This document is available in two formats: this web page (for browsing content) and PDF (comparable to original document formatting). To view the PDF you will need Acrobat Reader, which may be downloaded from the Adobe site. For an official signed copy, please contact the Antitrust Documents Group.

Page 1

1UNITED STATES FEDERAL TRADE COMMISSION

2and

3UNITED STATES DEPARTMENT OF JUSTICE

4

5

6

7SHERMAN ACT SECTION 2 JOINT HEARING

8PREDATORY PRICING

9THURSDAY, JUNE 22, 2006

10

11

12

13

14HELD AT:

15600 PENNSYLVANIA AVENUE, N.W.

16WASHINGTON, D.C.

17UNITED STATES FEDERAL TRADE COMMISSION

18HEADQUARTERS BUILDING, ROOM 432

199:30 A.M. TO 4:00 P.M.

20

21

22

23

24Reported and transcribed by:

25Susanne Bergling, RMR-CLR

Page 2

1MODERATORS:

2ROBERT POTTER

3Chief, Legal Policy Section

4Antitrust Division, Department of Justice

5and

6PATRICIA SCHULTHEISS

7Attorney

8Bureau of Competition, Federal Trade Commission

9

10PANELISTS:

11Morning Session:

12Patrick Bolton

13Kenneth G. Elzinga

14A. Douglas Melamed

15Janusz Ordover

16

17Afternoon Session:

18Tim Brennan

19John Kirkwood

20Janet L. McDavid

21Steven C. Salop

22Frederick R. Warren-Boulton

23

24

25

Page 3

1C O N T E N T S

2

3MORNING SESSION (SELLING):

4Introduction

5Presentations:

6     Kenneth Elzinga

7     Janusz Ordover

8     Patrick Bolton

9     A. Douglas Melamed

10Moderated Discussion

11Lunch Recess

12

13AFTERNOON SESSION (BUYING):

14Introduction

15Presentations:

16     Tim Brennan

17     John Kirkwood

18     Janet L. McDavid

19     Steven C. Salop

20     Frederick R. Warren-Boulton

21Moderated Discussion

22

23

24

25

Page 4

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

2- - - - -

3MR. POTTER: Thank you for coming, everybody.

4This is the first substantive hearing on predatory

5pricing from the Section 2 hearings. My name is Bob

6Potter. I'm the Chief, Legal Policy Section, Antitrust

7Division, Department of Justice, and I will be the lead

8moderator for this morning's session. Sitting to my

9left is Pat Schultheiss, an attorney with the Federal

10Trade Commission's Bureau of Competition's Office of

11Policy and Coordination. She will be the co-moderator

12for this morning and the lead moderator this afternoon

13on the buy-side predatory pricing.

14Before we start, just a couple of housekeeping

15things that I need to say. One, for the courtesy of the

16audience and the panelists, please turn off any cell

17phones, Blackberries or other devices that may make

18noise during the hearing.

19Second, the restrooms. The men's restroom is

20out the double doors to the left, on your left. The

21ladies restroom is out the double doors, past the

22elevator bank, to the left, and I saw this morning that

23neither of them had hot water, so, if you want hot

24water, you're out of luck.

25MS. SCHULTHEISS: There is no place in the

Page 5

1building that has it right now.

2MR. POTTER: Third, and perhaps most important,

3in the unlikely event that there is an emergency in the

4building, please calmly and quickly go out the doors to

5your right and down the stairs. The Federal Trade

6Commission has a policy of meeting in the Sculpture

7Garden, which is on Constitution Avenue. If you don't

8know where it is, just follow the line of people leaving

9the building, and I am sure you will get there.

10This morning, we are very grateful for having a

11very distinguished panel to talk with us about predatory

12pricing and Section 2. Our panelists are Ken Elzinga,

13Professor Ken Elzinga of the University of Virginia;

14Professor Janusz Ordover of New York University;

15Professor Patrick Bolton of Columbia University; and

16Doug Melamed of the law firm Wilmer Hale and former

17Deputy Assistant Attorney General of the Antitrust

18Division and Acting Assistant Attorney General of the

19Antitrust Division.

20The format for this morning is each of the

21panelists will give a 10 to 15-minute presentation, then

22we will have a short break, and then we will have sort

23of a moderated round table discussion for the rest of

24the time.

25We want to thank the panelists. I'll introduce

Page 6

1them each before their speech as opposed to giving

2everybody's introduction right now, and for the first

3instance, I will tell you that although I'll give you a

4short description, a much longer and better description

5is contained in the biographical information that we

6have.

7Our first speaker this morning is Professor Ken

8Elzinga of the University of Virginia. Professor

9Elzinga is the Robert C. Taylor Professor of Economics

10at UVA. He has a long and distinguished teaching career

11at UVA, having been a faculty member there, although I'm

12sure it doesn't look like it, for over 40 years.

13Even more importantly for today's purposes,

14Professor Elzinga is a creative and prolific academic

15writer, having authored more than 70 economic articles,

16a number of which have focused on predatory pricing.

17In addition, perhaps even more importantly,

18Professor Elzinga has been an expert witness in some of

19the most important predatory pricing cases in the

20history of antitrust, including Brooke Group,

21Matsushita, and most recently, Spirit Airlines.

22With that, please join me in welcoming Professor

23Elzinga.

24DR. ELZINGA: Thank you, Bob. I am going to

25speak from the table here if that's all right, and I

Page 7

1have got 15 minutes, max, to talk about predatory

2pricing. That's a big topic. So, hold on to your

3seats.

4As was mentioned, I was the economic expert for

5the defendants in the last two Supreme Court cases on

6predation, the first one being Matsushita -- that really

7dates me for some people in this crowd -- and then

8Brooke Group or what I still call Liggett v. Brown &

9Williamson, and then also, as was mentioned, I was

10involved more recently in a predatory pricing case,

11Spirit Airlines v. Northwest. I did an economic

12analysis for Spirit, a so-called low-cost carrier. This

13case had a happy landing for Professor Ordover at the

14district court level, it had a happy landing for me at

15the circuit court level, and the final destination of

16this case is still unknown, but I hope to make a few

17remarks about it later.

18When I first started speaking about this

19subject, before a number of you in this room were even

20born, there was not much economic analysis embedded in a

21predatory pricing case. You basically answered two

22questions. Were prices declining in the market -- not

23necessarily below cost, mind you, just going down -- and

24did the defendant generate documents with pugilistic or

25militaristic metaphors? "We are going to cut off their

Page 8

1air supply. We are going to squish them like a bug."

2If I had to pick two events, I am just doing a

3brief intellectual history here, if I had to pick two

4events that changed all this, it would be the Court's

5opinion in Matsushita with its famous line that

6predatory pricing schemes are rarely tried and even more

7rarely successful. That statement was based on the

8Court's exegesis of research about predatory pricing in

9the economics literature. Almost all of this research

10suggested that predation would be a strategy that would

11be difficult to pull off.

12And the second event was the publication of an

13article by Don Turner -- the first Assistant AG to

14enlist an academic economist in the front office, that

15should always be pointed out -- and Phil Areeda in the

16Harvard Law Review. It's the most often cited article

17in antitrust scholarship, led to the Areeda-Turner Test.

18Now, for this audience, I don't need to review

19that article or that test, but let me mention for the

20record how powerful was the hidden economic logic in

21this famous test by using an iconic product from

22Matsushita, a 19-inch black and white portable TV set, a

23consumer electronic products my students today cannot

24even imagine.

25Let's say -- and these numbers are not way

Page 9

1off -- that this set was sold by Toshiba, one of the

2defendants, to Sears for $95, and the average total cost

3was $100, but the average variable cost was $90. So, we

4have ATC equals 100, P equals 95, AVC equals 90. Almost

5everyone at the time believed Toshiba was selling below

6cost. After all, how could Toshiba survive with that

7type of price-cost relationship? And it took an

8instinct for economic reasoning or a recollection of a

9price theory course to realize that such a price was

10above the shut-down point, it was cash flow positive,

11and that Toshiba was better off making the sale to Sears

12than not making that sale, and the Areeda-Turner article

13convinced a lot of people, including a lot of people in

14this building and a building nearby, of something that

15economists have known since Alfred Marshall, and that

16is, in economics, what happens at the margin really does

17matter.

18What was missing from Areeda-Turner was a way of

19thinking about the period of recoupment. They set the

20stage for a more sophisticated -- I did not say highly

21sophisticated -- but a more sophisticated economic

22analysis that the Court adopts in Brooke Group. The

23Court in Brooke Group recognized that even if a firm

24charged a price below cost, whatever was the cost

25benchmark, if the firm couldn't recover its losses, it

Page 10

1was difficult to make a case for antitrust enforcement,

2because the aspiring predator would simply shoot itself

3in the foot if there was no recoupment, and this

4economic logic behind plausible recoupment entailed two

5analytical constructs.

6The first one is real clear in Brooke Group and

7the second one is not transparent. The first is the

8recognition that predation is like a capital

9expenditure. In Brooke Group, the Court cites a paper

10by David Mills and me entitled "Investment in

11Predation." Economists have always recognized that a

12dollar invested today requires more than a dollar in

13future products because of the time value of money, and

14Brooke Group understood that and applied that logic to

15predatory pricing, that losses from predation need to be

16recouped and not just on dollar-for-dollar basis.

17The second point follows from the first: Unless

18entry and exit conditions are symmetrical, the

19recoupment returns for the aspiring monopolist must be

20enjoyed for a longer time period than the time frame in

21which the aspiring monopolist shouldered the cost of the

22predation strategy, and I could just do a footnote here

23on Matsushita and how much the world has changed.

24The plaintiffs in Matsushita thought they were

25making a good case for their side by arguing that the

Page 11

1Japanese charged prices below cost for years and years

2and years, over a decade, not recognizing that the

3longer is that time period, the more difficult it would

4be -- indeed, I think mathematically impossible given

5the power of compound interest -- to ever make up the

6gains.

7For those of you who are attorneys, and that

8would be most of you in this room, I'll tell you what I

9find to be a fascinating war story from Matsushita. I

10did some back-of-the-envelope calculations as to what a

1119-inch black and white TV set would have to sell for

12under the plaintiff's argument that predation had gone

13on for 15 years, that is, these sets had been sold below

14cost for 15 years. What would a 19-inch black and white

15TV set have to sell for? And I found it would be like

16$800 into infinity.

17Now, I don't know if this is one of the things

18that economists talk about when we are not in the

19presence of antitrust lawyers. The antitrust lawyers

20thought, don't ever make that argument on the stand,

21because the plaintiffs will say, well, even the

22professor on the other side says the television sets

23will sell for $800 a year into infinity because of this

24case. And I said, no, that can't be. They can't sell

25for that much. They sell for $100 now. They are not

Page 12

1going to go up to 800, trust me.

2But my point is, the predator wants the period

3of recoupment to be long, not the period of predation to

4be long. The financial rewards that a successful

5predator is going to enjoy is the present value of the

6sum of each period's future return once the target has

7conceded the battle.

8Now, remember, a business firm has some hurdle

9rate or internal rate of return before it signs onto any

10investment project. Signing on for a predatory pricing

11strategy to an economist is no different. The higher is

12the hurdle rate, the bigger and longer the monthly

13returns have to be during the period of recoupment.

14And Grant, if you could show my first slide,

15please.

16In my experience, if one plays with the math

17that I have at the top, which shows the monthly

18sacrifice and the hurdle rate and the time period versus

19the monthly return, it's hard to look at past episodes

20of predation and come up with examples where recoupment

21is mathematically possible. To my mind, when I try to

22teach my students just the basic economics of the

23elementary price theory level class, the important

24asymmetry for predation is the one in the little box at

25the bottom, if you can see it on the slide, slow entry

Page 13

1but quick exits by target firms.

2Putting the math aside, putting even the

3diagrams aside, if there is slow entry but quick exits

4by target firms, then there's a possibility that

5predation can be successful. There's got to be, in

6other words, an economic asymmetry between exit and

7entry conditions in the market, and think about what

8that means. In most markets where entry is easy, exit

9is easy. So, predation simply won't work in those

10markets. And in like fashion, in markets where entry is

11difficult, that helps an aspiring predator, exit will be

12slow, and that is bad news for an aspiring predator.

13So, what the successful predator needs is a market

14setting where exit is quick, but entry or supply

15expansion is slow.

16Now, in the Spirit-Northwest case, one of the

17factors persuading me that predatory pricing was

18plausible or rational for Northwest was because the exit

19of Spirit, that was the target LCC, the target low-cost

20carrier, took place quickly, but re-entry and supply

21expansion was difficult. Spirit Airlines pulled

22capacity out of Detroit quickly when Northwest cut its

23fares in the two markets that Spirit served, but Spirit

24could not enter and expand rapidly during Northwest's

25recoupment period, because Spirit faced an entry barrier

Page 14

1in the form of access to gates at the Detroit Airport.

2Now, I went into the Spirit Airline case as

3someone from Missouri or Chicago, maybe either metaphor

4fits, but I ended up concluding that Spirit was a victim

5of predatory pricing by Northwest, and I'll just say as

6an aside, this is a case in which Fred Kahn should have

7testified and not myself. Professor Kahn knows more

8about the economics of airlines than most any group of

9economists combined, but he was unable to participate,

10though he was convinced that predation took place, as I

11slowly -- kicking and screaming -- came to conclude.

12The pricing trends in the Spirit case are a

13textbook example of what predatory pricing would look

14like. If I could have the first slide, this shows the

15prices in the Philadelphia area -- I think the first

16one -- yes, in Philadelphia. There were two city pairs,

17Detroit, Boston and Philadelphia, and you will see that

18Northwest prices in both of these are high. Spirit

19enters; Northwest prices fall dramatically. Spirit

20exits; Northwest prices jump up. If you show the other

21slide, you will see basically the same scenario.

22Now, these price trends -- I want to stress

23this -- they are merely suggestive. They are not

24dispositive of predatory pricing. Once a pricing

25scenario like this is observed, then there follows the

Page 15

1mind-numbing exercise of comparing revenues with some

2measure of variable costs, and this is a difficult task

3in the best of circumstances. It is by no means simple

4in the airline passenger industry. In the Spirit case,

5this was a battle between Professor Ordover, Janusz, for

6Northwest, and Dr. Dan Kaplan was the economist for

7Spirit. There was also a recoupment analysis done by my

8colleague David Mills.

9Briefly, from my perspective, going back to the

10little box on the bottom of my first slide, one key to

11the success for Northwest was simply how quickly Spirit

12exited and the duration of the recoupment period, and

13that's consistent with the first slide that I presented.

14I was going to show one more slide, but in the

15interests of time, I am going to pass on that.

16Let me conclude this way: Antitrust always has

17surprises. That is one of the reasons I have enjoyed

18being an antitrust economist all these years. Let me

19close by mentioning the surprise for me in the Spirit

20case.

21At the last minute, Spirit's attorneys suggested

22that a price below average total cost but above average

23variable cost could be predatory, and the Circuit Court,

24at the tail end of its opinion, seems to suggest that at

25least in the market circumstances of this case,

Page 16

1Northwest's conduct may have been predatory even if its

2fare structure exceeds, as the Circuit Court put it, and

3I'm quoting here, "an appropriate measure of average

4variable costs."

5Now, Spirit's attorneys were pleased with this

6little present, I am sure. I can restrain my enthusiasm

7for the way the Circuit Court closed out its opinion.

8This might take us into a more European view of

9predation under Article 82, where prices greater than

10average variable costs might be construed as predatory

11and where, as I understand that in Europe, there is a

12continued interest in intent documents and there is no

13recoupment requirement, again, as I understand it.

14Like most economists, I can restrain my

15enthusiasm for the misuse of intent documents. I hold

16the opposite view here of what had been the conventional

17view in antitrust. To me, pugilistic and militaristic

18metaphors are a welcome signal, not of predation, but of

19competition in a market that doesn't have a stodgy "live

20and let live" oligopoly setting, and where you see those

21documents, to me, the prima facie case is that

22consumers, albeit not rivals, but consumers are the

23beneficiaries of head-to-head competition and not

24predation.

