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

1

1UNITED STATES FEDERAL TRADE COMMISSION

2and

3UNITED STATES DEPARTMENT OF JUSTICE

4

5

6

7SHERMAN ACT SECTION 2 JOINT HEARING

8UNDERSTANDING SINGLE-FIRM BEHAVIOR:

9MONOPOLY POWER SESSION

10WEDNESDAY, MARCH 7, 2007

11

12

13

14

15HELD AT:

16UNITED STATES FEDERAL TRADE COMMISSION

17601 NEW JERSEY AVENUE, N.W.

18WASHINGTON, D.C.

199:30 A.M. TO 4:30 P.M.

20

21

22

23

24Reported and transcribed by:

25Susanne Bergling, RMR-CLR

2

1MODERATORS:

2DENNIS W. CARLTON

3Deputy Assistant Attorney General for Economic Analysis

4Antitrust Division, Department of Justice

5and

6JOEL L. SCHRAG

7Economist

8Bureau of Economics, Federal Trade Commission

9

10PANELISTS:

11

12Morning Session:

13Andrew J. Gavil

14Richard J. Gilbert

15Michael L. Katz

16Philip B. Nelson

17Joseph J. Simon

18Lawrence J. White

19

20Afternoon Session:

21Simon Bishop

22Thomas G. Krattenmaker

23Miguel de la Mano

24Joe Sims

25Irwin M. Stelzer

3

1C O N T E N T S

2

3MORNING SESSION:

4Introduction

5Presentations:

6     Andrew J. Gavil

7     Richard J. Gilbert

8     Michael L. Katz

9     Philip B. Nelson

10     Joseph J. Simons

11     Lawrence J. White

12Moderated Discussion

13Lunch Recess

14

15AFTERNOON SESSION:

16Introduction

17Presentations:

18     Simon Bishop

19     Thomas G. Krattenmaker

20     Miguel de la Mano

21     Joe Sims

22     Irwin M. Stelzer

23Moderated Discussion

24Conclusion

25

4

1P R O C E E D I N G S

2- - - - -

3MR. SCHRAG: Good morning. Sorry about the

4technical issues. Welcome.

5My name is Joel Schrag. I am an economist at

6the Bureau of Economics here at the Federal Trade

7Commission, and I am one of the moderators for this

8panel. My co-moderator, standing next to me, is Dennis

9Carlton, Deputy Assistant Attorney General for Economic

10Analysis at the Antitrust Division of the Department of

11Justice.

12Before we get into the substance of the program,

13on behalf of the FTC staff who have worked on this

14session, I would like to take the opportunity to thank

15all of our colleagues from DOJ for their hard work and

16their efforts to jointly present this session.

17In addition, after today's and tomorrow's

18hearings on monopoly power, the hearings will next turn

19to issues involving remedies later this month, and so I

20urge you all to be sure to check our agencies'

21respective web sites for updates on these future

22hearings.

23As the FTC representative, I do have just a few

24housekeeping matters to cover before we begin. First of

25all, please turn off all of your cell phones,

5

1BlackBerries and other noise-making electronic devices.

2Second, the restrooms are located out through the double

3doors and across the lobby. If you need help to find

4them, there are signs that should guide you.

5Third, one safety tip, especially for visitors,

6in the unlikely event that the building alarms go off,

7please proceed calmly and quickly as instructed. If we

8must leave the building, you exit out the New Jersey

9Avenue doors by the guard station. Please follow the

10stream of FTC employees to a gathering point across the

11street and await further instruction, but hopefully that

12won't be necessary.

13Finally, we request that you please not make

14comments or ask questions during the session. Thank

15you.

16Let me just say a few things about the session.

17Many of the prior sessions of the hearings addressed

18particular conduct that's been challenged under Section

192 of the Sherman Act. Today, the hearings turn to

20issues of monopoly power and market definition, and

21these issues we believe are very important.

22In fact, if you were at the opening day of the

23hearings back in June, both Herbert Hovenkamp and my

24co-moderator, Dennis Carlton, were given the opportunity

25to place the issues for the subsequent hearings in

6

1context. Both identified monopoly power and market

2definition as areas where there are difficult, uncertain

3questions that must be addressed in many cases, and I

4expect that today's panel will help to clarify, if not

5completely resolve, these difficult questions.

6The hearings will be organized as follows:

7First, we'll hear an approximately 15-minute

8presentation from each of our six distinguished

9panelists. We'll probably take a break after the fourth

10panelist and then come back from the break and hear from

11the two remaining panelists. After that, the panelists

12will have an opportunity to comment on each other's

13presentations, and we'll have a moderated discussion.

14So, I think I'd now like to turn things over to

15my co-moderator, Dennis Carlton, who will introduce our

16distinguished panelists.

17Thank you very much.

18DR. CARLTON: Okay, thank you. I am Dennis

19Carlton. I am a Deputy Assistant Attorney General in

20the Antitrust Division, and it is a pleasure to welcome

21all of you to these joint FTC/DOJ hearings.

22I had the privilege of participating in the

23opening session of the hearings, and one of the topics I

24said that needed clarification was precisely the topic

25of the panels today and tomorrow, a focus on what we

7

1mean by "market power" and "market definition" in

2Section 2 cases I think is really important.

3I am also a Commissioner on the Antitrust

4Modernization Commission, and despite my attempting to

5do so was not able to convince the Commission to study

6in depth the definition of market power and market

7definition in Section 2 cases and to report on it. So,

8that, I think, emphasizes all the more how important

9this session, this panel discussion, is today, and the

10real question is, can we reach consensus on any of the

11hard questions or at least can we reach a consensus that

12there's a lot of ambiguity and arbitrariness in what is

13going on?

14I am honored to chair such a distinguished

15panel. All of the members of the panel have extensive

16experience, both academic and nonacademic, in antitrust

17and have served both in the private sector and in the

18government sector.

19In the interest of saving time, I am going to

20introduce them all at once and hopefully by that time

21the computer will work. So, starting with Phil, Phil

22Nelson is a principal at Economists, Inc., an economic

23consulting firm. Previously, he served as the Assistant

24Director for Competition Analysis at the FTC and as an

25Adjunct Professor at Fordham Law School. He has written

8

1numerous articles and two books on antitrust topics, and

2he edited the ABA's antitrust section of market power --

3The Market Power Handbook. He currently is the

4vice-chair of the section's Healthcare and

5Pharmaceuticals Committee.

6Beside Phil is Joe Simons. Joe is a well-known

7attorney. He's a partner and co-chair of the antitrust

8group at Paul Weiss. Previous to that, Joe was the

9chief antitrust enforcer at the Federal Trade

10Commission, serving as the Director of the Bureau of

11Competition from June 2001 until August of 2003. He has

12the interesting characteristic of once being the tenth

13largest wireless carrier in the country, because I

14believe he was a trustee and had a lot of wireless

15licenses, but in addition to that, he has achieved

16something that's actually quite rare for attorneys to

17do, and that is he's written an article that economists

18cite all the time and is associated with critical loss

19analysis.

20Beside Joe, in a missing seat, is Larry White,

21who I am sure is on his way. Larry is the Arthur

22Imperatore Professor of Economics at NYU School of

23Business. He's the Deputy Chair of the Department of

24Economics. Previously, in the early eighties, Larry

25served as the Director of the Economic Policy Office in

9

1the Antitrust Division. Larry has written several books

2and articles, one of which is well-known to antitrust

3practitioners called The Antitrust Revolution:

4Economics Competition and Policy. He's currently the

5editor of The Review of Industrial Organization. Prior

6to serving at the Justice Department, he did extensive

7government service both for the Federal Home Loan Bank

8Board and for the Council of Economic Advisers.

9Andy Gavil is a Professor of Law at Howard

10University where he not only teaches antitrust, but he

11has also extensively written on antitrust many articles

12and has a very well-known case book with Bill Kovacic

13and Jonathan Baker, Antitrust Law in Perspective. He is

14about to publish or co-author a book called Microsoft

15and the Globalization of Competition Policy, which I am

16sure has focused on Section 2 type behavior. He's

17currently the articles editor of The Antitrust Magazine

18and serves on the ABA Antitrust Section's Liaison Task

19Force to the Antitrust Modernization Commission. He is

20Of counsel to the Sonnenschein Law Firm.

21To Andy's left is Rich Gilbert. Rich is a

22Professor of Economics at the University of California

23at Berkeley. He served as the Deputy Assistant Attorney

24General in the Antitrust Division in the mid-nineties,

25and at that time, he led the effort to write the

10

1Antitrust Guidelines for the Licensing of Intellectual

2Property. He has written widely on antitrust topics.

3He is currently the Director of the Competition Policy

4Center at Berkeley and is associated with the economic

5consulting firm of COMPASS.

6Finally, Mike Katz at the end of the table.

7Mike is currently the holder of the Sarin Chair in

8Strategy and Leadership at Berkeley, the Business

9School, and also holds an appointment in the Economics

10Department. Mike served as the Deputy Assistant

11Attorney General in the early 2000s, and he also served

12as the Chief Economist at the Federal Communications

13Commission. He's written numerous articles on economics

14and antitrust and has specialized in many topics,

15including network industries.

16So, with that introduction, I'll turn it over to

17our first speaker, Phil, and just let me remind the

18speakers, we're kind of running tight because we started

19late, so if you could keep to the 15 minutes, that would

20be good. The organization of this is going to be four

21speakers will go, 15 minutes, we'll take a 10-minute

22break, we'll have two more speakers. We will give the

23speakers a brief opportunity to talk to each other, and

24then I'll moderate a discussion for about an hour or so.

25Thank you.

11

1MR. NELSON: So, we have a -- are we moments

2away or should I just proceed without slides?

3DR. CARLTON: Is the computer still not working?

4MR. NELSON: Well, okay, the reason they put me

5first is the slides that you can't see are really sort

6of a background deck that gives you the background on

7market power. The first slide cites the definition of

8market power that's at the front of the monograph that

9the ABA published that was referred to earlier, which is

10market power is the ability of a firm or a group of

11firms within a market to profitably charge prices above

12the competitive level for a sustained period of time,

13and as you can't see on the screen, the word

14"profitably" is in italics, and so one of the important

15things in the definition is that a monopolist profit by

16doing this.

17If entry is easy, you may be able to raise

18prices, but not profitably, because somebody will enter,

19and if there are a lot of competitors, they can steal

20customers away from you, so you can't profit. That may

21become of importance in some of the discussion as to

22what type of performance evidence one might use in

23determining whether a firm has market power or not.

24A price above the competitive level, the

25"competitive level" was in italics, because people talk

12

1about the standard monopoly raising prices, and if you

2are not raising them above the competitive level,

3usually people don't care.

4Then a "sustained period of time" is in the

5definition because you may be able to opportunistically

6raise prices for a little bit, but again, entry or

7something might undermine the ability to do that.

8Now, in some of the legal cases, you see

9reference to the ability to exclude competition, and I

10will suggest that is something worth consideration,

11because in some contexts -- and there were FTC hearings

12many years ago about standard-setting organizations

13where there might be a collection of, let's say, 10 or

14more people making a particular product, and there might

15be enough competitors that they compete and charge a

16competitive price because there's so many people

17operating under that standard.

18Well, somebody may develop a new technology that

19would come in and completely take the market away from

20the incumbent competitors with the older technology.

21Acting jointly in that case, they might be able to block

22entry by controlling the standard-setting organization.

23Are they raising prices above the competitive level?

24They're excluding an entrant, somebody that would

25dynamically help the market with a new technology that

13

1might have better performance characteristics and be

2able to be sold at a lower price. What they get out of

3it is where their profit is, is that they get to earn

4the competitive rate of return rather than being maybe

5in bankruptcy court.

6So, while I gave you the standard definition,

7there are other things and other contexts, as you can

8see from the get-go, that you have to worry about in

9deciding whether a firm or a group of firms have market

10power. And today, largely we'll focus on dominant

11firms, but there are contexts where a group of firms

12acting together might have trouble. And if a dominant

13firm has control over a patent that's a blocking patent

14that blocks a new technology, he might have an interest

15in blocking the new technology just like the group of

16firms that ran the standard-setting organization has an

17incentive to block technology. So, that's one thing.

18The other thing that I wanted to highlight at

19the beginning is, some people talk about market power;

20some people talk about monopoly power. Often,

21economists mean the same thing, but in some contexts,

22people have defined them differently. Greg Werden is

23sitting there, and he's drawn a distinction in one of

24his articles and alludes to other people that

25distinguish market power and monopoly power perhaps in

14

1terms of the time period over which people have the

2ability to raise prices and the like.

3There are articles out there that talk about

4antitrust monopoly power, again, trying to make a

5distinction. And there, often the thought is if you

6have a differentiated product and thus have a

7downward-sloping demand curve for your product, you

8might have some degree of ability to raise prices above

9costs and you might in that sense have market power, but

10you might not have a substantial ability to do it.

11Because there are a lot of products out there that are

12roughly close substitutes, not exactly the same thing,

13and you might in that context have some market power but

14not antitrust monopoly power or antitrust market power,

15because you don't really have substantial ability to

16earn substantial profits and the like. So, some people

17try to distinguish that downward-sloping demand curve

18idea by talking about antitrust monopoly power.

19I think with that background, we're talking

20about antitrust market power. Something that's somewhat

21significant. And then different panelists may have

22different degrees of market power in mind when deciding

23how you go about measuring whether it is significant

24enough market power. So, with that sort of definition

25of market power, the next slide was going to lay out

15

1sort of the touchstones in a typical market power proof

2that you sort of run through, and the first thing people

3often define is product market definition. Then they go

4to defining geographic market definition.

5Once you have a relevant product/geographic

6market combination, often it is standard to look at

7market concentration in a monopoly case, and once you

8clear that hurdle and see that maybe it is substantially

9concentrated or a firm has a dominant market share, a

10high-level market share, you then start looking at

11things like entry conditions, other structural

12characteristics of the market. Maybe you look at in

13some contexts, you know, the structure of the buyer-side

14of the market, and if it is a collusion case type of

15monopoly power issue, maybe you look at the

16characteristics of the market that make it easier or

17harder for firms to collude in that market.

18Then finally, in a lot of the monopolization

19cases, you see a consideration of market performance

20evidence, and that's where you start having things like

21profit rates of return, profit margins, looking at

22prices over time or across geographic areas. You look

23at output patterns and how they vary with prices. And

24you look at new product introductions. You can either

25look at them in terms of formal econometric analysis or

16

1often you look at events -- market events that allow you

2to sort of control for some things -- and look at how if

3the events give you insights either directly into the

4market power or at some of the related issues like

5market definition.

6Now, increasingly, because of the success of the

7Merger Guidelines, you see references to the approach

8used in the Merger Guidelines of developing a relevant

9market in the context of monopolization cases, and there

10were a couple slides that sort of just quoted the

11Guidelines. I suspect with this audience, there is no

12reason to go through it, but it is the hypothetical

13monopolist test. Can the monopolist raise prices above

14the -- in the Guidelines, they talk more or less about

15the current level as opposed to the competitive level

16and see if that's profitable.

17Now, one thing that is worth pointing out,

18especially in transferring that concept, is that in the

19Guidelines themselves, Section 1.11 says that while you

20might look at prevailing prices in the Guidelines, there

21is a caveat that says if pre-merger circumstances are

22strongly suggestive of coordinated interaction, in that

23situation, the agency will use a price more reflective

24of the competitive price. So, there is a caveat in

25there where they don't always use prevailing prices.

17

1One sort of footnote is that the original

2guidelines were focused on coordinated effects, and then

3they later on added more information about unilateral

4effects. I think there's a little glitch here, because

5I think the Merger Guidelines actually should make a

6reference not only to coordinated interaction, but also

7if the dominant firms raise prices above the competitive

8level, then you might want to look at the competitive

9price level.

10Why might you want to do that? Well, that is

11because you get a different elasticity and different

12substitutes depending on at what price level you measure

13the substitution. And this is where the lack of slides

14really hurt us the most, because I put together an

15illustrative example of a demand curve with a concrete

16slope and all the rest, calculated the marginal revenue

17curve from that, showed where the competitive price

18would be, where basically price equals marginal cost,

19then showed where the monopolist would operate, which is

20at a higher price, and then estimated the elasticities

21of a couple of the different points along the demand

22curve.

23What you see is that even though a demand curve

24is a straight line and thus the slope is constant over

25the whole curve, the elasticity changes. And at the

18

1higher prices, the demand is more elastic. And, the

2reason that that makes sense is that a monopolist is

3going to keep raising its price, you know, and find a

4price that is more profitable. And in the monopolistic

5equilibrium, he has got a high enough price that demand

6becomes elastic and a further price increase would lose

7a lot of customers to other products. That is called

8the Cellophane fallacy -- that sets up the Cellophane

9fallacy, which is if you measure the elasticities at the

10monopoly price, you are going to run into problems

11because there are a lot of substitutes out there that

12are not substitutes at the competitive price. You can

13do all the econometrics you want and estimate the

14elasticities, but if you do not know whether you were at

15a competitive price or a monopoly price, that elasticity

16estimate does not tell you anything when you are doing a

17monopolization case particularly.

18So, then you get into this tautological

19situation. If you think about the paradigm of starting

20with a monopoly case and saying, "Well, do I have a

21monopoly here?" And you have to define the market, and

22you have to define a monopoly price to define the

23market, then why bother defining the market? So, you

24have got a couple of issues here that suggest, what do

25you do about it? And the rest of my deck talks about

19

1the sorts of things that one might look at. But, the

2basic thing that I wanted to suggest is -- while I think

3there are great problems with a simplistic analysis of

4the standard paradigm I outlined -- I think there are

5elements of it that, if you can go through it all, can

6help you in many circumstances unravel this thing and

7cross-check your conclusion.

8So, it is a way of organizing your story.

9Making sure that you look at your story or your analysis

10as consistent, and that it gives you insights into what

11you might look at. And where it leads you, I think, is

12looking more and more at some of the performance

13evidence. But you have got to be careful in looking at

14the performance evidence, because as economists have

15shown, things like profits and accounting data are

16tricky.

17Having said that, I also think that how

18difficult a problem it is varies a lot from market

19circumstance to market circumstance. I think it is

20probably trickiest when you are dealing with

21consumer-differentiated products, like Cellophane

22wrapping paper or something. It may be less of a

23problem when you are dealing with an input into an

24industrial process, where you can look at substitutes in

25a more maybe engineering approach type of way.

20

1My time is basically up, so to keep us on

2schedule, I would recommend -- they are going to post

3the slides later on, and they are written in a way that

4they are readable -- so I suggest you look at the slides

5for the rest of the story.

6Thanks.

7(Applause.)

8DR. CARLTON: Thank you.

9Our next speaker is Joe Simons.

10MR. SIMONS: Thanks, and good morning, everyone.

11I would like to start out by complimenting the

12FTC and the Department of Justice in holding these

13hearings and doing a terrific job. I am really quite

14encouraged that something really valuable will come out

15of this.

16So, one of the first things that happened this

17morning is the audience was instructed not to ask any

18questions or make any comments. So, I thought, well,

19gee, I was planning to hear you violate that restriction

20right away, but maybe we'll try something a little bit

21different.

22Perhaps by a show of hands, who in here would

23say that the 1982 Department of Justice Merger

24Guidelines market definition paradigm was the most

25significant development in market definition in the last

21

130 years?

2So, we have got most of the panelists and maybe

3half of the audience. That is pretty good for one

4thing. I would have expected it might have been a

5little bit higher.

6But in any case, what that showing would

7demonstrate is an enormous amount of success for that

8effort, I think by any standard, and why is that the

9case? Why were those guidelines on the market

10definition paradigm so successful?

11In my view, it is because those guidelines

12reflected an understanding that the tools of antitrust

13analysis should be designed for a specific purpose.

14Previously, you had market definition which was pulled

15out of the economics literature and it was not designed

16to do an antitrust analysis. The merger guidelines

17market definitions was done specifically for that

18purpose.

19The other thing that was really important is

20that the agency, the DOJ in that case, was willing to be

21out in front of the case law. I think there was a

22pretty good argument that those guidelines, the market

23definition therein, did not really reflect the case law.

24So, I thought it would be useful to do a little case

25study and talk about first principles and market

22

1definition.

2The Guidelines, the Merger Guidelines, were

3built around the goals defined right in the Guidelines

4of preventing mergers from creating or increasing market

5power -- initially through coordinated interaction and

6then later unilateral effects. And as I said, they

7geared this market definition specifically to this

8overall goal of the Merger Guidelines. So, it was

9designed to identify that universe of firms that were

10necessary to profitably engage in coordinated

11interaction or in unilateral effects. Then for the

12unilateral effects, arguably the analysis could collapse

13the market definition into the competitive effects

14analysis. The market definition in the Guidelines is

15rigorous, it is logical, and it is transparent.

16Now, sitting here today, 25 years later, and

17seeing what a success this was, you might forget what it

18was like when these things were first issued. There

19were hoots and howls from all sectors of the Antitrust

20Bar and the academic community. These guidelines were

21ivory tower nonsense; they were completely hypothetical;

22they were totally inoperable and just downright

23impractical; a complete waste of time. These were

24comments that people made very regularly, and some

25people even said it was a conspiracy to do away with the

23

1antitrust laws.

2There was a little bit of a kernel of truth to

3some of those complaints, not the conspiracy stuff, but

4to the practicality of this test. There were a lot of

5people who saw the initial attempts to implement this by

6the agencies in the following way. One of the staff

7lawyers would have a conversation with the customers and

8say, "Gee, do you think that the sellers in this market

9could profitably raise price 5 or 10 percent?" You are

10shaking your head, but I heard people do that, Greg, and

11the customer has no concept of what it takes for it to

12be profitable. There is no context to the question.

13So, there were reasonable criticisms.

14But what happened is that because the algorithm

15was rigorous and logical and transparent, it enabled the

16development of applications basically, tools, to

17implement this approach, econometric tools. Examples

18are Baker and Bresnahan and Scheffman and Spiller, Greg

19Werden as well, something near and dear to me, critical

20loss. These things did not exist when those guidelines

21were first issued, and that really was an important

22lesson to learn, that if you have the right structure,

23then you have created a platform on which you can build

24something that really works.

25So, what does this translate into in terms of

24

1what we should do for section 2? Well, what are the

2goals of section 2? What are we trying to accomplish?

3Is there a consensus? You know, there has been a lot of

4ink been spilled in relation to the Trinko case, for

5example. There are differences already between the way

6the DOJ and the FTC look at this. There is the profit

7sacrifice test, the no economic sense test; there is the

8disproportionate harm relative to efficiencies test.

9So, where does that leave us for market

10definition? Does that create a problem? Can we rely on

11what is in the case law? Reasonable interchangeability,

12what does that mean? How much interchangeability is

13reasonable? It is basically relying on

14cross-elasticities of demand. How high does the

15cross-elasticity have to be? Is that even something you

16can look at? Can we rely on the Merger Guidelines

17market definition? Does the hypothetical monopolist

18paradigm, as applied in the Merger Guidelines, really

19work for section 2? And one of the issues in section 2

20is, are we focused on the same phenomenon that we are

21for section 7?

22The Merger Guidelines, the Horizontal Merger

23Guidelines, are basically focused on collusion, an

24extreme form of which is unilateral behavior. So you

25are talking about situations in which a group of firms

25

1is trying to restrict their own output, whereas in

2section 2, what you are dealing with is a situation in

3which one firm, the large firm, the dominant firm, is

4trying to restrict the output of somebody else in most

5cases and maybe sometimes themselves as well. So, what

6do we do with all of that?

7One possible thing to do -- and I am just

8throwing this out -- would be to come up with a set of

9goals for section 2, what is the purpose, what are we

10trying to do, and then work through various scenarios as

11 to what the market definition would be under each of

12those. So, one potential scenario is, we are going to

13say that the goal of section 2 is to prevent unilateral

14conduct that is reasonably likely to significantly raise

15price or reduce quality. Reasonably significantly, you

16can come up with other adjectives, number one.

17Number two, and you are going to focus on

18conduct that either, A, has no efficiencies, B, has

19disproportionately low efficiencies relative to their

20exclusionary effect, or C, would make no economic sense

21in the absence of exclusionary effect, and potentially

22D, permits recoupment of the exclusionary conduct. So,

23kind of a menu from which to choose.

24Well, one could argue that the first condition,

25that the unilateral conduct be such that it is

26

1reasonably likely to significantly raise price and/or

2reduce quality, may be a necessary condition. That

3defines the universe in which something bad can happen.

4If you do not have that condition, then you might be

5able to say that nothing bad can really happen. So, you

6can use market definition in that sense, to focus on

7that aspect as a screen.

8You then could ask, "Well, gee, would the market

9definition need to change depending on your choice of 2A

10through D?" And at least at a first cut, I would say

11probably not, that these factors relate to what might be

12considered defenses or separate prongs of the analysis.

13They would not be necessary to worry about in the first

14market power screen, where you use market power or

15market definition as the screen.

16All right, so what would be the relevant

17context, then, for measuring profitability of a price

18increase? Well, obviously the options are before,

19during or after the execution of the alleged conduct.

20Well, we are concerned with the price going up as a

21result of this conduct, so it seems to me you want to

22focus on whether there might be a significant price

23increase, whether a significant price increase might be

24profitable during or after this alleged conduct.

25Then similarly, if the conduct is already in

27

1place, so you cannot observe it over time, then the

2question might be the reverse, which is, absent this

3conduct, would the price be lower, right?

4You see, I think there is the same problem here

5that you have -- not really a problem, but an issue in

6the Merger Guidelines -- where for the most part, you

7are measuring the profitability of a price increase

8going forward. You are not looking at the current

9level. You are really looking at a change in the

10current level that is brought about by the conduct that

11you are worried about. So, in the merger case, it is

12the merger; in this case, it would be the alleged

13exclusionary conduct.

14You know, one of the things that is near and

15dear to me, critical loss, might be a tool to help in

16this analysis, and it would not be exclusive by any

17means. Just like in the Merger Guidelines you can use

18critical loss, you can use all kinds of other estimation

19techniques, and they are not exclusive.

20So, one way to think about this would be that

21the burden would be on the plaintiff to show the likely

22extent to which the alleged conduct restrains

23third-party producers; in other words, whatever the

24conduct is, exclusive dealing, refusal to deal,

25whatever, what is the likely impact on third-party

28

1producers? How much restraint does this have on their

2ability to supply the market?

3Then the plaintiff would have to show that it

4would be profitable for the monopolist to raise price

5significantly -- whatever the number is, 5, 10 percent,

6whatever -- as a result of that exclusionary conduct.

7You could calculate a critical loss for the monopolist

8that would be based on margins, and you could estimate

9whether a 10 percent price increase after or during the

10alleged conduct would leave sufficient residual supply

11such that a monopolist would lose in excess of the

12critical loss. So, that would get you the market

13definition part of this. Then what do you do?

14One strategy would be to not even bother with

15shares, because you have basically concluded that the

16single firm was able to engage in this alleged conduct

17and get the price up, and in terms of that, one could

18say, "Well, that's what we needed to know," and we will

19now we go through the rest of the analysis and determine

20what are the efficiencies, and maybe you want to talk

21about recoupment as well. So, one could reasonably say,

22"Well, we don't really need a market share threshold."

23Other people could say, "Well, gee, it is in the case

24law. We want to try to make it consistent. It is

25really important. So, we need a market share

29

1threshold." How would that work?

2Well, one way to think about it in the context

3that I have just outlined would be you could say, "Well,

4the firms in the market would be obviously the alleged

5predator, and then potentially also other firms that

6have also benefitted from a price increase as a result

7of this exclusionary conduct," and you might base their

8share calculations on their sales of that product for

9which the price increase was experienced.

10But then you ask the question, "Well, why have a

11share requirement? What does that do for you?" You

12might say, "Well, it gives us some comfort because

13predatory conduct is only likely to occur where the

14shares are high." Well, there is an issue about that,

15because some exclusionary conduct is really cheap, and

16some exclusionary conduct is really expensive. So, if

17you are going to engage in really expensive exclusionary

18conduct, yes, then you probably want to have a big

19share, because you need to recover that expense that you

20laid out to execute the exclusionary conduct, but if you

21are executing really cheap exclusion involving, a

22Hatch-Waxman type of scenario or something like that,

23which costs virtually nothing, well, then, what does the

24market share do for you? So, that is unclear.

