Understanding Single Firm Behavior : Monopoly Power Session

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	UNDERSTANDING SINGLE-FIRM BEHAVIOR:
9	MONOPOLY POWER SESSION
10	WEDNESDAY, MARCH 7, 2007
11
12
13
14
15	HELD AT:
16	UNITED STATES FEDERAL TRADE COMMISSION
17	601 NEW JERSEY AVENUE, N.W.
18	WASHINGTON, D.C.
19	9:30 A.M. TO 4:30 P.M.
20
21
22
23
24	Reported and transcribed by:
25	Susanne Bergling, RMR-CLR

2

1	MODERATORS:
2	DENNIS W. CARLTON
3	Deputy Assistant Attorney General for Economic Analysis
4	Antitrust Division, Department of Justice
5	and
6	JOEL L. SCHRAG
7	Economist
8	Bureau of Economics, Federal Trade Commission
9
10	PANELISTS:
11
12	Morning Session:
13	Andrew J. Gavil
14	Richard J. Gilbert
15	Michael L. Katz
16	Philip B. Nelson
17	Joseph J. Simon
18	Lawrence J. White
19
20	Afternoon Session:
21	Simon Bishop
22	Thomas G. Krattenmaker
23	Miguel de la Mano
24	Joe Sims
25	Irwin M. Stelzer

3

1	C O N T E N T S
2
3	MORNING SESSION:
4	Introduction
5	Presentations:
6	Andrew J. Gavil
7	Richard J. Gilbert
8	Michael L. Katz
9	Philip B. Nelson
10	Joseph J. Simons
11	Lawrence J. White
12	Moderated Discussion
13	Lunch Recess
14
15	AFTERNOON SESSION:
16	Introduction
17	Presentations:
18	Simon Bishop
19	Thomas G. Krattenmaker
20	Miguel de la Mano
21	Joe Sims
22	Irwin M. Stelzer
23	Moderated Discussion
24	Conclusion
25

4

1	P R O C E E D I N G S
2	- - - - -
3	MR. SCHRAG: Good morning. Sorry about the
4	technical issues. Welcome.
5	My name is Joel Schrag. I am an economist at
6	the Bureau of Economics here at the Federal Trade
7	Commission, and I am one of the moderators for this
8	panel. My co-moderator, standing next to me, is Dennis
9	Carlton, Deputy Assistant Attorney General for Economic
10	Analysis at the Antitrust Division of the Department of
11	Justice.
12	Before we get into the substance of the program,
13	on behalf of the FTC staff who have worked on this
14	session, I would like to take the opportunity to thank
15	all of our colleagues from DOJ for their hard work and
16	their efforts to jointly present this session.
17	In addition, after today's and tomorrow's
18	hearings on monopoly power, the hearings will next turn
19	to issues involving remedies later this month, and so I
20	urge you all to be sure to check our agencies'
21	respective web sites for updates on these future
22	hearings.
23	As the FTC representative, I do have just a few
24	housekeeping matters to cover before we begin. First of
25	all, please turn off all of your cell phones,

5

1	BlackBerries and other noise-making electronic devices.
2	Second, the restrooms are located out through the double
3	doors and across the lobby. If you need help to find
4	them, there are signs that should guide you.
5	Third, one safety tip, especially for visitors,
6	in the unlikely event that the building alarms go off,
7	please proceed calmly and quickly as instructed. If we
8	must leave the building, you exit out the New Jersey
9	Avenue doors by the guard station. Please follow the
10	stream of FTC employees to a gathering point across the
11	street and await further instruction, but hopefully that
12	won't be necessary.
13	Finally, we request that you please not make
14	comments or ask questions during the session. Thank
15	you.
16	Let me just say a few things about the session.
17	Many of the prior sessions of the hearings addressed
18	particular conduct that's been challenged under Section
19	2 of the Sherman Act. Today, the hearings turn to
20	issues of monopoly power and market definition, and
21	these issues we believe are very important.
22	In fact, if you were at the opening day of the
23	hearings back in June, both Herbert Hovenkamp and my
24	co-moderator, Dennis Carlton, were given the opportunity
25	to place the issues for the subsequent hearings in

6

1	context. Both identified monopoly power and market
2	definition as areas where there are difficult, uncertain
3	questions that must be addressed in many cases, and I
4	expect that today's panel will help to clarify, if not
5	completely resolve, these difficult questions.
6	The hearings will be organized as follows:
7	First, we'll hear an approximately 15-minute
8	presentation from each of our six distinguished
9	panelists. We'll probably take a break after the fourth
10	panelist and then come back from the break and hear from
11	the two remaining panelists. After that, the panelists
12	will have an opportunity to comment on each other's
13	presentations, and we'll have a moderated discussion.
14	So, I think I'd now like to turn things over to
15	my co-moderator, Dennis Carlton, who will introduce our
16	distinguished panelists.
17	Thank you very much.
18	DR. CARLTON: Okay, thank you. I am Dennis
19	Carlton. I am a Deputy Assistant Attorney General in
20	the Antitrust Division, and it is a pleasure to welcome
21	all of you to these joint FTC/DOJ hearings.
22	I had the privilege of participating in the
23	opening session of the hearings, and one of the topics I
24	said that needed clarification was precisely the topic
25	of the panels today and tomorrow, a focus on what we

7

1	mean by "market power" and "market definition" in
2	Section 2 cases I think is really important.
3	I am also a Commissioner on the Antitrust
4	Modernization Commission, and despite my attempting to
5	do so was not able to convince the Commission to study
6	in depth the definition of market power and market
7	definition in Section 2 cases and to report on it. So,
8	that, I think, emphasizes all the more how important
9	this session, this panel discussion, is today, and the
10	real question is, can we reach consensus on any of the
11	hard questions or at least can we reach a consensus that
12	there's a lot of ambiguity and arbitrariness in what is
13	going on?
14	I am honored to chair such a distinguished
15	panel. All of the members of the panel have extensive
16	experience, both academic and nonacademic, in antitrust
17	and have served both in the private sector and in the
18	government sector.
19	In the interest of saving time, I am going to
20	introduce them all at once and hopefully by that time
21	the computer will work. So, starting with Phil, Phil
22	Nelson is a principal at Economists, Inc., an economic
23	consulting firm. Previously, he served as the Assistant
24	Director for Competition Analysis at the FTC and as an
25	Adjunct Professor at Fordham Law School. He has written

8

1	numerous articles and two books on antitrust topics, and
2	he edited the ABA's antitrust section of market power --
3	The Market Power Handbook. He currently is the
4	vice-chair of the section's Healthcare and
5	Pharmaceuticals Committee.
6	Beside Phil is Joe Simons. Joe is a well-known
7	attorney. He's a partner and co-chair of the antitrust
8	group at Paul Weiss. Previous to that, Joe was the
9	chief antitrust enforcer at the Federal Trade
10	Commission, serving as the Director of the Bureau of
11	Competition from June 2001 until August of 2003. He has
12	the interesting characteristic of once being the tenth
13	largest wireless carrier in the country, because I
14	believe he was a trustee and had a lot of wireless
15	licenses, but in addition to that, he has achieved
16	something that's actually quite rare for attorneys to
17	do, and that is he's written an article that economists
18	cite all the time and is associated with critical loss
19	analysis.
20	Beside Joe, in a missing seat, is Larry White,
21	who I am sure is on his way. Larry is the Arthur
22	Imperatore Professor of Economics at NYU School of
23	Business. He's the Deputy Chair of the Department of
24	Economics. Previously, in the early eighties, Larry
25	served as the Director of the Economic Policy Office in

9

1	the Antitrust Division. Larry has written several books
2	and articles, one of which is well-known to antitrust
3	practitioners called The Antitrust Revolution:
4	Economics Competition and Policy. He's currently the
5	editor of The Review of Industrial Organization. Prior
6	to serving at the Justice Department, he did extensive
7	government service both for the Federal Home Loan Bank
8	Board and for the Council of Economic Advisers.
9	Andy Gavil is a Professor of Law at Howard
10	University where he not only teaches antitrust, but he
11	has also extensively written on antitrust many articles
12	and has a very well-known case book with Bill Kovacic
13	and Jonathan Baker, Antitrust Law in Perspective. He is
14	about to publish or co-author a book called Microsoft
15	and the Globalization of Competition Policy, which I am
16	sure has focused on Section 2 type behavior. He's
17	currently the articles editor of The Antitrust Magazine
18	and serves on the ABA Antitrust Section's Liaison Task
19	Force to the Antitrust Modernization Commission. He is
20	Of counsel to the Sonnenschein Law Firm.
21	To Andy's left is Rich Gilbert. Rich is a
22	Professor of Economics at the University of California
23	at Berkeley. He served as the Deputy Assistant Attorney
24	General in the Antitrust Division in the mid-nineties,
25	and at that time, he led the effort to write the

10

1	Antitrust Guidelines for the Licensing of Intellectual
2	Property. He has written widely on antitrust topics.
3	He is currently the Director of the Competition Policy
4	Center at Berkeley and is associated with the economic
5	consulting firm of COMPASS.
6	Finally, Mike Katz at the end of the table.
7	Mike is currently the holder of the Sarin Chair in
8	Strategy and Leadership at Berkeley, the Business
9	School, and also holds an appointment in the Economics
10	Department. Mike served as the Deputy Assistant
11	Attorney General in the early 2000s, and he also served
12	as the Chief Economist at the Federal Communications
13	Commission. He's written numerous articles on economics
14	and antitrust and has specialized in many topics,
15	including network industries.
16	So, with that introduction, I'll turn it over to
17	our first speaker, Phil, and just let me remind the
18	speakers, we're kind of running tight because we started
19	late, so if you could keep to the 15 minutes, that would
20	be good. The organization of this is going to be four
21	speakers will go, 15 minutes, we'll take a 10-minute
22	break, we'll have two more speakers. We will give the
23	speakers a brief opportunity to talk to each other, and
24	then I'll moderate a discussion for about an hour or so.
25	Thank you.

11

1	MR. NELSON: So, we have a -- are we moments
2	away or should I just proceed without slides?
3	DR. CARLTON: Is the computer still not working?
4	MR. NELSON: Well, okay, the reason they put me
5	first is the slides that you can't see are really sort
6	of a background deck that gives you the background on
7	market power. The first slide cites the definition of
8	market power that's at the front of the monograph that
9	the ABA published that was referred to earlier, which is
10	market power is the ability of a firm or a group of
11	firms within a market to profitably charge prices above
12	the competitive level for a sustained period of time,
13	and as you can't see on the screen, the word
14	"profitably" is in italics, and so one of the important
15	things in the definition is that a monopolist profit by
16	doing this.
17	If entry is easy, you may be able to raise
18	prices, but not profitably, because somebody will enter,
19	and if there are a lot of competitors, they can steal
20	customers away from you, so you can't profit. That may
21	become of importance in some of the discussion as to
22	what type of performance evidence one might use in
23	determining whether a firm has market power or not.
24	A price above the competitive level, the
25	"competitive level" was in italics, because people talk

12

1	about the standard monopoly raising prices, and if you
2	are not raising them above the competitive level,
3	usually people don't care.
4	Then a "sustained period of time" is in the
5	definition because you may be able to opportunistically
6	raise prices for a little bit, but again, entry or
7	something might undermine the ability to do that.
8	Now, in some of the legal cases, you see
9	reference to the ability to exclude competition, and I
10	will suggest that is something worth consideration,
11	because in some contexts -- and there were FTC hearings
12	many years ago about standard-setting organizations
13	where there might be a collection of, let's say, 10 or
14	more people making a particular product, and there might
15	be enough competitors that they compete and charge a
16	competitive price because there's so many people
17	operating under that standard.
18	Well, somebody may develop a new technology that
19	would come in and completely take the market away from
20	the incumbent competitors with the older technology.
21	Acting jointly in that case, they might be able to block
22	entry by controlling the standard-setting organization.
23	Are they raising prices above the competitive level?
24	They're excluding an entrant, somebody that would
25	dynamically help the market with a new technology that

13

1	might have better performance characteristics and be
2	able to be sold at a lower price. What they get out of
3	it is where their profit is, is that they get to earn
4	the competitive rate of return rather than being maybe
5	in bankruptcy court.
6	So, while I gave you the standard definition,
7	there are other things and other contexts, as you can
8	see from the get-go, that you have to worry about in
9	deciding whether a firm or a group of firms have market
10	power. And today, largely we'll focus on dominant
11	firms, but there are contexts where a group of firms
12	acting together might have trouble. And if a dominant
13	firm has control over a patent that's a blocking patent
14	that blocks a new technology, he might have an interest
15	in blocking the new technology just like the group of
16	firms that ran the standard-setting organization has an
17	incentive to block technology. So, that's one thing.
18	The other thing that I wanted to highlight at
19	the beginning is, some people talk about market power;
20	some people talk about monopoly power. Often,
21	economists mean the same thing, but in some contexts,
22	people have defined them differently. Greg Werden is
23	sitting there, and he's drawn a distinction in one of
24	his articles and alludes to other people that
25	distinguish market power and monopoly power perhaps in

14

1	terms of the time period over which people have the
2	ability to raise prices and the like.
3	There are articles out there that talk about
4	antitrust monopoly power, again, trying to make a
5	distinction. And there, often the thought is if you
6	have a differentiated product and thus have a
7	downward-sloping demand curve for your product, you
8	might have some degree of ability to raise prices above
9	costs and you might in that sense have market power, but
10	you might not have a substantial ability to do it.
11	Because there are a lot of products out there that are
12	roughly close substitutes, not exactly the same thing,
13	and you might in that context have some market power but
14	not antitrust monopoly power or antitrust market power,
15	because you don't really have substantial ability to
16	earn substantial profits and the like. So, some people
17	try to distinguish that downward-sloping demand curve
18	idea by talking about antitrust monopoly power.
19	I think with that background, we're talking
20	about antitrust market power. Something that's somewhat
21	significant. And then different panelists may have
22	different degrees of market power in mind when deciding
23	how you go about measuring whether it is significant
24	enough market power. So, with that sort of definition
25	of market power, the next slide was going to lay out

15

1	sort of the touchstones in a typical market power proof
2	that you sort of run through, and the first thing people
3	often define is product market definition. Then they go
4	to defining geographic market definition.
5	Once you have a relevant product/geographic
6	market combination, often it is standard to look at
7	market concentration in a monopoly case, and once you
8	clear that hurdle and see that maybe it is substantially
9	concentrated or a firm has a dominant market share, a
10	high-level market share, you then start looking at
11	things like entry conditions, other structural
12	characteristics of the market. Maybe you look at in
13	some contexts, you know, the structure of the buyer-side
14	of the market, and if it is a collusion case type of
15	monopoly power issue, maybe you look at the
16	characteristics of the market that make it easier or
17	harder for firms to collude in that market.
18	Then finally, in a lot of the monopolization
19	cases, you see a consideration of market performance
20	evidence, and that's where you start having things like
21	profit rates of return, profit margins, looking at
22	prices over time or across geographic areas. You look
23	at output patterns and how they vary with prices. And
24	you look at new product introductions. You can either
25	look at them in terms of formal econometric analysis or

16

1	often you look at events -- market events that allow you
2	to sort of control for some things -- and look at how if
3	the events give you insights either directly into the
4	market power or at some of the related issues like
5	market definition.
6	Now, increasingly, because of the success of the
7	Merger Guidelines, you see references to the approach
8	used in the Merger Guidelines of developing a relevant
9	market in the context of monopolization cases, and there
10	were a couple slides that sort of just quoted the
11	Guidelines. I suspect with this audience, there is no
12	reason to go through it, but it is the hypothetical
13	monopolist test. Can the monopolist raise prices above
14	the -- in the Guidelines, they talk more or less about
15	the current level as opposed to the competitive level
16	and see if that's profitable.
17	Now, one thing that is worth pointing out,
18	especially in transferring that concept, is that in the
19	Guidelines themselves, Section 1.11 says that while you
20	might look at prevailing prices in the Guidelines, there
21	is a caveat that says if pre-merger circumstances are
22	strongly suggestive of coordinated interaction, in that
23	situation, the agency will use a price more reflective
24	of the competitive price. So, there is a caveat in
25	there where they don't always use prevailing prices.

17

1	One sort of footnote is that the original
2	guidelines were focused on coordinated effects, and then
3	they later on added more information about unilateral
4	effects. I think there's a little glitch here, because
5	I think the Merger Guidelines actually should make a
6	reference not only to coordinated interaction, but also
7	if the dominant firms raise prices above the competitive
8	level, then you might want to look at the competitive
9	price level.
10	Why might you want to do that? Well, that is
11	because you get a different elasticity and different
12	substitutes depending on at what price level you measure
13	the substitution. And this is where the lack of slides
14	really hurt us the most, because I put together an
15	illustrative example of a demand curve with a concrete
16	slope and all the rest, calculated the marginal revenue
17	curve from that, showed where the competitive price
18	would be, where basically price equals marginal cost,
19	then showed where the monopolist would operate, which is
20	at a higher price, and then estimated the elasticities
21	of a couple of the different points along the demand
22	curve.
23	What you see is that even though a demand curve
24	is a straight line and thus the slope is constant over
25	the whole curve, the elasticity changes. And at the

18

1	higher prices, the demand is more elastic. And, the
2	reason that that makes sense is that a monopolist is
3	going to keep raising its price, you know, and find a
4	price that is more profitable. And in the monopolistic
5	equilibrium, he has got a high enough price that demand
6	becomes elastic and a further price increase would lose
7	a lot of customers to other products. That is called
8	the Cellophane fallacy -- that sets up the Cellophane
9	fallacy, which is if you measure the elasticities at the
10	monopoly price, you are going to run into problems
11	because there are a lot of substitutes out there that
12	are not substitutes at the competitive price. You can
13	do all the econometrics you want and estimate the
14	elasticities, but if you do not know whether you were at
15	a competitive price or a monopoly price, that elasticity
16	estimate does not tell you anything when you are doing a
17	monopolization case particularly.
18	So, then you get into this tautological
19	situation. If you think about the paradigm of starting
20	with a monopoly case and saying, "Well, do I have a
21	monopoly here?" And you have to define the market, and
22	you have to define a monopoly price to define the
23	market, then why bother defining the market? So, you
24	have got a couple of issues here that suggest, what do
25	you do about it? And the rest of my deck talks about

19

1	the sorts of things that one might look at. But, the
2	basic thing that I wanted to suggest is -- while I think
3	there are great problems with a simplistic analysis of
4	the standard paradigm I outlined -- I think there are
5	elements of it that, if you can go through it all, can
6	help you in many circumstances unravel this thing and
7	cross-check your conclusion.
8	So, it is a way of organizing your story.
9	Making sure that you look at your story or your analysis
10	as consistent, and that it gives you insights into what
11	you might look at. And where it leads you, I think, is
12	looking more and more at some of the performance
13	evidence. But you have got to be careful in looking at
14	the performance evidence, because as economists have
15	shown, things like profits and accounting data are
16	tricky.
17	Having said that, I also think that how
18	difficult a problem it is varies a lot from market
19	circumstance to market circumstance. I think it is
20	probably trickiest when you are dealing with
21	consumer-differentiated products, like Cellophane
22	wrapping paper or something. It may be less of a
23	problem when you are dealing with an input into an
24	industrial process, where you can look at substitutes in
25	a more maybe engineering approach type of way.

20

1	My time is basically up, so to keep us on
2	schedule, I would recommend -- they are going to post
3	the slides later on, and they are written in a way that
4	they are readable -- so I suggest you look at the slides
5	for the rest of the story.
6	Thanks.
7	(Applause.)
8	DR. CARLTON: Thank you.
9	Our next speaker is Joe Simons.
10	MR. SIMONS: Thanks, and good morning, everyone.
11	I would like to start out by complimenting the
12	FTC and the Department of Justice in holding these
13	hearings and doing a terrific job. I am really quite
14	encouraged that something really valuable will come out
15	of this.
16	So, one of the first things that happened this
17	morning is the audience was instructed not to ask any
18	questions or make any comments. So, I thought, well,
19	gee, I was planning to hear you violate that restriction
20	right away, but maybe we'll try something a little bit
21	different.
22	Perhaps by a show of hands, who in here would
23	say that the 1982 Department of Justice Merger
24	Guidelines market definition paradigm was the most
25	significant development in market definition in the last

21

1	30 years?
2	So, we have got most of the panelists and maybe
3	half of the audience. That is pretty good for one
4	thing. I would have expected it might have been a
5	little bit higher.
6	But in any case, what that showing would
7	demonstrate is an enormous amount of success for that
8	effort, I think by any standard, and why is that the
9	case? Why were those guidelines on the market
10	definition paradigm so successful?
11	In my view, it is because those guidelines
12	reflected an understanding that the tools of antitrust
13	analysis should be designed for a specific purpose.
14	Previously, you had market definition which was pulled
15	out of the economics literature and it was not designed
16	to do an antitrust analysis. The merger guidelines
17	market definitions was done specifically for that
18	purpose.
19	The other thing that was really important is
20	that the agency, the DOJ in that case, was willing to be
21	out in front of the case law. I think there was a
22	pretty good argument that those guidelines, the market
23	definition therein, did not really reflect the case law.
24	So, I thought it would be useful to do a little case
25	study and talk about first principles and market

22

1	definition.
2	The Guidelines, the Merger Guidelines, were
3	built around the goals defined right in the Guidelines
4	of preventing mergers from creating or increasing market
5	power -- initially through coordinated interaction and
6	then later unilateral effects. And as I said, they
7	geared this market definition specifically to this
8	overall goal of the Merger Guidelines. So, it was
9	designed to identify that universe of firms that were
10	necessary to profitably engage in coordinated
11	interaction or in unilateral effects. Then for the
12	unilateral effects, arguably the analysis could collapse
13	the market definition into the competitive effects
14	analysis. The market definition in the Guidelines is
15	rigorous, it is logical, and it is transparent.
16	Now, sitting here today, 25 years later, and
17	seeing what a success this was, you might forget what it
18	was like when these things were first issued. There
19	were hoots and howls from all sectors of the Antitrust
20	Bar and the academic community. These guidelines were
21	ivory tower nonsense; they were completely hypothetical;
22	they were totally inoperable and just downright
23	impractical; a complete waste of time. These were
24	comments that people made very regularly, and some
25	people even said it was a conspiracy to do away with the

23

1	antitrust laws.
2	There was a little bit of a kernel of truth to
3	some of those complaints, not the conspiracy stuff, but
4	to the practicality of this test. There were a lot of
5	people who saw the initial attempts to implement this by
6	the agencies in the following way. One of the staff
7	lawyers would have a conversation with the customers and
8	say, "Gee, do you think that the sellers in this market
9	could profitably raise price 5 or 10 percent?" You are
10	shaking your head, but I heard people do that, Greg, and
11	the customer has no concept of what it takes for it to
12	be profitable. There is no context to the question.
13	So, there were reasonable criticisms.
14	But what happened is that because the algorithm
15	was rigorous and logical and transparent, it enabled the
16	development of applications basically, tools, to
17	implement this approach, econometric tools. Examples
18	are Baker and Bresnahan and Scheffman and Spiller, Greg
19	Werden as well, something near and dear to me, critical
20	loss. These things did not exist when those guidelines
21	were first issued, and that really was an important
22	lesson to learn, that if you have the right structure,
23	then you have created a platform on which you can build
24	something that really works.
25	So, what does this translate into in terms of

24

1	what we should do for section 2? Well, what are the
2	goals of section 2? What are we trying to accomplish?
3	Is there a consensus? You know, there has been a lot of
4	ink been spilled in relation to the Trinko case, for
5	example. There are differences already between the way
6	the DOJ and the FTC look at this. There is the profit
7	sacrifice test, the no economic sense test; there is the
8	disproportionate harm relative to efficiencies test.
9	So, where does that leave us for market
10	definition? Does that create a problem? Can we rely on
11	what is in the case law? Reasonable interchangeability,
12	what does that mean? How much interchangeability is
13	reasonable? It is basically relying on
14	cross-elasticities of demand. How high does the
15	cross-elasticity have to be? Is that even something you
16	can look at? Can we rely on the Merger Guidelines
17	market definition? Does the hypothetical monopolist
18	paradigm, as applied in the Merger Guidelines, really
19	work for section 2? And one of the issues in section 2
20	is, are we focused on the same phenomenon that we are
21	for section 7?
22	The Merger Guidelines, the Horizontal Merger
23	Guidelines, are basically focused on collusion, an
24	extreme form of which is unilateral behavior. So you
25	are talking about situations in which a group of firms

25

1	is trying to restrict their own output, whereas in
2	section 2, what you are dealing with is a situation in
3	which one firm, the large firm, the dominant firm, is
4	trying to restrict the output of somebody else in most
5	cases and maybe sometimes themselves as well. So, what
6	do we do with all of that?
7	One possible thing to do -- and I am just
8	throwing this out -- would be to come up with a set of
9	goals for section 2, what is the purpose, what are we
10	trying to do, and then work through various scenarios as
11	to what the market definition would be under each of
12	those. So, one potential scenario is, we are going to
13	say that the goal of section 2 is to prevent unilateral
14	conduct that is reasonably likely to significantly raise
15	price or reduce quality. Reasonably significantly, you
16	can come up with other adjectives, number one.
17	Number two, and you are going to focus on
18	conduct that either, A, has no efficiencies, B, has
19	disproportionately low efficiencies relative to their
20	exclusionary effect, or C, would make no economic sense
21	in the absence of exclusionary effect, and potentially
22	D, permits recoupment of the exclusionary conduct. So,
23	kind of a menu from which to choose.
24	Well, one could argue that the first condition,
25	that the unilateral conduct be such that it is

26

1	reasonably likely to significantly raise price and/or
2	reduce quality, may be a necessary condition. That
3	defines the universe in which something bad can happen.
4	If you do not have that condition, then you might be
5	able to say that nothing bad can really happen. So, you
6	can use market definition in that sense, to focus on
7	that aspect as a screen.
8	You then could ask, "Well, gee, would the market
9	definition need to change depending on your choice of 2A
10	through D?" And at least at a first cut, I would say
11	probably not, that these factors relate to what might be
12	considered defenses or separate prongs of the analysis.
13	They would not be necessary to worry about in the first
14	market power screen, where you use market power or
15	market definition as the screen.
16	All right, so what would be the relevant
17	context, then, for measuring profitability of a price
18	increase? Well, obviously the options are before,
19	during or after the execution of the alleged conduct.
20	Well, we are concerned with the price going up as a
21	result of this conduct, so it seems to me you want to
22	focus on whether there might be a significant price
23	increase, whether a significant price increase might be
24	profitable during or after this alleged conduct.
25	Then similarly, if the conduct is already in

27

1	place, so you cannot observe it over time, then the
2	question might be the reverse, which is, absent this
3	conduct, would the price be lower, right?
4	You see, I think there is the same problem here
5	that you have -- not really a problem, but an issue in
6	the Merger Guidelines -- where for the most part, you
7	are measuring the profitability of a price increase
8	going forward. You are not looking at the current
9	level. You are really looking at a change in the
10	current level that is brought about by the conduct that
11	you are worried about. So, in the merger case, it is
12	the merger; in this case, it would be the alleged
13	exclusionary conduct.
14	You know, one of the things that is near and
15	dear to me, critical loss, might be a tool to help in
16	this analysis, and it would not be exclusive by any
17	means. Just like in the Merger Guidelines you can use
18	critical loss, you can use all kinds of other estimation
19	techniques, and they are not exclusive.
20	So, one way to think about this would be that
21	the burden would be on the plaintiff to show the likely
22	extent to which the alleged conduct restrains
23	third-party producers; in other words, whatever the
24	conduct is, exclusive dealing, refusal to deal,
25	whatever, what is the likely impact on third-party

28

1	producers? How much restraint does this have on their
2	ability to supply the market?
3	Then the plaintiff would have to show that it
4	would be profitable for the monopolist to raise price
5	significantly -- whatever the number is, 5, 10 percent,
6	whatever -- as a result of that exclusionary conduct.
7	You could calculate a critical loss for the monopolist
8	that would be based on margins, and you could estimate
9	whether a 10 percent price increase after or during the
10	alleged conduct would leave sufficient residual supply
11	such that a monopolist would lose in excess of the
12	critical loss. So, that would get you the market
13	definition part of this. Then what do you do?
14	One strategy would be to not even bother with
15	shares, because you have basically concluded that the
16	single firm was able to engage in this alleged conduct
17	and get the price up, and in terms of that, one could
18	say, "Well, that's what we needed to know," and we will
19	now we go through the rest of the analysis and determine
20	what are the efficiencies, and maybe you want to talk
21	about recoupment as well. So, one could reasonably say,
22	"Well, we don't really need a market share threshold."
23	Other people could say, "Well, gee, it is in the case
24	law. We want to try to make it consistent. It is
25	really important. So, we need a market share

29

1	threshold." How would that work?
2	Well, one way to think about it in the context
3	that I have just outlined would be you could say, "Well,
4	the firms in the market would be obviously the alleged
5	predator, and then potentially also other firms that
6	have also benefitted from a price increase as a result
7	of this exclusionary conduct," and you might base their
8	share calculations on their sales of that product for
9	which the price increase was experienced.
10	But then you ask the question, "Well, why have a
11	share requirement? What does that do for you?" You
12	might say, "Well, it gives us some comfort because
13	predatory conduct is only likely to occur where the
14	shares are high." Well, there is an issue about that,
15	because some exclusionary conduct is really cheap, and
16	some exclusionary conduct is really expensive. So, if
17	you are going to engage in really expensive exclusionary
18	conduct, yes, then you probably want to have a big
19	share, because you need to recover that expense that you
20	laid out to execute the exclusionary conduct, but if you
21	are executing really cheap exclusion involving, a
22	Hatch-Waxman type of scenario or something like that,
23	which costs virtually nothing, well, then, what does the
24	market share do for you? So, that is unclear.
25	I have got about 30 seconds left, and I just

30

1	wanted to sum up by saying I think there are some really
2	important lessons to be learned from the Horizontal
3	Merger Guidelines market definition, and I am hopeful
4	that what will come out of this is we will get a bunch
5	of smart people in a room, maybe Greg and some of his
6	colleagues from the Antitrust Division and the FTC will
7	sit in a room, take all of this together, and come out
8	with an algorithm that is of similar significance to
9	what they did with the Merger Guidelines -- use the
10	first principles integrated approach, not worry about
11	the fact that what they might come out with is a
12	theoretic framework, theoretic algorithm that is not
13	immediately implementable, and then not be afraid to
14	consider a market definition guideline that deviates
15	from traditional case law, because what happened with
16	the Merger Guidelines is people originally said, "Oh,
17	this is nothing like the original case law," and now we
18	have been able to bring the two together, and the courts
19	have seemed to have adopted what is in the Guidelines.
20	Thanks very much.
21	(Applause.)
22	DR. CARLTON: Thank you, Joe.
23	Our next speaker is Larry White, who has arrived
24	in time. You have already been introduced, Larry.
25	DR. WHITE: Well, thank you.

