforcement actions, including
mergers and nonmergers. It considers the approach to remedy in
innovation markets, networks, and vertical mergers, and addresses the
use of several alternatives, including licensing, divestiture,
firewalls, and trustees. It then addresses relief in several nonmerger
cases.
The Commission’s Remedial Authority
The Federal Trade Commission’s remedies are designed to be equitable,
not to punish.1 The relief is generally prospective, in order to stop
unlawful conduct and deter future violations of the antitrust laws.
One aspect of deterrence, of course, is removing any gain from illegal
conduct. The Commission generally pursues three types of remedies in
competition cases: structural (divestiture and licensing), conduct
(cease and desist orders), and monetary (civil penalties,
disgorgement, restitution).
When addressing the competitive problems raised by a merger, the
foremost objective of the Bureau of Competition (“Bureau”) is to
provide relief that will return competition to the status quo ante.
The Bureau recognizes, as do the courts, that relief from a Section 7
violation should be tailored to alleviate the likely anticompetitive
effects in the relevant market. Thus, while it is true that the
Commission has “wide discretion in its choice of a remedy,”2 it seeks
to assure that its remedy is reasonably related to the unlawful
practices.
The Bureau is sensitive to the need not to upset transactions that may
enhance efficiency if effective remedies may be achieved through
settlement. Crafting an effective remedy depends, in part, on a
careful analysis of market structure, especially entry barriers. It
also depends on what kind of relief would be most effective in
restoring competition. To be an acceptable alternative to litigation,
a settlement must resolve the competitive concerns uncovered during
the investigations. That may require, for example in the case of a
merger, that the settlement provide that the divested assets will
create a viable competitor with market share that can replace the
competition lost by the merger, or, at least, facilitate entry (or
expansion) of a successful competitor.
Over the past decade the Bureau has faced several challenges in
devising remedies in high technology cases. In the past, divestitures
of physical assets in areas of competitive overlap have been the
principal form of remedy. In high technology markets, that form of
remedy sometimes appeared overinclusive and, on occasion, inadequate.
Thus, the Bureau has sought a wide range of other remedies, including
licensing arrangements.3 In some cases, these arrangements must be
supplemented with a variety of assets and continuing relationships in
order to assure that effective technology transfers take place.
As in traditional industries, it is helpful to divide the remedy
discussion in high technology industries into merger and nonmerger
sections.
Mergers
Since mergers are a structural event, the remedies to cure the
anticompetitive part of an acquisition are typically, and preferably,
structural. If the competitive overlap constitutes most or all of the
acquisition or merger in question, the Commission will often seek a
preliminary injunction to halt the transaction (and the likely harm to
consumers) before it happens; if the competitive overlap is a smaller
part of the overall deal, divestiture or licensing, if they can be
drafted with sufficient protections, may suffice to remedy the
competitive concern.
The merger remedies favored by the Commission in high technology
acquisitions are as innovative as the markets they protect.
When examining a merger that has both potential efficiencies and
anticompetitive effects, the Commission will ordinarily attempt to
find a structural solution that will preserve the former and eliminate
the latter. As Justice Brennan recognized over thirty years ago, the
“key to the whole question of antitrust remedy is of course the
discovery of measures necessary to preserve competition.”4 That
usually requires a remedy as substantial as divestiture. As the
Supreme Court acknowledged in Du Pont, divestiture is the “most
drastic” yet “most effective” form of remedy. The Court also advised
that divestiture could not be denied the government “because economic
hardship, however severe, may result.”5 Thus, the Court concluded that
divestiture is a “natural remedy” for a violation of Section 7 and
always should be “in the forefront of a court’s mind when a violation
. . . has been found.”6
Divestiture of a manufacturing plant or other physical assets is
something that the courts are familiar with and that the antitrust
agencies are comfortable doing in order to preserve competition,
particularly where the assets constitute an ongoing business.7 If the
competing assets can be sold to a third party or spun off into a
viable competitor, competition can be maintained with a one-time
event. The Commission is most comfortable with an asset divestiture
when it can identify an acceptable buyer in advance. If that cannot be
done, the Commission order will maintain the status quo during the
divestiture process by requiring a tight deadline for the sale, a
trustee to oversee the assets to be divested, and the use of a crown
jewel provision to ensure that the divestiture incentive is strong.
Where much of the value of a firm resides in its intellectual property
and the ability to innovate, however, divestiture may be somewhat more
problematic. The complex interplay of human capital from which
innovation springs is much more fragile than typical physical assets.
In addition, intellectual property is often embodied in patents, and
the patent system has its own complex system of legal protections.
Thus, in some cases involving high technology markets the Commission
will consider licensing as an alternative. Since the output of the
innovation process can be consumed by more than one entity, the
licensing of that output has the potential to replace lost competition
without physically restructuring the merging firms. However, the
Commission has always been concerned with licensing as a remedy.8 It
is by nature more regulatory than divestiture, and may require ongoing
oversight to ensure effectiveness.9
This section reviews merger remedies in innovation markets, computer
markets, network mergers, and vertical mergers. It then discusses two
special toots—firewalls and trustees—and closes with a practical road
map to merger remedies.
Innovation Markets
The Commission’s “innovation market” merger cases provide good
examples of its thinking about remedies in high technology industries
where only certain parts of a merger would have probable
anticompetitive effects. In these cases, the Commission alleged that
competition for the innovation of new products would be harmed by the
acquisitions. Remedies were needed that could replace the competition
lost in the merger without blocking any procompetitive benefits. Both
divestiture and licensing were used in appropriate circumstances.
Although divestiture is the preferred remedy in all merger cases, as a
remedy in innovation markets it requires special care because the
success of research and development efforts often depends on a complex
array of expertise and sustained knowledge. Sometimes, even in cases
where divestiture is the appropriate remedy, it may be necessary to
require ongoing obligations of the divesting party to assure that the
purchaser has some probability of successful completion of the
research effort.
The appropriateness of divestiture as a remedy, and its case-by-case
flexibility, are illustrated by the Glaxo case, where the merging
parties were the two firms farthest along in developing non-injectable
agonists, which are oral drugs used to treat migraine attacks.10
Although injectable drugs were already approved by the FDA, they were
not sufficiently substitutable to be included in the relevant market.
Both Glaxo and the acquired firm, Wellcome, competed to develop the
new drugs, and barriers to entry, based on the necessity to acquire
substantial specialized human capital resources, and the necessity of
completing the FDA approval process, were high. No other firm besides
the two merging firms was close to producing a non-injectable agonist.
The Commission thus faced a highly concentrated market for the
research and development of an important new drug where the merging
parties were each other’s closest competitors. The complaint alleged
that after the merger Glaxo could unilaterally reduce output in the
relevant market by decreasing the number of research and development
efforts to develop a non-injectable drug. It would have the incentive
to do so because the remaining research effort would presumably
produce a monopoly product until the third-best effort could complete
the FDA approval process some years hence.