25I have got about 30 seconds left, and I just

30

1wanted to sum up by saying I think there are some really

2important lessons to be learned from the Horizontal

3Merger Guidelines market definition, and I am hopeful

4that what will come out of this is we will get a bunch

5of smart people in a room, maybe Greg and some of his

6colleagues from the Antitrust Division and the FTC will

7sit in a room, take all of this together, and come out

8with an algorithm that is of similar significance to

9what they did with the Merger Guidelines -- use the

10first principles integrated approach, not worry about

11the fact that what they might come out with is a

12theoretic framework, theoretic algorithm that is not

13immediately implementable, and then not be afraid to

14consider a market definition guideline that deviates

15from traditional case law, because what happened with

16the Merger Guidelines is people originally said, "Oh,

17this is nothing like the original case law," and now we

18have been able to bring the two together, and the courts

19have seemed to have adopted what is in the Guidelines.

20Thanks very much.

21(Applause.)

22DR. CARLTON: Thank you, Joe.

23Our next speaker is Larry White, who has arrived

24in time. You have already been introduced, Larry.

25DR. WHITE: Well, thank you.

31

1DR. CARLTON: And the ground rules are we are

2running a little late, so if you could keep to 15

3minutes.

4DR. WHITE: Right.

5DR. CARLTON: Does the computer work?

6UNIDENTIFIED SPEAKER: Yes.

7DR. CARLTON: And you are the first person who

8has the use of the computer.

9DR. WHITE: All right, great. Well, thank you.

10I am very pleased to be here this morning, and sorry for

11the delay of my arrival. I flew down from New York this

12morning, and every once in a while you get hit with a --

13I do not know whether it is the right-hand tale or

14left-hand tale on variance, but we were an hour late

15taking off. So, here I am. I am very pleased to be

16here.

17I think this is a terrifically important issue,

18and it is an issue where unfortunately too many mistakes

19have been made, too many mistakes continue to be made,

20and I want to walk you through what I consider to be

21some important issues. I have got a few call it partial

22answers. I do not have the complete answer. At the

23end, I am going to be echoing Joe Simons' call. We need

24a new paradigm; a paradigm is missing.

25So, like any good business school professor, I

32

1am going to tell you what I am going to say, and then I

2am going to say it, and then I am going to tell you what

3I said. I will frame the issue, I will remind you what

4the standard monopoly model looks like, I will remind

5you what the implications of that model are, I will

6point out the loose language that has been used by

7people who do know better or who ought to know better,

8and I'll tell you about the danger of that loose

9language. That will bring me to the Cellophane fallacy.

10Everybody is going to talk -- you cannot not talk about

11the Cellophane fallacy when we're addressing this topic,

12remind you of an ongoing dilemma, put out some partial

13suggestions, and wrap it up.

14What's the issue? I am not going to get into

15this market power versus monopoly power. The way I was

16taught, it is all the same thing, and the exercise of

17this thing, call it monopoly power or market power, is

18the seller can sell at prices above marginal cost and

19earn rents, and I should have added for a sustained

20period of time, but I will go ahead with my story. That

21is the picture that we carry around in our head of what

22monopoly power, market power, is about, the sustained

23charging of a price above marginal cost, maintaining --

24I am going to use that word over and over again --

25maintaining a price substantially above marginal cost.

33

1All right, now, what also gets talked about,

2especially in an antitrust context, is actions --

3exclusionary, predatory actions -- that can create or

4enhance market power. So, somebody who did not have it,

5can create it. Somebody who has it through an

6exclusionary or predatory action can enhance it, make

7the demand curve yet less elastic or inelastic and earn

8even higher rents.

9If the seller is engaging in this kind of

10activity, whether he is exercising the market power or

11enhancing, a likely precondition is that the seller has

12a large share of its market. So, that is not necessary.

13You can come up with examples where if the overall

14supply is limited, where other suppliers cannot expand

15their output very much, where demand is quite inelastic,

16even somebody with a relatively small share of a

17commodity market by his unilateral actions can affect

18the price, but more generally, a large share of

19something called a market is going to be necessary. But

20that then raises this threshold or safe harbor issue,

21what is the market, and there is no standard paradigm

22for that determination.

23So, this is the picture we carry around in our

24head, and the implications of that picture, the

25monopolist maintains its price at a level above the

34

1competitive price. He would not want to raise his price

2any further unless demand changed or costs changed. He

3is already where he wants to be. In trying to raise his

4price, he would lose too many customers to sellers of

5something else, and, of course, if the market changes

6from a competitive structure to a monopoly -- because of

7cartelization, because of exclusion -- then the price

8changes, then the price increases, the seller, newly

9feeling this market power, raises the price from the

10competitive to the noncompetitive monopoly level, but as

11a characterization of what is going on when we take a

12snapshot of the market, he is maintaining the price at a

13level above the competitive level. That is clear in

14this standard model.

15About 40 years ago, George Stigler developed an

16expanded version of this, the dominant firm and the

17inverted price umbrella, where he described a firm that

18was not strictly a monopolist, he faced a reactive

19fringe of smaller firms that were limited in their

20supply response, and he showed basically you get a

21similar type of outcome. The dominant firm is able to

22charge, maintain a price above competitive levels, but

23he doesn't want to go any higher because -- and there,

24in the Stigler model, it is implicit -- he would lose

25too many sales to that competitive fringe.

35

1Okay, why am I making such a big deal out of

2this? Because there has been loose language out there,

3first by my colleagues, all of whom do know better, and

4they describe the phenomenon of monopoly power, market

5power, in terms of the ability of the firm to raise

6prices. In other words, I have put in italics over and

7over again, this language of "raise prices," or in the

8context of the Microsoft case, Fisher and Rubinfeld

9making this claim that, "Gee, Microsoft could have

10raised its price substantially and wouldn't have lost

11customers," and you have got to scratch your head, how

12come they didn't? Then Evans and Schmalensee on the

13other side, again, talking the language of "raise."

14Even earlier, as I walked in the door, I heard

15Phil Nelson talking about the monopolist "raising" the

16price. Maintaining is what we're talking about, but I

17am sure I in my looser moments fall into this "raising."

18It is an easy thing to do, but I am going to show you

19the dangers of it in just a minute.

20I'll go over to some noted legal cases and legal

21opinions, and again, you have got the same -- oh, did

22I -- no, I forgot to put the italics in there, but you

23can see the word "raise" in each of those -- in each of

24those quotations from those cases.

25All right, what is the danger? The danger in

36

1the "raise" terminology is that if we think market power

2and monopoly power are the ability to raise the price,

3then it is easy to then think, "Ah, well, the test of

4whether somebody has market power or not is whether the

5seller can raise prices above currently observed

6levels." Remember, that is what Fisher and Rubinfeld

7were talking about there.

8Conversely, if the seller is constrained from

9raising prices because of its fears of losing too many

10customers, then does that imply that it does not have

11market power? The trouble is, even in the standard

12paradigm where the monopolist is maintaining a price

13above competitive levels, it cannot profitably raise its

14price because it would lose too many customers to

15sellers of something else.

16That, of course, then leads us to the Cellophane

17fallacy, the U.S. v. Dupont case, where the issue was,

18was the market a narrow market of cellophane, in which

19case it is clear, Dupont had market power. There was

20one other seller of cellophane, Sylvania. It was under

21license from Dupont, and so, effectively, no question.

22If the market was cellophane, Dupont had market power.

23Or was it, as Dupont claimed, flexible wrapping

24materials, in which case Dupont only had a 17.9 percent

25share and didn't have market power?

37

1The Supreme Court majority said it was

2interchangeability that carried the day, that cellophane

3was interchangeable with other materials mentioned --

4there was wax paper and brown wrapping paper and

5aluminum foil and glassine and lots of other things --

6and the majority said, "Ah, look, it is interchangeable.

7Dupont can't raise its price. So, it must be part of

8that larger market."

9The minority pointed out the fallacy of that

10reasoning and also pointed out the comparison with

11rayon, where Dupont also faced 15 to 18 other producers,

12also had a market share that was below 20 percent, and

13made much less profits. They also pointed out that

14Dupont's price of cellophane did not move around when

15those other flexible materials' prices changed.

16So, we have this ongoing dilemma. Profit data

17nowadays are relied on a whole lot less than was the

18case back in the fifties when Stocking and Mueller were

19writing, when the Supreme Court minority relied on those

20profit data. The Horizontal Merger Guidelines cannot be

21used, because they are a forward look, as you have heard

22already, they are a forward-looking test.

23The one exception, which Greg Werden has pointed

24out, is that if we are talking about a practice that is

25not yet in place, say an exclusive dealing plan that is

38

1going to be put in place. A plaintiff comes in, asks

2for an injunction. We are talking about something where

3it is a prospective practice. Then the prospective,

4forward-looking paradigm of the Merger Guidelines will

5work. To the extent that that is what we are looking

6at, fine, we have got an answer, but lots of instances

7are not of that kind.

8As Phil remarked earlier, elasticities do not

9help us very much. You cannot tell the difference

10between a true monopolist and just a different -- a

11seller of a differentiated product, a Chamberlin/

12Robinson monopolistic competitor.

13Okay, what to do? Well, sometimes a complaint

14will involve a prospective practice, and then we have

15got the Merger Guidelines. Sometimes there will be

16cross-sectional or time-series evidence involving prices

17where we can tell that concentration matters, and when

18concentration matters, you have got a market, and retail

19services are an area where cross-sectional data may be

20available.

21I harken back now ten years to the Staples case,

22where cross-section data showed that prices were

23different, higher where only Staples or Office Depot was

24present in the market, lower when both were there, yet

25lower when they and a third office superstore were

39

1there. That evidence carried the day, and I think

2correctly, that there was a problem -- there would be a

3problem if the two firms merged, and it told us office

4superstores were a market.

5Think of the American Airlines predatory

6behavior case. Why do we think that city pairs are a

7market, city pairs airline transportation? Because

8there is lots of cross-sectional evidence that shows

9that, controlling for other things, prices matter and

10prices are related to concentration. Sometimes profit

11data will be useful.

12I mean, if you think the Microsoft case was a

13good case, if you thought that Microsoft's behavior was

14a problem, why did you think that? And I think at least

15part of the story was those profits. They were so large

16that even with all the problems that we know about

17profits, they were telling us something. But what if

18none of these possibilities are available?

19Well, Phil Nelson and I a few years ago made a

20proposal. It turns out similar language can be found in

21a 20-year-old article by Tom Krattenmaker. Greg had a

22version of this proposal in an article he wrote in 2000,

23where basically it is asking in the presence of an

24allegation of exclusion, what would have been the

25consequences of the absence of exclusion? It requires a

40

1two-step investigation.

2First you have got to ask, in the absence of

3exclusion, what would the plaintiff's sales have been?

4And then you have got to ask, what would the price

5consequences of those additional sales have been as

6well?

7Now, as was indicated earlier, this would focus

8directly on effect, and it implicitly delineates a

9market, but if you think about what the unilateral

10effects analysis under the Horizontal Merger Guidelines

11does, it is basically doing the same thing. It is

12looking for an effect, and then, if somebody goes ahead

13and then tries to delineate a market, that is sort of

14redundant. You have already found the effect.

15Implicitly, you have said there must be a market there,

16and that is basically what the Nelson and White proposal

17does as well.

18But I think the best approach would be let's try

19to develop -- you know, I have thought hard about it.

20The best I could come up with was this joint proposal

21with Phil. It may not be good enough. Can the world

22come up -- can the Division, can the FTC, can a bunch of

23smart people out there -- come up with a paradigm that

24will have the power and eventual universality of the

25Horizontal Merger Guidelines?

41

1I urge you, remember what the world looked like

2before 1982. Remember what 1981, 1980 and 1979 looked

3like. We did not have a paradigm. We had

4Elzinga-Hogarty. We had Ira Horowitz's suggestion.

5There were other ideas out there. George Hay was going

6around talking about how the Division defined markets,

7and he would say, "Well, we would look for whether there

8was a specialized trade journal that the sellers in a

9marketplace all submitted their data to." Those were

10the kinds of indicia that people looked to. The Merger

11Guidelines brushed all that stuff away, and we have now

12got a powerful paradigm. I hope that some smart people

13out there somewhere will be able to develop something

14with similar power.

15So, winding up, we have got an unsatisfactory

16state for market definition. I would hope we are in

171981, and next year, somebody is going to come up with

18something that will have the same kind of power as the

19Horizontal Merger Guidelines. I have shown you some

20partial remedies, but the best remedy would be a new

21paradigm.

22Thank you very much. I am very pleased to have

23this opportunity today.

24(Applause.)

25DR. CARLTON: Okay, thank you, Larry.

42

1Our next speaker is Andy Gavil.

2DR. GAVIL: Good morning, everyone. Thank you

3to the organizers for inviting me to join everyone

4today. I am delighted to be here and agree with

5everyone else that these are some very important --

6indeed, fundamental -- issues to how we go about

7analyzing antitrust cases, and in truth, they are not at

8all unique to section 2. Questions of power and effects

9really cut across all kinds of cases today. So,

10resolving one area clearly is going to influence and

11affect the others just as the Merger Guidelines has

12affected many areas.

13 So, I start with my first slide in talking about

14it is all about anticompetitive effects, and I think I

15would add to that, and legal process. At the end of the

16day -- that is a great phrase, "At the end of the

17day" -- "At the end of the day, in the final

18analysis" -- but at the end of the day, in the final

19analysis, whatever we conclude as a matter of economics

20is the right approach, we have to translate that into a

21legal system of decision-making. It has to work in

22courts. It has to work in a context where we have

23burdens of pleading and burdens of production and

24burdens of proof. It has to work in a context where we

25have various methods for discovery of evidence, where we

43

1have a role for expert witnesses, where we have judges

2and juries, and if it cannot work in that context, then

3perhaps there is a problem with what we have come up

4with as a theoretical matter.

5I forget who it was, I think it was Joe talking

6earlier about how the Merger Guidelines were originally

7received. Well, part of the problem in how they were

8received is that they were received by a legal community

9accustomed to looking at cases in one particular way.

10They suggested that we needed to look at those cases in

11a very different way, and it was very unclear in 1982

12how you would translate, how you would take something

13like SSNIP and what evidence would you need?

14The lawyers that were asking the questions of,

15what witness am I going to need to do this? What

16evidence will I need from my client, from the other

17parties? How will I assemble it? How will I present

18it? There can be no doubt at all I think in anybody's

19mind that the Merger Guidelines and subsequent

20developments have been an economist's full employment

21act, and certainly that has been evidenced in the

22antitrust area. It is hard to imagine today proving any

23kind of case, plaintiff or defense, without the role of

24economists, and that is a result of the writing into our

25substantive standards various economic ideas.

44

1So, as I go through these slides, I want you to

2sort of keep that in mind. The focus I have tried to

3bring to my comments today is, how do we make it work in

4this legal system? Well, common issues in antitrust are

5effects, and we have certain ways that we go about

6establishing them. We have irrebuttable presumptions --

7that is what the per se rule is all about -- and we have

8rebuttable presumptions; whether we are using direct

9evidence or circumstantial evidence -- and that is going

10to be an important issue that I am going to look at

11today -- we have different ways that we go about trying

12to establish effects.

13Direct evidence, defined here, is the actual

14exercise of market power. It may come out in

15performance evidence. It may come out in before and

16after studies of price. It is reflected to some degree

17in our use of "quick look." The "inherently suspect"

18formulation is also a way of looking at things that are

19obvious, and a question I will be asking today is, do we

20have equivalents for section 2 and would it make sense

21to use them in section 2?

22On the circumstantial evidence side, we have

23something that I have called a "double inference." We

24define a market, we calculate market shares from a

25certain level of market share, we infer market power,

45

1and in truth, from that, we then infer the capacity for

2anticompetitive effect. In litigating terms, we are

3dealing with two very standard paradigms of how to go

4about proving something.

5Well, power, of course, is a condition precedent

6of effects, but if you look in the cases, there is a lot

7of confusion -- again, loose language -- about how it is

8used. Some cases say, "Well, what we need is market

9power," and even in cases like NCAA and Indiana

10Federation of Dentists that really were out in the

11forefront in this quick look idea and the use of direct

12evidence of actual effects, there is confusing language

13about what "market power" means.

14Well, power is the condition precedent of

15effects. If you have the effects, the power is there.

16So, part of the point of Indiana Federation, talking

17about market definition and market power as surrogates,

18was to make the point that when you have the actual

19effects evidence, going sort of back around the

20circumstantial evidence route, trying to define a market

21and determine whether there are large market shares, may

22be beside the point. Those things are surrogates for

23direct evidence.

24Well, as in many areas of antitrust, that leads

25us to a point where we can identify easy cases and hard

46

1cases. A good example I think of the easy cases, when

2the direct and circumstantial evidence are aligned, when

3they are pointing in the same direction, when you have

4evidence of actual effects and you have high market

5shares, those are easy cases. We do not argue about

6those very much. The D.C. Circuit in Microsoft actually

7structured its discussion of monopoly power that way,

8looking at both direct evidence, circumstantial

9evidence, they are both pointing in the same direction,

10easy case.

11On the other hand, for safe harbor ideas, if you

12have de minimus evidence and no effects and you have low

13market shares, again, pointing in the same direction,

14and I would make this point -- I'll raise it a little

15bit later -- in terms of safe harbors, I do not think

16you can rely just on market shares alone. It has to be

17market shares plus certain other factors, and I will

18also suggest that if we are going to have safe harbors,

19we need some danger zones, and again, it might be market

20share plus some other characteristics.

21But evidence and power effects are interrelated,

22and I think this is what makes part of our current

23framework very difficult to think about. Courts do

24think, because of years and years of case law, first

25monopoly power, then willful acquisition or maintenance,

47

1when in truth, the evidence of conduct and effects in

2the evidence of power is going to be very interrelated.

3Well, again, thinking about direct and

4circumstantial evidence, the benchmark for

5circumstantial evidence is clearly the Horizontal Merger

6Guidelines. They really did advance the science of

7thinking in terms of circumstantial evidence. Recall,

8though, that Cellophane was a section 2 case, and maybe

9there are some different problems that come up when we

10are doing prospective predictions about likely market

11power versus retrospective methods when we have, you

12know, the before and after ability to actually look at

13the effect of conducts, but the Merger Guidelines in any

14paradigm we come up with are probably going to have some

15continuing significance. They have been cited by courts

16outside of section 7. They are cited in section 1 cases

17and section 2 cases. Basic ideas and concepts are

18clearly interrelated.

19So, my suggestion at this stage of our

20development is we need something of a similar to the

21Merger Guidelines to refine "actual exercise" standards

22and to harmonize those standards across different

23offenses. A critical question, I think, is how much and

24what kinds of effects evidence should be sufficient to

25shift a burden? And here I remind, again, that outside

48

1the area of exercising prosecutorial discretion, outside

2the walls of the agencies when they are deciding whether

3to bring a case, if the decision to bring a case is made

4and the economists agree, the next question the lawyers

5are going to have is, "Well, how do we meet our burden

6of production? What evidence are we going to assemble?

7What is going to make us win this case?"

8I think when you are thinking about direct

9effects evidence, and market share as well, a critical

10question in section 2 is, what does it take to shift a

11burden? Frequently what you see defendants arguing in

12cases is the burden didn't shift, the burden didn't

13shift, the burden didn't shift. Well, what does that

14mean?

15It means that there is no requirement on the

16part of the defendants to actually justify their

17conduct. If they claim there are efficiencies, where is

18the evidence of efficiencies? That does not happen

19until the burden shift takes place. That is a critical

20stage. It is a critical stage that has to be focused

21on, and I have given some examples here of various cases

22that raise some of those questions.

23I think we are also feeling the weight of the

24Alcoa paradigm. In looking back at Alcoa and the cases

25that preceded it, all Judge Hand did was he surveyed the

49

1previous cases and looked at winners and losers to come

2up with his three famous sort of -- you know, 33, not

3enough; 66, maybe; over 90, definitely. Well, where did

4he get that from?

5If you look at the prior Supreme Court cases,

6you will see that there were cases falling into each of

7those categories. He sort of synthesized them and came

8up with this benchmark. I think an important question

9for us is, are we ready to move beyond the total

10reliance on market shares, which sends us off in this

11direction of conflicting evidence, plaintiffs and

12defendants having experts -- the market is big, the

13market is small -- and is that really where we want to

14be? What can the role of direct evidence be?

15The Re/Max case was an example of a court

16relying on direct evidence, actual price effects

17evidence in a section 2 case. The 7th Circuit in

18Republic Tobacco rejected such an approach, said that

19Indiana Federation did not apply and NCAA did not apply

20to a vertical case. Is Staples -- and the unilateral

21effects that people have alluded to already -- is it

22related? I think it is. It is a way of trying to more

23directly gauge. We have talked about the monopoly

24versus market power being kind of a silly distinction.

25So, I will move on.

50

1I think there is an important role here for

2decision theory, which obviously has begun to influence

3our thinking. The emphasis tends to be on fear of error

4costs, and often that motivates calls for more and

5better evidence. We need more before that burden

6shifts. One point I would like to walk away with today

7is urging that we also consider the second half of

8decision theory, which is process and information costs.

9Is more evidence really always better?

10In that regard, I sort of suggest -- and it is

11not really new, there is a lot of general literature out

12there on the economics of evidence. Richard Posner has

13a long article on an economic analysis of evidence, and

14I put forward the question, "When does the marginal

15value of additional evidence in terms of economic

16certainty (minimizing error costs) outweigh the costs of

17obtaining and processing that evidence, taking into

18account whether it is reasonably accessible to the party

19bearing the risk of non-persuasion?" What I tried to do

20in that question is integrate some economic ideas and

21some legal process ideas from both the rules of

22procedure and the rules of evidence. It is always easy

23to demand more. It is always easy to pursue some kind

24of level of absolute certainty and minimal error costs.

25The question is, as a legal standard, when we take that

51

1into court, is that really going to strike the right

2balance for us in resolving cases?

3Antitrust is not always rocket science, and I

4think we need to get over the idea that it always is.

5Yes, we need safe harbors to guard against false

6positives. I think we also should be emphasizing

7equally defining danger zones where we might be running

8into false negatives.

9Is monopoly power all that puzzling? I would

10point out to everyone that neither 3M nor U.S. Tobacco,

11in two U.S. Courts of Appeals monopolization cases, even

12contested that they had monopoly power. In the

13Microsoft case, they contested it, but rather

14unpersuasively, and every agency and every court to look

15at it has concluded that yes, indeed, Microsoft had

16monopoly power.

17We could go on with a couple other examples,

18American Airlines, Dentsply. Were these really such

19difficult cases? If they were not, then why were they

20so difficult? Why would parties not even litigate the

21point about their power? There must be, there must be

22cases where -- again, market share plus -- where there

23must be additional factors, information on entry

24barriers. Entry barriers will always, for example, be

25important.

52

1Finally on this slide, sliding scales, not all

2burden shifts are created equally. You see this in

3cases like Baker Hughes and Heinz in the merger area,

4the realization that sometimes when a burden shifts, it

5really shifts, and the presumption is very strong, and

6other times, it kind of is just enough to shift. Well,

7in responding to those sorts of cases, we might want to

8respond in different ways by considering what is

9required to shift and what is required to shift back a

10burden in different ways.

11On legal standards and decision-making, I think

12that the balancing of effects idea is a straw man. We

13could cite, as Larry White did, we could put up lots of

14slides with courts saying, "Anticompetitive effects; the

15burden shifts. Efficiencies; then we balance one

16against the other." We do not really do that. I have

17looked; you can all look. If you can find me a Section

181 litigated case in which the case was actually decided

19on balancing effects versus efficiency effects, consumer

20surplus diminution versus increased producer surplus,

21find me such a case. I would like to see it. It is not

22what we do.

23What we do is weigh evidence. What juries do is

24they compare the evidence of anticompetitive effects

25with the evidence of efficiencies, and they make a

53

1decision about where the weight of the evidence is.

2That has to do with credibility; it has to do with

3persuasiveness. It does not have to do with $10 of

4anticompetitive effect and $11 of efficiency.

5Finally, a word about caricatures and corrosion

6of the rule of law. The level of discourse and the

7level of criticism of antitrust, as we all know, has

8continued for quite some time. It has continued despite

9the fact that in the last 40 years, we have seen some

10pretty major corrections to antitrust.

11I say caricature -- and this is not my

12caricature -- but this is what you see in a lot of the

13criticisms of antitrust, and I think it is a caricature

14that ignores this last period of adjustment over the

15last 30 years. Incompetence -- judges, just

16incompetent. They can do habeus corpus, they can do

17environmental, they can do securities law, but antitrust

18is rocket science, keep them away.

19The same thing with juries. They just do not

20know the difference between somebody who is full of it

21and somebody who really knows what they are doing. They

22cannot tell the difference between economists in this

23case and, of course, neither can they decide any other

24possible case.

25And, of course, enforcers. I have the asterisk

54

1there just to remind me to say that. Typically it is

2enforcers themselves who make this argument, God, we are

3so stupid. You shouldn't really trust us to make any

4decisions, and although we may -- and it gets very

5personal -- we may be able to make the decision, but

6other enforcers are really stupid, especially those guys

7at the offices of the states.

8Who are the untrustworthy self-interesteds, the

9self-interesteds who are untrustworthy? Rivals, oh,

10they are always full of it. They are always complaining

11about more competition. Dealers, yeah, what's that

12freedom of dealer stuff? You know, manufacturers,

13consumers, aligned; dealers, out in left field. And

14plaintiffs pretty much all are full of it, especially

15class action reps.

16Ah, but who can we trust? Dominant firms.

17Dominant firms articulating efficiencies. Fear of error

18cost? That's truth. We need to put a lot of weight in

19that. We need to be concerned about it. Other

20defendants, generally yeah, and especially efficiencies.

21Two problems I have with this sort of

22caricaturing of antitrust. One is, I don't think it is

23true. I would like to see the list of false positives

24in the last 25 years. We have been moving towards

25reduced error costs, and here I think it would be

55

1helpful to have the economists really define what they

2mean as "false positive." It is not a case on which

3reasonable people can differ. It is a case that sort

4of -- again, borrowing from procedure -- no reasonable

5party could have come out that way. To me, that would

6be a false positive or a false negative. It is not a

7case that we simply disagree about.

8LePage's has, you know, been frequently used as

9sort of this example of a false positive. Be reminded

10that 3M did not contest its market power, and if it did

11offer any evidence of efficiencies, nobody who looked at

12it found it very convincing. Did the Court of Appeals

13give us a useful standard for bundled pricing? No, but

14neither has anybody else yet. So, to call that a false

15positive and say, "This is an example of how we're going

16to inhibit all kinds of other cases," I am not sure that

17that is justified.

18The final point and I will sit down. As I said

19at the start, Larry said at the end, you say what you

20said at the beginning. At the end of the day, these

21cases have to go to court sometimes, and this kind of

22rhetoric of criticism ultimately is corrosive of the

23rule of law. I think it is heard in curious ways

24outside the United States. These criticisms really go

25to the heart of whether we are willing, at the end of

56

1the day, to rely on courts to make decisions.

2We have numerous procedural devices, summary

3judgment, judgment as a matter of law, Daubert

4standards, appeal rights. If after all of that has

5occurred a plaintiff actually wins a case, which does

6not happen very often, I think we ought to be a little

7bit more cautious about tossing the rhetoric around

8about the incompetence and the untrustworthy

9self-interesteds, all right?

10Thanks very much.

11(Applause.)

12DR. CARLTON: Thank you very much, Andy. I was

13pleased to hear I am not as incompetent as once

14enforcers were thought to be, and to prove that I am

15still competent, we are going to have a break, and it

16will be a 10-minute break, and we will reconvene

17promptly so that we can try and stay roughly on

18schedule. Thank you.

19(A brief recess was taken.)

20DR. CARLTON: Why don't we try and start. Our

21next speaker is Rich Gilbert.

22DR. GILBERT: I would like to thank the

23organizers for the opportunity to be here. I was

24invited to talk about technology markets, so if any ink

25gets spilled on this issue as a result of my comments,

57

1you can be sure it will be Independent Ink, though I

2will not talk about the presumption of market power for

3patents. I thought we resolved that issue in the IP

4Guidelines, although it is not the case that the Supreme

5Court immediately adopts everything that the agencies

6come up with.

7Now, when we talk about market definition, there

8is a real sense in which we are talking about either

9guide posts or lamp posts. Now, lamp posts, as you

10know, shed light on a subject but do not necessarily

11shed truth about the subject. A lamp post might

12illuminate the ground, but that does not mean that the

13dollar that we are looking for is around the lamp post,

14even though if it were, perhaps we could see it.

15Guide posts, on the other hand, serve to focus

16the analysis. The guide posts lead the way. The way

17may be very foggy and very complicated and very

18difficult, but can be very useful.

