1

1	UNITED STATES FEDERAL TRADE COMMISSION
2	and
3	UNITED STATES DEPARTMENT OF JUSTICE
4
5
6
7	SHERMAN ACT SECTION 2 JOINT HEARING
8	REFUSALS TO DEAL PANEL
9	TUESDAY, JULY 18, 2006
10
11
12
13
14	HELD AT:
15	UNITED STATES FEDERAL TRADE COMMISSION
16	CONFERENCE CENTER CONFERENCE ROOM C
17	601 NEW JERSEY AVENUE, N.W.
18	WASHINGTON, D.C.
19	1:30 P.M. to 5:13 P.M.
20
21
22
23
24	Reported and transcribed by:
25	Sally Jo Bowling

2

1	MODERATOR:
2	ALDEN F. ABBOTT
3	Federal Trade Commission
4	J. BRUCE McDONALD
5	Department of Justice
6
7	PANELISTS:
8
9	William J. Kolasky
10	R. Hewitt Pate
11	Robert Pitofsky
12	Steven C. Salop
13	Thomas F. Walton
14	Mark Whitener
15
16
17
18
19
20
21
22
23
24
25

3

1	C O N T E N T S
2
3	Introduction
4
5	Presentations:
6	William J. Kolasky
7	Robert Pitofsky
8	R. Hewitt Pate
9	Steven C. Salop
10	Thomas F. Walton
11	Mark Whitener
12
13	Moderated Discussion
14
15	Adjournment
16
17
18
19
20
21
22
23
24
25

4

1	P R O C E E D I N G S
2	-  -  -  -  -
3	MR. ABBOTT: Good afternoon. I'm Alden Abbott,
4	Associate Director of the Bureau of Competition of the
5	Federal Trade Commission. I wish to join my
6	co-moderator, Deputy Assistant Attorney General for
7	Antitrust, Bruce McDonald, to welcome you to today's
8	session of the FTC/Justice Department hearings on the
9	antitrust implications of single firm conduct.
10	This is the fourth session in the ongoing
11	hearings. Prior sessions involved an introductory
12	overview of the topic, and sessions on predatory pricing
13	and buying.
14	Before we start, I need to cover a few
15	housekeeping matters. First, please turn off cell
16	phones, Blackberries and any other electronic devices.
17	Second, and most important, the restrooms are outside
18	the double doors and across the lobby. There are signs
19	to guide you. Third, in the unlikely event building
20	alarms go off, please proceed calmly and quickly as
21	instructed. If we must leave the building, go out the
22	New Jersey Avenue entrance by the guard's desk, follow
23	the crowd of FTC employees to a gathering point and
24	await further instruction. Finally, we request you not
25	make comments or ask questions during the session.

5

1	Thank you.
2	Now, before turning the podium over to my
3	colleague, Bruce McDonald, I'll briefly mention, prior
4	to giving more fullsome introductions, we're honored to
5	have six of the most distinguished leading lights of
6	antitrust here today. Bill Kolasky, Wilmer Cutler &
7	Pickering, former deputy assistant Attorney General;
8	professor and former dean and FTC chairman Robert
9	Pitofsky of Georgetown University Law Center, and Arnold
10	& Porter; Hew Pate, former assistant Attorney General
11	and currently partner at Hunton & Williams; Professor
12	Steven Salop, Georgetown University Law Center,
13	Consultant CRA International, and also an FTC alumnus;
14	Thomas Walton, director economic policy analysis,
15	General Motors Corporation, and also an FTC alumnus; and
16	Mark Whitener, senior counsel, competition law and
17	policy, General Electric Company, and also an FTC
18	alumnus. So we see there's a certain FTC flavor to the
19	distinguished speakers here today, but I won't say
20	anything more about that.
21	Bruce?
22	MR. McDONALD: If counting, there is a distinct
23	DOJ flavor on the panel, too. Let me say my welcome to
24	the joint DOJ/FTC single firm conduct hearings. The
25	hearings opened on June 20 with an overview of the

6

1	issues presented by single firm conduct and the
2	enforcement of Sherman Act Section 2. At the opening
3	hearings, both FTC Chairman Debbie Majoras and antitrust
4	AAG Tom Barnett emphasized the challenges in identifying
5	what conduct threatens long-term harm to competition and
6	the importance of developing clear rules to guide
7	business and that both underdeterrence and
8	overenforcement need to be considered.
9	Today is our fourth session, and our third day
10	of hearings. Our topic today is refusals to deal, which
11	is hard fought ground in the single firm conduct debate.
12	Our distinguished panel will focus on the circumstances
13	in which a firm's unilateral refusal to deal with a
14	competitor violates or should or should not violate
15	Section 2, addressing issues raised by Colgate, Otter
16	Tail, Kodak, Aspen, Microsoft and Trinko. The views of
17	our panelists have been influential in this debate, and
18	we appreciate the time that they have devoted to these
19	hearings.
20	Let me outline the agenda for you this
21	afternoon. Each of the panelists will take about 15
22	minutes to outline the issues and things critical, then
23	we'll take a 15-minute break, and then we'll dig deeper
24	into a discussion, giving the panelists an opportunity
25	to respond to each other's presentations and to consider

7

1	several propositions and hypotheticals that we hope will
2	initiate further discussion. The hearing will end at
3	about 5:00.
4	Let me turn the podium back to Alden Abbott to
5	introduce the presenters. Thank you.
6	MR. ABBOTT: Thank you, Bruce. Our first
7	speaker, Bill Kolasky, is cochair of Wilmer Hale Cutler
8	& Pickering, actually Wilmer Cutler Pickering Hale &
9	Dorr, it's a problem with all of these law firm mergers.
10	He co-chairs the firm's antitrust and competition
11	practice group. He's also had a distinguished record of
12	public service. From September 2001 through December
13	2002, he served as Deputy Assistant Attorney General for
14	International Antitrust at the Justice Department, at
15	which time he spoke out vociferously on the benefits of
16	an economic approach to antitrust in the international
17	forum and was very active in helping launch the
18	International Competition Network. His private practice
19	covers a full range of antitrust matters and Bill has
20	also taught antitrust law at American University, and he
21	speaks regularly on antitrust topics.
22	Bill?
23	MR. KOLASKY: Thank you very much, Alden, and
24	thank you, Bruce, as well, for inviting me to
25	participate in this. I have to say that it's somewhat

8

1	intimidating to be the first speaker in this afternoon's
2	session, especially given that I think all of the other
3	members of the panel, and probably most of you in the
4	audience, have thought longer and harder about these
5	issues than I have.
6	The other disadvantage of speaking first, of
7	course, is that everyone gets the chance to shoot at
8	what I'm about to say. I do think that I have, perhaps,
9	one comparative advantage, and only one, I'm going to
10	try to take full advantage of that, and that is my age,
11	and therefore, in fact, I've been doing this a lot
12	longer than most of the people in the room.
13	I've titled my talk refusals to deal with
14	rivals, because I want to distinguish very clearly
15	between refusals to deal with competitors as opposed to
16	refusals to deal with customers.
17	Refusals to deal with customers, I think involve
18	very different competitive concerns. The exclusionary
19	effects are more likely to be direct and immediate, and
20	there's a long line of cases running from Lorain Journal
21	to Dentsply that deal with refusals to deal with
22	customers. As I understand it, we're not here to
23	discuss those, we're here today to discuss refusals to
24	deal with rivals.
25	In structuring my remarks, I felt that I made

9

1	one of the classic rookie mistakes, I have far too many
2	slides and so I'm going to have to skip around somewhat,
3	but I wanted to touch on five basic topics. The first
4	is the pre-Trinko refusal to deal cases. Next I want to
5	talk briefly about Trinko. Then I want to talk about
6	the current dialogue that is going on, among others,
7	between Steve Salop and my partner, Doug Melamed over
8	the various standards for applying Section 2 generally.
9	I then want to stake out my own position as to what
10	analytical framework I think should be applied to
11	Section 2, and it's basically a step-wise rule reason
12	approach, applying the California Dental sliding scale.
13	And then I propose to talk about how they apply to
14	refusals to deal with rivals.
15	Focusing first on the pre-Trinko refusal to deal
16	law, there are basically, I think, four distinct lines
17	of cases. The first line of cases, and the oldest, are
18	the vertical integration cases from the 1970s and early
19	80s. The second line of cases are the essential
20	facilities cases, largely from the 1980s and early
21	1990s. The third line of cases are the intellectual
22	property cases, most recently the Federal Circuit's
23	decision in CSU. And then finally there is Aspen, which
24	because it's a Supreme Court case, I think deserves
25	particular mention and focus.