Thus, divestiture of Wellcome’s non-injectable R&D assets was
appropriate but not sufficient because of the difficulty of competing
in this relevant market. In order to fully restore competition, the
order had to make certain that the buyer of the divested assets could
effectively use them to mount a strong research and development
effort. The assets themselves might not have been sufficient, without
the “complex array of expertise and sustained knowledge” embodied in
human capital, to enable an acquirer to get to market expeditiously.11
Thus, the order in this case required not only the divestiture of
Wellcome’s non-injectable R&D assets, it also imposed significant
obligations on Glaxo to assist the acquirer in its efforts to continue
the research and development effort successfully. Glaxo had to provide
information, technical assistance, and advice to the acquirer about
the R&D efforts, including consultation with, and training by, Glaxo
employees knowledgeable about the project.12 Additionally, under
certain conditions Glaxo had to produce more of the experimental drug
for the acquirer if it was unable to manufacture sufficient quantities
on its own. A trustee was appointed and given the power to sell either
the Wellcome or the Glaxo non-injectable assets if Glaxo had not
divested the Wellcome assets within nine months. The divestiture in
this case was a success: both Glaxo and the acquirer of its
intellectual property now have oral migraine drugs on the market. With
the required assistance from Glaxo, the acquiring firm, Zeneca,
received complete FDA approval in only fifteen months.
In the Ciba-Geigy/Sandoz case, the Commission alleged a market for the
development of gene therapy products, despite the fact that there were
no such current products licensed by the FDA.13 This $63 billion
acquisition was the largest pharmaceutical merger in history,
combining two Swiss firms that were the leading developers of gene
therapy products. The technology at issue involved the treatment of
disease through manipulation of genetic material and insertion or
reinsertion into a patient’s cells. Although there were many firms
doing pioneering research into gene therapies for various disease
states, the merging firms were two of only a few entities with the
intellectual property rights and other assets necessary for
commercialization of such therapies. The firms’ combined position in
gene therapy research was so dominant that other firms doing research
in this area needed to enter into joint ventures or contract with
either Ciba-Geigy or Sandoz in order to have any hope of
commercializing their own research efforts.
Although divestiture is the preferred remedy
in all merger cases, as a remedy in innovation markets it requires
special care . . . .
The remedy in the Ciba-Geigy/Sandoz case was designed to protect
competition both in the particular products then being researched and
the broader market for gene therapy research and development. For the
identifiable products under development, the order required the
licensing of certain key intellectual property rights held by the
combined firm, and also required that an acceptable buyer be
identified “up front.” Rhone Poulenc Rorer was identified as the
licensee before the order was accepted by the Commission. For the
broader gene therapy research and development market, the order
required the companies to grant to all gene therapy researchers who
applied non-exclusive licenses to all essential gene therapy
technologies, along with access to drug master files and safety data
filed with the FDA. Because adequate human capital was present in the
identified buyer, as well as in other companies with R&D programs in
gene therapy, ongoing technical assistance was unnecessary in this
case.
Although not usually ordered in merger cases, licensing was deemed
necessary here to restore partnering prospects for other firms lost by
the merger.14 Commissioner Azcuenaga dissented as to the licensing
aspect of this order, noting that divestiture would cure the
anticompetitive problem in a “simple, complete, and easily reviewable”
manner.15 This is the crux of the decision to use licensing or
divestiture to cure an anticompetitive merger. While divestiture is
certainly an easier remedy to impose and monitor, it may not always be
the most effective way of restoring competition. Because licensing is
more flexible and can more easily be tailored to unusual fact
situations, it may be the preferred remedy in innovation cases where
divestiture could interrupt potentially successful research efforts.
In this case, the majority of the Commission determined that the gene
therapy research efforts, which contained a number of joint efforts
with third parties, would be too difficult to disentangle from the
merging firms, and would thus “not only . . . hamper efficiency but
also could be less effective in restoring competition if it led to
coordinated interaction or left the divested business at the mercy of
the merged firm.”16
The Ciba-Geigy/Sandoz case and the settlement reached there highlight
both the procompetitive and the anticompetitive potential of
technology licensing. These agreements can increase competition if the
licensing is used to create additional competitors. An older
Commission case, however, shows the potential for competitive abuse
when licensing agreements become too restrictive.
Yamaha, a Japanese firm, and Brunswick, an American company, were two
of the leading outboard motor manufacturers in the world.17 In the
1970s, the two firms formed a joint venture that effectively divided
the world markets into exclusive territories so that they would not
compete with each other. Yamaha held exclusive rights to sales in
Japan, Brunswick in the United States and Australia, and the joint
venture would sell in other countries under the joint venture
trademark. Included in the joint venture agreements were technology
licensing arrangements that led to the exchange of patents and other
intellectual property regarding marine engines. The licensing
agreements were restricted in that each company was forbidden to use
the technology to make or sell products that would compete with the
products of the other. Where, as here, the technology licensing
agreements were nakedly anticompetitive, the only solution is to
dissolve them. The Commission’s order dissolved both the joint venture
and its attendant licensing agreements.
Sometimes technology licensing cannot solve all of the competitive
concerns in a merger. A joint venture between Shell Oil and
Montedison, an Italian chemical company, to produce and market
polypropylene was such a case.18 The relevant markets, in which both
Shell and the joint venture participated, involved polypropylene
technology and technology licensing. Both parties were willing to
attempt to alleviate the Commission’s competitive concerns by
excluding their technology licensing businesses from the joint
venture. Based on the investigation, the Commission determined that
exclusion from the joint venture of the parents’ respective technology
licensing businesses did not solve antitrust concerns for two
principal reasons: first, the shared interest created by the joint
venture might temper competition between the parent companies in
research and development and in technology licensing; second,
meaningful innovation in polypropylene technology was dependent on,
and inextricable from, firsthand experience in using the technology
and the proprietary catalyst to produce and sell polypropylene resin
and in working with customers to develop new resin formulations. The
joint venture would in the future be the principal means by which
Shell, and the only means by which Montedison, would produce and sell
polypropylene resin. Under these circumstances, it was simply
implausible that Shell and Montedison would continue to compete with
each other, let alone with their joint venture, in innovation in
polypropylene technology.
The Commission’s staff rejected a licensing-only remedy in that case
because the evidence showed that a viable and competitive technology
business could not be maintained without production, sales, and
distribution assets for both polypropylene resin and polypropylene
catalyst. Thus, comprehensive structural relief was necessary in order
to preserve continued competition between the two companies in
innovation and in licensing of polypropylene technology and to assure
the viability and competitiveness of the technology licensing
businesses. Solutions short of divestiture that would leave the
technology licensor and its licensees dependent on the
Shell-Montedison joint venture for supply of suitable raw materials
and for information regarding new product formulations were rejected.19
Accordingly, the Commission’s order required divestiture by Shell of
an integrated polypropylene company sufficient to support continued
research and development in polypropylene technology and to be a
credible licensor of technology.
Computer Mergers
The first industry most people think of when high tech is mentioned is
the computer industry, and indeed, both the hardware and software
components of the industry epitomize many of the characteristics that
apply across the board to high tech industries. The technology of
these products may change with unusual rapidity. Network effects are
prevalent in both hardware and software because computers are
essentially products designed to facilitate communication, and these
effects can result in high barriers to entry. The Commission has
recently investigated a number of mergers in this industry that have
raised significant competitive concerns.