19Now, my take, sort of in the spirit of Andy's

20comments, the courts and defendants like the market

21definition exercise, even though it is often used much

22more as a lamp post than a guide post. They like the

23exercise because, of course, for a defendant, if you can

24show the market is very broad, chances are there is no

25antitrust case there. For a court, they are all very

58

1busy. They have full dockets. If you can show the

2market is very broad, they do not have to worry about

3it.

4Plaintiffs also seem to like market definition

5or many of them like market definition, because if you

6can prove that or demonstrate or make a convincing case

7that the market is narrow, well, chances are then there

8will be an issue, but as I think everybody on this panel

9is implying, none of those conditions, whether it is

10broad or narrow, presumptive of a case or not

11presumptive of a case, none of them are really relevant

12directly to the analysis. We would rather have market

13definition serve as the guide post to lead the way to

14the right analysis rather than defining whether there is

15or is not a case.

16Now, so, if we talk about markets for

17technology -- first I should distinguish, I am going to

18focus more on technology markets than on markets for

19technology. Markets for technology can be analyzed

20using conventional goods markets, often using

21conventional goods markets, which are sufficient for

22analysis in many high technology industries, whereas

23technology markets are useful when what is at issue is a

24right or rights to a technology that are licensed rather

25than embodied in a patent. So, I am focusing more on

59

1technology markets than markets for technology, although

2maybe in discussion, we will get to that distinction,

3whether there should be a distinction.

4Technology markets are defined in the IP

5Guidelines. Technology markets consist of the

6intellectual property that is licensed that are close

7substitutes. Of course, here now, as in all market

8definition exercises, the issue is, what are the close

9substitutes? And when you are talking about technology

10markets, the close substitutes are not only other

11intellectual property rights, but also goods and

12services that may substitute for those intellectual

13property rights.

14It adds another layer of difficulty and

15complexity to the analysis, because just like in

16conventional -- other section 2 goods market definition,

17exactly what to sweep into that analysis and how, it

18depends upon the prices, prevailing prices, and whether

19the conduct is prospective or retrospective, these are

20all challenging issues, which I am not going to entirely

21resolve.

22Now, technology markets also are -- there is an

23upstream analysis for inputs which I think raises some

24interesting questions by itself. Technology markets

25have been used I think with some success to analyze the

60

1licensing of technology to manufacture float glass, for

2blending clean gasoline in the UNOCAL case, the float

3glass with the Pilkington case, for designing fast

4computer memory chips, as in the DRAM cases, perform

5laser eye surgery, or to incorporate genetically

6modified traits into agricultural seeds. These are all

7some examples, I think, of markets that have been

8analyzed using basically an upstream analysis for

9licensed inputs.

10Now, on the geographic market side, this is an

11area where using technology markets in some cases

12simplifies things. It is fair, I believe, to presume in

13many cases -- not all, of course -- the geographic scope

14for technology markets is very wide, because it is not

15very difficult for a potential licensee to negotiate

16with even quite distant licensors unless there are legal

17or regulatory or some other restrictions that prevent

18the use of licensed technology in different locations,

19as there was, for example, with the UNOCAL case, but in

20these other cases, the enforcement agencies I think have

21correctly concluded that the technology markets are very

22broad, U.S.-wide and sometimes worldwide.

23Now, technology fees, should these be indicators

24of market power? Interesting question which has not

25been quite directly addressed. Marginal cost of

61

1licensing is typically very low. It suggests that there

2is market power if we define market power as the ability

3to sustain prices above marginal cost, then looking at

4technology fees, gives you an immediate read on whether

5or not there is market power, not necessarily monopoly

6power, but, as economists have said, that is a difficult

7line to draw between market power and monopoly power.

8Now, again, the relevant question is the ability

9to increase or maintain technology fees significantly

10above marginal cost for an extended time. In this

11dispute about market power versus monopoly power, I am

12certainly in the camp that says that monopoly power is a

13lot of market power and that there is no clear dividing

14line between the two, and the question is, the relevant

15question is, is there conduct that leads to either

16increasing or maintaining technology fees significantly

17above marginal cost for an extended period of time and

18whether it is prospective or retrospective?

19If it is prospective, perhaps the ability is to

20increase technology fees. If it is retrospective, then

21the question is more has conduct contributed to the

22ability to maintain technology fees significantly above

23marginal cost?

24This is now more in the spirit of what Larry

25White was saying in his approach to section 2 market

62

1definition. Also for technology fees, a related and

2relevant question in a section 2 context is whether

3competition, whether injecting more competition, would

4result in a significantly lower technology fee if the

5competition were not excluded.

6I also agree that this opens up a lot of

7interesting and unresolved issues, as in how much

8competition should be enough to consider? What should

9the price effect of that competition be in order to

10define a relevant market? Is an elasticity of demand

11faced at the existing prices for the fees and other

12goods and services? Is an elasticity of demand minus

13two, is that low enough, small enough in magnitude, or

14does it have to be minus 1.1 or 1.5 or is minus 3

15enough?

16These are very important and serious questions

17that need to be addressed if we are going to do this

18kind of hypothetical decrease in price through a

19hypothetical increase in output as a way to identify a

20relevant market.

21So, the focus on that additional competition and

22whether it lowers the fee I do believe can get around

23the Cellophane fallacy, and I think another important

24aspect of that approach is that it focuses the analysis,

25the definition of the market, on the analysis of the

63

1competitive effects of the conduct. So, I think

2sometimes it is a criticism of the hypothetical decrease

3in price approach that it is too related to the conduct

4that is being alleged as anticompetitive.

5I turn it around and say that no, I think it is

6an advantage of this approach, because it connects the

7conduct at issue to the analysis of the impacts at

8issue. Too often, I think many of us would agree that

9the market definition exercise puts the cart in front of

10the horse. We should be thinking about where are the

11competitive effects, how significant can the competitive

12effects be, and then let the market definition respond

13to that rather than defining where the competitive

14effects are. Again, this stems from the problem of the

15Cellophane fallacy that a profit-maximizing firm has no

16incentive to raise or lower its technology fee.

17Another question about analysis of inputs, in

18principle, the antitrust analysis for a technology input

19is not qualitatively different from the analysis of any

20other upstream good or service. The demand for the

21input is derived from the demand for the final good or

22service, and one thing to point out is the

23Hicks-Marshall Law of derived demand, which says that

24the elasticity of Derived Demand is proportional to the

25cost share of the input. It is roughly the cost share

64

1of the input times the elasticity of demand for the

2output. That will generally lead to a conclusion that

3the elasticity of demand is pretty small in magnitude.

4Indeed, in the Microsoft case, Microsoft made

5the argument that if you do this calculation, the

6profit-maximizing price for Windows was I think, like,

7$1,500 or something like that, and therefore, we could

8not have market power because we are not charging

9$1,500. I think it was an argument that was never

10really entirely responded to, but one does find that as

11you go upstream, you are going to generally get less

12elastic demands, derived demands; therefore, more

13potential to raise prices; therefore, more possibility

14of competitive effects.

15Of course, while it implies relatively inelastic

16demand for inputs and the ability to affect the input

17price, the input price has only an indirect effect on

18the final consumer prices, which is why the elasticity

19of demand is low. So, it turns around and gets you the

20other way. So, upstream analysis can overstate the

21ability to affect consumer prices.

22As you move downstream, though, the question is,

23how far downstream do you go? If you go far enough

24downstream, almost everything competes with everything

25else. If you move all the way downstream, eventually

65

1you are competing for the consumer's entire budget

2allocation, and whether you are talking about movies or

3sports or buying a car or whatever, everything competes,

4and the overall elasticity of demand for all goods and

5services is one, because you only have so much income.

6So, it is my view -- my strong view, but it is a

7view -- that analysis should take place where the firm

8has the ability and incentive to raise or maintain a

9price paid for an input or a final good, and the

10question should be, is the conduct the type of conduct

11that we want to prevent? And if it is, let's analyze it

12where the conduct might have an effect and let the

13market definition follow from where the conduct could

14have an impact.

15I just have a very quick example to end with of

16genetically modified seeds, which express a desired

17characteristic, like insect resistance in corn or

18tolerance of some herbicide. Do conventional seeds

19compete with licenses for seed traits? So, that gets

20back to the IP Guidelines definition, where do the goods

21come in to compete with the traits? It is a complicated

22question, not one I am here to answer, but I would just

23point out that these agricultural markets are moving

24increasingly to genetically modified traits, which is

25now way above 80 percent in soybeans and up above 50

66

1percent in corn, and if you are looking at questions

2about whether conduct is maintaining high prices for

3these characteristics and ultimately higher corn prices,

4it is my view that you should look at the trait markets

5for where this conduct is expressed, because that is

6where the effect could be.

7It may be that the conduct is not the type of

8conduct that should be subject to an antitrust sanction,

9but that is the right place to look. It goes back to

10the lamp post and the guide post. Let's look where

11there could be effects. Let's let the market definition

12exercise follow from the inquiry into competitive

13effects. Let's not use market definition as a lamp post

14to illuminate a problem that may or may not exist.

15Thank you.

16(Applause.)

17DR. CARLTON: Okay, thanks, Rich.

18Our last speaker is Michael Katz.

19DR. KATZ: I would like to thank the organizers

20for inviting me here, but I do not have time.

21I want to talk about -- it is a bit of a grab

22bag, but I will start about something systematic, which

23addresses the question of why delineate relevant markets

24in a section 2 case, and what I want to start with is

25really the first principle, what is the point of all

67

1this, and we really try to answer this question of a

2given practice harms competition and consumers, and what

3I want to talk about for a few minutes is how that gets

4us to talking about relevant markets, and I am going to

5talk about at least three things that relevant markets

6might be doing to help us answer that question.

7Okay, so the first one is you can think of --

8what you are trying to do is you are defining relevant

9markets so you can calculate market shares and

10concentrations, and we are doing that because we want to

11know whether the defendant or the firm under

12investigation, whether the defendant currently has

13monopoly power. Now, as everyone has been talking

14about, this is where the hypothetical monopolist test

15breaks down, so there is an issue there.

16It seems to me where we have gotten, actually,

17in a bunch of the recent cases -- and maybe this also

18goes to Andy Gavil's point about somebody showing me a

19false positive -- but I think if you look at Dentsply

20and Microsoft, there was plenty of expert testimony, but

21in the end it just came down to hard core pornography,

22the thing is you know it when you see it. People have a

23good idea that false teeth are a product and they are

24not really worried about a lot of other substitutes. I

25mean, there is sewing your lips shut and things like

68

1that, but I think in both of those, that was not really

2the issue.

3Now, I want to make a couple of other points

4about concentration as an indicator of market power

5here. One, if we are going to look at market shares, I

6think we really ought to ask ourselves, where did the

7market shares come from? Because it matters. Think

8about it. In some cases it is because of product

9differentiation, and some producers have much more

10successful products that match up with consumer tastes.

11There can be very different managements of different

12producers have different strategies, and one of the

13firms decided to have a high-volume, low-price strategy.

14I think the conclusions one would typically want

15to draw about the implications of them for competition

16are very different, and so I think it is important to

17try to have such a theory, and I think it is often

18lacking in antitrust cases. People just talk about the

19shares but not what they really mean or where they came

20from.

21The other thing is we want to ask ourselves why

22we care whether the defendant currently has monopoly

23power, and I will say at least I think one reason is you

24can think of it as a one-sided test in a monopoly

25maintenance case, which is to say, if you are in there

69

1arguing that some particular practice successfully

2maintained a monopoly and you come up with a credible

3analysis that says the firm had a very low share, that

4is likely to undermine the case. Now, it certainly does

5not work in the other direction, right? Just because

6you have a high market share does not mean you are

7guilty of any sort of offense at all and it may be that

8you got it because you deserve it. Okay, so that is a

9particular use.

10Now, I want to distinguish that from some

11others, because I think they often get rolled together,

12and they really are different, although they are

13related. Okay, so another one that is related is

14concentration as a screen for potential harm to

15competition. Now, in a sense what I just said, it is a

16screen, the one I just said, which is you are saying,

17look, if they have a tiny market share, is it really

18plausible that they have harmed competition

19significantly in the past, but I also want to worry

20about it going forward, and there it is not at all

21clear -- in fact, I think it is not a general

22proposition -- that you want to look at concentrations

23to understand the potential for harm to competition,

24because if you are looking at a case on a going-forward

25basis, sometimes the current share of the defendant is

70

1relevant, but other times, it is not, right? You are

2not worried about their share now. You are worried

3about what their share is going to become or what the

4state of competition will become going forward.

5Okay, so, notice I hedged it. I am an

6economist, so lots of "on the one hands, but on the

7other hands." So, in some cases where you are looking

8on a going-forward basis, current shares may be largely

9irrelevant. In other cases -- and I have the example

10here of exclusive dealing -- even when you are looking

11on a going-forward basis, market shares could be

12relevant, and I would think that would have been true in

13Dentsply.

14Now, as it turns out, in Dentsply, it was

15looking backwards, but if one had brought the case much

16earlier, I think what one could have done is say, look,

17Dentsply has this large market share, and I think by any

18sensible measure they had a huge market share, and we

19could argue about the source, but let me just

20hypothesize here without anyone arguing that it is

21because they did have teeth that were more popular and

22more attractive, that there was something about their

23product that they did have an advantage, and others were

24not able to imitate, and then you could use that fact to

25say, okay, that is going to tell us something about how

71

1exclusive dealing is going to work going forward, and

2even exclusive dealing with at will contracts, which is

3what was present in Dentsply, because this one firm's

4products were such a better fit with consumer tastes,

5that if you have exclusive dealing, that is where the

6dealers are going to go.

7So, in that case, concentration would be

8relevant as a screen or a way to think about what is

9going to happen but through a much more complex chain of

10reasoning than to just say, well, they have a high

11market share; therefore, they must have market power.

12It is really a very different kind of analysis, and that

13is the kind of analysis that I think needs to be done.

14Okay, the third one -- and actually, this is the

15one that is my favorite -- is say, look, we need to

16identify relevant markets, because if we are talking

17about harm to competition, we need to have some sense of

18who the competitors are, and actually, I think that is

19what the role should be in the merger analysis I will

20say as well, this really should be about identifying the

21competitors and then seeing where that takes us in terms

22of the but-for world, what needs to be the scope of the

23but-for world, and this is an unfashionable view,

24because it is low tech and it does not drive you to come

25up with algorithms, but I think it is important to

72

1remember in the end, this really is what we are trying

2to do.

3We are trying to figure out who are the

4competitors, because then we can ask, does this practice

5harm them? And if it does, does that matter for

6competition and does it matter for consumer welfare?

7Okay, so again, I think this takes us in a somewhat

8different direction, and notice, in this one, you may

9not be worrying about concentration very much directly

10at all.

11Also, since I had promised -- but so far have

12not done it -- the organizers that I would talk about

13innovation, let me say a little bit about that. When

14innovation competition is really significant, and this

15is not a point that is new to me by any stretch of the

16imagination, current market shares may not tell us very

17much, right, the extreme model being Schumpeterian

18competition, where we see a string of product market

19monopolies, but the real way competition works in the

20industry would be that you have firms that come in with

21major innovations, become the new monopolist, but then

22there is this battle for the next round of drastic

23innovation. If you are looking at a market like that,

24looking at market shares is not going to tell you very

25much.

73

1Okay, a couple things about market definition

2and uncertainty. First off, we have talked about burden

3shifting a little bit. As everyone in this room who

4works for the Government knows, right, meeting the

5market definition burden can be difficult, and that is

6true even if you do not have innovation, and I will come

7back to innovation in a minute. One of the difficulties

8is when courts say we want a zero-one boundary. Every

9firm is either in the market or they are out of the

10market; none of this wishy-washy stuff.

11The problem with that is it can be really hard

12to do. I know Oracle is a merger case, but it is really

13striking because it is a case where Judge Walker said,

14all right, look, here are the economics of why you

15cannot draw zero-one boundaries. You have got product

16differentiation. You have got a continuum of products.

17There is no way there is going to be a sensible

18boundary. He did not say, "And oh, guess what, that

19means you lose."

20I mean, I think Judge Walker was right about the

21first part. It is just the notion that that is where it

22takes you I think is a little troublesome. It is

23particularly troublesome as well because if you believe

24that these are differentiated product markets and you

25believe competition is localized, then you really have

74

1to ask yourself, why are we worrying about a broader

2market anyway? I mean, what is the relevance of this

3alleged relevant market if what really matters is

4defined structure?

5So, it seems to me that where we have gone with

6a lot of -- just to jump back to mergers for a second

7where I think there is a broader lesson here -- with

8mergers, is worrying about unilateral effects cases in

9markets with differentiation -- and everyone seems to

10have conveniently forgotten that you can have a

11unilateral effects case with homogenous products -- but

12we have spent all this time worrying about market

13definitions in precisely the wrong places.

14Now, although this gets worse if you have

15innovation, because you can have things constantly

16changing, you can have products -- the characteristics

17of products are changing, I just want to make two points

18on this and then move on quickly. One, there are a lot

19of people who seem to be of the belief that what

20innovation means is markets are constantly getting

21broader, okay, and there is a set of people who will

22say, look, you have got all these things, you have got

23innovation, markets are always going to be so broad

24because new products can keep coming in, that really,

25there is nothing for antitrust to do. I would just like

75

1to remind people that, in fact, markets could be getting

2narrower, because these products are evolving, they are

3moving targets, and it is quite possible that some

4products or the producers of those products are falling

5behind in terms of innovation and they are dropping out

6of the relevant market.

7Okay, the point I have already made, that if you

8are looking at differentiated products and then you

9throw in the complexities of innovation, you just really

10may make it impossible to meet the burden. As we have

11talked about, since I think there is a fairly broad

12consensus, you do not really need to have a rigid market

13definition. That is unfortunate, but that is how a case

14would be decided.

15Now, I have to have a diagram. So, what this

16one shows, just very quickly, suppose there is

17disagreement on the scope of the relevant market here,

18and I am interested in a case where I will just suppose

19that one has beaten up on two, okay, these are suppliers

20markets, and this line represents some notion of product

21differentiation, and there is a debate. It is hard to

22know whether the market boundaries are -- the ones who

23have the narrow subscripts, so only include one and two,

24or they have the broad, and then they would include

25producer three as well. Suppose we get the debate down

76

1to that level. This is a dramatic oversimplification.

2Well, you can imagine a court, Judge Walker

3saying, "Look, Government, you cannot tell me whether it

4is the narrow definition or the broad one with

5certainty, so you lose." But suppose it does not make

6any difference whether you include three in the market,

7okay, to what you think are the competitive effects,

8then why does it matter that you cannot say which one is

9which, okay? So, what you really want to ask is not

10whether or not the plaintiffs can prove a market

11definition with certainty, but you want to ask can they

12tell you, "Look, we know well enough where it matters

13with a high degree of certainty."

14So, the approach to this would be to then ask,

15"Where does the dividing line matter," okay? Go back to

16this, "Does it matter whether we include five in or

17not?" If it turns out what is critical in the end is

18whether three is in the market, let's fight about that.

19Let's not fight about, no, you have to come up with the

20definition.

21Okay, a quick thing on decision theory. I have

22a pretty picture, I have to show it. What this is

23saying is -- I just want to make the following point, I

24probably will not actually go through the picture, so

25just admire it while I talk. It was not easy drawing

77

1this on the train while it was jerking around -- but is

2the following, that there is a lot of focus, I think, in

3court cases, at least, in actual legal decision-making

4on doing things like asking are probabilities above

5certain thresholds or is one probability higher than the

6other, something like that.

7This would be a diagram where if you weigh

8evidence, you would just ask, is the probability of harm

9bigger or less than the probability of efficiencies?

10So, you would get in that red zone, because that's where

11the probability of harm, P, would be viewed as being

12higher than the probability of the efficiencies, Q, and

13you would just sort of -- that is one interpretation of

14weighing the evidence. There are others, I will note,

15and if I had a longer time, I would tell you some of the

16others.

17Now, but if you try to balance the effects, you

18do not just look at the probabilities. You have also

19got to look at the magnitudes, and I have given the

20example here where the harms, denoted by H, are bigger

21than the efficiencies. So, in fact, you want to condemn

22not just practices where the harm is more likely or

23equally likely as the efficiencies. You actually want

24to condemn some where the harm is less likely, but the

25problem is, well, it is less likely, but when it

78

1happens, it is a worse thing, and that is where you get

2that purple area.

3I would say in the end, since we are worried

4about effects, the right thing to do, and if we do all

5this stuff, would be to condemn this bluish-purple area

6plus the red, but if you simply weigh the evidence, you

7are only going to get rid of the red. So, you are going

8to -- if there is enforcement, you are going to have

9false negatives. So, I think what is important in all

10of this, and there are many other interpretations of

11this, but the central point is I think we do have to

12worry about magnitudes more than we have in the last --

13okay, are you going to unplug this? This is like the

14Academy Awards, they start playing the music.

15Innovation, I will say one thing in support of

16innovation markets as a broad concept, because certainly

17they have been controversial in terms of actually using

18them, but if we are worrying about markets where

19innovation competition is really critical, then we need

20to worry about what is driving innovation, who the

21potential innovators are, and looking at markets in a

22product market may not tell you very much about it. It

23may be much more informative to look at the distribution

24of R&D capabilities and assets.

25As some people, one of them sitting near me,

79

1have pointed out, that can be really hard, because it

2may not even be in this industry, but that is

3conceptually the right thing to do, and so I think we

4ought to be asking ourselves, how do we get there? If

5we conclude it is too hard to do, fine, but I don't

6think it makes sense to say -- and persons near me

7didn't say this -- "Oh, it is too hard to do; therefore,

8let's go and do something else that does not make any

9sense but is easier." I think we want to keep in mind,

10though, that the R&D capabilities and the distribution

11of the assets there may be much more important than

12current market shares in terms of understanding

13innovation.

14Okay, last thing, which does not have anything

15to do with anything except people always screw it up. I

16will make what has actually turned out to be a

17controversial statement in practice, that geographic

18markets are markets, by which I mean since they are

19markets, they have buyers and sellers, okay? In

20practice, at least my experience has been that people

21often forget about the buyers part of that description

22of markets, and then if we are going to talk about

23geographic markets, we need to think about the buyers

24and where they are and the sellers and where they are.

25Now, in some markets, in the end, there may be

80

1global markets and those do not matter, but other times

2you want to ask something like, particularly in

3retailing, say, or certain kinds of manufacturing, you

4would want to say, let's look at a set of customers in a

5particular city and ask what producers, and in

6particular the producers' plants, can serve those

7customers, and look at it that way.

8Now, that may mean that a firm is in a lot of

9different geographic markets, and a single plant, by the

10way, could be in different geographic markets

11simultaneously, which drives people crazy, but if you

12want to think about what is really going on and take

13markets seriously, you have got to remember, markets are

14bringing together buyers and sellers, so we need to

15discuss or describe the locations of both of those.

16With that, I will stop.

17(Applause.)

18DR. CARLTON: Okay, thank you. The person close

19to you says, "Thank you very much."

20Okay, I would like to ask the panelists some

21questions. We have about 45 questions left, and I have

22a series of questions. I have about ten questions. I

23do not know if we will be able to get through them all.

24What I will do is I will ask the question, and then I

25will ask two of you to comment. If you could keep your

81

1answers relatively brief, that would be good. If

2someone on the panel who we have not asked feels they

3want to comment, they should do so, but since there is

4an opportunity cost, that just means you may not get to

5answer a later question.

6Here is what it seems to me that the purpose of

7these hearings are. One, we want to define market

8power. Can we agree on a definition? If we can, do we

9think defining the market and then taking market shares

10helps us in a section case? Then, what are the hard

11questions where we think that that may or may not help?

12Then the ultimate question really is -- and this

13I will ask everybody to answer, it will be the last

14question -- do we really need market definition and is

15it more of a hindrance than a help?

16So, let me just start off on first asking the

17question about market definition. In the legal

18literature and in the cases, they stress not just the

19 ability to control prices, which is what economists

20focus on, but they always add, "or the ability to

21 exclude competition," and it is that second prong I want

22to focus on for a second.

23I understand -- and Andy spoke a little bit

24about this -- that a joint venture can get together and

25exclude people. Let's just talk about single-firm

82

1behavior, and I am interested, in particular, from both

2Andy's point of view and Joe's point of view, with their

3sort of combined economic/legal backgrounds, if they

4could comment on whether they think the exclusion prong

5of the market power definition that is used in legal

6cases is useful. Do we need it? Can we do without it?

7For example, can we do without it by saying,

8"Well, it is the ability to control price, and if you

9say keeping it above the competitive level, obviously

10the competitive level is the level that arises when you

11do not exclude competition." If we can simplify the

12definition, it seems to me that helps things rather than

13complicates things. So, is your view that we need that

14second prong, exclude competition, in the definition of

15market power or not?

16So, let me first ask Andy and then I will ask

17Joe, and if you could keep your answers sort of

18relatively brief, that would be good.

19DR. GAVIL: I think in exclusion cases, the

20answer is yes, but it winds up being a first step. The

21ability to exclude competition -- I guess the "or" is

22the problem. Why do we have monopolization? We have

23monopolization cases because we want to prevent not just

24any exclusion of competition; it is exclusion of

25competition followed by the ability to either maintain

83

1price, maybe raise price, but the two to me go hand in

2 hand.

3In any section 2 case, the first step is going

4to be evidence of some exclusion, but I do not think you

5can stop there and conclude from that that there would

6automatically be monopoly power. You have to ask the

7second question of whether or not the exclusion will in

8some way facilitate the maintenance or the enhancement

9of the market power. So, I think it winds up being

10circular. You do come back to power over price.

11DR. CARLTON: Okay, Joe?

12MR. SIMONS: I agree with what Andy said, and

13also, just to follow up on what Rich said about the

14guide posts and the lamp posts. You know, what you see

15in the case law is an example of a lamp post. It is not

16an example of a guide post. That kind of definition is

17drawn generally from whatever you guys refer to in the

18equilibrium analysis or partial equilibrium analysis or

19whatever it is, and they just moved it over and said,

20"Here, this is what we are going to do," without

21thinking about why we really want to do it.

22The statute talks about monopoly, so you tend to

23have to have a big share and so it is natural that a

24share requirement gets imported into the law. But it

25does so without thinking, and so I do not think that

84

1focusing on that question based on the case law is going

2to be terribly helpful.

3I think Andy is right, you want to focus on why

4are we asking this question, what are we trying actually

5to prevent, what is the goal.

6DR. CARLTON: Okay, I think I agree with that.

7I think probably that is a fair summary of what you

8said, that I think both of you say we can get rid of

9that second prong as long as you keep your eye on the

10ball. In effects cases, obviously you have done

11something bad, and then did you raise price. So, if you

12are wanting to define market power alone, it is whether

13you can raise the price above what it would otherwise

14be.

15MR. NELSON: Or prevent it from falling.

16DR. CARLTON: Or prevent it from falling, that

17is right.

18DR. GAVIL: I think the exclusion does tell you

19something. I would not eliminate it entirely. I think

20the problem is it does not tell you whether or not you

21have monopoly power, but it is like the first red flag.

22It is the first guide post that tells you there may be

23reason to be concerned about a particular situation, but

24you cannot stop there. You have to ask the second

25question.

85

1Even going back to Salop and Krattenmaker, the

2title of the article was Raising Rivals' Costs to Obtain

3Power Over Price. So, the two really do go hand in

4hand, but the first sign of a problem may be the

5evidence of exclusion.

6DR. CARLTON: Right, but what is a mechanism to

7achieve the control of price? I agree, it is important

8to have both, but I am just trying to distinguish the

9two. One of the things that goes on in a section 2 case

10is you define markets and you have exclusion -- and I

11will come back to this later -- and the question is how

12you link the two. I am trying to keep them separate for

13a second.

14MR. SIMONS: I think in what Krattenmaker and

15Salop do with their article is they are linked. It is

16the exclusion that gives you the power over the price.

17What is the impact of the exclusion? Not kind of in a

18general sense, have you been able to exclude people, all

19right? Because maybe you have because you have such a

20terrific product or you have a patent or whatever it is.

21That is legal. The question then becomes, did you do

22something in addition to that that may not be so legal,

23and does that give you power over price?

24DR. CARLTON: Right.

25DR. GAVIL: Think of instances where the act of

86

1exclusion raises entry barriers.

2DR. CARLTON: Yes.

3DR. GAVIL: That leads you to the second part of

4it.

5DR. CARLTON: Yes and no. What that tells you

6is that but for the act, which we are trying to claim is

7illegal, the price would have been lower, and therefore,

8you have the power to set price above the but-for price.