10

1	In the debate over refusals to deal, I've been
2	surprised in the recent publications how little
3	attention has been paid to the vertical integration
4	cases, which is really where a lot of the law in this
5	area was first developed. And when you go back and read
6	those cases, I believe, at least, that the analytical
7	framework that they used is a surprisingly sound one,
8	given that these cases were decided largely in the 1970s
9	and early 80s as we were just emerging from what Doug
10	Ginsburg refers to as the dark ages of antitrust.
11	Many of the cases, some of which my firm was
12	involved in, involved refusals to deal by monopoly
13	newspapers that were vertically integrating into
14	distribution. The obvious reason why these papers were
15	vertically integrating into distribution was to get
16	around the problem that was created by Albrecht, by the
17	rule that maximum resale price by principles is per se
18	unlawful. Since it was obviously efficient to have a
19	single delivery person covering each block, newspapers
20	found themselves basically with the situation where they
21	were dealing with independent dealers, giving those
22	dealers a monopoly, and they had no way to prevent those
23	dealers from charging monopoly prices higher than what
24	the newspaper itself would have charged.
25	It's not surprising, therefore, that the cases

11

1	for the most part ended up with the courts ruling in
2	favor of the newspapers and upholding their refusal to
3	continue to deal with independent dealers and vertically
4	integrating into the distribution themselves.
5	When you go back and read the cases, and most
6	notable the Paschall versus Kansas City Star decision,
7	in 1984, which was an en banc decision of the Eighth
8	Circuit, what you find is that the courts applied
9	essentially a Section 1 rule of reason standard in
10	evaluating these unilateral refusals to deal. In that
11	sense, I would argue that they are in a way ahead of
12	their time, because it was really not until the
13	Microsoft decision in 2001 that a court of appeals here
14	in the D.C. Circuit affirmatively embraced the rule of
15	reason as the applicable standard for Section 2.
16	Applying that Section 1 rule of reason
17	framework, the Eighth Circuit found that the
18	anticompetitive effects from the alleged loss of
19	potential competition as claimed by the plaintiffs were
20	slight, and that the newspaper had offered several
21	legitimate business reasons for its decision to
22	vertically integrate into distribution.
23	One of the most interesting things about the
24	case is that the newspaper did not rely on the argument
25	that I relied on in my opening remarks about this case,

12

1	namely the need to get around Albrecht. Instead, the
2	newspaper focused on the desire to be more responsive to
3	subscribers and have more uniform pricing in order to
4	facilitate advertising.
5	Quite frankly, those are relatively weak
6	justifications for what the newspaper was doing, and yet
7	nevertheless the court held without scrutinizing those
8	justifications very closely, that they outweighed the
9	rather minimal showing of anticompetitive injury that
10	the plaintiffs had made.
11	One of the key factors in causing the court to
12	reach that decision was its determination -- and this is
13	consistent with what I said earlier on Albrecht -- that
14	a vertically integrated newspaper was likely to charge
15	lower prices than if you had unintegrated monopolists at
16	both the publication level and the distribution level.
17	The essential facilities cases, I'm going to
18	skip over lightly, because others are going to be
19	speaking about those in more detail. There are two
20	things that I want to note about them. The mother of
21	essential facilities cases, at least with respect to
22	unilateral conduct, is of course the Supreme Court's
23	decision, Otter Tail. What people often don't comment
24	on is that that was a decision in the mid-1970s, again,
25	as we were just emerging from the dark ages, it was a

13

1	four to three opinion written by Justice Douglas, who
2	probably wrote more decisions that antitrust lawyers now
3	try to distance themselves from than almost any other
4	Justice.
5	The other thing that's important about the key
6	essential facilities cases such as Otter Tail and the
7	Seventh Circuit's decision in MCI v. AT&T is that these
8	cases do not involve just a simple refusal to deal by a
9	monopolist. Rather, they were cases in which the
10	monopolist had engaged in a whole pattern of conduct
11	that was designed to exclude rivals from these monopoly
12	markets.
13	The next line of cases, as I mentioned, are the
14	cases involving intellectual property rights, the First
15	Circuit's decision in Data General, the Ninth Circuit's
16	decision in Kodak and the Federal Circuit's decision in
17	CSU. There's been an enormous amount of ink spilled
18	about these decisions, including a very good article by
19	Hew Pate, and I'm sure Hew will have something to say
20	about this line of cases.
21	The important point, I think, that one draws
22	from these line of cases is the Second Circuit's
23	recognition, which was endorsed even by the Ninth
24	Circuit, that an author's or inventor's desire to
25	exclude others from the use of copyrighted or patented

14

1	work is a presumptively valid business justification for
2	any immediate harm to consumers that might result from a
3	refusal to license.
4	The debate really, then, is between the Ninth
5	Circuit and the Federal Circuit under what's necessary
6	to rebut that presumption, with the Federal Circuit
7	taking probably the most restrictive view that the
8	presumption is virtually irrebuttable unless there is
9	additional conduct beyond just the simple refusal to
10	license, such as an illegal tie, fraud on the Patent &
11	Trademark Office, or sham litigation. And I think that
12	is consistent, in fact, with cases like MCI and Otter
13	Tail, if you go back and read those decisions.
14	That brings me to Aspen Ski, which was the first
15	serious effort, I would argue, by the Supreme Court to
16	deal with the question of what standards should apply to
17	refusals by monopolists to deal with its rivals, and the
18	key points here that I want to bring out are that the
19	Court focused not just on the impact on the rival, but
20	also on the impact of the refusal on consumers, and the
21	Court also made it clear that what it was looking at
22	under Section 2 was whether the defendant was seeking to
23	exclude rivals on some basis other than efficiency, that
24	is other than through competition on the merits. And I
25	think that's a very important strand that needs to be

15

1	kept in mind as one thinks about these cases.
2	The other point that's important to make about
3	Aspen requires really looking at the facts of the case
4	and what the conduct was. Again, as in Otter Tail and
5	MCI, the conduct was not a simple refusal to deal.
6	There was a lot of other conduct going on there,
7	including to me most significantly the fact that Ski Co.
8	discontinued its own three-day, three mountain pass so
9	that the only way somebody could get a discount on a
10	multi-day, multi-mountain pass was to buy a six-day
11	pass, and that meant that if the vacationer wanted to
12	ski the Highlands, they almost certainly had to pay
13	twice, both for the day ticket to the Highlands and the
14	six-day pass to the Highlands. The other thing that's
15	important is that, while the court described Ski Co.'s
16	justification as pretextual, the court also gave fairly
17	close scrutiny to those justifications before reaching
18	that conclusion.
19	Trinko, I'm not going to spend very much time
20	on, because others are going to spend a lot of time on
21	it. The key message point, of course, is that the Court
22	appeared to adopt a very restrictive view as to when a
23	monopolist might have a refusal to deal and cooperate
24	with its rivals.
25	Because I'm running out of time, I'm going to

16

1	jump ahead to the contending standards. As I say, there
2	are basically three sets of contending standards out
3	there now, in this area. One is what I would call the
4	Section 2 rule of reason approach, taken by the D.C.
5	Circuit in Microsoft and by the Eighth Circuit in
6	Paschall, the profit sacrifice or no economic sense test
7	that Greg Werden from the Justice Department and Doug
8	Melamed have been advocating and I think Hew from time
9	to time has advocated it as well, and then finally the
10	essential facilities doctrine.
11	Again, because we're running out of time, I'm
12	going to skip ahead to my proposed synthesis. I come
13	down, as I think about this, in favor of basically the
14	Microsoft step-wise rule of reason test for exclusionary
15	conduct. I think that test involves, as the court said,
16	basically four steps. First, an examination of whether
17	the monopolist's conduct, in this case its refusal to
18	deal, had the requisite anticompetitive effect.
19	Second, a requirement that the monopolist, if
20	the plaintiff establishes a prima facie case, proffer
21	some nonpretextual procompetitive justification for its
22	action, and if it does so, the burden then slides back
23	to the plaintiffs to rebut that justification. And it's
24	only if the plaintiff meets that burden that you move on
25	to the fourth and final stage, which is balancing.