31

1	DR. CARLTON: And the ground rules are we are
2	running a little late, so if you could keep to 15
3	minutes.
4	DR. WHITE: Right.
5	DR. CARLTON: Does the computer work?
6	UNIDENTIFIED SPEAKER: Yes.
7	DR. CARLTON: And you are the first person who
8	has the use of the computer.
9	DR. WHITE: All right, great. Well, thank you.
10	I am very pleased to be here this morning, and sorry for
11	the delay of my arrival. I flew down from New York this
12	morning, and every once in a while you get hit with a --
13	I do not know whether it is the right-hand tale or
14	left-hand tale on variance, but we were an hour late
15	taking off. So, here I am. I am very pleased to be
16	here.
17	I think this is a terrifically important issue,
18	and it is an issue where unfortunately too many mistakes
19	have been made, too many mistakes continue to be made,
20	and I want to walk you through what I consider to be
21	some important issues. I have got a few call it partial
22	answers. I do not have the complete answer. At the
23	end, I am going to be echoing Joe Simons' call. We need
24	a new paradigm; a paradigm is missing.
25	So, like any good business school professor, I

32

1	am going to tell you what I am going to say, and then I
2	am going to say it, and then I am going to tell you what
3	I said. I will frame the issue, I will remind you what
4	the standard monopoly model looks like, I will remind
5	you what the implications of that model are, I will
6	point out the loose language that has been used by
7	people who do know better or who ought to know better,
8	and I'll tell you about the danger of that loose
9	language. That will bring me to the Cellophane fallacy.
10	Everybody is going to talk -- you cannot not talk about
11	the Cellophane fallacy when we're addressing this topic,
12	remind you of an ongoing dilemma, put out some partial
13	suggestions, and wrap it up.
14	What's the issue? I am not going to get into
15	this market power versus monopoly power. The way I was
16	taught, it is all the same thing, and the exercise of
17	this thing, call it monopoly power or market power, is
18	the seller can sell at prices above marginal cost and
19	earn rents, and I should have added for a sustained
20	period of time, but I will go ahead with my story. That
21	is the picture that we carry around in our head of what
22	monopoly power, market power, is about, the sustained
23	charging of a price above marginal cost, maintaining --
24	I am going to use that word over and over again --
25	maintaining a price substantially above marginal cost.

33

1	All right, now, what also gets talked about,
2	especially in an antitrust context, is actions --
3	exclusionary, predatory actions -- that can create or
4	enhance market power. So, somebody who did not have it,
5	can create it. Somebody who has it through an
6	exclusionary or predatory action can enhance it, make
7	the demand curve yet less elastic or inelastic and earn
8	even higher rents.
9	If the seller is engaging in this kind of
10	activity, whether he is exercising the market power or
11	enhancing, a likely precondition is that the seller has
12	a large share of its market. So, that is not necessary.
13	You can come up with examples where if the overall
14	supply is limited, where other suppliers cannot expand
15	their output very much, where demand is quite inelastic,
16	even somebody with a relatively small share of a
17	commodity market by his unilateral actions can affect
18	the price, but more generally, a large share of
19	something called a market is going to be necessary. But
20	that then raises this threshold or safe harbor issue,
21	what is the market, and there is no standard paradigm
22	for that determination.
23	So, this is the picture we carry around in our
24	head, and the implications of that picture, the
25	monopolist maintains its price at a level above the

34

1	competitive price. He would not want to raise his price
2	any further unless demand changed or costs changed. He
3	is already where he wants to be. In trying to raise his
4	price, he would lose too many customers to sellers of
5	something else, and, of course, if the market changes
6	from a competitive structure to a monopoly -- because of
7	cartelization, because of exclusion -- then the price
8	changes, then the price increases, the seller, newly
9	feeling this market power, raises the price from the
10	competitive to the noncompetitive monopoly level, but as
11	a characterization of what is going on when we take a
12	snapshot of the market, he is maintaining the price at a
13	level above the competitive level. That is clear in
14	this standard model.
15	About 40 years ago, George Stigler developed an
16	expanded version of this, the dominant firm and the
17	inverted price umbrella, where he described a firm that
18	was not strictly a monopolist, he faced a reactive
19	fringe of smaller firms that were limited in their
20	supply response, and he showed basically you get a
21	similar type of outcome. The dominant firm is able to
22	charge, maintain a price above competitive levels, but
23	he doesn't want to go any higher because -- and there,
24	in the Stigler model, it is implicit -- he would lose
25	too many sales to that competitive fringe.

35

1	Okay, why am I making such a big deal out of
2	this? Because there has been loose language out there,
3	first by my colleagues, all of whom do know better, and
4	they describe the phenomenon of monopoly power, market
5	power, in terms of the ability of the firm to raise
6	prices. In other words, I have put in italics over and
7	over again, this language of "raise prices," or in the
8	context of the Microsoft case, Fisher and Rubinfeld
9	making this claim that, "Gee, Microsoft could have
10	raised its price substantially and wouldn't have lost
11	customers," and you have got to scratch your head, how
12	come they didn't? Then Evans and Schmalensee on the
13	other side, again, talking the language of "raise."
14	Even earlier, as I walked in the door, I heard
15	Phil Nelson talking about the monopolist "raising" the
16	price. Maintaining is what we're talking about, but I
17	am sure I in my looser moments fall into this "raising."
18	It is an easy thing to do, but I am going to show you
19	the dangers of it in just a minute.
20	I'll go over to some noted legal cases and legal
21	opinions, and again, you have got the same -- oh, did
22	I -- no, I forgot to put the italics in there, but you
23	can see the word "raise" in each of those -- in each of
24	those quotations from those cases.
25	All right, what is the danger? The danger in

36

1	the "raise" terminology is that if we think market power
2	and monopoly power are the ability to raise the price,
3	then it is easy to then think, "Ah, well, the test of
4	whether somebody has market power or not is whether the
5	seller can raise prices above currently observed
6	levels." Remember, that is what Fisher and Rubinfeld
7	were talking about there.
8	Conversely, if the seller is constrained from
9	raising prices because of its fears of losing too many
10	customers, then does that imply that it does not have
11	market power? The trouble is, even in the standard
12	paradigm where the monopolist is maintaining a price
13	above competitive levels, it cannot profitably raise its
14	price because it would lose too many customers to
15	sellers of something else.
16	That, of course, then leads us to the Cellophane
17	fallacy, the U.S. v. Dupont case, where the issue was,
18	was the market a narrow market of cellophane, in which
19	case it is clear, Dupont had market power. There was
20	one other seller of cellophane, Sylvania. It was under
21	license from Dupont, and so, effectively, no question.
22	If the market was cellophane, Dupont had market power.
23	Or was it, as Dupont claimed, flexible wrapping
24	materials, in which case Dupont only had a 17.9 percent
25	share and didn't have market power?

37

1	The Supreme Court majority said it was
2	interchangeability that carried the day, that cellophane
3	was interchangeable with other materials mentioned --
4	there was wax paper and brown wrapping paper and
5	aluminum foil and glassine and lots of other things --
6	and the majority said, "Ah, look, it is interchangeable.
7	Dupont can't raise its price. So, it must be part of
8	that larger market."
9	The minority pointed out the fallacy of that
10	reasoning and also pointed out the comparison with
11	rayon, where Dupont also faced 15 to 18 other producers,
12	also had a market share that was below 20 percent, and
13	made much less profits. They also pointed out that
14	Dupont's price of cellophane did not move around when
15	those other flexible materials' prices changed.
16	So, we have this ongoing dilemma. Profit data
17	nowadays are relied on a whole lot less than was the
18	case back in the fifties when Stocking and Mueller were
19	writing, when the Supreme Court minority relied on those
20	profit data. The Horizontal Merger Guidelines cannot be
21	used, because they are a forward look, as you have heard
22	already, they are a forward-looking test.
23	The one exception, which Greg Werden has pointed
24	out, is that if we are talking about a practice that is
25	not yet in place, say an exclusive dealing plan that is

38

1	going to be put in place. A plaintiff comes in, asks
2	for an injunction. We are talking about something where
3	it is a prospective practice. Then the prospective,
4	forward-looking paradigm of the Merger Guidelines will
5	work. To the extent that that is what we are looking
6	at, fine, we have got an answer, but lots of instances
7	are not of that kind.
8	As Phil remarked earlier, elasticities do not
9	help us very much. You cannot tell the difference
10	between a true monopolist and just a different -- a
11	seller of a differentiated product, a Chamberlin/
12	Robinson monopolistic competitor.
13	Okay, what to do? Well, sometimes a complaint
14	will involve a prospective practice, and then we have
15	got the Merger Guidelines. Sometimes there will be
16	cross-sectional or time-series evidence involving prices
17	where we can tell that concentration matters, and when
18	concentration matters, you have got a market, and retail
19	services are an area where cross-sectional data may be
20	available.
21	I harken back now ten years to the Staples case,
22	where cross-section data showed that prices were
23	different, higher where only Staples or Office Depot was
24	present in the market, lower when both were there, yet
25	lower when they and a third office superstore were

39

1	there. That evidence carried the day, and I think
2	correctly, that there was a problem -- there would be a
3	problem if the two firms merged, and it told us office
4	superstores were a market.
5	Think of the American Airlines predatory
6	behavior case. Why do we think that city pairs are a
7	market, city pairs airline transportation? Because
8	there is lots of cross-sectional evidence that shows
9	that, controlling for other things, prices matter and
10	prices are related to concentration. Sometimes profit
11	data will be useful.
12	I mean, if you think the Microsoft case was a
13	good case, if you thought that Microsoft's behavior was
14	a problem, why did you think that? And I think at least
15	part of the story was those profits. They were so large
16	that even with all the problems that we know about
17	profits, they were telling us something. But what if
18	none of these possibilities are available?
19	Well, Phil Nelson and I a few years ago made a
20	proposal. It turns out similar language can be found in
21	a 20-year-old article by Tom Krattenmaker. Greg had a
22	version of this proposal in an article he wrote in 2000,
23	where basically it is asking in the presence of an
24	allegation of exclusion, what would have been the
25	consequences of the absence of exclusion? It requires a

40

1	two-step investigation.
2	First you have got to ask, in the absence of
3	exclusion, what would the plaintiff's sales have been?
4	And then you have got to ask, what would the price
5	consequences of those additional sales have been as
6	well?
7	Now, as was indicated earlier, this would focus
8	directly on effect, and it implicitly delineates a
9	market, but if you think about what the unilateral
10	effects analysis under the Horizontal Merger Guidelines
11	does, it is basically doing the same thing. It is
12	looking for an effect, and then, if somebody goes ahead
13	and then tries to delineate a market, that is sort of
14	redundant. You have already found the effect.
15	Implicitly, you have said there must be a market there,
16	and that is basically what the Nelson and White proposal
17	does as well.
18	But I think the best approach would be let's try
19	to develop -- you know, I have thought hard about it.
20	The best I could come up with was this joint proposal
21	with Phil. It may not be good enough. Can the world
22	come up -- can the Division, can the FTC, can a bunch of
23	smart people out there -- come up with a paradigm that
24	will have the power and eventual universality of the
25	Horizontal Merger Guidelines?

41

1	I urge you, remember what the world looked like
2	before 1982. Remember what 1981, 1980 and 1979 looked
3	like. We did not have a paradigm. We had
4	Elzinga-Hogarty. We had Ira Horowitz's suggestion.
5	There were other ideas out there. George Hay was going
6	around talking about how the Division defined markets,
7	and he would say, "Well, we would look for whether there
8	was a specialized trade journal that the sellers in a
9	marketplace all submitted their data to." Those were
10	the kinds of indicia that people looked to. The Merger
11	Guidelines brushed all that stuff away, and we have now
12	got a powerful paradigm. I hope that some smart people
13	out there somewhere will be able to develop something
14	with similar power.
15	So, winding up, we have got an unsatisfactory
16	state for market definition. I would hope we are in
17	1981, and next year, somebody is going to come up with
18	something that will have the same kind of power as the
19	Horizontal Merger Guidelines. I have shown you some
20	partial remedies, but the best remedy would be a new
21	paradigm.
22	Thank you very much. I am very pleased to have
23	this opportunity today.
24	(Applause.)
25	DR. CARLTON: Okay, thank you, Larry.

42

1	Our next speaker is Andy Gavil.
2	DR. GAVIL: Good morning, everyone. Thank you
3	to the organizers for inviting me to join everyone
4	today. I am delighted to be here and agree with
5	everyone else that these are some very important --
6	indeed, fundamental -- issues to how we go about
7	analyzing antitrust cases, and in truth, they are not at
8	all unique to section 2. Questions of power and effects
9	really cut across all kinds of cases today. So,
10	resolving one area clearly is going to influence and
11	affect the others just as the Merger Guidelines has
12	affected many areas.
13	So, I start with my first slide in talking about
14	it is all about anticompetitive effects, and I think I
15	would add to that, and legal process. At the end of the
16	day -- that is a great phrase, "At the end of the
17	day" -- "At the end of the day, in the final
18	analysis" -- but at the end of the day, in the final
19	analysis, whatever we conclude as a matter of economics
20	is the right approach, we have to translate that into a
21	legal system of decision-making. It has to work in
22	courts. It has to work in a context where we have
23	burdens of pleading and burdens of production and
24	burdens of proof. It has to work in a context where we
25	have various methods for discovery of evidence, where we

43

1	have a role for expert witnesses, where we have judges
2	and juries, and if it cannot work in that context, then
3	perhaps there is a problem with what we have come up
4	with as a theoretical matter.
5	I forget who it was, I think it was Joe talking
6	earlier about how the Merger Guidelines were originally
7	received. Well, part of the problem in how they were
8	received is that they were received by a legal community
9	accustomed to looking at cases in one particular way.
10	They suggested that we needed to look at those cases in
11	a very different way, and it was very unclear in 1982
12	how you would translate, how you would take something
13	like SSNIP and what evidence would you need?
14	The lawyers that were asking the questions of,
15	what witness am I going to need to do this? What
16	evidence will I need from my client, from the other
17	parties? How will I assemble it? How will I present
18	it? There can be no doubt at all I think in anybody's
19	mind that the Merger Guidelines and subsequent
20	developments have been an economist's full employment
21	act, and certainly that has been evidenced in the
22	antitrust area. It is hard to imagine today proving any
23	kind of case, plaintiff or defense, without the role of
24	economists, and that is a result of the writing into our
25	substantive standards various economic ideas.

44

1	So, as I go through these slides, I want you to
2	sort of keep that in mind. The focus I have tried to
3	bring to my comments today is, how do we make it work in
4	this legal system? Well, common issues in antitrust are
5	effects, and we have certain ways that we go about
6	establishing them. We have irrebuttable presumptions --
7	that is what the per se rule is all about -- and we have
8	rebuttable presumptions; whether we are using direct
9	evidence or circumstantial evidence -- and that is going
10	to be an important issue that I am going to look at
11	today -- we have different ways that we go about trying
12	to establish effects.
13	Direct evidence, defined here, is the actual
14	exercise of market power. It may come out in
15	performance evidence. It may come out in before and
16	after studies of price. It is reflected to some degree
17	in our use of "quick look." The "inherently suspect"
18	formulation is also a way of looking at things that are
19	obvious, and a question I will be asking today is, do we
20	have equivalents for section 2 and would it make sense
21	to use them in section 2?
22	On the circumstantial evidence side, we have
23	something that I have called a "double inference." We
24	define a market, we calculate market shares from a
25	certain level of market share, we infer market power,

45

1	and in truth, from that, we then infer the capacity for
2	anticompetitive effect. In litigating terms, we are
3	dealing with two very standard paradigms of how to go
4	about proving something.
5	Well, power, of course, is a condition precedent
6	of effects, but if you look in the cases, there is a lot
7	of confusion -- again, loose language -- about how it is
8	used. Some cases say, "Well, what we need is market
9	power," and even in cases like NCAA and Indiana
10	Federation of Dentists that really were out in the
11	forefront in this quick look idea and the use of direct
12	evidence of actual effects, there is confusing language
13	about what "market power" means.
14	Well, power is the condition precedent of
15	effects. If you have the effects, the power is there.
16	So, part of the point of Indiana Federation, talking
17	about market definition and market power as surrogates,
18	was to make the point that when you have the actual
19	effects evidence, going sort of back around the
20	circumstantial evidence route, trying to define a market
21	and determine whether there are large market shares, may
22	be beside the point. Those things are surrogates for
23	direct evidence.
24	Well, as in many areas of antitrust, that leads
25	us to a point where we can identify easy cases and hard

46

1	cases. A good example I think of the easy cases, when
2	the direct and circumstantial evidence are aligned, when
3	they are pointing in the same direction, when you have
4	evidence of actual effects and you have high market
5	shares, those are easy cases. We do not argue about
6	those very much. The D.C. Circuit in Microsoft actually
7	structured its discussion of monopoly power that way,
8	looking at both direct evidence, circumstantial
9	evidence, they are both pointing in the same direction,
10	easy case.
11	On the other hand, for safe harbor ideas, if you
12	have de minimus evidence and no effects and you have low
13	market shares, again, pointing in the same direction,
14	and I would make this point -- I'll raise it a little
15	bit later -- in terms of safe harbors, I do not think
16	you can rely just on market shares alone. It has to be
17	market shares plus certain other factors, and I will
18	also suggest that if we are going to have safe harbors,
19	we need some danger zones, and again, it might be market
20	share plus some other characteristics.
21	But evidence and power effects are interrelated,
22	and I think this is what makes part of our current
23	framework very difficult to think about. Courts do
24	think, because of years and years of case law, first
25	monopoly power, then willful acquisition or maintenance,

47

1	when in truth, the evidence of conduct and effects in
2	the evidence of power is going to be very interrelated.
3	Well, again, thinking about direct and
4	circumstantial evidence, the benchmark for
5	circumstantial evidence is clearly the Horizontal Merger
6	Guidelines. They really did advance the science of
7	thinking in terms of circumstantial evidence. Recall,
8	though, that Cellophane was a section 2 case, and maybe
9	there are some different problems that come up when we
10	are doing prospective predictions about likely market
11	power versus retrospective methods when we have, you
12	know, the before and after ability to actually look at
13	the effect of conducts, but the Merger Guidelines in any
14	paradigm we come up with are probably going to have some
15	continuing significance. They have been cited by courts
16	outside of section 7. They are cited in section 1 cases
17	and section 2 cases. Basic ideas and concepts are
18	clearly interrelated.
19	So, my suggestion at this stage of our
20	development is we need something of a similar to the
21	Merger Guidelines to refine "actual exercise" standards
22	and to harmonize those standards across different
23	offenses. A critical question, I think, is how much and
24	what kinds of effects evidence should be sufficient to
25	shift a burden? And here I remind, again, that outside

48

1	the area of exercising prosecutorial discretion, outside
2	the walls of the agencies when they are deciding whether
3	to bring a case, if the decision to bring a case is made
4	and the economists agree, the next question the lawyers
5	are going to have is, "Well, how do we meet our burden
6	of production? What evidence are we going to assemble?
7	What is going to make us win this case?"
8	I think when you are thinking about direct
9	effects evidence, and market share as well, a critical
10	question in section 2 is, what does it take to shift a
11	burden? Frequently what you see defendants arguing in
12	cases is the burden didn't shift, the burden didn't
13	shift, the burden didn't shift. Well, what does that
14	mean?
15	It means that there is no requirement on the
16	part of the defendants to actually justify their
17	conduct. If they claim there are efficiencies, where is
18	the evidence of efficiencies? That does not happen
19	until the burden shift takes place. That is a critical
20	stage. It is a critical stage that has to be focused
21	on, and I have given some examples here of various cases
22	that raise some of those questions.
23	I think we are also feeling the weight of the
24	Alcoa paradigm. In looking back at Alcoa and the cases
25	that preceded it, all Judge Hand did was he surveyed the

49

1	previous cases and looked at winners and losers to come
2	up with his three famous sort of -- you know, 33, not
3	enough; 66, maybe; over 90, definitely. Well, where did
4	he get that from?
5	If you look at the prior Supreme Court cases,
6	you will see that there were cases falling into each of
7	those categories. He sort of synthesized them and came
8	up with this benchmark. I think an important question
9	for us is, are we ready to move beyond the total
10	reliance on market shares, which sends us off in this
11	direction of conflicting evidence, plaintiffs and
12	defendants having experts -- the market is big, the
13	market is small -- and is that really where we want to
14	be? What can the role of direct evidence be?
15	The Re/Max case was an example of a court
16	relying on direct evidence, actual price effects
17	evidence in a section 2 case. The 7th Circuit in
18	Republic Tobacco rejected such an approach, said that
19	Indiana Federation did not apply and NCAA did not apply
20	to a vertical case. Is Staples -- and the unilateral
21	effects that people have alluded to already -- is it
22	related? I think it is. It is a way of trying to more
23	directly gauge. We have talked about the monopoly
24	versus market power being kind of a silly distinction.
25	So, I will move on.

50

1	I think there is an important role here for
2	decision theory, which obviously has begun to influence
3	our thinking. The emphasis tends to be on fear of error
4	costs, and often that motivates calls for more and
5	better evidence. We need more before that burden
6	shifts. One point I would like to walk away with today
7	is urging that we also consider the second half of
8	decision theory, which is process and information costs.
9	Is more evidence really always better?
10	In that regard, I sort of suggest -- and it is
11	not really new, there is a lot of general literature out
12	there on the economics of evidence. Richard Posner has
13	a long article on an economic analysis of evidence, and
14	I put forward the question, "When does the marginal
15	value of additional evidence in terms of economic
16	certainty (minimizing error costs) outweigh the costs of
17	obtaining and processing that evidence, taking into
18	account whether it is reasonably accessible to the party
19	bearing the risk of non-persuasion?" What I tried to do
20	in that question is integrate some economic ideas and
21	some legal process ideas from both the rules of
22	procedure and the rules of evidence. It is always easy
23	to demand more. It is always easy to pursue some kind
24	of level of absolute certainty and minimal error costs.
25	The question is, as a legal standard, when we take that

51

1	into court, is that really going to strike the right
2	balance for us in resolving cases?
3	Antitrust is not always rocket science, and I
4	think we need to get over the idea that it always is.
5	Yes, we need safe harbors to guard against false
6	positives. I think we also should be emphasizing
7	equally defining danger zones where we might be running
8	into false negatives.
9	Is monopoly power all that puzzling? I would
10	point out to everyone that neither 3M nor U.S. Tobacco,
11	in two U.S. Courts of Appeals monopolization cases, even
12	contested that they had monopoly power. In the
13	Microsoft case, they contested it, but rather
14	unpersuasively, and every agency and every court to look
15	at it has concluded that yes, indeed, Microsoft had
16	monopoly power.
17	We could go on with a couple other examples,
18	American Airlines, Dentsply. Were these really such
19	difficult cases? If they were not, then why were they
20	so difficult? Why would parties not even litigate the
21	point about their power? There must be, there must be
22	cases where -- again, market share plus -- where there
23	must be additional factors, information on entry
24	barriers. Entry barriers will always, for example, be
25	important.

52

1	Finally on this slide, sliding scales, not all
2	burden shifts are created equally. You see this in
3	cases like Baker Hughes and Heinz in the merger area,
4	the realization that sometimes when a burden shifts, it
5	really shifts, and the presumption is very strong, and
6	other times, it kind of is just enough to shift. Well,
7	in responding to those sorts of cases, we might want to
8	respond in different ways by considering what is
9	required to shift and what is required to shift back a
10	burden in different ways.
11	On legal standards and decision-making, I think
12	that the balancing of effects idea is a straw man. We
13	could cite, as Larry White did, we could put up lots of
14	slides with courts saying, "Anticompetitive effects; the
15	burden shifts. Efficiencies; then we balance one
16	against the other." We do not really do that. I have
17	looked; you can all look. If you can find me a Section
18	1 litigated case in which the case was actually decided
19	on balancing effects versus efficiency effects, consumer
20	surplus diminution versus increased producer surplus,
21	find me such a case. I would like to see it. It is not
22	what we do.
23	What we do is weigh evidence. What juries do is
24	they compare the evidence of anticompetitive effects
25	with the evidence of efficiencies, and they make a

53

1	decision about where the weight of the evidence is.
2	That has to do with credibility; it has to do with
3	persuasiveness. It does not have to do with $10 of
4	anticompetitive effect and $11 of efficiency.
5	Finally, a word about caricatures and corrosion
6	of the rule of law. The level of discourse and the
7	level of criticism of antitrust, as we all know, has
8	continued for quite some time. It has continued despite
9	the fact that in the last 40 years, we have seen some
10	pretty major corrections to antitrust.
11	I say caricature -- and this is not my
12	caricature -- but this is what you see in a lot of the
13	criticisms of antitrust, and I think it is a caricature
14	that ignores this last period of adjustment over the
15	last 30 years. Incompetence -- judges, just
16	incompetent. They can do habeus corpus, they can do
17	environmental, they can do securities law, but antitrust
18	is rocket science, keep them away.
19	The same thing with juries. They just do not
20	know the difference between somebody who is full of it
21	and somebody who really knows what they are doing. They
22	cannot tell the difference between economists in this
23	case and, of course, neither can they decide any other
24	possible case.
25	And, of course, enforcers. I have the asterisk

54

1	there just to remind me to say that. Typically it is
2	enforcers themselves who make this argument, God, we are
3	so stupid. You shouldn't really trust us to make any
4	decisions, and although we may -- and it gets very
5	personal -- we may be able to make the decision, but
6	other enforcers are really stupid, especially those guys
7	at the offices of the states.
8	Who are the untrustworthy self-interesteds, the
9	self-interesteds who are untrustworthy? Rivals, oh,
10	they are always full of it. They are always complaining
11	about more competition. Dealers, yeah, what's that
12	freedom of dealer stuff? You know, manufacturers,
13	consumers, aligned; dealers, out in left field. And
14	plaintiffs pretty much all are full of it, especially
15	class action reps.
16	Ah, but who can we trust? Dominant firms.
17	Dominant firms articulating efficiencies. Fear of error
18	cost? That's truth. We need to put a lot of weight in
19	that. We need to be concerned about it. Other
20	defendants, generally yeah, and especially efficiencies.
21	Two problems I have with this sort of
22	caricaturing of antitrust. One is, I don't think it is
23	true. I would like to see the list of false positives
24	in the last 25 years. We have been moving towards
25	reduced error costs, and here I think it would be

55

1	helpful to have the economists really define what they
2	mean as "false positive." It is not a case on which
3	reasonable people can differ. It is a case that sort
4	of -- again, borrowing from procedure -- no reasonable
5	party could have come out that way. To me, that would
6	be a false positive or a false negative. It is not a
7	case that we simply disagree about.
8	LePage's has, you know, been frequently used as
9	sort of this example of a false positive. Be reminded
10	that 3M did not contest its market power, and if it did
11	offer any evidence of efficiencies, nobody who looked at
12	it found it very convincing. Did the Court of Appeals
13	give us a useful standard for bundled pricing? No, but
14	neither has anybody else yet. So, to call that a false
15	positive and say, "This is an example of how we're going
16	to inhibit all kinds of other cases," I am not sure that
17	that is justified.
18	The final point and I will sit down. As I said
19	at the start, Larry said at the end, you say what you
20	said at the beginning. At the end of the day, these
21	cases have to go to court sometimes, and this kind of
22	rhetoric of criticism ultimately is corrosive of the
23	rule of law. I think it is heard in curious ways
24	outside the United States. These criticisms really go
25	to the heart of whether we are willing, at the end of

56

1	the day, to rely on courts to make decisions.
2	We have numerous procedural devices, summary
3	judgment, judgment as a matter of law, Daubert
4	standards, appeal rights. If after all of that has
5	occurred a plaintiff actually wins a case, which does
6	not happen very often, I think we ought to be a little
7	bit more cautious about tossing the rhetoric around
8	about the incompetence and the untrustworthy
9	self-interesteds, all right?
10	Thanks very much.
11	(Applause.)
12	DR. CARLTON: Thank you very much, Andy. I was
13	pleased to hear I am not as incompetent as once
14	enforcers were thought to be, and to prove that I am
15	still competent, we are going to have a break, and it
16	will be a 10-minute break, and we will reconvene
17	promptly so that we can try and stay roughly on
18	schedule. Thank you.
19	(A brief recess was taken.)
20	DR. CARLTON: Why don't we try and start. Our
21	next speaker is Rich Gilbert.
22	DR. GILBERT: I would like to thank the
23	organizers for the opportunity to be here. I was
24	invited to talk about technology markets, so if any ink
25	gets spilled on this issue as a result of my comments,

57

1	you can be sure it will be Independent Ink, though I
2	will not talk about the presumption of market power for
3	patents. I thought we resolved that issue in the IP
4	Guidelines, although it is not the case that the Supreme
5	Court immediately adopts everything that the agencies
6	come up with.
7	Now, when we talk about market definition, there
8	is a real sense in which we are talking about either
9	guide posts or lamp posts. Now, lamp posts, as you
10	know, shed light on a subject but do not necessarily
11	shed truth about the subject. A lamp post might
12	illuminate the ground, but that does not mean that the
13	dollar that we are looking for is around the lamp post,
14	even though if it were, perhaps we could see it.
15	Guide posts, on the other hand, serve to focus
16	the analysis. The guide posts lead the way. The way
17	may be very foggy and very complicated and very
18	difficult, but can be very useful.
19	Now, my take, sort of in the spirit of Andy's
20	comments, the courts and defendants like the market
21	definition exercise, even though it is often used much
22	more as a lamp post than a guide post. They like the
23	exercise because, of course, for a defendant, if you can
24	show the market is very broad, chances are there is no
25	antitrust case there. For a court, they are all very