9It is just defining what the but-for price is.

10Okay, let me go on, because I am going to come

11back to this benchmark point. The definition that

12economists use a lot is that market power is the ability

13to set price profitably above the competitive level,

14presumably by a significant amount, for some significant

15amount of time. So, first, I have two parts to this

16question, and I am going to ask Phil and Larry.

17Assume that there are constant returns to scale,

18so competition is possible. So, first, do you agree

19that the definition I gave you is a reasonable one --

20put aside whether it is implementable, but is it a

21reasonable one -- and if so, what is a significant

22amount of the price increase and what is a significant

23amount of time?

24In particular, when you are answering, if you

25could talk about why we do not pay attention to dead

87

1weight loss and why we just talk about numbers. I mean,

2we are economists, and 5 percent, 10 percent, we know

3that may not be meaningful depending upon the size of

4the market. So, if you could just address those.

5DR. WHITE: Are you looking at me? Look, you

6know, where do 5 and 10 percent come from? As Bill

7Baxter used to say, from these (indicating hands), and

8there's nothing magical about that. You know, it partly

9would also depend on how much noise you think is out

10there protecting ourselves against error that might be

11harmful. So, the real answer -- the first part is yes,

12under constant returns to scale, a price significantly

13above marginal cost, sustained for a sustained amount of

14time, would in my mind constitute an exercise of market

15power, and how much and for how long, I do not know.

16Sure, 10 percent sounds like a number to be

17thinking about and two years sounds like a number to be

18thinking about, but I have just picked those out of the

19air, and I do not have any further basis.

20DR. CARLTON: Okay, let me just say one thing.

21My preference would be it is probably better -- even

22though it is hard to choose a number, someone is going

23to choose a number, so you should think, as to your

24willingness to choose a number, would you rather some

25random judge choose a number or this panel? So, that is

88

1why I am asking.

2DR. KATZ: I mean, I disagree with the premise.

3Why should you choose a number? I am almost

4certainly -- if you thought the court was going to do

5enough of the analysis -- and we would have to talk

6about the cost of the court and the time they have --

7but almost certainly you would say the number depends on

8the market. I mean, there are some markets where

9worrying about a price change within 5 or 10 percent, I

10mean it is completely lost in the noise, because the

11prices are changing 40 percent every year, so it does

12not mean a 10 percent price increase could not matter,

13but it becomes less plausible you could actually tell.

14In other markets, it might be that you could reliably

15predict a 3 percent price change.

16DR. CARLTON: Following that same logic,

17wouldn't you be concerned about a 1 percent change in a

18market that is huge?

19DR. KATZ: If you believed you could actually

20make reliable predictions at that level, yes. So, I

21think you need to look, as you were saying, at the

22magnitudes of the effects, and some of it comes within

23when do you want to bring cases and how to allocate

24resources and then also the various characteristics of

25the market that are going to affect the reliability of

89

1your projections and whether you think that you really

2can discern at those levels, but I think it would be

3pretty clear that holding aside -- which is obviously a

4big thing to hold aside -- the various sorts of

5processing costs, there is no reason to think there is

6one right number, and, in fact, there certainly isn't.

7DR. CARLTON: The question is, should we give

8any guidance to the courts when they are trying to

9decide whether a firm has market power, and if you just

10say it is up to the discretion of the judge based on a

11lot of things -- I mean, I agree with you, it is hard to

12come up with one number. The question is, is it better

13leaving it completely to the discretion of the courts,

14or should we not -- I think one of the advantages of the

15Merger Guidelines, even though they make the point that

16the 5 percent is just a suggestion, is that it has

17focused thinking and clarified thinking. So, I agree

18with everything you have said, but in light of the

19decision-making of the court process, there can be a

20benefit to articulating some standards, maybe flexible

21standards.

22DR. KATZ: I would agree with that, but I think

23a question would be -- and this is just thinking off the

24top of my head -- could you say something like -- have

25some sort of relatively easily observable data, say like

90

1the annual price changes or something, or try and do

2something that says that the standard you use should be

3proportional to some characteristic in the market? We

4would have to think a lot about what that is, and I

5think ideally, for the reasons you bring up, it would be

6something fairly mechanical, but it would still be an

7improvement over a one-size-fits-all.

8DR. CARLTON: Rich?

9DR. GILBERT: Well, I certainly agree that the

10number, however you define this number, depends on the

11nature of the conduct, the efficiencies that can be

12presumed to go along with that conduct, and maybe the

13size of the market and all of that, but I also think

14there is the case that can be made for shifting the

15inquiry to something like the firm-specific elasticity

16of demand, which often can be measured in many

17instances. I think Greg has pointed this out in some of

18his writings.

19It is not that hard to say if the elasticity of

20demand is bigger than 10, maybe we shouldn't be worried

21about this. On the other hand, if it is in the range of

222 to 3, maybe we should be worried about this.

23DR. CARLTON: Yeah, that raises a point I am

24always puzzled about, that if you are thinking about

25what is a magnitude that is important, an elasticity of,

91

1say, 20, which everybody would say is a really high

2elasticity, that gives you a 5 percent upcharge over the

3competitive price. So, that should tell us something

4about our intuition versus sort of practical --

5DR. GILBERT: Well, on that, maybe I am

6differing from other people, I think of that 5 percent

7rule as being a derivative, not an absolute amount. So,

8we ask, if quantity goes down by 5 percent, will the

9price go up by 5 percent, that sort of thing, and rather

10than because we are really worried about the price going

11up by 5 percent. Now, some people I know would disagree

12with that and would say that that 5 percent is a

13threshold of concern. I think of it more as an

14elasticity test.

15DR. CARLTON: Okay.

16MR. NELSON: Since I was one of the original --

17DR. CARLTON: I am going to give you another

18question, okay? It is actually a harder question now.

19We are going to move on to something else. That was an

20easier question. If you remember, that was premised on

21constant returns to scale. So, it could actually define

22a competitive price.

23Let's suppose now that I am in an industry where

24there cannot be competition. There is a fixed cost of

25entry. There are constant returns to scale, and it is

92

1Cournot competition, okay? What is the meaning of that

2common phrase that we use, can you profitably price

3above the competitive level? What in the world should

4we take as the competitive level in that situation? Is

5it the zero profit equilibrium or is it price equaling

6marginal cost?

7So, let's see, maybe Phil, if you want to take a

8crack at that.

9MR. NELSON: Well, one of the things that sort

10of concerns me about taking sort of the current level as

11opposed to something like marginal cost is you do have

12some of these monopolization cases that are really

13entrenchment theories, and is the question whether the

14entry is going to drive you significantly back towards

15competition, or this guy already has some market power,

16and he is going to --

17DR. CARLTON: Try to define the market

18equilibrium, free entry, fixed costs, constant returns

19to scale, Cournot equilibrium, do we want to call that

20market power?

21MR. NELSON: I guess I am saying that to answer

22that, you want to know sort of what your benchmark is as

23to where you're going.

24DR. CARLTON: Right, that is what I am asking

25you.

93

1MR. NELSON: Yeah, and what I was going to say

2is that I think you would start to look, as N goes up,

3what happens to the equilibrium price? Then as N gets

4high enough, are you still at a price where somebody

5could make an economic profit? I mean, you are going to

6want to see if that is a tenable number of firms and

7start to use something like that as the equilibrium,

8which is higher. It is going to be a lower price and

9define market power in some circumstances where you

10might not find it if you are at your starting point.

11DR. CARLTON: So, the point of the question is

12to show that there is a difficulty in defining market

13power when you cannot define the competitive price. You

14can define a rate of return, and you can define marginal

15cost in this example and prices above marginal cost in

16this example, but profit is zero, and there seems to be

17a complete ambiguity between the willingness of people

18to distinguish which of those two definitions they are

19using.

20Is it price above marginal cost that is market

21power? Is it rate of return above a competitive level,

22or which of the two, or are those two different things?

23They obviously from an economic point of view are two

24different things, yet often, in the writings and in case

25law, they in my mind do not get distinguished.

94

1MR. NELSON: I mean, yeah, you want to have

2profit -- you want to be able to make a monopoly profit.

3I mean, if you have got easy entry, as some of the

4different -- you know, if you don't have any profits,

5then they are not going to have enough -- but I --

6DR. CARLTON: Larry, did you want to say

7something?

8DR. WHITE: But why would we be interested in

9your hypothetical? If it is somebody coming in and

10saying, "That guy is charging an outrageously high

11price, Judge, find him guilty of a section 2 violation

12and mandate that he charge a lower price," we do not see

13that all that often, but that would be a problem. If it

14is, "Judge, that guy has excluded me from offering my

15rivalrous product, and had he not excluded me, I could

16have come in and the price could have been lower,"

17that's a different --

18DR. CARLTON: I agree, but that is mixing

19together two different questions. The first is, what is

20the effect of this action? If you can answer that

21question, you have answered the section 2 -- you have

22resolved the section 2 issue.

23DR. WHITE: And then we do not have to worry

24 about it.

25DR. CARLTON: And then we do not have to worry

95

1about market definition; however, the way the courts

2seem to use market definition in section 2 cases is not

3like that at all. Courts seem to do the following:

4Unlike a merger context where you ask, as a

5result of a merger, is market power going to go up, the

6courts define a market and then look to define market

7share. Courts do it. They do not look at the change in

8the market shares that arise as a result of the bad act.

9They do not do that. That would be an analogy to a

10merger case.

11Instead what they do is they ask, is there

12market power? They do not ask about the change in

13market power, but they ask, is there market power? They

14use that as a screen whether to then further

15investigate, and that distinction, that asymmetry

16between a merger case and a section 2 case, I think

17leads to peculiar discussions, but it also I think leads

18to exactly why I am asking this question, which is, if

19the courts are going to go this route and use market

20definition -- I agree with you, Larry, if you do an

21effects-based analysis, you can solve the problem -- but

22the first question the court is going to be asking, is

23there market power, and I am just trying to figure out,

24can we even define what we mean by that in this Cournot

25example?

96

1MR. SIMONS: I think what you want to ask is why

2are we doing this, why are we engaged in an exercise,

3before you can even think about answering the question.

4DR. CARLTON: This firm has been sued, there is

5a bad act, and the first question is, does he have

6market power? And I am trying to find out -- I cannot

7answer that -- begin to answer that question unless we

8can agree on a definition of market power. So, is the

9definition price above marginal cost or is the

10definition rate of return above the competitive level?

11Mike?

12DR. KATZ: The problem is if you are going to

13say this has to be a screen that works for everything,

14then the most useful definition of market power would be

15does the firm have at least one employee or something

16that is equivalent of it so we throw this screen out,

17because what Larry has pointed out -- I think what in

18most cases makes sense is something that says -- and

19actually, I make a different distinction, and I think,

20actually, a lot of economists writing not as part of

21antitrust make a different distinction. I think a lot

22of people, economists, would say that market power would

23be facing downward-sloping demand curve and not having

24it perfectly elastic, which then would end up giving you

25the profit-maximizing price of that firm above marginal

97

1cost. I think that is a useful definition of market

2power.

3Then I try to reserve monopoly power for being

4two parts. One, that you have a lot of market power,

5which is to say the price would be -- and again, I will

6be vague -- but significantly above marginal cost, and I

7would typically put in a test saying for I think most

8purposes or a lot of them, we do care whether or not the

9price is above average cost, whether or not there are

10profits, but what Larry has pointed out, I think

11correctly, is that test does not always work, that if

12what you are worried about is somebody who is in there

13now and is just breaking even but is narrowly keeping

14all sorts of more efficient entrants out who could make

15a profit, I think in that case, saying, "Well, look,

16they are not making money, there can't be a problem,"

17would give you a misleading answer.

18MR. NELSON: That was my standards example I

19gave in my opening talk.

20DR. CARLTON: I think, again, that is really

21asking the but-for price; in other words, price may

22equal marginal cost and price may be above average cost

23in the present environment, but but for the bad act,

24that would not occur, okay?

25The distinction you make between price above

98

1marginal cost and then whether the rate of -- the price

2above average cost, the rate of return is above the

3anticipated return, is exactly the distinction that I

4made between market power and monopoly power. It is a

5logical distinction. I am not sure -- we may be the

6only two people who make that distinction, because I do

7not see the legal cases going in that direction.

8So, I guess I do have a question, and I think it

9is a relevant question, as to whether the distinction

10between monopoly power and market power that we do see

11in the cases, is that a useful distinction, and is it a

12useful legal distinction? Is it a useful economic

13distinction?

14So, maybe, Andy, you could answer that and maybe

15Rich.

16DR. GAVIL: Yeah, one point I wanted to make

17earlier and I think I can make it now in answering the

18question, I think historically the association of market

19power and monopoly power as being different things was

20linked to market share. It was linked to circumstantial

21evidence as the basic mind set that we used to approach

22cases, and I think a concrete example of this is the

23Supreme Court decision in Copperweld, where it says --

24it is known for the parent/subsidiary enterprise

25conspiracy issue, but it has a very interesting

99

1discussion of the relationship between section 1 and

2section 2, and it says, "An unreasonable restraint of

3trade by a single firm is not reached under section 2,

4and therefore, the drafters of the Sherman Act left a

5gap between section 1 and section 2, and the implication

6was that for a section 2 case, you need something more."

7At that moment in time, the "something more" was the 70

8percent or more market share as opposed to the 40 to 60

9percent that was typical in rule of reason cases.

10If you let go of the commitment to the Alcoa

11framework and the market share associations and start

12thinking about market power in different ways as

13expressing itself in different ways, that kind of mind

14set of distinguishing market and monopoly power based on

15market shares goes away, and I think that that would

16make a big difference in how we think about antitrust

17generally.

18But you have said it several times, Dennis, and

19it is clearly the case, that courts say, "Okay, the

20first element under section 2, do you have monopoly

21power?" On your no profit example, if I could just

22throw in, what if the purpose of the conduct was to make

23that firm profitable and that is what it was trying to

24do? So, currently it is not profitable, but the whole

25point of the conduct, maybe it affects entry barriers,

100

1was that they are trying to get profitable by engaging

2in conduct. Again, I think it shows the link between

3the conduct and the power increase.

4DR. CARLTON: Rich?

5DR. CARLTON: Rich?

6DR. GILBERT: Yeah, if I can answer this, as has

7been said before, in some sense I subscribe to the

8argument that monopoly power is a lot of market power,

9but it is also market power that is durable. Now,

10whether you define durable market power in terms of the

11ability to raise price above average cost, the ability

12to maintain price above average cost, or the ability to

13maintain price above long-run marginal cost, I do not

14think that is really critical. To me, it is the ability

15to exclude, and as you have noted, Dennis, yourself,

16that when you are talking about exclusion, obviously it

17also depends on thinking about entry barriers, and then

18when you think about entry barriers, you have to think

19occur.

20So, if you had an extremely competitive market

21post-entry, maybe a little bit of exclusion is enough to

22maintain a monopoly position, but I think the key issue

23is the ability to exclude is important to me, because it

24says something about the ability to maintain price above

25some measure of long-run profitability of an efficient

101

1competitor.

2I want to add one other comment that I think is

3related to all of this, which is we are very good when

4we talk about impacts on competition to understand that

5impacts on competition is different from impacts on a

6competitor, I think we have learned that one, but when

7we talk about section 2 cases, we are often talking

8about the market share or monopoly power of a single

9firm. Shouldn't we be talking in many of these cases,

10at least, if not all of them, about the power in the

11market, not just the power of this firm, because

12obviously if the firm reduces supply so that its market

13share is below the Alcoa threshold, but in doing so,

14raises market power generally in the industry, that is a

15problem, and we want to look at that, not just what the

16firm's market share is or focusing on the firm.

17Now, if you did this firm-specific residual

18demand analysis, then you pick that up by looking at the

19elasticity of the residual demand. So, I think it is

20all right in that context.

21DR. CARLTON: Let me go back to something that

22sort of was a common theme in some of the presentations.

23Let me restate it as follows: It really has to do with

24what the benchmark price is.

25If you look at a firm in a section 2 case and it

102

1is engaged in a bad act, can you then ask, "Well, does

2that firm have market power," which is what the courts

3first ask, and if the answer is no, they throw it out.

4In order to answer that question, you have to ask,

5"Well, what would the price --" depending on your

6definition of market power, you want to ask, "Does the

7firm have market power?" Whatever your definition is,

8whether it is pricing above the competitive level after

9the bad act, are they pricing above the level but for

10the bad act, whatever definition you want to use, and I

11think it is the latter definition that makes more sense,

12it is not obvious why market shares and market

13definition help you answer that question, because if you

14know the current price and you knew the benchmark price,

15it is just a comparison of two prices. So, calculating

16market share in that case does not advance the ball.

17If that is the typical case that we see in

18section 2, what really are we talking about when we are

19doing market definition? Are we really doing an

20analytic economic exercise, or are we doing something --

21or are the courts doing something much more -- I don't

22want to say sensible, but much more using common sense,

23which is there are five guys doing the same thing, don't

24bother me, and they're just using their common sense.

25Now, how you define "five guys doing the same

103

1thing" may be hard, but it seems to me that is what a

2lot of courts are doing, and I am wondering if we are

3worrying too hard about defining markets in cases where

4market definition is just this seat-of-the-pants thing

5that the courts then use, and as long as they understand

6it is real seat-of-the-pants, don't bother me with

7details about market shares and get on to your

8competitive effects analysis.

9So, maybe, Joe and Mike, you could comment on

10that.

11MR. SIMONS: Yeah, I think that what the courts

12will do is not just say, well, let's get on with it and

13let's get to the competitive effects. It is a real

14screening event, a big one, and it also seriously

15impacts what happens when lawyers counsel their clients.

16If there is some chance that your client is going to be

17deemed to have a big market share, at least most lawyers

18I know will give advice that is much more conservative

19than if their shares are 30 percent. So, it makes a big

20difference in the real world.

21I think the judges do focus on it, and it is

22important in court now, and there is a serious question

23in my mind about how important it should be, certainly

24with respect to how the Antitrust Division and the FTC

25exercise their prosecutorial discretion -- whether they

104

1really need to get hung up on this or whether they

2really need to make a decision about what is the impact

3of whatever conduct we are worried about. Did it have a

4significant impact, and then, when we are proving in our

5case in court, it is a different exercise.

6DR. CARLTON: Okay, who did I say? Mike?

7DR. KATZ: You are not supposed to remind me of

8that. No, I would say a couple things. Part of it -- I

9will come back to what I said in my presentation,

10though, is that we can be using market definition in a

11number of different ways and that the level of -- we

12want to understand who the competitors are, because we

13want to figure out that is where we are going to see the

14competitive effects, are they harmed or not, does it

15matter if they are harmed, does it matter for

16competition and for consumer welfare. So, that level, I

17think we would certainly want to do market definition,

18but that may not be through a formal algorithm.

19In terms of your question about the alternative

20prices, I think there is a difference between asking

21about a but-for price and asking about certain

22interpretations of the competitive benchmark, because

23you can take a competitive benchmark to be some formula

24based on marginal cost or average cost or something like

25that, and you could ask a market power question, but you

105

1could well conclude from that that yes, this firm has

2market power, but it has not done anything wrong, okay?

3It has market power because it has been a great

4innovator. So, I think that that is a different

5question than asking is it charging a higher price than

6the but-for price, because there you are asking about

7what would happen as a result of the specific piece of

8conduct. So, I think if one wants to go through the

9market definition exercise in that form and to have the

10competitive effects analysis be different than the

11market definition analysis, I think you can do it. One

12would ask about almost this more abstract or formulaic

13competitive price as the benchmark for market

14definition, and then the competitive effects analysis

15would look for a but-for price and would take into

16account a specific practice.

17DR. CARLTON: I just don't know how to do the

18first in an analytic way that is other than comparing

19the two prices. If I cannot compare the two prices and

20I have to do a competitive effects, say an econometric

21analysis, I do not really need market definition. So,

22that is why I think what judges often do is, as Joe

23described, is do a seat-of-the-pants analysis or I

24described as a seat-of-the-pants analysis, but as Joe

25described, that is their screen.

106

1DR. KATZ: Well, the screen makes sense if what

2the plaintiff is saying is there has been successful

3monopolization and you end up coming out of this being

4convinced that here is a sensible market definition that

5tells me how competition works, and this particular firm

6does not seem to be dominant in any sense or doing well,

7and if you see that, then it seems to me it does pretty

8heavily undermine the claim that there was successful

9monopolization.

10DR. GILBERT: But that's Cellophane. I mean, it

11is Cellophane, did not look like they were --

12DR. KATZ: No, Cellophane, they did not have the

13sensible market definition.

14DR. GILBERT: Then you are back down to how do

15you define the market.

16MR. SIMONS: Dennis, think about it this way:

17Someone had mentioned a gap earlier, and maybe it was

18Andy. If you think about under section 1, right, if you

19prove an effect, you win, right? Under section 2, the

20way you are describing the state of the law, which is

21accurate, is it is unilateral conduct. No contract.

22There may be an effect, and then liability is going to

23turn on whether there is some high market share.

24DR. CARLTON: Right.

25MR. SIMONS: So, the question to me would be,

107

1why do you want to do that?

2DR. CARLTON: I think that is right. Let me

3flip the question a bit.

4It seems to me this emphasis on market

5definition in section 2 cases is coming precisely

6because of the way judges apply these screens and

7that -- I cannot remember -- I think Andy mentioned it,

8that just like you might want to have decision processes

9based on market shares, you might also want to immunize

10certain types of safe -- have safe harbors and as well

11as have danger zones, and it is the fact that it seems

12to me that the sequential decision-making in Section 2

13cases is first to look at market definition as a screen

14and then you go to competitive effects that causes this

15undue emphasis on market definition, and one way around

16that might be -- and this is going to be a question I

17will ask Andy and Phil -- suppose we also allow a screen

18based on safe harbors and said, "No section 2 cases if

19you're doing X, Y and Z; no section 2 cases if market

20share is -- you do not have market power."

21Wouldn't that be a way to de-emphasize the role

22of market definition, which I think we are all agreeing

23is difficult to define in a section 2 case?

24Let me first ask Phil, and then he --

25MR. NELSON: Okay, wait a second, no market --

108

1DR. CARLTON: Either you do not have market --

2the current screen, but I am going to add to that

3current screen that there are certain safe harbors, and

4what we should do is spend more time on defining the

5safe harbors for conduct rather than trying to figure

6out can we define markets any better.

7MR. NELSON: So, it is conduct safe harbors,

8not --

9DR. CARLTON: Correct, yes.

10MR. NELSON: -- not structural safe harbors.

11DR. CARLTON: Correct.

12MR. NELSON: Okay, there was an "and" there, and

13I was starting to think that maybe where you wanted to

14go was a combination of a structural safe harbor with a

15conduct safe harbor, because there are certain types of

16conduct that might mean a lower market share, you could

17still have a problem, like some of these -- but I think

18there is a -- as I was alluding to, the importance of

19performance evidence, which is another way maybe of

20saying conduct, that you would want to start looking at

21some of that conduct evidence.

22However, I am a little worried that the problem

23is that a lot of this conduct is not so easy to

24categorize, so that when you start to try to define a

25safe harbor using conduct evidence, I am not sure that

109

1you are going to find a lot of situations where you can

2say for sure that this is conduct that is absolutely

3okay, because in other contexts, it may not be.

4DR. CARLTON: I agree. Safe harbors make

5mistakes, but that is what decision theory tells us is

6the right thing to do.

7Andy?

8DR. GAVIL: I think the idea of defining safe

9harbors and danger zones, as I said, is useful, and I

10think you cannot do it just by using market share

11numbers, which has been our tendency in the past.

12Now, once you use a market share number, you are

13stuck in the, "Okay, we need to define a relevant market

14problem." So, I think that reducing it to certain

15characteristics of the market, maybe it is structural

16characteristics, performance characteristics, but trying

17to look at other kinds of measures might make the safe

18harbor and the danger zones a little bit more meaningful

19and move the attention away from market share.

20But one last comment, Joe said how this affects

21you in advising clients. That 70 percent share that has

22become pretty fixed for monopolization cases, that is

23perceived, even by defendants who do not like the market

24definition and market share approach analysis, that is

25perceived as a pretty big, significant safe harbor when

110

1you are advising single-firm clients, and keep that in

2mind with all the monopolization cases.

3If there really is not any good, strong argument

4that you could be in a market with a market share that

5is up in that range, you are pretty much free to do

6whatever you want. So, if we moved away from it, I

7suspect you would actually hear some objections from

8large firms that perceive that it is actually a very

9useful benchmark.

10So, where I would come out is, I do not know

11that you can completely get away from the market shares,

12but maybe we need something like a market share plus,

13and not that it is a great model that we would ever want

14to rely on, but the concept from the Sentencing

15Guidelines that you have a guideline, but then you have

16factors that allow you to depart upward or downward, and

17that is sort of what Michael was talking about earlier

18when he was answering one of your questions, is certain

19factors might lead you to be cautious or less cautious

20in certain circumstances.

21DR. CARLTON: Let's see, let me skip a few

22questions since I am going to try and end roughly on

23time or maybe at most five minutes late, but let me ask

24a question about technology, and I am going to direct

25this to Rich and Mike, because you have done a lot of

111

1work in these areas.

2It seems like a really hard question is where

3you have industries where marginal cost is low, product

4innovation is the key, and new products are coming out

5every so many years. It kind of came up a bit in the

6Microsoft case, and Dick will probably talk about this

7tomorrow, Schmalensee, but what do you mean by "market

8power" in those industries unless "durable" really means

9more than a year or two? Is that the right thing to be

10focusing on? If it is, if it is focused on in those

11industries, is it -- let me rephrase it. Is our focus

12on price misleading us and should we be focusing on

13other things?

14DR. GILBERT: Well, I do not view that -- you

15have asked a lot of hard questions. I do not think this

16is one of the hardest questions. I find it relatively

17straightforward in the sense that when you are talking

18about dynamic competition, Schumpeterian spiral

19competition, it is very much like thinking about entry

20analysis. There is some probability that entry will

21occur at some date. The question is how soon will it

22be, what will be the magnitude of it.

23There is also I think a legitimate question that

24even if entry is going to occur, is that going to

25neutralize the type of conduct we are concerned about,

112

1or does the conduct we are concerned about promote entry

2or retard entry? It is my view that these are questions

3that can be addressed within the context of the way we

4do antitrust analysis generally.

5Now, it is, of course, the case that in high

6technology industries, you are more likely to get very

7high price-cost margins, so you are more likely to be

8worried about market power, but it is often benevolent

9market power, and if it is benevolent, you should not be

10doing an antitrust case. So, it is more like magnifying

11the things that we are concerned about but not changing

12the qualitative way that in my view you should take them

13into account when you are doing an antitrust analysis.

14DR. KATZ: I have a couple of things and maybe

15tie it to Microsoft. I mean, one of the things to

16remember is when the Government was looking at

17Microsoft, when the Government was dealing with them in

18the mid-nineties, everybody pointed out, "Well, look,

19it's such a fast changing market, and yes, it is true

20that Microsoft dominates personal computer software

21today and Apple is a distant second, and there are these

22other things that aficionados use, but they really have

23not caught on, and now here we are 12 years later and,

24okay, all of that's the same." So, this whole thing

25about the fast-paced -- and certainly Linux is doing

113

1better than, you know, "the operating system" did, but

2sometimes we do tend to exaggerate the rate at which we

3think markets change and certainly relative to the pace

4of antitrust enforcement.

5But the other thing is, I think, Microsoft I

6think is an interesting case, and maybe it comes back to

7one of Larry White's points, that the Microsoft case, I

8think it is fair to say that both sides took a

9Schumpeterian view. Microsoft said, "Look, this is

10Schumpeterian competition, someone else could come

11along, they will displace us, because of network

12effects, you would expect the winner to get a very high

13share in operating systems, and so leave us alone,

14because that's how competition occurs," and the

15Government said, "Okay, look, this is Schumpeterian

16competition. If you guys didn't do bad things --" Greg

17is shaking his head. Now, it is true, sometimes the

18Government didn't say that, but I think the only

19sensible interpretation of what the Government was

20saying was, "This is a Schumpeterian market, and,

21Microsoft, you are trying to stop the next wave of

22innovation that would have displaced you," and I am

23saying that's somewhat like Larry's point about saying

24it is not just how well you are doing in some absolute

25sense, but whether there is a threat that you are trying

114

1to stop that would leave you worse off.

2So, I mean, I think in that one is the

3Schumpeterian view was consistent with saying we have to

4intervene, although it does shift what you look at.

5DR. CARLTON: All right. Well, we are about out

6of time, so I want to end with this, to get everybody to

7comment on this question.