17

1	That's the reason why I don't particularly like to have
2	this test described as the balancing test, because in
3	fact, you rarely reach the fourth balancing step in the
4	test.
5	In applying the step-wise rule of reason under
6	Section 2, I would argue that the courts should do just
7	as they do in Section 1, and as I believe they do in
8	practice under Section 2, and that is apply a sliding
9	scale. That is to say, as Justice Souter wrote in
10	California Dental, what is required is an enquiry need
11	for the case. In other words, the stronger the evidence
12	of anticompetitive harm, the closer the scrutiny of
13	proper justifications.
14	Going back to, I'm not sure how to go to a
15	previous slide, I want to go back to Microsoft for a
16	second, because -- I'm sorry about this. I hope I get a
17	minute for my technological ineptitude. Here we go.
18	In Microsoft, if you read the decision closely,
19	you will see that the court, in fact, applied exactly
20	this kind of a sliding scale. When it came to the
21	license restrictions that Microsoft imposed on OEMs, the
22	court subjected Microsoft's proposed justifications to
23	very close scrutiny. When it came, however, to the
24	integration of Internet Explorer and Windows, the court
25	expressed at the very outset of that section of its

18

1	opinion a general deference to the dominant firm's
2	product design decisions, and the only reason it found
3	Microsoft's conduct unlawful, to the extent it did, is
4	that Microsoft proffered no justification whatever for
5	its decisions.
6	What I found interesting, and I credit this to
7	one of our summer associates, Tian Mayimin, who is in
8	the audience today, is how similar the California Dental
9	sliding scale approach to the rule of reason is to what
10	the courts do in the constitutional area, both under the
11	First Amendment, and under equal protection, where over
12	the years, what began back in the 1960s as a balancing
13	test, has evolved instead to three different levels of
14	review, strict scrutiny, intermediate scrutiny, and weak
15	scrutiny, in which the degree to which the court
16	subjects the proffered justifications for the
17	government's action depends on how objectionable the
18	conduct is in terms of First Amendment principles and/or
19	equal protection.
20	And I would suggest that the analogy in the
21	antitrust area is to the test we use for determining
22	whether or not the proper justifications justify the
23	conduct at issue. We often talk about needing to find
24	that the conduct is reasonably necessary, that's a
25	relatively tough standard.

19

1	A more relaxed standard would be to find that
2	it's reasonably related, and an even more relaxed
3	standard would be that it's plausibly related, which is
4	the standard the Supreme Court adopted in Broadcast
5	Music in determining whether or not the per se rule
6	should be applied. I would argue that you could use
7	that same sliding scale under Section 2, where the
8	degree of scrutiny depends on the nature of the conduct
9	in question.
10	Why do I prefer the rule of reason approach to
11	the profit sacrifice test? I think basically four
12	simple reasons. One is that it focuses directly on
13	competitive effects, whereas the profit sacrifice test
14	focuses more on the effect on the monopolist, rather
15	than the effect on consumers. Second, because, as Steve
16	Salop has pointed out quite persuasively, exclusionary
17	conduct can be profitable, even in the short-term, and
18	in fact, if you read the facts of Aspen Ski, I suspect
19	that even there, Aspen's conduct was profitable in the
20	short-term, even though it degraded the attractiveness
21	of its product to the skiers, and that's because it
22	would have shifted skiers from Highlands to the Aspen
23	mountains, thereby increasing its revenues, i.e., even
24	if the total number of skiers coming to the Aspen area
25	generally declined.

20

1	Third, at least as I have read the articles, the
2	profit sacrifice test, as it has been articulated,
3	doesn't acknowledge the need to calibrate the degree of
4	scrutiny of the business justifications based on the
5	strength of the evidence of competitive injury. Doug
6	Melamed, for example, has argued that one can look at a
7	refusal to deal as basically a make-or-buy decision, and
8	that it should be unlawful if it would be more
9	profitable for the monopolist to buy the downstream
10	services than to vertically integrate them. I would
11	argue that that is too high a degree of scrutiny for the
12	courts to impose on those kinds of decisions.
13	And then finally, there is no obvious reason why
14	courts should be any less able to evaluate competitive
15	injury and business justifications in a Section 2 versus
16	a Section 1 setting. What should differ is how strictly
17	they scrutinize the justifications, not the test that
18	they apply.
19	Thank you.
20	(Applause.)
21	MR. ABBOTT: Thank you, Bill. Now I have the
22	honor of introducing Robert Pitofsky, a name known
23	certainly to all of you and throughout the antitrust
24	world, former FTC Chairman, Commissioner and Bureau of
25	Consumer Protection Director, distinguished background

21

1	in private practice, currently of counsel at Arnold &
2	Porter, and of course very distinguished academic,
3	former NYU law professor, then dean of Georgetown Law
4	School, currently Sheehy Professor in Antitrust and
5	Trade Regulation Law at Georgetown University Law
6	Center. His writings are many. He has co-authored,
7	Cases and Materials on Trade Regulations, which is in
8	its fifth edition, one of the most widely used antitrust
9	and trade regulation case books.
10	Bob Pitofsky.
11	(Applause.)
12	MR. PITOFSKY: Thank you all and good afternoon.
13	It's great to be back at the FTC, and to see that the
14	DOJ and the FTC are continuing the tradition of taking
15	on the toughest issues and addressing them not
16	necessarily by litigation, but by hearings like this.
17	And I do regard the definition of exclusion under
18	Section 2, and refusals to deal in particular, as about
19	the toughest issues that an antitrust lawyer is required
20	to face today.
21	I'm going to do three things here. One, I want
22	to put refusals to deal in a broader context, and I
23	believe that's what Trinko's majority opinion was
24	designed to do. Secondly, I want to say a little bit
25	about the general universal test that Bill talked about

22

1	in such an interesting way. I just have one question,
2	because I agree with virtually all that he had to say.
3	And then I'm going to discuss, the antitrust concept of
4	essential facilities and whether essential facilities is
5	such an unwise doctrine that it ought to be abolished.
6	Let's start with Trinko, because I don't think
7	Trinko is just about the facts of that particular case.
8	It was a unanimous opinion. I would have voted to
9	reverse the Second Circuit, too. I had no problem with
10	the holding. It's the dicta in Trinko that went on and
11	on and on, and I'm disappointed that other judges on the
12	court didn't concur separately, and write that they were
13	not ready to go along with all this additional talk.
14	More broadly, I think Justice Scalia was saying, very
15	directly, that he's uncomfortable, he's skeptical about
16	enforcement of Section 2, and thinks that Section 2,
17	certainly compared to Section 1 of the Sherman Act,
18	causes more harm than good. His reasons were that there
19	are too many false positives, as he put it, in Section
20	2, that Section 2 enforcement tends to chill the
21	incentives of aggressive and innovative companies, that
22	he's uncomfortable with a generalist antitrust court
23	taking on issues like those raised by Section 2
24	enforcement, and the remedy, especially with refusal to
25	deal, is at least difficult and may be impossible.

23

1	Let me just go through these. First of all,
2	what is this false positives thing? I didn't agree with
3	the Second Circuit either, but I didn't conclude that
4	Section 2 raised many false positives as a result of
5	that wrong decision. Is the meaning that lots of
6	Section 2 cases have been brought by the government and
7	private parties and have been thrown out on motions to
8	dismiss, not stating a legitimate case? Well, let's go
9	back and review the record: Lorain Journal, Walker
10	Process, Otter Tail, Kodak, Xerox, Aspen, and Intel.
11	The plaintiff won every one of those Section 2 cases.
12	Now you might say yes, but they were false positives,
13	Otter Tail should have been decided the other way. But
14	the Supreme Court decided Otter Tail in favor of the
15	plaintiff, and the Court has not subsequently overruled
16	the decision.
17	Now there have been mistakes that have been
18	made, but the idea that there's just constant false
19	positives in Section 2 enforcement, I don't know where
20	that's coming from.
21	Second, Section 2 enforcement chills incentives
22	for innovative companies. I'm agnostic on that. Maybe
23	that's true. Just show me the data. Show me anyone who
24	has done a study which demonstrates that once a company
25	is aware that it may have to engage in mandatory

24

1	licensing, at a reasonable royalty, they cut back on
2	their investment in innovation. I haven't seen it. But
3	I'm uncomfortable with all these ex cathedra statements
4	that that would occur.
5	Third, uncomfortable because generalist
6	antitrust judges are deciding these cases? Well, who
7	are the judges deciding joint venture cases? Merger
8	cases? Rule of reason cases? They all involve
9	trade-offs, just like Section 2; they all involve
10	generalist judges. Up until now, I thought U.S.
11	antitrust was doing a pretty good job, and I'm not
12	troubled that district judges are making a botch out of
13	these trials.
14	On refusal to deal, if you mandate disclosure,
15	you have not just the decision about mandating, you have
16	a decision about at what royalty, what terms, what
17	timing, and so forth. And there's no question, that
18	complicates this issue immensely. It was worked out in
19	Aspen Ski, it was worked out in Otter Tail, although
20	there was a Federal Power Commission at the time Otter
21	Tail was decided to help to work out the remedy. The
22	question for me is, given the fact that the remedies in
23	these cases are difficult, do you throw up your hands
24	and say, impossible, therefore the monopolist can do
25	anything it wants, or do you try to work out the best

