Understanding Single Firm Behavior : Monopoly Power Session


This document is available in three formats: this web page (for browsing content), PDF (comparable to original document formatting), and WordPerfect.



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 June 25, 2015