One of the Commission’s more prominent merger investigations during
the last year involved the settlement of Digital Equipment’s patent
litigation with Intel, which resulted in cross-licensing of technology
and the acquisition of a fabrication facility.20 Digital and Intel are
aggressive rivals in the present and future development of
microprocessors, and Digital’s Alpha microprocessor is a significant
competitor both to Intel’s Pentium microprocessor and to Intel’s next
generation IA-64 microprocessor. Digital sued Intel in May 1997,
alleging that Intel’s Pentium microprocessors infringed some of
Digital’s Alpha microprocessor patents. Intel countersued, claiming
that Digital’s Alpha microprocessors infringed some of Intel’s
patents. Digital and Intel settled their suit in October 1997 by
agreeing to broad patent cross-licenses, the sale of Digital’s
microprocessor production facilities to Intel, and an agreement that
Intel would produce Alpha microprocessors for Digital. The sale of
facilities to Intel did not include the intellectual property rights,
design assets, or employees for the Alpha chip. Digital also agreed to
design computer systems based on Intel’s IA-64 architecture and chips.
The Commission’s complaint alleged that certain aspects of the
settlement were likely to reduce competition in three relevant
markets: (1) the manufacture and sale of high-performance, general
purpose microprocessors capable of running Windows NT in native mode;
(2) the manufacture and sale of all general purpose microprocessors;
and (3) the design and development of future generations of
high-performance, general purpose microprocessors.21 Antitrust
concerns arose because in each of these markets Digital’s Alpha chips
were the highest performance and most technologically advanced threat
facing Intel’s own microprocessors. Barriers to entry were significant
because any new microprocessor firm would need to convince computer
manufacturers to design their systems around the new microprocessor, a
daunting task given the network externalities embodied in current
microprocessors. While recognizing that the settlement would free
Digital from the enormous cost of owning and operating its own chip
fabrication facilities, the Commission was nonetheless concerned that
Alpha would not remain competitively viable under the original terms
of the Digital-Intel agreement. Intel would have the ability to
interfere with Digital’s supply of Alpha chips by withholding
necessary cooperation at the manufacturing stage. Digital might also
lack the ability and incentive to continue to actively develop and
promote Alpha.
To resolve these concerns, Digital entered into a consent agreement
under which it would grant perpetual and generally non-terminable
licenses of the Alpha architecture to Samsung and AMD or other
suitable partners approved by the Commission so that they would be
able to produce and develop Alpha chips. This is an “architectural”
license under which the licensee will be free to create its own
implementations and derivative works—that is, to design original chips
around the architecture—with the caveat that the licensee maintain
backward compatibility with the Alpha architecture. In that fashion,
the licensees will have the incentives to develop their own variations
of the Alpha chip and thus restore innovation competition. An
architectural license is important for two reasons. First, the
licensee can become completely independent of the licensor by
developing its own product line. Second, customers may be more willing
to commit to the Alpha architecture knowing there are alternative
sources of supply. The order also requires Digital to provide
technical support to the licensees for a period of up to two years.
Digital also agreed to begin the process of certifying IBM as a
foundry for Alpha chips, thereby establishing a manufacturing
alternative to Intel. Structural relief was not necessary in this case
because the strength, experience, and market position of the licensees
made them highly formidable potential competitors. Taken together,
these provisions were intended to preserve Alpha as a viable product
and competitor to Intel’s microprocessors.
In Adobe, the Commission faced a merger to monopoly of two software
firms.22 Both Adobe Systems and the acquired firm, Aldus, designed
professional illustration software. These products are used by
graphics professionals to efficiently and reliably create and print
high-quality illustrations, a $60 million market. The Adobe product
and the Aldus product were the only two professional illustration
products available for use on Apple Macintosh and Power Macintosh
computers, which are particularly favored by professional
illustrators. Even if the market were expanded to include
IBM-compatible computers, Adobe and Aldus would have 35 percent of the
market and would be each other’s closest competitors. No new entry has
occurred and barriers are high due to high developmental and
reputational barriers, and a large installed base that makes
penetration difficult and time consuming. In particular, the network
externalities associated with both firms’ installed user base posed a
particularly substantial barrier to entry.
The consent agreement settling this case required Adobe to divest
Aldus’s “Freehand” professional illustration software within six
months. The parties presented and the Commission approved Altsys
Corporation as an up-front buyer, which had developed the Freehand
software. The divestiture package included customer names and
addresses, as well as marketing, advertising, training, and technical
support information and materials. The merged firm was also required
to maintain Freehand’s viability and marketability in the interim. The
Commission also required a ten-year prior approval agreement from the
merged firm before acquiring any interest in any firm engaged in the
development or sale of professional illustration software for the
Macintosh or Power Macintosh, or before directly acquiring any such
software if the purchase price was $2 million or more.
As noted above, network effects can be a particularly thorny barrier
to entry. If a new entrant’s product is not compatible with the
dominant network, consumers will be unlikely to switch unless the new
product offers a huge advantage in price or product attributes.
Additionally, compatibility can be difficult to achieve due to
intellectual property rights held by the incumbent network owner. The
Commission recently had a merger case in which the owner of a product
with dominant network effects attempted to purchase a firm that was on
the verge of entry with a competitive and compatible product,
Autodesk’s acquisition of Softdesk.23 Autodesk had 70 percent of the
market and 70 percent of the installed base for computer-aided design
(“CAD”) software engines for Windows-based computers. These products
are used in the architecture, engineering, and construction industries
to automate the design process, and are used by over 1.4 million
consumers. The large installed base of Autodesk users necessitates
that any new CAD engine developed and offered in the market offer file
compatibility and transferability with Autodesk’s product in order to
be an effective competitor. Users of the Autodesk product have a large
number of drawings in the Autodesk format, and many users must share
files they create with others who must be able to read and edit those
files using the CAD software. This situation creates barriers to entry
to CAD engines that cannot read the Autodesk files without losing data
or information. A smaller firm, Softdesk was about to introduce its
own CAD engine, IntelliCADD, that would have provided substantial
direct competition to Autodesk because it offered compatibility and
transferability with Autodesk-generated files and application
software.
In this case, when apprised by the Commission staff of the competitive
implications of the merger, Softdesk divested its product to a third
party, Boomerang Technology. The order settling the case barred
Autodesk or Softdesk from reacquiring IntelliCADD or any entity that
owns or controls the IntelliCADD technology for a period of ten years
without prior notification to the Commission. In addition, Autodesk
and Softdesk were prohibited from interfering with the ability of
Boomerang to recruit or hire employees of Softdesk who worked on the
development of IntelliCADD.
Both the Adobe/Aldus and the Autodesk/Softdesk acquisitions suggest
the importance of structural relief where network externalities are
present. In those cases, even though technology may have been
evolving, network effects created such a formidable barrier that
remedy short of divestiture may have been inadequate.
Network Mergers
The antitrust enforcement agencies have been increasingly vigilant in
investigating and seeking relief when appropriate in network mergers.