58

1	busy. They have full dockets. If you can show the
2	market is very broad, they do not have to worry about
3	it.
4	Plaintiffs also seem to like market definition
5	or many of them like market definition, because if you
6	can prove that or demonstrate or make a convincing case
7	that the market is narrow, well, chances are then there
8	will be an issue, but as I think everybody on this panel
9	is implying, none of those conditions, whether it is
10	broad or narrow, presumptive of a case or not
11	presumptive of a case, none of them are really relevant
12	directly to the analysis. We would rather have market
13	definition serve as the guide post to lead the way to
14	the right analysis rather than defining whether there is
15	or is not a case.
16	Now, so, if we talk about markets for
17	technology -- first I should distinguish, I am going to
18	focus more on technology markets than on markets for
19	technology. Markets for technology can be analyzed
20	using conventional goods markets, often using
21	conventional goods markets, which are sufficient for
22	analysis in many high technology industries, whereas
23	technology markets are useful when what is at issue is a
24	right or rights to a technology that are licensed rather
25	than embodied in a patent. So, I am focusing more on

59

1	technology markets than markets for technology, although
2	maybe in discussion, we will get to that distinction,
3	whether there should be a distinction.
4	Technology markets are defined in the IP
5	Guidelines. Technology markets consist of the
6	intellectual property that is licensed that are close
7	substitutes. Of course, here now, as in all market
8	definition exercises, the issue is, what are the close
9	substitutes? And when you are talking about technology
10	markets, the close substitutes are not only other
11	intellectual property rights, but also goods and
12	services that may substitute for those intellectual
13	property rights.
14	It adds another layer of difficulty and
15	complexity to the analysis, because just like in
16	conventional -- other section 2 goods market definition,
17	exactly what to sweep into that analysis and how, it
18	depends upon the prices, prevailing prices, and whether
19	the conduct is prospective or retrospective, these are
20	all challenging issues, which I am not going to entirely
21	resolve.
22	Now, technology markets also are -- there is an
23	upstream analysis for inputs which I think raises some
24	interesting questions by itself. Technology markets
25	have been used I think with some success to analyze the

60

1	licensing of technology to manufacture float glass, for
2	blending clean gasoline in the UNOCAL case, the float
3	glass with the Pilkington case, for designing fast
4	computer memory chips, as in the DRAM cases, perform
5	laser eye surgery, or to incorporate genetically
6	modified traits into agricultural seeds. These are all
7	some examples, I think, of markets that have been
8	analyzed using basically an upstream analysis for
9	licensed inputs.
10	Now, on the geographic market side, this is an
11	area where using technology markets in some cases
12	simplifies things. It is fair, I believe, to presume in
13	many cases -- not all, of course -- the geographic scope
14	for technology markets is very wide, because it is not
15	very difficult for a potential licensee to negotiate
16	with even quite distant licensors unless there are legal
17	or regulatory or some other restrictions that prevent
18	the use of licensed technology in different locations,
19	as there was, for example, with the UNOCAL case, but in
20	these other cases, the enforcement agencies I think have
21	correctly concluded that the technology markets are very
22	broad, U.S.-wide and sometimes worldwide.
23	Now, technology fees, should these be indicators
24	of market power? Interesting question which has not
25	been quite directly addressed. Marginal cost of

61

1	licensing is typically very low. It suggests that there
2	is market power if we define market power as the ability
3	to sustain prices above marginal cost, then looking at
4	technology fees, gives you an immediate read on whether
5	or not there is market power, not necessarily monopoly
6	power, but, as economists have said, that is a difficult
7	line to draw between market power and monopoly power.
8	Now, again, the relevant question is the ability
9	to increase or maintain technology fees significantly
10	above marginal cost for an extended time. In this
11	dispute about market power versus monopoly power, I am
12	certainly in the camp that says that monopoly power is a
13	lot of market power and that there is no clear dividing
14	line between the two, and the question is, the relevant
15	question is, is there conduct that leads to either
16	increasing or maintaining technology fees significantly
17	above marginal cost for an extended period of time and
18	whether it is prospective or retrospective?
19	If it is prospective, perhaps the ability is to
20	increase technology fees. If it is retrospective, then
21	the question is more has conduct contributed to the
22	ability to maintain technology fees significantly above
23	marginal cost?
24	This is now more in the spirit of what Larry
25	White was saying in his approach to section 2 market

62

1	definition. Also for technology fees, a related and
2	relevant question in a section 2 context is whether
3	competition, whether injecting more competition, would
4	result in a significantly lower technology fee if the
5	competition were not excluded.
6	I also agree that this opens up a lot of
7	interesting and unresolved issues, as in how much
8	competition should be enough to consider? What should
9	the price effect of that competition be in order to
10	define a relevant market? Is an elasticity of demand
11	faced at the existing prices for the fees and other
12	goods and services? Is an elasticity of demand minus
13	two, is that low enough, small enough in magnitude, or
14	does it have to be minus 1.1 or 1.5 or is minus 3
15	enough?
16	These are very important and serious questions
17	that need to be addressed if we are going to do this
18	kind of hypothetical decrease in price through a
19	hypothetical increase in output as a way to identify a
20	relevant market.
21	So, the focus on that additional competition and
22	whether it lowers the fee I do believe can get around
23	the Cellophane fallacy, and I think another important
24	aspect of that approach is that it focuses the analysis,
25	the definition of the market, on the analysis of the

63

1	competitive effects of the conduct. So, I think
2	sometimes it is a criticism of the hypothetical decrease
3	in price approach that it is too related to the conduct
4	that is being alleged as anticompetitive.
5	I turn it around and say that no, I think it is
6	an advantage of this approach, because it connects the
7	conduct at issue to the analysis of the impacts at
8	issue. Too often, I think many of us would agree that
9	the market definition exercise puts the cart in front of
10	the horse. We should be thinking about where are the
11	competitive effects, how significant can the competitive
12	effects be, and then let the market definition respond
13	to that rather than defining where the competitive
14	effects are. Again, this stems from the problem of the
15	Cellophane fallacy that a profit-maximizing firm has no
16	incentive to raise or lower its technology fee.
17	Another question about analysis of inputs, in
18	principle, the antitrust analysis for a technology input
19	is not qualitatively different from the analysis of any
20	other upstream good or service. The demand for the
21	input is derived from the demand for the final good or
22	service, and one thing to point out is the
23	Hicks-Marshall Law of derived demand, which says that
24	the elasticity of Derived Demand is proportional to the
25	cost share of the input. It is roughly the cost share

64

1	of the input times the elasticity of demand for the
2	output. That will generally lead to a conclusion that
3	the elasticity of demand is pretty small in magnitude.
4	Indeed, in the Microsoft case, Microsoft made
5	the argument that if you do this calculation, the
6	profit-maximizing price for Windows was I think, like,
7	$1,500 or something like that, and therefore, we could
8	not have market power because we are not charging
9	$1,500. I think it was an argument that was never
10	really entirely responded to, but one does find that as
11	you go upstream, you are going to generally get less
12	elastic demands, derived demands; therefore, more
13	potential to raise prices; therefore, more possibility
14	of competitive effects.
15	Of course, while it implies relatively inelastic
16	demand for inputs and the ability to affect the input
17	price, the input price has only an indirect effect on
18	the final consumer prices, which is why the elasticity
19	of demand is low. So, it turns around and gets you the
20	other way. So, upstream analysis can overstate the
21	ability to affect consumer prices.
22	As you move downstream, though, the question is,
23	how far downstream do you go? If you go far enough
24	downstream, almost everything competes with everything
25	else. If you move all the way downstream, eventually

65

1	you are competing for the consumer's entire budget
2	allocation, and whether you are talking about movies or
3	sports or buying a car or whatever, everything competes,
4	and the overall elasticity of demand for all goods and
5	services is one, because you only have so much income.
6	So, it is my view -- my strong view, but it is a
7	view -- that analysis should take place where the firm
8	has the ability and incentive to raise or maintain a
9	price paid for an input or a final good, and the
10	question should be, is the conduct the type of conduct
11	that we want to prevent? And if it is, let's analyze it
12	where the conduct might have an effect and let the
13	market definition follow from where the conduct could
14	have an impact.
15	I just have a very quick example to end with of
16	genetically modified seeds, which express a desired
17	characteristic, like insect resistance in corn or
18	tolerance of some herbicide. Do conventional seeds
19	compete with licenses for seed traits? So, that gets
20	back to the IP Guidelines definition, where do the goods
21	come in to compete with the traits? It is a complicated
22	question, not one I am here to answer, but I would just
23	point out that these agricultural markets are moving
24	increasingly to genetically modified traits, which is
25	now way above 80 percent in soybeans and up above 50

66

1	percent in corn, and if you are looking at questions
2	about whether conduct is maintaining high prices for
3	these characteristics and ultimately higher corn prices,
4	it is my view that you should look at the trait markets
5	for where this conduct is expressed, because that is
6	where the effect could be.
7	It may be that the conduct is not the type of
8	conduct that should be subject to an antitrust sanction,
9	but that is the right place to look. It goes back to
10	the lamp post and the guide post. Let's look where
11	there could be effects. Let's let the market definition
12	exercise follow from the inquiry into competitive
13	effects. Let's not use market definition as a lamp post
14	to illuminate a problem that may or may not exist.
15	Thank you.
16	(Applause.)
17	DR. CARLTON: Okay, thanks, Rich.
18	Our last speaker is Michael Katz.
19	DR. KATZ: I would like to thank the organizers
20	for inviting me here, but I do not have time.
21	I want to talk about -- it is a bit of a grab
22	bag, but I will start about something systematic, which
23	addresses the question of why delineate relevant markets
24	in a section 2 case, and what I want to start with is
25	really the first principle, what is the point of all

67

1	this, and we really try to answer this question of a
2	given practice harms competition and consumers, and what
3	I want to talk about for a few minutes is how that gets
4	us to talking about relevant markets, and I am going to
5	talk about at least three things that relevant markets
6	might be doing to help us answer that question.
7	Okay, so the first one is you can think of --
8	what you are trying to do is you are defining relevant
9	markets so you can calculate market shares and
10	concentrations, and we are doing that because we want to
11	know whether the defendant or the firm under
12	investigation, whether the defendant currently has
13	monopoly power. Now, as everyone has been talking
14	about, this is where the hypothetical monopolist test
15	breaks down, so there is an issue there.
16	It seems to me where we have gotten, actually,
17	in a bunch of the recent cases -- and maybe this also
18	goes to Andy Gavil's point about somebody showing me a
19	false positive -- but I think if you look at Dentsply
20	and Microsoft, there was plenty of expert testimony, but
21	in the end it just came down to hard core pornography,
22	the thing is you know it when you see it. People have a
23	good idea that false teeth are a product and they are
24	not really worried about a lot of other substitutes. I
25	mean, there is sewing your lips shut and things like

68

1	that, but I think in both of those, that was not really
2	the issue.
3	Now, I want to make a couple of other points
4	about concentration as an indicator of market power
5	here. One, if we are going to look at market shares, I
6	think we really ought to ask ourselves, where did the
7	market shares come from? Because it matters. Think
8	about it. In some cases it is because of product
9	differentiation, and some producers have much more
10	successful products that match up with consumer tastes.
11	There can be very different managements of different
12	producers have different strategies, and one of the
13	firms decided to have a high-volume, low-price strategy.
14	I think the conclusions one would typically want
15	to draw about the implications of them for competition
16	are very different, and so I think it is important to
17	try to have such a theory, and I think it is often
18	lacking in antitrust cases. People just talk about the
19	shares but not what they really mean or where they came
20	from.
21	The other thing is we want to ask ourselves why
22	we care whether the defendant currently has monopoly
23	power, and I will say at least I think one reason is you
24	can think of it as a one-sided test in a monopoly
25	maintenance case, which is to say, if you are in there

69

1	arguing that some particular practice successfully
2	maintained a monopoly and you come up with a credible
3	analysis that says the firm had a very low share, that
4	is likely to undermine the case. Now, it certainly does
5	not work in the other direction, right? Just because
6	you have a high market share does not mean you are
7	guilty of any sort of offense at all and it may be that
8	you got it because you deserve it. Okay, so that is a
9	particular use.
10	Now, I want to distinguish that from some
11	others, because I think they often get rolled together,
12	and they really are different, although they are
13	related. Okay, so another one that is related is
14	concentration as a screen for potential harm to
15	competition. Now, in a sense what I just said, it is a
16	screen, the one I just said, which is you are saying,
17	look, if they have a tiny market share, is it really
18	plausible that they have harmed competition
19	significantly in the past, but I also want to worry
20	about it going forward, and there it is not at all
21	clear -- in fact, I think it is not a general
22	proposition -- that you want to look at concentrations
23	to understand the potential for harm to competition,
24	because if you are looking at a case on a going-forward
25	basis, sometimes the current share of the defendant is

70

1	relevant, but other times, it is not, right? You are
2	not worried about their share now. You are worried
3	about what their share is going to become or what the
4	state of competition will become going forward.
5	Okay, so, notice I hedged it. I am an
6	economist, so lots of "on the one hands, but on the
7	other hands." So, in some cases where you are looking
8	on a going-forward basis, current shares may be largely
9	irrelevant. In other cases -- and I have the example
10	here of exclusive dealing -- even when you are looking
11	on a going-forward basis, market shares could be
12	relevant, and I would think that would have been true in
13	Dentsply.
14	Now, as it turns out, in Dentsply, it was
15	looking backwards, but if one had brought the case much
16	earlier, I think what one could have done is say, look,
17	Dentsply has this large market share, and I think by any
18	sensible measure they had a huge market share, and we
19	could argue about the source, but let me just
20	hypothesize here without anyone arguing that it is
21	because they did have teeth that were more popular and
22	more attractive, that there was something about their
23	product that they did have an advantage, and others were
24	not able to imitate, and then you could use that fact to
25	say, okay, that is going to tell us something about how

71

1	exclusive dealing is going to work going forward, and
2	even exclusive dealing with at will contracts, which is
3	what was present in Dentsply, because this one firm's
4	products were such a better fit with consumer tastes,
5	that if you have exclusive dealing, that is where the
6	dealers are going to go.
7	So, in that case, concentration would be
8	relevant as a screen or a way to think about what is
9	going to happen but through a much more complex chain of
10	reasoning than to just say, well, they have a high
11	market share; therefore, they must have market power.
12	It is really a very different kind of analysis, and that
13	is the kind of analysis that I think needs to be done.
14	Okay, the third one -- and actually, this is the
15	one that is my favorite -- is say, look, we need to
16	identify relevant markets, because if we are talking
17	about harm to competition, we need to have some sense of
18	who the competitors are, and actually, I think that is
19	what the role should be in the merger analysis I will
20	say as well, this really should be about identifying the
21	competitors and then seeing where that takes us in terms
22	of the but-for world, what needs to be the scope of the
23	but-for world, and this is an unfashionable view,
24	because it is low tech and it does not drive you to come
25	up with algorithms, but I think it is important to

72

1	remember in the end, this really is what we are trying
2	to do.
3	We are trying to figure out who are the
4	competitors, because then we can ask, does this practice
5	harm them? And if it does, does that matter for
6	competition and does it matter for consumer welfare?
7	Okay, so again, I think this takes us in a somewhat
8	different direction, and notice, in this one, you may
9	not be worrying about concentration very much directly
10	at all.
11	Also, since I had promised -- but so far have
12	not done it -- the organizers that I would talk about
13	innovation, let me say a little bit about that. When
14	innovation competition is really significant, and this
15	is not a point that is new to me by any stretch of the
16	imagination, current market shares may not tell us very
17	much, right, the extreme model being Schumpeterian
18	competition, where we see a string of product market
19	monopolies, but the real way competition works in the
20	industry would be that you have firms that come in with
21	major innovations, become the new monopolist, but then
22	there is this battle for the next round of drastic
23	innovation. If you are looking at a market like that,
24	looking at market shares is not going to tell you very
25	much.

73

1	Okay, a couple things about market definition
2	and uncertainty. First off, we have talked about burden
3	shifting a little bit. As everyone in this room who
4	works for the Government knows, right, meeting the
5	market definition burden can be difficult, and that is
6	true even if you do not have innovation, and I will come
7	back to innovation in a minute. One of the difficulties
8	is when courts say we want a zero-one boundary. Every
9	firm is either in the market or they are out of the
10	market; none of this wishy-washy stuff.
11	The problem with that is it can be really hard
12	to do. I know Oracle is a merger case, but it is really
13	striking because it is a case where Judge Walker said,
14	all right, look, here are the economics of why you
15	cannot draw zero-one boundaries. You have got product
16	differentiation. You have got a continuum of products.
17	There is no way there is going to be a sensible
18	boundary. He did not say, "And oh, guess what, that
19	means you lose."
20	I mean, I think Judge Walker was right about the
21	first part. It is just the notion that that is where it
22	takes you I think is a little troublesome. It is
23	particularly troublesome as well because if you believe
24	that these are differentiated product markets and you
25	believe competition is localized, then you really have

74

1	to ask yourself, why are we worrying about a broader
2	market anyway? I mean, what is the relevance of this
3	alleged relevant market if what really matters is
4	defined structure?
5	So, it seems to me that where we have gone with
6	a lot of -- just to jump back to mergers for a second
7	where I think there is a broader lesson here -- with
8	mergers, is worrying about unilateral effects cases in
9	markets with differentiation -- and everyone seems to
10	have conveniently forgotten that you can have a
11	unilateral effects case with homogenous products -- but
12	we have spent all this time worrying about market
13	definitions in precisely the wrong places.
14	Now, although this gets worse if you have
15	innovation, because you can have things constantly
16	changing, you can have products -- the characteristics
17	of products are changing, I just want to make two points
18	on this and then move on quickly. One, there are a lot
19	of people who seem to be of the belief that what
20	innovation means is markets are constantly getting
21	broader, okay, and there is a set of people who will
22	say, look, you have got all these things, you have got
23	innovation, markets are always going to be so broad
24	because new products can keep coming in, that really,
25	there is nothing for antitrust to do. I would just like

75

1	to remind people that, in fact, markets could be getting
2	narrower, because these products are evolving, they are
3	moving targets, and it is quite possible that some
4	products or the producers of those products are falling
5	behind in terms of innovation and they are dropping out
6	of the relevant market.
7	Okay, the point I have already made, that if you
8	are looking at differentiated products and then you
9	throw in the complexities of innovation, you just really
10	may make it impossible to meet the burden. As we have
11	talked about, since I think there is a fairly broad
12	consensus, you do not really need to have a rigid market
13	definition. That is unfortunate, but that is how a case
14	would be decided.
15	Now, I have to have a diagram. So, what this
16	one shows, just very quickly, suppose there is
17	disagreement on the scope of the relevant market here,
18	and I am interested in a case where I will just suppose
19	that one has beaten up on two, okay, these are suppliers
20	markets, and this line represents some notion of product
21	differentiation, and there is a debate. It is hard to
22	know whether the market boundaries are -- the ones who
23	have the narrow subscripts, so only include one and two,
24	or they have the broad, and then they would include
25	producer three as well. Suppose we get the debate down

76

1	to that level. This is a dramatic oversimplification.
2	Well, you can imagine a court, Judge Walker
3	saying, "Look, Government, you cannot tell me whether it
4	is the narrow definition or the broad one with
5	certainty, so you lose." But suppose it does not make
6	any difference whether you include three in the market,
7	okay, to what you think are the competitive effects,
8	then why does it matter that you cannot say which one is
9	which, okay? So, what you really want to ask is not
10	whether or not the plaintiffs can prove a market
11	definition with certainty, but you want to ask can they
12	tell you, "Look, we know well enough where it matters
13	with a high degree of certainty."
14	So, the approach to this would be to then ask,
15	"Where does the dividing line matter," okay? Go back to
16	this, "Does it matter whether we include five in or
17	not?" If it turns out what is critical in the end is
18	whether three is in the market, let's fight about that.
19	Let's not fight about, no, you have to come up with the
20	definition.
21	Okay, a quick thing on decision theory. I have
22	a pretty picture, I have to show it. What this is
23	saying is -- I just want to make the following point, I
24	probably will not actually go through the picture, so
25	just admire it while I talk. It was not easy drawing

77

1	this on the train while it was jerking around -- but is
2	the following, that there is a lot of focus, I think, in
3	court cases, at least, in actual legal decision-making
4	on doing things like asking are probabilities above
5	certain thresholds or is one probability higher than the
6	other, something like that.
7	This would be a diagram where if you weigh
8	evidence, you would just ask, is the probability of harm
9	bigger or less than the probability of efficiencies?
10	So, you would get in that red zone, because that's where
11	the probability of harm, P, would be viewed as being
12	higher than the probability of the efficiencies, Q, and
13	you would just sort of -- that is one interpretation of
14	weighing the evidence. There are others, I will note,
15	and if I had a longer time, I would tell you some of the
16	others.
17	Now, but if you try to balance the effects, you
18	do not just look at the probabilities. You have also
19	got to look at the magnitudes, and I have given the
20	example here where the harms, denoted by H, are bigger
21	than the efficiencies. So, in fact, you want to condemn
22	not just practices where the harm is more likely or
23	equally likely as the efficiencies. You actually want
24	to condemn some where the harm is less likely, but the
25	problem is, well, it is less likely, but when it

78

1	happens, it is a worse thing, and that is where you get
2	that purple area.
3	I would say in the end, since we are worried
4	about effects, the right thing to do, and if we do all
5	this stuff, would be to condemn this bluish-purple area
6	plus the red, but if you simply weigh the evidence, you
7	are only going to get rid of the red. So, you are going
8	to -- if there is enforcement, you are going to have
9	false negatives. So, I think what is important in all
10	of this, and there are many other interpretations of
11	this, but the central point is I think we do have to
12	worry about magnitudes more than we have in the last --
13	okay, are you going to unplug this? This is like the
14	Academy Awards, they start playing the music.
15	Innovation, I will say one thing in support of
16	innovation markets as a broad concept, because certainly
17	they have been controversial in terms of actually using
18	them, but if we are worrying about markets where
19	innovation competition is really critical, then we need
20	to worry about what is driving innovation, who the
21	potential innovators are, and looking at markets in a
22	product market may not tell you very much about it. It
23	may be much more informative to look at the distribution
24	of R&D capabilities and assets.
25	As some people, one of them sitting near me,

79

1	have pointed out, that can be really hard, because it
2	may not even be in this industry, but that is
3	conceptually the right thing to do, and so I think we
4	ought to be asking ourselves, how do we get there? If
5	we conclude it is too hard to do, fine, but I don't
6	think it makes sense to say -- and persons near me
7	didn't say this -- "Oh, it is too hard to do; therefore,
8	let's go and do something else that does not make any
9	sense but is easier." I think we want to keep in mind,
10	though, that the R&D capabilities and the distribution
11	of the assets there may be much more important than
12	current market shares in terms of understanding
13	innovation.
14	Okay, last thing, which does not have anything
15	to do with anything except people always screw it up. I
16	will make what has actually turned out to be a
17	controversial statement in practice, that geographic
18	markets are markets, by which I mean since they are
19	markets, they have buyers and sellers, okay? In
20	practice, at least my experience has been that people
21	often forget about the buyers part of that description
22	of markets, and then if we are going to talk about
23	geographic markets, we need to think about the buyers
24	and where they are and the sellers and where they are.
25	Now, in some markets, in the end, there may be

80

1	global markets and those do not matter, but other times
2	you want to ask something like, particularly in
3	retailing, say, or certain kinds of manufacturing, you
4	would want to say, let's look at a set of customers in a
5	particular city and ask what producers, and in
6	particular the producers' plants, can serve those
7	customers, and look at it that way.
8	Now, that may mean that a firm is in a lot of
9	different geographic markets, and a single plant, by the
10	way, could be in different geographic markets
11	simultaneously, which drives people crazy, but if you
12	want to think about what is really going on and take
13	markets seriously, you have got to remember, markets are
14	bringing together buyers and sellers, so we need to
15	discuss or describe the locations of both of those.
16	With that, I will stop.
17	(Applause.)
18	DR. CARLTON: Okay, thank you. The person close
19	to you says, "Thank you very much."
20	Okay, I would like to ask the panelists some
21	questions. We have about 45 questions left, and I have
22	a series of questions. I have about ten questions. I
23	do not know if we will be able to get through them all.
24	What I will do is I will ask the question, and then I
25	will ask two of you to comment. If you could keep your

81

1	answers relatively brief, that would be good. If
2	someone on the panel who we have not asked feels they
3	want to comment, they should do so, but since there is
4	an opportunity cost, that just means you may not get to
5	answer a later question.
6	Here is what it seems to me that the purpose of
7	these hearings are. One, we want to define market
8	power. Can we agree on a definition? If we can, do we
9	think defining the market and then taking market shares
10	helps us in a section case? Then, what are the hard
11	questions where we think that that may or may not help?
12	Then the ultimate question really is -- and this
13	I will ask everybody to answer, it will be the last
14	question -- do we really need market definition and is
15	it more of a hindrance than a help?
16	So, let me just start off on first asking the
17	question about market definition. In the legal
18	literature and in the cases, they stress not just the
19	ability to control prices, which is what economists
20	focus on, but they always add, "or the ability to
21	exclude competition," and it is that second prong I want
22	to focus on for a second.
23	I understand -- and Andy spoke a little bit
24	about this -- that a joint venture can get together and
25	exclude people. Let's just talk about single-firm

82

1	behavior, and I am interested, in particular, from both
2	Andy's point of view and Joe's point of view, with their
3	sort of combined economic/legal backgrounds, if they
4	could comment on whether they think the exclusion prong
5	of the market power definition that is used in legal
6	cases is useful. Do we need it? Can we do without it?
7	For example, can we do without it by saying,
8	"Well, it is the ability to control price, and if you
9	say keeping it above the competitive level, obviously
10	the competitive level is the level that arises when you
11	do not exclude competition." If we can simplify the
12	definition, it seems to me that helps things rather than
13	complicates things. So, is your view that we need that
14	second prong, exclude competition, in the definition of
15	market power or not?
16	So, let me first ask Andy and then I will ask
17	Joe, and if you could keep your answers sort of
18	relatively brief, that would be good.
19	DR. GAVIL: I think in exclusion cases, the
20	answer is yes, but it winds up being a first step. The
21	ability to exclude competition -- I guess the "or" is
22	the problem. Why do we have monopolization? We have
23	monopolization cases because we want to prevent not just
24	any exclusion of competition; it is exclusion of
25	competition followed by the ability to either maintain

83

1	price, maybe raise price, but the two to me go hand in
2	hand.
3	In any section 2 case, the first step is going
4	to be evidence of some exclusion, but I do not think you
5	can stop there and conclude from that that there would
6	automatically be monopoly power. You have to ask the
7	second question of whether or not the exclusion will in
8	some way facilitate the maintenance or the enhancement
9	of the market power. So, I think it winds up being
10	circular. You do come back to power over price.
11	DR. CARLTON: Okay, Joe?
12	MR. SIMONS: I agree with what Andy said, and
13	also, just to follow up on what Rich said about the
14	guide posts and the lamp posts. You know, what you see
15	in the case law is an example of a lamp post. It is not
16	an example of a guide post. That kind of definition is
17	drawn generally from whatever you guys refer to in the
18	equilibrium analysis or partial equilibrium analysis or
19	whatever it is, and they just moved it over and said,
20	"Here, this is what we are going to do," without
21	thinking about why we really want to do it.
22	The statute talks about monopoly, so you tend to
23	have to have a big share and so it is natural that a
24	share requirement gets imported into the law. But it
25	does so without thinking, and so I do not think that

84

1	focusing on that question based on the case law is going
2	to be terribly helpful.
3	I think Andy is right, you want to focus on why
4	are we asking this question, what are we trying actually
5	to prevent, what is the goal.
6	DR. CARLTON: Okay, I think I agree with that.
7	I think probably that is a fair summary of what you
8	said, that I think both of you say we can get rid of
9	that second prong as long as you keep your eye on the
10	ball. In effects cases, obviously you have done
11	something bad, and then did you raise price. So, if you
12	are wanting to define market power alone, it is whether
13	you can raise the price above what it would otherwise
14	be.
15	MR. NELSON: Or prevent it from falling.
16	DR. CARLTON: Or prevent it from falling, that
17	is right.
18	DR. GAVIL: I think the exclusion does tell you
19	something. I would not eliminate it entirely. I think
20	the problem is it does not tell you whether or not you
21	have monopoly power, but it is like the first red flag.
22	It is the first guide post that tells you there may be
23	reason to be concerned about a particular situation, but
24	you cannot stop there. You have to ask the second
25	question.