25MR. POTTER: Thank you.

Page 17

1DR. ELZINGA: Sure, thank you.

2(Applause.)

3MR. POTTER: Our second speaker today is

4Professor Janusz Ordover. Professor Ordover is a

5Professor of Economics and a former Director of the

6Masters in Economics Program at New York University,

7also Director of Competition Policy Associates in

8Washington, D.C. I first met him when he was the Deputy

9Assistant Attorney General for Economics in the

10Antitrust Division.

11While at the Antitrust Division, Janusz was a

12member of the White House deregulation task force. He

13guided economic analysis of antitrust enforcement and

14acted as a major liaison between the Justice Department

15and various regulatory agencies.

16Professor Ordover has written extensively about

17predatory pricing and has a great deal of experience as

18an expert witness in predatory pricing cases. He was an

19expert for the defendant in the Division's American

20Airlines case, and he is, as Professor Elzinga said, an

21expert in the Spirit versus Northwest case on

22Northwest's behalf.

23Professor Ordover, welcome.

24(Applause.)

25DR. ORDOVER: Well, while we're getting set up,

Page 18

1thank you very much. It's always a pleasure to

2participate in these kinds of hearings.

3Predation, of course, marks a lot of my

4antitrust life. The first time I unveiled my thinking

5on the subject of predation was about 1980 at the FTC

6hearings on predatory conduct, and at that time, I think

7I was attacked -- Professor Willig and I were attacked

8by Frank Easterbrook, David Scheffman and Mike Scherer,

9so essentially from left to right, everybody thought we

10were completely foolish, and Mike Scherer said it was

11the worst antitrust paper ever written, unlike the

12Areeda-Turner paper, obviously, has its own different

13reputation.

14And then, just a few years ago, it was my

15misfortune to fly into Ken Elzinga, who has never seen

16predation other than the case that I was involved in.

17Something is wrong here. So, I don't know what's wrong,

18but I guess maybe I will switch careers in my waning

19years.

20In any case, what I wanted to do today is to

21quickly run through some of the ideas that I have been

22toying with in the antitrust predation field for some

23past 20-some odd years and perhaps follow up on some of

24the comments that Ken made, although I will not try to

25relitigate Spirit versus Northwest. This will have to

Page 19

1await Northwest's exit out of bankruptcy. So, unless

2they get into bankrupt predating, but you never know.

3So, now we are in a holding pattern until somebody

4coughs up some money and we can actually go back and

5litigate the antitrust part of the airline life.

6In any case, what I wanted to do was just go

7through a few slides focusing essentially on some of the

8issues that have been discussed over the years, and that

9is how to analyze challenged conduct from

10monopolization, particularly paying some attention to

11predatory behavior.

12I was going to simply jettison this whole talk

13by simply saying one should have no price predation

14cases, but I thought that would be too quick an exit, so

15I have to torture you for a bit longer to convince you

16maybe that we should think about it as a possible

17solution to our woes in this antitrust patch, without,

18at the same time, suggesting that we should throw out

19all kinds of scrutiny of firms' conduct, which consists

20of much more sophisticated pricing from other aspects of

21what they do, behavior, what I would often call

22competitive response package, which is I think a term I

23coined for my testimony in U.S. V. American Airlines,

24where actually American Airlines' behavior was not just

25simply pricing but involved a lot of other things that

Page 20

1the Government alleged were designed to, in fact, retain

2or maintain or defend American Airlines' position at its

3hub, Dallas-Fort Worth.

4So, I'm always thinking about competitive

5response packages as strategies designed either to exit

6the rival or to prevent the rival from coming in or

7possibly designed to contain the rival, and I think it's

8the last category of strategies which I believe is of

9great interest and perhaps should be given a little bit

10more time than we often do.

11But in any case, the question becomes, how

12should the decision-maker delineate conduct that does

13not harm competition by harming scarce rivals from

14standard, day-to-day market interactions? And

15economists have been pulling their hairs out since

16Areeda-Turner, 1975 paper, so we are now in 31 years, 30

17years of thinking about it, and there is no solution as

18evidenced by the articles in the latest Antitrust Law

19Journal, where everybody is still fighting over

20important things but without actually coming to any

21particular conclusion.

22I have been associated over the years with

23something called the sacrifice test, but I always

24thought of sacrifice test actually as a version of the

25welfare test. In other words, what attracted me and

Page 21

1Professor Willig to thinking about the so-called profit

2sacrifice approach to delineating procompetitive versus

3anticompetitive conduct, or at least neutral from

4anticompetitive conduct, was the notion that at least in

5some well-defined range of circumstances, these two

6tests ought to give a pretty close set of answers.

7In other words, that one was not -- that is, the

8sacrifice test -- was not somehow biased, setting aside

9the difficulties of implementation, but it somehow was

10not biased one way or the other against deterring what

11would be anticompetitive conduct or what would not be

12anticompetitive conduct, relating to too much conduct

13that, in fact, would be harmful. We have been able to

14show in a variety of circumstances -- in fact, these two

15tests coincide for the very simple reason that a pursuit

16of profit, which is the engine of market economies, in

17fact, is a kind of behavior that generally or frequently

18does, in fact, conduce to welfare maximization. Seeking

19profit is a good thing, it is not a bad thing, and

20therefore, it is not surprising that if you write down

21your economic model correctly, or at least correctly for

22the purposes at hand here, that in many circumstances,

23these two tests will give you the same kind of answer.

24So, there might be, however, a range of

25circumstances in which these two tests fall apart by

Page 22

1virtue of the fact that the basic condition under which

2they do coincide is potentially difficult to meet, and

3that condition is incomprehensively stated as the third

4bullet on this slide, but the basic idea here is that if

5the incumbent firm can effectively, without creating

6additional distortions, extract profits by its pricing

7strategies and other strategies, then any strategy that

8actually lowers the profits relative to that extraction

9ought to signal, at least as a first step to the

10analysis, ought to signal that a firm may have some

11other aims in pursuing that strategy, something that I

12think Bernheim and Whinston have now been calling over

13the years as trying to create market power in what's

14called a noncoincident market, okay?

15So, the action takes place in market A, assuming

16we have it well defined, but the goal essentially turns

17out to be gaining incremental power or preventing

18erosion of power in some other market, which Schullman

19called a noncoincident market, let's say, which could

20be, in Areeda-Turner world, it could be the same market

21but in the future day, okay? So, what's the meaning of

22noncoincident market is actually a little loose, but

23that's the term that at least Berheim and Whinston in

24their fine unpublished monograph on exclusionary

25behavior utilized as a view for analyzing this kind of

Page 23

1situation.

2So, it could be a setting in which the whole

3thing works beautifully. An example is an inferior

4source of supply, this is the second thing which I think

5is quite ubiquitous, in which the incumbent firm is

6faced with competition from another firm or a firm that

7constrains its ability to exercise pricing freedom,

8which provides an inferior product, and therefore,

9enables the incumbent firm to earn supra-competitive

10profits, at least profits higher than some rents, but

11getting rid of that firm would, in fact, lift the

12ceiling and therefore would enable the firm to raise the

13price even higher.

14The problem turns out to be that maybe exiting

15that firm may be just very difficult; however, a

16circumstance that we have analyzed, Willig and I, under

17the rubric of systems competition, informs a view of the

18circumstance in which actually disabling a component

19that the other firm needs in order to be a full-fledged

20parcel, bundle and bundle competitor with the incumbent

21firm will, indeed, lift the ceiling and therefore enable

22the firm to exercise incremental market power. So, the

23idea that we have pursued, and the idea which I think is

24actually fruitful, is that in many circumstances, the

25goal of the competitive response package is not

Page 24

1necessarily to kill or to weaken or to disable the

2person that or the firm against whom this conduct is

3being perpetrated, but rather, to try to lessen or

4weaken some other kind of restraint which cooperates

5complimentarily with the firm whose market presence is

6being weakened.

7I think if you look at these Microsoft cases,

8some of which were discussed along the same lines, this

9is a fruitful way of thinking about it, but you can

10immediately see that the economics of the situation is

11much more difficult than the one instance in the

12Areeda-Turner case, which is drop the price below some

13level of cost, you go perhaps profit-negative, assuming

14you know how to calculate profits, you know how to

15calculate revenues maybe, you know how to calculate

16costs maybe, and you can compare the two and see what

17happens, you are losing money, and as a result of which,

18it is anticompetitive.

19But in the situation like this, you don't have

20to be losing money on anything unless you try to look at

21the situation in a somewhat different way, which is

22where the efficient component pricing rule tells you how

23to look at that situation that I have just described.

24The efficient component pricing rule for those of you

25who are not regulatory freaks like myself is a rule that

Page 25

1tells you what the price of a scarce bottleneck should

2be if it does not involve any kind of profit sacrifice,

3okay? So, ECPR is a way of thinking about pricing

4access, pricing access to the component that is needed

5in order for the firm to be a viable component system or

6system competitor.

7Another example along the same lines, which

8again focuses on a complicated pricing strategy, not

9simply dropping price below some measure of cost, was

10discussed in Ortho v. Abbott. Actually, I worked for

11Ortho in that case, and there the situation was, again,

12packaged pricing of a very interesting kind, in part

13interesting because the buyer insisted on firms offering

14not only unbundled pricing, but also bundled pricing

15with a different number of components put in. The buyer

16needed to buy five tests. There was a regulatory

17presence out there that required that every blood

18screening used five tests at that time, I think now it's

19six, and Abbott was the one that could offer five of

20them, Ortho could only offer three.

21Then the question was, could Ortho compete

22against Abbott if it did not get the access to the

23remaining two, either because the buyer could create the

24bundle or because Ortho could buy the necessary input

25and then resell the bundle? Again, in this case, it

Page 26

1turns out that there is some discussion that potentially

2Abbott was pricing the incremental two tests at levels

3that were unprofitable, that violated some version of

4what we called a second ago the efficient component

5pricing rule.

6What was very complicated in that case was, A,

7that Ortho did not give me any cost data. So, I

8couldn't say anything, whether it was true or not, but I

9did derive the test on a napkin, so other than the

10Laffer Curve that also was derived on a napkin, this is

11probably the second most famous napkin in the history of

12economics. But in any case, the point I'm making is

13that in this case at least we had a way of dealing with

14an issue, but we had no reason to explain why this was

15going on, and I think that's a very important aspect of

16any predation case, which is that the plaintiff makes a

17clear connection between the conduct that is at issue

18and the anticompetitive impact that is being challenged

19as leading to this anticompetitive outcome down the

20road.

21Virgin versus BA, another complicated case that

22pitted Bernheim against Schmalensee, actually a

23beautiful battle -- I think it was Schmalensee -- of

24battle in IO, in which, again, there was no simple

25pricing strategy, but rather, a complex pricing strategy

Page 27

1that Bernheim showed leads to an equilibrium in which

2there's relatively cheap exclusion but in which no price

3is technically below marginal cost, simply understood,

4yet as we know, all of these tests that we have in front

5of us do have some flavor of comparing something to

6cost, and again, what Bernheim tried to demonstrate in

7that case, that a simple comparison of price to

8something like marginal cost may be a flawed way to go

9if you put that pricing in a strategy that British

10Airways allegedly developed in a broader context.

11Quantity-forcing contracts, I think we will skip

12that, only because we have to, A, rush, and B, we will

13talk about it in the fall, so I am going to skip that

14unless it comes up in questions.

15Just because I don't believe that true price

16predation is an antitrust offense that is of great

17interest, it does not mean that we as economists and you

18as enforcers do not have plenty to focus on. I believe

19that business strategies, these competitive response

20packages, that have a strong commitment value, are

21actually a more relevant focus than just simply pure

22price predation, which creates all types of problems as

23these papers in ALJ demonstrate.

24Commitment to discount, which is Virgin versus

25BA, commitment to product design, commitment to defend

Page 28

1lucrative markets, which I call the "new era" tying

2models, network economies, commitments to effectively

3raising rival's cost of competing, are the types of

4strategies that we are now slowly beginning to

5understand with the help of very fancy economic models

6and beautiful game theory.

7The question that I think we will have to leave

8for Patrick to help us answer is whether or not we can

9actually fashion workable tests that will take into

10account these kinds of complications that economists

11have been focusing on.

12Thank you.

13(Applause.)

14MR. POTTER: Thank you, Janusz.

15Our next speaker is Professor Patrick Bolton.

16Professor Bolton is the David Zalaznick Professor of

17Business. He began as Assistant Professor at U-Cal

18Berkeley, then moved to Harvard. Then he was the John

19H. Scully Professor of Finance and Economics at

20Princeton University.

21Professor Bolton's research and areas of

22interest are in contract theory and contracting issues

23in corporate finance and industrial organization. One

24of his particular areas of research is the impact of

25strategic economic game models on predatory pricing

Page 29

1theory.

2Professor Bolton, welcome.

3DR. BOLTON: Thank you, Bob. It's a pleasure

4and an honor to be on this panel.

5Unlike Professors Elzinga and Ordover, I have no

6experience as an expert, haven't had that pleasure, and

7if you want, I'm a new entrant. We will see whether

8this will elicit predatory response from the economists.

9So, my interest, as Robert just alluded to, my

10interest in this topic came from reading the original

11McGee article, which claimed that predation couldn't be

12a rational economic strategy, and, you know, I read this

13article again and again, and I just was not convinced,

14and this led me later on to write a theory piece with

15David Scharfstein where we outlined how predation could

16be a rational strategy if it took the form of financial

17predation, and I will say a little bit more about that

18in my presentation.

19And then later, I had the good fortune of

20meeting with Joe Brodley, who introduced me to the new

21developments in policy under Brooke Group and

22highlighted some of the problems with the new policy and

23also some of the new opportunities and challenges, and

24that then led to our, in my view, very fruitful

25collaboration on our article, which I will make the

Page 30

1centerpiece of my brief presentation today.

2So, I thought I would start by saying first, you

3know, where are potential areas of agreement among

4economists and legal scholars and where there are still

5areas of disagreement. I would argue that this is

6relatively easy, that we are all in agreement on the

7general definition on predatory pricing. Namely, it's a

8price reduction that is only profitable because of the

9added market power the predator gains from eliminating,

10disciplining or inhibiting the competitive conduct, and

11to summarize what both Professors Elzinga and Ordover

12said earlier, you can distinguish two phases in any

13predatory pricing episode, a sacrifice phase and a

14recoupment phase. As Professor Elzinga wrote elsewhere,

15you can think of predation as an investment in market

16power. So, I would say that there is general agreement

17on this characterization.

18Where there is more disagreement is on policy,

19and, well, there had been long disagreements on basic

20economic premise, whether predation is an economically

21rational strategy and how prevalent predatory pricing

22episodes are. My sense is that this is an area of

23convergence, at least on the first bullet point. I

24think nowadays it is more and more widely accepted that

25predation can be an economically rational strategy.

Page 31

1On the second bullet point, I think there are

2still some areas of disagreement, but I would argue that

3over time, things have moved in the direction of

4thinking of predatory pricing as being more prevalent

5than we thought before and also more likely to succeed

6than we thought before, in part because our initial

7beliefs were built on writing, McGee's writing,

8suggesting that it couldn't be rational, and those

9writings, I would argue, are now obsolete.

10There are, however, still very sharp

11disagreements on the legal standard. Some people argue

12that we should have simple rules. Others have argued

13that we should always err on the side of

14under-deterrence to reduce the risk of false positives,

15and the policy under Brooke Group is characterized as

16both being simple and under-deterring. I would argue

17against this.

18Now, let me skip the description of Brooke

19Group, because I imagine most of you are familiar with

20it, so it involves both a cost test and a recoupment

21test, and let me emphasize potential problems first with

22the new policy, and namely, when we look at the facts on

23what happened post-Brooke, what we find is that since

24Brooke, plaintiffs have not prevailed in a single case,

25and almost all cases have been decided by summary

Page 32

1judgment, and it is only very recently that we are

2seeing some action on predatory pricing, particularly in

3the case of Spirit versus Northwest.