25

1	remedy you can? Sometimes the remedy is easy. Perhaps
2	the monopolist has already been licensing other people,
3	but refuses to license potential competitors. It's not
4	common, but it happens.
5	Sometimes the monopolist has been selling in
6	other markets at a price it was comfortable with.
7	That's the beginning of negotiation for this remedy. I
8	grant immediately, it's difficult, the question is, does
9	that mean free reign for the monopolist?
10	Second, on proposals for a general rule, first
11	of all, I want to compliment Hew Pate, now Bill Kolasky,
12	Steve Salop, Doug Melamed, Greg Werden, all of whom are
13	trying to come up with a rule that lends certainty and
14	predictability to Section 2 generally and refusals to
15	deal specifically. But in the end, I think the
16	balancing test as advocated in Aspen and Microsoft is
17	where you have to end up. I'm uncomfortable with the
18	universal rule that focuses on the welfare of the
19	monopolist. That's the profit sacrifice test. I'm more
20	concerned about the consumer, not whether the monopolist
21	sacrificed profits.
22	On the approach that asks if there was any
23	plausible economic reason for doing something, you know,
24	I think lawyers can always come up with a plausible
25	economic reason. That's not the issue. The issue is

26

1	whether that reason is good enough to outweigh the
2	anticompetitive effects. And that, it seems to me, is
3	what you have to do.
4	I would welcome a clearer rule, but in the end,
5	you have to take into account the redeeming virtues, the
6	business reasons, the justification, but if the
7	anticompetitive effects are large and the efficiencies
8	small, you can't stop with step one, you have to get to
9	as many steps as you can, and that's the question that I
10	would like to address to Bill. His third step is: what
11	was your justification? Suppose the defendant states
12	it, and then the other side comes in and let's say fails
13	to show that your justification was not plausible,
14	substantial, significant -- that is, there was some
15	justification. Do we stop there? Or do we go on to the
16	question of maybe you had a good justification, but it
17	didn't outweigh the anticompetitive effects?
18	Let me return finally return to the issues
19	relating to essential facilities. Let me start with the
20	proposition that the general rule is and must be no
21	general duty to deal. You don't have to disclose these
22	kinds of information except under a very rare exception,
23	and the exception is where a monopolist has a bottleneck
24	monopoly. The scholars are suppose to all say let's get
25	rid of the doctrine. That's really not what they say.

27

1	They say it should be rare and extremely narrow, that's
2	Areeda, that's Hovenkamp. I say the same thing. It
3	should be very rare, and very narrow.
4	But I think it should be an exception to the
5	general rule. I think the best summary of the
6	limitations on essential facility claims is in the MCI
7	case, which I notice virtually every lower court that
8	either sustains or overrules the essential facilities
9	claim, they all use the MCI test. The test is as
10	follows: one, it only applies to a monopolist; two,
11	other potential rivals cannot duplicate the facility or
12	the service. It's not just that it would be hard to
13	duplicate it, it's they can't do it at all. Three, the
14	monopolist denies access to the service or the facility;
15	and four, that it's feasible to make use of the facility
16	available.
17	I remember there was a throw-away line in Otter
18	Tail, and that's not my favorite case in this area, but
19	there's a throw-away line saying, you know, if you had
20	said that there's an engineering reason why you couldn't
21	wheel power to those municipalities, this would be a
22	different case. The problem with Otter Tail is there
23	was no plausible explanation except anticompetitive
24	purpose for refusing to wheel the power.
25	The EU has added a few additional

28

1	qualifications: The refusal to deal must eliminate all
2	competition, and that the product that the person
3	seeking access would make is not just a clone of the
4	first product, I don't think you need those two
5	additional restrictions, although they do narrow the
6	doctrine.
7	I think with the general qualifications stated
8	in MCI, we're in good shape. And I do want to emphasize
9	here -- the idea is not that the monopolist is giving
10	anything away, it's receiving reasonable royalties that
11	a court or an expert witness figured out was acceptable.
12	Finally, it has been said that there's Terminal
13	Railways, there's Otter Tail, there's Associated Press,
14	and there aren't many cases that address the essential
15	facility issue. That's just not true. There are scores
16	of lower court cases, including lower court cases since
17	Trinko kicked a lot of mud on the essential facilities
18	doctrine, which have addressed the claim of essential
19	facilities.
20	Let me conclude by saying that while Section 2
21	enforcement is an area that deserves to be addressed, at
22	least for the time being, I think Aspen Ski is the best
23	approach to it. It applies a rule of reason, and the
24	Court looked at and rejected any plausible business
25	justification. It seems to me a monopolist ought to

29

1	have some reason for refusing to do business with a
2	potential rival. I just don't think of antitrust as
3	being so narrowly confined when it comes to the market
4	power of a monopolist. I look forward to the
5	discussion. Thank you.
6	(Applause. )
7	MR. ABBOTT: Well, so far we've heard one
8	endorsement of the Cal Dental sliding scale approach and
9	an endorsement of an approach based on Aspen Ski,
10	variations on balancing approaches, and it will be
11	interesting to see what our next speaker has to say
12	about such approaches.
13	Hew Pate, partner and head of Hunton & Williams'
14	Global Competition Practice Group, is a former Assistant
15	Attorney General for antitrust, until relatively
16	recently. Hew's practice involves all aspects of
17	competition law, counseling and litigation. Hew has
18	served as Ewald Distinguished Visiting Professor of Law
19	at Virginia, from which he graduated first in his class.
20	Hew clerked for two Supreme Court Justices, Justice
21	Powell and Justice Kennedy.
22	Hew?
23	(Applause.)
24	MR. PATE: Thank you very much, Alden. It is
25	great to be here at the Commission's conference facility

30

1	for these hearings. I appreciate the opportunity to
2	take a part in them. I have submitted some written
3	testimony, which I have prepared on behalf of the United
4	States Telecom Association. That, as I understand it,
5	will be available on the website for these hearings. As
6	to my elaborations on that and what I say in the
7	exchange, you've just got me, and all the views I
8	express, both in the written testimony and here, are my
9	own.
10	The general point of the testimony I'm going to
11	give is that independent competition among competitors
12	who are not relying upon one another for assistance or
13	even for pulled punches in the competitive process is
14	what best produces innovative products at low prices.
15	Government-imposed duties to assist competitors force
16	courts into setting prices, a task for which they are
17	not very well equipped, particularly in capital
18	intensive or high technology fields. The uncertainty
19	that is caused by indeterminate liability rules and
20	duties to assist competitors are likely to retard
21	desirable investment.
22	And the U.S. system of private litigation, which
23	uniquely puts decisions on these types of issues in the
24	hands of general judges, as has been mentioned, and in
25	the hands of juries, sometimes with very vague

31

1	instructions, exacerbates the problem. And I would
2	suggest that recent experience in the telecommunications
3	field provides a good illustration of this point.
4	This testimony, my testimony is first going to
5	talk about refusals to deal and essential facilities.
6	The question is where after Trinko these doctrines
7	should go in the future, and my suggestion is not much
8	of anywhere. These doctrines inherently generate
9	uncertainty, they threaten returns on investment, and by
10	doing so, they discourage investment from taking place.
11	With respect to refusals to deal, or as I prefer
12	to think of it, duties to assist competitors, all have
13	the right to take a different tack. I think in the wake
14	of Trinko, as we have seen lower courts try to make
15	sense of, and cabin the Aspen decision, that the time
16	has come for Aspen to be overruled, and that the law
17	would be better with it off the books, and that the
18	Commission and the Division would do a service to the
19	law by advocating that in their report from these
20	hearings.
21	The second major point I want to make, while I
22	don't at least in this presentation want to debate the
23	variety of standards, as has been mentioned, I think the
24	no economic sense test has a good deal to be commended.
25	At the Antitrust Modernization Commission, I have

32

1	responded to some criticisms and made a general defense
2	of that test, but for today, I simply want to suggest
3	that the agencies would do a service by continuing to
4	push for more objective standards in this area. And to
5	my mind, while a general balancing test is flexible,
6	because it can apply in a wide variety of circumstances,
7	it is inherently lacking in any objective content that
8	businesses can apply in a predictable manner to make
9	their decisions. And while there may be different
10	formulations of it, some variation of a price-cost
11	comparison in my judgment is going to be necessary if
12	objectivity is going to be brought to the inquiry.
13	With respect to the telecommunications industry
14	experience, I think it does shed some light on whether
15	duties with forced sharing are likely to produce
16	desirable results. Telecommunications is an area where
17	huge capital expenditures and great risk need to be
18	undertaken to provide the product, and before any
19	profits can be made. I had a good deal of experience in
20	this industry in working on DOJ's implementation of the
21	1996 Act. And my experience there was that the DOJ
22	staff worked tremendously hard to try to implement that
23	act. But my experience in that process also left me
24	convinced that forced sharing of assets with competitors
25	is not a sound foundation for promoting competition.