85

1	Even going back to Salop and Krattenmaker, the
2	title of the article was Raising Rivals' Costs to Obtain
3	Power Over Price. So, the two really do go hand in
4	hand, but the first sign of a problem may be the
5	evidence of exclusion.
6	DR. CARLTON: Right, but what is a mechanism to
7	achieve the control of price? I agree, it is important
8	to have both, but I am just trying to distinguish the
9	two. One of the things that goes on in a section 2 case
10	is you define markets and you have exclusion -- and I
11	will come back to this later -- and the question is how
12	you link the two. I am trying to keep them separate for
13	a second.
14	MR. SIMONS: I think in what Krattenmaker and
15	Salop do with their article is they are linked. It is
16	the exclusion that gives you the power over the price.
17	What is the impact of the exclusion? Not kind of in a
18	general sense, have you been able to exclude people, all
19	right? Because maybe you have because you have such a
20	terrific product or you have a patent or whatever it is.
21	That is legal. The question then becomes, did you do
22	something in addition to that that may not be so legal,
23	and does that give you power over price?
24	DR. CARLTON: Right.
25	DR. GAVIL: Think of instances where the act of

86

1	exclusion raises entry barriers.
2	DR. CARLTON: Yes.
3	DR. GAVIL: That leads you to the second part of
4	it.
5	DR. CARLTON: Yes and no. What that tells you
6	is that but for the act, which we are trying to claim is
7	illegal, the price would have been lower, and therefore,
8	you have the power to set price above the but-for price.
9	It is just defining what the but-for price is.
10	Okay, let me go on, because I am going to come
11	back to this benchmark point. The definition that
12	economists use a lot is that market power is the ability
13	to set price profitably above the competitive level,
14	presumably by a significant amount, for some significant
15	amount of time. So, first, I have two parts to this
16	question, and I am going to ask Phil and Larry.
17	Assume that there are constant returns to scale,
18	so competition is possible. So, first, do you agree
19	that the definition I gave you is a reasonable one --
20	put aside whether it is implementable, but is it a
21	reasonable one -- and if so, what is a significant
22	amount of the price increase and what is a significant
23	amount of time?
24	In particular, when you are answering, if you
25	could talk about why we do not pay attention to dead

87

1	weight loss and why we just talk about numbers. I mean,
2	we are economists, and 5 percent, 10 percent, we know
3	that may not be meaningful depending upon the size of
4	the market. So, if you could just address those.
5	DR. WHITE: Are you looking at me? Look, you
6	know, where do 5 and 10 percent come from? As Bill
7	Baxter used to say, from these (indicating hands), and
8	there's nothing magical about that. You know, it partly
9	would also depend on how much noise you think is out
10	there protecting ourselves against error that might be
11	harmful. So, the real answer -- the first part is yes,
12	under constant returns to scale, a price significantly
13	above marginal cost, sustained for a sustained amount of
14	time, would in my mind constitute an exercise of market
15	power, and how much and for how long, I do not know.
16	Sure, 10 percent sounds like a number to be
17	thinking about and two years sounds like a number to be
18	thinking about, but I have just picked those out of the
19	air, and I do not have any further basis.
20	DR. CARLTON: Okay, let me just say one thing.
21	My preference would be it is probably better -- even
22	though it is hard to choose a number, someone is going
23	to choose a number, so you should think, as to your
24	willingness to choose a number, would you rather some
25	random judge choose a number or this panel? So, that is

88

1	why I am asking.
2	DR. KATZ: I mean, I disagree with the premise.
3	Why should you choose a number? I am almost
4	certainly -- if you thought the court was going to do
5	enough of the analysis -- and we would have to talk
6	about the cost of the court and the time they have --
7	but almost certainly you would say the number depends on
8	the market. I mean, there are some markets where
9	worrying about a price change within 5 or 10 percent, I
10	mean it is completely lost in the noise, because the
11	prices are changing 40 percent every year, so it does
12	not mean a 10 percent price increase could not matter,
13	but it becomes less plausible you could actually tell.
14	In other markets, it might be that you could reliably
15	predict a 3 percent price change.
16	DR. CARLTON: Following that same logic,
17	wouldn't you be concerned about a 1 percent change in a
18	market that is huge?
19	DR. KATZ: If you believed you could actually
20	make reliable predictions at that level, yes. So, I
21	think you need to look, as you were saying, at the
22	magnitudes of the effects, and some of it comes within
23	when do you want to bring cases and how to allocate
24	resources and then also the various characteristics of
25	the market that are going to affect the reliability of

89

1	your projections and whether you think that you really
2	can discern at those levels, but I think it would be
3	pretty clear that holding aside -- which is obviously a
4	big thing to hold aside -- the various sorts of
5	processing costs, there is no reason to think there is
6	one right number, and, in fact, there certainly isn't.
7	DR. CARLTON: The question is, should we give
8	any guidance to the courts when they are trying to
9	decide whether a firm has market power, and if you just
10	say it is up to the discretion of the judge based on a
11	lot of things -- I mean, I agree with you, it is hard to
12	come up with one number. The question is, is it better
13	leaving it completely to the discretion of the courts,
14	or should we not -- I think one of the advantages of the
15	Merger Guidelines, even though they make the point that
16	the 5 percent is just a suggestion, is that it has
17	focused thinking and clarified thinking. So, I agree
18	with everything you have said, but in light of the
19	decision-making of the court process, there can be a
20	benefit to articulating some standards, maybe flexible
21	standards.
22	DR. KATZ: I would agree with that, but I think
23	a question would be -- and this is just thinking off the
24	top of my head -- could you say something like -- have
25	some sort of relatively easily observable data, say like

90

1	the annual price changes or something, or try and do
2	something that says that the standard you use should be
3	proportional to some characteristic in the market? We
4	would have to think a lot about what that is, and I
5	think ideally, for the reasons you bring up, it would be
6	something fairly mechanical, but it would still be an
7	improvement over a one-size-fits-all.
8	DR. CARLTON: Rich?
9	DR. GILBERT: Well, I certainly agree that the
10	number, however you define this number, depends on the
11	nature of the conduct, the efficiencies that can be
12	presumed to go along with that conduct, and maybe the
13	size of the market and all of that, but I also think
14	there is the case that can be made for shifting the
15	inquiry to something like the firm-specific elasticity
16	of demand, which often can be measured in many
17	instances. I think Greg has pointed this out in some of
18	his writings.
19	It is not that hard to say if the elasticity of
20	demand is bigger than 10, maybe we shouldn't be worried
21	about this. On the other hand, if it is in the range of
22	2 to 3, maybe we should be worried about this.
23	DR. CARLTON: Yeah, that raises a point I am
24	always puzzled about, that if you are thinking about
25	what is a magnitude that is important, an elasticity of,

91

1	say, 20, which everybody would say is a really high
2	elasticity, that gives you a 5 percent upcharge over the
3	competitive price. So, that should tell us something
4	about our intuition versus sort of practical --
5	DR. GILBERT: Well, on that, maybe I am
6	differing from other people, I think of that 5 percent
7	rule as being a derivative, not an absolute amount. So,
8	we ask, if quantity goes down by 5 percent, will the
9	price go up by 5 percent, that sort of thing, and rather
10	than because we are really worried about the price going
11	up by 5 percent. Now, some people I know would disagree
12	with that and would say that that 5 percent is a
13	threshold of concern. I think of it more as an
14	elasticity test.
15	DR. CARLTON: Okay.
16	MR. NELSON: Since I was one of the original --
17	DR. CARLTON: I am going to give you another
18	question, okay? It is actually a harder question now.
19	We are going to move on to something else. That was an
20	easier question. If you remember, that was premised on
21	constant returns to scale. So, it could actually define
22	a competitive price.
23	Let's suppose now that I am in an industry where
24	there cannot be competition. There is a fixed cost of
25	entry. There are constant returns to scale, and it is

92

1	Cournot competition, okay? What is the meaning of that
2	common phrase that we use, can you profitably price
3	above the competitive level? What in the world should
4	we take as the competitive level in that situation? Is
5	it the zero profit equilibrium or is it price equaling
6	marginal cost?
7	So, let's see, maybe Phil, if you want to take a
8	crack at that.
9	MR. NELSON: Well, one of the things that sort
10	of concerns me about taking sort of the current level as
11	opposed to something like marginal cost is you do have
12	some of these monopolization cases that are really
13	entrenchment theories, and is the question whether the
14	entry is going to drive you significantly back towards
15	competition, or this guy already has some market power,
16	and he is going to --
17	DR. CARLTON: Try to define the market
18	equilibrium, free entry, fixed costs, constant returns
19	to scale, Cournot equilibrium, do we want to call that
20	market power?
21	MR. NELSON: I guess I am saying that to answer
22	that, you want to know sort of what your benchmark is as
23	to where you're going.
24	DR. CARLTON: Right, that is what I am asking
25	you.

93

1	MR. NELSON: Yeah, and what I was going to say
2	is that I think you would start to look, as N goes up,
3	what happens to the equilibrium price? Then as N gets
4	high enough, are you still at a price where somebody
5	could make an economic profit? I mean, you are going to
6	want to see if that is a tenable number of firms and
7	start to use something like that as the equilibrium,
8	which is higher. It is going to be a lower price and
9	define market power in some circumstances where you
10	might not find it if you are at your starting point.
11	DR. CARLTON: So, the point of the question is
12	to show that there is a difficulty in defining market
13	power when you cannot define the competitive price. You
14	can define a rate of return, and you can define marginal
15	cost in this example and prices above marginal cost in
16	this example, but profit is zero, and there seems to be
17	a complete ambiguity between the willingness of people
18	to distinguish which of those two definitions they are
19	using.
20	Is it price above marginal cost that is market
21	power? Is it rate of return above a competitive level,
22	or which of the two, or are those two different things?
23	They obviously from an economic point of view are two
24	different things, yet often, in the writings and in case
25	law, they in my mind do not get distinguished.

94

1	MR. NELSON: I mean, yeah, you want to have
2	profit -- you want to be able to make a monopoly profit.
3	I mean, if you have got easy entry, as some of the
4	different -- you know, if you don't have any profits,
5	then they are not going to have enough -- but I --
6	DR. CARLTON: Larry, did you want to say
7	something?
8	DR. WHITE: But why would we be interested in
9	your hypothetical? If it is somebody coming in and
10	saying, "That guy is charging an outrageously high
11	price, Judge, find him guilty of a section 2 violation
12	and mandate that he charge a lower price," we do not see
13	that all that often, but that would be a problem. If it
14	is, "Judge, that guy has excluded me from offering my
15	rivalrous product, and had he not excluded me, I could
16	have come in and the price could have been lower,"
17	that's a different --
18	DR. CARLTON: I agree, but that is mixing
19	together two different questions. The first is, what is
20	the effect of this action? If you can answer that
21	question, you have answered the section 2 -- you have
22	resolved the section 2 issue.
23	DR. WHITE: And then we do not have to worry
24	about it.
25	DR. CARLTON: And then we do not have to worry

95

1	about market definition; however, the way the courts
2	seem to use market definition in section 2 cases is not
3	like that at all. Courts seem to do the following:
4	Unlike a merger context where you ask, as a
5	result of a merger, is market power going to go up, the
6	courts define a market and then look to define market
7	share. Courts do it. They do not look at the change in
8	the market shares that arise as a result of the bad act.
9	They do not do that. That would be an analogy to a
10	merger case.
11	Instead what they do is they ask, is there
12	market power? They do not ask about the change in
13	market power, but they ask, is there market power? They
14	use that as a screen whether to then further
15	investigate, and that distinction, that asymmetry
16	between a merger case and a section 2 case, I think
17	leads to peculiar discussions, but it also I think leads
18	to exactly why I am asking this question, which is, if
19	the courts are going to go this route and use market
20	definition -- I agree with you, Larry, if you do an
21	effects-based analysis, you can solve the problem -- but
22	the first question the court is going to be asking, is
23	there market power, and I am just trying to figure out,
24	can we even define what we mean by that in this Cournot
25	example?

96

1	MR. SIMONS: I think what you want to ask is why
2	are we doing this, why are we engaged in an exercise,
3	before you can even think about answering the question.
4	DR. CARLTON: This firm has been sued, there is
5	a bad act, and the first question is, does he have
6	market power? And I am trying to find out -- I cannot
7	answer that -- begin to answer that question unless we
8	can agree on a definition of market power. So, is the
9	definition price above marginal cost or is the
10	definition rate of return above the competitive level?
11	Mike?
12	DR. KATZ: The problem is if you are going to
13	say this has to be a screen that works for everything,
14	then the most useful definition of market power would be
15	does the firm have at least one employee or something
16	that is equivalent of it so we throw this screen out,
17	because what Larry has pointed out -- I think what in
18	most cases makes sense is something that says -- and
19	actually, I make a different distinction, and I think,
20	actually, a lot of economists writing not as part of
21	antitrust make a different distinction. I think a lot
22	of people, economists, would say that market power would
23	be facing downward-sloping demand curve and not having
24	it perfectly elastic, which then would end up giving you
25	the profit-maximizing price of that firm above marginal

97

1	cost. I think that is a useful definition of market
2	power.
3	Then I try to reserve monopoly power for being
4	two parts. One, that you have a lot of market power,
5	which is to say the price would be -- and again, I will
6	be vague -- but significantly above marginal cost, and I
7	would typically put in a test saying for I think most
8	purposes or a lot of them, we do care whether or not the
9	price is above average cost, whether or not there are
10	profits, but what Larry has pointed out, I think
11	correctly, is that test does not always work, that if
12	what you are worried about is somebody who is in there
13	now and is just breaking even but is narrowly keeping
14	all sorts of more efficient entrants out who could make
15	a profit, I think in that case, saying, "Well, look,
16	they are not making money, there can't be a problem,"
17	would give you a misleading answer.
18	MR. NELSON: That was my standards example I
19	gave in my opening talk.
20	DR. CARLTON: I think, again, that is really
21	asking the but-for price; in other words, price may
22	equal marginal cost and price may be above average cost
23	in the present environment, but but for the bad act,
24	that would not occur, okay?
25	The distinction you make between price above

98

1	marginal cost and then whether the rate of -- the price
2	above average cost, the rate of return is above the
3	anticipated return, is exactly the distinction that I
4	made between market power and monopoly power. It is a
5	logical distinction. I am not sure -- we may be the
6	only two people who make that distinction, because I do
7	not see the legal cases going in that direction.
8	So, I guess I do have a question, and I think it
9	is a relevant question, as to whether the distinction
10	between monopoly power and market power that we do see
11	in the cases, is that a useful distinction, and is it a
12	useful legal distinction? Is it a useful economic
13	distinction?
14	So, maybe, Andy, you could answer that and maybe
15	Rich.
16	DR. GAVIL: Yeah, one point I wanted to make
17	earlier and I think I can make it now in answering the
18	question, I think historically the association of market
19	power and monopoly power as being different things was
20	linked to market share. It was linked to circumstantial
21	evidence as the basic mind set that we used to approach
22	cases, and I think a concrete example of this is the
23	Supreme Court decision in Copperweld, where it says --
24	it is known for the parent/subsidiary enterprise
25	conspiracy issue, but it has a very interesting

99

1	discussion of the relationship between section 1 and
2	section 2, and it says, "An unreasonable restraint of
3	trade by a single firm is not reached under section 2,
4	and therefore, the drafters of the Sherman Act left a
5	gap between section 1 and section 2, and the implication
6	was that for a section 2 case, you need something more."
7	At that moment in time, the "something more" was the 70
8	percent or more market share as opposed to the 40 to 60
9	percent that was typical in rule of reason cases.
10	If you let go of the commitment to the Alcoa
11	framework and the market share associations and start
12	thinking about market power in different ways as
13	expressing itself in different ways, that kind of mind
14	set of distinguishing market and monopoly power based on
15	market shares goes away, and I think that that would
16	make a big difference in how we think about antitrust
17	generally.
18	But you have said it several times, Dennis, and
19	it is clearly the case, that courts say, "Okay, the
20	first element under section 2, do you have monopoly
21	power?" On your no profit example, if I could just
22	throw in, what if the purpose of the conduct was to make
23	that firm profitable and that is what it was trying to
24	do? So, currently it is not profitable, but the whole
25	point of the conduct, maybe it affects entry barriers,

100

1	was that they are trying to get profitable by engaging
2	in conduct. Again, I think it shows the link between
3	the conduct and the power increase.
4	DR. CARLTON: Rich?
5	DR. CARLTON: Rich?
6	DR. GILBERT: Yeah, if I can answer this, as has
7	been said before, in some sense I subscribe to the
8	argument that monopoly power is a lot of market power,
9	but it is also market power that is durable. Now,
10	whether you define durable market power in terms of the
11	ability to raise price above average cost, the ability
12	to maintain price above average cost, or the ability to
13	maintain price above long-run marginal cost, I do not
14	think that is really critical. To me, it is the ability
15	to exclude, and as you have noted, Dennis, yourself,
16	that when you are talking about exclusion, obviously it
17	also depends on thinking about entry barriers, and then
18	when you think about entry barriers, you have to think
19	occur.
20	So, if you had an extremely competitive market
21	post-entry, maybe a little bit of exclusion is enough to
22	maintain a monopoly position, but I think the key issue
23	is the ability to exclude is important to me, because it
24	says something about the ability to maintain price above
25	some measure of long-run profitability of an efficient

101

1	competitor.
2	I want to add one other comment that I think is
3	related to all of this, which is we are very good when
4	we talk about impacts on competition to understand that
5	impacts on competition is different from impacts on a
6	competitor, I think we have learned that one, but when
7	we talk about section 2 cases, we are often talking
8	about the market share or monopoly power of a single
9	firm. Shouldn't we be talking in many of these cases,
10	at least, if not all of them, about the power in the
11	market, not just the power of this firm, because
12	obviously if the firm reduces supply so that its market
13	share is below the Alcoa threshold, but in doing so,
14	raises market power generally in the industry, that is a
15	problem, and we want to look at that, not just what the
16	firm's market share is or focusing on the firm.
17	Now, if you did this firm-specific residual
18	demand analysis, then you pick that up by looking at the
19	elasticity of the residual demand. So, I think it is
20	all right in that context.
21	DR. CARLTON: Let me go back to something that
22	sort of was a common theme in some of the presentations.
23	Let me restate it as follows: It really has to do with
24	what the benchmark price is.
25	If you look at a firm in a section 2 case and it

102

1	is engaged in a bad act, can you then ask, "Well, does
2	that firm have market power," which is what the courts
3	first ask, and if the answer is no, they throw it out.
4	In order to answer that question, you have to ask,
5	"Well, what would the price --" depending on your
6	definition of market power, you want to ask, "Does the
7	firm have market power?" Whatever your definition is,
8	whether it is pricing above the competitive level after
9	the bad act, are they pricing above the level but for
10	the bad act, whatever definition you want to use, and I
11	think it is the latter definition that makes more sense,
12	it is not obvious why market shares and market
13	definition help you answer that question, because if you
14	know the current price and you knew the benchmark price,
15	it is just a comparison of two prices. So, calculating
16	market share in that case does not advance the ball.
17	If that is the typical case that we see in
18	section 2, what really are we talking about when we are
19	doing market definition? Are we really doing an
20	analytic economic exercise, or are we doing something --
21	or are the courts doing something much more -- I don't
22	want to say sensible, but much more using common sense,
23	which is there are five guys doing the same thing, don't
24	bother me, and they're just using their common sense.
25	Now, how you define "five guys doing the same

103

1	thing" may be hard, but it seems to me that is what a
2	lot of courts are doing, and I am wondering if we are
3	worrying too hard about defining markets in cases where
4	market definition is just this seat-of-the-pants thing
5	that the courts then use, and as long as they understand
6	it is real seat-of-the-pants, don't bother me with
7	details about market shares and get on to your
8	competitive effects analysis.
9	So, maybe, Joe and Mike, you could comment on
10	that.
11	MR. SIMONS: Yeah, I think that what the courts
12	will do is not just say, well, let's get on with it and
13	let's get to the competitive effects. It is a real
14	screening event, a big one, and it also seriously
15	impacts what happens when lawyers counsel their clients.
16	If there is some chance that your client is going to be
17	deemed to have a big market share, at least most lawyers
18	I know will give advice that is much more conservative
19	than if their shares are 30 percent. So, it makes a big
20	difference in the real world.
21	I think the judges do focus on it, and it is
22	important in court now, and there is a serious question
23	in my mind about how important it should be, certainly
24	with respect to how the Antitrust Division and the FTC
25	exercise their prosecutorial discretion -- whether they

104

1	really need to get hung up on this or whether they
2	really need to make a decision about what is the impact
3	of whatever conduct we are worried about. Did it have a
4	significant impact, and then, when we are proving in our
5	case in court, it is a different exercise.
6	DR. CARLTON: Okay, who did I say? Mike?
7	DR. KATZ: You are not supposed to remind me of
8	that. No, I would say a couple things. Part of it -- I
9	will come back to what I said in my presentation,
10	though, is that we can be using market definition in a
11	number of different ways and that the level of -- we
12	want to understand who the competitors are, because we
13	want to figure out that is where we are going to see the
14	competitive effects, are they harmed or not, does it
15	matter if they are harmed, does it matter for
16	competition and for consumer welfare. So, that level, I
17	think we would certainly want to do market definition,
18	but that may not be through a formal algorithm.
19	In terms of your question about the alternative
20	prices, I think there is a difference between asking
21	about a but-for price and asking about certain
22	interpretations of the competitive benchmark, because
23	you can take a competitive benchmark to be some formula
24	based on marginal cost or average cost or something like
25	that, and you could ask a market power question, but you

105

1	could well conclude from that that yes, this firm has
2	market power, but it has not done anything wrong, okay?
3	It has market power because it has been a great
4	innovator. So, I think that that is a different
5	question than asking is it charging a higher price than
6	the but-for price, because there you are asking about
7	what would happen as a result of the specific piece of
8	conduct. So, I think if one wants to go through the
9	market definition exercise in that form and to have the
10	competitive effects analysis be different than the
11	market definition analysis, I think you can do it. One
12	would ask about almost this more abstract or formulaic
13	competitive price as the benchmark for market
14	definition, and then the competitive effects analysis
15	would look for a but-for price and would take into
16	account a specific practice.
17	DR. CARLTON: I just don't know how to do the
18	first in an analytic way that is other than comparing
19	the two prices. If I cannot compare the two prices and
20	I have to do a competitive effects, say an econometric
21	analysis, I do not really need market definition. So,
22	that is why I think what judges often do is, as Joe
23	described, is do a seat-of-the-pants analysis or I
24	described as a seat-of-the-pants analysis, but as Joe
25	described, that is their screen.

106

1	DR. KATZ: Well, the screen makes sense if what
2	the plaintiff is saying is there has been successful
3	monopolization and you end up coming out of this being
4	convinced that here is a sensible market definition that
5	tells me how competition works, and this particular firm
6	does not seem to be dominant in any sense or doing well,
7	and if you see that, then it seems to me it does pretty
8	heavily undermine the claim that there was successful
9	monopolization.
10	DR. GILBERT: But that's Cellophane. I mean, it
11	is Cellophane, did not look like they were --
12	DR. KATZ: No, Cellophane, they did not have the
13	sensible market definition.
14	DR. GILBERT: Then you are back down to how do
15	you define the market.
16	MR. SIMONS: Dennis, think about it this way:
17	Someone had mentioned a gap earlier, and maybe it was
18	Andy. If you think about under section 1, right, if you
19	prove an effect, you win, right? Under section 2, the
20	way you are describing the state of the law, which is
21	accurate, is it is unilateral conduct. No contract.
22	There may be an effect, and then liability is going to
23	turn on whether there is some high market share.
24	DR. CARLTON: Right.
25	MR. SIMONS: So, the question to me would be,

107

1	why do you want to do that?
2	DR. CARLTON: I think that is right. Let me
3	flip the question a bit.
4	It seems to me this emphasis on market
5	definition in section 2 cases is coming precisely
6	because of the way judges apply these screens and
7	that -- I cannot remember -- I think Andy mentioned it,
8	that just like you might want to have decision processes
9	based on market shares, you might also want to immunize
10	certain types of safe -- have safe harbors and as well
11	as have danger zones, and it is the fact that it seems
12	to me that the sequential decision-making in Section 2
13	cases is first to look at market definition as a screen
14	and then you go to competitive effects that causes this
15	undue emphasis on market definition, and one way around
16	that might be -- and this is going to be a question I
17	will ask Andy and Phil -- suppose we also allow a screen
18	based on safe harbors and said, "No section 2 cases if
19	you're doing X, Y and Z; no section 2 cases if market
20	share is -- you do not have market power."
21	Wouldn't that be a way to de-emphasize the role
22	of market definition, which I think we are all agreeing
23	is difficult to define in a section 2 case?
24	Let me first ask Phil, and then he --
25	MR. NELSON: Okay, wait a second, no market --

108

1	DR. CARLTON: Either you do not have market --
2	the current screen, but I am going to add to that
3	current screen that there are certain safe harbors, and
4	what we should do is spend more time on defining the
5	safe harbors for conduct rather than trying to figure
6	out can we define markets any better.
7	MR. NELSON: So, it is conduct safe harbors,
8	not --
9	DR. CARLTON: Correct, yes.
10	MR. NELSON: -- not structural safe harbors.
11	DR. CARLTON: Correct.
12	MR. NELSON: Okay, there was an "and" there, and
13	I was starting to think that maybe where you wanted to
14	go was a combination of a structural safe harbor with a
15	conduct safe harbor, because there are certain types of
16	conduct that might mean a lower market share, you could
17	still have a problem, like some of these -- but I think
18	there is a -- as I was alluding to, the importance of
19	performance evidence, which is another way maybe of
20	saying conduct, that you would want to start looking at
21	some of that conduct evidence.
22	However, I am a little worried that the problem
23	is that a lot of this conduct is not so easy to
24	categorize, so that when you start to try to define a
25	safe harbor using conduct evidence, I am not sure that

109

1	you are going to find a lot of situations where you can
2	say for sure that this is conduct that is absolutely
3	okay, because in other contexts, it may not be.
4	DR. CARLTON: I agree. Safe harbors make
5	mistakes, but that is what decision theory tells us is
6	the right thing to do.
7	Andy?
8	DR. GAVIL: I think the idea of defining safe
9	harbors and danger zones, as I said, is useful, and I
10	think you cannot do it just by using market share
11	numbers, which has been our tendency in the past.
12	Now, once you use a market share number, you are
13	stuck in the, "Okay, we need to define a relevant market
14	problem." So, I think that reducing it to certain
15	characteristics of the market, maybe it is structural
16	characteristics, performance characteristics, but trying
17	to look at other kinds of measures might make the safe
18	harbor and the danger zones a little bit more meaningful
19	and move the attention away from market share.
20	But one last comment, Joe said how this affects
21	you in advising clients. That 70 percent share that has
22	become pretty fixed for monopolization cases, that is
23	perceived, even by defendants who do not like the market
24	definition and market share approach analysis, that is
25	perceived as a pretty big, significant safe harbor when

110

1	you are advising single-firm clients, and keep that in
2	mind with all the monopolization cases.
3	If there really is not any good, strong argument
4	that you could be in a market with a market share that
5	is up in that range, you are pretty much free to do
6	whatever you want. So, if we moved away from it, I
7	suspect you would actually hear some objections from
8	large firms that perceive that it is actually a very
9	useful benchmark.
10	So, where I would come out is, I do not know
11	that you can completely get away from the market shares,
12	but maybe we need something like a market share plus,
13	and not that it is a great model that we would ever want
14	to rely on, but the concept from the Sentencing
15	Guidelines that you have a guideline, but then you have
16	factors that allow you to depart upward or downward, and
17	that is sort of what Michael was talking about earlier
18	when he was answering one of your questions, is certain
19	factors might lead you to be cautious or less cautious
20	in certain circumstances.
21	DR. CARLTON: Let's see, let me skip a few
22	questions since I am going to try and end roughly on
23	time or maybe at most five minutes late, but let me ask
24	a question about technology, and I am going to direct
25	this to Rich and Mike, because you have done a lot of

111

1	work in these areas.
2	It seems like a really hard question is where
3	you have industries where marginal cost is low, product
4	innovation is the key, and new products are coming out
5	every so many years. It kind of came up a bit in the
6	Microsoft case, and Dick will probably talk about this
7	tomorrow, Schmalensee, but what do you mean by "market
8	power" in those industries unless "durable" really means
9	more than a year or two? Is that the right thing to be
10	focusing on? If it is, if it is focused on in those
11	industries, is it -- let me rephrase it. Is our focus
12	on price misleading us and should we be focusing on
13	other things?
14	DR. GILBERT: Well, I do not view that -- you
15	have asked a lot of hard questions. I do not think this
16	is one of the hardest questions. I find it relatively
17	straightforward in the sense that when you are talking
18	about dynamic competition, Schumpeterian spiral
19	competition, it is very much like thinking about entry
20	analysis. There is some probability that entry will
21	occur at some date. The question is how soon will it
22	be, what will be the magnitude of it.
23	There is also I think a legitimate question that
24	even if entry is going to occur, is that going to
25	neutralize the type of conduct we are concerned about,

112

1	or does the conduct we are concerned about promote entry
2	or retard entry? It is my view that these are questions
3	that can be addressed within the context of the way we
4	do antitrust analysis generally.
5	Now, it is, of course, the case that in high
6	technology industries, you are more likely to get very
7	high price-cost margins, so you are more likely to be
8	worried about market power, but it is often benevolent
9	market power, and if it is benevolent, you should not be
10	doing an antitrust case. So, it is more like magnifying
11	the things that we are concerned about but not changing
12	the qualitative way that in my view you should take them
13	into account when you are doing an antitrust analysis.
14	DR. KATZ: I have a couple of things and maybe
15	tie it to Microsoft. I mean, one of the things to
16	remember is when the Government was looking at
17	Microsoft, when the Government was dealing with them in
18	the mid-nineties, everybody pointed out, "Well, look,
19	it's such a fast changing market, and yes, it is true
20	that Microsoft dominates personal computer software
21	today and Apple is a distant second, and there are these
22	other things that aficionados use, but they really have
23	not caught on, and now here we are 12 years later and,
24	okay, all of that's the same." So, this whole thing
25	about the fast-paced -- and certainly Linux is doing

113

1	better than, you know, "the operating system" did, but
2	sometimes we do tend to exaggerate the rate at which we
3	think markets change and certainly relative to the pace
4	of antitrust enforcement.
5	But the other thing is, I think, Microsoft I
6	think is an interesting case, and maybe it comes back to
7	one of Larry White's points, that the Microsoft case, I
8	think it is fair to say that both sides took a
9	Schumpeterian view. Microsoft said, "Look, this is
10	Schumpeterian competition, someone else could come
11	along, they will displace us, because of network
12	effects, you would expect the winner to get a very high
13	share in operating systems, and so leave us alone,
14	because that's how competition occurs," and the
15	Government said, "Okay, look, this is Schumpeterian
16	competition. If you guys didn't do bad things --" Greg
17	is shaking his head. Now, it is true, sometimes the
18	Government didn't say that, but I think the only
19	sensible interpretation of what the Government was
20	saying was, "This is a Schumpeterian market, and,
21	Microsoft, you are trying to stop the next wave of
22	innovation that would have displaced you," and I am
23	saying that's somewhat like Larry's point about saying
24	it is not just how well you are doing in some absolute
25	sense, but whether there is a threat that you are trying