8In light of all the difficulties and ambiguities

9with the use of market definition in section 2 cases, is

10it your view that we should still rely on it as we do,

11put less reliance on it, or go to a competitive effects

12and forget about market definition? So, why don't we

13just go down the table in order.

14MR. NELSON: Okay, I think I am halfway between

15your two extremes, because I think there are -- as I say

16in my slides, I think that there are organizing

17principles and things that the exercise -- the market

18definition exercise helps you understand what is going

19on and tell either a story or an analysis that is

20internally consistent, but that is not to say you have

21to do it in every case, and there are numerous cases

22where you may be able to expedite things by going

23straight to the competitive effects bottom line.

24MR. SIMONS: My take would be that the DOJ and

25the FTC should try to come up with something that

115

1focuses only on competitive effects, does not worry

2about market share, and then see what happens over time

3in terms of what they come up with and how operable it

4is. And if the thing really works, terrific, then try

5to get it into the courts, but not worry about that at

6the outset.

7DR. WHITE: Yes, we ought to be looking -- I

8have a feeling we are going to be having all of the

9divergence of opinion ranging from A to B. Yes, you

10ought to look at competitive effects more than we have,

11but I think there is still going to be a role for market

12definition. Think about the private suits, not the

13government suits, but the private suits that were

14brought against MasterCard and Visa, and these were --

15you know, the -- say take a WalMart case. This was a

16tying case, but they were not -- it was -- you could

17tell some stories about how if the tie was not there,

18there were -- there would have been more entry somewhere

19in -- in the credit card markets, but it was primarily

20the tie is preventing us merchants from doing something

21we would like to do.

22I am not sure a competitive effects analysis is

23going to tell you about market definition in that

24particular case. Of course, MasterCard and Visa were

25telling you, oh, all kinds of transaction media are in

116

1the market, you know, cash and checks and everything, we

2do not have market power, and they were actually trying

3to say it with a straight face.

4A market definition paradigm I think would help

5in that kind of case, and so yes, I think we still have

6need. I am hoping this is 1981, and next year, some

7smart people are going to come in with a useful

8paradigm.

9DR. CARLTON: Andy?

10DR. GAVIL: I think I agree with Larry. I think

11it still has a role to play, but I think as you stated,

12I think I would agree also that we over-rely on it, and

13I think somewhat complex is the problem of

14over-reliance. I think it can lead to both false

15positives and false negatives, but I think with the

16false positives, if somebody is found to have monopoly

17power, to some degree, you have the backstop of the

18conduct inquiry.

19My concern is because of the high process costs

20in trying to prove monopoly power in this -- as you

21described it accurately -- sequential model, you get

22false negatives, and there is no backstop to that. The

23case ends, and the court does not look at conduct, does

24not look at effects. So, I think this is an example

25where the over-reliance may actually increase the threat

117

1of false negatives more so than false positives.

2DR. GILBERT: I would join the chorus that we

3need more emphasis on competitive effects. A good

4example, not necessarily really in the section 2

5context, is the Oracle merger case where in my view

6there was some certainly interesting evidence, if not

7dispositive evidence, about competitive effects, but

8once Judge Walker determined that he could not define a

9market, he then concluded that there was no market, and

10the competitive effects were almost ignored in that

11case, and to me, it is like saying that I do not know

12exactly where downtown Los Angeles is, and therefore,

13there is not one. But I also can sympathize that if we

14did away with market definition completely, it could be

15highly problematic in leading to a lot of cases.

16DR. KATZ: Okay, well, I guess I will say, at

17the risk of sounding like Bill Clinton, it depends on

18what one means by doing market definition, and I think a

19lot of times what people mean is they mean applying the

20hypothetical monopolist test, they mean doing a

21concentration analysis, and they mean trying to come up

22with boundaries with certainty, and then slavishly

23applying that, and if you cannot meet that, you throw

24the case out.

25I think that way of doing things is surely

118

1wrong, but I think we also surely do want to do some

2sort of market definition exercise in the sense of

3identifying the competitors, and I think where we have

4come up short is what is the right way to do it in the

5middle in terms of I think we still do not have a very

6good sense of what is the right algorithm and the right

7approach in different situations.

8We have not mapped out, so, here is exactly

9where you could do the hypothetical monopolist test,

10here is where we need to do some alternative

11methodology. We do not have that, and I think the

12courts -- sometimes, the fact that we do not have that

13has become an obstacle to good decision-making, as Rich

14was just saying in the Oracle case, but I think the

15bottom line is we need to figure out a better way to do

16market definition, and that way we will recognize that

17it should not be taken overly seriously or applied too

18mechanically.

19DR. CARLTON: Thank you very much. I want to

20thank the people at the Department of Justice and the

21FTC who did all of the legwork in putting this together,

22and I am sure, although I have not checked with them,

23all of the panelists, myself included, thank Larry White

24for not including us in his slide of dumb quotes, and I

25want to personally thank everybody on the panel for

119

1coming and giving us the benefit of their substantial

2expertise. I have learned a lot, and I thank you all.

3(Applause.)

4(Whereupon, at 12:36 p.m., a lunch recess was

5taken.)

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

120

1AFTERNOON SESSION

2(2:01 p.m.)

3MR. WALES: Well, good afternoon, everybody.

4Thanks so much for braving the cold and the snow. I

5think we have a very exciting panel on tap for this

6afternoon. For those of you who were here for this

7morning, I hear it was very lively, and I am hoping that

8we can live up to that and be lively as well.

9My name is Dave Wales. I am a Deputy Director

10here in the Bureau of Competition at the Federal Trade

11Commission. I will be moderating today, along with Greg

12Werden, the panel. Greg is the Policy Project Director

13At the Economic Analysis Group at the Antitrust Division

14of the Department of Justice. That is his official

15title, and you will be hearing more from him shortly.

16Have you been elevated perhaps?

17DR. WERDEN: I have never even heard of that

18title.

19MR. WALES: Is it better than the one you have?

20DR. WERDEN: No, not really.

21MR. WALES: Not really? Sorry about that, Greg.

22Before we get into substance, it is my job to do

23a little bit of the housekeeping. First off, what I

24would like to do is on behalf of the FTC, to really

25thank our friends at DOJ in putting this together, and I

121

1think it has been phenomenal to date, and I am sure it

2will continue to be that way. I would also like to

3thank each of our five panelists today for, again,

4braving the weather and coming out, and I think we have

5got some great issues to discuss.

6Today and tomorrow, I guess today and tomorrow,

7we have the hearings on monopoly power, and I guess what

8I wanted to point out is that next month we will be

9turning to the issue of remedies, which should also be

10pretty interesting. Stay tuned for that. I guess what

11we typically do is post the schedule on each agency's

12web page, so look out for those.

13A couple housekeeping items. First, I guess we

14ask that people turn off their cell phones,

15BlackBerries, any other electronic devices that make

16annoying noises. Second, importantly, restrooms are out

17across the lobby. In case someone needs to use those,

18follow the signs, you cannot miss them.

19Third, a safety tip for everybody, I guess in

20the event the fire alarms do go off, please do not

21panic. Please walk towards the exit, and we will guide

22you to I guess a safe place across the street where we

23will gather, hopefully with warmer coats.

24Lastly, I guess the way we set this up is we ask

25that the audience please not ask questions, and we are

122

1going to have a lively discussion today, so you can look

2forward to that.

3Many of the prior sessions talked about

4obviously the conduct involved in section 2 challenges,

5and today, what we would like to talk about is a

6separate topic, which is, of course, monopoly power and

7defining markets when monopoly power has been alleged,

8and I think that is a pretty important topic, one that I

9think when the hearings kicked off, that Herb Hovenkamp

10and Dennis Carlton identified as being one of the two

11that they thought were probably the toughest and two

12that needed a lot of discussion. So, hopefully we will

13be able to accomplish that today.

14I think that is pretty much what I wanted to

15say, Greg. I don't know, maybe you want to give your

16title and anything else you want to say.

17DR. WERDEN: Yes. Hi, I am Greg Werden, Senior

18Economic Counsel in the Antitrust Division, Department

19of Justice. I just want to add my thanks to our

20panelists and the staffs of the two agencies for

21organizing this session.

22MR. WALES: Great, thanks.

23The way the format is going to work today is

24similar to what we have done previously and did this

25morning. First, we are going to have each presenter

123

1give about a 15 to 20-minute oral presentation, and then

2what we will probably do is take a break either in the

3middle of that or towards the end of that, we will see

4how long things go, after which we are going to have a

5moderated discussion where we will give each panelist an

6opportunity to respond to the other panelists and also

7for Greg and I to pose some questions and some

8principles that we think we might be able to move

9towards convergence on.

10With that, I guess what I would like to do is

11introduce Simon, our first speaker. Simon is a partner

12and co-founder of RBB Economics. He has been advising

13clients on competition policy issues since 1991 and has

14particular expertise in applying empirical techniques in

15the context of merger investigations. In addition to

16his private sector work, Simon has been seconded for a

17short period of time to the German Federal Cartel

18Office, where he gave a series of seminars on use of

19economics in competition law. Simon has published

20widely on virtually all aspects of competition law

21economics and is a regular speaker at competition law

22conferences. He is a co-author of The Economics of EC

23Competition Law and has worked on several hundred

24competition law matters spanning virtually all sectors

25of the economy. Thanks, Simon.

124

1MR. BISHOP: Long intro, and you forgot to say I

2am from Europe. I am up first, and given that it was

3probably a lively session this morning, I think probably

4the reason I have been chosen to go first is because my

5topic this afternoon is probably the dryest and most

6technical, which is my remarks are really going to

7concern the Cellophane fallacy and what implications

8that has for a structural approach to assessing

9monopolization or what we Europeans talk about as an

10abuse of dominant position.

11I am also going to say that my remarks are also

12sort of Euro-centric in the sense of this really

13reflects my experience of EU cases and European national

14competition law cases and not really the U.S. case law;

15however, given that we are all in this facade, one might

16say, or claims that there is increasing convergence

17between Europe and the U.S., I hope that some of these

18remarks will actually carry over to U.S. antitrust.

19Now, in order to give this some sort of context,

20in Europe, as in the U.S., we are engaged in an ongoing

21reform of Article 82, which is the equivalent of the

22monopolization act, and last year, the European

23Commission, of which Miguel was heavily involved, issued

24guidelines on how to reform Article 82 and to move the

25current system away from a form-based approach to a much

125

1more effects-based approach.

2Now, that is all very admirable, but it seems to

3me that within these guidelines of reform, there is the

4elephant in the room which no one really wants to talk

5about; namely, the concept of dominance. Really,

6dominance is also based on structural market share. In

7Europe, we have the two-step process, whether a firm is

8found to be dominant and then whether, if that firm is

9found to be dominant, whether that behavior constitutes

10an abuse of a dominant position.

11Now, as I said, all the focus has been on the

12second step, but the problem is is that within Europe,

13certainly how the courts have interpreted dominance and,

14indeed, some of Miguel's colleagues in the Commission

15also, is that if you are dominant, then any behavior

16which affects or harms competitors is almost deemed to

17necessarily harm competition, and if you take that

18approach, then that really means there is no role for an

19effects-based analysis within European antitrust under

20Article 82.

21It also means that the dominance, and therefore

22the market share calculations and market definition, are

23absolutely paramount in the whole assessment. Now, we

24all know that from the sort of 1982 U.S. Merger

25Guidelines, there has been pretty wide acceptance that

126

1the hypothetical monopolist test or the SSNIP test has

2provided the appropriate framework for assessing and

3defining relevant markets. We also know that, even

4though we have the framework and that is well accepted,

5in actual individual cases, it is actually quite hard

6sometimes to actually implement that test. Actually, in

7monopolization cases or abuse of dominance cases, things

8are even more difficult because of the existence of the

9Cellophane fallacy.

10Now, what is the Cellophane fallacy? What are

11the problems? Well, in a merger context, which is where

12the SSNIP test or hypothetical monopolist test was first

13proposed, we are interested in what has the merger

14changed? Is it going to relax competitive constraints

15at prevailing price levels? Now, that has an important

16implication, because that says we can use existing data,

17observed data, to assess the strength of existing

18competitive constraints between products supplied by the

19merging parties.

20However, when we talk about monopolization

21cases, in many cases -- and some might argue in all --

22the relevant issue is not whether the prices can go up

23even further from prevailing levels, but actually, have

24prices already been increased above competitive levels?

25Now, the important implication of that is that using

127

1observed data will tend to overstate the competitive

2constraint, because as we know from the famous DuPont

3Cellophane case, is that even a monopolist, if you put

4up prices far enough, something is going to start

5looking like an effective substitute at some point,

6because even monopolists face some constraint. So, the

7real issue here I think from the Cellophane fallacy is

8what implications does it have for the use of empirical

9analysis?

10Now, that is a case of sort of, well, what do we

11do about this? We know that the Cellophane fallacy

12 exists, and we know that that has a big impact on how we

13can interpret and use existing data. Well, there are a

14number of approaches which have been put forward to try

15and address the Cellophane fallacy. One which has been

16put forward in a number of cases both by the European

17Commission and some national competition authorities in

18Europe is to say, "Well, the hypothetical monopolist

19test is only one way of defining a relevant market."

20Well, the question or the problem with that sort

21of line of argument is, they do not actually tell you

22what the alternative ways of defining a relevant market

23are, and what is good about the hypothetical monopolist

24test is it is forcing people to at least think about the

25scope for demand-side substitution and supply-side

128

1substitution, and if those two things are not part of

2the approach of defining a relevant market, it seems to

3me that, indeed, any other alternative approach is

4useless for antitrust purposes.

5The second approach to trying to deal with the

6Cellophane fallacy is, "Well, let's just recalculate

7everything from the competitive price." Well, great

8idea, but if we knew what the competitive price is, then

9we would not need to be defining what the relevant

10market is. We could just say, "Well, we observed that

11Firm X is charging 10, we know the competitive price is

125; therefore, there must be some sort of exercise of

13market power going on." But that is not how the real

14world works.

15So, as a slight anecdote here, I was reading in

16some of the trial transcripts in the Microsoft case, one

17of the economists I think it was for the DOJ was asked,

18"Well, what is the competitive price that Microsoft

19should charge?" The answer was, "Lower than they

20currently charge," which seems to me sort of just

21demonstrate the difficulties of actually re-adjusting

22what the competitive price is. So, that is not going to

23get us anywhere.

24The third approach is, "Well, let's just ignore

25the Cellophane fallacy; pretend it does not happen."

129

1Well, again, that is not going to work, because if you

2ignore it and just apply empirical analysis, you are

3going to tend to define relevant markets too widely, and

4therefore, not capture some market power when we should

5be capturing it.

6The fourth approach which has been proposed,

7which is, "Well, let's do away with market definition

8altogether. It is difficult -- we have got the

9Cellophane fallacy, you know, we are very smart economic

10professors or consultants, and we can just sort out --

11you know, market definition is just an interim step. We

12can go straight to the answer." Well, personally, I am

13a bit more humble than that, and I think if we see the

14relevant market definition and the structural analysis

15for what it is, i.e., an intermediate step, I think it

16is important that we keep that step.

17Secondly, if you do away with it, it actually

18introduces real scope for ad hoc analysis. There is a

19real -- we know that particularly in the areas of

20exclusionary abuses or exclusionary power, trying to

21discriminate between behavior which just merely harms

22competitors and is therefore procompetitive from that

23which harms competitors and drives them out of the

24market and leads to harm to consumers is very, very

25difficult, and really, the market definition structural

130

1screen provides a good touchstone to prevent people from

2adopting "I know abusive behavior when I see it."

3So, I think that sort of a summary of this is

4really we are stuck with the hypothetical monopolist

5test, the SSNIP test, as a framework for thinking about

6how we define relevant markets, and we are also stuck

7with the Cellophane fallacy. We need to accept that it

8exists. So, what are my proposed implications for the

9sort of policy?

10Well, a structural analysis still can provide a

11very useful filter, and even recognizing the existence

12of the Cellophane fallacy, I think we can go through a

13number of steps, that we can define relevant markets

14which are consistent with the basic principles of the

15hypothetical monopolist test. So, if someone proposes a

16relevant market and that it does not seem to be

17consistent with the principles of demand-side

18substitution or supply-side substitution, then it is not

19a relevant market. So, I think even just using the

20SSNIP test as a thought process can actually provide a

21useful discipline on how to define relevant markets.

22Secondly, we know the Cellophane fallacy exists,

23but if the parties are arguing for a wide market

24definition, then they at least ought to be able to

25demonstrate that at prevailing prices, there is

131

1substitution. Now, that means that it does not stop

2with saying, "Well, the price has already been increased

3above competitive levels and is subject to the

4Cellophane fallacy," but at least they should be able to

5show that at prevailing prices, there is a competitive

6constraint between product A and product B if they are

7arguing they are in the same relevant market.

8The third element I think is we can look at

9product characteristics in the marketplace, but again, I

10think we should be careful about how we look at that,

11and this really goes back to my first point, which is

12consistency with basic principles, is it is not saying

13that these two products are not in the same relevant

14market because they look different or have different

15product characteristics. We are saying they are not in

16the same relevant market because the differences in

17product characteristics imply that demand-side

18substitution or supply-side substitution is unlikely.

19The fourth element is that there are some cases

20and there is some evidence which is not subject to the

21Cellophane fallacy which we can use to discriminate

22between competing claims, and as always, there is a

23paper by Greg Werden, who addresses this, and I think it

24was from about 2000.

25The second policy issue is, well, we have

132

1defined the relevant market, we have calculated some

2market shares, and clearly we need to put that into

3context. We need to take into account the scope of all

4barriers for entry, expansion, the scope of buyer power,

5whether the market is subject to bidding competition,

6and also general market dynamics.

7My final comment was really, well, let's be

8humble here, because we can go through all of these

9steps, but in a lot of cases, the available evidence

10will not allow us to discriminate between the wide

11market definition which the parties are putting forward

12and the narrow market definition which the agencies are

13going to be putting forward. Everything may be

14consistent with the basic principles of the hypothetical

15monopolist test, the parties can show that at prevailing

16prices there is substitution and so on and so on, and

17where you have got these two competing potential market

18definitions, sometimes that will not be a problem,

19because the market shares in both of those may be low,

20and then unlikely to have market power. Alternatively,

21market shares in both of those could be high, and then

22that is not really a problem, because the market power

23is reasonably high. The difficulty or the problem,

24potential problem, is where in one market, the narrow

25market, the firm has a high market share, and in the

133

1wide market, it has a low.

2It seems to me when you are in those situations,

3all it says is, well, then we really need to have some

4pretty good evidence and examination of the business

5conduct, and this I think brings me back to where we are

6in Europe, is that a lot of times in Europe, with the

7current situation, the business conduct is not assessed

8on the market effects, but actually on the form of the

9business conduct. So, the reform in Europe is certainly

10going in the right direction in focusing on the form,

11and that is the end of my comments.

12Thank you.

13MR. WALES: Thanks very much, Simon.

14(Applause.)

15MR. WALES: Our second speaker is Miguel de la

16Mano. Miguel joined the European Commission in 2001 and

17is currently a member of the Chief Competition

18Economist's Team. He carries out economic analysis in

19mergers and commercial practices by dominant companies

20and their impact on the competitive structure of the

21markets. He is also responsible for drafting

22guidelines, setting the Commission's analytical

23framework in these areas, a key area. He completed

24graduate studies in economics at The Institute For the

25World Economics in Kiel, Germany and The European

134

1Institute at Saarbrucken University, Germany. He

2conducted his Ph.D. research at Oxford.

3With that, Miguel? Thanks.

4MR. de la MANO: Thank you very much. It is

5definitely a pleasure and also a great honor to

6participate on this panel today together with so many

7distinguished and well-experienced practitioners.

8I will try to contribute to this issue basically

9by offering a view or an assessment of the way in which

10dominance or the role that dominance plays today in

11competition policy assessment in Europe and which, as

12you know, is enshrined in Article 82, which is the

13equivalent of section 2 here in the U.S.

14As you also know and as Simon has remarked, the

15Commission is in the process of reviewing its policy in

16the area of Article 82, and like every type of reform,

17it is somewhat case-dependent, and we are constrained by

18case law and case practice; however, we believe that the

19time is right basically to align the implementation or

20the enforcement of Article 82 to current thinking and

21current economic knowledge, and, of course, to a more

22modern analytical framework.

23So, I will basically start by making a somewhat

24obvious remark, yet actually crucial, which is that in

25the context of the analysis of monopolization in Europe,

135

1dominance is a necessary condition. That is how the

2system has been set up, and the EU Treaty actually

3prohibits single-firm conduct that harms consumers only

4when undertaken by a dominant company, and normally, to

5ensure the efficiency of the decision-making process,

6this also means that the first step of the analysis is

7to establish whether or not a single firm actually is in

8a dominant position or not. It is not a must, but that

9is just the best way forward. If a single firm is not

10dominant, then there is no need to proceed any further.

11At the same time, a somewhat more subtle point,

12this also rules out what in the U.S. is attempted

13monopolization. If you are not dominant in the first

14place in the EU, basically there is nothing you can do

15that will violate Article 82, and I think this is an

16important point, because it somewhat dispels the myth

17that in the EU, there is a serious concern or serious

18worry with type II errors; namely, false acquittals. I

19think personally that is not the case.

20But what are the reasons for this institutional

21setup? I can think basically of two primary reasons.

22Number one is to provide legal certainty. Surely it is

23better for firms to know in what circumstances they may

24be liable to and they are obligated to. There is also

25another reason, which is that we should not forget, it

136

1was member states that have delegated the powers to a

2rather independent body, namely the European Commission,

3to enforce competition policy in their name, and when

4delegating such powers, member states want some

5assurances that these powers will not be abused, and

6therefore, forcing the Commission to start off by

7assessing whether or not a firm is dominant imposes some

8sort of discipline, which understandably was necessary

9for member states to delegate such powers.

10However, unfortunately, despite the best wishes

11of everybody at the time, maybe 30 years ago, it has not

12fully worked, and I think there are three reasons why it

13has not fully worked, which I would like to share with

14you and hopefully also in doing so contribute to the

15thinking that is taking place here in the U.S. with

16respect to monopolization.

17The first reason why they do not work is the

18concept of dominance is somewhat elusive. The member

19states put it into Article 82; however, no definition

20was actually provided. That was left for the courts to

21develop one over time.

22However, as is normal, the courts were actually

23reacting to cases that were brought to them, and they

24were not necessarily thinking in the abstract, well,

25what is it that dominance should mean? How should it be

137

1defined? But instead, we are reacting to the

2circumstances.

3Of course, it became increasingly complex and

4increasingly difficult to understand exactly what

5dominance is as time went by and European courts were

6issuing rulings where the concept of dominance was

7mentioned or in some cases defined.

8Of course, what happened ultimately is that,

9before the Commission, it became increasing difficult to

10identify what is dominance, and therefore, the more

11difficult it was, the more elements which would normally

12go into the competitive assessment creeped into the

13assessment of dominance, up to a point where it seems,

14at least to me, that as Simon pointed out before,

15assessing dominance became an end in itself to the

16extent that once dominance had been established, it was

17not just a necessary condition but almost sufficient for

18a finding of abuse.

19Now, I think that these three concerns can be

20corrected, and this is, of course, the rationale for the

21review process, and I would just like to share with you

22the three ways in which I think this can be done.

23So, first of all, again, a rather obvious

24statement, but somewhat important in a context where

25European courts have said that the dominant firms have a

138

1special responsibility, whatever that might mean,

2dominance should be defined or equated with substantial

3market power. Now, of course, all firms have some

4market power, but most of them have very little, and

5accordingly, the relevant question in antitrust cases

6should not be whether market power is present or not --

7it always is -- but whether it is important, that is,

8whether it is substantial.

9In going back to the sort of most established

10definition of dominance by the ECJ, dominance is said to

11be a situation where a company has the power to behave

12to an appreciable extent independently of its

13competitors, customers and ultimately its consumers, and

14a close look at this definition suggests, indeed, that

15dominance can be equated to significant market power,

16and this is because a firm is dominant if its decisions

17are fairly insensitive to reactions of competitors and

18customers. That is what the "to an appreciable extent"

19actually means. Of course, no firm is fully independent

20of customers and competitors, that we know from economic

21theory, but to an appreciable extent, it might well be.

22The measure of this sensitivity, of the

23sensitivity to the actions of others, is given by an

24elasticity, which is, again, the other side of the way

25that economists would measure market power in practice.

139

1So, we end up with a situation where to behave

2independently to an appreciable extent can be definitely

3equated to an ability to significantly and profitably

4durably increase prices, and therefore, there should be

5no more debate about what is dominance, just substantial

6market power.

7Now, how is this substantial market power to be

8established? Well, again, this is not new to anyone,

9but I would argue that first market shares have to be

10significant, and significant in two senses. First, they

11have to be significant in that they must be important,

12high, but also significant in that they are actually

13providing a good proxy for the relative insensitivity of

14the single firm to the actions and reactions of its

15competitors and customers. There is, again, a good

16paper by Greg Werden which talks about assigning market

17share and how difficult this process actually is.

18The second point is that barriers to entry and

19expansion have to be significant, and by this I would

20like to emphasize that we mean in the absence of the

21conduct, not barriers to entry in general, but in the

22absence of the conduct, if the conduct itself actually

23increases barriers to entry or barriers to expansion.

24Now, that is part of the anticompetitive effects of such

25conduct, and therefore, it should not be seen as an

140

1element that plays a role in establishing dominance.

2Of course, there are other elements like

3dynamics of the market, there should be no technological

4leapfrogging, and buyer power, it cannot be shown that

5the customers have very little countervailing buyer

6power.

7Now, I will try to make here a rather

8provocative statement, but in my view, the acid test,

9the way to ensure whether a company is dominant or not,

10is to ask, well, is it the most efficient in the market?

11Because if it is, it is likely to have high market

12shares, it is likely to be very difficult to enter

13successfully and profitably, and it is also going to be

14very difficult possibly to leapfrog.

15However, one might argue, well, isn't this just

16the old efficiency offense? Well, I do not think this

17is an offense, because I personally think there is

18nothing wrong with being dominant. There is no offense

19in being dominant, and companies should not feel that an

20assessment of dominance actually implies that this is

21going to lead to a finding of anticompetitive behavior

22on their part. Quite the opposite, a finding of

23dominance should in most cases just mean that they are

24probably the most efficient company out there.

25This takes me to the final point, which is that

141

1dominance is not only a screen. It is not an end in

2itself. It is just a screen to try and filter out, as

3Simon was saying, those situations where there might be

4scope for significant harm to consumers resulting from

5certain conduct from other situations where this is very

6unlikely to happen.

7Now, it is clear that if a practice is shown to

8be anticompetitive, the firm must be dominant, but

9proving that a practice is anticompetitive is hard, and

10it takes a lot of time and resources. Therefore, it

11seems like assessing dominance can play a very important

12role in acting as a screen, and it is also a screen that

13bites. It bites because large firms may not necessarily

14be dominant if innovation is taking place at a rapid

15place, if there is fierce competition between large

16players, for instance, in the concept of bidding

17markets, or if there is strong disciplining by potential

18entrants or customers.

19Now, I am just going to briefly go into a

20non-hypothetical example, which unfortunately I am not

21allowed to get into further details of the market, but

22where actually I will be able to show to you that the

23Commission takes very seriously the dominance screen and

24it actually works in practice.

25We had a case not long ago where the defendant

142

1had very high market shares in a homogenous good market,

2above 60 percent. There were very important and

3significant barriers to entry, like large overcapacity

4on the part of the dominant company, declining demand,

5high fixed costs to establish new facilities, but also

6strong learning effects in the process. It was common

7practice in the industry to use very long-term

8contracts, which, of course, we argued would limit

9customer switching, and not the least of which the

10defendant seemed to be in a very strong position to fend

11entry given that it had the broadest product range and

12the largest financial resources. So, with this criteria

13on the table, one would very easily conclude that this

14company is actually dominant.

15Well, actually, the Commission concluded it was

16not, and this was on five counts. First, there was

17significant buyer concentration. The top three

18customers took 70 percent of the market. There was

19product homogeneity, which allowed them to switch

20suppliers without incurring significant switching costs,

21and buyers, indeed, have dual sourcing strategy to shift

22volumes between suppliers with little transaction costs.

23Rival suppliers had significant overcapacity which they

24could use to expand, and therefore, there were no

25barriers to expansion. It was also found that the

143

1competition mechanism was bidding for large and very

2occasional contracts, just every few years.