33

1	As you all you are aware, the unbundling
2	obligations of the 1996 Act were premised on a so-called
3	stepping stone theory, the idea that if competitive
4	local exchange providers were given mandated wholesale
5	price access to incumbent local exchange providers'
6	facilities, this would allow so-called CLACs to enter
7	these markets officially without building facilities,
8	without undergoing that inherent risk. This would bring
9	immediate competition of a sort, and importantly, it
10	would then allow CLACs to build their own facilities so
11	that facility-based competition could follow thereafter.
12	A lot of water has gone under the bridge since
13	the passage of that Act in attempts to administer it. I
14	think the basic lessons are difficult to deny at this
15	point. Rather than provide a stepping stone to
16	independent competition, sharing obligations led to
17	demands for ever greater and more complicated sharing
18	obligations, many of which were found unlawful by the
19	courts in ensuing litigation.
20	One writer who has actually supported forced
21	sharing as a part of the antitrust laws recently summed
22	it up this way: "The 1996 Act is arguably a good
23	example of the questionable effectiveness of legally
24	mandated sharing. After eight years, the FCC has failed
25	to produce a legal system of access, and has instead

34

1	furthered a disastrous $50 billion Telecom boom and bust
2	in local telecommunications."
3	The experience there, I would suggest, is
4	illustrative of what happens when -- even when an
5	agency, but when an agency and parties who can be
6	protected want to litigate over the agency's rulings and
7	what the forced sharing obligation will mean, I think
8	provides an illustration of what is likely to ensue.
9	I think it also appears clear at this point that
10	the Act's forced sharing obligation has in many
11	instances slowed investment that otherwise would have
12	been made. Bob asked, and other speakers wonder what is
13	the empirical case for suggesting that incentives would
14	be chilled. Among one collection of studies, I would
15	point you to one by Scott Wallsten at the AEI-Brookings
16	Joint Center For Regulatory Studies, which can be found
17	on their website, and in summarizing the work in this
18	area, he suggests that although there are a few
19	dissenting voices, most economists and most studies
20	conclude that unbundling obligations in the U.S. reduced
21	incentives to invest in high-speed Internet
22	infrastructure. Cable companies which weren't bound by
23	these sort of unbundling obligations deployed more
24	quickly. DSL has lagged behind cable in terms of
25	deployment. That's the opposite situation we see in

35

1	many other countries.
2	The telecommunications industry recently has
3	rebounded, perhaps not coincidentally, with a diminution
4	of forced sharing obligations, and where reform of the
5	1996 Act is headed, is not entirely clear. But I do
6	think that antitrust generally can learn some lessons
7	from the experience, and the most important is that
8	forced sharing discourages and slows innovation.
9	Second, I certainly do believe that the many
10	complex and unforeseeable consequences of a forced
11	sharing regime are extremely difficult to administer.
12	It may be that in certain circumstances a regulatory
13	framework can administer forced sharing obligations in
14	some circumstances, or that a regulatory judgment will
15	be made that it should, but as a general matter, as a
16	general antitrust principle, and this is a point Justice
17	Stewart made in his dissent in Otter Tail, the rare
18	situations where that would be necessarily are not very
19	easily translated into a general duty of antitrust to be
20	applied across all industries. So, certainly in my
21	judgment, the transaction costs that come with a broad
22	sharing obligation are likely to outweigh the benefits.
23	Let me turn to refusals to deal and essential
24	facilities under the antitrust laws. We've heard some
25	comment about Trinko, and Aspen, already, and the three

36

1	rationales that the Court in Trinko offered for
2	limiting, very severely, any duty to assist competitors.
3	The Court did that in granting a motion to dismiss,
4	holding that the plaintiff's claim in Trinko was so
5	lacking in traditional antitrust merit that it does not
6	even require discovery before dismissal of the case.
7	And the three rationales, as you know, were the
8	negative incentive effects, both on the incumbent, the
9	high-market share incumbent, and on potential new
10	entrants from a sharing rule. Yes, skepticism of
11	generalist courts and juries' ability to manage sharing
12	obligations to set terms and prices. And then finally,
13	this idea of false positives. I think false positives
14	doesn't necessarily mean that we go to the Supreme Court
15	or even to lower courts and figure out whether the
16	defendants or the plaintiffs were winning, or whether
17	cases were rightly decided, but it does require some
18	consideration of the duties of those who are charged
19	with risking capital and conducting business, about
20	whether, in fact, their potential competitive activities
21	are chilled by the fear of being embroiled in litigation
22	under sharing duty types of rules, and for that reason,
23	I think that the risk of false positives is significant.
24	As to Aspen, while I think Aspen, as I have said
25	elsewhere, can be reconciled with a no economic sense

37

 

1	approach to the law and as consistent with it, since
2	Trinko, a number of courts, and some commentators have
3	come to view Aspen as standing for the proposition that
4	once a course of sharing conduct begins, that it
5	shouldn't be stopped. And if that's what Aspen is going
6	to stand for, then I think we would all be better off if
7	the case were overruled.
8	The reason for that, I think is pretty simple,
9	that while it is a way to distinguish the fact pattern
10	in Aspen from the fact pattern in Trinko, there's
11	nothing in economics that would suggest that the facts
12	are not likely to change in a pre-existing relationship.
13	There's no particular reason to believe that a course of
14	conduct that was once entered into remains efficient
15	forever.
16	So, it may be true that a voluntary course of
17	dealing provides an initial benchmark to set a price
18	that presumably the parties wouldn't have entered into
19	the relationship unless it were mutually profitable, all
20	that's true, and mitigates to some extent the concerns
21	that were in existence in Trinko, but it does not
22	eliminate them.
23	The other serious problem I think with a duty of
24	continued sharing is that it can prevent voluntary
25	sharing from taking place in the first place. This is a

38

1	point Judge Posner made in the Olympia Equipment Leasing
2	Company case, a case where Western Union had initially
3	assisted Olympia, decided to stop, got sued for doing
4	so, and as Judge Posner put it, if Western Union had
5	known that it was undertaking a journey from which there
6	could be no turning back, a journey it could not even
7	interrupt momentarily, it would have been foolish to
8	have embarked. And I think that's the real risk of a
9	developing idea that Aspen stands for the proposition
10	that you just can't stop sharing if you ever start.
11	Essential facilities, I won't spend too much
12	time on. I certainly do not think it adds anything as a
13	stand-alone theory of liability. I think Professors
14	Areeda and Hoenkamp said it well, the doctrine is
15	harmful because, I quote, "Forcing a firm to share its
16	monopoly is inconsistent with antitrust basic goals for
17	two reasons. First, consumers are no better off when a
18	monopoly is shared. Ordinarily a price and output are
19	the same as they were when one monopolist used the input
20	alone. And second, the right to share monopoly
21	discourages firms from developing their own alternative
22	inputs."
23	I will conclude, and time is running out, simply
24	by renewing a call for the agencies to participate in
25	advocating more objective standards. I think we're at a

39

1	high water mark now of criticisms leveled at the
2	standard-less nature of Section 2 generally. The OECD
3	competition committee recently issued a background note
4	that collects a number of these. I recall Elhauge has
5	described the exclusionary conduct law that exists today
6	as using a barrage of conclusory labels to cover for a
7	lack of any well-defined -- for any well-defined
8	criteria for sorting out desirable from undesirable
9	conduct. Even Eleanor Fox, with whom I often disagree
10	on panels like this, states that a number of the
11	contemporary cases tend to be noncommittal and rely on
12	obfuscatory language in their use of terms, such as
13	anticompetitive.
14	So, I think uncertain legal and regulatory
15	regimes, like limits on investment, are likely to prove
16	strong deterrents to investment, and innovation.
17	Certainly the continued reliance in some cases on intent
18	is one example of the type of subjective standards that
19	can lead to uncertainty and retard investment.
20	There is some positive sign, I think, on the
21	horizon that the Supreme Court may continue to look into
22	this area in the Weyerhaeuser case that they've granted
23	recently, where liability was imposed on the basis of
24	purchasing more saw logs than were needed. I would
25	suggest that we're really not going to do very well in a

40

1	regime where juries make a determination based on what
2	is right and wrong in log buying, without any more
3	objective basis for decision.
4	I'll stop there. As to the empirical basis for
5	all this, I would simply suggest that if the government
6	is going to intervene, if it's going to decide to
7	require sharing of a facility, if it's going to decide
8	not to use a property rule for determining how assets
9	are going to be used, but instead use a liability rule
10	to take from the Doug Melamed paradigm from the famous
11	law review article he authored with Judge Calabresi a
12	long time ago, that it ought to have some pretty serious
13	grounding for believing that the situation is going to
14	be made better. I don't think right now that an
15	empirical case can be made that forced sharing, that
16	this aspect of antitrust used to assist competitors is
17	going to leave consumers better off. I suggested some
18	time before I left government that the Modernization
19	Commission could do a study by trying to look into the
20	empirical basis for different areas of antitrust.
21	That's a hard thing to do, as they quickly decided, but
22	without it, in an area where the economics don't produce
23	a real consensus, I think the basis for government
24	intervention is lacking.
25	Bob asked whether we should just throw up our