114

1	to stop that would leave you worse off.
2	So, I mean, I think in that one is the
3	Schumpeterian view was consistent with saying we have to
4	intervene, although it does shift what you look at.
5	DR. CARLTON: All right. Well, we are about out
6	of time, so I want to end with this, to get everybody to
7	comment on this question.
8	In light of all the difficulties and ambiguities
9	with the use of market definition in section 2 cases, is
10	it your view that we should still rely on it as we do,
11	put less reliance on it, or go to a competitive effects
12	and forget about market definition? So, why don't we
13	just go down the table in order.
14	MR. NELSON: Okay, I think I am halfway between
15	your two extremes, because I think there are -- as I say
16	in my slides, I think that there are organizing
17	principles and things that the exercise -- the market
18	definition exercise helps you understand what is going
19	on and tell either a story or an analysis that is
20	internally consistent, but that is not to say you have
21	to do it in every case, and there are numerous cases
22	where you may be able to expedite things by going
23	straight to the competitive effects bottom line.
24	MR. SIMONS: My take would be that the DOJ and
25	the FTC should try to come up with something that

115

1	focuses only on competitive effects, does not worry
2	about market share, and then see what happens over time
3	in terms of what they come up with and how operable it
4	is. And if the thing really works, terrific, then try
5	to get it into the courts, but not worry about that at
6	the outset.
7	DR. WHITE: Yes, we ought to be looking -- I
8	have a feeling we are going to be having all of the
9	divergence of opinion ranging from A to B. Yes, you
10	ought to look at competitive effects more than we have,
11	but I think there is still going to be a role for market
12	definition. Think about the private suits, not the
13	government suits, but the private suits that were
14	brought against MasterCard and Visa, and these were --
15	you know, the -- say take a WalMart case. This was a
16	tying case, but they were not -- it was -- you could
17	tell some stories about how if the tie was not there,
18	there were -- there would have been more entry somewhere
19	in -- in the credit card markets, but it was primarily
20	the tie is preventing us merchants from doing something
21	we would like to do.
22	I am not sure a competitive effects analysis is
23	going to tell you about market definition in that
24	particular case. Of course, MasterCard and Visa were
25	telling you, oh, all kinds of transaction media are in

116

1	the market, you know, cash and checks and everything, we
2	do not have market power, and they were actually trying
3	to say it with a straight face.
4	A market definition paradigm I think would help
5	in that kind of case, and so yes, I think we still have
6	need. I am hoping this is 1981, and next year, some
7	smart people are going to come in with a useful
8	paradigm.
9	DR. CARLTON: Andy?
10	DR. GAVIL: I think I agree with Larry. I think
11	it still has a role to play, but I think as you stated,
12	I think I would agree also that we over-rely on it, and
13	I think somewhat complex is the problem of
14	over-reliance. I think it can lead to both false
15	positives and false negatives, but I think with the
16	false positives, if somebody is found to have monopoly
17	power, to some degree, you have the backstop of the
18	conduct inquiry.
19	My concern is because of the high process costs
20	in trying to prove monopoly power in this -- as you
21	described it accurately -- sequential model, you get
22	false negatives, and there is no backstop to that. The
23	case ends, and the court does not look at conduct, does
24	not look at effects. So, I think this is an example
25	where the over-reliance may actually increase the threat

117

1	of false negatives more so than false positives.
2	DR. GILBERT: I would join the chorus that we
3	need more emphasis on competitive effects. A good
4	example, not necessarily really in the section 2
5	context, is the Oracle merger case where in my view
6	there was some certainly interesting evidence, if not
7	dispositive evidence, about competitive effects, but
8	once Judge Walker determined that he could not define a
9	market, he then concluded that there was no market, and
10	the competitive effects were almost ignored in that
11	case, and to me, it is like saying that I do not know
12	exactly where downtown Los Angeles is, and therefore,
13	there is not one. But I also can sympathize that if we
14	did away with market definition completely, it could be
15	highly problematic in leading to a lot of cases.
16	DR. KATZ: Okay, well, I guess I will say, at
17	the risk of sounding like Bill Clinton, it depends on
18	what one means by doing market definition, and I think a
19	lot of times what people mean is they mean applying the
20	hypothetical monopolist test, they mean doing a
21	concentration analysis, and they mean trying to come up
22	with boundaries with certainty, and then slavishly
23	applying that, and if you cannot meet that, you throw
24	the case out.
25	I think that way of doing things is surely

118

1	wrong, but I think we also surely do want to do some
2	sort of market definition exercise in the sense of
3	identifying the competitors, and I think where we have
4	come up short is what is the right way to do it in the
5	middle in terms of I think we still do not have a very
6	good sense of what is the right algorithm and the right
7	approach in different situations.
8	We have not mapped out, so, here is exactly
9	where you could do the hypothetical monopolist test,
10	here is where we need to do some alternative
11	methodology. We do not have that, and I think the
12	courts -- sometimes, the fact that we do not have that
13	has become an obstacle to good decision-making, as Rich
14	was just saying in the Oracle case, but I think the
15	bottom line is we need to figure out a better way to do
16	market definition, and that way we will recognize that
17	it should not be taken overly seriously or applied too
18	mechanically.
19	DR. CARLTON: Thank you very much. I want to
20	thank the people at the Department of Justice and the
21	FTC who did all of the legwork in putting this together,
22	and I am sure, although I have not checked with them,
23	all of the panelists, myself included, thank Larry White
24	for not including us in his slide of dumb quotes, and I
25	want to personally thank everybody on the panel for

119

1	coming and giving us the benefit of their substantial
2	expertise. I have learned a lot, and I thank you all.
3	(Applause.)
4	(Whereupon, at 12:36 p.m., a lunch recess was
5	taken.)
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

120

1	AFTERNOON SESSION
2	(2:01 p.m.)
3	MR. WALES: Well, good afternoon, everybody.
4	Thanks so much for braving the cold and the snow. I
5	think we have a very exciting panel on tap for this
6	afternoon. For those of you who were here for this
7	morning, I hear it was very lively, and I am hoping that
8	we can live up to that and be lively as well.
9	My name is Dave Wales. I am a Deputy Director
10	here in the Bureau of Competition at the Federal Trade
11	Commission. I will be moderating today, along with Greg
12	Werden, the panel. Greg is the Policy Project Director
13	At the Economic Analysis Group at the Antitrust Division
14	of the Department of Justice. That is his official
15	title, and you will be hearing more from him shortly.
16	Have you been elevated perhaps?
17	DR. WERDEN: I have never even heard of that
18	title.
19	MR. WALES: Is it better than the one you have?
20	DR. WERDEN: No, not really.
21	MR. WALES: Not really? Sorry about that, Greg.
22	Before we get into substance, it is my job to do
23	a little bit of the housekeeping. First off, what I
24	would like to do is on behalf of the FTC, to really
25	thank our friends at DOJ in putting this together, and I

121

1	think it has been phenomenal to date, and I am sure it
2	will continue to be that way. I would also like to
3	thank each of our five panelists today for, again,
4	braving the weather and coming out, and I think we have
5	got some great issues to discuss.
6	Today and tomorrow, I guess today and tomorrow,
7	we have the hearings on monopoly power, and I guess what
8	I wanted to point out is that next month we will be
9	turning to the issue of remedies, which should also be
10	pretty interesting. Stay tuned for that. I guess what
11	we typically do is post the schedule on each agency's
12	web page, so look out for those.
13	A couple housekeeping items. First, I guess we
14	ask that people turn off their cell phones,
15	BlackBerries, any other electronic devices that make
16	annoying noises. Second, importantly, restrooms are out
17	across the lobby. In case someone needs to use those,
18	follow the signs, you cannot miss them.
19	Third, a safety tip for everybody, I guess in
20	the event the fire alarms do go off, please do not
21	panic. Please walk towards the exit, and we will guide
22	you to I guess a safe place across the street where we
23	will gather, hopefully with warmer coats.
24	Lastly, I guess the way we set this up is we ask
25	that the audience please not ask questions, and we are

122

1	going to have a lively discussion today, so you can look
2	forward to that.
3	Many of the prior sessions talked about
4	obviously the conduct involved in section 2 challenges,
5	and today, what we would like to talk about is a
6	separate topic, which is, of course, monopoly power and
7	defining markets when monopoly power has been alleged,
8	and I think that is a pretty important topic, one that I
9	think when the hearings kicked off, that Herb Hovenkamp
10	and Dennis Carlton identified as being one of the two
11	that they thought were probably the toughest and two
12	that needed a lot of discussion. So, hopefully we will
13	be able to accomplish that today.
14	I think that is pretty much what I wanted to
15	say, Greg. I don't know, maybe you want to give your
16	title and anything else you want to say.
17	DR. WERDEN: Yes. Hi, I am Greg Werden, Senior
18	Economic Counsel in the Antitrust Division, Department
19	of Justice. I just want to add my thanks to our
20	panelists and the staffs of the two agencies for
21	organizing this session.
22	MR. WALES: Great, thanks.
23	The way the format is going to work today is
24	similar to what we have done previously and did this
25	morning. First, we are going to have each presenter

123

1	give about a 15 to 20-minute oral presentation, and then
2	what we will probably do is take a break either in the
3	middle of that or towards the end of that, we will see
4	how long things go, after which we are going to have a
5	moderated discussion where we will give each panelist an
6	opportunity to respond to the other panelists and also
7	for Greg and I to pose some questions and some
8	principles that we think we might be able to move
9	towards convergence on.
10	With that, I guess what I would like to do is
11	introduce Simon, our first speaker. Simon is a partner
12	and co-founder of RBB Economics. He has been advising
13	clients on competition policy issues since 1991 and has
14	particular expertise in applying empirical techniques in
15	the context of merger investigations. In addition to
16	his private sector work, Simon has been seconded for a
17	short period of time to the German Federal Cartel
18	Office, where he gave a series of seminars on use of
19	economics in competition law. Simon has published
20	widely on virtually all aspects of competition law
21	economics and is a regular speaker at competition law
22	conferences. He is a co-author of The Economics of EC
23	Competition Law and has worked on several hundred
24	competition law matters spanning virtually all sectors
25	of the economy. Thanks, Simon.

124

1	MR. BISHOP: Long intro, and you forgot to say I
2	am from Europe. I am up first, and given that it was
3	probably a lively session this morning, I think probably
4	the reason I have been chosen to go first is because my
5	topic this afternoon is probably the dryest and most
6	technical, which is my remarks are really going to
7	concern the Cellophane fallacy and what implications
8	that has for a structural approach to assessing
9	monopolization or what we Europeans talk about as an
10	abuse of dominant position.
11	I am also going to say that my remarks are also
12	sort of Euro-centric in the sense of this really
13	reflects my experience of EU cases and European national
14	competition law cases and not really the U.S. case law;
15	however, given that we are all in this facade, one might
16	say, or claims that there is increasing convergence
17	between Europe and the U.S., I hope that some of these
18	remarks will actually carry over to U.S. antitrust.
19	Now, in order to give this some sort of context,
20	in Europe, as in the U.S., we are engaged in an ongoing
21	reform of Article 82, which is the equivalent of the
22	monopolization act, and last year, the European
23	Commission, of which Miguel was heavily involved, issued
24	guidelines on how to reform Article 82 and to move the
25	current system away from a form-based approach to a much

125

1	more effects-based approach.
2	Now, that is all very admirable, but it seems to
3	me that within these guidelines of reform, there is the
4	elephant in the room which no one really wants to talk
5	about; namely, the concept of dominance. Really,
6	dominance is also based on structural market share. In
7	Europe, we have the two-step process, whether a firm is
8	found to be dominant and then whether, if that firm is
9	found to be dominant, whether that behavior constitutes
10	an abuse of a dominant position.
11	Now, as I said, all the focus has been on the
12	second step, but the problem is is that within Europe,
13	certainly how the courts have interpreted dominance and,
14	indeed, some of Miguel's colleagues in the Commission
15	also, is that if you are dominant, then any behavior
16	which affects or harms competitors is almost deemed to
17	necessarily harm competition, and if you take that
18	approach, then that really means there is no role for an
19	effects-based analysis within European antitrust under
20	Article 82.
21	It also means that the dominance, and therefore
22	the market share calculations and market definition, are
23	absolutely paramount in the whole assessment. Now, we
24	all know that from the sort of 1982 U.S. Merger
25	Guidelines, there has been pretty wide acceptance that

126

1	the hypothetical monopolist test or the SSNIP test has
2	provided the appropriate framework for assessing and
3	defining relevant markets. We also know that, even
4	though we have the framework and that is well accepted,
5	in actual individual cases, it is actually quite hard
6	sometimes to actually implement that test. Actually, in
7	monopolization cases or abuse of dominance cases, things
8	are even more difficult because of the existence of the
9	Cellophane fallacy.
10	Now, what is the Cellophane fallacy? What are
11	the problems? Well, in a merger context, which is where
12	the SSNIP test or hypothetical monopolist test was first
13	proposed, we are interested in what has the merger
14	changed? Is it going to relax competitive constraints
15	at prevailing price levels? Now, that has an important
16	implication, because that says we can use existing data,
17	observed data, to assess the strength of existing
18	competitive constraints between products supplied by the
19	merging parties.
20	However, when we talk about monopolization
21	cases, in many cases -- and some might argue in all --
22	the relevant issue is not whether the prices can go up
23	even further from prevailing levels, but actually, have
24	prices already been increased above competitive levels?
25	Now, the important implication of that is that using

127

1	observed data will tend to overstate the competitive
2	constraint, because as we know from the famous DuPont
3	Cellophane case, is that even a monopolist, if you put
4	up prices far enough, something is going to start
5	looking like an effective substitute at some point,
6	because even monopolists face some constraint. So, the
7	real issue here I think from the Cellophane fallacy is
8	what implications does it have for the use of empirical
9	analysis?
10	Now, that is a case of sort of, well, what do we
11	do about this? We know that the Cellophane fallacy
12	exists, and we know that that has a big impact on how we
13	can interpret and use existing data. Well, there are a
14	number of approaches which have been put forward to try
15	and address the Cellophane fallacy. One which has been
16	put forward in a number of cases both by the European
17	Commission and some national competition authorities in
18	Europe is to say, "Well, the hypothetical monopolist
19	test is only one way of defining a relevant market."
20	Well, the question or the problem with that sort
21	of line of argument is, they do not actually tell you
22	what the alternative ways of defining a relevant market
23	are, and what is good about the hypothetical monopolist
24	test is it is forcing people to at least think about the
25	scope for demand-side substitution and supply-side

128

1	substitution, and if those two things are not part of
2	the approach of defining a relevant market, it seems to
3	me that, indeed, any other alternative approach is
4	useless for antitrust purposes.
5	The second approach to trying to deal with the
6	Cellophane fallacy is, "Well, let's just recalculate
7	everything from the competitive price." Well, great
8	idea, but if we knew what the competitive price is, then
9	we would not need to be defining what the relevant
10	market is. We could just say, "Well, we observed that
11	Firm X is charging 10, we know the competitive price is
12	5; therefore, there must be some sort of exercise of
13	market power going on." But that is not how the real
14	world works.
15	So, as a slight anecdote here, I was reading in
16	some of the trial transcripts in the Microsoft case, one
17	of the economists I think it was for the DOJ was asked,
18	"Well, what is the competitive price that Microsoft
19	should charge?" The answer was, "Lower than they
20	currently charge," which seems to me sort of just
21	demonstrate the difficulties of actually re-adjusting
22	what the competitive price is. So, that is not going to
23	get us anywhere.
24	The third approach is, "Well, let's just ignore
25	the Cellophane fallacy; pretend it does not happen."

129

1	Well, again, that is not going to work, because if you
2	ignore it and just apply empirical analysis, you are
3	going to tend to define relevant markets too widely, and
4	therefore, not capture some market power when we should
5	be capturing it.
6	The fourth approach which has been proposed,
7	which is, "Well, let's do away with market definition
8	altogether. It is difficult -- we have got the
9	Cellophane fallacy, you know, we are very smart economic
10	professors or consultants, and we can just sort out --
11	you know, market definition is just an interim step. We
12	can go straight to the answer." Well, personally, I am
13	a bit more humble than that, and I think if we see the
14	relevant market definition and the structural analysis
15	for what it is, i.e., an intermediate step, I think it
16	is important that we keep that step.
17	Secondly, if you do away with it, it actually
18	introduces real scope for ad hoc analysis. There is a
19	real -- we know that particularly in the areas of
20	exclusionary abuses or exclusionary power, trying to
21	discriminate between behavior which just merely harms
22	competitors and is therefore procompetitive from that
23	which harms competitors and drives them out of the
24	market and leads to harm to consumers is very, very
25	difficult, and really, the market definition structural

130

1	screen provides a good touchstone to prevent people from
2	adopting "I know abusive behavior when I see it."
3	So, I think that sort of a summary of this is
4	really we are stuck with the hypothetical monopolist
5	test, the SSNIP test, as a framework for thinking about
6	how we define relevant markets, and we are also stuck
7	with the Cellophane fallacy. We need to accept that it
8	exists. So, what are my proposed implications for the
9	sort of policy?
10	Well, a structural analysis still can provide a
11	very useful filter, and even recognizing the existence
12	of the Cellophane fallacy, I think we can go through a
13	number of steps, that we can define relevant markets
14	which are consistent with the basic principles of the
15	hypothetical monopolist test. So, if someone proposes a
16	relevant market and that it does not seem to be
17	consistent with the principles of demand-side
18	substitution or supply-side substitution, then it is not
19	a relevant market. So, I think even just using the
20	SSNIP test as a thought process can actually provide a
21	useful discipline on how to define relevant markets.
22	Secondly, we know the Cellophane fallacy exists,
23	but if the parties are arguing for a wide market
24	definition, then they at least ought to be able to
25	demonstrate that at prevailing prices, there is

131

1	substitution. Now, that means that it does not stop
2	with saying, "Well, the price has already been increased
3	above competitive levels and is subject to the
4	Cellophane fallacy," but at least they should be able to
5	show that at prevailing prices, there is a competitive
6	constraint between product A and product B if they are
7	arguing they are in the same relevant market.
8	The third element I think is we can look at
9	product characteristics in the marketplace, but again, I
10	think we should be careful about how we look at that,
11	and this really goes back to my first point, which is
12	consistency with basic principles, is it is not saying
13	that these two products are not in the same relevant
14	market because they look different or have different
15	product characteristics. We are saying they are not in
16	the same relevant market because the differences in
17	product characteristics imply that demand-side
18	substitution or supply-side substitution is unlikely.
19	The fourth element is that there are some cases
20	and there is some evidence which is not subject to the
21	Cellophane fallacy which we can use to discriminate
22	between competing claims, and as always, there is a
23	paper by Greg Werden, who addresses this, and I think it
24	was from about 2000.
25	The second policy issue is, well, we have

132

1	defined the relevant market, we have calculated some
2	market shares, and clearly we need to put that into
3	context. We need to take into account the scope of all
4	barriers for entry, expansion, the scope of buyer power,
5	whether the market is subject to bidding competition,
6	and also general market dynamics.
7	My final comment was really, well, let's be
8	humble here, because we can go through all of these
9	steps, but in a lot of cases, the available evidence
10	will not allow us to discriminate between the wide
11	market definition which the parties are putting forward
12	and the narrow market definition which the agencies are
13	going to be putting forward. Everything may be
14	consistent with the basic principles of the hypothetical
15	monopolist test, the parties can show that at prevailing
16	prices there is substitution and so on and so on, and
17	where you have got these two competing potential market
18	definitions, sometimes that will not be a problem,
19	because the market shares in both of those may be low,
20	and then unlikely to have market power. Alternatively,
21	market shares in both of those could be high, and then
22	that is not really a problem, because the market power
23	is reasonably high. The difficulty or the problem,
24	potential problem, is where in one market, the narrow
25	market, the firm has a high market share, and in the

133

1	wide market, it has a low.
2	It seems to me when you are in those situations,
3	all it says is, well, then we really need to have some
4	pretty good evidence and examination of the business
5	conduct, and this I think brings me back to where we are
6	in Europe, is that a lot of times in Europe, with the
7	current situation, the business conduct is not assessed
8	on the market effects, but actually on the form of the
9	business conduct. So, the reform in Europe is certainly
10	going in the right direction in focusing on the form,
11	and that is the end of my comments.
12	Thank you.
13	MR. WALES: Thanks very much, Simon.
14	(Applause.)
15	MR. WALES: Our second speaker is Miguel de la
16	Mano. Miguel joined the European Commission in 2001 and
17	is currently a member of the Chief Competition
18	Economist's Team. He carries out economic analysis in
19	mergers and commercial practices by dominant companies
20	and their impact on the competitive structure of the
21	markets. He is also responsible for drafting
22	guidelines, setting the Commission's analytical
23	framework in these areas, a key area. He completed
24	graduate studies in economics at The Institute For the
25	World Economics in Kiel, Germany and The European

134

1	Institute at Saarbrucken University, Germany. He
2	conducted his Ph.D. research at Oxford.
3	With that, Miguel? Thanks.
4	MR. de la MANO: Thank you very much. It is
5	definitely a pleasure and also a great honor to
6	participate on this panel today together with so many
7	distinguished and well-experienced practitioners.
8	I will try to contribute to this issue basically
9	by offering a view or an assessment of the way in which
10	dominance or the role that dominance plays today in
11	competition policy assessment in Europe and which, as
12	you know, is enshrined in Article 82, which is the
13	equivalent of section 2 here in the U.S.
14	As you also know and as Simon has remarked, the
15	Commission is in the process of reviewing its policy in
16	the area of Article 82, and like every type of reform,
17	it is somewhat case-dependent, and we are constrained by
18	case law and case practice; however, we believe that the
19	time is right basically to align the implementation or
20	the enforcement of Article 82 to current thinking and
21	current economic knowledge, and, of course, to a more
22	modern analytical framework.
23	So, I will basically start by making a somewhat
24	obvious remark, yet actually crucial, which is that in
25	the context of the analysis of monopolization in Europe,

135

1	dominance is a necessary condition. That is how the
2	system has been set up, and the EU Treaty actually
3	prohibits single-firm conduct that harms consumers only
4	when undertaken by a dominant company, and normally, to
5	ensure the efficiency of the decision-making process,
6	this also means that the first step of the analysis is
7	to establish whether or not a single firm actually is in
8	a dominant position or not. It is not a must, but that
9	is just the best way forward. If a single firm is not
10	dominant, then there is no need to proceed any further.
11	At the same time, a somewhat more subtle point,
12	this also rules out what in the U.S. is attempted
13	monopolization. If you are not dominant in the first
14	place in the EU, basically there is nothing you can do
15	that will violate Article 82, and I think this is an
16	important point, because it somewhat dispels the myth
17	that in the EU, there is a serious concern or serious
18	worry with type II errors; namely, false acquittals. I
19	think personally that is not the case.
20	But what are the reasons for this institutional
21	setup? I can think basically of two primary reasons.
22	Number one is to provide legal certainty. Surely it is
23	better for firms to know in what circumstances they may
24	be liable to and they are obligated to. There is also
25	another reason, which is that we should not forget, it

136

1	was member states that have delegated the powers to a
2	rather independent body, namely the European Commission,
3	to enforce competition policy in their name, and when
4	delegating such powers, member states want some
5	assurances that these powers will not be abused, and
6	therefore, forcing the Commission to start off by
7	assessing whether or not a firm is dominant imposes some
8	sort of discipline, which understandably was necessary
9	for member states to delegate such powers.
10	However, unfortunately, despite the best wishes
11	of everybody at the time, maybe 30 years ago, it has not
12	fully worked, and I think there are three reasons why it
13	has not fully worked, which I would like to share with
14	you and hopefully also in doing so contribute to the
15	thinking that is taking place here in the U.S. with
16	respect to monopolization.
17	The first reason why they do not work is the
18	concept of dominance is somewhat elusive. The member
19	states put it into Article 82; however, no definition
20	was actually provided. That was left for the courts to
21	develop one over time.
22	However, as is normal, the courts were actually
23	reacting to cases that were brought to them, and they
24	were not necessarily thinking in the abstract, well,
25	what is it that dominance should mean? How should it be

137

1	defined? But instead, we are reacting to the
2	circumstances.
3	Of course, it became increasingly complex and
4	increasingly difficult to understand exactly what
5	dominance is as time went by and European courts were
6	issuing rulings where the concept of dominance was
7	mentioned or in some cases defined.
8	Of course, what happened ultimately is that,
9	before the Commission, it became increasing difficult to
10	identify what is dominance, and therefore, the more
11	difficult it was, the more elements which would normally
12	go into the competitive assessment creeped into the
13	assessment of dominance, up to a point where it seems,
14	at least to me, that as Simon pointed out before,
15	assessing dominance became an end in itself to the
16	extent that once dominance had been established, it was
17	not just a necessary condition but almost sufficient for
18	a finding of abuse.
19	Now, I think that these three concerns can be
20	corrected, and this is, of course, the rationale for the
21	review process, and I would just like to share with you
22	the three ways in which I think this can be done.
23	So, first of all, again, a rather obvious
24	statement, but somewhat important in a context where
25	European courts have said that the dominant firms have a

138

1	special responsibility, whatever that might mean,
2	dominance should be defined or equated with substantial
3	market power. Now, of course, all firms have some
4	market power, but most of them have very little, and
5	accordingly, the relevant question in antitrust cases
6	should not be whether market power is present or not --
7	it always is -- but whether it is important, that is,
8	whether it is substantial.
9	In going back to the sort of most established
10	definition of dominance by the ECJ, dominance is said to
11	be a situation where a company has the power to behave
12	to an appreciable extent independently of its
13	competitors, customers and ultimately its consumers, and
14	a close look at this definition suggests, indeed, that
15	dominance can be equated to significant market power,
16	and this is because a firm is dominant if its decisions
17	are fairly insensitive to reactions of competitors and
18	customers. That is what the "to an appreciable extent"
19	actually means. Of course, no firm is fully independent
20	of customers and competitors, that we know from economic
21	theory, but to an appreciable extent, it might well be.
22	The measure of this sensitivity, of the
23	sensitivity to the actions of others, is given by an
24	elasticity, which is, again, the other side of the way
25	that economists would measure market power in practice.