Perhaps a decade ago, most such mergers received little attention, in
part because it was difficult to perceive significant barriers to
entry. Now as the understanding of network effects, including network
externalities, installed base, and switching costs, has become more
sophisticated, network mergers are receiving more careful scrutiny.24
The antitrust enforcement agencies have been increasingly vigilant in
investigating and seeking relief when appropriate in network mergers.
For example, in 1995 the FTC challenged the acquisition of the Western
Union consumer money transfer system (owned by First Financial
Management Corp.) by First Data Corp., the owner of the MoneyGram
system.25 Consumer wire money transfer systems involve one-way money
transfers, typically between two consumers. Wire transfer agents
include a wide variety of retail outlets, including grocery stores and
check cashing outlets. Historically, entry into the consumer wire
transfer service market was difficult for two reasons: (1) the need to
develop a minimum viable scale nationwide network of money transfer
agents; and (2) the need to establish name recognition and customer
acceptance through large-scale advertising and promotion. Long-term
agent contracts utilized by Western Union made acquiring a sufficient
agent network difficult, and building brand name recognition required
a substantial investment over a number of years. First Data
(MoneyGram) was the only firm that had succeeded in overcoming those
barriers and had begun to offer significant competition to Western
Union, until then a monopolist. No other firm was likely to enter.
To address these concerns the consent order required First Data to
divest the MoneyGram network, including its trademark and
relationships with a sufficient number of geographically dispersed
agents (over 10,000) to create a viable network. Since the agent base
and trademark were critical factors, the order had two novel
provisions. First, First Data could not solicit any MoneyGram agents
until the term of their contract expired. Second, during the
divestiture period, an interim trustee was appointed to ensure the
network did not deteriorate in value. Specifically, the trustee had
the ability to increase the level of MoneyGram advertising by 20
percent, and there were financial incentives to sign up additional
MoneyGram agents to build the agent base.
Another recent network merger involved the acquisition of most of the
assets of AutoInfo by Automated Data Processing (“ADP”) in the market
for a used auto parts exchange network.26 Auto exchange networks
facilitate the trade and purchase of auto parts between salvage yards.
They consist of several components, including (1) a yard management
system, a computerized inventory-control and management system for
salvage yards; (2) an interchange, a numeric system for identifying
interchangeable used automobile parts that is included in electronic
form within the AutoInfo yard management system; and (3) a
communications network, an electronic, satellite, or land-based
communications system to effectuate trading of parts between salvage
yards. The ADP/AutoInfo transaction resulted in a monopoly in the
relevant market.
The merger was challenged after its consummation, in part because of
withholding of relevant Hart-Scott-Rodino information by ADP.27 The
antitrust case was ultimately settled after administrative litigation
had begun. The consent order required the divestiture of the AutoInfo
assets, including the communications network and the yard management
system. There was a problem with a simple divestiture, however,
because ADP had switched all AutoInfo customers to use the ADP
interchange after the acquisition and failed to update the AutoInfo
interchange, which quickly became obsolete. In effect the ADP
interchange had become the industry standard since ADP’s conversion of
all users to its own interchange rendered impracticable any attempt to
ask users of the assets to be divested to change back to the AutoInfo
Interchange.
Thus, the Commission faced a problem: how to provide access to the ADP
interchange to the acquirer of the AutoInfo assets while providing
incentives for them to develop their own interchange? The answer was a
requirement that ADP provide a paid-up, perpetual, non-exclusive
license (with no continuing royalties and with unlimited rights to
sub-license) to ADP’s interchange. This license arrangement required
no post-divestiture royalties and did not limit the licensee’s right
to sublicense. The order also required ADP to license all interchange
updates. Moreover, the order provided that ADP had to give the
acquirer a copy and non-exclusive license to all computer programs,
databases, and information sources that ADP has used to update its
interchange.
As in the First Data case, the Commission recognized that the
divesting network could take actions to drive customers from the new
network, and therefore implemented provisions to protect the customer
base. ADP could have prevented customers from dealing with the new
network or create barriers to interoperability between networks. To
address that issue the order contained an open access provision that
prevented ADP from reinforcing its installed base advantage by
preventing its customers from accepting or transmitting inventory data
by way of other vendors’ services. The order also provided that ADP
could not use any contracts or other means to hinder its customers’
attempts to interconnect with the acquirer’s products or services.
This provision ensured that ADP’s customers could freely choose to
send inventory data generated from ADP systems or sent over ADP
communications networks to other vendors’ systems or networks.
Moreover, to facilitate the transition of ADP customers to the new
network the order provided that, for one year following divestiture,
ADP facilitate conversion of any customers (acquired post-acquisition)
to the acquirer’s system by foregoing any penalties otherwise
available and by permitting customers to exit existing contracts.
Vertical Mergers
Vertical merger cases can also have anticompetitive effects, although
they may raise different issues from the horizontal cases already
mentioned.28 The main potential anticompetitive effect from vertical
mergers is that market opportunities may be foreclosed and barriers to
entry may be raised by the necessity of entering at more than one
level in order to complete effectively. The Commission recently
settled two vertical merger cases in high technology industries. In
both of these cases, the acquisitions exhibited potential efficiencies
along with the probable anticompetitive effects from potential
foreclosure. Thus, the challenge in each case was to tailor the relief
to restore pre-acquisition competition while allowing the efficiencies
to be realized.
In Silicon Graphics, the acquiring firm had 90 percent of the market
for workstations that run entertainment graphics software and the
acquired firms, Alias Research and Wavefront Technologies, were two of
the three leading entertainment graphic software firms in the world.29
The software is used to produce high-resolution two- and
three-dimensional digital images for movies and other media. The two
interdependent acquisitions totaled nearly $500 million. The complaint
alleged that the acquisitions would foreclose access by other
workstation producers to the relevant software as well as give Silicon
Graphics nonpublic information about its competitors’ products. In
addition, the acquisition could foreclose, or increase costs to,
potential entertainment software firms, raising barriers to entry by
making it necessary to enter on two levels. The complaint also alleged
a reduction in innovation in both hardware and software markets.
The consent order required Silicon Graphics to enter into a “porting
agreement” with a Commission-approved partner that would allow the
partner to run the two most important software programs of Alias. This
porting requirement would ensure that at least one other workstation
manufacturer would have access to the important Alias graphics
packages being acquired by Silicon Graphics. The settlement identified
Digital Equipment, Hewlett-Packard, IBM, or Sun Microsystems as
possible porting agreement partners, as long as the agreement itself
was approved by the Commission. The Commission also required Silicon
Graphics to maintain an open architecture and to publish its
application programming interfaces so that other software developers
could write entertainment graphics software for Silicon Graphics
workstations. Silicon Graphics was also required to offer independent
entertainment graphics software companies participation in its
software development programs on terms no less favorable than it
offers other types of software companies. To address the potential for
disclosure of competitors’ information, the order prohibited the
release of non-public information from the platform partner porting
the Alias software to those Silicon Graphics or Alias employees not
participating in the porting process.