4So, what are the problems with the present

5policy? Well, first of all, and I think we will discuss

6this later on in the question time, I would argue that

7the basic problem with the present policy is that the

8cost test is highly unreliable. Professor Elzinga

9earlier qualified proving a cost test as a mind-numbing

10exercise. I would fully agree with that. I would say

11that when you go into the details of trying to prove a

12cost test, you will lose track of the economics of the

13problem and of the case, and in particular, a very

14narrow interpretation of the cost test, price being

15below average variable cost, is a very poor proxy for

16measuring profit sacrifice, which is what we are trying

17to go after.

18Another problem with current policy, we have

19never gone to a point where we had to ask about a

20possible efficiency defense on the part of the

21defendant. There has never been any talk of applying

22the same rigorous recoupment criteria that the plaintiff

23has to fulfill on the defendant in proving an efficiency

24defense. I would argue we should go in that direction.

25But just to emphasize, I think that the major

Page 33

1problem with present policy is its failure to focus on

2the main issues, and those are what is the predatory

3strategy, what strategy drives alleged predation, first

4of all, and second, what are the possible dynamic

5efficiencies and how do you balance procompetitive and

6predatory effects? And this is where our article takes

7off and proposes an alternative approach, which I would

8summarize as taking away some weight off the cost test

9and emphasizing instead intent, bringing back intent,

10but intent as structured by an economic analysis, and so

11this is what in my short time I want to briefly go into.

12So, specifically, we are thinking that any

13approach based on intent should be based on strategic

14analysis of predatory pricing, and in our article, we

15emphasize at least two well-proven strategies, which are

16financial market predation and reputation effect

17predation. We also discuss test market predation. Of

18course, as Professor Ordover highlighted, predation can

19take many different and complex forms, and one should

20not necessarily reduce one's self to just those few

21strategies, and one should allow for any

22well-articulated and rational strategy that might be

23used. I might comment on that later on.

24Anyway, so what we argue in our paper is that

25this approach has two advantages. One is that it can

Page 34

1reduce the risk of false positives, and second, that it

2puts the spotlight back on what we are really trying to

3determine, which is discriminate between procompetitive

4and anticompetitive effects, and there we can use intent

5as our guide, evidence of intent as a guide to possible

6defense, and what I mean by intent is not what Professor

7Elzinga has referred to as militaristic and pugilistic

8language, but evidence of a deliberate effort to exclude

9and evidence of pursuit of a predatory strategy.

10So, in our article, we outline five legal

11elements to a predatory pricing test. Let me enumerate

12them first, and then I will go into some of them in more

13detail. The first element, which is straightforward, is

14there should be a facilitating market structure. The

15second element is the scheme of predation and supporting

16evidence. Third, probable recoupment. Fourth, price

17below cost. And those four elements would constitute a

18prima facie case of predatory pricing.

19I have put the fourth element in brackets here

20to emphasize the fact that we try to de-emphasize the

21cost test, and we would agree with the appeals court

22opinion in the Spirit Airlines case that predatory

23pricing which is above some measure of average variable

24cost but below average total cost, that kind of pricing

25could be predatory. Then, however, we add, if you

Page 35

1de-emphasize the cost test, we want to add as a safe

2harbor the -- allow for an efficiency defense.

3So, how do we prove those elements? Well, some

4of them are straightforward, and I will not go into -

5so, facilitating market structure is any evidence of

6market power. The scheme of predation and supporting

7evidence, I want to give you an example of how you go

8about doing this. So, I will in particular take out of

9our article the example we have on financial market

10predation, and so under this element, what is important

11is to establish that the conditions to implement a given

12strategy are present and to provide direct or

13circumstantial evidence showing that this strategy is

14being implemented.

15Recoupment, again, this is relatively

16straightforward. You would want to show evidence of

17exclusion and disciplining of rivals, and we stressed

18the idea that second, that you should emphasize probable

19recoupment instead of actual recoupment, because what

20matters is whether at the time when this strategy was

21being chosen, whether at that time, at the time of the

22information the incumbent had at that time, whether it

23made sense to implement such a strategy, and we know, as

24in our own investments in finance, we know that at the

25time when we make a decision of investment, we make an

Page 36

1analysis using this kind of cash flows that suggests

2that we have a positive net present value investment,

3but that is no guarantee that when we actually undertake

4the investment, it will end up being profitable. So, we

5would emphasize probable recoupment, and in particular,

6put a lot of weight on market structure that makes

7recoupment likely in the future.

8Let me also emphasize here the "or related" in

9brackets, and this is a point that Professor Ordover

10emphasized, that recoupment shouldn't just be seen in

11the narrow market where predation takes place. It could

12be obtained in a related -- I forget the term you

13used -

14DR. ORDOVER: Noncoincident.

15DR. BOLTON: -- noncoincident market.

16On price below cost, I do not have much to add

17to what I have said already except that in the paper we

18emphasize that a better measure than average variable

19cost would be average avoidable cost, and a better

20measure for long-run average cost would be long-run

21average incremental cost. I do not want to go further

22into this, because making fine distinctions about these

23definitions could end up being a mind-numbing exercise,

24and it just highlights the difficulty with applying the

25cost test.

Page 37

1So, what I would like to emphasize, though, is

2that we would argue that failure to meet the cost test,

3in particular, failure to establish pricing below

4average variable cost, should not be grounds for a

5dismissal on summary judgment and that, in fact, the way

6to go would be to balance the cost test with an

7efficiency defense. So, I would argue that if you are

8able to show that there was pricing below average total

9cost but above average variable cost but that there was

10absolutely no efficiency defense, plausible efficiency

11defense provided, that that would then make a strong

12case for predatory pricing.

13So, the efficiency defense, we spent a lot of

14time in the paper on that, because one of the weaknesses

15of the policy under Brooke Group and the Areeda-Turner

16Test is that it really neglects looking at efficiencies,

17and so we would argue that an efficiency defense does

18provide safe harbor in itself for price competition that

19benefits consumers, and we distinguish between defensive

20defenses and market-expanding defenses and provide in

21the paper an approach to proving those defenses. So,

22defensive defenses, we mean by unilateral best response

23mainly and minimizing losses from unexpected market

24developments, and as for market-expanding defenses, we

25really mean here promotional pricing, learning-by-doing,

Page 38

1and network externalities.

2So, let me move on perhaps in the few minutes

3that I have left with an illustrative example. How do

4you prove financial market predation in a particular

5case? So, very briefly, the theory here, you know, what

6is financial market predation, why does it work?

7Well, the reason why it works is because in

8corporate finance, there are imperfections -- and there

9is enormous literature on this -- there are

10imperfections in capital markets due to agency problems

11in lending, and as I have argued and have written in my

12paper with David Scharfstein, a predator can take

13advantage of those imperfections and drive out an

14entrant by basically drying up financing.

15So, how do you go about proving financial market

16predation? So, we distinguish five essential

17preconditions. One, the prey's dependent on outside

18financing. The prey's outside funding depends on its

19cash flow. Three, predation will reduce the prey's cash

20flow sufficient to threaten its continued viability.

21All these are fairly straightforward. Four, the

22predator knows of the prey's dependence on outside

23funding or can be assumed to know based on easily

24accessible facts or rational conjecture. And five, the

25predator can finance predation internally or has

Page 39

1substantially better access to external credit than the

2prey.

3I think in the Spirit Airlines case, I quickly

4looked at it, most of these elements you would be able

5to establish.

6So, the example we have in the paper is about

7entry into the cable TV market in Sacramento. This is a

8case that predates Brooke, and here are the facts. So,

9this is an entrant with outside financing amounting to

10$6 million, entered in a small district in the

11Sacramento area, the Arden District, serving 5000 homes,

12and the entrant's intention was, of course, to reach a

13bigger market share and expand gradually in the

14Sacramento area. The incumbent Sacramento TV company

15responded to this entrant with drastic price-cutting,

16and after eight months, the entrant exited. So, how

17would we prove a scheme of financial predation here?

18Well, first of all, the dependence on outside

19funding, what do we know? What are the facts here?

20Well, first of all, the prey obtained funds through a

21loan, and the entrant's owners were unwilling to commit

22more capital than they had initially. Secondly, outside

23financing depends on cash flow. Well, the incumbent

24targeted its price reductions on the entrant's customers

25and potential customers, and that obviously had the

Page 40

1effect of reducing cash flow. Predation will reduce

2cash flow and threaten viability. Again, that is easy

3to establish in this case.

4The predator knows of the prey's dependence on

5outside funding. Well, here it turns out there is

6evidence, intent evidence, memorandum from the

7incumbent's files that speaks of sending a message to

8the entrant's bankers. Well, that's relatively easy to

9establish here. And then finally, the predator has

10better access to credit than the prey. Again, that is

11an easy proof in this particular case.

12So, let me -- sorry for having stepped over my

13time -- so, let me just quickly conclude with

14highlighting one potential concern with our approach,

15and that is something that Posner mentions in his second

16edition of his antitrust book, and he argues that one

17concern one might have with evidence of intent is that

18it's really "a function of luck and of defendant's legal

19sophistication." So, we would argue that this concern

20is reduced if the plaintiff is also required to prove,

21as we articulate in our article, all the other elements,

22and if what you are required to establish is the

23implementability of a rational predatory strategy.

24So, let me end with that.

25MR. POTTER: Thank you very much.

Page 41

1(Applause.)

2MR. POTTER: Our final speaker today is also the

3only lawyer on the panel, although Doug is very used to

4dealing with economists, so I am sure it will not be a

5problem for him to follow them.

6Doug is a partner at Wilmer Hale, and he is the

7co-chair of the firm's Antitrust and Competition

8Department. He has significant experience in a number

9of government investigations, both government and

10private litigation, substantial antitrust counseling,

11and some of that counseling in investigatory work, in

12litigation work, has involved predatory pricing.

13From 1996 to 2001, Mr. Melamed served as the

14Principal Deputy Assistant Attorney General in the

15Antitrust Division and then as the Acting Assistant

16Attorney General in the Antitrust Division. He's a

17prolific writer, a frequent speaker, always has

18interesting viewpoints that are well thought out. His

19most recent -

20MR. MELAMED: Don't raise the bar, please.

21MR. POTTER: -- his most recent article, which

22appears in the summer 2006 Antitrust Law Journal

23provides a thought-provoking commentary on whether there

24are unifying principles under Section 2.

25Mr. Melamed, welcome.

Page 42

1MR. MELAMED: Thank you.

2Well, I am a lawyer, and much though I enjoy

3listening to economists and talking to them, I am going

4to be talking as a lawyer now and giving a lawyer's

5perspective on some aspects of the predatory pricing

6issue.

7Let me start by saying, I think Brooke Group was

8correctly decided, an important decision, it brought

9needed rigor and order to predatory pricing law, but I

10am concerned about what has happened to it in the life

11of the law. There is a kind of -- I do not know if this

12is the right word -- a kind of rarefaction of Brooke

13Group that I think has done some mischief, and let me

14tell you what I mean.

15As everyone knows, Brooke Group has proven to be

16a defendant friendly standard. As Professor Bolton

17noted, no plaintiff has won a predatory pricing case

18post-Brooke Group. Not surprisingly, therefore, when

19price is an element of the allegedly unlawful strategy,

20the defendant argues that the standard to be applied by

21the Court should be Brooke Group, and, of course, they

22are entitled to do that, because if that's the law, they

23ought to make that argument, and certainly I have done

24that myself.

25But if it is not a straightforward price-cutting

Page 43

1case, if it is a little complicated, the plaintiff says,

2"No, no, no, this was different, bundled discounts,

3aggressive buying, low prices conditioned on exclusivity

4or other preferential treatment and so on." So, you

5have a legal dispute. Does Brooke Group apply? Is this

6the right category, predatory pricing, in which Brooke

7Group applies, or does the conduct at issue belong in a

8different category?

9And there is a kind of a notion that there is an

10apparent precision of Brooke Group, the price-cost test

11and the recoupment test, that is uniquely valuable but

12uniquely applicable to predatory pricing, and one

13consequence of this is that when the Court decides in

14this kind of stovepipe analysis that the conduct before

15it really should not be considered predatory pricing,

16too often, courts seem to find themselves in a kind of

17"deer in the headlights, what do I do now" posture, and

18the result is incoherent decisions like LePage's or

19courts affirming nonsensical and meaningless jury

20instructions like Weyerhaeuser and basically a casting

21about in the way that Professor Elhauge had spoke of

22Section 2 as a kind of incoherent mess.

23I think this stovepipe or essentialist way of

24looking at predatory pricing has created these kinds of

25dichotomies as categorization, and it has inhibited the

Page 44

1development of a more robust antitrust jurisprudence,

2one that can help courts make reasoned decisions about

3conduct that they do not think falls into a precise,

4well-established category, whether it be exclusive

5dealing or predatory pricing or whatever.

6Put differently, instead of inducing from Brooke

7Group principles of broader application in the kind of

8common law tradition which antitrust has in other

9contexts involved, the process seems to have separated

10predatory pricing from other forms of exclusionary

11conduct, and it's done so because there has been in what

12I call this rarefaction a number of propositions about

13predatory pricing that are taken for granted or thought

14to be true or thought to be unique to predatory pricing,

15and I want to express some skepticism about that. There

16is a lot of these propositions I have in mind, four or

17maybe three depending on my time, and I want to express

18skepticism either that they are true or that they are

19unique to predatory pricing or perhaps both.

20So, proposition one, to apply the price-cost

21test, we need to select some term of art from the

22economists as our measure of cost, average variable cost

23or something like that. Now, this is a big topic. I

24will make just a couple observations.

25Almost everyone seems to agree that some kind of

Page 45

1incremental cost is the right measure, because we want

2to know whether the allegedly predatory sales cost so

3much that either the defendant must have intended some

4predatory scheme or, at the very least, that the cost of

5the sales exceeded the amount consumers were willing to

6pay for them and therefore resulted in a welfare loss.

7Areeda and Turner say, "Well, marginal cost is

8the right test, but it's hard to prove, so let's use

9average variable cost as a proxy," and now we have this

10debate for 30 years, "Well, average variable cost really

11isn't a good proxy, we should use average long-run

12incremental cost or average total cost, may depend on

13the circumstances," and you all probably read the

14article, too, the discussion paper which went through

15this discussion at great length.

16Why are we even having this conversation? Why

17are we debating these categories about technical

18economic jargon that might have made sense in the

19Areeda-Turner world in 1975, a simple static price

20series model, and you can draw the ABC curve, the

21marginal cost curve, and you can talk about these

22metaphors, what's going on in the real word, but that

23doesn't make any sense in the real world as I have

24experienced it as a lawyer.

25Areeda talked about additional increments of

Page 46

1output. I have rarely had a client say to me, "I'm

2thinking of pushing more widgets off my production line.

3How low can I go in price?" That's not how the problem

4comes up in the real world, and if it looks like that,

5there's a lot more going on.

6The kind of predatory pricing problems I've

7counseled clients on in recent years are things like

8this: Price offerings to early adopters in a de facto

9standards war; prices in two-sided markets; decisions to

10assign a plant or an airplane to one market or one

11segment rather than another. In these situations, I

12think these terms of art that economists have, they are

13very valuable in their models and their heuristic

14exercises, don't have much value, and even if they have

15value to the economists, they don't have much value to

16the lawyer and the client.

17What I find is valuable is saying to the client,

18when I'm talking about costs, "What are the costs you

19are incurring to engage in the strategy at issue that

20you wouldn't otherwise have incurred?" Clients

21understand that question, and it's not always a trivial

22question, but I think it's one they can answer. So, I

23think avoidable costs -- and I don't mean that as some

24technical term, I mean simply as the but for costs of

25the allegedly unlawful conduct -- is the cost measure,

Page 47

1okay?