139

1	So, we end up with a situation where to behave
2	independently to an appreciable extent can be definitely
3	equated to an ability to significantly and profitably
4	durably increase prices, and therefore, there should be
5	no more debate about what is dominance, just substantial
6	market power.
7	Now, how is this substantial market power to be
8	established? Well, again, this is not new to anyone,
9	but I would argue that first market shares have to be
10	significant, and significant in two senses. First, they
11	have to be significant in that they must be important,
12	high, but also significant in that they are actually
13	providing a good proxy for the relative insensitivity of
14	the single firm to the actions and reactions of its
15	competitors and customers. There is, again, a good
16	paper by Greg Werden which talks about assigning market
17	share and how difficult this process actually is.
18	The second point is that barriers to entry and
19	expansion have to be significant, and by this I would
20	like to emphasize that we mean in the absence of the
21	conduct, not barriers to entry in general, but in the
22	absence of the conduct, if the conduct itself actually
23	increases barriers to entry or barriers to expansion.
24	Now, that is part of the anticompetitive effects of such
25	conduct, and therefore, it should not be seen as an

140

1	element that plays a role in establishing dominance.
2	Of course, there are other elements like
3	dynamics of the market, there should be no technological
4	leapfrogging, and buyer power, it cannot be shown that
5	the customers have very little countervailing buyer
6	power.
7	Now, I will try to make here a rather
8	provocative statement, but in my view, the acid test,
9	the way to ensure whether a company is dominant or not,
10	is to ask, well, is it the most efficient in the market?
11	Because if it is, it is likely to have high market
12	shares, it is likely to be very difficult to enter
13	successfully and profitably, and it is also going to be
14	very difficult possibly to leapfrog.
15	However, one might argue, well, isn't this just
16	the old efficiency offense? Well, I do not think this
17	is an offense, because I personally think there is
18	nothing wrong with being dominant. There is no offense
19	in being dominant, and companies should not feel that an
20	assessment of dominance actually implies that this is
21	going to lead to a finding of anticompetitive behavior
22	on their part. Quite the opposite, a finding of
23	dominance should in most cases just mean that they are
24	probably the most efficient company out there.
25	This takes me to the final point, which is that

141

1	dominance is not only a screen. It is not an end in
2	itself. It is just a screen to try and filter out, as
3	Simon was saying, those situations where there might be
4	scope for significant harm to consumers resulting from
5	certain conduct from other situations where this is very
6	unlikely to happen.
7	Now, it is clear that if a practice is shown to
8	be anticompetitive, the firm must be dominant, but
9	proving that a practice is anticompetitive is hard, and
10	it takes a lot of time and resources. Therefore, it
11	seems like assessing dominance can play a very important
12	role in acting as a screen, and it is also a screen that
13	bites. It bites because large firms may not necessarily
14	be dominant if innovation is taking place at a rapid
15	place, if there is fierce competition between large
16	players, for instance, in the concept of bidding
17	markets, or if there is strong disciplining by potential
18	entrants or customers.
19	Now, I am just going to briefly go into a
20	non-hypothetical example, which unfortunately I am not
21	allowed to get into further details of the market, but
22	where actually I will be able to show to you that the
23	Commission takes very seriously the dominance screen and
24	it actually works in practice.
25	We had a case not long ago where the defendant

142

1	had very high market shares in a homogenous good market,
2	above 60 percent. There were very important and
3	significant barriers to entry, like large overcapacity
4	on the part of the dominant company, declining demand,
5	high fixed costs to establish new facilities, but also
6	strong learning effects in the process. It was common
7	practice in the industry to use very long-term
8	contracts, which, of course, we argued would limit
9	customer switching, and not the least of which the
10	defendant seemed to be in a very strong position to fend
11	entry given that it had the broadest product range and
12	the largest financial resources. So, with this criteria
13	on the table, one would very easily conclude that this
14	company is actually dominant.
15	Well, actually, the Commission concluded it was
16	not, and this was on five counts. First, there was
17	significant buyer concentration. The top three
18	customers took 70 percent of the market. There was
19	product homogeneity, which allowed them to switch
20	suppliers without incurring significant switching costs,
21	and buyers, indeed, have dual sourcing strategy to shift
22	volumes between suppliers with little transaction costs.
23	Rival suppliers had significant overcapacity which they
24	could use to expand, and therefore, there were no
25	barriers to expansion. It was also found that the

143

1	competition mechanism was bidding for large and very
2	occasional contracts, just every few years.
3	So, I would just like to conclude with two
4	remarks, one on market shares and one on market
5	definition, linking it to what Simon has said. First,
6	on market shares, it is, often said that there should be
7	a bright line safe harbor, and also that, only firms who
8	are market leaders can ever be dominant. I think the
9	latter makes no economic sense, and this is clear given
10	the application of unilateral effects in the area of
11	merger control, and, of course, at least in the context
12	of European competition policy, the dominance concept
13	plays a role both in mergers and antitrust, so they are
14	interlinked.
15	However, bright line safe harbors do make sense;
16	however, I believe the threshold should be set rather
17	low, and this is for four reasons. First of all, rivals
18	might be constrained. For example, in the electricity
19	industry, this happens very often. You might have
20	strong multi-market presence, like in the airline
21	industry, if you have one company who is number two in a
22	number of routes and the number one company in each one
23	of the routes is a different one, one can argue that
24	this company who was number two everywhere is actually
25	more dominant or has more significant market power given

144

1	this multi-market context than anyone who just has a
2	leadership position in one individual route.
3	Market leaders are more constrained by
4	regulation than nonleaders, and that can be the case in
5	certain industries, such as telecoms, and the leader may
6	be more constrained by close substitutes or by new
7	entry, for example, in the case of pharma. There was a
8	case of AstraZeneca in the EU not long ago where this
9	was clearly an issue.
10	So, what are the policy implications for not
11	arguing that only if you are the market leader, you can
12	be dominant? There are at least two. One is that for
13	consistency, I will just mention unilateral effects in
14	mergers, but also, to leave the door slightly open for
15	attempted monopolization in the EU, in the EU policy.
16	Then just one very short and final remark on
17	market delineation, which I will just start by saying
18	that I agree with everything that Simon has said, but
19	unfortunately, even though I think we ought to be humble
20	and I definitely agree with that, the EU Commission is
21	forced to be arrogant, because in a sense, we are
22	obliged to take decisions. We have to say what we think
23	about the market. We cannot leave definitions open. We
24	have to say whether we think it is narrow or we think it
25	is wide, whether or not we win the case, and this is a

145

1	problem.
2	This is a problem because we cannot just say,
3	well, you know, let's ignore the Cellophane fallacy or
4	let's think about the Cellophane fallacy as something
5	that plays a very significant role and there is nothing
6	we can do about it, so we just be humble. That we
7	cannot afford to do.
8	However, I think we do not have to lose all
9	hope, because when thinking about the role or the
10	assessment that dominance plays, particularly thinking
11	of dominance as a screen, I think that even if we
12	recognize that the hypothetical monopoly test, the SSNIP
13	test, is, indeed, a useful conceptual framework to
14	identify competitors that are constraining a single
15	firm, the assessment of dominance actually goes a step
16	further, and not just ask the question, well, which are
17	the firms that are there constraining the incumbent, but
18	actually asking, well, how much are they constraining
19	the incumbent?
20	So, in trying to figure out how much is the
21	incumbent being constrained or the defendant being
22	constrained, we can also have a good glimpse into which
23	other firms that are part of that particular market, and
24	therefore, market delineation can in some cases -- not
25	always, but in some cases -- be a by-product of the

146

1	dominance assessment, and this obviously simply reflects
2	that market definition is a means to an end, and what
3	the real issue is is market power.
4	Thank you very much for your attention.
5	(Applause.)
6	MR. WALES: Thank you, Miguel.
7	Next up we have Tom Krattenmaker. Dean
8	Krattenmaker is currently Of Counsel in the Washington,
9	D.C. office of Wilson Sonsini Goodrich & Rosati, where
10	he focuses on antitrust, telecommunications and trade
11	regulation issues. Immediately prior to joining Wilson
12	Sonsini, Tom was an attorney in the Federal Trade
13	Commission's Bureau of Competition, Office of Policy and
14	Coordination, where I had the pleasure of working with
15	him for too short a time, but really enjoyed my time
16	working with him. In that role he principally served as
17	legal adviser to the bureau directors and to attorneys
18	investigating and litigating antitrust cases and advised
19	on several Bureau and Commission public reports.
20	Previously he served as senior counsel in the Department
21	of Justice's Antitrust Division and held positions at
22	the Federal Communications Commission, including Chief
23	of Telecommunications Merger Review and Director of
24	Research and Co-Director of the Network Inquiry Special
25	Staff. Tom has spent more than 30 years in legal

147

1	education. He was a Professor at the University of
2	Connecticut, Professor and Associate Dean at Georgetown
3	University and the Dean of William & Mary School of Law.
4	Thanks, Tom.
5	MR. KRATTENMAKER: Hello. I'd like to begin by
6	thanking Dave and Greg for giving me this monopoly
7	platform for 15 or 20 minutes and am particularly
8	appreciative for you surrounding the platform with the
9	entry barriers with your declaration that there be no
10	questions from the audience.
11	I also would love to be able to take refuge in
12	the defense offered by Miguel that he was forced to be
13	arrogant. The problem is that there is at least one
14	member of the audience I see here who was one of my law
15	school classmates, so he knows darn well that I have
16	chosen to be arrogant. So, what I would like to say
17	honestly is that I am going to sound more assured about
18	my views than I am. I have asked that on my tombstone
19	they write something like, "Often wrong but never in
20	doubt," so if you really do not like what I am saying,
21	say, "Oh, Tom's just trying to be provocative again."
22	Dave can tell you that he has said that many times and
23	enabled himself to get home without going home in a funk
24	or thinking that they have to let me go the next day.
25	The other thing I want to say at the beginning

148

1	is that aside from the fact that I am quite honestly
2	flattered to have been invited to join this group, I am
3	more interested in trying to respond to questions than
4	saying anything in particular, so please do send up a
5	flag after 10 or 15 minutes, and I will just stop. I
6	have four points to make, and if we only get three of
7	them out, I am sure I will be able to smuggle the fourth
8	one in somewhere later on.
9	I am speaking largely off a text -- I am not
10	going to read it to you -- of an article that I
11	published with a couple of really outstanding antitrust
12	lawyers and scholars, Bob Lande and Steve Salop in the
13	Georgetown Law Journal in 1987 called Monopoly Power and
14	Market Power in Antitrust Law. It turns out that even
15	though that is 20 years ago, I think it is still right,
16	so if you want to have a look at that, that is where I
17	am coming from.
18	The first point I wanted to make I think is one
19	where we could say I am preaching to the choir, so I
20	will go through it quickly, but it is not a trivial
21	point, and that is, what do we mean by market power? I
22	think my sense is that in this room, we are all
23	co-religionists; that is, we all think that market power
24	is the ability to price profitably for a significant
25	period of time above the competitive level for that

149

1	market.
2	I might just stop to observe that that has
3	hardly been the history, the unbroken history, of
4	antitrust. We have had many other tests of whether
5	something is anticompetitive or not. Justice Douglas
6	once opined that a merger was anticompetitive because it
7	would lead to moving the corporate headquarters of the
8	firm from a small town on the West Coast to big, bad New
9	York City. Justice Black once told us that a merger was
10	anticompetitive because there would be fewer
11	single-store grocery stores in Los Angeles.
12	We seem to have, at least at this point in time,
13	a consensus that we have an economic concept of market
14	power, and it is the ability profitably to price above
15	competitive levels for a significant period of time, and
16	I know that for crystallizing that definition, one of
17	the people we really have to thank for that is Greg.
18	Another question that I think I was asked to
19	address is what is the difference between monopoly power
20	and market power? Now, syntactically, "monopoly" sounds
21	like -- it says, well, how can you have monopoly power
22	unless you have complete control over a relevant market?
23	You must have to have a 100 percent share of a relevant
24	antitrust market that is surrounded by entry barriers.
25	I suppose -- I do not know, I didn't look at a

150

1	dictionary, I should have -- you could say that is it.
2	That is certainly not the case law definition,
3	and I think, again, within the current antitrust
4	community, nobody would doubt that. I think the right
5	answer is that it is the same as market power. There
6	are some cases out there where there is noise in the
7	opinions that suggests that there is some kind of
8	difference between market power in monopoly power, but
9	it does not seem to make any sense. That is, market
10	power and monopoly power and antitrust law are and
11	should be synonymous. They can occur in various
12	degrees, but they are qualitatively the same.
13	Of course, the analogy that came to my mind was
14	basketball. I am supposed to leave here tonight and
15	play in a basketball game. Yes, you can tell by looking
16	at me I am our team's power forward, and monopoly power
17	and market power are the same in the same sense that a
18	shot is the same. It goes in or it does not go in. It
19	goes in the basket or does not go in the basket.
20	Now, some are worth one point, some are worth
21	two points, some are worth three points. You could have
22	lots of market power or little bits of market power, but
23	it is the same thing. It is not like being tall. You
24	could be very tall or not very tall or sort of tall,
25	but -- no, this is like shots in basketball.

151

1	I guess I have waited long enough for some wag
2	to say, "Well, what about goal tending?" The answer to
3	that, "If you figure that out, you have got the whole
4	rationale for the per se rule," but you did not want me
5	to talk about per se rules? Okay, I will go to the next
6	thing.
7	Market power, monopoly power, are really the
8	same thing. They are qualitatively the same thing. We
9	mean the same thing by it. It is helpful to distinguish
10	between I think two types of market power. The DuPont
11	formulation that is quoted a lot is that monopoly power
12	is -- DuPont is the same one that introduced the
13	Cellophane fallacy -- the power to control prices or
14	exclude competition.
15	That sounds like it is two things, doesn't it?
16	Power to control price or the power to exclude
17	competition, arguing it is really the same, but the
18	reason you see that or the reason you sometimes see this
19	noise in the cases about there are these different
20	things is that the intuition the judges have is that it
21	might make a difference what kind of market power you
22	have or how you are exercising it. We put names on them
23	in the paper, but I do not have to use names.
24	One way to exercise market power is by
25	restricting your own output, cutting your own output,

152

1	sometimes in concert with that of other people in the
2	market who are happy to join with you. I would call
3	that collusive market power. We called it Stiglerian in
4	honor of George Stigler because it is the kind of market
5	power he wrote about.
6	The other way that one might exercise market
7	power is not by restricting one's own output but by
8	restricting rivals' output, letting market output
9	decline and letting your price rise through no
10	restriction in your own output. That I would call
11	exclusionary market power or market power obtained or
12	exercised by exclusionary means. In the paper we called
13	it Bainian, after Joe Bain, an economist who had written
14	a lot about entry barriers and exclusionary issues.
15	My second submission to you is that -- while
16	market power and monopoly power are the same kind of
17	concept and that we do have a notion of what it means
18	that we tend to agree on -- that it will help us if we
19	distinguish between whether we are talking about
20	collusion or exclusion, or if you like the little
21	labels, Stiglerian or Bainian market power. Now, why is
22	that the case?
23	The article is about market power in antitrust
24	law. We are here talking about section 2. So, let me
25	try to explain with respect to section 2 cases, monopoly

153

1	or attempted monopoly cases, why it might make a
2	difference to think about the source of the market power
3	or the type of market power that we are talking about.
4	Point one, market and monopoly power include the
5	power to keep prices from falling to competitive levels.
6	I do not think we forget this a lot. We usually just
7	say it is the ability to raise prices, but when
8	confronted with the ability to keep prices from falling,
9	we usually recognize that as market power, but you
10	should in case you did not.
11	If you had a horse and buggy industry that was
12	perfectly competitive, a hundred firms each producing 1
13	percent of all horse and buggy output, if they managed
14	to exclude one firm and that firm was the first firm
15	that was going to produce the automobile, they have
16	nevertheless exercised market power even though it was a
17	completely competitively organized industry. It is
18	market power. It is market power to be able to keep
19	prices from falling to competitive levels. Fencing out
20	rivals who have the ability to bring in a new technology
21	or simply be able to produce products at a much lower
22	cost is an exercise of market power.
23	Secondly, and connected to that, I believe it is
24	not correct to insist on a threshold showing of market
25	power if the conduct complained of is acquisition and

154

1	exercise of market power by excluding rivals. If you
2	are talking about a section 2 monopoly case, and you are
3	saying what they are going to do is restrict their own
4	output and profitably price for a long time above
5	competitive levels, it is probably correct for the
6	reasons that Simon and Miguel have already talked
7	about -- although it was not the principal purpose of
8	their talk, but they explained it -- to insist on some
9	kind of threshold of market power. It is kind of hard
10	to imagine how a firm with only 40 percent of the market
11	can restrict its own output profitably for a long period
12	of time and thereby price above competitive levels all
13	by itself.
14	That is not true if you are talking about
15	exclusionary behavior. Exclusionary behavior can create
16	the market power. You do not necessarily need to
17	already control a market in order to be able to engage
18	in exclusionary behavior that winds up creating
19	effective market power. You might still have a
20	threshold.
21	If you do the math, he said -- referring to
22	other people because he is not a mathematician -- but if
23	I understand the literature right, the raising rivals'
24	costs literature, you may want to have kind of a
25	threshold that does have to do with size, like relative

155

1	disparity in size. It is unlikely that a firm that has
2	got 5 percent of the market is going to be able to,
3	through exclusionary tactics, drive out rivals who are
4	two and three times as big if it were the smallest firm
5	in the market, but to say that one needs to have a kind
6	of a dominant firm presence before one could ever be
7	tagged with the offense of monopolization under section
8	2 is just not right unless you are -- because you appear
9	to be forgetting what I've called Bainian or
10	exclusionary market power.
11	A third lesson from this that is relevant to
12	section 2 cases, I think, is that it seems to me that we
13	frequently hear it said that the mere exercise of market
14	power is not prohibited by antitrust, and I think there
15	is a statement to that effect in the Trinko decision by
16	the Supreme Court a year and a half ago. Indeed, if I
17	recall correctly, Justice Scalia not only said it, but
18	he said you sort of welcome that kind of behavior
19	because it is a signal to people to come enter the
20	market. There are high prices. You can come in and do
21	something. There is nothing wrong with exercising
22	market power if you have got it. The question is how
23	you got it.
24	Well, once again, I think that probably is true
25	for collusive or Stiglerian market power. It is

156

1	probably correct that a firm that has got 90 percent of
2	the market, if they acquired it lawfully, to say that
3	when they raised -- when they restrict output and raise
4	price, that is an antitrust offense, that is a very
5	tough nut to crack, a very hard argument to make,
6	because what are you going to do about it? What is your
7	remedy? How are you going to decide whether they raised
8	price too high?
9	But if you have in mind the possibility that you
10	might be talking about a section 2 case based on
11	exclusionary market power, it is just not right, because
12	you would attack the exclusionary act, and sometimes you
13	can distinguish between the exclusionary act and other
14	types of behavior with respect to the market power. The
15	most obvious example would be, I think, if I could build
16	off Miguel's example.
17	He gave that terrific example of the industry
18	where, when you first looked at it, you might think
19	dominance, and then you find all these other aspects
20	here. If this had been an industry in which the issue
21	had been an exclusive dealing arrangement that was
22	having the effect of denying vital inputs to rivals, not
23	only does it not require, in order for that to be a
24	successful antitrust strategy, that the firm have a
25	dominant share to begin with, but it is also not the

157

1	case that if it has got a position of dominance, if it
2	is a monopoly, that then the mere exercise of monopoly
3	power is permissible. It is not the case at all, and,
4	indeed, that is an area where I think the European law
5	is ahead of ours, because it clearly reflects that is
6	the abuse of dominance.
7	Finally, I had one more. It is relevant to
8	antitrust law, but it is not relevant to the Federal
9	Trade Commission or the Department of Justice. One of
10	the lovely things about working for the -- there are
11	many nice things about working for the FTC and the
12	Department of Justice that I think, you know, the most
13	are that you always think you are on the right side and
14	you have these wonderful people to work with, but
15	another thing is you never have to worry about standing,
16	because if you see something wrong out there, you can go
17	after it.
18	Out in the private sector, you have got to have
19	standing, and I think another lesson that you learn from
20	distinguishing between these types of market power or
21	these types of means of acquiring or exercising market
22	power is relevant to competitor standing. Competitor
23	standing should not be an issue in most section 2 cases
24	involving Bainian or exclusionary market power, because
25	the action is actually targeted at the competitor.

158

1	On the other hand and for the same reason,
2	consumer standing, even though the person who may suffer
3	the effects is the consumer, consumer standing may be
4	quite risky, both because there is a more direct subject
5	of the harm, that is, the competitor, and therefore,
6	there is the risk of double damages, and so following
7	things like Associated General contractors and Illinois
8	Brick, consumer standing in monopoly cases may be
9	difficult, and consumer standing in attempted monopoly
10	cases I don't think the Supreme Court has ever addressed
11	it, but there is a growing body of case law in the lower
12	courts now that consumers just do not have standing to
13	bring attempted monopoly cases.
14	Most section 2 cases are these Bainian
15	exclusionary power type, and you can see the reason for
16	that is that the harm is not directed at the consumer,
17	and if it is merely an attempted monopoly, there is no
18	follow-through on the part of the consumer.
19	Well, enough for that commercial. Again, I have
20	tried to suggest really just two things to you. One is
21	that we have a concept of market power that we are at
22	least presently comfortable with, and that is no
23	different from the notion of monopoly power for the same
24	reason that we are comfortable with the conception of
25	market power. We are talking about what is the goal of

159

1	antitrust, what are we trying to target our antitrust
2	rules to do, and it is to prevent undue concentrations
3	of power where power means the ability to profitably
4	price above competitive levels for a significant period
5	of time.
6	Secondly, that it will help to keep your eye on
7	the ball, to dig a little bit deeper and say, are we
8	talking about market power that is going to be
9	manifested by restricting one's own output, either by
10	one's self or in concert with one's competitors, or are
11	we talking about market power that is going to be
12	manifested or acquired by driving one's rivals out of
13	the market and thereby gaining the power to exercise
14	higher prices without necessarily restricting one's own
15	output? I think it has a number of potential lessons
16	for section 2, and maybe we will explore some more about
17	that as we talk through the questions.
18	MR. WALES: Thanks, Tom.
19	(Applause.)
20	MR. WALES: Okay, next up we have Irwin Stelzer.
21	Irwin is a Senior Fellow and Director of Hudson
22	Institute's Economic Policy Studies Group. Prior to
23	joining the Hudson Institute, Dr. Stelzer was Resident
24	Scholar and Director of Regulatory Policy Studies at the
25	American Enterprise Institute. He also is a U.S.

160

1	economic and political columnist for The Sunday Times
2	and The Courier Mail, a contributing editor of The
3	Weekly Standard, and a member of the board of the
4	Regulatory Policy Institute at Oxford, a member of the
5	Advisory Board of the American Antitrust Institute, and
6	adviser to the U.S. Trade Representative.
7	Dr. Stelzer founded National Economic Research
8	Associates, NERA, and served as its president for many
9	years. He also served as a Managing Director of the
10	investment banking firm Rothschild, Inc., and Director
11	of the Energy and Environmental Policy Center at
12	Harvard. His academic career includes teaching
13	appointments at Cornell, the University of Connecticut
14	and NYU. He has been elected a visiting fellow at
15	Nuffield College, Oxford, and he is a former member of
16	the Faculty of Practicing Law.
17	DR. STELZER: Thank you very much. Can you hear
18	me in the back? Thank you for inviting me to this,
19	although I fear I may be sailing under false pretenses.
20	Let me clear up one of them. Although I am at the
21	Hudson Institute, I do not want to appear here as
22	somebody who is a disinterested scholar. I do have
23	clients, some of whom are accused of being dominant,
24	others of whom think dominant firms pick on them, but my
25	views go back before most of you were born. I, too,

161

1	still play basketball, but I have learned a trick, which
2	is I yell "Get that rebound" to other people.
3	I am going to leave any comment on specific
4	cases to my co-panelists, because they are more familiar
5	with them than I. I will say, if I am permitted one
6	vignette, I gave up trying to be involved in specific
7	cases when I was sitting on a witness stand in Tucson,
8	Arizona, and the judge summoned counsel to the bench and
9	said, "We have to talk about schedule." The first
10	lawyer said, "Well, you know, my daughter's getting
11	married in May, and that's going to tie me up." The
12	other guy said, "Well, you know, in June, I really was
13	planning a fishing trip." The judge said, "Well,
14	September, I cannot really do," and so they put
15	everything off about a year. In the middle of this, I
16	said, "Can I tell you something about my schedule, Your
17	Honor?" He said, "Don't be ridiculous." I suddenly
18	realized three lunatics were deciding how I was going to
19	live my life for the next year, and I am not doing this
20	anymore. So, I speak to you as a person who used to
21	testify in these cases.
22	I have submitted a much longer, unconscionably
23	long paper, which I assume is available to those who
24	want it, and I will therefore restrict my comments to a
25	very few, and also, I want to try out ideas. I am not

162

1	wedded to what I am about to say. I assumed we were
2	here to try out ideas, not to hand down edicts, and I
3	thought that is why I would try concentrating on pricing
4	practices by dominant firms.
5	Simon Bishop said if you are dominant the
6	practice is questionable; my feeling is if the practices
7	are questionable, you are probably dominant. Simon says
8	he is a bit more humble than doing away with market
9	definition. Those of you who have ever tried to do any
10	market definition know that only the non-humble would
11	attempt the elasticity measurements and the other things
12	involved in it. So, the notion that we must begin with
13	market definition because that is somehow a constraint,
14	and anybody who has read any decisions of the EU knows
15	that it is a very, at best -- you defined it as a loose
16	constraint. I think it is looser even than that.
17	I am not certain that going through the agony of
18	market definition gives you a degree of precision, some
19	sort of constraint on the examiner. It may, but
20	given -- if you go through it, I am not so sure that
21	beauty is in the sight of the beholder as with any other
22	part of economic analysis. I am not wedded to market
23	definition, and I would like to explore the possibility
24	that we might want to do away with that exercise
25	altogether in deciding about dominance.

163

1	I recognize that that would unemploy half of the
2	economics profession, leaving only that part that knows
3	about exclusionary practices still existing, but I do
4	think we should think -- think -- about the possibility
5	that defining relevant markets, defining product
6	characteristics, all of that is a kind of very elastic
7	process that we could do away with.
8	Let me suggest instead -- and I really mean
9	suggest. There is this kind of academic politeness
10	about "let me suggest," meaning "I really know that." I
11	do not use the language that way. I really mean to
12	suggest that we consider that it is the practices that
13	reveal dominance and not dominance that reveals the
14	practices.
15	I have read some of the proceedings, and it
16	seems to me there is a great deal of sort of motherhood
17	and apple pie stuff in this record. It is certainly
18	true, we do not want to prevent vigorous competition
19	that results in lower prices to consumers. Who would
20	want to prevent vigorous competition? Certainly
21	Microsoft did not want to prevent vigorous competition,
22	it says. Yes, we want firms to develop pricing plans
23	that benefit consumers; yes, we want to give businessmen
24	as much certainty as possibility; and yes, we want to
25	reduce the role of lawyers in the board room and leave

164

1	it to businessmen. But I do not think that means that
2	pricing practices should be unscrutinied by antitrust
3	enforcement authorities regardless of any finding of
4	dominance.
5	What we do not want to condone is long-term harm
6	to the competitive process, therefore to consumers, by
7	approving short-run price reductions aimed at creating
8	barriers to entry or preserving market positions that
9	are unrelated to efficiency. Now, again, I would not
10	try to measure efficiency of a firm, because I do not
11	think I know how to do that. There may be people who
12	know how to do that, and when people say to you they are
13	going to measure costs, they are going to compare costs,
14	I would urge any one of you who agrees that that is a
15	terrific idea to determine any cost of any large firm,
16	and you tell me what range you think would make you feel
17	comfortable in that determination, especially since you
18	are usually dealing with someone who does not want you
19	to find out, and so I think you are going to have a very
20	difficult problem.
21	What you have to do is examine a firm's pricing
22	practices in the context of the firm's total behavior.
23	You cannot look at a thread in a tapestry in order to
24	get a picture of whether or not a firm is engaging in
25	exclusionary practices.

165

1	I will give you an example. If you had in the
2	record that a firm had offered a million dollars to a
3	customer not to deal with a competitor, you would say,
4	"Well, gee, we can't tolerate that." But it is very
5	easy to manipulate a pricing schedule in a large
6	multi-product firm to accomplish the exact same thing,
7	to reduce the cost of the incremental order to pretty
8	close to nil by simply manipulating the pricing
9	schedules and the relationship of past to future
10	deliveries.
11	In other words, it seems to me, again, that
12	firms spend millions, hundreds of millions, on discovery
13	in antitrust cases, and the discovery is really
14	discovery that will tell you whether the firm is
15	dominant, whether the firm is engaging in exclusionary
16	practices, with far greater certainty than would any
17	measure of its market share.
18	I think, also, you can tell -- I hate to use
19	this word because I think it is old-fashioned -- you can
20	divine intent from looking at what discovery turns up.
21	Now, by that I do not mean that the statement by an
22	enthusiastic salesman who says "I just rubbed out the
23	competition in Florida" or something like that should be
24	taken at face value, but I think you can determine the
25	intent of a variety of competitive weapons wielded by a

166

1	firm by examining the entire record of its behavior,
2	which brings me to the last question -- I said I would
3	not take my full time -- and that is, has what I just
4	said reduced certainty?
5	A lot of my clients talk about certainty, they
6	want certainty, so you say, "Well, you want certainty?
7	There are two kinds of certainty you can have.
8	Everything you do is subject to a per se rule. That is
9	certainty. How about that?"
10	"No, that is not what I particularly had in mind
11	by 'certainty.'"
12	"Well, the other form of certainty is to say,
13	'Well, almost everything you do is okay.'"
14	"Well, I think that is lousy public policy."
15	Certainty is simply not available in this
16	business. That is it. It is good for the lawyers. It
17	is bad for the businessmen. In making their decisions,
18	they have to listen for counsel and decide what to do
19	about the legal advise they get. It is simply one
20	aspect of the many risks they take, just like guessing
21	at interest rates. Certainty is not there. It cannot
22	be had unless some of the more distinguished members of
23	this panel can give it. I cannot.
24	Thank you very much.
25	(Applause.)

167

1	MR. WALES: Last, but not least, we have Joe
2	Sims. Joe is a senior antitrust partner at Jones Day
3	here in D.C. His practice is concentrated on antitrust
4	and related areas of governmental regulation and
5	includes litigation counseling, agency practice before
6	state and federal courts, antitrust enforcement agencies
7	and various specialized agencies where competition
8	policy or antitrust issues arise. Joe is a member of
9	the American Bar Association, Antitrust Law Section, and
10	has served as chair of numerous committees on the
11	Antitrust Law Section. He's a Fellow of the American
12	Bar Foundation and a member of the American Law
13	Institute. He regularly writes and lectures on
14	antitrust and related subjects and is listed in The Best
15	Lawyers in America, The World's Leading Lawyers, and
16	Who's Who Legal.
17	Joe, thanks.
18	MR. SIMS: Thank you, Dave.
19	Let me start with a point about my perspective,
20	which will also be true for at least Irwin and Tom. I
21	had the revelation when preparing for this and looking
22	back at some of the older cases that I have been
23	practicing antitrust law for about a third of the time
24	that we have had antitrust laws, which is kind of a
25	scary thought if you think about it, but it is true

168

1	nonetheless. A little depressing, too.
2	During that time, no one has ever confused me,
3	unlike most of these people on the panel, as a scholar.
4	I do not cite footnotes in cases. Sometimes I cannot
5	even remember what a case holding was. I do not write
6	law review articles. I write commentaries, not as
7	eloquent as Irwin's commentaries, but it is a less
8	taxing discipline than law review articles. So, I view
9	my role here as offering the practice perspective. I
10	know Tom is a practicing lawyer, but his scholarship is
11	so impressive that I have always viewed him as an
12	academic at heart. So, I am going to approach what I
13	have to say in that light, focusing not on the theory,
14	but on the practice.
15	Fortunately, jurisprudence and for that matter
16	economics and antitrust is very heavily fact-weighted.
17	The jurisprudence and the economics almost always take a
18	back seat to the facts, at least in the long run.
19	Antitrust law in the United States, where it is really
20	law enforcement and not regulation, is mostly about the
21	facts and how the facts are presented. This is true
22	whether you are talking about agencies or judges. It is
23	certainly true when you are talking about juries.
24	Of course, the case law is important. Bad case
25	law is not desirable. It is a good idea, if we can,

169

1	which we do now and have from time to time, have
2	competent, intelligent people running the antitrust
3	agencies, but all of that fades in importance to the
4	unique facts at play in any particular case.
5	During at least my practicing lifetime, we have
6	moved steadily away from what we used to spend a lot of
7	time at, which was antitrust by sloganeering, to more
8	careful analysis of the facts. If you remember, Derek
9	Bok called for more certainty and bright line rules in
10	section 7 cases more than 30 years ago. Well, that
11	actually had some resonance for a while, but that
12	concept was seriously injured by Bill Baxter's Merger
13	Guidelines and probably finally killed by the 1992
14	edition of the Guidelines. When the analysis focuses on
15	competitive effects and not on market shares or
16	concentration or other slogans, the notion of broadly
17	applicable bright lines disappears.
18	So, today, in merger cases, we do not really
19	have any clear rules. All the facts are in play. Every
20	case is unique, and while the outcome needs to comport
21	generally with stated case law and regulatory guidance,
22	the operative word is "generally."
23	This is equally true in section 2 matters. We
24	have come a long way from American Tobacco or Alcoa or
25	even Grinnell, which I was shocked to see was decided

170

1	just four years before I graduated from law school. It
2	seems like a very old case, and with some obvious
3	exceptions, like, Aspen Ski and maybe Kodak, the general
4	direction of Supreme Court decisions over my lifetime
5	has been to gradually cabin in the reach of section 2,
6	in significant part by insisting upon a focus on the
7	facts as opposed to reliance on the mostly populist
8	rhetoric about market dominance and relative size that
9	dominated section 2 jurisprudence in earlier times.
10	A good deal of this, of course, reflects the
11	fact that our markets have matured -- that many more
12	markets today, maybe most markets, are truly
13	contestable, which was not always the case -- but
14	nevertheless, we do not have very many clear rules in
15	section 2 today.
16	I think this is generally a good thing, but it
17	does, as Irwin pointed out, inevitably carry with it
18	uncertainty of outcomes in particular cases. I noticed
19	in looking back at some of the earlier hearings that the
20	Microsoft representative, perhaps understandably, took
21	the position of wishing that there was more clarity in
22	the law. It is a common business position. I think it
23	is a short-sighted business position.
24	To pick up on Irwin's point, if we really did
25	have more clarity, we would have more restrictive rules.