3So, I would just like to conclude with two

4remarks, one on market shares and one on market

5definition, linking it to what Simon has said. First,

6on market shares, it is, often said that there should be

7a bright line safe harbor, and also that, only firms who

8are market leaders can ever be dominant. I think the

9latter makes no economic sense, and this is clear given

10the application of unilateral effects in the area of

11merger control, and, of course, at least in the context

12of European competition policy, the dominance concept

13plays a role both in mergers and antitrust, so they are

14interlinked.

15However, bright line safe harbors do make sense;

16however, I believe the threshold should be set rather

17low, and this is for four reasons. First of all, rivals

18might be constrained. For example, in the electricity

19industry, this happens very often. You might have

20strong multi-market presence, like in the airline

21industry, if you have one company who is number two in a

22number of routes and the number one company in each one

23of the routes is a different one, one can argue that

24this company who was number two everywhere is actually

25more dominant or has more significant market power given

144

1this multi-market context than anyone who just has a

2leadership position in one individual route.

3Market leaders are more constrained by

4regulation than nonleaders, and that can be the case in

5certain industries, such as telecoms, and the leader may

6be more constrained by close substitutes or by new

7entry, for example, in the case of pharma. There was a

8case of AstraZeneca in the EU not long ago where this

9was clearly an issue.

10So, what are the policy implications for not

11arguing that only if you are the market leader, you can

12be dominant? There are at least two. One is that for

13consistency, I will just mention unilateral effects in

14mergers, but also, to leave the door slightly open for

15attempted monopolization in the EU, in the EU policy.

16Then just one very short and final remark on

17market delineation, which I will just start by saying

18that I agree with everything that Simon has said, but

19unfortunately, even though I think we ought to be humble

20and I definitely agree with that, the EU Commission is

21forced to be arrogant, because in a sense, we are

22obliged to take decisions. We have to say what we think

23about the market. We cannot leave definitions open. We

24have to say whether we think it is narrow or we think it

25is wide, whether or not we win the case, and this is a

145

1problem.

2This is a problem because we cannot just say,

3well, you know, let's ignore the Cellophane fallacy or

4let's think about the Cellophane fallacy as something

5that plays a very significant role and there is nothing

6we can do about it, so we just be humble. That we

7cannot afford to do.

8However, I think we do not have to lose all

9hope, because when thinking about the role or the

10assessment that dominance plays, particularly thinking

11of dominance as a screen, I think that even if we

12recognize that the hypothetical monopoly test, the SSNIP

13test, is, indeed, a useful conceptual framework to

14identify competitors that are constraining a single

15firm, the assessment of dominance actually goes a step

16further, and not just ask the question, well, which are

17the firms that are there constraining the incumbent, but

18actually asking, well, how much are they constraining

19the incumbent?

20So, in trying to figure out how much is the

21incumbent being constrained or the defendant being

22constrained, we can also have a good glimpse into which

23other firms that are part of that particular market, and

24therefore, market delineation can in some cases -- not

25always, but in some cases -- be a by-product of the

146

1dominance assessment, and this obviously simply reflects

2that market definition is a means to an end, and what

3the real issue is is market power.

4Thank you very much for your attention.

5(Applause.)

6MR. WALES: Thank you, Miguel.

7Next up we have Tom Krattenmaker. Dean

8Krattenmaker is currently Of Counsel in the Washington,

9D.C. office of Wilson Sonsini Goodrich & Rosati, where

10he focuses on antitrust, telecommunications and trade

11regulation issues. Immediately prior to joining Wilson

12Sonsini, Tom was an attorney in the Federal Trade

13Commission's Bureau of Competition, Office of Policy and

14Coordination, where I had the pleasure of working with

15him for too short a time, but really enjoyed my time

16working with him. In that role he principally served as

17legal adviser to the bureau directors and to attorneys

18investigating and litigating antitrust cases and advised

19on several Bureau and Commission public reports.

20Previously he served as senior counsel in the Department

21of Justice's Antitrust Division and held positions at

22the Federal Communications Commission, including Chief

23of Telecommunications Merger Review and Director of

24Research and Co-Director of the Network Inquiry Special

25Staff. Tom has spent more than 30 years in legal

147

1education. He was a Professor at the University of

2Connecticut, Professor and Associate Dean at Georgetown

3University and the Dean of William & Mary School of Law.

4Thanks, Tom.

5MR. KRATTENMAKER: Hello. I'd like to begin by

6thanking Dave and Greg for giving me this monopoly

7platform for 15 or 20 minutes and am particularly

8appreciative for you surrounding the platform with the

9entry barriers with your declaration that there be no

10questions from the audience.

11I also would love to be able to take refuge in

12the defense offered by Miguel that he was forced to be

13arrogant. The problem is that there is at least one

14member of the audience I see here who was one of my law

15school classmates, so he knows darn well that I have

16chosen to be arrogant. So, what I would like to say

17honestly is that I am going to sound more assured about

18my views than I am. I have asked that on my tombstone

19they write something like, "Often wrong but never in

20doubt," so if you really do not like what I am saying,

21say, "Oh, Tom's just trying to be provocative again."

22Dave can tell you that he has said that many times and

23enabled himself to get home without going home in a funk

24or thinking that they have to let me go the next day.

25The other thing I want to say at the beginning

148

1is that aside from the fact that I am quite honestly

2flattered to have been invited to join this group, I am

3more interested in trying to respond to questions than

4saying anything in particular, so please do send up a

5flag after 10 or 15 minutes, and I will just stop. I

6have four points to make, and if we only get three of

7them out, I am sure I will be able to smuggle the fourth

8one in somewhere later on.

9I am speaking largely off a text -- I am not

10going to read it to you -- of an article that I

11published with a couple of really outstanding antitrust

12lawyers and scholars, Bob Lande and Steve Salop in the

13Georgetown Law Journal in 1987 called Monopoly Power and

14Market Power in Antitrust Law. It turns out that even

15though that is 20 years ago, I think it is still right,

16so if you want to have a look at that, that is where I

17am coming from.

18The first point I wanted to make I think is one

19where we could say I am preaching to the choir, so I

20will go through it quickly, but it is not a trivial

21point, and that is, what do we mean by market power? I

22think my sense is that in this room, we are all

23co-religionists; that is, we all think that market power

24is the ability to price profitably for a significant

25period of time above the competitive level for that

149

1market.

2I might just stop to observe that that has

3hardly been the history, the unbroken history, of

4antitrust. We have had many other tests of whether

5something is anticompetitive or not. Justice Douglas

6once opined that a merger was anticompetitive because it

7would lead to moving the corporate headquarters of the

8firm from a small town on the West Coast to big, bad New

9York City. Justice Black once told us that a merger was

10anticompetitive because there would be fewer

11single-store grocery stores in Los Angeles.

12We seem to have, at least at this point in time,

13a consensus that we have an economic concept of market

14power, and it is the ability profitably to price above

15competitive levels for a significant period of time, and

16I know that for crystallizing that definition, one of

17the people we really have to thank for that is Greg.

18Another question that I think I was asked to

19address is what is the difference between monopoly power

20and market power? Now, syntactically, "monopoly" sounds

21like -- it says, well, how can you have monopoly power

22unless you have complete control over a relevant market?

23You must have to have a 100 percent share of a relevant

24antitrust market that is surrounded by entry barriers.

25I suppose -- I do not know, I didn't look at a

150

1dictionary, I should have -- you could say that is it.

2That is certainly not the case law definition,

3and I think, again, within the current antitrust

4community, nobody would doubt that. I think the right

5answer is that it is the same as market power. There

6are some cases out there where there is noise in the

7opinions that suggests that there is some kind of

8difference between market power in monopoly power, but

9it does not seem to make any sense. That is, market

10power and monopoly power and antitrust law are and

11should be synonymous. They can occur in various

12degrees, but they are qualitatively the same.

13Of course, the analogy that came to my mind was

14basketball. I am supposed to leave here tonight and

15play in a basketball game. Yes, you can tell by looking

16at me I am our team's power forward, and monopoly power

17and market power are the same in the same sense that a

18shot is the same. It goes in or it does not go in. It

19goes in the basket or does not go in the basket.

20Now, some are worth one point, some are worth

21two points, some are worth three points. You could have

22lots of market power or little bits of market power, but

23it is the same thing. It is not like being tall. You

24could be very tall or not very tall or sort of tall,

25but -- no, this is like shots in basketball.

151

1I guess I have waited long enough for some wag

2to say, "Well, what about goal tending?" The answer to

3that, "If you figure that out, you have got the whole

4rationale for the per se rule," but you did not want me

5to talk about per se rules? Okay, I will go to the next

6thing.

7Market power, monopoly power, are really the

8same thing. They are qualitatively the same thing. We

9mean the same thing by it. It is helpful to distinguish

10between I think two types of market power. The DuPont

11formulation that is quoted a lot is that monopoly power

12is -- DuPont is the same one that introduced the

13Cellophane fallacy -- the power to control prices or

14exclude competition.

15That sounds like it is two things, doesn't it?

16Power to control price or the power to exclude

17competition, arguing it is really the same, but the

18reason you see that or the reason you sometimes see this

19noise in the cases about there are these different

20things is that the intuition the judges have is that it

21might make a difference what kind of market power you

22have or how you are exercising it. We put names on them

23in the paper, but I do not have to use names.

24One way to exercise market power is by

25restricting your own output, cutting your own output,

152

1sometimes in concert with that of other people in the

2market who are happy to join with you. I would call

3that collusive market power. We called it Stiglerian in

4honor of George Stigler because it is the kind of market

5power he wrote about.

6The other way that one might exercise market

7power is not by restricting one's own output but by

8restricting rivals' output, letting market output

9decline and letting your price rise through no

10restriction in your own output. That I would call

11exclusionary market power or market power obtained or

12exercised by exclusionary means. In the paper we called

13it Bainian, after Joe Bain, an economist who had written

14a lot about entry barriers and exclusionary issues.

15My second submission to you is that -- while

16market power and monopoly power are the same kind of

17concept and that we do have a notion of what it means

18that we tend to agree on -- that it will help us if we

19distinguish between whether we are talking about

20collusion or exclusion, or if you like the little

21labels, Stiglerian or Bainian market power. Now, why is

22that the case?

23The article is about market power in antitrust

24law. We are here talking about section 2. So, let me

25try to explain with respect to section 2 cases, monopoly

153

1or attempted monopoly cases, why it might make a

2difference to think about the source of the market power

3or the type of market power that we are talking about.

4Point one, market and monopoly power include the

5power to keep prices from falling to competitive levels.

6I do not think we forget this a lot. We usually just

7say it is the ability to raise prices, but when

8confronted with the ability to keep prices from falling,

9we usually recognize that as market power, but you

10should in case you did not.

11If you had a horse and buggy industry that was

12perfectly competitive, a hundred firms each producing 1

13percent of all horse and buggy output, if they managed

14to exclude one firm and that firm was the first firm

15that was going to produce the automobile, they have

16nevertheless exercised market power even though it was a

17completely competitively organized industry. It is

18market power. It is market power to be able to keep

19prices from falling to competitive levels. Fencing out

20rivals who have the ability to bring in a new technology

21or simply be able to produce products at a much lower

22cost is an exercise of market power.

23Secondly, and connected to that, I believe it is

24not correct to insist on a threshold showing of market

25power if the conduct complained of is acquisition and

154

1exercise of market power by excluding rivals. If you

2are talking about a section 2 monopoly case, and you are

3saying what they are going to do is restrict their own

4output and profitably price for a long time above

5competitive levels, it is probably correct for the

6reasons that Simon and Miguel have already talked

7about -- although it was not the principal purpose of

8their talk, but they explained it -- to insist on some

9kind of threshold of market power. It is kind of hard

10to imagine how a firm with only 40 percent of the market

11can restrict its own output profitably for a long period

12of time and thereby price above competitive levels all

13by itself.

14That is not true if you are talking about

15exclusionary behavior. Exclusionary behavior can create

16the market power. You do not necessarily need to

17already control a market in order to be able to engage

18in exclusionary behavior that winds up creating

19effective market power. You might still have a

20threshold.

21If you do the math, he said -- referring to

22other people because he is not a mathematician -- but if

23I understand the literature right, the raising rivals'

24costs literature, you may want to have kind of a

25threshold that does have to do with size, like relative

155

1disparity in size. It is unlikely that a firm that has

2got 5 percent of the market is going to be able to,

3through exclusionary tactics, drive out rivals who are

4two and three times as big if it were the smallest firm

5in the market, but to say that one needs to have a kind

6of a dominant firm presence before one could ever be

7tagged with the offense of monopolization under section

82 is just not right unless you are -- because you appear

9to be forgetting what I've called Bainian or

10exclusionary market power.

11A third lesson from this that is relevant to

12section 2 cases, I think, is that it seems to me that we

13frequently hear it said that the mere exercise of market

14power is not prohibited by antitrust, and I think there

15is a statement to that effect in the Trinko decision by

16the Supreme Court a year and a half ago. Indeed, if I

17recall correctly, Justice Scalia not only said it, but

18he said you sort of welcome that kind of behavior

19because it is a signal to people to come enter the

20market. There are high prices. You can come in and do

21something. There is nothing wrong with exercising

22market power if you have got it. The question is how

23you got it.

24Well, once again, I think that probably is true

25for collusive or Stiglerian market power. It is

156

1probably correct that a firm that has got 90 percent of

2the market, if they acquired it lawfully, to say that

3when they raised -- when they restrict output and raise

4price, that is an antitrust offense, that is a very

5tough nut to crack, a very hard argument to make,

6because what are you going to do about it? What is your

7remedy? How are you going to decide whether they raised

8price too high?

9But if you have in mind the possibility that you

10might be talking about a section 2 case based on

11exclusionary market power, it is just not right, because

12you would attack the exclusionary act, and sometimes you

13can distinguish between the exclusionary act and other

14types of behavior with respect to the market power. The

15most obvious example would be, I think, if I could build

16off Miguel's example.

17He gave that terrific example of the industry

18where, when you first looked at it, you might think

19dominance, and then you find all these other aspects

20here. If this had been an industry in which the issue

21had been an exclusive dealing arrangement that was

22having the effect of denying vital inputs to rivals, not

23only does it not require, in order for that to be a

24successful antitrust strategy, that the firm have a

25dominant share to begin with, but it is also not the

157

1case that if it has got a position of dominance, if it

2is a monopoly, that then the mere exercise of monopoly

3power is permissible. It is not the case at all, and,

4indeed, that is an area where I think the European law

5is ahead of ours, because it clearly reflects that is

6the abuse of dominance.

7Finally, I had one more. It is relevant to

8antitrust law, but it is not relevant to the Federal

9Trade Commission or the Department of Justice. One of

10the lovely things about working for the -- there are

11many nice things about working for the FTC and the

12Department of Justice that I think, you know, the most

13are that you always think you are on the right side and

14you have these wonderful people to work with, but

15another thing is you never have to worry about standing,

16because if you see something wrong out there, you can go

17after it.

18Out in the private sector, you have got to have

19standing, and I think another lesson that you learn from

20distinguishing between these types of market power or

21these types of means of acquiring or exercising market

22power is relevant to competitor standing. Competitor

23standing should not be an issue in most section 2 cases

24involving Bainian or exclusionary market power, because

25the action is actually targeted at the competitor.

158

1On the other hand and for the same reason,

2consumer standing, even though the person who may suffer

3the effects is the consumer, consumer standing may be

4quite risky, both because there is a more direct subject

5of the harm, that is, the competitor, and therefore,

6there is the risk of double damages, and so following

7things like Associated General contractors and Illinois

8Brick, consumer standing in monopoly cases may be

9difficult, and consumer standing in attempted monopoly

10cases I don't think the Supreme Court has ever addressed

11it, but there is a growing body of case law in the lower

12courts now that consumers just do not have standing to

13bring attempted monopoly cases.

14Most section 2 cases are these Bainian

15exclusionary power type, and you can see the reason for

16that is that the harm is not directed at the consumer,

17and if it is merely an attempted monopoly, there is no

18follow-through on the part of the consumer.

19Well, enough for that commercial. Again, I have

20tried to suggest really just two things to you. One is

21that we have a concept of market power that we are at

22least presently comfortable with, and that is no

23different from the notion of monopoly power for the same

24reason that we are comfortable with the conception of

25market power. We are talking about what is the goal of

159

1antitrust, what are we trying to target our antitrust

2rules to do, and it is to prevent undue concentrations

3of power where power means the ability to profitably

4price above competitive levels for a significant period

5of time.

6Secondly, that it will help to keep your eye on

7the ball, to dig a little bit deeper and say, are we

8talking about market power that is going to be

9manifested by restricting one's own output, either by

10one's self or in concert with one's competitors, or are

11we talking about market power that is going to be

12manifested or acquired by driving one's rivals out of

13the market and thereby gaining the power to exercise

14higher prices without necessarily restricting one's own

15output? I think it has a number of potential lessons

16for section 2, and maybe we will explore some more about

17that as we talk through the questions.

18MR. WALES: Thanks, Tom.

19(Applause.)

20MR. WALES: Okay, next up we have Irwin Stelzer.

21Irwin is a Senior Fellow and Director of Hudson

22Institute's Economic Policy Studies Group. Prior to

23joining the Hudson Institute, Dr. Stelzer was Resident

24Scholar and Director of Regulatory Policy Studies at the

25American Enterprise Institute. He also is a U.S.

160

1economic and political columnist for The Sunday Times

2and The Courier Mail, a contributing editor of The

3Weekly Standard, and a member of the board of the

4Regulatory Policy Institute at Oxford, a member of the

5Advisory Board of the American Antitrust Institute, and

6adviser to the U.S. Trade Representative.

7Dr. Stelzer founded National Economic Research

8Associates, NERA, and served as its president for many

9years. He also served as a Managing Director of the

10investment banking firm Rothschild, Inc., and Director

11of the Energy and Environmental Policy Center at

12Harvard. His academic career includes teaching

13appointments at Cornell, the University of Connecticut

14and NYU. He has been elected a visiting fellow at

15Nuffield College, Oxford, and he is a former member of

16the Faculty of Practicing Law.

17DR. STELZER: Thank you very much. Can you hear

18me in the back? Thank you for inviting me to this,

19although I fear I may be sailing under false pretenses.

20Let me clear up one of them. Although I am at the

21Hudson Institute, I do not want to appear here as

22somebody who is a disinterested scholar. I do have

23clients, some of whom are accused of being dominant,

24others of whom think dominant firms pick on them, but my

25views go back before most of you were born. I, too,

161

1still play basketball, but I have learned a trick, which

2is I yell "Get that rebound" to other people.

3I am going to leave any comment on specific

4cases to my co-panelists, because they are more familiar

5with them than I. I will say, if I am permitted one

6vignette, I gave up trying to be involved in specific

7cases when I was sitting on a witness stand in Tucson,

8Arizona, and the judge summoned counsel to the bench and

9said, "We have to talk about schedule." The first

10lawyer said, "Well, you know, my daughter's getting

11married in May, and that's going to tie me up." The

12other guy said, "Well, you know, in June, I really was

13planning a fishing trip." The judge said, "Well,

14September, I cannot really do," and so they put

15everything off about a year. In the middle of this, I

16said, "Can I tell you something about my schedule, Your

17Honor?" He said, "Don't be ridiculous." I suddenly

18realized three lunatics were deciding how I was going to

19live my life for the next year, and I am not doing this

20anymore. So, I speak to you as a person who used to

21testify in these cases.

22I have submitted a much longer, unconscionably

23long paper, which I assume is available to those who

24want it, and I will therefore restrict my comments to a

25very few, and also, I want to try out ideas. I am not

162

1wedded to what I am about to say. I assumed we were

2here to try out ideas, not to hand down edicts, and I

3thought that is why I would try concentrating on pricing

4practices by dominant firms.

5Simon Bishop said if you are dominant the

6practice is questionable; my feeling is if the practices

7are questionable, you are probably dominant. Simon says

8he is a bit more humble than doing away with market

9definition. Those of you who have ever tried to do any

10market definition know that only the non-humble would

11attempt the elasticity measurements and the other things

12involved in it. So, the notion that we must begin with

13market definition because that is somehow a constraint,

14and anybody who has read any decisions of the EU knows

15that it is a very, at best -- you defined it as a loose

16constraint. I think it is looser even than that.

17I am not certain that going through the agony of

18market definition gives you a degree of precision, some

19sort of constraint on the examiner. It may, but

20given -- if you go through it, I am not so sure that

21beauty is in the sight of the beholder as with any other

22part of economic analysis. I am not wedded to market

23definition, and I would like to explore the possibility

24that we might want to do away with that exercise

25altogether in deciding about dominance.

163

1I recognize that that would unemploy half of the

2economics profession, leaving only that part that knows

3about exclusionary practices still existing, but I do

4think we should think -- think -- about the possibility

5that defining relevant markets, defining product

6characteristics, all of that is a kind of very elastic

7process that we could do away with.

8Let me suggest instead -- and I really mean

9suggest. There is this kind of academic politeness

10about "let me suggest," meaning "I really know that." I

11do not use the language that way. I really mean to

12suggest that we consider that it is the practices that

13reveal dominance and not dominance that reveals the

14practices.

15I have read some of the proceedings, and it

16seems to me there is a great deal of sort of motherhood

17and apple pie stuff in this record. It is certainly

18true, we do not want to prevent vigorous competition

19that results in lower prices to consumers. Who would

20want to prevent vigorous competition? Certainly

21Microsoft did not want to prevent vigorous competition,

22it says. Yes, we want firms to develop pricing plans

23that benefit consumers; yes, we want to give businessmen

24as much certainty as possibility; and yes, we want to

25reduce the role of lawyers in the board room and leave

164

1it to businessmen. But I do not think that means that

2pricing practices should be unscrutinied by antitrust

3enforcement authorities regardless of any finding of

4dominance.

5What we do not want to condone is long-term harm

6to the competitive process, therefore to consumers, by

7approving short-run price reductions aimed at creating

8barriers to entry or preserving market positions that

9are unrelated to efficiency. Now, again, I would not

10try to measure efficiency of a firm, because I do not

11think I know how to do that. There may be people who

12know how to do that, and when people say to you they are

13going to measure costs, they are going to compare costs,

14I would urge any one of you who agrees that that is a

15terrific idea to determine any cost of any large firm,

16and you tell me what range you think would make you feel

17comfortable in that determination, especially since you

18are usually dealing with someone who does not want you

19to find out, and so I think you are going to have a very

20difficult problem.

21What you have to do is examine a firm's pricing

22practices in the context of the firm's total behavior.

23You cannot look at a thread in a tapestry in order to

24get a picture of whether or not a firm is engaging in

25exclusionary practices.

165

1I will give you an example. If you had in the

2record that a firm had offered a million dollars to a

3customer not to deal with a competitor, you would say,

4"Well, gee, we can't tolerate that." But it is very

5easy to manipulate a pricing schedule in a large

6multi-product firm to accomplish the exact same thing,

7to reduce the cost of the incremental order to pretty

8close to nil by simply manipulating the pricing

9schedules and the relationship of past to future

10deliveries.

11In other words, it seems to me, again, that

12firms spend millions, hundreds of millions, on discovery

13in antitrust cases, and the discovery is really

14discovery that will tell you whether the firm is

15dominant, whether the firm is engaging in exclusionary

16practices, with far greater certainty than would any

17measure of its market share.

18I think, also, you can tell -- I hate to use

19this word because I think it is old-fashioned -- you can

20divine intent from looking at what discovery turns up.

21Now, by that I do not mean that the statement by an

22enthusiastic salesman who says "I just rubbed out the

23competition in Florida" or something like that should be

24taken at face value, but I think you can determine the

25intent of a variety of competitive weapons wielded by a

166

1firm by examining the entire record of its behavior,

2which brings me to the last question -- I said I would

3not take my full time -- and that is, has what I just

4said reduced certainty?

5A lot of my clients talk about certainty, they

6want certainty, so you say, "Well, you want certainty?

7There are two kinds of certainty you can have.

8Everything you do is subject to a per se rule. That is

9certainty. How about that?"

10"No, that is not what I particularly had in mind

11by '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."

15Certainty is simply not available in this

16business. That is it. It is good for the lawyers. It

17is bad for the businessmen. In making their decisions,

18they have to listen for counsel and decide what to do

19about the legal advise they get. It is simply one

20aspect of the many risks they take, just like guessing

21at interest rates. Certainty is not there. It cannot

22be had unless some of the more distinguished members of

23this panel can give it. I cannot.

24Thank you very much.

25(Applause.)

167

1MR. WALES: Last, but not least, we have Joe

2Sims. Joe is a senior antitrust partner at Jones Day

3here in D.C. His practice is concentrated on antitrust

4and related areas of governmental regulation and

5includes litigation counseling, agency practice before

6state and federal courts, antitrust enforcement agencies

7and various specialized agencies where competition

8policy or antitrust issues arise. Joe is a member of

9the American Bar Association, Antitrust Law Section, and

10has served as chair of numerous committees on the

11Antitrust Law Section. He's a Fellow of the American

12Bar Foundation and a member of the American Law

13Institute. He regularly writes and lectures on

14antitrust and related subjects and is listed in The Best

15Lawyers in America, The World's Leading Lawyers, and

16Who's Who Legal.

17Joe, thanks.

18MR. SIMS: Thank you, Dave.

19Let me start with a point about my perspective,

20which will also be true for at least Irwin and Tom. I

21had the revelation when preparing for this and looking

22back at some of the older cases that I have been

23practicing antitrust law for about a third of the time

24that we have had antitrust laws, which is kind of a

25scary thought if you think about it, but it is true

168

1nonetheless. A little depressing, too.

2During that time, no one has ever confused me,

3unlike most of these people on the panel, as a scholar.

4I do not cite footnotes in cases. Sometimes I cannot

5even remember what a case holding was. I do not write

6law review articles. I write commentaries, not as

7eloquent as Irwin's commentaries, but it is a less

8taxing discipline than law review articles. So, I view

9my role here as offering the practice perspective. I

10know Tom is a practicing lawyer, but his scholarship is

11so impressive that I have always viewed him as an

12academic at heart. So, I am going to approach what I

13have to say in that light, focusing not on the theory,

14but on the practice.

15Fortunately, jurisprudence and for that matter

16economics and antitrust is very heavily fact-weighted.

17The jurisprudence and the economics almost always take a

18back seat to the facts, at least in the long run.

19Antitrust law in the United States, where it is really

20law enforcement and not regulation, is mostly about the

21facts and how the facts are presented. This is true

22whether you are talking about agencies or judges. It is

23certainly true when you are talking about juries.

24Of course, the case law is important. Bad case

25law is not desirable. It is a good idea, if we can,

169

1which we do now and have from time to time, have

2competent, intelligent people running the antitrust

3agencies, but all of that fades in importance to the

4unique facts at play in any particular case.

5During at least my practicing lifetime, we have

6moved steadily away from what we used to spend a lot of

7time at, which was antitrust by sloganeering, to more

8careful analysis of the facts. If you remember, Derek

9Bok called for more certainty and bright line rules in

10section 7 cases more than 30 years ago. Well, that

11actually had some resonance for a while, but that

12concept was seriously injured by Bill Baxter's Merger

13Guidelines and probably finally killed by the 1992

14edition of the Guidelines. When the analysis focuses on

15competitive effects and not on market shares or

16concentration or other slogans, the notion of broadly

17applicable bright lines disappears.

18So, today, in merger cases, we do not really

19have any clear rules. All the facts are in play. Every

20case is unique, and while the outcome needs to comport

21generally with stated case law and regulatory guidance,

22the operative word is "generally."

23This is equally true in section 2 matters. We

24have come a long way from American Tobacco or Alcoa or

25even Grinnell, which I was shocked to see was decided

170

1just four years before I graduated from law school. It

2seems like a very old case, and with some obvious

3exceptions, like, Aspen Ski and maybe Kodak, the general

4direction of Supreme Court decisions over my lifetime

5has been to gradually cabin in the reach of section 2,

6in significant part by insisting upon a focus on the

7facts as opposed to reliance on the mostly populist

8rhetoric about market dominance and relative size that

9dominated section 2 jurisprudence in earlier times.

10A good deal of this, of course, reflects the

11fact that our markets have matured -- that many more

12markets today, maybe most markets, are truly

13contestable, which was not always the case -- but

14nevertheless, we do not have very many clear rules in

15section 2 today.

16I think this is generally a good thing, but it

17does, as Irwin pointed out, inevitably carry with it

18uncertainty of outcomes in particular cases. I noticed

19in looking back at some of the earlier hearings that the

20Microsoft representative, perhaps understandably, took

21the position of wishing that there was more clarity in

22the law. It is a common business position. I think it

23is a short-sighted business position.