2Proposition two, price-cutting is beneficial to

3consumers, so we should therefore have a standard that

4errs in favor of avoiding false positives. Then Judge

5Breyer, in the wonderful "bird in the hand" metaphor, I

6think most famously perhaps articulated that.

7Here is my concern: Sure, price-cutting is good

8for consumers, no question about that. So are all sorts

9of other things that companies do for consumers. In

10fact, as I understand, from what the economists tell us,

11that innovation does a lot more for welfare than

12improving allocative efficiency by some price cuts and

13supra-competitive down toward competitive levels. So,

14why don't we -- and innovation, by the way, could be

15inventing the PC or it could be coming up with an

16improved method of distribution because of tying

17arrangements or because of exclusive dealing. It could

18be anything that improves the value of the product to

19consumers.

20In fact, cost of sale reductions could be

21beneficial certainly to a total welfare sense and

22ultimately to consumers as well. So, why do we single

23out price-cutting, which I don't think has any unique

24benefits to consumers?

25Now, there is one thing about price-cutting that

Page 48

1is different, and that is it's unambiguously in the

2interests of consumers. A product improvement, you

3know, the car with the air conditioner might look like

4it's better, but maybe consumers would rather have

5better mileage. So, there is some ambiguity about

6whether other forms of conduct benefit consumers, but

7why do we have a legal superstructure built on the

8premise that pricing is unique?

9At some point, if we do that simply because it's

10easier to identify the consumer benefit, don't we begin

11to look like the economists searching for the keys under

12the light post? At the very least, when the defendant

13is able to show that his conduct is benefiting

14consumers, why treat predatory pricing any differently?

15Proposition three, the recoupment requirement is

16central to and a great contribution to predatory pricing

17law. Let me be clear. I strongly believe there should

18be something like a recoupment requirement at least in

19the sense of a market power screen; that is to say, a

20plaintiff ought to have to prove that the allegedly

21predatory scheme will pay off for the defendant by

22creating additional market power or preserving market

23power that will guard against -- kind of belt and

24suspenders -- a mistake in the application of the

25price-cost test, and it will preserve antitrust

Page 49

1violations for those cases where there is competitive

2harm, and we won't worry about the others.

3I think, in fact, there should be such a screen

4in all cases of exclusionary conduct. The problem is, I

5think in many quarters, including some of my

6predecessors this morning, the recoupment test is

7understood to mean that the plaintiff should prove,

8should quantify, the defendant's investment in the

9predatory strategy and then quantify his

10supracompetitive returns during the recoupment period,

11discount them by risk and uncertainty and time, and

12conclude that the recoupment exceeds the investment.

13Now, I think evidence of that sort, on that

14issue, whether introduced by the plaintiff or the

15defendant, should be relevant in a predatory pricing

16case, because it certainly illuminates the likelihood

17that what is going on here is some exclusionary conduct,

18but I am very skeptical of the notion that that should

19be an element of the offense. It clearly complicates

20the proceedings, increases costs. It may be an

21impossible burden for the plaintiff in a multi-market

22reputation effect recoupment story.

23If taken literally, you would have to go to a

24profit-maximizing standard to figure out the defendant's

25investment in the predatory strategy, because you

Page 50

1wouldn't be asking simply what did it cost him below

2cost, you would be asking how much in profits did he

3sacrifice. It's not necessary in order to identify

4anticompetitive conduct, because if we think we got the

5price-cost test right and the guy is selling below cost,

6you can actually, it seems to me, infer that he expects

7to recoup. It's not needed, because the market power

8screen will identify the cases of competitive harm. And

9finally -- and this is a point that I don't know that

10it's original to me, but I haven't seen it before -- I

11think it is an illusion that we're measuring something

12about the welfare effects of the conduct when we use a

13recoupment screen.

14The welfare question in predatory pricing is

15whether the welfare gains, consumer or total, during the

16rivalry period, the competitive period, are greater than

17or less than the welfare costs, consumer or total,

18during the recoupment period, but the recoupment test

19doesn't measure either of those. The recoupment test

20measures producer surplus in the competitive period

21versus producer surplus in the recoupment period, and it

22doesn't take a whole lot of imagination to think of

23situations where the results could be different, where

24you could have, for example, recoupment but no welfare

25loss from an allegedly predatory strategy.

Page 51

1So, proposition four, in applying the Brooke

2Group price-cost test, the price of the product at issue

3is the appropriate price to compare to cost. That in my

4view is only partially correct. Obviously you look at

5the price of or the revenues generated by the additional

6sales attributable to the predatory conduct. You don't

7look at the price of, of course, the inframarginal

8units, the units that would have been sold anyhow,

9because those units didn't exclude the rival or at least

10they didn't exclude them by reason of the

11anticompetitive conduct.

12But that's not all there is to it. Suppose

13we're in a two-sided market. Suppose you're cutting

14price on circulation of the newspaper in order to

15generate more readers and therefore more advertising

16revenues. Surely you want to take into account the

17incremental advertising revenues.

18Suppose you have complimentary revenues. You

19know, the Government didn't accuse Microsoft of

20predatory pricing because the browser was free when

21bundled with the operating system. Because it was a

22plausible story that it increased revenues, we didn't -

23increased revenues for the operating system. What about

24revenues lost from inframarginal sales; that is to say,

25the sales that the defendant would have made anyhow even

Page 52

1if he had not engaged in the predatory scheme, but he

2would have made them at a higher price if he hadn't cut

3prices?

4To me, another way of putting that question is,

5are we concerned with the incremental revenues or the

6revenues from incremental sales? The law chooses wisely

7in my view the latter. It ignores the loss of

8inframarginal revenues, I think -- I know Professor

9Bolton may disagree with this -- I think the law wisely

10ignores that, because if you want to go into those lost

11inframarginal revenues, you have to have a profit

12maximization test, you know, what would have been the

13profit-maximizing outcome of the strategy, and that is

14in most cases going to be virtually impossible it seems

15to me for the Court to figure out and surely impossible

16for the firm to figure out in real time when it's trying

17to comply with the law.

18As implied by my discussion a minute ago with

19the recoupment test, it's not going to correlate with

20the welfare trade-offs you are looking at, although it

21may illuminate a little bit, but most important, it

22seems to me, is that price cuts on the inframarginal

23purchases, price cuts until they are below some measure

24of cost for the incremental units, enhance welfare, and

25they enhance efficiency, and we all know that story,

Page 53

1right, going toward the competitive outcome, and you

2reduce dead weight loss.

3So, it seems to me that we ought to ignore

4inframarginal revenues. I didn't mention this earlier,

5I meant to say it, but costs ought to include

6opportunity costs, ought to include the cost of moving

7allocating assets to the predatory scheme rather than

8somewhere else. That's part of the avoidable cost it

9seems to me.

10Forgoing inframarginal revenues in my view

11shouldn't be treated as an opportunity cost, at least

12not for this purpose, because they are not a cost. They

13don't involve the consumption of any resources. They

14are simply a transfer payment actually from producer to

15consumer, and I don't see why we should take that into

16account in the calculation.

17Okay, so what does all this come down to? It

18comes down to, I think, predatory pricing law ought to

19be looked at in a common sense way. Predatory pricing

20law ought to be looked at straightforwardly as pricing

21that is not efficient, that is to say, pricing whose

22avoidable costs exceeds the revenues generated by the

23sales in question, and thus, pricing that reduces

24welfare during the rivalry period.

25If it's efficient pricing, if it increases

Page 54

1rivalry during the welfare period, the competitive

2period, because consumers value the marginal units, or

3the compliments that they generate, more than the

4avoidable costs of those units, it seems to me we ought

5to call this competition on the merits, and it ought to

6be lawful.

7Now, looked at this way, it seems to me,

8predatory pricing isn't all that special. If we think

9of it in this common sense way and simply ask where the

10conduct is efficient in this sense, we have both in my

11view a sound approach to predatory pricing and the

12beginning of a more general theory of exclusionary

13conduct that can avoid the pitfalls of the stovepipe

14analysis to predatory pricing.

15MR. POTTER: Thank you, Doug.

16(Applause.)

17MR. POTTER: Before we begin our round table

18discussion, we will take a short maybe ten-minute break

19to let people use the facilities and stretch, and if

20they have to call their offices, call their offices.

21MS. SCHULTHEISS: And coffee upstairs if they

22need it, 7th floor, if you need coffee or water.

23(A brief recess was taken.)

24MR. POTTER: In deciding how to handle the round

25table discussion, I thought maybe one of the effective

Page 55

1ways of doing this would be to put up various

2propositions on the screen and then ask for agreement or

3disagreement among the panelists. If there's agreement,

4fine, we have reached a consensus point, we can go on,

5and we have solved the issue, and if there is

6disagreement, we can debate the issue. The panelists

7all have this in front of them, so they do not have to

8turn around and look at the screen every three seconds.

9Do you want to put the first one up?

10MR. POTTER: I think Professor Bolton already

11indicated there might be convergence around this point,

12but there used to be economic literature saying that

13predatory pricing was an irrational business strategy.

14The proposition for the consideration of the panelists

15is that predatory pricing can be a rational business

16strategy. Is there anyone on the panel who disagrees

17with this?

18DR. ELZINGA: There is no disagreement. I would

19like to correct one matter for the record, at least I

20think this is a correction. Patrick indicated that his

21reading of John McGee's classic article on predatory

22pricing in the Standard Oil case or the lack thereof

23indicated that predatory pricing was always irrational.

24I think that's unfair to Professor McGee. That is not

25my exegesis of the article.

Page 56

1I think the position of McGee and the Chicago

2School generally is that predatory pricing can be a

3rational business strategy, it's just it's a very

4unusual one, defined where it's successful, where it

5works.

6DR. ORDOVER: Well, I certainly agree with the

7statement, with a couple of -- I don't know how many

8caveats, but first -- five caveats -- one, two, three -

9the first caveat is we have got to define what predation

10means. Second, we have to figure out what the price is.

11Third, we have got to figure out how to engage

12rationality. Other than that, I think it's all fine.

13Other than that, how was the performance, right? So,

14this is exactly the way I see it.

15I mean, this is surely a statement that has a

16meaning as long as we can agree on the meaning of the

17terms or words that go into the statement. None of

18these things are relatively or clearly defined. We

19already have different standards for predation. In the

20airline case that Ken and I are in, pricing to whom is

21an issue that -- average price on the aircraft? Is it

22the price to the business passengers, the leisure

23passengers? A huge amount of disagreement. Is it the

24price of the incremental unit? Is the price averaged

25out over the volume that is being sold?

Page 57

1Anyway, what's rational? I guess

2profit-maximizing, over what horizon, what discount

3rates we are going to use? All of these things enter

4into what we have been struggling with, which is to say

5that we have something -- we understand a basic core set

6of issues, but I think these remaining areas of

7disagreement are really needed to breathe light -

8MR. POTTER: And with the later slides, we will

9get into those specific areas.

10DR. ORDOVER: I haven't looked at them. Ex ante

11assessment, huh?

12MR. MELAMED: Just a comment provoked by what

13Janusz said, there is always -- at least in my

14experience in cases I have dealt with, I am not involved

15in Spirit -- difficult questions about what are the

16products you're talking about, what prices are you

17talking about, is it the leisure passengers or whoever

18it may be. There is, if there is discipline in the

19overall case, however, some discipline on the parties on

20that issue, if the plaintiff wants to argue that the

21price is predatory because he found one passenger in

22seat 14B where the price was below cost, he is probably

23not going to be able to prove that he was driven out of

24the market on account of the predatory price, and so if

25the courts are rigorous in connecting the allegedly

Page 58

1predatory activity with the requirement of proving a

2causal connection with the creation or maintenance of

3market power, some of the sort intellectual concerns

4that Janusz has may become less practically important.

5DR. ORDOVER: Actually, it was 15C that's at

6issue.

7MR. POTTER: Patrick?

8DR. BOLTON: So, you know, I may well have read

9too much into McGee's article. Having said that, I

10think it does -- the legacy that's left is tremendous

11skepticism, and what I wanted to say was that there has

12been new scholarship started in the 1980s, rigorous

13economic scholarship based on rigorous game theory

14analysis showing exactly how predatory pricing strategy

15could be rational, and I think what I want to say is

16that where things have changed is that slowly, this

17literature is being brought in, is being acknowledged,

18and is being recognized, and so what I wanted to say is

19that, if anything, today, we should be less skeptical

20about the rationale for predatory pricing than we have

21been and that the Supreme Court has been in its Brooke

22decision and its Matsushita decision, which was based on

23older writing which couldn't be articulated using the

24tools of the modern game theory.

25MR. POTTER: Okay, subject to Janusz's caveats,

Page 59

1I will take that as agreement among the panel.

2The second proposition, this is a quote from the

3Supreme Court in 1986, two decades old now. "Predatory

4pricing is 'rarely tried, and even more rarely

5successful,'" was repeated in Brooke Group in '93. Does

6the panel think that this is still a correct statement?

7Doug?

8MR. MELAMED: Well, I don't know. I will leave

9it to the economists. The question is whether it means

10anything. You know, murder may be rare, too, in some

11statistical sense.

12But I wanted to say something about that,

13because I think in my own thinking, at least, until

14yesterday when I was preparing for this, there was some

15sloppiness, and maybe that's true of others, as well.

16In Matsushita, interestingly, when I looked at it, that

17was when the proposition was first set, it was used as a

18factual proposition to aid the Court's assessment of the

19evidence and to say is the predatory theory here

20sufficiently plausible that we should let it go to the

21jury?

22It morphed into something else by the time of

23Brooke Group. It morphed into the rationale for

24defining predation a particular way. If it's used that

25way, we have to be very careful about what we mean. If

Page 60

1we mean pricing below cost is rarely tried and even more

2rarely successful, it's rationally then used in

3Matsushita, but it doesn't support Brooke Group, because

4that would be like saying killing with an ice pick is

5very rare, so let's define murder as consisting solely

6of killing with an ice pick.

7The question, if you want to justify or explore

8the wisdom of defining predatory pricing as pricing

9below cost, the question is, what about the conduct that

10isn't deemed to be predatory pricing by that definition,

11some kind of profit sacrifice at above cost levels, is

12that rarely tried and rarely successful? And I'm just

13not sure that there has been rigor in thinking about

14what this statement means.

15MR. POTTER: Building off of this slide, does

16anybody have a view on whether predatory pricing is more

17or less likely in certain industries because of the

18characteristics of those industries?

19DR. ELZINGA: Yes, I certainly do. I have a

20belief that predatory pricing is more likely to occur

21where the target firm will exit quickly and be unlike -

22either the target firm or other capacity will be

23unlikely to enter again, and just picking up on

24something that Doug said, where you are trying to look

25for some more simple benchmark, he suggested just

Page 61

1focusing on where avoidable costs exceed the revenues of

2the practice, well, that's a very helpful way of

3thinking about predation.

4I think it's just as powerful, maybe even more

5illuminating, to focus on entry and exit conditions as a

6kind of filter, and I am a little surprised that Doug

7never mentioned focusing on exit and entry. That is

8kind of the mirror image of what he is getting at, but I

9think it is clearer and analytically more robust.

10MR. POTTER: Janusz?

11DR. ORDOVER: Well, I think certainly by the

12basic principle of self-selection, you at least observe

13an attempt to induce an exit in the industry in which

14exit is likely to be relatively quick or not too

15costly -- it will be not too costly to engage in such a

16strategy and in which, as Bobby and I said, re-entry is

17very difficult entry or re-entry is very difficult. If

18re-entry is trivial, as it generally could be in the

19airline industry, setting aside the question of

20signaling predation, setting aside gate constraints and

21those kind of things -- which were not present in

22Detroit, just by the way.