41

1	hands because it's so difficult. Emil Paulis, who works
2	at the European Commission, used to make the same
3	comment after he heard me speak, and he would always
4	say, well, Hew, you just want to throw the baby out with
5	the bath water, because the standards are so difficult.
6	And I always would respond by saying, well, Emil, if
7	I've got a baby, and I've got to dip it into some bath
8	water, I would like to have some reason to believe that
9	the baby is going to be cleaner after I take it out than
10	it was before I put it in. And I don't think in this
11	area of the law that we have that.
12	Thanks, I look forward to the discussion.
13	(Applause.)
14	MR. ABBOTT: The people who are standing in the
15	back, there are some seats up front, so don't be shy,
16	there are seats. Thanks, Hew.
17	So, now we have two rational balancers and one
18	antitrust skeptic, and now we're going to turn to our
19	first academically trained economist on the panel, Steve
20	Salop, professor of economics and law at Georgetown
21	University Law Center, where he teaches antitrust law
22	and economics, economic reasoning for lawyers, and in
23	addition maintains an active consulting practice at CRA
24	International. Steve is no stranger to government,
25	having worked at the Civil Aeronautics Board, the

42

1	Federal Reserve Board and the Federal Trade Commission.
2	Now I remember him giving tutorials to young staffers on
3	economics at the FTC, young bright staffers, I was one
4	of them. And he did a very impressive job in that
5	regard. Steve has written widely in leading antitrust
6	journals, on this topic of Section 2, and I, for one,
7	look forward eagerly to hear his comments.
8	Steve?
9	(Applause. )
10	MR. SALOP: Thank you. I'm really pleased to be
11	here. I'm thrilled that Bill Kolasky seems to agree
12	with me. That's one down at Wilmer Cutler and several
13	to go I guess.
14	I want to talk a little bit about the general
15	exclusion standards, but just for a moment, and then go
16	on and talk about the application of refusals to deal.
17	As you know, there are two standards that people
18	have been talking about, what I call the consumer
19	welfare effects standard, I just want to focus on the
20	fact that that's really the effective price and quantity
21	effect, not some complicated balancing, and then the
22	profit and no economic sense test. I favor the consumer
23	welfare effect test. You know, it's focused on the goal
24	of antitrust, it's flexible, it is an enquiry meet for
25	the case, I agree with Bill on that. It implies a

43

1	tailored structural enquiry for each type of
2	exclusionary conduct.
3	It's not an open-ended balancing of the sort
4	that was suggested in Chicago Board of Trade, but rather
5	there's a series of steps that one must go through and
6	those series of steps differ for different types of
7	exclusionary conduct.
8	For example, I spoke at the -- at this panel the
9	FTC had last month on timber overbuying and so on, and I
10	distinguished between predatory overbuying and raising
11	rivals costs overbuying and depending on the
12	characterization of the conduct, there was a different
13	test that was used.
14	Should be still a different test for predatory
15	pricing, still a different test for refusals to deal,
16	still a different set of tests for exclusive dealing,
17	but all within the umbrella of a focus on consumer
18	welfare and this consumer welfare approach.
19	So, I don't think that the consumer welfare
20	standard leads to balancing. I also don't think it
21	leads to false positives. Indeed the sacrifice test is
22	usually criticized for causing false negatives, but as I
23	discuss in my article, it also causes false positives,
24	and indeed I'll argue that with refusals to deal, the
25	sacrifice standard would be more likely to cause false

44

1	positives than would the consumer welfare test.
2	We've talked a little bit about whether the
3	innovation incentives are a reason to cut back Section
4	2. I'm going to talk about this before we get to
5	refusals to deal, but just basically, you know, firms
6	have incentives to compete, incentives to innovate in
7	competitive markets. I believe it's the consensus of
8	economists that innovation incentives are greater in
9	competitive markets than in monopoly markets,
10	monopolists have weaker innovation incentives than
11	competitors. I would cite you to Mike Scherer's
12	article, which is cited in my antitrust law journal
13	article. And of course, you know, if a monopolist, if
14	the dominant firm knocks the entrants out of business,
15	then it will, of course, reduce the innovation
16	incentives of the entrants as well.
17	Well, now, how would you apply this to refusals
18	to deal? Well, here, you've got the consumer welfare
19	test, we've got the first -- the profit sacrifice, or
20	NES test, and then of course per se legality. What I
21	want to say about this is that the consumer welfare test
22	and the sacrifice test actually have a lot of
23	similarities. They both require a price benchmark, and
24	a lot of people say the price benchmark is the fatal
25	flaw in anything other than per se legality. I'm going

45

1	to explain why I don't think that's true. And I'll also
2	talk about why I think the sacrifice test is more likely
3	to lead to false positives, because it does not have any
4	or may not have any anticompetitive effects prong. And
5	of course I say legality leads to false negatives.
6	Okay, so what should the rule be under the
7	consumer welfare test? I'm going to talk about the
8	rule. I have a hand-out, which you can pick up at the
9	break, which sets out the rule I've composed in detail,
10	but we can talk a little bit about that now.
11	There will be basically three pieces to it.
12	First of all you have to show that the defendant has
13	monopoly power, and that would be monopoly power in the
14	input market and actual or likely monopoly power in the
15	output market, so we're talking about a vertically
16	integrated monopolist.
17	You would have to show that the plaintiff has
18	made a genuine offer to buy at or above some benchmark
19	price, and I'll talk in a bit about how you determine
20	that benchmark price. So, this is not a matter of
21	saying that the monopolist has to sell at cost, I'm
22	going to come up with a benchmark that's going to
23	compensate the monopolist adequately, and the plaintiff
24	would have the burden of showing that it made an offer.
25	So, the plaintiff can't go to the court first, the

46

1	plaintiff has to go to the monopolist and try to get the
2	product, and if it fails, and the defendant, you know,
3	refuses to deal, then there is at least potential for a
4	case.
5	This test I use, which I call a compensation
6	test, is going to compensate the monopolist for its lost
7	profits for the customers that it loses to the entrant,
8	and this is very much a sacrifice test, a no economic
9	sense test. But under the consumer welfare analysis,
10	you also require the plaintiff to prove anticompetitive
11	harm. And that would be during the output market, or
12	the input market, or some other -- some other market
13	where the firms are actual or potential competitors.
14	It's not clear to me that the sacrifice standard
15	requires this third step, and that's why I think it's
16	going to lead to false positives. I think it only
17	requires the first two. Now, if you actually parse the
18	literature, Greg Werden probably does not have this
19	third step. He has some type of incipiency standard for
20	the third step. I think Doug Melamed, I think, adds
21	this third prong.
22	In which market do I have to show
23	anticompetitive effects? Well, that's going to depend
24	on the case. But, you know, a refusal to deal could
25	cover up, you know, a naked noncompete. For example,

47

1	you know, a contemporary example might be suppose
2	Halliburton, which has a monopoly over certain
3	transportation services in Iraq, suppose it says to a
4	firm, I will only provide you transportation services in
5	Iraq which you need in order to sell other commodities
6	to the armed forces, I will only provide that input to
7	you if you promise not to compete with me in providing
8	oil field services in Louisiana.
9	Well, that's a refusal to deal, the harm would
10	not be in the geographic market in whatever Halliburton
11	competes in in Iraq, but rather some other unrelated
12	market. So, it's possible that this litigation could be
13	brought here.
14	Or, you know, more generally, if it's not the
15	input or output market, it's going to be a complementary
16	product, it's going to be a complementary product
17	market.
18	So, notice, this consumer welfare test, it's not
19	an open-ended Chicago Board of Trade inquiry, have to
20	show market power, have to show anticompetitive effects
21	in a particularized way, and you have to show that the
22	price offered by the plaintiff meets the compensation
23	test.
24	Okay. Well, the real issue is, what about this
25	price benchmark? This is where the controversy is. And

48

1	there are several candidates, as Hew pointed out.
2	There's the prior price paid by the plaintiff, as in the
3	case of Aspen. It could be the price charged to other
4	buyers, which also was an issue in Aspen, where they
5	were willing to deal with other mountains in other ski
6	resorts. Or there could be some benchmark, if the first
7	two don't work, either because there's no course of --
8	previous course of dealing, or because of some reason
9	they're not appropriate, and I agree with you that they
10	may not be appropriate, then you need another benchmark
11	and the benchmark that I've come up with is a benchmark
12	I call protected profits benchmark, and it's a price
13	that compensates the defendant for the monopoly profits
14	lost to plaintiff from losing -- from customers that
15	shift from the defendant to the plaintiff.
16	I'll give you an example. So, it is a sacrifice
17	test, it is giving the defendant the monopoly profits
18	that it's earned, and I think that's a key issue. You
19	might want to adjust this benchmark. For example,
20	suppose dealing with the plaintiff raises the
21	defendant's production costs. Well then you would have
22	to take that into account in setting the benchmark.
23	Suppose the plaintiff creates real reputational
24	free-riding, you know, suppose it says, well, we've
25	used -- we've used this input that we got from GE, and