In the other merger, the Commission alleged that Cadence Design
System’s acquisition of Cooper & Chyan Technology would reduce
competition for software used to automate the design of integrated
circuits.30 Cadence is the leading supplier of microchip layout
environments, and Cooper & Chyan was the only firm with a commercially
viable constraint-driven, shape-based router, which is used to
automate the solution of unique engineering problems associated with
the increasingly smaller scale of microchip design. Cadence thus stood
in a vertical relationship with Cooper & Chyan as a purchaser of its
products, and the merger could lead to efficiencies due to tighter
integration between Cooper & Chyan’s integrated circuit design tools
and Cadence’s product line, which would allow the tighter coupling of
physical design software with logical design tools that is necessary
to design integrated circuits at the submicron level. The complaint
alleged that the acquisition would reduce the incentives of Cadence to
permit competing suppliers of routing tools to obtain access to
Cadence’s layout environments, thereby increasing barriers to entry
for routing tool developers because they would have to simultaneously
enter at the router and layout environment levels. Innovation would
suffer as a result. The evidence indicated that Cadence was no
stranger to attempts to block access to its layout environments in
order to raise costs to potential competitors. In the past, Cadence
had thwarted attempts by firms with potentially competitive technology
to develop interfaces.
The consent order required Cadence to allow developers of commercial
integrated circuit routing tools to participate in the Cadence
“Connections Program” and any other Cadence independent software
interface programs that enable independent software developers to
develop and sell interfaces to Cadence layout tools and environments.
Cadence was required to offer participation to independent software
developers on terms no less favorable than those applicable to any
other participants in the program. Cadence’s Connections Program had
over 100 participants, and thus the nondiscrimination requirement was
easy to enforce and it allowed actual and potential competitors of
Cooper & Chyan products to have equal access to Cadence’s layout
environment.
The key aspect of the relief in both of these cases was the use of
nondiscrimination provisions aimed at preventing self-favoritism.
These provisions are typically not easy to draft or monitor. However,
in both the Silicon Graphics and Cadence cases, there were available
benchmarks that could be used to judge how the companies would have
treated rivals if the market were more competitive. Thus, in Silicon
Graphics, the respondent must treat independent entertainment graphics
software developers in the same way it treats independent developers
of industrial and scientific software applications doing business in
more competitive markets. Similarly, in Cadence, the respondent must
allow router developers to participate in existing software
development interface programs available to developers of other
software tools. As Howard Morse has observed, in both cases “the
application of an existing benchmark resulted in an order that is
easier to monitor compared to other nondiscrimination provisions.”31
In other cases, where these benchmarks are unavailable, stronger
remedy alternatives may be necessary.
Firewalls
Vertical merger cases with both potential efficiencies and
anticompetitive problems may present an additional remedial
difficulty. If the merger is allowed to proceed, the combined firm may
be in a position to gain access to proprietary information of a
competitor through a supply relationship with that company. One form
of relief to obviate this problem is the so-called “firewall,” an
order provision requiring some segregation of information within the
merged entity. The Commission has taken this approach in several
vertical merger cases in defense industries32 and pharmaceuticals.33
Firewalls may be appropriate when, because of its position in two
levels of the market, a newly vertically integrated firm may be both a
competitor and a customer or supplier of a particular firm or group of
firms. In its activities as customer or supplier, the merged firm may,
in the nature of things, acquire access to competitively sensitive
information about these competitors. If this information flows freely
from the customer/supplier segment of the integrated firm to the
competitor segment, the non-integrated competitor may be harmed in
several ways. The integrated firm may be able to appropriate the
fruits of the competitor’s research and development and decipher
product road maps. If the integrated competitor gets access to a
non-integrated competitor’s technical information, the latter’s
incentives to innovate or engage in research and development may be
sharply reduced because the rewards for such investment may be
appropriated by the integrated firm. Further, in a market where firms
bid for business, the integrated firm also may be able to forecast, or
at least approximate, the competitor’s bids, and itself bid less
aggressively than it would have in a state of greater uncertainty.
One example of a firewall in a defense-industry case is found in
Martin Marietta/General Dynamics, a 1994 consent order. That case
involved the acquisition by Martin—now Lockheed Martin—of General
Dynamics’ Space Systems Division. The Commission’s complaint alleged
that as a result of the acquisition, Martin’s launch-vehicle division
would have access to detailed, non-public information from competitors
in the satellite industry, which it could then pass on to its
satellite division. That is because launch vehicle designers must have
a lot of information about the satellites their vehicles are to carry.
This information could be misappropriated by Martin to make better
bids for satellite business or even to build better satellites. The
consent permitted the acquisition, but prohibits Martin’s launch
division from disclosing to its satellite division any non-public
information it receives from a satellite manufacturer.
Vertical merger cases with both potential efficiencies and
anticompetitive problems may present an additional remedial
difficulty.
The first use of a firewall provision in the pharmaceuticals industry
was a consent in Eli Lilly and Company/PCS Health Systems, in 1995.
That case involved Lilly’s acquisition of McKesson Corporation’s PCS
Health Systems, a pharmacy benefit management (“PBM”) business. PCS,
in carrying out its PBM activities, maintains a “drug formulary,” a
list of drugs it gives to pharmacies, physicians, and third-party
payors to guide them in prescribing and dispensing prescriptions to
health plan beneficiaries, and it negotiates with drug manufacturers
about rebates, discounts, and prices to be paid for drugs. Obviously,
this involves obtaining a great deal of proprietary information from
manufacturers that compete with Lilly, and the concern was that, in
the absence of a firewall provision, such information could be shared
with Lilly. The consent order prohibited such information-sharing.34
The Commission’s experience with these cases indicates that, if the
troublesome information flow can be prevented, the problems of
misappropriation of competitively sensitive information can be
avoided.35 In this regard, competitors serve as an extra set of eyes
for the Commission, and can alert it if there has been a breach of the
firewalls. These firewalls are rather straightforward and they are a
“clean” solution, in the sense that they should not have any negative
impact on legitimate potential efficiencies. Firewalls appear to be
effective and seem not to be creating unplanned negative effects.
Nonetheless, firewalls are not appropriate to every situation in which
disclosure of competitively sensitive information is potentially
anticompetitive. A firewall provision may be adequate where there is
some neutral third-party oversight, for instance the Department of
Defense or third-party payors like insurance companies. In other
cases, a firewall will be far less than sufficient.
One might suggest that firewalls are unnecessary because private
parties can negotiate for similar agreements. But private parties
might not be able to accomplish the same level of protection with
private non-disclosure agreements. There also may be some instances in
which the incentives of parties to negotiate private confidentiality
agreements would not be fully aligned with the Commission’s concerns
about the health of competition.
Trustees
One important element in some of the remedies in high technology
markets is the use of a trustee. Trustees have been used for a number
of purposes. For example, in First Data,36 an interim trustee was used
to ensure that the value of the MoneyGram network did not dissipate
during the divestiture period. The trustee was responsible for
monitoring advertising expenditures and solicitation of agents. In
Glaxo,37 a trustee was used to ensure that an adequate package of
intellectual property was divested, that the research and development
continued to be pursued in an appropriate fashion, and that the
acquiring party could manage the regulatory drug approval process. In
Digital,38 a trustee was used to monitor both the process of
finalizing the implementing agreements for the licenses and the
continuing relationships between Digital and the licensees.