23I think that obviously nobody in his right mind

24is going to try to exit somebody who has invested

25hundreds of millions of dollars of sunk capital that is

Page 62

1simply impossible to take out, but you can try very

2aggressively and actively to prevent that person from

3putting in another hundred million dollars of to-be-sunk

4capital. So, you can try to accomplish something

5different, but actually self-selection and rational

6business behavior that we have all accepted as a premise

7of what firms do, such as that you are not going to try

8it when it is not likely to be successful, which is why

9when we get to the recoupment phase of this whole thing,

10we will probably have different views from what the

11slides will ask us to say, but it is all part and parcel

12of the same aspect of the analysis, which is to say, you

13have to look at the entry and re-entry barriers and the

14exit barriers or problems with trying to dislodge the

15rival or problems with the ability to increase the entry

16or impediment facing the incumbent. If you cannot

17accomplish entry-enhancing creation of a barrier, then

18you are not likely to go after that, because somebody is

19going to come back sooner or later. How soon is

20unpredictable.

21DR. BOLTON: I have very little to add, just two

22remarks. There used to be a time when economists

23characterized the airline industry as a contestable

24market. I just want to remark that we have come a long

25way from that conclusion. Now we are I think defining

Page 63

1the airline industry as particularly prone to predatory

2pricing.

3And on the rarely tried and even more rarely

4successful, I want to be even more outrageous by saying

5that, you know, nuclear bombs have been rarely tried,

6but they have been very successful. We have to look at

7the deterrent effect of episodic, very rare predatory

8pricing. So, you know, you look back at predatory

9pricing in the telecom industry at the beginning of the

10century or in the tobacco industry, it was followed by

11prolonged periods of lack of entry and oligopolistic

12pricing with very high returns to the firms, which is

13evidence that consumers were not getting the low prices

14that they deserved.

15MR. POTTER: Proposition three, because lower

16prices immediately benefit consumers, we should be

17extremely careful not to adopt legal rules that can

18result in false positives; that is, condemn legitimate

19price-cutting. This seems to be a fundamental basis of

20Brooke Group, at least. Anybody have any agreement or

21disagreement with this? First say agreement.

22DR. ELZINGA: Agreement certainly for me.

23DR. BOLTON: I beg to disagree on the following

24grounds, not in principle, but on the basis of the

25evidence. How concerned should we be about false

Page 64

1positives today after a quarter century of systematic

2rejection of predatory pricing allegations? How worried

3should we be today that firms will be very cautious in

4their pricing and will refrain from aggressive pricing

5after this record?

6I think in principle, we should be worried about

7this, but I am not sure that with the past history of

8predatory pricing enforcement that this is still a major

9concern.

10MR. POTTER: Ken, I think you wanted to comment.

11DR. ELZINGA: Yes, let me comment at two

12different levels.

13First of all, there is no doubt, since

14Matsushita, that the economists have taught us things

15that we did not know at the time about models in which

16predatory pricing can be successful for the predator

17under conditions of certain financial asymmetries or

18information asymmetries or information effects, but if

19you look at some of the cases, the most recent, I think,

20or if I'm mistaken, the most recent predatory pricing

21case brought by the FTC, a long time ago, was the coffee

22case, General Foods Coffee case, and the staff was

23unsuccessful on that.

24When we look at the record, did Maxwell House,

25which had all the things that would fit nicely into this

Page 65

1model of reputation effects, signaling and so on, where

2you might think, boy, this looks like predatory pricing,

3the way the game theorists would structure the world,

4and people like Milgrom and Roberts have referred to

5that case as illustrative of applying their models to

6predatory pricing.

7Well, that was a case in which Maxwell House was

8trying to keep Folgers from moving east. They were

9singularly unsuccessful. Folgers rolled out nationally,

10and if you walk around a bit, you just don't see Maxwell

11House of having visual evidence of being a monopolist in

12the coffee industry today. You are much more apt to see

13Starbucks than Maxwell House.

14Matsushita, you think about the signaling

15effects or the reputation effects that the Japanese had

16and the popular culture at that time about being

17price-aggressive. You look at the television industry

18today -- now remember, this is a case the Japanese

19won -- they have less than 40 percent of the television

20business, total, all the companies combined. The

21largest television producer in the world is in China.

22Brooke Group, the idea there was the majors, led

23by Brown & Williamson, would dial down the discount

24segment. That was a term used over and over again in

25Brooke Group. The discount segment would be dialed

Page 66

1down. Everybody would be left buying a full revenue

2cigarette if they were a smoker.

3The discount segment continues to grow. It's

4about 40 percent of the industry now. So, if all of

5these cases had been decided differently using game

6theoretic approach or a concern that Patrick expresses,

7I think consumers would be worse off. I really do.

8MR. POTTER: Janusz?

9DR. ORDOVER: One comment. I think that there

10is an issue that we may want to talk about a little bit

11more, and that is to say, the rigor and the reviews of

12the galaxy of predation models that are based on really

13state-of-the-art game theory, and the question, what

14follows from those in terms of public policy? To me,

15that is the biggest problem that I have been totally

16incapable of resolving in my own head, but in the end,

17coming down on the proposition that while we cannot be

18as perhaps lackadaisical about anticompetitive

19exclusionary behavior as the Court in this famous quote

20was, we still need to take some kind of tools that the

21courts can use to say, yes, yes, I agree, things can

22happen, and Milgrom and Roberts and Kreps-Wilson, they

23all have shown all those things, and many others follow

24and, you know, your lovely paper with Scharfstein on -

25what, signal jamming or -- it was, signal jamming paper,

Page 67

1which was the coffee case, and all those things are all

2true.

3And then we come back to the question, what to

4do with that, how to translate it into something that a

5businessperson, who has to be counseled, will be able to

6understand in day-to-day operations, and how will the

7Court be able to take these principles of game theory,

8subgame perfect, Nash equilibria and all these things,

9and translate it into some simple rules that, you know,

10thou shall not do what? Thou shall not signal that you

11are going to be a tough guy? You can't say that. You

12have to be able to translate it into something. "Look,

13you can write any memos you want, you can do anything

14you want, but you cannot do X."

15I think that it is absolutely essential that we

16take these models and we translate them into principles

17that are implementable by the business people, by the

18lawyers and by the courts. Otherwise, we are nowhere,

19and I think what we have been struggling with is trying

20to come to articulation of some principles that are

21actually understandable, and I think Doug went a long

22way in proposing that we actually take the learning of

23these models as implying we should not dismiss these

24cases, but we should take the learning of these models

25and figure out what they mean in terms of implementable

Page 68

1rules by all the stakeholders, and that includes, of

2course, consumers as well.

3MR. POTTER: Doug, in your dealings with your

4clients, without a rule that is under-inclusive by

5protecting against false positives, is it your belief

6that monopolists wouldn't price close to the line?

7MR. MELAMED: Ah, I'm not sure I understood the

8question. I think you are asking, should we worry about

9over-deterrence?

10MR. POTTER: Well, if we don't protect against

11false positives, will the chilling effect of getting too

12close to the line lead people with monopoly power not to

13lower their prices to consumers because they're worried

14about false liability?

15MR. MELAMED: Sometimes. I do not know whether

16the overall economy, with the relative magnitude, what

17its effects are. I particularly agree with how Janusz

18started. The signals you send to the business community

19are much more important frankly than whether the cases

20are right or wrong. If every case at the margin were

21wrongly decided but we were generally setting a useful

22set of standards, the law would be pretty good. So, the

23question is the false positives versus the false

24negatives.

25Generally speaking, with the state of the law

Page 69

1today, you have a slide later on, is it hard to counsel

2your client? No, I say not to worry about it, because

3the -- but actually -- actually -

4DR. ORDOVER: Can we go home now?

5MR. MELAMED: But actually, I say more than

6that. First I say you may lose the characterization

7issue, you may not be able to prove predatory pricing,

8but then I say, "Wait a minute, there are certain

9settings in which you could get hurt. Is your target

10likely not only to withdraw from this market but, for

11example, to go out of business and become bankrupt and

12his only asset may be a lawsuit? How litigious is he?

13Is this a part of some broader commercial strategy?"

14So, there are situations I think even with the

15law today totally in favor of the weight of false

16positives where it probably does deter some

17procompetitive pricing. Whether on balance at this rule

18or at some other rule that harm is greater than the harm

19of false negatives I'll leave to the economists.

20MR. POTTER: All right.

21Next one, establishing a reasonable prospect of

22recoupment should be essential in any analysis of

23predatory pricing. Is there anyone who disagrees with

24this statement?

25MR. MELAMED: Only to the extent I already said

Page 70

1so.

2MR. POTTER: Janusz?

3DR. ORDOVER: Oh, I think the point I want to

4make is that from my perspective, this recoupment

5component is really part and parcel of a prior filter.

6Now, you can try to do it at the later stage. My

7preference is to ask the question whether the particular

8markets, market or markets, in which this

9anticompetitive conduct is alleged to be exclusionary,

10anti-consumer, whatever characterizations you want to

11attach, is acceptable to incremental exercise of market

12power, and if the answer is no because, you know, you

13get rid of this particular rival, but, you know, quick

14as a bunny, somebody else is going to show up who may be

15even more competitively advantaged rival, then there is

16no need to somehow construct this potentially

17complicated analytics.

18As is clear from Ken Elzinga's net present value

19calculation, it is a very, very difficult step, possibly

20as difficult as the step of measuring revenues to costs,

21which costs which revenues and so on and so forth. So,

22I would say that as a filter, you certainly would want

23to implement a step during which the parties will slug

24it out, one saying, "Look, I get rid of you, there is

25ten more coming. I get rid of you, that will carry no

Page 71

1visible signal for the rest of the players that may be

2sitting out on the outskirts and waiting what to do."

3Or it could be that the firm which you are

4trying to induce to exit or to restrain its expansion is

5what I called in the first slide a scarce competitor,

6and, in fact, there is something very special, very

7particular about that rival which cannot be replicated,

8and in that case, yes, you get to the point in which the

9assessment of this later recoupment or the implications

10of this strategy is critical, and if you cannot show -

11you, the plaintiff -- that if you exit the marketplace

12or if you get cut back in the marketplace, economic

13welfare is going to be hurt in some way, then I think

14you have gone very far in challenging the conduct at

15issue.

16MR. POTTER: Next slide -- oh, I'm sorry.

17DR. ELZINGA: I was just going to say, I think I

18am saying just the same thing that Janusz said but

19perhaps in just a couple words. I do not think you need

20to do a recoupment analysis for many predation

21allegations, because entry conditions or prices and

22costs will tell you you needn't take that extra step.

23DR. BOLTON: Can I just add -

24MR. POTTER: Sure, go ahead.

25DR. BOLTON: -- one comment? So, I agree with

Page 72

1Janusz that in principle, recoupment is important, it is

2the right question to ask, but in terms of how do you

3administer a recoupment test, I think the weight has to

4be on what you call the reasonable prospect, and I think

5a narrow reading of a recoupment test, as you criticized

6earlier, I would criticize as well.

7MR. POTTER: Okay, fair enough.

8Next slide. Prices above some measure of cost,

9and you can all pick your own measure of cost that you

10think is the best cost, whatever it is, should not be

11considered predatory. Is there anyone who disagrees

12with this?

13Patrick, do you want to say anything?

14DR. BOLTON: So, from the -- well, we know that

15a policy of -- after Brooke Group is that a price -- at

16least a price above average total cost should not be

17considered as predatory. I am happy to live with that,

18although I am not sure that it is always a wise policy.

19MR. POTTER: When it is -

20DR. BOLTON: I would disagree, though, with the

21statement that prices above average variable cost should

22not be considered as predatory.

23MR. POTTER: You just mentioned that you might

24disagree in certain instances that even prices above

25average total cost should not be predatory.

Page 73

1DR. BOLTON: Could be predatory. In principle,

2in theory, there are situations where prices above -

3even price above total cost can be predatory.

4MR. POTTER: Can you give an example of those?

5DR. BOLTON: Well, an example of a large

6incumbent with increasing returns, scaled technology,

7facing a small entrant that has not been able to reach

8minimum cost capacity, you could exclude that entrant by

9pricing lower than monopoly price but still above your

10average total cost and exclude the entrant.

11DR. ORDOVER: Maybe I could just ask you a

12question. Would you comment on the cost principles -- I

13have been puzzled by them myself -- that follow from

14these various game theory like models of, say,

15Kreps-Wilson? They do not seem to give clear cost

16benchmarks. Is that true? Is that your reading as

17well?

18DR. BOLTON: Yes, that is correct. They do not

19give a clear reading on cost benchmarks, and I think

20there is a whole group of economists who have been

21working on predatory pricing who think that costs are a

22very poor way of discriminating between anticompetitive

23effects and procompetitive effects, that there are as

24likely to be false positives as there are to be false

25negatives. There are many situations where pricing is

Page 74

1below even average variable cost, and it is efficient.

2It is not predatory. So, a lot of economists feel it is

3just a poor test.

4MR. POTTER: Doug?

5MR. MELAMED: Let me ask Patrick this question.

6I understand the theory, even if I cannot understand the

7game theory, of why an above cost, even above total

8cost, but below profit -- monopoly profit-maximizing

9test could be predatory in the sense that it could

10exclude a rival and in the long run we are all going to

11be worse off for it.

12What I don't understand and I am interested in

13your reaction to is how one turns that into a legal rule

14that companies can comply with. I mean, how do you -

15you know, if -- sure, if you imagine -- if you posit a

16stable market on day one and then the entrant comes and

17maybe you have a good historic benchmark and you can

18say, "Gee, he's changed his pricing," but even then you

19have to ask the question, "Well, what would the monopoly

20profit maximizing price be with the new entrant?" Is

21each company supposed to hire a game theorist and work

22out the game and figure out -

23DR. ORDOVER: Yes.

24MR. MELAMED: -- what the price is? In other

25words, how do we implement that test?

Page 75

1MR. POTTER: Patrick is looking for future

2employment.

3DR. ORDOVER: We all are.

4DR. BOLTON: Administerability is a serious

5concern, I take that. I'm happy with a rule that

6says -- I would not object to a rule that says price

7above average total cost is per se legal as a way of

8implementing an easily administrable rule.

9As for determining whether it is procompetitive

10or anticompetitive conduct, I think there -- while

11business decisions are taken on average in a rational

12way, and you have to get justifications for the kind of

13policy you are implementing, these justifications of ten

14find their form in written documents in the company,

15whether it is emails or other board room records, and as

16I emphasized in my presentation, this is evidence of

17intent, which is extremely valuable. Intent here, that

18kind of intent evidence, is a very good guide to the

19kind of effects a policy can have, and there, I think we

20can be on pretty firm ground, and we do not have to

21do -- we do not have to hire a game theorist to do that

22kind of analysis.

23MR. MELAMED: Sometimes companies adapt to the

24law, and if they are well counseled, they know how to

25write pieces of paper that perhaps articulate an

Page 76

1economic rationale rather than intent.

2DR. ELZINGA: And in like fashion, Doug, I

3suspect you have encountered clients who just aren't

4aware that when they write things, they have to be

5written with an eye towards antitrust enforcement, and

6so you do find documents coming -- to have militaristic

7or powerful metaphors that have nothing to do with

8consumer welfare and may, in fact, represent exaggerated

9views of the company's prowess and stature in the

10marketplace.

11MR. MELAMED: I find almost invariably they are

12written by lunatic middle managers, but -

13MR. POTTER: That wasn't your position when you

14were Deputy Assistant Attorney General.

15MR. MELAMED: Well, you know, you learn.

16DR. ORDOVER: Or those documents are usually by

17investment bankers who are pedaling a particular deal or

18something like that, alleging that as a result of action

19X or Y, the firm would be able to leverage its market

20power from one market to another. If one were to take

21these arguments -- take these documents seriously, that

22would be the end of most of the Chicago Business School,

23presumably, investment banking, but also, the ability of

24business people to compete in the marketplace, because

25this is what these guys are selling. So, you have to

Page 77

1read their signal, which is the investment advice or

2business advice which they proffer, as being an attempt

3to market the product at above competitive price to the,

4you know, willing or unwilling buyers, and I think that

5that is why I am very worried about reading all kinds -

6I mean, I have seen documents probably as good or as bad

7as Doug or anyone, and I try to discount their value

8because they are frequently misleading.