49

1	suppose their product is no good, and that hurts GE's
2	reputation, well that could would be a reason why GE
3	should be permitted not to deal with them or charge them
4	a higher price.
5	And lastly, suppose the monopoly, we've been
6	acting up until now that these monopolies are attained
7	legitimately. If they're not obtained legitimately,
8	then it's not clear that you want to give someone
9	protection from the monopolist. Not clear that you
10	would worry so much about protecting those monopoly
11	profits or protecting the incentives.
12	Finally, the other adjustment I would make is
13	this is a rule intended to generate negotiation, so if
14	the defendant just has a flat refusal to deal, a
15	non-negotiable refusal to deal, or only makes sham
16	offers, as they did in Aspen, then the burden is going
17	to shift to the defendant to show that the plaintiff's
18	price offer was good.
19	So, for example, in Aspen, it's not as if
20	Highlands said, I'll pay you ten cents for the daily
21	tickets, and Ski Co. said, no, no, no, I want $44,
22	that's much more reasonable, and Highlands said, I'm
23	going to sue you. It wasn't like that at all. In fact,
24	Highlands made an offer, in fact the retail price, but
25	Ski Co. made a counteroffer designed for Highlands to

50

1	turn down. I mean, it was not a real counteroffer, it
2	was one that Highlands would be forced to reject. So, I
3	place some burden on the defendant in those
4	circumstances.
5	Okay, so how do you calculate this? Well, this
6	is the part with the math, but as I tell my law
7	students, this is not really math, it's just shorthand,
8	it's just abbreviations. So, my benchmark has two
9	important properties to it. One is it compensates the
10	defendant for the monopoly profits that it loses on the
11	customers that it loses to the plaintiff. However, it
12	does not get compensation for price competition that's
13	induced by entry by a firm that has lower costs or
14	superior product.
15	So, I'm compensating them for their monopoly
16	profits they have, but I'm not allowing them to deter
17	entry by a more efficient competitor, one that has lower
18	costs or a better product. Where did I get the standard
19	from? Well, I didn't invent it. This goes way back.
20	It's called the efficient components pricing standard,
21	first started in the late 70s or early 80s. It's been
22	referred to in the context and there's been a lot of
23	commentary on this basic standard by people, among
24	others, John Vickers, who just left heading up the OFT
25	in Europe.

51

1	The way you calculate this, this benchmark
2	price, is the monopolist's input cost, plus it gets its
3	margin, plus its margin times the fraction of the
4	plaintiff's customers that get diverted from the
5	monopolist. This is not -- it's not a lot of letters,
6	it looks like algebra, but it's not really so
7	complicated.
8	So, let me give you an example to show that, and
9	I'll use -- suppose the Trinko case were not in the
10	context of regulation, how would you, you know, how
11	would you use this protected profit standard? Well,
12	here's the data. Suppose Verizon's incremental cost of
13	providing DSL, wholesale DSL, suppose that were $10.
14	Suppose Verizon's margin on retail DSL, their monopoly
15	margin, suppose that were $50. And suppose that if
16	Verizon sells wholesale DSL to AT&T, half the customers
17	AT&T gets will come out of the hide of Verizon, and the
18	other half will come from cable and dial-up. And yes, I
19	know Verizon provides dial-up in its own territory, but
20	they probably don't make much money there, so I am just
21	leaving that out for now. But if you will, you could
22	make it more complicated to take into account the
23	dial-up margin, but I think Verizon probably sells at a
24	negative margin on dial-up anyway.
25	So, under these circumstances, half of AT&T's

52

1	retail DSL customers are going to come out of Verizon,
2	half are going to come out of the hide of Comcast, Time
3	Warner and so on. So, this diversion rate would be 50
4	percent. Diversion rate, you know, it's something we
5	use in mergers all the time.
6	What would be the benchmark price? It would be
7	$35. Verizon's $10 cost, plus they get a monopoly
8	margin of $50, they lose that monopoly margin on half
9	their customers, so half of $50 is $25, you have to
10	compensate them for those expected losses, that gives us
11	$35. Okay?
12	If AT&T were going to get all its customers out
13	of the hide of Verizon, then the benchmark would be a
14	lot higher, it would be $60, Verizon would have to be
15	compensated for its costs, plus the margin that it lost.
16	Okay? Not so difficult to do this at all.
17	Under this standard, and this is another sort of
18	key aspect, I probably should have put it on the
19	previous slide. The entrant will not be able to succeed
20	in the market under this standard, unless it has lower
21	costs or a superior product for at least some consumers.
22	So, this is not a prescription for inducing inefficient
23	entry, the only kind of entry that gets induced as a
24	result of this test is efficient entry, and therefore I
25	think it meets the -- I think it meets the standard.

53

1	So, for that reason, I think this, you know,
2	this consumer welfare standard, look at how much the
3	plaintiff has to prove. Monopoly power in the input
4	market, you know, if the entrant's got an alternative,
5	then they're out. The defendant has to have actual or
6	potential monopoly power in the output market, or else
7	the plaintiff loses.
8	A lot of things for plaintiffs to prove. It's
9	got to prove that the price offered exceeds the test, a
10	test that I don't think is very difficult for a firm,
11	certainly not a firm like Verizon, to calculate. I
12	don't think it's hard for any firm.
13	This is the same sort of data we routinely use
14	for merger analysis, and that a firm needs to run its
15	own business. A firm needs to know its margin. And in
16	fact, it can look up its margin, it can ask the CFO for
17	their margin, it's on the profit and loss statement and
18	should be on the profit and loss statement for each
19	division. And they just need to know the extent to
20	which they compete with the plaintiff.
21	And the plaintiff here also has to prove
22	anticompetitive effects. So, there's big barriers for
23	the plaintiff here. So, this is not -- this is not a
24	standard that's going to lead to overwhelming amount of
25	litigation.

54

1	Now, this is the standard, how do we deal, what
2	do we have to say about Trinko? Well, Trinko raises a
3	number of cautions that have been discussed by the
4	earlier speakers. They pointed out that there's no
5	general Sherman Act duty to deal, and they said forced
6	share, I guess red flags is my term, the justice
7	division did not use the term red flags, but it raises a
8	number of red flags. Lessens investment incentives,
9	requires courts to act as central planners, that's the
10	red flag. And the compelling negotiation can facilitate
11	collusion. All of this adds up to the concern with
12	false positives.
13	Well, let me go through these and look at these
14	in a little more detail. Well, first of all, the no
15	general Sherman Act duty to deal, that's true. I teach
16	antitrust, every antitrust professor knows that. I wish
17	that in the Trinko opinion, however, they had quoted
18	Colgate correctly. They said Colgate stands for no duty
19	to deal. The proper quote says, i.e., in the absence of
20	any purpose to create a monopoly, there's no duty to
21	deal. So, Colgate is limited and in that Justice Scalia
22	tried to change the meaning of Colgate.
23	So, what about these more detailed questions?
24	Well, first is this investment incentives, this has been
25	alluded to by several speakers. I think the first

55

1	point, the key point is the benchmark price compensates
2	the defendant for the monopoly profits that it loses on
3	customers that it loses to the plaintiff. So, in terms
4	of reducing their investment incentives, we're making,
5	and I thought Hew was exactly right, it is a liability
6	standard. It's making them whole on the profits they
7	lose, on the customers that they would lose to the
8	plaintiff.
9	But there's other reasons why I think it will
10	not reduce investment incentive. First of all, Scalia
11	worries about reducing the entrant's investment
12	standards, that the entrant would otherwise enter the
13	input market on its own. But that is a very weak
14	statement. I mean, you don't get into one of these
15	cases unless the defendant's got monopoly power in the
16	input market, and what we mean by monopoly power is
17	durable monopoly power. What we mean by durable
18	monopoly power is that there are high barriers to entry.
19	So, unlikely that the plaintiff otherwise would
20	have entered the input market. It also means you can't
21	get into the -- you can't enter one market at a time,
22	you're unlikely to see leapfrog competition. Secondly,
23	we know the competitive markets increase the defendant's
24	innovation incentives. Monopolists have weaker
25	innovation incentives than do competitors and, you know,