The Commission’s increasing use of trustees in recent high technology
mergers is based on several recurring complexities in high technology
industries. Where the Commission does not possess the technical
expertise to draft and monitor effective licenses, an experienced
trustee may be able to do so. Many trustees are experts with hands-on
experience in certain industries and technologies and can effectively
determine whether a particular remedy is adequate to restore
competition. Ongoing relationships between the respondent and the
acquirer may provide the opportunity for strategic behavior by the
respondent to thwart the acquirer’s ability to compete effectively.
The trustee can monitor the transfer of know-how and other agreements
to deter strategic behavior. Also, the constant presence of a trustee
may be necessary in order to ensure that research and development
continues to be pursued and to facilitate the seamless exchange of
information necessary to contract for a license. Finally, Commission
orders that contain licensing provisions usually contain “crown jewel”
provisions requiring the divestiture of a more valuable set of assets
if problems arise with the license. Trustees are available to enforce
these provisions and activate the crown jewel clause, if necessary.
A Practical Road Map to Merger Remedies
The merger remedies favored by the Commission in high technology
acquisitions are as innovative as the markets they protect.39 While
each remedy is different, molded to the precise competitive situation
in the affected markets, there are certain key factors that are
considered:
# Divestiture of an Ongoing Business
Divestiture of an ongoing business with customer and supplier
relationships with an up-front buyer or under a strict timetable
remains the preferred remedy; it is quick and has a high probability
of success.
# Divestiture of Selected Assets
Where it can be determined that a divestiture of selected assets is
adequate, with the right acquirer, to facilitate entry, the Bureau
will consider this as a possibly less burdensome requirement. In high
technology markets, such selected assets may include patents and
technology portfolios, and research and development assets.
# Licensing
If divestiture appears to be inappropriate for whatever reason,
licensing will be considered. Generally, the Bureau prefers exclusive
licenses, so that the licensee has the greatest economic stake in the
technology being transferred. In addition, the Bureau will generally
prefer that the licensee obtain the right to use any improvements to
the technology that the licensee develops, including uses of the
technology in fields outside the alleged innovation product market.
Often, licensing must be supplemented with other assets such as supply
agreements for an interim period, critical personnel, facilities, or
customer relationships.
# Use of an Up-Front Buyer
Presentation of an up-front buyer is always useful, regardless of the
form of remedy. Where the divestiture is less than an ongoing business
or where licensing is considered, it is critical for the respondent to
present an up-front buyer or licensee during consent negotiations. An
up-front buyer or licensee will help to assure that the remedy is
adequate to restore competition. The Bureau will always attempt to
determine if the up-front buyer will be able to effectively utilize
the divested assets or the proffered license. This gives the Bureau
more confidence that competition will be restored for the long run.
# Crown Jewel Provision
Also, the Bureau will often insist on a “crown jewel” provision, with
enforcement power in the trustee, to assure that the respondent will
make every effort to resolve the remedy issue in a timely and
effective fashion. A crown jewel provision requires a company to sell
a different, more valuable group of assets if the assets to be
divested are not sold in a given time period. Such provisions increase
the incentive for the respondent to sell the initial package during
the initial period and provide a more attractive package for the
trustee to divest if the respondent is unsuccessful. A crown jewel
will be particularly important where licensing is used; it creates
incentives for the respondent to cooperate fully in technology
transfers and supply agreements and may dampen the incentives to
engage in strategic behavior.
# Interim Trustees
Very often, the Bureau will require the appointment of an
auditor-trustee to oversee the respondent’s efforts in transferring
technology and performing under any supply agreement. The orders will
also frequently contain a hold-separate and an interim maintenance
requirement, to make certain that the active research and development
continues on schedule and that the assets are not dissipated during
transfer. In such instances, an auditor-trustee can prevent the
dissipation of assets and information-exchange problems that might
occur.
Nonmergers
When the Commission considers the need for effective remedies for
anticompetitive conduct in high technology industries, it faces a
broad set of issues. In part, this merely reflects the greater number
of merger cases, but it also highlights the wider range of conduct
under review in nonmerger cases. While all mergers are alike in that
they are a melding of separately owned assets at a particular point in
time, potentially anticompetitive nonmerger conduct is more diverse,
ranging from per se horizontal division of markets to monopolization
to vertical distribution agreements analyzed under the rule of reason.
In addition, divestiture and other structural remedies may be more
difficult to fashion in a situation where assets are not already being
combined or spun-off. To be sure, conduct remedies such as injunctive
relief and cease and desist orders are also important remedies in
nonmerger cases, but even these are more difficult to analyze and put
in place than a simple prohibition of an acquisition. Finally,
nonmerger conduct might call for a third type of remedy that is not
often appropriate in mergers, that is, monetary relief. Monetary
remedies might include civil penalties, restitution, or perhaps even
disgorgement.40
Several of the perceived differences in high technology competition
may have an impact on the type of remedies sought in nonmerger cases.
For instance, the prevalence of network effects may raise issues of
access to the network by competitors or by producers of complementary
goods. In such a case, the owner of a proprietary system who committed
anticompetitive or exclusionary acts could be forced to open the
system in order to restore competitive conditions.
The abuse of intellectual property licensing in order to maintain
monopoly power can also give rise to an antitrust violation. In the
Pilkington case, the Department of Justice alleged that the British
firm, the world’s largest producer of float glass, violated Sections 1
and 2 of the Sherman Act by maintaining agreements and understandings
that unreasonably restrained interstate and foreign trade in the
construction and operation of float glass plants and in float glass
process technology, and by monopolizing the world market for the
design and construction of float glass plants.41 The agreements in
question were put in place in the 1960s, but the patents and trade
secrets they were initially based-on had long since expired and had
moved into the public domain. Despite the erosion of the intellectual
property protections, over 90 percent of float glass worldwide
continued to be manufactured under a Pilkington license agreement. The
complaint specifically alleged that, without sufficiently valuable
intellectual property rights, Pilkington and other float glass
manufacturers allocated territories for, and limited the use of, float
glass technology worldwide, required competitors to prove that all of
the licensed technology had become publicly known before being
relieved of the territorial and use restrictions, and imposed
limitations on the sublicensing of float glass technology. As a result
of the restrictions, existing licensees, including those in the United
States, could not design and build new float plants without
Pilkington’s permission.
The consent decree entered in Pilkington enjoined defendants from
enforcing license provisions that restrain their U.S.-based licensees’
freedom to use float glass technology anywhere in the world, and from
enforcing license restrictions against their other licensees that
restrain the licensees’ freedom to use float glass technology in the
United States. It also enjoined defendants from asserting any
proprietary know-how rights in such technology against individuals or
firms in the United States who are not licensees.
The settlement freed competitors worldwide to pursue the best
technology and to innovate to produce new technology. The Department
estimated that thirty to fifty float glass plants were planned or
projected worldwide out to the year 2000, amounting to expenditures of
as much as $5 billion. Once defendants were enjoined from imposing the
licensing restrictions, firms in the United States could compete for
contracts to design and manufacture these plants. The judgment also
enjoined conduct that had the effect of restricting exports of float
glass and float glass technology into the United States.