9Now, this is not to say that people who run the

10companies do not have an insight into the marketplaces

11in which they are competing, but I think there are

12limits of the kind of inferences you can really make

13from those types of documents, especially when they are

14also written by third parties with a very special agenda

15of their own in my view.

16MR. POTTER: Next slide, as long as we're on

17costs, let's throw this out, a variety of cost tests.

18The proposition for the panel is, average avoidable

19cost, which for definition, cost per incremental unit

20that does not have to be paid if the incremental units

21are not produced, is the best cost measure to use if

22forced to use the Brooke Group analysis. Is there any

23disagreement with that?

24Doug?

25MR. MELAMED: Yes, I actually disagree with it

Page 78

1as phrased. I think it would be more useful to use

2avoidable cost compared -- and then add up the revenues,

3because when you say average avoidable cost compared to

4price, you are limiting yourself to the revenues from

5the product in question, and you can't take account by

6that formula, at least, of two-sided markets and

7everything else.

8DR. ORDOVER: And networks. I think the issue,

9just to pitch it to the folks if you want to raise it or

10discuss it further, I think in the American Airlines

11case, there was a lot of debate as to what the right

12benchmark of cost was in my view, and at least the

13Government had I think proposed four, if I am not

14mistaken -

15MR. POTTER: Correct.

16DR. ORDOVER: -- that may be three too many,

17but, you know, we have offered at least one or two

18ourselves.

19The same thing in Spirit, I think we have come

20up with two different measures of cost, but they would

21apply to different types of outputs, all passengers

22rather than leisure versus business. So, there is a lot

23of wiggle room as to what it is that this cost measure

24is going to be applied to, and as Doug pointed out, it

25is also key to figure out what is the measure of

Page 79

1revenues against which these costs are to be assessed.

2So, I think that this is not a bad -- again,

3this is not a bad standard, but I think it is important

4to understand what is avoidability here that is at

5stake.

6For example, in the American Airlines case, we

7thought, at least I thought quite strongly, that the

8right set of costs would be those that if -- the airline

9would avoid if it were to exit or substantially cut back

10on a particular route, and that actually includes a lot

11of costs that would be avoided, because it would include

12avoiding the aircraft costs which were at issue, those

13would be significant, much more than many other costs,

14and it could be cutting back at the hub, perhaps,

15cutting back at the stations from which the airline

16would exit.

17So, these avoidable costs which we looked at at

18the route level are typically the kind of costs business

19people look at when they make business decisions in the

20airline business, and I thought there was a good measure

21of avoidable cost. It happens so that the increment of

22output over which we are looking at was that of the

23route as opposed to -- it could be a seat or it could be

24an aircraft or it could be a flight or something or an

25aircraft day, because these aircraft, they fly in

Page 80

1strange ways around the globe, but there is all these

2things that can be taken into account that could confuse

3or could illuminate the matter.

4MR. POTTER: I wanted to follow up on this slide

5for a couple other questions. One, Doug, I think you

6talked about this in your presentation, but I was

7wondering if anybody had a view of whether opportunity

8costs should be considered in viewing the cost test.

9MR. MELAMED: Yes.

10MR. POTTER: Doug, I thought you were a yes,

11Doug. Does anybody else on the panel have any views on

12that?

13DR. ORDOVER: Well, as I said, I think in

14American Airlines, we had an internal discussion of

15what's the meaning -- what to do about the aircraft. I

16mean, there is no denial that American Airlines brought

17in additional flights. You could say, well, is there an

18opportunity cost of that aircraft, and if there is, how

19are you going to measure it? And one measure, which I

20thought was the most easy to implement, would be to look

21at the lease rates as opposed to trying to understand

22what is another route that this aircraft could have been

23flying or was this aircraft sitting somewhere in the

24desert in Arizona and doing nothing? Maybe American has

25a lot of aircraft like that, they could fly them at very

Page 81

1low incremental capital cost, but I thought that the

2most reasonable assumption would be to assume that the

3airline uses its aircraft properly and it could actually

4deploy an aircraft by leasing a new one, which is why an

518-month time horizon I thought would reflect the

6leasing strategies and the fact that heavy and costly

7equipment was deployed, and there is no way of avoiding

8counting the assets that are not being used, which

9aircraft were one of them, and there are others that I

10thought were appropriately included as well. That was

11not the view necessarily of all of my colleagues on that

12case.

13MR. POTTER: Ken?

14DR. ELZINGA: I do not think you could trust any

15economist who would say opportunity costs should not be

16considered. I mean, opportunity cost is the main

17analytical construct that we bring to the social

18sciences, that the cost of something is the highest

19valued opportunity forgone, but the problem is that in

20antitrust, opportunity cost is a Promethean expression,

21and it is very difficult to unpack it, and one of the

22sobering things for me, who has worked in this area for

23a while and tried to think about it, is how fragile some

24of the cases are, some of them that I have been involved

25in, some of them that I have studied, to the taxonomy of

Page 82

1cost.

2The case that Janusz and I were involved in,

3what do you mean by the price of the product? Is it to

4all passengers, or is it to a group of passengers who

5are called leisure and price-sensitive and business and

6insensitive to price? The case can pivot upon that

7taxonomy.

8In Brooke Group, after hundreds of hours of

9thought, one of the things that distressed me is that

10price above average variable cost pivoted upon -- it

11could pivot upon how you counted layers of tobacco.

12Tobacco is stored for years to age, and if you used a

13LIFO method, it looks like it violated Areeda-Turner.

14If you used LOFI, it looked like it was okay, and I

15don't like living in a world where it pivots upon what

16accounting standard you use.

17One of the things economists supposedly also

18bring to the table is to get people out of using

19accounting data and to think in terms of opportunity

20costs, but even trying to apply that standard is

21problematic, and when you live in a world where there's

22a predatory pricing allegation, you probably are in a

23world where prices are close to costs, by whatever

24measure, and so then you start to figure out, well, how

25close and above or below, you inherently, if you are

Page 83

1going to use a cost-based standard, have to deal with

2accounting data, but you try to always put it through

3the filter of opportunity costs.

4DR. ORDOVER: Just as an anecdote, in the

5American Airlines case, I lived through like three days

6of deposition and a number of questions related to the

7question of how we treated these carts that people put

8their luggage on when they pick it up at the station

9that was going to be exited potentially, so I finally in

10desperation said, "Look, if the whole goddamn case turns

11on how we treat these carts, then the Government

12shouldn't be bringing a case like that." There has to

13be something more to it than that, right?

14And I think somewhere -- I mean, tobacco is

15probably more important to cigarettes than the carts are

16to the airline, but again, this is really demonstrating

17very well the kind of deep-level accounting issues and

18cost treatment of issues that can go whichever way

19somebody wants them to go, and therefore, to say that

20something is average avoidable cost is, again, the same

21thing. It is average, it is avoidable, it is a cost,

22but other than that...

23MR. POTTER: If you are going to require

24opportunity costs to be considered, how does that differ

25from requiring the defendant to maximize his profit?

Page 84

1Because one view of opportunity costs is he had an

2opportunity to get more profit. How do you distinguish

3the two, if at all?

4Patrick?

5DR. BOLTON: Yes, so, this is where Doug drew a

6very clear line that you should not count lost revenues

7and inframarginal sales as an opportunity cost. I would

8say as a matter of theory, you should count that as an

9opportunity cost. The real question is just one of

10administerability.

11The other point I would make in this respect is

12that -

13DR. ELZINGA: I am sorry, could you just explain

14for the benefit of at least me why, why you would count

15that, the inframarginal?

16DR. BOLTON: Well, because the question we are

17trying to answer is, what is driving the price

18reduction? Is this a move by the incumbent that will

19raise profits irrespective of its anticompetitive

20effects or not? To be able to answer that question, we

21need to understand the nature of the profit sacrifice,

22the size of the profit sacrifice and what justifies it.

23Is there an efficiency rationale or is there an

24anticompetitive rationale? So, we cannot avoid it, and

25that is what I was going to say with respect to

Page 85

1recoupment.

2Recoupment is the right question to ask. When

3you try and make a recoupment analysis, you are

4comparing profit sacrifice and return on investment.

5So, it is inescapable, you have to look at lost revenue

6and inframarginal sales when you do your recoupment

7analysis.

8Now, if you do it for your recoupment analysis,

9I don't see why we should not take that into account for

10the cost test, but I will leave that for -- I think I

11see that as a practical problem and not a conceptual

12problem.

13MR. POTTER: Doug, do you want any extra time or

14not?

15MR. MELAMED: Well, the woman who is

16transcribing this said I spoke so fast that maybe my

17words were not caught before, but I will assume she got

18them.

19MR. POTTER: Okay. What about, how do we

20distinguish in situations the appropriate costs when -

21essentially price discrimination or nonlinear pricing?

22Does anybody have any views on that?

23Doug? No, that is not on the slide.

24MR. MELAMED: What are the avoidable costs?

25MR. POTTER: Fine, let's go to the next slide.

Page 86

1Out-of-market reputation effects are so hard to

2assess they should not be considered in an analysis, and

3let me just give you another minute on this. I am

4thinking of a scenario whereby -- let's take the airline

5industry. An airline allegedly predates on, let's say,

6Dallas-Wichita. The airline has a bunch of other

7markets where it's in, and perhaps it's got monopoly

8power in a number of those markets. Maybe it doesn't

9know, maybe we don't know, who potential entrants will

10be in those other markets, but when the management team

11sits down to determine what they are going to do in

12Wichita-Dallas, they sit there and say, "Well, let's

13drive the guy out, because future people then won't

14challenge us in our other markets." How does that, if

15at all, get analyzed in determining recoupment?

16Patrick?

17DR. BOLTON: Yes. So, we elaborate on this

18point in our article. So, reputation effects do come

19into a recoupment analysis to the extent that reputation

20effects may raise the barriers to re-entry into the

21market, and they do raise them in the form of making

22other competitors aware that should they enter this

23market, the first thing they will be facing is a tough

24price war.

25How do you prove reputation effects? That is

Page 87

1the harder question. But I do not think that that is

2necessarily insurmountable. Again, there can be

3analysis by the incumbent suggesting that this is a

4profitable strategy, that reputation effects work, that

5if we drive out this rival in this market, we will

6benefit because there will not be other rivals or we

7will be slowing down the growth of this rival.

8So, this analysis is recognized by the

9incumbent. If you find circumstantial evidence

10suggesting that financiers think in those terms, that

11they will raise the cost of funding of a new entrant

12because they recognize that the market that the new

13entrant is about to enter is one where there have been

14past price war episodes, then I think this is all

15evidence that this is a problem.

16MR. POTTER: Ken?

17DR. ELZINGA: This is a very strange statement,

18and it could be easily misunderstood, but we have to

19remember that the case for new entry and the enthusiasm

20that we often have in antitrust for new entry can be

21exaggerated. New entrants can inefficiently use

22society's scarce resources. There are lots of

23businesses that have no business entering an industry

24because they use the resources inefficiently, and one of

25the good things that keeps inefficient entrants out is

Page 88

1the reputation of incumbent firms for being tough,

2aggressive, low-cost competitors.

3My concern -- and again, so much of the

4difference, perhaps, between Patrick and myself is one

5of administerability -- is once you start bringing in

6reputation effects as a potential hammer for antitrust

7plaintiffs, what is the consequence of that for all the

8good things that reputations do of incumbent firms to

9keep people, even for their own good, out of markets in

10which they have no business competing because they will

11not be efficient utilizers of society's scarce resources

12in those settings?

13DR. ORDOVER: I think that the reputation

14effects are almost a cornerstone of the new game

15theoretic model of anticompetitive behavior of the sort

16that Patrick summarized in his talk and his paper, and

17there is no question that firms act in ways that try to

18convey signals to the outside world and to the inside

19world, and I think I would agree with Ken, not going so

20far as we should not find entrants to be all such great

21participants -- I like entrants. I think they should -

22you know, let them slug it out.

23Let's not create a presumption that some of them

24may be inefficient ones or some of them are efficient

25ones. Who the hell knows? It is the crucible of the

Page 89

1marketplace that ultimately will determine that.

2But my issue is, again, I guess where we agree

3is administerability, and then to say, yes, indeed, it

4is plausible to postulate the reputation effects. We

5have the economic models. What we don't know in real

6life is how many of these new entrants do you have to

7kill in the airline business before somebody finally

8realizes, hey, I'm not coming in, and empirical work

9shows that no matter how many of them you squish, they

10always come back, and so you say, am I still in the

11reputation-building way or am I in the recoupment phase

12or how am I going to account for that other than to say,

13look, you go ahead and do what you want to do, compete

14as hard as you want, but you should not break the

15following simple rules, whatever they might be, because

16I cannot account for all the other additional

17considerations.

18However, I think it is appropriate to say that

19these reputational effects that we are encountering in

20economic theoretical literature, but also in some

21empirical stuff. In fact, there is some beautiful work

22by Canadians on the supermarkets in Canada, indeed,

23indicating that some reputational effects that have been

24established. I just don't see how I can translate that

25into an administrable test for the courts and for

Page 90

1counsel, because what can Doug say to somebody who says,

2"Look, I think I want to kill three guys because I think

3that will be enough," and he goes, "No, only kill two."

4I mean, what? What do you say? I just don't know how

5to do it. I wouldn't know how to do it.

6MR. MELAMED: Look, if the client comes to me

7and says, in effect, I want to cut my prices to below my

8avoidable costs, I might say, "Why are you doing that?

9You are going to run the risk of losing an antitrust

10case." And if he says to me, "I'll beat the rap because

11they will never hang me with the recoupment thing,"

12because if my recoupment is going to be my reputation, I

13might say, "That ought to be illegal."

14That is to say, as Patrick says, if you can

15prove a plausible recoupment story, a reputational

16story, that, in fact, you are gaining market power

17because you are gaining reputation and it is not just

18the lawyers -- the plaintiff's lawyer's fantasy, then I

19don't know why that's not enough to satisfy the market

20power screen. In the Microsoft case, for example, which

21I believe was rightly decided, the proof of competitive

22effects was, you know, rather conjectural, but you had

23conduct that unequivocally didn't do anybody any good

24and you had a plausible theory of a market power entry

25barrier story and the fact that Microsoft believed that.

Page 91

1Why wouldn't that be enough just because we have price

2here as opposed to the set of conduct that was at issue

3in Microsoft?

4DR. ORDOVER: I think that there are substantial

5sunk costs of coming in, combined with the signaling, I

6think you have a plausible story to tell, say, look, you

7know, you are trying to convince these people that if

8they come in, there is going to be an aggressive price,

9and with the substantial sunk costs at issue, that might

10be something that will take you over the edge, and they

11say, I'll stay out of the relevant market, but it is the

12combination of the informational aspects of behavior

13coupled with the structural features, which is

14substantial up-front costs, which you require and that

15the market power screen really -- to say, ah, this

16market is susceptible to anticompetitive behavior, it's

17susceptible to recoupment or to price elevation if you

18protect it.

19So, this acts as a part of the analytical story

20that is being told, but again -- and I am perfectly

21happy to accept it. What I am trying to say, it has to

22go hand in hand with another aspect of the proof, which

23is to say that the informational aspects are conjoined

24with the real exposure that the entrants will face if

25things go wrong, so that when he comes in, it loses a

Page 92

1lot of money if it stays, because it cannot exit, and

2therefore, it is not willing to try. If exit and entry

3are easy, then I don't believe there is much power to

4that informational signal.

5MR. MELAMED: I agree completely with that.

6MR. POTTER: Next slide. I know at least one of

7our panelists said in his presentation that he would

8disagree with this, so, Patrick, I believe that is you,

9but we will see what the others think. Meeting

10competition should not be a defense to predatory

11pricing. Is there anybody who agrees with that?