171

1	I do not have any doubt that if you have to choose
2	between clear restrictive rules and clear unrestrictive
3	rules, it is where that line would be drawn. I do not
4	think that would be useful for the public interest in
5	the long term, and it would not even be useful for
6	business at least in the medium to long term. It would
7	make the advisory job easier, but that is about it.
8	So, with this context, these kinds of hearings
9	are really a great idea, especially if they try, as I
10	think they have, to take the long view of an important
11	area of law. More discussion will produce more
12	understanding and will also demonstrate, as these
13	hearings pretty clearly have, that there is an enormous
14	variety of views on section 2 jurisprudence and policy.
15	Indeed, I would argue that this might be more true today
16	than it has been in my practicing lifetime.
17	We still have, of course, the strong populist
18	supporters of very aggressive section 2 enforcement. We
19	still have plenty of conservative "let the market work"
20	advocates. But we also today have an incredible variety
21	of economists and law professors and others who
22	articulate an amazing range of interesting approaches to
23	the identification and analysis of market power. Tom
24	Krattenmaker and Steve Salop obviously are responsible
25	for maybe the single most visible effort in this field,

172

1	but there are a lot of people keeping them company with
2	new and interesting ideas, including, of course, Greg
3	Werden and others on this panel.
4	So, there is no end to possible options for new
5	section 2 approaches, but there is also clearly no
6	consensus on any particular approach, with the possible
7	exception that we really ought to pay attention to the
8	facts. It is very hard for me to imagine how we can
9	productively create clear rules or safe harbors for
10	section 2 using market shares or, for that matter,
11	anything else. Given this lack of consensus on where we
12	ought to draw the lines and the truism, that, at least
13	over the long run, markets are a lot better at
14	identifying and responding to consumer demand than
15	courts or regulators or most academics, the chances of
16	finding consensus bright lines that really do advance
17	the public interest are pretty low. But it is
18	nonetheless worth talking about, and so these hearings
19	are a good idea.
20	Any legal discipline like antitrust where the
21	operative legal standard is in one form or another the
22	rule of reason is going to be messy and unpredictable.
23	Facts are highly variable, and their perception and
24	analysis by humans is even more so. There is the
25	additional problem that courts and regulators, even very

173

1	thoughtful ones that take the time to think about and
2	listen to various points of view, are inevitably better
3	at evaluating the past than they are at predicting the
4	future. They are too often focused on fixing
5	yesterday's problems without really having a very clear
6	picture of how that is going to affect tomorrow.
7	Because of this, we ought to try to be cautious
8	about interfering with markets, doing so only when we
9	are pretty darn confident that the intervention will
10	make things better. I have written on this for 25
11	years, describing (in very gross and simplistic terms,
12	of course) the two basic approaches in antitrust as "do
13	no harm" and "can we help". The "can we help" school
14	tends to be a lot more confident about their and a
15	court's ability to improve market performance than I am,
16	but the "do no harm" school has been in clear ascendency
17	in the past several years, both at the federal agencies
18	and at the Supreme Court.
19	This certainly does not mean that it would not
20	be great if these hearings could find a way to produce
21	some clear consensus and let us feel comfortable in
22	drawing some more bright lines like we have in the per
23	se rule against price fixing, or in the section 2
24	analog, the below-cost requirement for finding predatory
25	pricing. But my reading of the results so far -- and I

174

1	have read at least summaries of all of the hearings --
2	does not leave me with the impression that we have yet
3	identified that consensus.
4	As I said, I am not sure this is a bad thing.
5	One of the most important -- maybe the most important --
6	reasons the antitrust laws have continued to serve us so
7	well after more than a century is that they are pretty
8	darn flexible. Congress, of course, passes a lot of
9	statutes where, in effect, buck the problem to the
10	courts or a regulatory agency, but it rarely works as
11	well as it has in this field.
12	I think that is because, in general and over the
13	long term, the rule of reason is a pretty accurate
14	description of what courts really do -- and regulators
15	too, for that matter. They generally try to figure out
16	what is reasonable under the circumstances with a strong
17	bias most of the time -- let's put the Robinson-Putman
18	Act to the side as an outlier -- toward leaving markets
19	free to work their magic.
20	As long as this is the operative legal regime
21	under section 2, we will have uncertainty about
22	particular cases and there will be uncertainty about how
23	a particular fact pattern is analyzed. This approach
24	has costs, of course, including, most importantly, the
25	inadvertent deterrence of procompetitive behavior, but I

175

1	suspect the costs are less than would be the case with
2	either bright line rules that miss the mark or
3	impractical tests that over-deter because of ambiguity.
4	So, I do not think we really need a whole bunch
5	of new rules; nevertheless, if we could come up with
6	them, we should, and so I am glad we are looking at it.
7	We have to remember, however, that there is a difference
8	between section 1 and section 2 and a very good reason
9	for the difference. Section 1 deals with joint conduct,
10	and while there are many times when joint conduct can be
11	neutral or procompetitive, there are obvious and very
12	real circumstances where there are competitive risks
13	from joint conduct, cartel behavior being the most
14	obvious. Given this, it is tolerable to have some
15	potentially overreaching penumbras of illegality,
16	although as we get more cases like Daugher, even this is
17	gradually reduced.
18	But Section 2, by contrast, is aimed at
19	unilateral conduct, and over-enforcement here would
20	threaten the very essence of competition. We want firms
21	to be monopolists or to try to be monopolists. The less
22	risky we make that effort, the less aggressively firms
23	will try. So, section 2 cases should be hard to bring;
24	they should be harder to win. Successful cases should
25	be rare, because true monopolists with durable monopoly

176

1	power are rare as determined by how hard it is to name
2	some. It is kind of hard to do, actually.
3	That's why Microsoft was such an attractive
4	case. It was one of the few instances where you could
5	look at it and say, "Doggone it, it looks like they do
6	have a monopoly." If we can devise some rules or
7	guidelines to help us advance this cause, that is great.
8	My guess is we cannot, so we ought to let the market --
9	in this case, the market for judicial decisions over the
10	long run -- create and enforce the rules, and the result
11	will be just fine.
12	Thanks.
13	(Applause.)
14	MR. WALES: Thanks, Joe.
15	Okay, as we said, we are going to take a
16	15-minute break. So, why don't we reconvene at 3:35.
17	Thanks.
18	(A brief recess was taken.)
19	DR. WERDEN: Okay, let's get started again.
20	What we are going to do for the next little
21	while is start by putting one or two questions to each
22	of the speakers, in turn, and then letting the other
23	panelists, if they like, comment on what has been said,
24	and we are going to take the panelists in the order that
25	they spoke, so I am going to start with Simon, and my

177

1	question, Simon, is, while there is clearly a dominance
2	threshold under Article 82, there really is an open
3	question as to how high the bar is for dominance, and I
4	think the way Miguel described it, the bar is and ought
5	to be quite low. What do you think about that?
6	MR. BISHOP: Okay, well, contrary to what Irwin
7	might have suggested, most of my clients are actually
8	dominant firms, so on that basis, I think, you know, the
9	40 percent threshold, which is enshrined in Article 82,
10	is a pretty reasonable threshold to have. I mean, if
11	your market share is below 40 percent, then you can do
12	whatever you like. If you are above that, then we move
13	into the effects and the assessment of the behavior
14	under consideration. It does not mean if you are above
15	40 percent, what you are doing is necessarily
16	anticompetitive.
17	DR. WERDEN: But you wouldn't say that all the
18	firms above 40 percent are dominant, of course, would
19	you?
20	MR. BISHOP: Absolutely not, and that is why I
21	said in my talk, you know, the market share is only one
22	factor. You have got to take into account a lot of
23	other factors to assess whether that 60 percent, say, is
24	representative of significant market power.
25	DR. WERDEN: Do any of the other panelists wish

178

1	to offer a view as to how high the bar should be set in
2	the United States where I think most observers think it
3	is set considerably higher than in Europe?
4	MR. KRATTENMAKER: Or whether there should be a
5	bar at all, I guess.
6	MR. SIMS: But, Tom, wouldn't you say that there
7	shouldn't be a bar, I would think?
8	MR. KRATTENMAKER: Yes.
9	MR. WALES: So, the answer is there is no bar.
10	MR. KRATTENMAKER: Or what I would say is, bar
11	to what?
12	DR. WERDEN: Bar to proceeding.
13	MR. KRATTENMAKER: You mean, like, a
14	post-behavior section 2 case where the claim is what I
15	called collusive or Stiglerian power? Sure.
16	DR. WERDEN: Well, if you want to go down that
17	road, in an actual monopolization case, where the
18	defendant is alleged to have acquired a monopoly, the
19	courts have set the bar fairly high on what it means to
20	have a monopoly and generally have required, in fact, a
21	70 percent share protected by pretty high barriers to
22	entry.
23	MR. KRATTENMAKER: Yes, right, right. I think
24	if they acquired that monopoly by, for example,
25	acquiring a lot of rivals by purchasing firms, that

179

1	would probably be an appropriate threshold to do. Now,
2	you do not see cases like that because we have had
3	section 7, so almost all section 2 cases now are what I
4	would call exclusionary or Bainian type, and yeah, that
5	is right.
6	I think it is not correct to say you could not
7	possibly have market power if you have got 66 percent of
8	the market.
9	DR. WERDEN: So, in the Microsoft case, if their
10	share had been 10 percent, you would have looked on
11	things pretty much the same way?
12	MR. KRATTENMAKER: You know, there were so many
13	facts at issue in the Microsoft case...
14	No, as I tried to indicate, it does not seem to
15	me that you utterly disregard market share, Greg, but as
16	I understand it -- and I am still learning this area --
17	the ability to exclude can oftentimes be a factor of
18	relative size, but the idea that it requires dominance
19	of the entire market I think is quite wrong.
20	DR. STELZER: Given what Microsoft did and
21	proved itself capable of doing, did you have to bother
22	measuring its market share? I mean, nobody who didn't
23	have huge market dominance, i.e., 90, 80, 40, could do
24	those things, could make an equipment manufacturer pay
25	them for stuff that was not in the machine. I mean, you

180

1	have got to have an awful lot of market power to do
2	that. You want to measure market power because lawyers
3	make you do it, but as a matter of policy, in the case
4	of any firm that can pull off what Microsoft pulled off,
5	you could skip the whole market share measurement stuff
6	and just say, if they did this, they have market power,
7	they have abused it.
8	MR. KRATTENMAKER: I probably ought to let Joe
9	pick up on that, but I will say -- I mean, I know a
10	little bit about Microsoft. I mean, you might be able
11	to say that, but if what you are doing is talking about
12	the part of the case where they allegedly misrepresented
13	whether their programs -- either how it interfaced with
14	Java, I do not know that you needed to have a dominant
15	market share in order to lie.
16	DR. STELZER: No, no, I was talking about where,
17	if you decided to put a competitor's product in the
18	machine, they charged you for each machine whether you
19	put their stuff in it or not.
20	MR. KRATTENMAKER: No, I've gotcha. I take
21	it -- I mean, I am sympathetic to your viewpoint, but it
22	is conduct-specific. For certain kinds of conduct, you
23	might infer market power from the fact of the behavior.
24	DR. STELZER: What they do, I shall know them.
25	MR. SIMS: On this point, I am more with Tom

181

1	than Irwin, I think, surprisingly enough. Market
2	definition and whatever you draw from that market
3	definition is a tool that you want to use when it is
4	necessary and useful to figure out what the competitive
5	effects of the conduct at issue are. So, there are some
6	where, careful market definition is not all that
7	important.
8	MR. BISHOP: But I think, I mean, some of the
9	difference between the U.S. people at that end of the
10	table and the Europeans down here is really -- sort of
11	reflects some of the sort of philosophical,
12	institutional differences, and I'll say institutional
13	because I think my personal philosophy is going to be
14	closer to that end of the table than a lot of Europeans,
15	and I think that that is a point which Joe talked about,
16	you know, is do no harm, which is, you know, very much a
17	high threshold before you would start intervening, then
18	sure, maybe you don't need a market share bright line
19	test, but in Europe, the institutional philosophy is
20	much more -- you know, there are a lot of markets, the
21	EU, the Commission or the competition authorities can
22	intervene in to make things better, and in that
23	situation, in that sort of institutional setup, then
24	having a bright line test which says, "If you do not
25

182

1	have a market share of above 40 percent or whatever, you
2	can do whatever you like," seems to me an important
3	safeguard to prevent people coming in and start messing
4	around with your industry, which is very costly and
5	potentially extremely disruptive to the firm's business
6	model if that firm has got no market power at all.
7	DR. STELZER: But that is kind of the "stop me
8	before I kill again" argument, right? You need --
9	because you know that you really could be irresponsible
10	and do bad things, you better have some sort of rule
11	that stops you from doing it on the theory that the
12	rule, is the lesser of the evils. It is a substitute
13	for judgment.
14	MR. BISHOP: No, it's not. It is a substitute
15	for deciding when a competition authority can bring an
16	action against a business.
17	DR. WERDEN: Or in the United States, substitute
18	for a jury trial.
19	MR. SIMS: Well, there is that pretty critical
20	difference between the U.S. and Europe in that in
21	Europe, the Commission generally gets to say yea or nay,
22	and in the United States, the FTC and the DOJ never get
23	to say yea or nay. Unlike the EU, they have to go to a
24	court and convince a court.
25	I think what Simon is postulating is that some

183

1	kind of -- if I could borrow the word -- durable
2	guidelines that, would last beyond a particular
3	administration of the Commission and thus constrain the
4	current occupant of those decision-making positions is a
5	good substitute, partial though it may be, for what we
6	have here in the courts.
7	DR. WERDEN: Okay, that was fun. Let's move on
8	to a question for Miguel.
9	I was very intrigued by your very clear point
10	that the suspect conduct in an Article 82 case cannot
11	itself be what creates the barrier to entry that is
12	required, in turn, for the firm to be dominant, so that
13	if it was possible to have a firm with a whopping share
14	protected only by the suspect conduct in the case,
15	otherwise you would be flooded with competition, then
16	that firm isn't dominant? Is that your submission?
17	MR. de la MANO: Indeed, and there is the
18	problem that we have in the EU, that we do not really
19	have a standard which allows us to pursue attempted
20	monopolization.
21	DR. WERDEN: No, let the firm be 80 percent. It
22	is 80 percent, but the only thing keeping out
23	competition is this guy's anticompetitive conduct. Now,
24	the guys at the end of the table would go after this guy
25	at 5 percent it sounds like, but let's put that aside.

184

1	He's 80 percent, and he's doing bad stuff, and he's
2	keeping the competition out. If he didn't keep doing
3	the bad stuff, the competition would come in. They
4	might even swamp him.
5	MR. de la MANO: So, let me now link that
6	question to the previous question to Simon, which is
7	where should we put the threshold for the finding of
8	dominance, and, of course, Simon has argued 40 percent
9	might be a good place. I am not sure it is a good
10	place, and there are a number of reasons why 40 percent
11	might be too high.
12	First of all, dominance is going to be a
13	necessary requirement, and in some cases, like the
14	situation you just presented, it may well be that if the
15	practice is preventing entry in the market, but in
16	assessing dominance, what we are ultimately assessing is
17	the situation without such practice. That's why
18	dominance is a screen. In a case like that, it would
19	not be possible to be brought forward by the European
20	Commission.
21	Now, that clearly -- you might say, "Well,
22	that's wrong," and that's why you have attempted
23	monopolization in the U.S. and we do not have it, but a
24	second reason why if dominance acts as a screen, we have
25	to be very careful in not setting the market share

185

1	threshold for a finding of dominance far too high.
2	There is a third reason, which is, as has
3	already been highlighted by Simon before, which is
4	market definition is an imprecise exercise. Now, I
5	think everybody here will argue that in some cases, if a
6	company has a share slightly above 40 percent, slightly
7	below 40 percent, you know, it probably doesn't make
8	much of a difference, but if you have a threshold at 40
9	percent, it is critical.
10	So, even though in practice, a firm with 35 or
11	45 percent is probably likely to have much more -- the
12	same kind of market power, in theory, this is a
13	threshold at which it either -- the Commission is going
14	to intervene or not, whereas if you had a lower
15	threshold -- and, of course, market definition is going
16	to be critical there. It is going to determine whether
17	or not the Commission is going to intervene or not. If
18	you have a lower threshold, then the precision of the
19	market definition exercise matters much less, because if
20	you had it wrong and the market definition was actually
21	too narrow or too wide, but you are wedding yourself
22	into the 20-30 percent threshold, it doesn't really
23	matter.
24	As long as you are below 25 percent, even if
25	you've got market definition wrong, it is for certain,

186

1	almost for certain, that there are going to be no
2	problems, and therefore, there should be no intervention
3	whatsoever.
4	DR. STELZER: To ask a practical question, what
5	makes you look at something in the first place? You go
6	into a bunch of market share studies and you say, "Oops,
7	here's a 40-percenter, I'll go after him"? Or is it
8	some practice that makes you look?
9	MR. de la MANO: The latter, essentially a
10	complainant would --
11	DR. STELZER: Simon says no.
12	MR. BISHOP: Well, Miguel said it right. It is
13	some complainant submits a case.
14	DR. STELZER: Right. Now, as I understand the
15	EU attitude, it differs from the American. Here my
16	economist friends believe that if the complaint comes
17	from a competitor, it is therefore tainted somehow. It
18	is the use of the legal system as a strategic device.
19	That is different from the EU, and I think the EU is
20	right but is the EU sticking with the notion that the
21	fact that a complaint comes from a competitor does not
22	taint the complaint?
23	MR. de la MANO: Well, practically in all
24	cases -- probably in all cases that I have been involved
25	in, the complaint has come from the competitor, some

187

1	outliers where a consumer may bring the case, but it is
2	very, very rare. When that happens, because we have an
3	opportunistic system, the Commission, of course, has to
4	take in mind the private interests of the complainant
5	and how that might taint their submissions, but
6	ultimately the Commission is obliged to give its
7	decision, whether it is a decision to intervene, and
8	therefore -- and that would be trying an independent
9	objection sent to the dominant company or allegedly
10	dominant company, or there would be a rejection of the
11	complaint, which would be a formal rejection, would be
12	written and sent to the complainant.
13	So, either way, the Commission basically has to
14	make up its mind, and in doing so, has to definitely
15	take into account to find out if the evidence that has
16	been brought forward to it is submitted by parties which
17	have their own interests at heart.
18	DR. WERDEN: Tom, I have a question for you.
19	You seem to be saying that the mere exercise of
20	exclusionary market power is a section 2 offense all of
21	the time, but I want to clarify if you mean without
22	regard to the potential of that conduct to create or
23	maintain something we would call monopoly power.
24	MR. KRATTENMAKER: I do not mean that.
25	DR. WERDEN: Okay, that's great.

188

1	MR. KRATTENMAKER: Thank you.
2	DR. WERDEN: Anybody want to follow up on that?
3	MR. KRATTENMAKER: Irwin says no.
4	DR. WERDEN: Well, say it out loud.
5	DR. STELZER: But brevity is so much the soul of
6	wit that I hated -- I just preferred to let your answer
7	hang out there.
8	MR. KRATTENMAKER: Sort of like a beautiful
9	arcing three-point shot that's probably right dead bang
10	through, nothing but the net, exactly, just let it sit
11	there.
12	DR. STELZER: Right, see, but I play basketball
13	at 10,000 feet.
14	MR. KRATTENMAKER: Of course you do. You are a
15	good guy.
16	DR. STELZER: I was trying out ideas. I am not
17	sure. Tom, tell me why you think about that.
18	MR. KRATTENMAKER: Oh.
19	DR. STELZER: How, as a practical matter, you
20	would tell in a case.
21	MR. KRATTENMAKER: Because there is lots of --
22	because the whole point about the competitive process is
23	to beat your rivals, and so inferring from the fact that
24	practice has an untoward effect on rivals, that it
25	therefore violates the antitrust laws, it is just too --

189

1	to coin a phrase -- over-inclusive.
2	DR. STELZER: Yeah, okay, but -- I guess I was
3	thinking in terms of defending the competitive process,
4	not competitors.
5	MR. KRATTENMAKER: Yeah, right.
6	DR. STELZER: And that's harder.
7	MR. KRATTENMAKER: Well, I agree. I mean, the
8	fact that you inflict some sort of inefficiency on your
9	rival, you could say, "Gee, that's bad, and we ought to
10	stop it," and that's kind of like the Klor's case.
11	That's Klor's against Broadway-Hale. I mean, they might
12	have done something bad, and we could care for less that
13	there were a hundred other stores in that city, and, I
14	mean, there is a way I used to tell that. I mean, I
15	went back to the record and examined that case, and it
16	turns out that the reason that there was this dispute
17	here was that the owner of Broadway-Hale had a
18	ne'er-do-well son who had impregnated and run away with
19	the daughter of Klor's, and this was an alienation of
20	affection suit brought as a Sherman Act case.
21	Now, of course, that is not true, but I tell
22	that story and the students believe it, and so that's
23	the long way of saying I do not think that section 1 --
24	of course, we are not talking about section 1 -- was
25	meant to federalize the tort of alienation of affection.

190

1	So, not only are you supposed to beat up on your rivals,
2	but not everything you do to your rivals is either
3	necessarily commercially motivated or motivated to drive
4	monopoly profits.
5	MR. SIMS: And, Irwin, if you don't demand that
6	the conduct have at least a high likelihood of creating
7	durable monopoly power, then you really do have a
8	serious risk of sticking your nose into the market where
9	you are going to do more harm than good, because
10	differentiating between exclusionary practices on some
11	grounds other than whether they have the potential to
12	create durable market power seems to me to be very hard.
13	DR. STELZER: But you used the term "durable"
14	about five times. What do you mean?
15	MR. SIMS: I mean more than temporary.
16	MR. KRATTENMAKER: There you go.
17	DR. WERDEN: Your turn, Irwin, as if you haven't
18	talked enough.
19	You seem not to at all be a fan of limiting
20	principles, and I want to push the limit on limiting
21	principles. Are you suggesting, for example, that the
22	Brooke Group rule was a really bad idea?
23	DR. STELZER: I don't have any idea.
24	DR. WERDEN: You don't think that in a predatory
25	pricing case, a plaintiff should have to show pricing

191

1	below some measure of cost?
2	DR. STELZER: Oh, no, I think that's ridiculous,
3	and I'll tell you why. First of all, I don't believe
4	you can measure marginal cost. I've spent a lot of time
5	trying to do that.
6	DR. WERDEN: The courts do not like marginal
7	cost either.
8	DR. STELZER: I'll take any kind of cost you
9	want. I don't think you can do it. I've been in enough
10	proceedings at regulatory agencies where people are
11	supposed to measure costs to know that.
12	Second of all, the real question with predatory
13	pricing is not whether the person prices below or at
14	some concept of cost and has a prospect of recoupment,
15	but think of it this way. You are walking along and you
16	want to have a picnic, and there's a sign that says, "No
17	trespassing." You figure, what the hell. You throw
18	down your blanket, you have a nice picnic, and you
19	leave, right?
20	Now you are walking along and there's another
21	field where you want to have a picnic and there's a no
22	trespassing sign, and there are about four or five
23	corpses lying around. Are you going to have a picnic
24	there? I don't think so.
25	So, what we are talking about is the kind of

192

1	practices that are entry-deterring in the technical
2	jargon, that scare the hell out of people, because
3	remember, this is more and more an age in which the
4	financing of new companies is done by venture
5	capitalists, and if you have ever been to a meeting with
6	a venture capitalist -- these are not very nice people,
7	many of them -- the first thing they want to know is
8	what is the range of practices available to the
9	incumbent competitors to keep you out or to destroy you
10	if you get in. That is what they want to know.
11	I mean, have you got a good idea? Yeah. Are
12	you a pretty good manager? Yeah. Can I suck most of
13	the value out of your enterprise? Yeah. And then they
14	want to know what are the incumbents going to do to you,
15	and if you go to enough meetings where people describe
16	what Microsoft might do to you or what other companies
17	might do to you, a lot of the stuff we are talking about
18	becomes irrelevant. Entry-deterrence is the problem.
19	Will they cut prices? Yes, they might. Is that okay?
20	Well, that's a tough one. That's very hard.
21	I know this sounds mushier than you'd like it to
22	be. People who say I am going to measure costs and then
23	I am going to measure market share -- in the Sirius/XM
24	merger, right, they are going to take one data point and
25	they are going to measure cross-elasticities and all

193

1	that other stuff? Ridiculous.
2	So, what I am saying is in a practical world in
3	which new firms are being created, in which technology
4	is increasingly important, in which small businesses and
5	new entrants are the manufacturers of macroeconomic
6	growth, I would lean pretty hard in the direction of
7	being very skeptical about the range of competitive
8	tools permitted to incumbents, to powerful incumbents,
9	for macroeconomic reasons, for microeconomic reasons,
10	and -- dare I say it, even though Judge Bork is a
11	colleague of mine -- for equity reasons.
12	DR. WERDEN: Are you suggesting that if the
13	incumbent is happily pricing at 100 and somebody has a
14	new idea and comes in and sells it at 80 and the
15	incumbent says, "Well, I better knock my price down to
16	80 or I am not going to make any sales," he's already in
17	trouble?
18	DR. STELZER: No, I am saying you have to look
19	at a lot of things. You see, that's the trouble. You
20	are trying to pick out one thing that will tell you what
21	the hell is going on in this industry. You can't do
22	that.
23	DR. WERDEN: Okay. Well, I concede that I can't
24	do that. So, what do I do?
25	DR. STELZER: You look at the entire range of

194

1	business practices of the company. You look at the
2	durability of its market share. You look at the history
3	of the notices it has posted in the past when
4	competitors try to come in, and you try to make a
5	decision as to whether those were imposing
6	inefficiencies on the potential competitors or not.
7	MR. WALES: Go ahead, Tom.
8	MR. KRATTENMAKER: I want to come to Irwin's
9	partial defense now --
10	DR. STELZER: Oh, God.
11	MR. KRATTENMAKER: -- on Brooke Group but make a
12	comment about -- to make a comment about what Joe said,
13	too.
14	On what Irwin said, you know, pricing below
15	cost, I am really not so sure. Recoupment, yes, and the
16	short answer to your question, Greg, is you have got to
17	show that they will be able to get their price back up.
18	When we all sit around and decide that we have this
19	common mantra and we decide to chant it, whatever this
20	antitrust religion is that we have, you have to be
21	careful to think about it once in a while.
22	Saying it has got to be below the pricing firm's
23	cost is to smuggle in the old efficient competitor rule
24	into the marketplace. If it is the case that the firm
25	can by pricing right down to its cost drive out four

195

1	firms and leave us with one firm instead of five in a
2	market, some people may say that drives us to more
3	efficient production, and other people will say that is
4	going to tend to drive prices further away from costs.
5	It depends on which value you think is important in
6	antitrust.
7	I think it would be better to have a discussion
8	about that than the silly stuff in Brooke Group about
9	what we happen to know because we happen to put on black
10	robes and so we are infallible, that people often try
11	predatory pricing and rarely succeed, a statement which
12	I believe had no support. There might be a footnote
13	there, but it doesn't cite any empirical work.
14	So, I don't mean to say that I am opposed to
15	Brooke Group, but what I mean to say is you don't look
16	askance at somebody and say, "You mean they wouldn't
17	price below cost?" Irwin is talking about a somewhat
18	different set of values and in this case a very
19	defensible set of values, particularly if you do keep
20	the recoupment link, I would say.
21	The other comment, I mean, I think this is the
22	right time to make it, I thought Joe had one of the most
23	interesting observations I've heard in a long time about
24	the bright line rules and fact-based rules, and that's
25	exactly what has happened to merger law in the whatever

196

1	years since Joe and I first started studying merger law,
2	but it's not what's going on in section 2, and these are
3	hearings about section 2.
4	You've got some cases that were sort of driven
5	down to fact-based. Aspen Ski is one of those where
6	they looked in the record and found that there were some
7	angry skiers in Atlanta, and Kodak copiers is one of
8	those, but we have some bright line cases, too,
9	Weyerhaeuser, Brooke Group, the 11th Circuit decision in
10	Schering-Plough, that say, do not tell me any facts.
11	All I want to hear is some theory.
12	So, in section 2, we are in -- I'll shut up here
13	now in a minute -- in section 2, we are at this funny
14	point where we haven't moved to Joe's Nirvana, and I
15	think we need to face that.
16	MR. SIMS: See, it is interesting. I agree with
17	you on Brooke Group and Weyerhaeuser. Those are
18	essentially safe harbor decisions.
19	MR. KRATTENMAKER: Yeah.
20	MR. SIMS: But I would vehemently disagree with
21	you on Aspen Ski and Schering-Plough. I think that
22	Aspen Ski is certainly not fact-based. You can't do a
23	fact-based analysis of Aspen Ski and conclude that there
24	was an antitrust violation there.
25	MR. KRATTENMAKER: No, the fact they found turns

197

1	out not to be a violation -- turns out not to be an
2	anticompetitive act, but --
3	MR. SIMS: Well, that's certainly true, and I
4	think Schering-Plough I think did focus on the facts,
5	and the fact that was determined -- that was found to be
6	determinative in Schering-Plough was the existence of
7	the patent and the scope of that patent. That's a
8	fact-based analysis to me, not rule-based.
9	DR. STELZER: Can I ask you something about
10	Aspen Ski, because I am not a lawyer --
11	MR. SIMS: Sure.
12	DR. STELZER: -- although I was involved in that
13	case just because I happened to be in Aspen at the time
14	and the plaintiff couldn't afford anybody and I was
15	free.
16	MR. SIMS: I remember actually visiting you in
17	Aspen periodically.
18	DR. STELZER: Right. Well, come this summer,
19	because I don't have judges setting my schedules
20	anymore.
21	Let me ask you something. There was an
22	unchallenged determination of the relevant market.
23	MR. SIMS: Yes, that was the --
24	DR. STELZER: Now, is that a fact or is that not
25	a fact?