24To pick up on Irwin's point, if we really did

25have more clarity, we would have more restrictive rules.

171

1I do not have any doubt that if you have to choose

2between clear restrictive rules and clear unrestrictive

3rules, it is where that line would be drawn. I do not

4think that would be useful for the public interest in

5the long term, and it would not even be useful for

6business at least in the medium to long term. It would

7make the advisory job easier, but that is about it.

8So, with this context, these kinds of hearings

9are really a great idea, especially if they try, as I

10think they have, to take the long view of an important

11area of law. More discussion will produce more

12understanding and will also demonstrate, as these

13hearings pretty clearly have, that there is an enormous

14variety of views on section 2 jurisprudence and policy.

15Indeed, I would argue that this might be more true today

16than it has been in my practicing lifetime.

17We still have, of course, the strong populist

18supporters of very aggressive section 2 enforcement. We

19still have plenty of conservative "let the market work"

20advocates. But we also today have an incredible variety

21of economists and law professors and others who

22articulate an amazing range of interesting approaches to

23the identification and analysis of market power. Tom

24Krattenmaker and Steve Salop obviously are responsible

25for maybe the single most visible effort in this field,

172

1but there are a lot of people keeping them company with

2new and interesting ideas, including, of course, Greg

3Werden and others on this panel.

4So, there is no end to possible options for new

5section 2 approaches, but there is also clearly no

6consensus on any particular approach, with the possible

7exception that we really ought to pay attention to the

8facts. It is very hard for me to imagine how we can

9productively create clear rules or safe harbors for

10section 2 using market shares or, for that matter,

11anything else. Given this lack of consensus on where we

12ought to draw the lines and the truism, that, at least

13over the long run, markets are a lot better at

14identifying and responding to consumer demand than

15courts or regulators or most academics, the chances of

16finding consensus bright lines that really do advance

17the public interest are pretty low. But it is

18nonetheless worth talking about, and so these hearings

19are a good idea.

20Any legal discipline like antitrust where the

21operative legal standard is in one form or another the

22rule of reason is going to be messy and unpredictable.

23Facts are highly variable, and their perception and

24analysis by humans is even more so. There is the

25additional problem that courts and regulators, even very

173

1thoughtful ones that take the time to think about and

2listen to various points of view, are inevitably better

3at evaluating the past than they are at predicting the

4future. They are too often focused on fixing

5yesterday's problems without really having a very clear

6picture of how that is going to affect tomorrow.

7Because of this, we ought to try to be cautious

8about interfering with markets, doing so only when we

9are pretty darn confident that the intervention will

10make things better. I have written on this for 25

11years, describing (in very gross and simplistic terms,

12of course) the two basic approaches in antitrust as "do

13no harm" and "can we help". The "can we help" school

14tends to be a lot more confident about their and a

15court's ability to improve market performance than I am,

16but the "do no harm" school has been in clear ascendency

17in the past several years, both at the federal agencies

18and at the Supreme Court.

19This certainly does not mean that it would not

20be great if these hearings could find a way to produce

21some clear consensus and let us feel comfortable in

22drawing some more bright lines like we have in the per

23se rule against price fixing, or in the section 2

24analog, the below-cost requirement for finding predatory

25pricing. But my reading of the results so far -- and I

174

1have read at least summaries of all of the hearings --

2does not leave me with the impression that we have yet

3identified that consensus.

4As I said, I am not sure this is a bad thing.

5One of the most important -- maybe the most important --

6reasons the antitrust laws have continued to serve us so

7well after more than a century is that they are pretty

8darn flexible. Congress, of course, passes a lot of

9statutes where, in effect, buck the problem to the

10courts or a regulatory agency, but it rarely works as

11well as it has in this field.

12I think that is because, in general and over the

13long term, the rule of reason is a pretty accurate

14description of what courts really do -- and regulators

15too, for that matter. They generally try to figure out

16what is reasonable under the circumstances with a strong

17bias most of the time -- let's put the Robinson-Putman

18Act to the side as an outlier -- toward leaving markets

19free to work their magic.

20As long as this is the operative legal regime

21under section 2, we will have uncertainty about

22particular cases and there will be uncertainty about how

23a particular fact pattern is analyzed. This approach

24has costs, of course, including, most importantly, the

25inadvertent deterrence of procompetitive behavior, but I

175

1suspect the costs are less than would be the case with

2either bright line rules that miss the mark or

3impractical tests that over-deter because of ambiguity.

4So, I do not think we really need a whole bunch

5of new rules; nevertheless, if we could come up with

6them, we should, and so I am glad we are looking at it.

7We have to remember, however, that there is a difference

8between section 1 and section 2 and a very good reason

9for the difference. Section 1 deals with joint conduct,

10and while there are many times when joint conduct can be

11neutral or procompetitive, there are obvious and very

12real circumstances where there are competitive risks

13from joint conduct, cartel behavior being the most

14obvious. Given this, it is tolerable to have some

15potentially overreaching penumbras of illegality,

16although as we get more cases like Daugher, even this is

17gradually reduced.

18But Section 2, by contrast, is aimed at

19unilateral conduct, and over-enforcement here would

20threaten the very essence of competition. We want firms

21to be monopolists or to try to be monopolists. The less

22risky we make that effort, the less aggressively firms

23will try. So, section 2 cases should be hard to bring;

24they should be harder to win. Successful cases should

25be rare, because true monopolists with durable monopoly

176

1power are rare as determined by how hard it is to name

2some. It is kind of hard to do, actually.

3That's why Microsoft was such an attractive

4case. It was one of the few instances where you could

5look at it and say, "Doggone it, it looks like they do

6have a monopoly." If we can devise some rules or

7guidelines to help us advance this cause, that is great.

8My guess is we cannot, so we ought to let the market --

9in this case, the market for judicial decisions over the

10long run -- create and enforce the rules, and the result

11will be just fine.

12Thanks.

13(Applause.)

14MR. WALES: Thanks, Joe.

15Okay, as we said, we are going to take a

1615-minute break. So, why don't we reconvene at 3:35.

17Thanks.

18(A brief recess was taken.)

19DR. WERDEN: Okay, let's get started again.

20What we are going to do for the next little

21while is start by putting one or two questions to each

22of the speakers, in turn, and then letting the other

23panelists, if they like, comment on what has been said,

24and we are going to take the panelists in the order that

25they spoke, so I am going to start with Simon, and my

177

1question, Simon, is, while there is clearly a dominance

2threshold under Article 82, there really is an open

3question as to how high the bar is for dominance, and I

4think the way Miguel described it, the bar is and ought

5to be quite low. What do you think about that?

6MR. BISHOP: Okay, well, contrary to what Irwin

7might have suggested, most of my clients are actually

8dominant firms, so on that basis, I think, you know, the

940 percent threshold, which is enshrined in Article 82,

10is a pretty reasonable threshold to have. I mean, if

11your market share is below 40 percent, then you can do

12whatever you like. If you are above that, then we move

13into the effects and the assessment of the behavior

14under consideration. It does not mean if you are above

1540 percent, what you are doing is necessarily

16anticompetitive.

17DR. WERDEN: But you wouldn't say that all the

18firms above 40 percent are dominant, of course, would

19you?

20MR. BISHOP: Absolutely not, and that is why I

21said in my talk, you know, the market share is only one

22factor. You have got to take into account a lot of

23other factors to assess whether that 60 percent, say, is

24representative of significant market power.

25DR. WERDEN: Do any of the other panelists wish

178

1to offer a view as to how high the bar should be set in

2the United States where I think most observers think it

3is set considerably higher than in Europe?

4MR. KRATTENMAKER: Or whether there should be a

5bar at all, I guess.

6MR. SIMS: But, Tom, wouldn't you say that there

7shouldn't be a bar, I would think?

8MR. KRATTENMAKER: Yes.

9MR. WALES: So, the answer is there is no bar.

10MR. KRATTENMAKER: Or what I would say is, bar

11to what?

12DR. WERDEN: Bar to proceeding.

13MR. KRATTENMAKER: You mean, like, a

14post-behavior section 2 case where the claim is what I

15called collusive or Stiglerian power? Sure.

16DR. WERDEN: Well, if you want to go down that

17road, in an actual monopolization case, where the

18defendant is alleged to have acquired a monopoly, the

19courts have set the bar fairly high on what it means to

20have a monopoly and generally have required, in fact, a

2170 percent share protected by pretty high barriers to

22entry.

23MR. KRATTENMAKER: Yes, right, right. I think

24if they acquired that monopoly by, for example,

25acquiring a lot of rivals by purchasing firms, that

179

1would probably be an appropriate threshold to do. Now,

2you do not see cases like that because we have had

3section 7, so almost all section 2 cases now are what I

4would call exclusionary or Bainian type, and yeah, that

5is right.

6I think it is not correct to say you could not

7possibly have market power if you have got 66 percent of

8the market.

9DR. WERDEN: So, in the Microsoft case, if their

10share had been 10 percent, you would have looked on

11things pretty much the same way?

12MR. KRATTENMAKER: You know, there were so many

13facts at issue in the Microsoft case...

14No, as I tried to indicate, it does not seem to

15me that you utterly disregard market share, Greg, but as

16I understand it -- and I am still learning this area --

17the ability to exclude can oftentimes be a factor of

18relative size, but the idea that it requires dominance

19of the entire market I think is quite wrong.

20DR. STELZER: Given what Microsoft did and

21proved itself capable of doing, did you have to bother

22measuring its market share? I mean, nobody who didn't

23have huge market dominance, i.e., 90, 80, 40, could do

24those things, could make an equipment manufacturer pay

25them for stuff that was not in the machine. I mean, you

180

1have got to have an awful lot of market power to do

2that. You want to measure market power because lawyers

3make you do it, but as a matter of policy, in the case

4of any firm that can pull off what Microsoft pulled off,

5you could skip the whole market share measurement stuff

6and just say, if they did this, they have market power,

7they have abused it.

8MR. KRATTENMAKER: I probably ought to let Joe

9pick up on that, but I will say -- I mean, I know a

10little bit about Microsoft. I mean, you might be able

11to say that, but if what you are doing is talking about

12the part of the case where they allegedly misrepresented

13whether their programs -- either how it interfaced with

14Java, I do not know that you needed to have a dominant

15market share in order to lie.

16DR. STELZER: No, no, I was talking about where,

17if you decided to put a competitor's product in the

18machine, they charged you for each machine whether you

19put their stuff in it or not.

20MR. KRATTENMAKER: No, I've gotcha. I take

21it -- I mean, I am sympathetic to your viewpoint, but it

22is conduct-specific. For certain kinds of conduct, you

23might infer market power from the fact of the behavior.

24DR. STELZER: What they do, I shall know them.

25MR. SIMS: On this point, I am more with Tom

181

1than Irwin, I think, surprisingly enough. Market

2definition and whatever you draw from that market

3definition is a tool that you want to use when it is

4necessary and useful to figure out what the competitive

5effects of the conduct at issue are. So, there are some

6where, careful market definition is not all that

7important.

8MR. BISHOP: But I think, I mean, some of the

9difference between the U.S. people at that end of the

10table and the Europeans down here is really -- sort of

11reflects some of the sort of philosophical,

12institutional differences, and I'll say institutional

13because I think my personal philosophy is going to be

14closer to that end of the table than a lot of Europeans,

15and I think that that is a point which Joe talked about,

16you know, is do no harm, which is, you know, very much a

17high threshold before you would start intervening, then

18sure, maybe you don't need a market share bright line

19test, but in Europe, the institutional philosophy is

20much more -- you know, there are a lot of markets, the

21EU, the Commission or the competition authorities can

22intervene in to make things better, and in that

23situation, in that sort of institutional setup, then

24having a bright line test which says, "If you do not

25

182

1have a market share of above 40 percent or whatever, you

2can do whatever you like," seems to me an important

3safeguard to prevent people coming in and start messing

4around with your industry, which is very costly and

5potentially extremely disruptive to the firm's business

6model if that firm has got no market power at all.

7DR. STELZER: But that is kind of the "stop me

8before I kill again" argument, right? You need --

9because you know that you really could be irresponsible

10and do bad things, you better have some sort of rule

11that stops you from doing it on the theory that the

12rule, is the lesser of the evils. It is a substitute

13for judgment.

14MR. BISHOP: No, it's not. It is a substitute

15for deciding when a competition authority can bring an

16action against a business.

17DR. WERDEN: Or in the United States, substitute

18for a jury trial.

19MR. SIMS: Well, there is that pretty critical

20difference between the U.S. and Europe in that in

21Europe, the Commission generally gets to say yea or nay,

22and in the United States, the FTC and the DOJ never get

23to say yea or nay. Unlike the EU, they have to go to a

24court and convince a court.

25I think what Simon is postulating is that some

183

1kind of -- if I could borrow the word -- durable

2guidelines that, would last beyond a particular

3administration of the Commission and thus constrain the

4current occupant of those decision-making positions is a

5good substitute, partial though it may be, for what we

6have here in the courts.

7DR. WERDEN: Okay, that was fun. Let's move on

8to a question for Miguel.

9I was very intrigued by your very clear point

10that the suspect conduct in an Article 82 case cannot

11itself be what creates the barrier to entry that is

12required, in turn, for the firm to be dominant, so that

13if it was possible to have a firm with a whopping share

14protected only by the suspect conduct in the case,

15otherwise you would be flooded with competition, then

16that firm isn't dominant? Is that your submission?

17MR. de la MANO: Indeed, and there is the

18problem that we have in the EU, that we do not really

19have a standard which allows us to pursue attempted

20monopolization.

21DR. WERDEN: No, let the firm be 80 percent. It

22is 80 percent, but the only thing keeping out

23competition is this guy's anticompetitive conduct. Now,

24the guys at the end of the table would go after this guy

25at 5 percent it sounds like, but let's put that aside.

184

1He's 80 percent, and he's doing bad stuff, and he's

2keeping the competition out. If he didn't keep doing

3the bad stuff, the competition would come in. They

4might even swamp him.

5 MR. de la MANO: So, let me now link that

6question to the previous question to Simon, which is

7where should we put the threshold for the finding of

8dominance, and, of course, Simon has argued 40 percent

9might be a good place. I am not sure it is a good

10place, and there are a number of reasons why 40 percent

11might be too high.

12First of all, dominance is going to be a

13necessary requirement, and in some cases, like the

14situation you just presented, it may well be that if the

15practice is preventing entry in the market, but in

16assessing dominance, what we are ultimately assessing is

17the situation without such practice. That's why

18dominance is a screen. In a case like that, it would

19not be possible to be brought forward by the European

20Commission.

21Now, that clearly -- you might say, "Well,

22that's wrong," and that's why you have attempted

23monopolization in the U.S. and we do not have it, but a

24second reason why if dominance acts as a screen, we have

25to be very careful in not setting the market share

185

1threshold for a finding of dominance far too high.

2There is a third reason, which is, as has

3already been highlighted by Simon before, which is

4market definition is an imprecise exercise. Now, I

5think everybody here will argue that in some cases, if a

6company has a share slightly above 40 percent, slightly

7below 40 percent, you know, it probably doesn't make

8much of a difference, but if you have a threshold at 40

9percent, it is critical.

10So, even though in practice, a firm with 35 or

1145 percent is probably likely to have much more -- the

12same kind of market power, in theory, this is a

13threshold at which it either -- the Commission is going

14to intervene or not, whereas if you had a lower

15threshold -- and, of course, market definition is going

16to be critical there. It is going to determine whether

17or not the Commission is going to intervene or not. If

18you have a lower threshold, then the precision of the

19market definition exercise matters much less, because if

20you had it wrong and the market definition was actually

21too narrow or too wide, but you are wedding yourself

22into the 20-30 percent threshold, it doesn't really

23matter.

24As long as you are below 25 percent, even if

25you've got market definition wrong, it is for certain,

186

1almost for certain, that there are going to be no

2problems, and therefore, there should be no intervention

3whatsoever.

4DR. STELZER: To ask a practical question, what

5makes you look at something in the first place? You go

6into a bunch of market share studies and you say, "Oops,

7here's a 40-percenter, I'll go after him"? Or is it

8some practice that makes you look?

9MR. de la MANO: The latter, essentially a

10complainant would --

11DR. STELZER: Simon says no.

12MR. BISHOP: Well, Miguel said it right. It is

13some complainant submits a case.

14DR. STELZER: Right. Now, as I understand the

15EU attitude, it differs from the American. Here my

16economist friends believe that if the complaint comes

17from a competitor, it is therefore tainted somehow. It

18is the use of the legal system as a strategic device.

19That is different from the EU, and I think the EU is

20right but is the EU sticking with the notion that the

21fact that a complaint comes from a competitor does not

22taint the complaint?

23MR. de la MANO: Well, practically in all

24cases -- probably in all cases that I have been involved

25in, the complaint has come from the competitor, some

187

1outliers where a consumer may bring the case, but it is

2very, very rare. When that happens, because we have an

3opportunistic system, the Commission, of course, has to

4take in mind the private interests of the complainant

5and how that might taint their submissions, but

6ultimately the Commission is obliged to give its

7decision, whether it is a decision to intervene, and

8therefore -- and that would be trying an independent

9objection sent to the dominant company or allegedly

10dominant company, or there would be a rejection of the

11complaint, which would be a formal rejection, would be

12written and sent to the complainant.

13So, either way, the Commission basically has to

14make up its mind, and in doing so, has to definitely

15take into account to find out if the evidence that has

16been brought forward to it is submitted by parties which

17have their own interests at heart.

18DR. WERDEN: Tom, I have a question for you.

19You seem to be saying that the mere exercise of

20exclusionary market power is a section 2 offense all of

21the time, but I want to clarify if you mean without

22regard to the potential of that conduct to create or

23maintain something we would call monopoly power.

24MR. KRATTENMAKER: I do not mean that.

25DR. WERDEN: Okay, that's great.

188

1MR. KRATTENMAKER: Thank you.

2DR. WERDEN: Anybody want to follow up on that?

3MR. KRATTENMAKER: Irwin says no.

4DR. WERDEN: Well, say it out loud.

5DR. STELZER: But brevity is so much the soul of

6wit that I hated -- I just preferred to let your answer

7hang out there.

8MR. KRATTENMAKER: Sort of like a beautiful

9arcing three-point shot that's probably right dead bang

10through, nothing but the net, exactly, just let it sit

11there.

12DR. STELZER: Right, see, but I play basketball

13at 10,000 feet.

14MR. KRATTENMAKER: Of course you do. You are a

15good guy.

16DR. STELZER: I was trying out ideas. I am not

17sure. Tom, tell me why you think about that.

18MR. KRATTENMAKER: Oh.

19DR. STELZER: How, as a practical matter, you

20would tell in a case.

21MR. KRATTENMAKER: Because there is lots of --

22because the whole point about the competitive process is

23to beat your rivals, and so inferring from the fact that

24practice has an untoward effect on rivals, that it

25therefore violates the antitrust laws, it is just too --

189

1to coin a phrase -- over-inclusive.

2DR. STELZER: Yeah, okay, but -- I guess I was

3thinking in terms of defending the competitive process,

4not competitors.

5MR. KRATTENMAKER: Yeah, right.

6DR. STELZER: And that's harder.

7MR. KRATTENMAKER: Well, I agree. I mean, the

8fact that you inflict some sort of inefficiency on your

9rival, you could say, "Gee, that's bad, and we ought to

10stop it," and that's kind of like the Klor's case.

11That's Klor's against Broadway-Hale. I mean, they might

12have done something bad, and we could care for less that

13there were a hundred other stores in that city, and, I

14mean, there is a way I used to tell that. I mean, I

15went back to the record and examined that case, and it

16turns out that the reason that there was this dispute

17here was that the owner of Broadway-Hale had a

18ne'er-do-well son who had impregnated and run away with

19the daughter of Klor's, and this was an alienation of

20affection suit brought as a Sherman Act case.

21Now, of course, that is not true, but I tell

22that story and the students believe it, and so that's

23the long way of saying I do not think that section 1 --

24of course, we are not talking about section 1 -- was

25meant to federalize the tort of alienation of affection.

190

1So, not only are you supposed to beat up on your rivals,

2but not everything you do to your rivals is either

3necessarily commercially motivated or motivated to drive

4monopoly profits.

5MR. SIMS: And, Irwin, if you don't demand that

6the conduct have at least a high likelihood of creating

7durable monopoly power, then you really do have a

8serious risk of sticking your nose into the market where

9you are going to do more harm than good, because

10differentiating between exclusionary practices on some

11grounds other than whether they have the potential to

12create durable market power seems to me to be very hard.

13DR. STELZER: But you used the term "durable"

14about five times. What do you mean?

15MR. SIMS: I mean more than temporary.

16MR. KRATTENMAKER: There you go.

17DR. WERDEN: Your turn, Irwin, as if you haven't

18talked enough.

19You seem not to at all be a fan of limiting

20principles, and I want to push the limit on limiting

21principles. Are you suggesting, for example, that the

22Brooke Group rule was a really bad idea?

23DR. STELZER: I don't have any idea.

24DR. WERDEN: You don't think that in a predatory

25pricing case, a plaintiff should have to show pricing

191

1below some measure of cost?

2DR. STELZER: Oh, no, I think that's ridiculous,

3and I'll tell you why. First of all, I don't believe

4you can measure marginal cost. I've spent a lot of time

5trying to do that.

6DR. WERDEN: The courts do not like marginal

7cost either.

8DR. STELZER: I'll take any kind of cost you

9want. I don't think you can do it. I've been in enough

10proceedings at regulatory agencies where people are

11supposed to measure costs to know that.

12Second of all, the real question with predatory

13pricing is not whether the person prices below or at

14some concept of cost and has a prospect of recoupment,

15but think of it this way. You are walking along and you

16want to have a picnic, and there's a sign that says, "No

17trespassing." You figure, what the hell. You throw

18down your blanket, you have a nice picnic, and you

19leave, right?

20Now you are walking along and there's another

21field where you want to have a picnic and there's a no

22trespassing sign, and there are about four or five

23corpses lying around. Are you going to have a picnic

24there? I don't think so.

25So, what we are talking about is the kind of

192

1practices that are entry-deterring in the technical

2jargon, that scare the hell out of people, because

3remember, this is more and more an age in which the

4financing of new companies is done by venture

5capitalists, and if you have ever been to a meeting with

6a venture capitalist -- these are not very nice people,

7many of them -- the first thing they want to know is

8what is the range of practices available to the

9incumbent competitors to keep you out or to destroy you

10if you get in. That is what they want to know.

11I mean, have you got a good idea? Yeah. Are

12you a pretty good manager? Yeah. Can I suck most of

13the value out of your enterprise? Yeah. And then they

14want to know what are the incumbents going to do to you,

15and if you go to enough meetings where people describe

16what Microsoft might do to you or what other companies

17might do to you, a lot of the stuff we are talking about

18becomes irrelevant. Entry-deterrence is the problem.

19Will they cut prices? Yes, they might. Is that okay?

20Well, that's a tough one. That's very hard.

21I know this sounds mushier than you'd like it to

22be. People who say I am going to measure costs and then

23I am going to measure market share -- in the Sirius/XM

24merger, right, they are going to take one data point and

25they are going to measure cross-elasticities and all

193

1that other stuff? Ridiculous.

2So, what I am saying is in a practical world in

3which new firms are being created, in which technology

4is increasingly important, in which small businesses and

5new entrants are the manufacturers of macroeconomic

6growth, I would lean pretty hard in the direction of

7being very skeptical about the range of competitive

8tools permitted to incumbents, to powerful incumbents,

9for macroeconomic reasons, for microeconomic reasons,

10and -- dare I say it, even though Judge Bork is a

11colleague of mine -- for equity reasons.

12DR. WERDEN: Are you suggesting that if the

13incumbent is happily pricing at 100 and somebody has a

14new idea and comes in and sells it at 80 and the

15incumbent says, "Well, I better knock my price down to

1680 or I am not going to make any sales," he's already in

17trouble?

18DR. STELZER: No, I am saying you have to look

19at a lot of things. You see, that's the trouble. You

20are trying to pick out one thing that will tell you what

21the hell is going on in this industry. You can't do

22that.

23DR. WERDEN: Okay. Well, I concede that I can't

24do that. So, what do I do?

25DR. STELZER: You look at the entire range of

194

1business practices of the company. You look at the

2durability of its market share. You look at the history

3of the notices it has posted in the past when

4competitors try to come in, and you try to make a

5decision as to whether those were imposing

6inefficiencies on the potential competitors or not.

7MR. WALES: Go ahead, Tom.

8MR. KRATTENMAKER: I want to come to Irwin's

9partial defense now --

10DR. STELZER: Oh, God.

11MR. KRATTENMAKER: -- on Brooke Group but make a

12comment about -- to make a comment about what Joe said,

13too.

14On what Irwin said, you know, pricing below

15cost, I am really not so sure. Recoupment, yes, and the

16short answer to your question, Greg, is you have got to

17show that they will be able to get their price back up.

18When we all sit around and decide that we have this

19common mantra and we decide to chant it, whatever this

20antitrust religion is that we have, you have to be

21careful to think about it once in a while.

22Saying it has got to be below the pricing firm's

23cost is to smuggle in the old efficient competitor rule

24into the marketplace. If it is the case that the firm

25can by pricing right down to its cost drive out four

195

1firms and leave us with one firm instead of five in a

2market, some people may say that drives us to more

3efficient production, and other people will say that is

4going to tend to drive prices further away from costs.

5It depends on which value you think is important in

6antitrust.

7I think it would be better to have a discussion

8about that than the silly stuff in Brooke Group about

9what we happen to know because we happen to put on black

10robes and so we are infallible, that people often try

11predatory pricing and rarely succeed, a statement which

12I believe had no support. There might be a footnote

13there, but it doesn't cite any empirical work.

14So, I don't mean to say that I am opposed to

15Brooke Group, but what I mean to say is you don't look

16askance at somebody and say, "You mean they wouldn't

17price below cost?" Irwin is talking about a somewhat

18different set of values and in this case a very

19defensible set of values, particularly if you do keep

20the recoupment link, I would say.

21The other comment, I mean, I think this is the

22right time to make it, I thought Joe had one of the most

23interesting observations I've heard in a long time about

24the bright line rules and fact-based rules, and that's

25exactly what has happened to merger law in the whatever

196

1years since Joe and I first started studying merger law,

2but it's not what's going on in section 2, and these are

3hearings about section 2.

4You've got some cases that were sort of driven

5 down to fact-based. Aspen Ski is one of those where

6they looked in the record and found that there were some

7angry skiers in Atlanta, and Kodak copiers is one of

8those, but we have some bright line cases, too,

9Weyerhaeuser, Brooke Group, the 11th Circuit decision in

10Schering-Plough, that say, do not tell me any facts.

11All I want to hear is some theory.

12So, in section 2, we are in -- I'll shut up here

13now in a minute -- in section 2, we are at this funny

14point where we haven't moved to Joe's Nirvana, and I

15think we need to face that.

16MR. SIMS: See, it is interesting. I agree with

17you on Brooke Group and Weyerhaeuser. Those are

18essentially safe harbor decisions.

19MR. KRATTENMAKER: Yeah.

20MR. SIMS: But I would vehemently disagree with

21you on Aspen Ski and Schering-Plough. I think that

22Aspen Ski is certainly not fact-based. You can't do a

23fact-based analysis of Aspen Ski and conclude that there

24was an antitrust violation there.

25MR. KRATTENMAKER: No, the fact they found turns

197

1out not to be a violation -- turns out not to be an

2anticompetitive act, but --

3MR. SIMS: Well, that's certainly true, and I

4think Schering-Plough I think did focus on the facts,

5and the fact that was determined -- that was found to be

6determinative in Schering-Plough was the existence of

7the patent and the scope of that patent. That's a

8fact-based analysis to me, not rule-based.

9DR. STELZER: Can I ask you something about

10Aspen Ski, because I am not a lawyer --

11MR. SIMS: Sure.

12DR. STELZER: -- although I was involved in that

13case just because I happened to be in Aspen at the time

14and the plaintiff couldn't afford anybody and I was

15free.

16MR. SIMS: I remember actually visiting you in

17Aspen periodically.

18DR. STELZER: Right. Well, come this summer,

19because I don't have judges setting my schedules

20anymore.

21Let me ask you something. There was an

22unchallenged determination of the relevant market.

23MR. SIMS: Yes, that was the --

24DR. STELZER: Now, is that a fact or is that not

25a fact?

198

1MR. SIMS: That was a lawyer error, actually.

2That was a stipulated market which any good antitrust

3lawyer wouldn't have done.