12MR. MELAMED: Agrees that it should not be?

13MR. POTTER: Should not be. Doug, one

14agreement. Patrick, you disagree. Okay, Doug, why -

15DR. ORDOVER: Wait, should not be -

16MR. POTTER: Should not be a defense. So, you

17have essentially a high-cost producer. A lower-cost

18producer comes in, more efficient, at lower price. The

19high-cost producer cuts his price, lowers cost to meet

20competition. Should that be protected or not in a

21predatory pricing case?

22DR. BOLTON: So, on the meeting competition

23defense, if meeting the competition is a best response,

24then this should be a defense. So, in principle, this

25is an admissible defense. Administerability, again,

Page 93

1concerns are important here. For example, what do we

2mean by meeting the competition? Is matching the price

3of the entrant meeting the competition? Is that how we

4define it? I would argue that's dangerous, because the

5products may not be the same. If the incumbent's

6product is higher quality than the entrant's, then

7matching the price of the entrant is not meeting

8competition. It's -

9MR. POTTER: So, a jury is going to decide what

10the quality-adjusted price is?

11DR. BOLTON: (Nodding.)

12MR. POTTER: Doug?

13MR. MELAMED: I think Patrick and I might not

14actually disagree but just use different words. He said

15if this is the best response. If it's the best

16response, then it would seem to me that the revenues

17generated by the response are in excess of the avoidable

18costs, in which case it passes the price-cost test, but

19if that's not the case, if it fails that test, it's an

20inefficient response. The fact that he's meeting

21competition I don't think should make it a safe harbor.

22MR. POTTER: On a more general basis, it's not

23one of the slides, but what role should efficiencies

24play as a defense to predatory pricing? I know,

25Patrick, you think they should play a central role. Any

Page 94

1of the other panelists have a view on that?

2(No response.)

3MR. POTTER: Seeing none, we will go on to the

4next slide.

5It is extremely difficult to craft an effective

6injunctive remedy in predatory pricing cases, and I'm

7thinking of the -- I think there was at least one TRO in

8a case where the judge said for purposes of the TRO, the

9company couldn't price above the price that it had set

10on August 1st, you know -

11DR. ORDOVER: Of any year?

12MR. POTTER: Well, it was a particular year, you

13know, subject to changes in raw material costs. You

14know, is that the remedy that -

15DR. ORDOVER: That was the Baumol Test, right,

16you can cut the price, but you can't raise it for five

17years?

18MR. POTTER: What is the injunctive remedy? I

19understand what the damage remedy is if it's a private

20case. What is the effective injunctive remedy? Is

21there any?

22Doug, have you given that some thought?

23MR. MELAMED: Yes, that question I have. I

24think it is very difficult, very dicey. There may be a

25circumstance in which it makes sense for a court to

Page 95

1specify a price in that sense, I can't offhand think of

2one, but I don't know why it's a terrible thing to

3simply say, "I declare the conduct to be illegal, and I

4order you to stop pricing below avoidable costs."

5DR. ORDOVER: Having first defined it.

6MR. MELAMED: Right, of course, assuming the law

7has been decided so that we know what that means,

8because I think the action in the government case, for

9example, is to help the law evolve into sensible

10principles, and then the deterrent effect might be

11served by the damages rather than having government

12regulate through injunctions.

13MR. POTTER: Ken?

14DR. ELZINGA: Well, probably like everybody on

15around the table and everybody on the other side of the

16table, I'm suspicious of having antitrust become a price

17regulatory regime. It may be that in a genuine

18predatory pricing case, as the Court has the authority,

19that you could get at some other part of the structure

20of the market that allows the predatory pricing to be a

21viable marketing strategy. Patrick gives the example in

22his article, which I found persuasive, of the Intel

23Communications, whether the Court would have the

24authority to get at the regulatory issue that allows the

25financial asymmetries and the resource asymmetries to

Page 96

1make the predation successful.

2I don't know, and I'm picking this not to pick a

3dispute with Janusz, but let's say for point of argument

4that -- I genuinely mean that -- but let's say that gate

5constraints are the one variable that might make

6predation successful in the airline industry, and if you

7can get away from gate restraints, then new entrants

8could always come in and unravel any successful

9predation scheme. I would much prefer in the setting

10like that for a court to say, "Well, instead of trying

11to monitor and manipulate prices of airline tickets to

12make sure they're above some measure of cost, that we do

13away with that particular structural constraint that

14keeps the new entrant from being viable at such and such

15an airport because they can't get gates."

16If that were the case, perhaps that would be the

17way to get at it. That would be more appealing to me

18than having the Court monitor prices over time.

19DR. ORDOVER: I cannot disagree on the gate

20issues. It has been recognized in the airline business

21that gates and slots are one of these assets that the

22contestability literature perhaps forgot about when it

23was first deployed in the airline business, capital and

24wings, but I think that is a very sound prescription. I

25just have, again, one little caveat.

Page 97

1One of the reasons these gate problems often

2arise is because the airlines actually could finance a

3large part of the construction of the airport, and that

4becomes an issue, who is going to -- unless the airport

5is a public resource that is not paid for by the

6airlines or by one airline investing in a -- that's in a

7part of the airport, if it's actually paying for these

8gates, then I think it becomes potentially expropriation

9of what could be a costly investment, and I think we

10will have to worry about the remedy from the standpoint

11of investment incentives, in other words, opening up the

12scarce asset.

13I don't have to worry about that so much perhaps

14in the airline industry, but other industries.

15Obviously Microsoft raised the question and said, "Wait

16a minute, we are investing -- the remedy is to open up

17the API. Hey, we are spending a huge amount of money

18innovating in this space, and now you are telling me to

19open up." So, this is again, you know, perhaps even

20more problematic than regulating price, to regulate

21access. It is an equally complex or even perhaps more

22complex issue.

23MR. POTTER: Let's go to the last slide, and we

24are running a little bit out of time, but I definitely

25wanted to get this question in and get a response from

Page 98

1each of the panelists.

2The final slide is, if there was one thing you

3could change with the current legal approach to

4predatory pricing, what would it be? And since we

5started with Ken, I think this time we will end with

6Ken, in reverse order. Doug, why don't you take this

7one first.

8MR. MELAMED: I think I would just try to

9demystify it a little bit and think of it simply as part

10of a complicated set of strategies that companies use

11that under some set of circumstances can be

12anticompetitive.

13MR. POTTER: Patrick?

14DR. BOLTON: Yes, I would agree with that, and I

15would also vote for de-emphasis of the cost test and

16putting intent back as a possible way of proving

17predatory pricing, and here, I think it would be helpful

18to maybe articulate the guidelines on how one would -

19what's a legitimate way of proving intent and perhaps,

20you know, move in that direction.

21MR. POTTER: Janusz?

22DR. ORDOVER: I don't think I have a favorite.

23I will just say that I will agree with Doug, that we

24need to get clarity on what are the public policy or

25economic principles that either underlie the tests that

Page 99

1are being proposed, where does the -- you know, the cost

2tests, where do they come from? There has to be -

3going back, I think at this point we have enough

4learning to try to go back to first principles and try

5to understand what it is that we are trying to

6accomplish, taking full account of the administerability

7of whatever provisions are going to ultimately be

8developed, but I think it would be foolish for us -- for

9me, anyway -- to vote for the least favorite aspect of

10what is out there at this point.

11MR. POTTER: Ken?

12DR. ELZINGA: Well, I certainly can't argue

13against Janusz's call for clarity, but I think we are in

14a pretty congenial equilibrium right now.

15MR. POTTER: Good. I just want to point out

16that over on the side, we have some of this afternoon's

17panelists that were kind enough to be here this morning.

18They are John Kirkwood, Tim Brennan and Rick

19Warren-Boulton, and just right before we leave, I will

20give each of you 30 seconds to say anything you wanted

21to say about this morning's panel, or if you just want

22to wait until this afternoon, feel free. Anyone want to

23take a go?

24DR. WARREN-BOULTON: I think I will wait for

25this afternoon, but -

Page 100

1MR. POTTER: You will all wait for this

2afternoon?

3Well, if you could join me in thanking the

4panelists.

5(Applause.)

6MR. POTTER: That will end the morning session,

7and the afternoon session begins at 1:30. Thank you

8very much.

9(Whereupon, at 11:56 a.m., a lunch recess was

10taken.)

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

Page 101

1AFTERNOON SESSION

2(1:28 p.m.)

3MS. SCHULTHEISS: Everybody ready? Okay, good

4afternoon, and for those of you who were not here this

5morning, my name is Pat Schultheiss. I'm an attorney

6with the Federal Trade Commission's Bureau of

7Competition in the Office of Policy and Coordination,

8and I am the lead moderator for this afternoon. My

9co-moderator is Bob Potter, who is the Chief of the

10Legal Policy Section at the Federal -- at the Department

11of Justice's Antitrust Division. I was going to put you

12at the FTC again, sorry, Bob. I can't just get those

13words out.

14Before we start, a few housekeeping matters.

15First, for everybody's benefit, please turn off your

16cell phones, Blackberries, any other device that might

17make noise during the session. We appreciate that.

18Second, the restrooms are out to -- men's room

19right to the left, and the women's room, past the

20elevators and to your left. There are little signs out

21there to guide you as well.

22Third, in the very rare event that the alarms

23happen to go off, please just calmly proceed down the

24staircase and follow the zillion FTC attorneys and staff

25that will be going towards 7th Street to the Sculpture

Page 102

1Garden. Those are the preliminaries.

2This afternoon's topic is dealing with predatory

3pricing but looking at it from the buying or bidding

4side. We have assembled a very distinguished panel who

5will discuss I think, among other things, just how

6common buy-side predatory pricing is in the real world,

7if at all, whether it's common at all, what Section 2 of

8the Sherman Act can and should be doing about predatory

9buying or bidding, and I'm sure there will be many other

10things. I know we will have a little bit of raising

11rivals' costs from Professor Salop and others. So, with

12that, let me introduce very briefly the panel, and then

13I will introduce with a little bit of greater detail

14each panelist right before they speak.

15Our panelists this afternoon, in the order they

16will be speaking, are Professor Jack Kirkwood from the

17University of Seattle; Professor Tim Brennan from the

18University of Maryland, Baltimore County; Professor

19Steve Salop from Georgetown University; Rick

20Warren-Boulton, a consultant with Microeconomic

21Consulting and Research Associates; and Janet McDavid, a

22partner with Hogan & Hartson. Each of the panelists

23will give a 10 to 15-minute presentation. After that,

24we will take a brief break and then have a panel

25discussion for the remainder of our time. I would like

Page 103

1to thank all the panelists for being here and for the

2morning panelists who have decided to stay and enjoy the

3rest of the discussion. We greatly appreciate the

4willingness of all of the panelists to give their time,

5not just here today, but also the time input into

6preparing for this session.

7Our first speaker today is going to be Jack

8Kirkwood, as I said. Jack is a professor of law at the

9Seattle University School of Law. Before joining the

10Seattle University, Professor Kirkwood was an attorney

11with the Federal Trade Commission. Before leaving the

12rarified air of Washington, D.C., he was the director of

13several policy offices here in the Premerger

14Notification Program, but then decided to head out to

15the Pacific Northwest, and he joined the Seattle

16Regional Office, where he headed up numerous antitrust

17investigations and cases.

18Professor Kirkwood has edited two books and

19published numerous articles, and he recently addressed

20the topic of today's hearing in his article in the

21Antitrust Law Journal, "Buyer Power and Exclusionary

22Conduct: Should Brooke Group Set the Standards For

23Buyer-Induced Price Discrimination and Predatory

24Bidding?"

25In addition, Jack Kirkwood is a consultant for

Page 104

1the plaintiffs in the Weyerhaeuser appeal against

2Ross-Simmons in the Ninth Circuit. I think we only have

3one panelist who is not somehow or another involved in

4the Weyerhaeuser case, but with that, Jack.

5MR. KIRKWOOD: Thanks, Pat, and thank you to

6both agencies for inviting me.

7This panel's focus is especially significant.

8It's not only an intellectual or antitrust policy

9question, but it is a question of how should the Supreme

10Court come out in a case that in my sense is they are

11very likely to take, Ross-Simmons versus Weyerhaeuser,

12and the central issue in that case, of course, will be

13should Brooke Group apply. Should Brooke Group's

14price-cost and recoupment tests apply to a practice that

15has been called predatory overbuying or predatory

16bidding.

17This is, as I've conceived it and as Steve and

18others have looked at it, is in major respects the

19mirror image of predatory pricing on the sell-side.

20With predatory bidding, a dominant buyer bids up the

21price of a critical input, forcing up the market price,

22and in certain circumstances, making it impossible for

23rival buyers to continue to compete, causing either

24their exit or their inability to constrain the dominant

25buyer's future exercise of power, and hence, the

Page 105

1dominant buyer captures monopsony power that it wouldn't

2otherwise have, and in some circumstances, this may lead

3to a long-run harm to welfare, most directly to supplier

4welfare but also possibly to consumer welfare as well.

5Given these similarities, all of the Supreme

6Court's stated rationales for applying Brooke Group

7apply to predatory bidding as well, at least to some

8degree. Furthermore, the only alternative test approved

9by the Ninth Circuit was an exceptionally vague jury

10instruction allowing the jury to find liability if it

11found that Weyerhaeuser bid more than "necessary" in

12order to prevent the plaintiff from buying at a "fair

13price." That strikes me as too vague and many others as

14well.

15So, if that test is not acceptable, should we

16resort to Brooke Group or try something in the middle?

17And what I am going to suggest today is that a

18middle-of-the-road test is more appropriate, a test that

19has two parts. One, the plaintiff would have to show

20harm to welfare, I will explain what that means, but the

21defendant would get a complete defense if it can show

22that it would pass the no-economic-sense test.

23Why a middle-of-the-road test, why not Brooke

24Group? In some significant respects, predatory bidding

25is different from predatory pricing. There is, of

Page 106

1course, one cosmetic difference. In predatory bidding,

2the first thing that happens is input prices go up,

3output prices to consumers do not go down, at least

4initially. So, there is this key difference in terms of

5what we normally think of as the central focus of

6antitrust on providing low prices to consumers, but that

7is essentially a cosmetic rather than an important

8difference, because when a buyer bids up input prices

9that benefit suppliers, it can be procompetitive, and

10through an output effect I will describe, it can benefit

11consumers as well. So, we should be concerned with

12chilling procompetitive bidding for inputs just as we

13are concerned about chilling procompetitive price cuts.

14There is, though, a more significant difference.

15Compared to predatory pricing cases, predatory bidding

16cases have been brought less frequently, have been won

17less frequently, and arguably, there have been no false

18positives, no liability findings where it appeared that

19the defendant had not, indeed, harmed welfare. That is

20arguable, not hardly proven.

21In the last two decades, since the mid-eighties,

22there have only been two cases in which the plaintiffs

23have won. Both have involved bidding up timber prices

24in the Pacific Northwest, and if the Supreme Court takes

25Weyerhaeuser, we are not even sure that the plaintiff

Page 107

1will have won the second case. Why is it that these

2cases seem to be rarer and that the proportion of

3successful predation may be higher? Rick Warren-Boulton

4will address that. I will defer to him on that. I have

5suggestions, but he can cover it.

6This track record suggests to me, at least, that

7it is a little too early to apply a Brooke Group

8price-cost safe harbor test to predatory bidding. To be

9sure, the number of successful cases is too small to

10produce a reliable conclusion that predatory bidding is

11more dangerous than predatory pricing to welfare. We

12are not there yet, yet the track record does suggest

13that the danger of deterring procompetitive bidding is

14less high than it is with predatory pricing, at least if

15there were a stiff rule that applied to a plaintiff as I

16will suggest.

17This is consistent with my experience at the FTC

18as head of the Planning Office, the Evaluation Office,

19and then as a staff attorney in Seattle. I received

20over the years many complaints about price-cutting but

21never, ever, a complaint about bidding up input prices.