Three recent high technology nonmerger cases brought by the FTC show
the diversity of remedies being sought. In the Intel case, which is
currently in administrative litigation, the Commission alleges that
Intel refused to provide to three companies technical information
necessary to make products that would work with Intel microprocessors,
in an attempt to force those companies to license their own technology
to Intel.42 The relief sought by the Commission would order Intel to
cease and desist from discriminating, or threatening to discriminate,
between customers that compete with it or who threaten to assert
intellectual property rights concerning computer technology against
it, and those that do not, unless Intel can demonstrate legitimate
business reasons for such discrimination. Discrimination would be
forbidden in the areas of the sale of microprocessors, terms of
nondisclosure agreements concerning microprocessors, providing
information concerning Intel’s computer technology, providing
prototypes of Intel’s products in development, and providing technical
assistance concerning Intel’s existing dominant products or products
in development.
The VISX case also involved abuse of market power based on
intellectual property.43 In that case, Summit Technology and VISX, the
only two FDA-approved manufacturers of lasers used in photo refractive
keratectomy (“PRK”) to treat vision disorders, formed a patent pool
whereby the two firms agreed to charge a $250 licensing fee that was
paid into the pool each time laser eye surgery was performed using
either firm’s equipment. The proceeds of the pool were split according
to a formula. The result was that prices were far higher than they
would have been if the two firms had been competing with each other,
in markets for the sale or lease of PRK equipment, and the licensing
of technology related to PRK. Both companies possessed patents that
would have allowed them to be horizontal competitors, and they had
planned to so act before the pool was formed.
The pool effectively fixed the minimum price for the technology
licenses under which doctors perform PRK. The pool also had the power
to exclude competitors. Any manufacturer that wanted to market a laser
for PRK needed a license from the pool for one of the patented laser
aiming mechanisms. The pool eliminated the option of receiving bids
from two competing firms to license the technology and it also
permitted either Summit or VISX to veto any potential licensee. The
veto power was exercised and third-party entry was prevented.
In this case, a cease and desist order and licensing of the
intellectual property of both companies were sufficient to restore
competitive conditions. Under the terms of the settlement of the
patent pool charge, the two firms are prohibited from fixing prices or
agreeing in any way to restrict each other’s sales or licensing of
their patents or lasers, including prohibition of the “per-procedure
fee” charged to doctors for use of the lasers. Since the companies had
already agreed to dissolve the patent pool, the order requires them to
take no action inconsistent with the dissolution, and also requires
them to license each other, on a royalty-free and non-exclusive basis,
the patents that each contributed to the pool. This settlement
embodies the principle found in the Intellectual Property Guidelines
that a patent pool that evidences anticompetitive effects may be
defended if it is reasonably necessary to achieve procompetitive
efficiencies.44 Here, the Commission found that the pool was
anticompetitive and not necessary to achieve efficiencies.
The Dell Computer case involved an anticompetitive acquisition of
market power through abuse of a standard-setting procedure.45 At issue
was a standard designed for the Video Electronics Standards
Association for a local bus to transfer instructions between a
computer’s CPU and peripherals. There would be considerable
efficiency-enhancing potential in a product that would let computer
and peripheral manufacturers know how to make products compatible with
one another. The agreement on the standard was premised on
representations by the participants that no firm would assert
intellectual property rights that might block others from developing
towards the standard. The anticompetitive potential of the
standard-setting activity surfaced when Dell alleged that the new
standard infringed on its patent, despite twice certifying, along with
other members of the Association, that it would not assert
intellectual property rights. Dell made its claim only after the bus
was highly successful, and its claim for royalties gave it effective
control of the standard.
Dell’s belated assertion of patent ownership in this case enabled it
to exercise market power that went beyond power that it could have
obtained in the absence of its misrepresentation. The Commission’s
complaint specifically alleged that industry acceptance of the new
standard was delayed and that uncertainty about the acceptance of the
design standard raised the cost of implementing the new design. Other
firms avoided using the new bus because they were concerned that the
patent dispute would reduce its acceptance as an industry standard. In
addition, willingness to participate in industry standard-setting
efforts was chilled. When Dell asserted ownership of a blocking
patent, this anticompetitive use of intellectual property was
sufficient to cause harm through suppression of innovation and delay
of introduction of the products that the standard was designed to
implement.
Dell Computer was a relatively straightforward case, and the resulting
remedy was also simple and easy to effectuate. The consent order
requires that Dell refrain from enforcing its patent against any
computer manufacturer using the new design in its products. In
addition, Dell is prohibited from enforcing any of its patent rights
that it intentionally fails to disclose upon request of any
standard-setting organization during a standard-setting process. Other
competitors are thus free to innovate and produce products to the
specifications of the standard, and Dell is precluded from duplicating
its anticompetitive behavior in the future.
Conclusion
Effective antitrust enforcement depends on a thoughtful and determined
approach to remedies.
Effective antitrust enforcement depends on a thoughtful and determined
approach to remedies. The unique innovation incentives and
intellectual property aspects of high technology industries have
challenged the Commission to craft remedies that work effectively in
difficult and unfamiliar situations. The same remedies of divestiture
and licensing that are used in traditional “smokestack” industries are
also used in high technology industries, but they must be applied with
unusual precision in order not to adversely affect the innovation and
new product development that are the hallmarks of competition for
these rapidly evolving industries. The Commission’s efforts to apply
intelligent and flexible remedies in high technology industries have
helped protect competitive markets and ensure open access to those
markets for new entrants, and thus have the ultimate effect of
increasing innovation incentives across all industries.
N o t e s
1. The Commission’s remedial authority is derived from the Federal
Trade Commission Act, 15 U.S.C. §§ 41-58, the Clayton Act, 15 U.S.C. §
§ 12-21, and more than 30 more specialized statutes, e.g.,
Magnuson-Moss Warranty Act, 15 U.S.C. § 2301, Fair Packaging and
Labeling Act, 15 U.S.C. §§ 1451-61.
2. FTC v. Rubberoid Co., 343 U.S. 470, 473 (1952).
3. By requiring licensing, the Commission can create a new competitor
or enhance the probable success of an existing competitor without
weakening the respondent by removing assets or intellectual property
that may be necessary to successfully compete in the relevant market.
4. United States v. E. I. du Pont de Nemours & Co., 366 U.S. 316, 326
(1960).
5. Id. at 327.
6. Id. at 331. See Ford Motor Co. v. United States, 405 U.S. 562, 573
(1972) (divestiture is “particularly appropriate” in merger cases);
Olin Corp. v. FTC, 986 F.2d 1295 (9th Cir. 1993), cert. denied, 510
U.S. 1110 (1994); RSR Corp v. FTC, 602 F.2d 1317, 1325-26 (9th Cir.
1979), cert. denied, 445 U.S. 927 (1980); Ash Grove Cement Co. v. FTC,
577 F.2d 1368, 1379-80 (9th Cir. 1978).
7. See, e.g., RSR Corp. v. FTC, 602 F.2d 1317 (9th Cir. 1979), cert.
denied, 445 U.S. 927 (1980) (partial divestiture sufficient to restore
competitive entity); OKC Corp. v. FTC, 455 F.2d 1159, 1161 (10th Cir.