4DR. STELZER: All right. So, we are now down

5to, if I understood it, it is not a fact if it is

6determined by a judge and a jury but it is a lawyering

7error. Is that right? So, that makes it not a fact.

8MR. KRATTENMAKER: That's our position and we

9are sticking to it.

10DR. STELZER: Okay, that's all right, I just

11wanted to know.

12 DR. WERDEN: Moving right along, Joe, I am not

13entirely sure I understand your position. I am not sure

14that you go so far as to say clarity is bad. I think

15your position more is that hoped for clarity isn't going

16to come in a useful way, to which my follow-up question

17is, well, aren't there things like the Brooke Group rule

18that would form conduct-based safe harbors that might be

19a good idea? For example, that it is okay to introduce

20a new product even if that causes your competitor to

21fail?

22MR. SIMS: Well, I wouldn't have any problem

23with that rule, but I think you'd have a lot of trouble

24getting broad consensus on it.

25DR. WERDEN: I am willing to try. Let's see

199

1what we can do here on the panel.

2MR. SIMS: You might find some people that think

3that's what Microsoft did and does and is doing --

4introducing new products that are creating competitive

5harms; at least I think that's the theory in the EU's

6current preoccupation with Microsoft. So, I am fine

7with a Brooke-type safe harbor for new product

8introductions. I am not exactly sure how you'd set it

9out so that you left it open for the one in a however

10many times that might be anticompetitive, but I'd be

11fine with that. I doubt seriously that you would get

12broad consensus on that.

13My point is that there is not incredibly broad

14consensus on the Brooke Group rule, which is I think

15about the only effective safe harbor in section 2 now.

16So, I am not sure that you would have a very easy time

17coming up with consensus on any others. I am happy to

18see you try, and I could come up with a number that I'd

19be comfortable with, but I doubt that I'd get everybody

20to join with me.

21DR. WERDEN: Well, we can give you 30 more

22seconds. How many can you give me in 30 seconds?

23MR. SIMS: Well, new product design would be

24fine. I mean, in general, new products and product

25design decisions, I am involved now in defending Apple

200

1in the iPod tying cases. We shouldn't have to go

2through all the hassle that we are going to have to go

3through to get rid of those cases. So, I am perfectly

4happy with that if you can find enough consensus to

5implement it.

6DR. WERDEN: Do I hear any dissenters?

7DR. STELZER: Well, I was just curious, Joe,

8what about what they call fighting brands in the

9cigarette industry?

10MR. SIMS: What about them?

11DR. STELZER: That's a new product.

12MR. SIMS: Is there anything wrong with that?

13DR. STELZER: Is there anything wrong with that?

14MR. SIMS: No, I don't see anything wrong with

15that. Did it impair competition in some way?

16DR. STELZER: It had very negative effects on

17some of the competitors who made the brands.

18MR. SIMS: That's different.

19DR. STELZER: But it sends a notice that you are

20going to come in --

21MR. SIMS: Look, I happen to know an awful lot

22about the cigarette business, unfortunately, because I

23just did a merger there a couple years ago. There are

24one heck of a lot of independent sellers of cigarettes

25in the cigarette business. In fact, they have driven

201

1the market share of the market leaders down, and more

2importantly, they have taken away a big part of their

3margin, which is why the FTC decided not to challenge

4the merger of the number two and number three players.

5DR. STELZER: Okay.

6MR. KRATTENMAKER: I worked on that case, too,

7and some of what Joe just said is true.

8DR. WERDEN: Probably some of what Joe says is

9always true; it is a question of how much.

10MR. KRATTENMAKER: I was on the other side, I'm

11sorry, I was doing it for the FTC.

12MR. de la MANO: I would defend, Greg, that

13particular bright line rule.

14DR. WERDEN: Okay. When is a new product

15introduction a bad thing for consumers?

16MR. de la MANO: I think that's the wrong way to

17put the question. I think no bright line rule is going

18to work unless you define it very, very carefully, and

19you will --

20DR. WERDEN: Of course. That's what your job

21is.

22MR. de la MANO: Well, that's what we found in

23the new product rule that we were given by the court in

24the area of refusal to supply, the new product test,

25that -- it sounds fine in the context of that particular

202

1case, I admit, but we just do not know what's a new

2product.

3DR. WERDEN: Well, but if we are going to take a

4European approach to this question, then perhaps we

5should appeal to our ordoliberal traditions where, what

6we say in English, competition on the merits was a

7fundamental principle. That was legal without regard to

8its effect, and there are reasons to believe that this

9concept is embraced by Article 82.

10Now, as far as I can tell, no European court has

11ever said that that actually means something, but it

12should mean something, shouldn't it?

13MR. de la MANO: Definitely.

14DR. WERDEN: Okay, what does it mean?

15MR. de la MANO: Well, the problem is that if

16you put the question in terms of would a new product

17ever constitute the situation where it could lead to

18consumer harm, I think the answer is always going to be

19no. That is competition on the merits. That is a

20situation where there's going to be traditional value to

21consumers, that's pretty obvious, but the difficult

22thing for a competition agency is to define or identify

23whether that product is, indeed, new, and there are many

24situations where what might appear on the face of it to

25be a new product, from the perspective of certain

203

1customers, but is just an extension or an additional

2feature that's added to an old product, but if that

3additional feature serves the purpose of preventing

4entry, then maybe there is a problem.

5DR. WERDEN: I agree there's always going to be

6a fine line, and Irwin correctly pointed out that the

7fine line is Brooke Group is a serious problem. We

8can't figure out costs well. But that doesn't mean

9there's something fundamentally wrong with the

10principle.

11MR. de la MANO: Absolutely not. It's not just

12a good bright line for enforcement.

13DR. WERDEN: You are coming to that decision

14awfully fast. How long have you been applying it?

15MR. de la MANO: I don't think we have had a

16single case in the IMS where we have actually been able

17to define a new product as of -- that's a few years.

18DR. WERDEN: Of course, the bright line rule

19there is that you can refuse to license. That solves

20that problem, doesn't it?

21MR. de la MANO: Yeah, solves that one, yeah.

22DR. WERDEN: Okay.

23MR. WALES: Should we move on to the principles?

24Go to the first one.

25DR. WERDEN: Okay, I hope you people can see

204

1this. We are going to read these.

2MR. WALES: I actually have them in hard copies

3and we can pass them out.

4DR. WERDEN: Okay. We are going to read them

5into the record in any event.

6We have in most of our sessions, but not this

7morning, gone through what we call the propositions

8where we put up a declarative sentence and ask the

9panelists whether they agree or disagree and why.

10The first one we have here is, "Monopoly power

11is the long-term ability of a firm to earn greater than

12a competitive return on investment."

13It's not the most orthodox definition of

14monopoly power, but it happens to be the almost verbatim

15the definition in one of the leading economics

16textbooks, and it focuses attention on something that in

17principle we might be able to figure out, although it's

18not going to be easy, whether a firm is earning more

19than a competitive rate of return.

20So, Tom, why don't you start.

21MR. KRATTENMAKER: I think it is good enough for

22government work.

23DR. WERDEN: Good enough for the courts of the

24United States of America?

25MR. KRATTENMAKER: Not having tried to do a case

205

1under this test, I would want to think some more about

2whether I'd rather be going and getting evidence about

3competitive returns than I would about prices and costs,

4Greg. So, I cannot answer your question. I am

5obviously -- as a lawyer, I am, of course, hind-bound, I

6am always looking backwards, and so I am happier with a

7test that focuses on price than competitive return if

8you give me 30 seconds to think about it, but --

9DR. WERDEN: Well, that's fine. It doesn't say

10here what the evidence would be, and I think it would be

11prices and costs in some cases, most cases, but the

12question then is going to be, what price and what cost?

13MR. KRATTENMAKER: Thank you for modifying this

14as we go. It has changed from long-term to long-run, it

15has changed from competitive return to pricing above

16costs. I think it is basically right, but I want to say

17the devil's in the details, but there are some details

18that would need to be worked out, but sure.

19DR. STELZER: Would you accept --

20MR. KRATTENMAKER: As you know, I'd also say

21that's also market power. I do not know, is that the

22next question? Do we have another question about that?

23DR. STELZER: Can I ask you a question?

24DR. WERDEN: Please.

25DR. STELZER: Would you substitute cost of

206

1capital for competitive return on investment?

2DR. WERDEN: Possibly.

3DR. STELZER: Okay. Have you ever been in a

4utility case where they're determining the cost of

5capital?

6DR. WERDEN: We hardly ever do that anymore,

7thank God.

8DR. STELZER: You hardly ever do it, but if you

9walk down the block, there's a lot of people doing it.

10There's economists doing it all the time and there's a

11huge dispute about it, but I think cost of capital is at

12least more precise as far as the literature goes than a

13competitive return on investment. So, if you want to

14play with this, I think you should do it in terms of

15cost of capital, because there are all sorts of ways of

16measuring cost of capital, and no one will know -- they

17won't know with as much precision what you are talking

18about when you talk about a competitive return.

19DR. WERDEN: Well, coming back to Tom's

20question, if you want to put this in terms of prices and

21costs, the question, as I said, is what price and what

22cost?

23MR. KRATTENMAKER: Sure.

24DR. WERDEN: And in particular, the difference

25between monopoly power and market power, it is

207

1conventional, at least, although there are some

2dissenters, to define market power as the ability to

3price above short-run marginal cost, but hardly anybody

4would say that the right definition of monopoly power is

5the ability to price above short-run marginal cost,

6because that would give us too many monopolists.

7MR. KRATTENMAKER: I think your second sentence

8is correct and your first sentence is wrong.

9DR. WERDEN: So, what is the definition of

10market power?

11MR. KRATTENMAKER: I believe that market power

12has a durability component as well, the last time I read

13the Guidelines, nontransitory.

14MR. WALES: So, shorter, Tom, is that the point?

15It is shorter than monopoly power?

16MR. KRATTENMAKER: No, it is the same.

17MR. WALES: So, both qualitative and

18quantitative? I guess you made the point that

19qualitatively, they're the same, but are they also

20quantitatively the same?

21MR. KRATTENMAKER: Oh, I think each of them

22comes in degrees, Dave, I'm sorry. To go back to my

23metaphor -- they could turn out to be a one-point shot,

24a two-point shot, a three-point shot. I don't think it

25would serve us any value to say, well, if it is a

208

1two-point shot, it is market power, and if it is a

2three-point shot, it is monopoly power. I don't -- as a

3matter of moving the cases along, I don't see the point.

4DR. WERDEN: Well, let me put the question,

5then, doesn't it make sense to have a significant

6threshold in a section 2 case that is different and

7higher than the threshold of market power in a section 1

8case? And don't the cases pretty much say that's the

9law now?

10MR. KRATTENMAKER: No. Yes.

11DR. WERDEN: Okay, at least that was clear.

12MR. BISHOP: But, I mean, the European

13perspective, I mean there is some debate in Europe about

14whether we can characterize firms which are dominant and

15those firms which are super-dominant, which is sort of,

16you know, similar to this, and my sense is that, you

17know, why bother introducing this new term, you know,

18"super-dominant"? If we are just going to use the

19dominance as a threshold step to deciding whether we

20need to investigate in more detail the competitive

21conduct, whether a firm is dominant or super-dominant

22doesn't really make any difference in that decision.

23DR. WERDEN: Okay, let's move to the second

24proposition. I think Joe spoke precisely these words,

25and I want to see how much consensus we have on the

209

1proposition that monopoly power is rare.

2MR. WALES: If we can go back to Miguel.

3MR. de la MANO: Well, in line with any

4consensus that monopoly -- it makes very little sense to

5distinguish between market power and monopoly power for

6the reasons that have been explained on both sides of

7where I am sitting, I would say monopoly power is fairly

8common. The key question is, however, how much of it do

9you really need to show or need to have before you

10decide to investigate any further? Being shown monopoly

11power is not anything in itself; it is the practice

12itself, the conduct.

13DR. WERDEN: I think you have identified one of

14the major differences in attitude between the European

15school and ours. Our courts are really hard sells on

16the subject of monopoly power. It is an empirical fact

17that it is very hard to convince a court that a firm has

18a monopoly in the United States, and it's not that hard,

19it seems, in Europe.

20I think you have already cast your vote that it

21is probably too hard in the United States. Anybody else

22want to weigh in on that?

23MR. KRATTENMAKER: Well, yeah. I mean, I think

24that Miguel has really laid his finger on it. If we

25then say that you possess market or monopoly power if

210

1you face a downward-sloping demand curve, I think it may

2well be that many, perhaps most firms, do, but the

3second thing I was going to say is this question,

4monopoly power is rare, is exactly why I went to law

5school instead of graduate school in economics. You

6have to ask an economist who does not I/O theory, but

7I/O reality, how often this happens. Isn't this what

8Joe Bain spent his life trying to do, but --

9DR. WERDEN: I don't think so, but --

10MR. KRATTENMAKER: Okay.

11DR. WERDEN: Anyone else?

12MR. de la MANO: Can I reverse the question?

13DR. WERDEN: Rare is power monopoly?

14MR. de la MANO: No. Do you think

15contestability of a market is rare?

16DR. WERDEN: I think it is unheard of.

17MR. de la MANO: Well, there you go.

18DR. WERDEN: I am not sure where I am.

19MR. BISHOP: How does that follow?

20MR. de la MANO: Well, it follows that if

21contestability is the opposite of monopoly power and

22contestability is unheard of, it must be because most

23firms have market power.

24DR. WERDEN: Well, but then you are equating

25market and monopoly power, and I am not buying into that

211

1one.

2MR. de la MANO: Okay.

3MR. BISHOP: And I guess it also relates to

4entry to a market. You can have firms with high market

5shares subject to effective competitive constraints

6because the small rivals could easily expand.

7DR. WERDEN: Okay, a third proposition, and this

8is something that Simon already said. "The Cellophane

9fallacy likely does not apply in attempt to monopolize

10cases." Of course, he didn't use that language, because

11that's American language, but here we have an offense of

12attempt to monopolize in which the defendant doesn't

13start out dominant, but it is alleged that he would end

14up dominant with a dangerous probability through the

15activities that he's engaged in, and in defining the

16market in such a case, the proposition is that the

17Cellophane fallacy probably isn't a problem.

18Simon I think already said yes, that's true. Do

19we have any other views?

20MR. BISHOP: Easy one.

21DR. WERDEN: I think that's an easy one. I like

22easy ones.

23Next, "When the Cellophane fallacy does apply,

24which is not a significant number of cases, the proper

25benchmark price in market delineation is the market

212

1price absent the challenged conduct, which is normally

2not the competitive price."

3It is often said, perhaps rashly and wrongly --

4we are going to find out -- that you should go down to

5the competitive price to do the market definition

6analysis. This proposition says no, you should look at

7some kind of but-for price, and Simon, what do you think

8about that?

9MR. BISHOP: Interesting theoretical question.

10The answer is sort of, maybe, but I think in the sort of

11practical reality, it makes no difference. You don't

12know what the but-for price is; you don't know what the

13competitive price is.

14DR. WERDEN: As a practical matter, you may be

15exactly right, but let us suppose you could actually

16figure these things out. What would you do?

17DR. STELZER: And if my grandmother had wheels,

18she'd be a bus.

19MR. BISHOP: If you think about these things,

20then all we need to do is be concerned with the

21Cellophane fallacy or anything. The whole antitrust

22would be very, very easy.

23MR. SIMS: And that is how we get ourselves into

24the messes that we get ourselves into, is pretending

25that we can ignore reality.

213

1MR. KRATTENMAKER: I think this is a very

2interesting concept, and it might be right, but I didn't

3understand the earlier question, and I don't mean this

4as a challenge, Greg, but if you -- if we know both the

5market price absent the challenged conduct and we also

6know the competitive price?

7DR. WERDEN: Yes.

8MR. KRATTENMAKER: And you are making two

9statements, which is that those are normally

10different --

11DR. STELZER: Right, and then which is the

12benchmark?

13MR. KRATTENMAKER: And then I would choose one?

14DR. WERDEN: Yeah. I am not saying these things

15are easy to figure out. They are not. I agree with

16Simon.

17DR. STELZER: They are impossible. It's not

18that they are not easy.

19MR. KRATTENMAKER: I am only clarifying the

20question. The question assumes that I know these two

21prices that are in here, and so you are asking -- you

22are making a statement and asking us about a statement

23and a value choice.

24DR. WERDEN: I'll let you know everything that

25you'd like to know.

214

1MR. KRATTENMAKER: Okay, I know the market price

2absent the challenged conduct, and I know the

3competitive price, and I know that the market price

4absent the challenged conduct is higher than the

5competitive price.

6DR. WERDEN: Yes.

7MR. KRATTENMAKER: Simon's the expert, but I'd

8be inclined to say that the right answer whatever the

9empirical fact is, that the right answer is you focus

10not on the price absent the challenged conduct but on

11the competitive price, but I thought his basic answer

12was correct --

13DR. WERDEN: Why?

14MR. KRATTENMAKER: -- which is, you know, I do

15not know either better than the other.

16DR. WERDEN: I don't want you to give an answer

17now. I want to know why.

18MR. KRATTENMAKER: Because that is what we are

19more likely to be able to assess the supply and demand

20responses to, that --

21MR. BISHOP: But doesn't --

22MR. KRATTENMAKER: -- as the market definition

23process asks us to do.

24MR. BISHOP: But this comes down to, I mean,

25there's practically no difference. I mean, if you knew

215

1what the competitive price was in every single industry,

2antitrust policy would be extremely easy, just go around

3and tell firms that you are not allowed to price more

4than the competitive price.

5MR. de la MANO: I wouldn't be so drastic on

6that, Simon. I think the question has merit. I do not

7know what the theoretical answer to this is, but I think

8from a practical standpoint, I actually think it could

9be easier in some cases to assess what the price would

10be in the absence of the conduct given that we are very

11unlikely to see, going back in time, a market which is

12currently not competitive that might have been

13competitive in the past, but it is very likely to see a

14situation that a few years ago, a market being a

15monopoly was one where that conduct was absent, and it

16might be possible to compare or even do some natural

17experiments across regions, even contemporaneously, to

18compare what is the precise situation where the conduct

19is absent. So, this theoretical conversation, were it

20to be valid, I think in practice, it could be very

21useful.

22MR. BISHOP: Well, I still think that, you know,

23either benchmark means that the inferences that you can

24draw from, you know, the available data is similar to

25the same issues, whether it is a competitive price or a

216

1price absent the conduct. Just seriously, from a

2practical point of view, I do not think it makes any

3difference at all. We can have a, you know, good, you

4know, theoretical debate in saying which one is the

5appropriate one, but from a practical point of view, I

6do not think there is any difference whatsoever.

7DR. WERDEN: We have pretty much covered this

8one, but we are going to put it up anyway, see if

9anybody has anything more to add.

10"A market-share based safe harbor is appropriate

11in monopoly cases."

12MR. BISHOP: Yes.

13DR. WERDEN: Okay, we have one yes.

14MR. de la MANO: Two.

15MR. SIMS: What's the number?

16DR. WERDEN: That's the next slide.

17MR. SIMS: I can't answer it without the number.

18DR. WERDEN: Pick your own number.

19MR. KRATTENMAKER: I say no to this sentence

20because it has a singular noun.

21MR. SIMS: If you give me -- if you give me a,

22you know, 70 percent or an 80 percent number, I might be

23very comfortable with that.

24DR. WERDEN: Okay, we have got a vote for 70 or

2580 percent. We might not have unanimity on 70 or 80

217

1percent.

2DR. STELZER: What is it appropriate to? If it

3is appropriate as a general prosecutorial guide for guys

4picking cases to bring, along with the feasibility of

5relief, then it might be useful, but --

6DR. WERDEN: If it is a safe harbor, it is a

7rule that courts are going to use on summary judgment to

8kick out cases.

9DR. STELZER: Then I would say no.

10MR. SIMS: And I know Tom says no. He has to

11say no.

12MR. KRATTENMAKER: Yes, I did. I already said

13no. I would say yes, it might make sense to have one

14safe harbor --

15DR. WERDEN: You're saying yes, but you're

16coming in with a low number, right, 25?

17MR. de la MANO: I find it hard to understand

18this myth, which I alluded to before, that in Europe we

19have a serious concern with type II errors, yet when it

20comes to using market share safe harbors, there is

21consensus here on this side of the table that they can

22be used. Isn't that a sign that you want to leave open

23the possibility to bring any type of case, irrespective

24of market shares being rather low?

25MR. SIMS: Well, no, that's not my reason at

218

1least. My reason for being nervous about safe harbors

2unless they're very high is the concern that the safe

3harbor set too low will end up with serious

4over-enforcement above that number.

5MR. BISHOP: Okay, but this comes back to the

6sort of philosophical or institutional, philosophical

7differences between the EU and the U.S., because

8personally, I would set the threshold at 70-80 percent,

9but I'd much prefer in the EU to have one of 40 percent

10than to have no threshold at all.

11MR. SIMS: Okay, and that's a fair point given

12the regulatory environment that you find yourself in.

13MR. WALES: I guess one question I had, Tom, is

14I thought I had read where you talked about the

15possibility of having different thresholds perhaps for

16different types of -- your two types of conduct. You

17had the conduct where someone acts to reduce output on

18their own as opposed to acting to exclude rivals, and I

19guess you kind of left open the proposition I thought

20that perhaps you might be willing to look for markets

21with the former and not the latter.

22MR. KRATTENMAKER: No, I might be willing to

23look for one for each. That's why I said, my objection

24to this is that it -- that the noun is singular.

25DR. WERDEN: Do you have some numbers in mind?

219

1MR. KRATTENMAKER: Do I have numbers in mind?

2No, but I think you might well be able to come up with

3market share based safe harbor for exclusionary conduct

4section 2 cases.

5MR. WALES: I have a question for --

6MR. KRATTENMAKER: But it wouldn't, in my view,

7be an appropriate -- it wouldn't be the same threshold

8that would be appropriate for collusion-based section 2

9type cases, which are generally rare but still can be

10out there.

11MR. WALES: A quick question for Miguel, I guess

12where does 40 come from in terms of setting the

13threshold level in the European Commission?

14MR. de la MANO: Well, as far as I know, it is

15from a case, but, I mean, I think the thing is -- I

16think the discussion is also highlighting this -- there

17is a question as to, you know, what is the threshold

18going to be used for? If you believe that once you are

19above the threshold, basically the case has been proven,

20then clearly you want to have as high a threshold as

21possible.

22If, on the other hand, you believe as I do, at

23least, that the threshold is just the first step, just

24the screen to sort of ditch the cases which are

25obviously not a problem, if you have sufficient

220

1discipline imposed upon yourself as a competition

2authority in what you need to prove further, there is no

3problem in having a low threshold. In fact, it is

4probably better to have a low threshold, because that

5makes the assessment of your facts credible.

6Otherwise, if you have a threshold at a sort of

7middle level, such as 40 or 50 percent, there is always

8going to be a group of people who think, a-ha, okay, so

9this discipline you say you have, that you are going to

10go after -- assessing the effects afterwards, after

11showing dominance, it is not really true, because as

12soon as you are above 50, it is really easy to assess

13the facts, and therefore, there is no credibility to the

14second discipline, as it were.

15MR. BISHOP: Okay, but I would take a different

16view, and sort of just to be clear here, when I said

17that dominance in Europe is then inferred to be an abuse

18of, you know, of that market power, that's not my

19position. That's the position of the European courts,

20that most of the issues we are talking about here are

21exclusionary, and the courts have held that any harm to

22a competitor necessarily leads to harm to competition,

23and therefore, given that sort of standard by the

24European courts, there is no room, really, for an

25effects-based system.

221

1So, as you lower the threshold from 40 percent

2to 25 percent, it makes things much worse in Europe

3unless the Commission is going to be very clear that

4they are going to take on the courts and that court

5reasoning, that you can infer harm to competitors

6necessarily translates to harms to competition, that,

7you know, the Commission is going to take that square

8on, because if they do not, any lowering away from the

940 percent to just come out of case law is just going to

10make things worse.

11MR. de la MANO: The court has already told us a

12few months ago that it is willing to reconsider its

13previous positions on this matter, and in the Glaxo

14decision -- and actually, it is actually an area of

15Article 81, cartels or agreements, but it has made it

16very clear that it is very open and willing to see a

17more effects-based analysis on the part of the

18Commission both in the area of assessing possible harm

19to consumers, but also in the area of assessing

20efficiencies. So, I think the courts are open to be

21challenged by the Commission on this point.

22MR. BISHOP: Well, I would just say, you know,

23let's wait and see, stick with 40 percent and then see

24how they move before lowering the threshold.

25MR. WALES: Let's go to the next one.

222

1DR. WERDEN: Skip the next one and go one

2further.

3MR. KRATTENMAKER: Can we mail in our answers to

4the one, number seven?

5DR. WERDEN: If you like. It is about

6econometrics. Did you want to handle it, Tom?

7MR. KRATTENMAKER: Of course. I mean, that's

8the most fun, is talking about something that we do not

9know. I thought it was a really interesting and

10provocative question. I think it is largely correct,

11but I would have some comments on it, but go ahead.

12DR. WERDEN: We are nearing our end point.

13MR. KRATTENMAKER: No, go ahead.

14DR. WERDEN: As our end point, we are going to

15take this last proposition from the Syufy case, one of

16our failures in court.

17"In evaluating monopoly power, it is not market

18share that counts, but the ability to maintain market

19share."

20 MR. KRATTENMAKER: Could there be anything more

21incorrect?

22DR. WERDEN: I imagine that there could, but let

23me just add that I think what the quote is trying to say

24is the point that Joe made several times, which is

25durability is crucial in monopoly power.

223

1MR. KRATTENMAKER: I see, okay.

2DR. WERDEN: And monopoly power requires much

3more durable power over price than market power does.

4MR. KRATTENMAKER: Gotcha.

5MR. SIMS: When I read this, my answer was, I do

6 not know exactly what these words mean --

7MR. KRATTENMAKER: Okay.

8MR. SIMS: -- but if they mean durable market

9power, then --

10 MR. KRATTENMAKER: If they mean entry barriers

11and -- okay, you are saying they're importing it, okay.

12DR. STELZER: As a practical problem with that,

13it is an easy matter in any case to find someone who

14will tell you why whatever monopoly power or market

15power you see is not durable. I have had people tell me

16that monopoly power in the transmission of electricity

17is not durable because they have some innovation in

18mind.

19 In other words, you can fill the courtroom with

20experts who will tell you why market power that has

21persisted for 150 years is really not durable given some

22new technology or given some new something, but --

23DR. WERDEN: But they're wrong, aren't they?

24But you are saying that they're wrong?

25DR. STELZER: They're wrong.

224

1DR. WERDEN: Okay.

2DR. STELZER: So I would be very careful about

3introducing a test that says not only do you have to

4have market power, but it has to be proved to be durable

5in order to create a problem, because that's an

6impossible test to meet.

7MR. SIMS: It is true, and I think everybody

8should admit that it is true, that the more you get away

9from slogans and general rhetorical concepts and the

10closer you get to careful analysis of the facts, the

11less enforcement you are going to have, because it is

12harder. It is harder for plaintiffs, whether they're

13the Government or private plaintiffs, to prove a case if

14they have to slog their way through the facts.

15That's why the per se rule is so attractive to

16plaintiffs' lawyers in damage cases, because they do not

17have to prove anything. So, you know, that's an

18inevitable result of being more wedded to factual

19analysis than setting up bright-line rules. I don't

20think it is a reason not to do it, but it is a result

21that we ought to be -- that we ought to recognize and

22accept.

23MR. WALES: Anybody else?

24DR. WERDEN: Well, we are a few minutes past our

25official end time, so why don't we wrap it up and take

225

1one last opportunity to thank our panelists.

2(Applause.)

3MR. WALES: Thank you very much. I guess we are

4adjourned.

5(Whereupon, at 4:34 p.m., the hearing was

6adjourned.)

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

226

1C E R T I F I C A T I O N O F R E P O R T E R

2DOCKET/FILE NUMBER: P062106

3CASE TITLE: SECTION 2 HEARING

4DATE: MARCH 7, 2007

5

6I HEREBY CERTIFY that the transcript contained

7herein is a full and accurate transcript of the notes

8taken by me at the hearing on the above cause before the

9FEDERAL TRADE COMMISSION to the best of my knowledge and

10belief.

11

12               DATED: 3/12/2007

13

14

15

16               SUSANNE BERGLING, RMR-CLR

17

18C 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

20I HEREBY CERTIFY that I proofread the transcript

21for accuracy in spelling, hyphenation, punctuation and

22format.

23

24

25               DIANE QUADE