56

1	I mean, the telephone companies have a million excuses
2	for why they never innovate, and we have just heard some
3	others.
4	I think that -- but I think if they had faced
5	more competition, they would have stronger innovation.
6	They are certainly innovated in trying to come in to
7	compete with cable, where they don't have -- where
8	Telecom is not -- where telephone companies do not have
9	a monopoly.
10	Of course entering the output market will
11	increase the entrant's innovation incentives. And
12	finally, and this is I think a key point, and I think in
13	Bill Kolasky's list of cases, Kodak was conveniently
14	left out. In Trinko, Kodak doesn't get mentioned.
15	Well, one very important point that was made in the
16	Kodak opinion is that you can't call the entrant a free
17	-rider if they only enter one market rather than all of
18	them.
19	Kodak says that this understanding of
20	free-riding is an argument made by -- made by Kodak, and
21	the Supreme Court said, this understanding of
22	free-riding has no support in the case law. So, you
23	know, I think that argument just does not add up.
24	The courts as central planners, I'm running out
25	of time, so let me go quickly. You know, I guess the

57

1	point I've been making all along is this isn't so hard.
2	Market prices often provide a good benchmark. I think
3	this protected profits compensation benchmark is not too
4	difficult to evaluate, and then the other point I want
5	to make here is, you know, if antitrust withdraws, it's
6	not clear that we're going to have laissez faire. This
7	has not been the way the United States economy has
8	worked.
9	When antitrust fails, we often get real formal
10	public utility commission regulation, real central
11	benefits, and so I just want to raise the question about
12	whether we're really going to get ourselves into the
13	federal operating system commission if antitrust drops
14	out. And of course the essential facility doctrine fits
15	in here.
16	Okay, finally is this issue about facilitating
17	collusion. I think that one is really silly. You know,
18	if you believed -- if you believed this argument that
19	letting people negotiate is going to facilitate
20	collusion, well then we also prohibit voluntary dealing,
21	we also prohibit joint ventures, we also prohibit patent
22	settlements, which we know from the FTC experience are
23	sometimes used to strike noncompetition agreements.
24	It's also, you know, the refusal to deal can be
25	used, if it's a threatened refusal to deal, can be used

58

1	to facilitate collusion. I'll sell to you, but only if
2	you promise not to compete with me. So, I think that
3	the -- that effect put out that dicta by the Trinko
4	court was really they -- it's either insignificant or
5	goes the other way.
6	Finally, I want to raise the question of if we
7	go down Hew's route for per se legality, where are we
8	going to stop? I note that's perhaps not a question
9	that Hew is worried about, but it's a question that I'm
10	worried about. If it's per se illegal -- per se legal
11	to refuse to deal with firms that compete with you, then
12	what about exclusive dealing? Why isn't that, per se,
13	legal, either with respect to whether if the firm wants
14	to buy stuff from you, sell it to your competitors, or
15	if they want to buy from your competitors? What about
16	the tie-in? Why doesn't it make tie-in per se legal,
17	because that's just basically refusal to deal. What
18	about noncompetition agreements? What if a firm says,
19	like in my little Halliburton example, we're going to
20	compete with you in some unrelated market, and they say,
21	well, in that case, I'm not going to sell to you. Well,
22	that would be -- that would be per se legal.
23	And finally, what if they use a refusal to deal
24	in order to force the firm to raise prices, either in
25	the market -- the output market that we're talking about

59

1	or some other market. Would that also be per se legal
2	for them to make that argument? So, I would be quite
3	concerned about that.
4	I'm out of time, thank you very much.
5	(Applause.)
6	MR. ABBOTT: Thank you, Steve, for presenting an
7	attempt to establish an administrative rule that will
8	undoubtedly bring forth some more discussion about the
9	rule that might apply in evaluations under the rule of
10	reason.
11	Now we have another economist who is going to
12	take a crack at this difficult set of topics. Tom
13	Walton, director of economic policy analysis, General
14	Motors Corporation, in which position he oversees the
15	analysis of costs, current and prospective governmental
16	policies and regulations, and their implications for
17	General Motors. Tom Walton received a Ph.D. in
18	economics from UCLA, was assistant professor at NYU,
19	before joining GM, and served briefly as special advisor
20	for regulatory affairs at the FTC. He's vice chair of
21	the Business Research Advisory Counsel for the U.S.
22	Bureau of Labor Statistics in Washington, D.C.
23	Tom?
24	(Applause.)
25	MR. WALTON: Thank you very much. I'm going to

60

1	try a little bit of a change of pace to give you an idea
2	of what it's like to be inside the fish bowl of
3	competition.
4	Well, it all began back in 1963 when the Federal
5	Trade Commission launched its first investigation into
6	the manufacturing and distribution practices of the
7	major auto makers with regard to the production and sale
8	of their single source crash parts. Now, these are the
9	parts that are most frequently damaged in the event of
10	auto accidents, and which also happen to be single
11	source. They include radiators, bumpers, fenders,
12	grills, all the sheet metal. They don't include glass,
13	because glass is multiple source.
14	At that time, Chrysler, Ford and GM, the major
15	manufacturers at that time, distributed these parts
16	exclusively through our franchised auto dealers. Our
17	franchised line-make auto dealers. That's an important
18	distinction. For example, Chevrolet parts we
19	distributed exclusively through Chevrolet. If an
20	independent body shop wanted to buy a part, it could
21	only get a Chevrolet brand part at Chevrolet, they could
22	not get it at Pontiac, for example.
23	Insurance companies instigated the
24	investigations. Congressional investigators had been
25	constantly pressing them to reduce their auto insurance

61

1	premiums. Insurance had a pretty good handle on the
2	labor rate at the auto shops, both at the auto dealers
3	and the independents, but they wanted to set up
4	independent warehouse distributors or wholesale
5	distributors so they could get similar concessions on
6	parts. They brought along with them the lobbying arm of
7	the independent body shops, or IBSs, as they called
8	themselves. They complained that GM and other auto
9	manufacturers, everyone used the same system at the
10	time, were discriminating against them because they --
11	because in the case of the independent body shop, they
12	had to buy the part from the dealer at a mark-up, or
13	have the dealer provide the part directly from the
14	manufacturer, General Motors or another manufacturer at
15	wholesale.
16	Of course, the auto dealers, like any other
17	retailer, have the wholesaling cost. They have the cost
18	of ordering, carrying, insuring and financing the
19	distribution of the parts. And of course they charge
20	for those wholesaling services. So, the IBSs, the
21	independent body shops and insurers went to the Congress
22	and went to the Federal Trade Commission to try to force
23	us to directly sell those parts, those single-sourced
24	crash parts to the body shops and to the independent
25	wholesalers.

62

1	Little interest was expressed by the large
2	warehouse distributors, and later they would testify
3	that they had no interest in taking on the business.
4	They also believed that there was no need to take on
5	additional wholesalers, additional customers. There was
6	no shortage of GM dealers to handle the business.
7	There's something like 12,000 dealers spread out in
8	every area of the country. They thought they could do
9	the best job of handling the bulky and complex repair
10	parts because in part, they shared our incentive to keep
11	the customer happy and make sure that the owner of a
12	Chevrolet vehicle was put quickly and efficiently back
13	on the road.
14	Sure, they shared our interest in the integrity
15	of the brand name. We believe that opening up the
16	system to tens of thousands of independent body shops
17	would reduce the availability of the parts and increase
18	the time necessary to get them to the customer. We knew
19	it would impose substantial additional administrative
20	and monitoring costs. We didn't feel we could derive
21	the monopoly profits from pricing the parts, because we
22	would be jeopardizing 95 percent of our business, that's
23	the vehicle business, by trying to achieve a monopoly on
24	the parts.
25	Higher priced parts would have meant driving up

63

1	the repair costs for our customers, and would have
2	reduced the likelihood that a Chevrolet vehicle owner
3	would become a repeat customer. We knew that one
4	company, Renault, had recently ceased doing business in
5	this country because of a faulty service repair system.
6	Another company, another competitor, Chrysler, had spent
7	something like $350 million to convert from the system
8	the FTC was proposing, this open warehousing, open
9	distribution system, back to the system of distributing
10	the parts exclusively through its franchised dealers.
11	We did offer subsidies for GM dealers to sell
12	the parts to the independent body shops at reduced
13	prices. In order to pacify them and to pacify the
14	Federal Trade Commission, in September 1967, we proposed
15	a plan in which we would offer a 12 percent discount on
16	the parts resold through the independents. A program we
17	then called wholesale compensation.
18	In February of 1968, the Commission, though,
19	told us that they intended to file a lawsuit in order to
20	bring about price parody between the GM dealer body
21	shops and the independent repair shops. Further
22	negotiations ensued and in the fall of 1968, the
23	Commission accepted our proposal to raise that subsidy,
24	that incentive for reselling to 23 percent. Accordingly
25	we increased our prices on all crash parts in order to