1972) (total divestiture of acquired assets necessary to restore
“viable, independent, local competitive entity”).
8. Although it may present difficulties, sometimes the divestiture of
intellectual property or innovation efforts is necessary to fully
restore competition. In these cases, the Commission will not hesitate
to order divestiture. See, for example, the discussion of Glaxo and
Montedison, infra.
9. Some commentators have criticized the adequacy of licensing as an
antitrust remedy. See Richard T. Rapp, The Misapplication of the
Innovation Market Approach to Merger Analysis, 64 Antitrust L.J. 19
(1995). For a response, see Thomas N. Dahdouh & James F. Mongoven, The
Shape of Things to Come: Innovation Markets in Merger Cases, 64
Antitrust L.J. 405, 438 (1996) (“Licensing is usually found to be an
appropriate remedy when market participants and innovators agree that
access to intellectual property is key”).
10. Glaxo plc, No. C-3586 (FTC June 14, 1995) (consent order).
11. Dahdouh & Mongoven, supra note 9, at 439.
12. See III P. Areeda & H. Hovenkamp, Antitrust Law ¶ 707i, at 184
(1996) (advocating “divestiture of sufficient assets to create viable
new firms with free access to the monopolist’s then-existing
technology . . . where an acquisition, or a series of acquisitions,
has probably made a substantial contribution to monopoly power”).
13. Novartis A.G., No. C-3725 (FTC Apr. 8, 1997) (consent order)
(Commissioner Azcuenaga concurring in part and dissenting in part).
14. For a discussion of the Commission’s use of licensing as a remedy,
see Mary Lou Steptoe & David A. Balto, Finding the Right Prescription:
The FTC’s Use of Innovative Merger Remedies, Antitrust, Fall 1995, at
16.
15. Separate Statement of Commissioner Azcuenaga at 1.
16. Statement of Chairman Pitofsky & Commissioners Steiger, Starek &
Varney at 2.
17. Brunswick Corp., 96 F.T.C. 151 (1980), aff’d as modified,
Brunswick Corp. v. FTC, 657 F.2d 971 (8th Cir. 1981), cert. denied,
456 U.S. 915 (1982).
18. Montedison S.p.A., No. C-3580 (FTC June 15, 1995) (consent order).
19. Potential future input supply problems will not always necessitate
divestiture. In the recent Exxon/Shell case, the Commission determined
that a long-term supply contract would solve the competitive problems.
Exxon Corp., FTC File No. 971-0007 (Aug. 20, 1998) (proposed consent
order).
20. Digital Equip. Corp., No. C-3818 (FTC July 14, 1998) (consent
order).
21. Digital Equip. Corp., Complaint at ¶¶ 11-13.
22. Adobe Sys., Inc., No. C-3536 (FTC Nov. 8, 1994) (consent order)
(Commissioner Owen dissenting).
23. Autodesk, Inc., No. C-3756 (FTC June 18, 1997) (consent order).
24. For a discussion of the evolution of government enforcement in
this area, see David Balto, The Murky World of Network Mergers:
Searching for the Opportunities for Network Competition, 42 Antitrust
Bull. 793 (1997).
25. First Data Corp., No. C-3635 (FTC Jan. 16, 1996) (consent order).
26. Automatic Data Processing, Inc., FTC Dkt. No. 9282 (June 18, 1997)
(consent order).
27. Automatic Data Processing, Inc., FTC File No. 951-0113 (Mar. 27,
1996) (consent order). ADP agreed to pay a $2.97 million civil
penalty, the third largest ever obtained for an HSR violation and the
largest ever involving the failure to submit documents. United States
v. Automatic Data Processing, Inc., No. 96-0606, 1996 U.S. Dist. LEXIS
21160 (D.D.C. Apr. 10, 1996).
28. For a description of recent government enforcement in this area,
see M. Howard Morse, Vertical Mergers: Recent Learning, 53 Bus. Law.
1217 (1998). For a more elaborate analysis of vertical merger
remedies, see Richard Parker & David Balto, The Merger Wave: Trends in
Enforcement and Litigation (forthcoming).
29. Silicon Graphics, Inc., No. C-3626 (FTC Nov. 16, 1995) (consent
order) (Commissioners Azcuenaga & Starek dissenting).
30. Cadence Design Sys., Inc., No. C-3761 (FTC Sept. 2, 1997) (consent
order) (Commissioner Azcuenaga concurring in part and dissenting in
part; Commissioner Starek dissenting).
31. Morse, supra note 28.
32. See, e.g., Martin Marietta/General Dynamics, No. C-3500 (FTC June
1994) (consent order).
33. See, e.g., Eli Lilly & Co./PCS Health Sys., No. C‑3594 (FTC July
1995) (consent order).
34. A similar prohibition of sharing non-public information is
contained in the order in Merck & Co., Inc., FTC File No. 971-0097
(Aug. 27, 1998) (proposed consent order), which arose from the
vertical merger between Merck and Medco Containment Services, a
pharmacy benefit manager.
35. The use of firewalls has raised questions in the past.
Commissioner Mary L. Azcuenaga issued a separate statement questioning
the extent to which the firewall adds to private contracts and the
FTC’s ability to effectively monitor compliance in one of the first
cases using a firewall. Alliant Techsystems, Inc., 5 Trade Reg. Rep.
(CCH) ¶ 23,714, at 23,474 (FTC Apr. 7, 1995).
36. First Data Corp., No. C-3635 (FTC Jan. 16, 1996) (consent order).
37. Glaxo plc, No. C-3586 (FTC June 14, 1995) (consent order).
38. Digital Equip. Corp., No. C-3818 (FTC July 14, 1998) (consent
order).
39. See Mary Lou Steptoe & David A. Balto, Finding the Right
Prescription: The FTC’s Use of Innovative Merger Remedies, Antitrust,
Fall 1995, at 16, 18.
40. See Andrew J. Strenio, Jr., FTC Commissioner, Why Thirteen Should
Be a Lucky Number for Victims of Price Fixing, Speech before the
Section of Antitrust Law, American Bar Association, 36th Annual Spring
Meeting, Washington, D.C. (Mar. 23, 1988) (advocating disgorgement in
antitrust cases in appropriate circumstances).
41. United States v. Pilkington plc, No. 94-345 (D. Ariz. May 25,
1994) (consent decree).
42. Intel Corp., FTC Dkt. No. 9288 (June 8, 1998) (complaint).
43. VISX, Inc. and Summit Technology, Inc., FTC Dkt. No. 9286 (Aug.
21, 1998) (proposed consent order).
44. U.S. Dep’t of Justice & Federal Trade Comm’n, Antitrust Guidelines
for the Licensing of Intellectual Property (1995), reprinted in
Antitrust Laws and Trade Regulation: Primary Source Pamphlet (Matthew
Bender 1998).
45. Dell Computer Co., No. C-3658 (FTC May 20, 1996) (consent order)
(Commissioner Azcuenaga dissenting). È
David A. Balto is Assistant Director, Office of Policy and Evaluation,
Federal Trade Commission. James F. Mongoven is an attorney in the
Office of Policy and Evaluation. The opinions expressed herein are the
authors and not necessarily those of the Commission or of any
individual Commissioner.