198

1	MR. SIMS: That was a lawyer error, actually.
2	That was a stipulated market which any good antitrust
3	lawyer wouldn't have done.
4	DR. STELZER: All right. So, we are now down
5	to, if I understood it, it is not a fact if it is
6	determined by a judge and a jury but it is a lawyering
7	error. Is that right? So, that makes it not a fact.
8	MR. KRATTENMAKER: That's our position and we
9	are sticking to it.
10	DR. STELZER: Okay, that's all right, I just
11	wanted to know.
12	DR. WERDEN: Moving right along, Joe, I am not
13	entirely sure I understand your position. I am not sure
14	that you go so far as to say clarity is bad. I think
15	your position more is that hoped for clarity isn't going
16	to come in a useful way, to which my follow-up question
17	is, well, aren't there things like the Brooke Group rule
18	that would form conduct-based safe harbors that might be
19	a good idea? For example, that it is okay to introduce
20	a new product even if that causes your competitor to
21	fail?
22	MR. SIMS: Well, I wouldn't have any problem
23	with that rule, but I think you'd have a lot of trouble
24	getting broad consensus on it.
25	DR. WERDEN: I am willing to try. Let's see

199

1	what we can do here on the panel.
2	MR. SIMS: You might find some people that think
3	that's what Microsoft did and does and is doing --
4	introducing new products that are creating competitive
5	harms; at least I think that's the theory in the EU's
6	current preoccupation with Microsoft. So, I am fine
7	with a Brooke-type safe harbor for new product
8	introductions. I am not exactly sure how you'd set it
9	out so that you left it open for the one in a however
10	many times that might be anticompetitive, but I'd be
11	fine with that. I doubt seriously that you would get
12	broad consensus on that.
13	My point is that there is not incredibly broad
14	consensus on the Brooke Group rule, which is I think
15	about the only effective safe harbor in section 2 now.
16	So, I am not sure that you would have a very easy time
17	coming up with consensus on any others. I am happy to
18	see you try, and I could come up with a number that I'd
19	be comfortable with, but I doubt that I'd get everybody
20	to join with me.
21	DR. WERDEN: Well, we can give you 30 more
22	seconds. How many can you give me in 30 seconds?
23	MR. SIMS: Well, new product design would be
24	fine. I mean, in general, new products and product
25	design decisions, I am involved now in defending Apple

200

1	in the iPod tying cases. We shouldn't have to go
2	through all the hassle that we are going to have to go
3	through to get rid of those cases. So, I am perfectly
4	happy with that if you can find enough consensus to
5	implement it.
6	DR. WERDEN: Do I hear any dissenters?
7	DR. STELZER: Well, I was just curious, Joe,
8	what about what they call fighting brands in the
9	cigarette industry?
10	MR. SIMS: What about them?
11	DR. STELZER: That's a new product.
12	MR. SIMS: Is there anything wrong with that?
13	DR. STELZER: Is there anything wrong with that?
14	MR. SIMS: No, I don't see anything wrong with
15	that. Did it impair competition in some way?
16	DR. STELZER: It had very negative effects on
17	some of the competitors who made the brands.
18	MR. SIMS: That's different.
19	DR. STELZER: But it sends a notice that you are
20	going to come in --
21	MR. SIMS: Look, I happen to know an awful lot
22	about the cigarette business, unfortunately, because I
23	just did a merger there a couple years ago. There are
24	one heck of a lot of independent sellers of cigarettes
25	in the cigarette business. In fact, they have driven

201

1	the market share of the market leaders down, and more
2	importantly, they have taken away a big part of their
3	margin, which is why the FTC decided not to challenge
4	the merger of the number two and number three players.
5	DR. STELZER: Okay.
6	MR. KRATTENMAKER: I worked on that case, too,
7	and some of what Joe just said is true.
8	DR. WERDEN: Probably some of what Joe says is
9	always true; it is a question of how much.
10	MR. KRATTENMAKER: I was on the other side, I'm
11	sorry, I was doing it for the FTC.
12	MR. de la MANO: I would defend, Greg, that
13	particular bright line rule.
14	DR. WERDEN: Okay. When is a new product
15	introduction a bad thing for consumers?
16	MR. de la MANO: I think that's the wrong way to
17	put the question. I think no bright line rule is going
18	to work unless you define it very, very carefully, and
19	you will --
20	DR. WERDEN: Of course. That's what your job
21	is.
22	MR. de la MANO: Well, that's what we found in
23	the new product rule that we were given by the court in
24	the area of refusal to supply, the new product test,
25	that -- it sounds fine in the context of that particular

202

1	case, I admit, but we just do not know what's a new
2	product.
3	DR. WERDEN: Well, but if we are going to take a
4	European approach to this question, then perhaps we
5	should appeal to our ordoliberal traditions where, what
6	we say in English, competition on the merits was a
7	fundamental principle. That was legal without regard to
8	its effect, and there are reasons to believe that this
9	concept is embraced by Article 82.
10	Now, as far as I can tell, no European court has
11	ever said that that actually means something, but it
12	should mean something, shouldn't it?
13	MR. de la MANO: Definitely.
14	DR. WERDEN: Okay, what does it mean?
15	MR. de la MANO: Well, the problem is that if
16	you put the question in terms of would a new product
17	ever constitute the situation where it could lead to
18	consumer harm, I think the answer is always going to be
19	no. That is competition on the merits. That is a
20	situation where there's going to be traditional value to
21	consumers, that's pretty obvious, but the difficult
22	thing for a competition agency is to define or identify
23	whether that product is, indeed, new, and there are many
24	situations where what might appear on the face of it to
25	be a new product, from the perspective of certain

203

1	customers, but is just an extension or an additional
2	feature that's added to an old product, but if that
3	additional feature serves the purpose of preventing
4	entry, then maybe there is a problem.
5	DR. WERDEN: I agree there's always going to be
6	a fine line, and Irwin correctly pointed out that the
7	fine line is Brooke Group is a serious problem. We
8	can't figure out costs well. But that doesn't mean
9	there's something fundamentally wrong with the
10	principle.
11	MR. de la MANO: Absolutely not. It's not just
12	a good bright line for enforcement.
13	DR. WERDEN: You are coming to that decision
14	awfully fast. How long have you been applying it?
15	MR. de la MANO: I don't think we have had a
16	single case in the IMS where we have actually been able
17	to define a new product as of -- that's a few years.
18	DR. WERDEN: Of course, the bright line rule
19	there is that you can refuse to license. That solves
20	that problem, doesn't it?
21	MR. de la MANO: Yeah, solves that one, yeah.
22	DR. WERDEN: Okay.
23	MR. WALES: Should we move on to the principles?
24	Go to the first one.
25	DR. WERDEN: Okay, I hope you people can see

204

1	this. We are going to read these.
2	MR. WALES: I actually have them in hard copies
3	and we can pass them out.
4	DR. WERDEN: Okay. We are going to read them
5	into the record in any event.
6	We have in most of our sessions, but not this
7	morning, gone through what we call the propositions
8	where we put up a declarative sentence and ask the
9	panelists whether they agree or disagree and why.
10	The first one we have here is, "Monopoly power
11	is the long-term ability of a firm to earn greater than
12	a competitive return on investment."
13	It's not the most orthodox definition of
14	monopoly power, but it happens to be the almost verbatim
15	the definition in one of the leading economics
16	textbooks, and it focuses attention on something that in
17	principle we might be able to figure out, although it's
18	not going to be easy, whether a firm is earning more
19	than a competitive rate of return.
20	So, Tom, why don't you start.
21	MR. KRATTENMAKER: I think it is good enough for
22	government work.
23	DR. WERDEN: Good enough for the courts of the
24	United States of America?
25	MR. KRATTENMAKER: Not having tried to do a case

205

1	under this test, I would want to think some more about
2	whether I'd rather be going and getting evidence about
3	competitive returns than I would about prices and costs,
4	Greg. So, I cannot answer your question. I am
5	obviously -- as a lawyer, I am, of course, hind-bound, I
6	am always looking backwards, and so I am happier with a
7	test that focuses on price than competitive return if
8	you give me 30 seconds to think about it, but --
9	DR. WERDEN: Well, that's fine. It doesn't say
10	here what the evidence would be, and I think it would be
11	prices and costs in some cases, most cases, but the
12	question then is going to be, what price and what cost?
13	MR. KRATTENMAKER: Thank you for modifying this
14	as we go. It has changed from long-term to long-run, it
15	has changed from competitive return to pricing above
16	costs. I think it is basically right, but I want to say
17	the devil's in the details, but there are some details
18	that would need to be worked out, but sure.
19	DR. STELZER: Would you accept --
20	MR. KRATTENMAKER: As you know, I'd also say
21	that's also market power. I do not know, is that the
22	next question? Do we have another question about that?
23	DR. STELZER: Can I ask you a question?
24	DR. WERDEN: Please.
25	DR. STELZER: Would you substitute cost of

206

1	capital for competitive return on investment?
2	DR. WERDEN: Possibly.
3	DR. STELZER: Okay. Have you ever been in a
4	utility case where they're determining the cost of
5	capital?
6	DR. WERDEN: We hardly ever do that anymore,
7	thank God.
8	DR. STELZER: You hardly ever do it, but if you
9	walk down the block, there's a lot of people doing it.
10	There's economists doing it all the time and there's a
11	huge dispute about it, but I think cost of capital is at
12	least more precise as far as the literature goes than a
13	competitive return on investment. So, if you want to
14	play with this, I think you should do it in terms of
15	cost of capital, because there are all sorts of ways of
16	measuring cost of capital, and no one will know -- they
17	won't know with as much precision what you are talking
18	about when you talk about a competitive return.
19	DR. WERDEN: Well, coming back to Tom's
20	question, if you want to put this in terms of prices and
21	costs, the question, as I said, is what price and what
22	cost?
23	MR. KRATTENMAKER: Sure.
24	DR. WERDEN: And in particular, the difference
25	between monopoly power and market power, it is

207

1	conventional, at least, although there are some
2	dissenters, to define market power as the ability to
3	price above short-run marginal cost, but hardly anybody
4	would say that the right definition of monopoly power is
5	the ability to price above short-run marginal cost,
6	because that would give us too many monopolists.
7	MR. KRATTENMAKER: I think your second sentence
8	is correct and your first sentence is wrong.
9	DR. WERDEN: So, what is the definition of
10	market power?
11	MR. KRATTENMAKER: I believe that market power
12	has a durability component as well, the last time I read
13	the Guidelines, nontransitory.
14	MR. WALES: So, shorter, Tom, is that the point?
15	It is shorter than monopoly power?
16	MR. KRATTENMAKER: No, it is the same.
17	MR. WALES: So, both qualitative and
18	quantitative? I guess you made the point that
19	qualitatively, they're the same, but are they also
20	quantitatively the same?
21	MR. KRATTENMAKER: Oh, I think each of them
22	comes in degrees, Dave, I'm sorry. To go back to my
23	metaphor -- they could turn out to be a one-point shot,
24	a two-point shot, a three-point shot. I don't think it
25	would serve us any value to say, well, if it is a

208

1	two-point shot, it is market power, and if it is a
2	three-point shot, it is monopoly power. I don't -- as a
3	matter of moving the cases along, I don't see the point.
4	DR. WERDEN: Well, let me put the question,
5	then, doesn't it make sense to have a significant
6	threshold in a section 2 case that is different and
7	higher than the threshold of market power in a section 1
8	case? And don't the cases pretty much say that's the
9	law now?
10	MR. KRATTENMAKER: No. Yes.
11	DR. WERDEN: Okay, at least that was clear.
12	MR. BISHOP: But, I mean, the European
13	perspective, I mean there is some debate in Europe about
14	whether we can characterize firms which are dominant and
15	those firms which are super-dominant, which is sort of,
16	you know, similar to this, and my sense is that, you
17	know, why bother introducing this new term, you know,
18	"super-dominant"? If we are just going to use the
19	dominance as a threshold step to deciding whether we
20	need to investigate in more detail the competitive
21	conduct, whether a firm is dominant or super-dominant
22	doesn't really make any difference in that decision.
23	DR. WERDEN: Okay, let's move to the second
24	proposition. I think Joe spoke precisely these words,
25	and I want to see how much consensus we have on the

209

1	proposition that monopoly power is rare.
2	MR. WALES: If we can go back to Miguel.
3	MR. de la MANO: Well, in line with any
4	consensus that monopoly -- it makes very little sense to
5	distinguish between market power and monopoly power for
6	the reasons that have been explained on both sides of
7	where I am sitting, I would say monopoly power is fairly
8	common. The key question is, however, how much of it do
9	you really need to show or need to have before you
10	decide to investigate any further? Being shown monopoly
11	power is not anything in itself; it is the practice
12	itself, the conduct.
13	DR. WERDEN: I think you have identified one of
14	the major differences in attitude between the European
15	school and ours. Our courts are really hard sells on
16	the subject of monopoly power. It is an empirical fact
17	that it is very hard to convince a court that a firm has
18	a monopoly in the United States, and it's not that hard,
19	it seems, in Europe.
20	I think you have already cast your vote that it
21	is probably too hard in the United States. Anybody else
22	want to weigh in on that?
23	MR. KRATTENMAKER: Well, yeah. I mean, I think
24	that Miguel has really laid his finger on it. If we
25	then say that you possess market or monopoly power if

210

1	you face a downward-sloping demand curve, I think it may
2	well be that many, perhaps most firms, do, but the
3	second thing I was going to say is this question,
4	monopoly power is rare, is exactly why I went to law
5	school instead of graduate school in economics. You
6	have to ask an economist who does not I/O theory, but
7	I/O reality, how often this happens. Isn't this what
8	Joe Bain spent his life trying to do, but --
9	DR. WERDEN: I don't think so, but --
10	MR. KRATTENMAKER: Okay.
11	DR. WERDEN: Anyone else?
12	MR. de la MANO: Can I reverse the question?
13	DR. WERDEN: Rare is power monopoly?
14	MR. de la MANO: No. Do you think
15	contestability of a market is rare?
16	DR. WERDEN: I think it is unheard of.
17	MR. de la MANO: Well, there you go.
18	DR. WERDEN: I am not sure where I am.
19	MR. BISHOP: How does that follow?
20	MR. de la MANO: Well, it follows that if
21	contestability is the opposite of monopoly power and
22	contestability is unheard of, it must be because most
23	firms have market power.
24	DR. WERDEN: Well, but then you are equating
25	market and monopoly power, and I am not buying into that

211

1	one.
2	MR. de la MANO: Okay.
3	MR. BISHOP: And I guess it also relates to
4	entry to a market. You can have firms with high market
5	shares subject to effective competitive constraints
6	because the small rivals could easily expand.
7	DR. WERDEN: Okay, a third proposition, and this
8	is something that Simon already said. "The Cellophane
9	fallacy likely does not apply in attempt to monopolize
10	cases." Of course, he didn't use that language, because
11	that's American language, but here we have an offense of
12	attempt to monopolize in which the defendant doesn't
13	start out dominant, but it is alleged that he would end
14	up dominant with a dangerous probability through the
15	activities that he's engaged in, and in defining the
16	market in such a case, the proposition is that the
17	Cellophane fallacy probably isn't a problem.
18	Simon I think already said yes, that's true. Do
19	we have any other views?
20	MR. BISHOP: Easy one.
21	DR. WERDEN: I think that's an easy one. I like
22	easy ones.
23	Next, "When the Cellophane fallacy does apply,
24	which is not a significant number of cases, the proper
25	benchmark price in market delineation is the market

212

1	price absent the challenged conduct, which is normally
2	not the competitive price."
3	It is often said, perhaps rashly and wrongly --
4	we are going to find out -- that you should go down to
5	the competitive price to do the market definition
6	analysis. This proposition says no, you should look at
7	some kind of but-for price, and Simon, what do you think
8	about that?
9	MR. BISHOP: Interesting theoretical question.
10	The answer is sort of, maybe, but I think in the sort of
11	practical reality, it makes no difference. You don't
12	know what the but-for price is; you don't know what the
13	competitive price is.
14	DR. WERDEN: As a practical matter, you may be
15	exactly right, but let us suppose you could actually
16	figure these things out. What would you do?
17	DR. STELZER: And if my grandmother had wheels,
18	she'd be a bus.
19	MR. BISHOP: If you think about these things,
20	then all we need to do is be concerned with the
21	Cellophane fallacy or anything. The whole antitrust
22	would be very, very easy.
23	MR. SIMS: And that is how we get ourselves into
24	the messes that we get ourselves into, is pretending
25	that we can ignore reality.

213

1	MR. KRATTENMAKER: I think this is a very
2	interesting concept, and it might be right, but I didn't
3	understand the earlier question, and I don't mean this
4	as a challenge, Greg, but if you -- if we know both the
5	market price absent the challenged conduct and we also
6	know the competitive price?
7	DR. WERDEN: Yes.
8	MR. KRATTENMAKER: And you are making two
9	statements, which is that those are normally
10	different --
11	DR. STELZER: Right, and then which is the
12	benchmark?
13	MR. KRATTENMAKER: And then I would choose one?
14	DR. WERDEN: Yeah. I am not saying these things
15	are easy to figure out. They are not. I agree with
16	Simon.
17	DR. STELZER: They are impossible. It's not
18	that they are not easy.
19	MR. KRATTENMAKER: I am only clarifying the
20	question. The question assumes that I know these two
21	prices that are in here, and so you are asking -- you
22	are making a statement and asking us about a statement
23	and a value choice.
24	DR. WERDEN: I'll let you know everything that
25	you'd like to know.

214

1	MR. KRATTENMAKER: Okay, I know the market price
2	absent the challenged conduct, and I know the
3	competitive price, and I know that the market price
4	absent the challenged conduct is higher than the
5	competitive price.
6	DR. WERDEN: Yes.
7	MR. KRATTENMAKER: Simon's the expert, but I'd
8	be inclined to say that the right answer whatever the
9	empirical fact is, that the right answer is you focus
10	not on the price absent the challenged conduct but on
11	the competitive price, but I thought his basic answer
12	was correct --
13	DR. WERDEN: Why?
14	MR. KRATTENMAKER: -- which is, you know, I do
15	not know either better than the other.
16	DR. WERDEN: I don't want you to give an answer
17	now. I want to know why.
18	MR. KRATTENMAKER: Because that is what we are
19	more likely to be able to assess the supply and demand
20	responses to, that --
21	MR. BISHOP: But doesn't --
22	MR. KRATTENMAKER: -- as the market definition
23	process asks us to do.
24	MR. BISHOP: But this comes down to, I mean,
25	there's practically no difference. I mean, if you knew

215

1	what the competitive price was in every single industry,
2	antitrust policy would be extremely easy, just go around
3	and tell firms that you are not allowed to price more
4	than the competitive price.
5	MR. de la MANO: I wouldn't be so drastic on
6	that, Simon. I think the question has merit. I do not
7	know what the theoretical answer to this is, but I think
8	from a practical standpoint, I actually think it could
9	be easier in some cases to assess what the price would
10	be in the absence of the conduct given that we are very
11	unlikely to see, going back in time, a market which is
12	currently not competitive that might have been
13	competitive in the past, but it is very likely to see a
14	situation that a few years ago, a market being a
15	monopoly was one where that conduct was absent, and it
16	might be possible to compare or even do some natural
17	experiments across regions, even contemporaneously, to
18	compare what is the precise situation where the conduct
19	is absent. So, this theoretical conversation, were it
20	to be valid, I think in practice, it could be very
21	useful.
22	MR. BISHOP: Well, I still think that, you know,
23	either benchmark means that the inferences that you can
24	draw from, you know, the available data is similar to
25	the same issues, whether it is a competitive price or a

216

1	price absent the conduct. Just seriously, from a
2	practical point of view, I do not think it makes any
3	difference at all. We can have a, you know, good, you
4	know, theoretical debate in saying which one is the
5	appropriate one, but from a practical point of view, I
6	do not think there is any difference whatsoever.
7	DR. WERDEN: We have pretty much covered this
8	one, but we are going to put it up anyway, see if
9	anybody has anything more to add.
10	"A market-share based safe harbor is appropriate
11	in monopoly cases."
12	MR. BISHOP: Yes.
13	DR. WERDEN: Okay, we have one yes.
14	MR. de la MANO: Two.
15	MR. SIMS: What's the number?
16	DR. WERDEN: That's the next slide.
17	MR. SIMS: I can't answer it without the number.
18	DR. WERDEN: Pick your own number.
19	MR. KRATTENMAKER: I say no to this sentence
20	because it has a singular noun.
21	MR. SIMS: If you give me -- if you give me a,
22	you know, 70 percent or an 80 percent number, I might be
23	very comfortable with that.
24	DR. WERDEN: Okay, we have got a vote for 70 or
25	80 percent. We might not have unanimity on 70 or 80

217

1	percent.
2	DR. STELZER: What is it appropriate to? If it
3	is appropriate as a general prosecutorial guide for guys
4	picking cases to bring, along with the feasibility of
5	relief, then it might be useful, but --
6	DR. WERDEN: If it is a safe harbor, it is a
7	rule that courts are going to use on summary judgment to
8	kick out cases.
9	DR. STELZER: Then I would say no.
10	MR. SIMS: And I know Tom says no. He has to
11	say no.
12	MR. KRATTENMAKER: Yes, I did. I already said
13	no. I would say yes, it might make sense to have one
14	safe harbor --
15	DR. WERDEN: You're saying yes, but you're
16	coming in with a low number, right, 25?
17	MR. de la MANO: I find it hard to understand
18	this myth, which I alluded to before, that in Europe we
19	have a serious concern with type II errors, yet when it
20	comes to using market share safe harbors, there is
21	consensus here on this side of the table that they can
22	be used. Isn't that a sign that you want to leave open
23	the possibility to bring any type of case, irrespective
24	of market shares being rather low?
25	MR. SIMS: Well, no, that's not my reason at

218

1	least. My reason for being nervous about safe harbors
2	unless they're very high is the concern that the safe
3	harbor set too low will end up with serious
4	over-enforcement above that number.
5	MR. BISHOP: Okay, but this comes back to the
6	sort of philosophical or institutional, philosophical
7	differences between the EU and the U.S., because
8	personally, I would set the threshold at 70-80 percent,
9	but I'd much prefer in the EU to have one of 40 percent
10	than to have no threshold at all.
11	MR. SIMS: Okay, and that's a fair point given
12	the regulatory environment that you find yourself in.
13	MR. WALES: I guess one question I had, Tom, is
14	I thought I had read where you talked about the
15	possibility of having different thresholds perhaps for
16	different types of -- your two types of conduct. You
17	had the conduct where someone acts to reduce output on
18	their own as opposed to acting to exclude rivals, and I
19	guess you kind of left open the proposition I thought
20	that perhaps you might be willing to look for markets
21	with the former and not the latter.
22	MR. KRATTENMAKER: No, I might be willing to
23	look for one for each. That's why I said, my objection
24	to this is that it -- that the noun is singular.
25	DR. WERDEN: Do you have some numbers in mind?

219

1	MR. KRATTENMAKER: Do I have numbers in mind?
2	No, but I think you might well be able to come up with
3	market share based safe harbor for exclusionary conduct
4	section 2 cases.
5	MR. WALES: I have a question for --
6	MR. KRATTENMAKER: But it wouldn't, in my view,
7	be an appropriate -- it wouldn't be the same threshold
8	that would be appropriate for collusion-based section 2
9	type cases, which are generally rare but still can be
10	out there.
11	MR. WALES: A quick question for Miguel, I guess
12	where does 40 come from in terms of setting the
13	threshold level in the European Commission?
14	MR. de la MANO: Well, as far as I know, it is
15	from a case, but, I mean, I think the thing is -- I
16	think the discussion is also highlighting this -- there
17	is a question as to, you know, what is the threshold
18	going to be used for? If you believe that once you are
19	above the threshold, basically the case has been proven,
20	then clearly you want to have as high a threshold as
21	possible.
22	If, on the other hand, you believe as I do, at
23	least, that the threshold is just the first step, just
24	the screen to sort of ditch the cases which are
25	obviously not a problem, if you have sufficient

220

1	discipline imposed upon yourself as a competition
2	authority in what you need to prove further, there is no
3	problem in having a low threshold. In fact, it is
4	probably better to have a low threshold, because that
5	makes the assessment of your facts credible.
6	Otherwise, if you have a threshold at a sort of
7	middle level, such as 40 or 50 percent, there is always
8	going to be a group of people who think, a-ha, okay, so
9	this discipline you say you have, that you are going to
10	go after -- assessing the effects afterwards, after
11	showing dominance, it is not really true, because as
12	soon as you are above 50, it is really easy to assess
13	the facts, and therefore, there is no credibility to the
14	second discipline, as it were.
15	MR. BISHOP: Okay, but I would take a different
16	view, and sort of just to be clear here, when I said
17	that dominance in Europe is then inferred to be an abuse
18	of, you know, of that market power, that's not my
19	position. That's the position of the European courts,
20	that most of the issues we are talking about here are
21	exclusionary, and the courts have held that any harm to
22	a competitor necessarily leads to harm to competition,
23	and therefore, given that sort of standard by the
24	European courts, there is no room, really, for an
25	effects-based system.

221

1	So, as you lower the threshold from 40 percent
2	to 25 percent, it makes things much worse in Europe
3	unless the Commission is going to be very clear that
4	they are going to take on the courts and that court
5	reasoning, that you can infer harm to competitors
6	necessarily translates to harms to competition, that,
7	you know, the Commission is going to take that square
8	on, because if they do not, any lowering away from the
9	40 percent to just come out of case law is just going to
10	make things worse.
11	MR. de la MANO: The court has already told us a
12	few months ago that it is willing to reconsider its
13	previous positions on this matter, and in the Glaxo
14	decision -- and actually, it is actually an area of
15	Article 81, cartels or agreements, but it has made it
16	very clear that it is very open and willing to see a
17	more effects-based analysis on the part of the
18	Commission both in the area of assessing possible harm
19	to consumers, but also in the area of assessing
20	efficiencies. So, I think the courts are open to be
21	challenged by the Commission on this point.
22	MR. BISHOP: Well, I would just say, you know,
23	let's wait and see, stick with 40 percent and then see
24	how they move before lowering the threshold.
25	MR. WALES: Let's go to the next one.

222

1	DR. WERDEN: Skip the next one and go one
2	further.
3	MR. KRATTENMAKER: Can we mail in our answers to
4	the one, number seven?
5	DR. WERDEN: If you like. It is about
6	econometrics. Did you want to handle it, Tom?
7	MR. KRATTENMAKER: Of course. I mean, that's
8	the most fun, is talking about something that we do not
9	know. I thought it was a really interesting and
10	provocative question. I think it is largely correct,
11	but I would have some comments on it, but go ahead.
12	DR. WERDEN: We are nearing our end point.
13	MR. KRATTENMAKER: No, go ahead.
14	DR. WERDEN: As our end point, we are going to
15	take this last proposition from the Syufy case, one of
16	our failures in court.
17	"In evaluating monopoly power, it is not market
18	share that counts, but the ability to maintain market
19	share."
20	MR. KRATTENMAKER: Could there be anything more
21	incorrect?
22	DR. WERDEN: I imagine that there could, but let
23	me just add that I think what the quote is trying to say
24	is the point that Joe made several times, which is
25	durability is crucial in monopoly power.

223

1	MR. KRATTENMAKER: I see, okay.
2	DR. WERDEN: And monopoly power requires much
3	more durable power over price than market power does.
4	MR. KRATTENMAKER: Gotcha.
5	MR. SIMS: When I read this, my answer was, I do
6	not know exactly what these words mean --
7	MR. KRATTENMAKER: Okay.
8	MR. SIMS: -- but if they mean durable market
9	power, then --
10	MR. KRATTENMAKER: If they mean entry barriers
11	and -- okay, you are saying they're importing it, okay.
12	DR. STELZER: As a practical problem with that,
13	it is an easy matter in any case to find someone who
14	will tell you why whatever monopoly power or market
15	power you see is not durable. I have had people tell me
16	that monopoly power in the transmission of electricity
17	is not durable because they have some innovation in
18	mind.
19	In other words, you can fill the courtroom with
20	experts who will tell you why market power that has
21	persisted for 150 years is really not durable given some
22	new technology or given some new something, but --
23	DR. WERDEN: But they're wrong, aren't they?
24	But you are saying that they're wrong?
25	DR. STELZER: They're wrong.

224

1	DR. WERDEN: Okay.
2	DR. STELZER: So I would be very careful about
3	introducing a test that says not only do you have to
4	have market power, but it has to be proved to be durable
5	in order to create a problem, because that's an
6	impossible test to meet.
7	MR. SIMS: It is true, and I think everybody
8	should admit that it is true, that the more you get away
9	from slogans and general rhetorical concepts and the
10	closer you get to careful analysis of the facts, the
11	less enforcement you are going to have, because it is
12	harder. It is harder for plaintiffs, whether they're
13	the Government or private plaintiffs, to prove a case if
14	they have to slog their way through the facts.
15	That's why the per se rule is so attractive to
16	plaintiffs' lawyers in damage cases, because they do not
17	have to prove anything. So, you know, that's an
18	inevitable result of being more wedded to factual
19	analysis than setting up bright-line rules. I don't
20	think it is a reason not to do it, but it is a result
21	that we ought to be -- that we ought to recognize and
22	accept.
23	MR. WALES: Anybody else?
24	DR. WERDEN: Well, we are a few minutes past our
25	official end time, so why don't we wrap it up and take

225

1	one last opportunity to thank our panelists.
2	(Applause.)
3	MR. WALES: Thank you very much. I guess we are
4	adjourned.
5	(Whereupon, at 4:34 p.m., the hearing was
6	adjourned.)
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

226

1	C E R T I F I C A T I O N O F R E P O R T E R
2	DOCKET/FILE NUMBER: P062106
3	CASE TITLE: SECTION 2 HEARING
4	DATE: MARCH 7, 2007
5
6	I HEREBY CERTIFY that the transcript contained
7	herein is a full and accurate transcript of the notes
8	taken by me at the hearing on the above cause before the
9	FEDERAL TRADE COMMISSION to the best of my knowledge and
10	belief.
11
12	DATED: 3/12/2007
13
14
15
16	SUSANNE BERGLING, RMR-CLR
17
18	C E R T I F I C A T I O N O F P R O O F R E A D E R
19
20	I HEREBY CERTIFY that I proofread the transcript
21	for accuracy in spelling, hyphenation, punctuation and
22	format.
23
24
25	DIANE QUADE

Updated January 2, 2024