22There are two other reasons to think we ought at

23this point to choose a more flexible test rather than a

24Brooke Group safe harbor. One is, there has been to

25date much less scholarly or judicial analysis of the

Page 108

1practice, though thanks to the agencies and thanks to

2Steve, we are working on that. Even so, even so, the

3pile of articles on predatory bidding does not compare

4to the mountain on predatory pricing.

5There is also a growing, though probably still

6minority, view that the Brooke Group average variable

7cost test, at least as interpreted that way, may not be

8the right standard even for predatory pricing where the

9concern with chilling procompetitive price competition

10is greater; that at least it counsels against extended

11Brooke Group predatory bidding.

12What would an alternative test look like? My

13suggestion, just a proposal, is to put a stiff burden on

14the plaintiff and to give the defendant a complete

15defense. The plaintiff's burden would be to show harm

16to welfare. So, a plaintiff would have to prove the

17elements of a welfare-reducing instance of predatory

18bidding. So, they would have to show that, yes, the

19price was bid up; yes, at least some significant rivals

20were constrained in their ability to hold up, since we

21are talking about bidding, hold up the alleged

22predator's price; as a result, the predator got

23monopsony power it would not otherwise have; and most

24important of all, the plaintiff would have to show that

25the long-run impact on welfare was negative. So, to

Page 109

1pick up on one of Ken's key points this morning, the

2plaintiff would have to show that it was relatively easy

3to induce exit, but either re-entry or new entry would

4be more difficult.

5How should we measure welfare? I have

6deliberately not put an adjective in front of it. It

7seems to me there are two principal possibilities: One,

8supplier welfare; two, consumer welfare. Steve is going

9to talk about either of those measures, particularly

10consumer welfare as opposed to total welfare. I will

11skip that debate.

12Between supplier welfare and consumer welfare,

13it seems that both precedent and ease of measurement

14favor supplier welfare. The cases that have looked at

15monopsony abuses and at buyer cartels tend, on balance,

16to focus on the impact on suppliers rather than on

17consumers.

18In addition, there can be instances of

19substantial harm with little or no measurable effect on

20consumers. So, if the plaintiff had to show some sort

21of significant, discernible, provable effect on

22consumers, that would be harder. So, I am tending to

23favor a supplier welfare test, but you could use a

24consumer welfare test at least for most cases. Why is

25that? Because successful monopsonization is likely to

Page 110

1harm consumers in two ways.

2One, as many economists have pointed out, if the

3predatory bidding produces a net increase in long-run

4monopsony power, then there is likely to be a reduction

5in output. The dominant buyer is likely, on balance,

6over time, to buy less, and if it buys less, it is

7likely to produce less, and that means there is likely

8to be less final product on the market. And so, if the

9demand curve is neither totally vertical nor totally

10horizontal, if it is the normal downward sloping type,

11then less output is going to put some upward pressure on

12price.

13So, the mere output effect will tend to harm

14consumers, again, maybe not noticeably, but there is

15that linkage, and that does not require market power.

16That is, the dominant buyer does not have to have market

17power in the final product market for this effect to

18occur. It occurs through the output reduction caused by

19the monopsonization.

20There could be, though, a market power effect,

21as Rick may emphasize in his talk. Suppose the relevant

22market downstream was limited to the product whose input

23price was bid up. Then, if the dominant buyer

24eliminates its key rivals as buyers, it will also

25eliminate them as sellers. So, it may gain both

Page 111

1monopsony power upstream and monopoly power downstream,

2and so then that would magnify the potential consumer

3effect.

4We could, therefore, use a consumer welfare

5test, as Steve may suggest. There might have to be

6exceptions, though, from such a test where predatory

7bidding leads to monopsonization but consumers are

8unaffected.

9Whatever criterion is used for welfare, it seems

10to me that a welfare test would provide substantial

11protection to defendants. First, successful

12monopsonization appears to be rare and appears to be

13limited to certain markets, as Rick will suggest,

14markets where there is inelastic supply, and that is not

15commonly observed, typically in labor or natural

16resource markets.

17The power buyers that we all know or suspect,

18the Wal-Marts, the Barnes & Nobles, the Costcos, they

19don't induce lower prices by monopsonization. That is,

20they do not go to their suppliers and say, "I am going

21to cut back my output a little bit, and I expect,

22because that will reduce your marginal cost, that you

23will give me a lower price." Rather, they engage in

24bargaining tactics, and at the risk of oversimplifying,

25the way they obtain a lower price is, in essence, saying

Page 112

1that I am going to increase my purchases over what they

2would otherwise be if and only if you give me a lower

3price. So, that is not monopsonization.

4So, one, you have the limited set of cases, and

5two, you have all of those elements that the plaintiff

6has to prove. I will not repeat them again, but showing

7net long run harm to welfare is not an easy task. This

8is essentially a full rule of reason analysis, and as

9you well know, private plaintiffs do not often prevail

10in full rule of reason cases. That has been the record

11under Section 1 and is likely to be the record under

12Section 2 as well.

13You might say, "Ah, but private plaintiffs

14prevailed in LePage's and below in Weyerhaeuser," but

15the difference is in neither of those cases did the

16courts insist on a full rule of reason net welfare test.

17Is such a test unworkable? It is certainly

18reasonable to contend it is, but we do use it in

19horizontal merger cases under Section 7 and in full rule

20of reason cases under Section 1, and as Steve has

21pointed out, we are not really balancing immeasurables

22when we use this long-run welfare test. We are not

23trying to decide what's more important, national

24security or freedom of speech. We are asking whether

25the long-run impact on our target group, let's say

Page 113

1suppliers, is positive or negative.

2Still, still, the inquiry would not be easy for

3a defendant to predict its outcome, and the inquiry

4would stretch over a longer time period than in the case

5of a Section 7 matter certainly and probably the typical

6Section 1 matter, because we are talking about a

7long-run impact on welfare, and the key issue there, as

8Ken has stressed this morning, is entry barriers, and

9that is an uncertain and controversial topic. So, my

10sense is that we should not rely solely on a welfare

11test, that we should create an efficiency defense for

12the defendant, and for that, I have borrowed from the

13no-economic-sense test advocated by the Division.

14It seems to me that if the defendant can show

15that bidding up input prices was profitable, without

16regard to any increase in monopsony power, that it

17should have a complete defense. This would put the

18burden on the party that best knows its own

19profitability and would give it an out if it could

20provide a good answer to the question, why did you do

21this?

22So, to conclude, let me give you just a simple

23example. Suppose, as is my understanding of

24Weyerhaeuser's theory, suppose the dominant buyer

25improved its production process, lowering its marginal

Page 114

1cost. Then the new profit-maximizing price, without

2regard to monopsony power, might be an increase in

3output. That would entail buying more input, selling

4more of the final product, so there would be a margin

5reduction from paying more for the input and getting a

6little less for the final product, but if the lower

7marginal cost more than compensated for that, without

8figuring in any increase in margin due to monopsony

9power, then the defendant would be excused.

10I am happy to talk more about that, but my time

11is up.

12MS. SCHULTHEISS: Thank you, Jack.

13Our next speaker is Tim Brennan, who is, as I

14indicated, a Professor of Public Policy and Economics at

15the University of Maryland, Baltimore County. Professor

16Brennan also has been serving as the 2006 T.D. MacDonald

17Chair in Industrial Economics at the Canadian

18Competition Bureau.

19Before joining the University of Maryland,

20Professor Brennan held a number of positions focusing on

21economics and antitrust, including staff economist at

22the Antitrust Division, senior economist for industrial

23organization and regulatory policy on the staff of the

24White House Council of Economic Advisers, and a

25consultant to the Bureau of Economics here at the

Page 115

1Federal Trade Commission.

2Professor Brennan's research areas related to

3antitrust include regulatory economics, monopolization

4law, exclusionary conduct, vertical integration, and the

5competition-regulation interface. His articles have

6appeared in numerous journals in economics, law, and

7other fields.

8Tim, would you care to start?

9DR. BRENNAN: Thank you.

10I am grateful to the Department of Justice and

11to the FTC for the invitation to participate on this

12predatory buying panel. It is a great honor for me to

13be here. I am especially grateful because I have been

14thinking for longer than I care to remember about how to

15support Section 2 of the Sherman Act, and yet reconcile

16it with less controversial, more accepted frameworks for

17prosecuting cartels and horizontal mergers.

18I will offer a suggestion along those lines

19today. Although I believe that my suggestion will make

20deserving exclusion cases easier to bring, some aspects

21may be significantly different from established

22jurisprudence. For that reason, I particularly

23recognize the privilege of having a place at this

24distinguished table.

25Before proceeding, I need to say my statement

Page 116

1today reflects solely my own opinions and does not

2represent those of the Competition Bureau or any of its

3staff.

4For this complex topic, I offer a series of

5recommendations.

6Predation or exclusion? Pick one or the

7other -- they are fundamentally different.

8When first asked to participate in a panel on

9"predatory buying," my response was to object to the

10title. We should recognize that "monopolization"

11entails two essentially different types of practices,

12one that for shorthand could be called "predation," and

13the other "exclusion." The most succinct distinction is

14that predation cases involve doing too much of a good

15thing to bring about a bad result later. There, the

16understandable concern is with deterring energetic

17competition -- not discouraging firms from charging low

18prices, adding product features, and the like.

19Exclusion cases, on the other hand, involve

20doing a bad thing now. One way or another they come

21down to acquiring control and effective market power

22over a supplier or access to an input or service needed

23to compete, what economists called complements. The

24most explicit way to accomplish such control would be

25through a series of exclusive contracts with the

Page 117

1complement's suppliers. It may involve overbuying

2inputs through explicit purchase or, as I'll suggest

3below, bundling, rebates, or other forms of "leaving

4money on the table." I call this practice "complement

5market monopolization," or CMM.

6The major problem with single-firm conduct law

7is the failure to recognize the essential difference

8between these two types of conduct, leading to the

9counterproductive imposition of predation standards on

10exclusion cases. Perhaps the failure arises from a

11presumption that one statute -- Section 2 -- must imply

12one principle. Perhaps it follows from the persistent

13belief that Section 2 must be premised on harm to

14rivals. Since competition also harms rivals, Section 2

15law is thus driven by fear of over-deterrence. Instead,

16exclusion cases should be recognized as different, where

17we can apply horizontal tools and not predation screens

18to the delineation and protection of complement markets.

19Two, genuine predatory buying cases will be

20rare; when they occur, validate necessary assumptions.

21I would have changed the title of this panel to

22"Exclusionary Buying," because the leading cases involve

23creating of market power over complements. The recent

24DOJ/FTC cert petition in Weyerhaeuser v. Ross-Simmons

25illustrates an exception that proves the rule. The

Page 118

1exception is unusual, in that the concern is not that a

2timber processor would acquire so much control over a

3relevant market in uncut trees to be able to raise their

4effective price. Rather, according to the petition, the

5allegation is that a mill would pay too much for trees

6to drive out other buyers, with subsequent recoupment by

7cutting prices paid for trees in the future.

8I have little to say about which market power,

9price-to-cost, and recoupment tests are appropriate for

10preventing over-deterrence in these rare predatory

11buying cases. I do suggest that courts demand not only

12evidence appropriate for such tests. They should also

13demand evidence that specific assumptions behind

14strategic models are satisfied, i.e., that the alleged

15predator either has a reputation for non-profit

16maximizing behavior to protect, or benefits from

17identified asymmetric failures in capital markets.

18Theoretical possibility alone does not make a practice

19harmful.

20Three, for exclusion cases, the first and

21crucial step is to delineate a complement market being

22monopolized using the Horizontal Merger Guidelines

23procedures.

24Market power is often characterized not just as

25the ability to raise price but also as "the ability to

Page 119

1exclude." This is a mistake of imprecision. Ability to

2raise the price of X depends upon entry barriers or

3other impediments to competition, but those do not

4depend upon the price of X. Higher X prices would, if

5anything, encourage entry. Rather, the ability to

6exclude depends upon control over the prices of Y, Z, W,

7or something else needed to enter and produce X.

8Delineation of that relevant complement market

9should therefore be the first step in all exclusion

10cases. Taking Dentsply as an example, the case rested

11on the premise that the national distributors constitute

12what in merger contexts we would regard as a relevant

13market, in this case for the distribution of teeth to

14dental labs. The Merger Guidelines provide the useful

15framework for testing this premise. They ask whether

16teeth manufacturers would turn to other distributors,

17whether there would be entry into that distribution

18market in response to a "small but significant

19nontransitory increase in price" of using such dealers.

20I do not know the facts of that case and thus

21the answers, but the Merger Guidelines ask exactly the

22right questions. Cases eventually turn to evidence of

23entry or substitution into the complement market, but

24they do not make such concerns central -- the best

25indicator being the continued identification of a

Page 120

1relevant market as that in which the alleged monopolizer

2is already dominant, not that over inputs or services

3competitors need to compete. Control over such a

4complement market is not only sufficient to raise

5competitive concerns; it is necessary for

6anticompetitive exclusion.

7Hence, plaintiffs should focus on identifying

8that complement market and showing that the practices at

9hand cover enough of it to raise the complement's price.

10In effect, one should ask if one would be troubled if

11the complement providers covered by the alleged

12exclusionary practice merged. Unlike usual

13characterizations of monopolization cases, this is one

14we know how to answer -- use the Merger Guidelines. If

15the answer is no, stop; if the answer is yes, go to the

16next step.

17Next, having delineated the relevant complement

18market, the second step should be to establish the price

19effect in that market.

20Barriers to entry cannot be raised, and

21competition impeded, by any more than the extent to

22which the price of the complement can be raised.

23Sometimes this higher price will be explicit, sometimes

24it will be only an inferred higher price -- Professor

25Carlton has usefully called it a "shadow price" -- if

Page 121

1the exclusionary practice so ties up the complement

2market that only higher priced substitutes, including

3self-provision, are available.

4Explicit exclusive dealing contracts offer one

5such standard: Firms wanting to use those dealers would

6have to cover the cost of breaching the contract. Other

7alleged exclusionary practices, such as bundle discounts

8or royalty rebates, may create a significant price

9increase -- once one has established the first step.

10Next, the standard for assessing the

11exclusionary effect of a bundle or rebate is not whether

12an incremental price is below incremental cost, but its

13effect on the price of the complement.

14Following the last point, one could ask whether

15bundles, rebates, or other programs have to increase the

16effective price of the complement as much as it would

17explicit contracts. I have no reason to believe it

18should. Were we to follow the Merger Guidelines, as we

19should for complement market definition, we might only

20need ask if the practice leads to a small but

21significant nontransitory increase in price of the

22complement.

23This tells us that whether a bundle is

24anticompetitive has nothing to do with a predation-like

25test. It does not depend on whether the incremental

Page 122

1price of adding a good to a bundle, or of supplying more

2of a product given a discount, is less than some measure

3of marginal or average variable cost. Rather, it

4depends only on the extent to which such practices

5create market power in order to raise the price others

6must pay for the services provided by retailers,

7distributors, or other complement providers getting the

8discount.

9Predation case screens -- profit sacrifice,

10equally efficient competitor, and prior dominance -- do

11not belong in exclusion cases.

12Even for predation, we have heard today, some

13commentators have noted that some or all of the screens

14need not increase competition and consumer benefit.

15Nevertheless, they may be appropriate to prevent

16over-deterrence of competition through low prices or

17added features. However, in exclusion cases,

18controlling a monopoly share of complement markets is

19not inherently procompetitive, and thus need not have

20high bars for its protection.

21The profit sacrifice or "no business sense"

22test -- the two are equivalent if one assumes that

23"business sense" means "maximize profits" -- substitutes

24concern with intent and tactics for concern with

25effects, as if whether someone had been murdered depends

Page 123

1upon the price paid for the gun. Others have noted that

2it creates an absolute efficiencies defense, in that a

3penny of gain from a practice excuses untold

4anticompetitive harms. As Rick Warren-Boulton has said,

5the test is notably inappropriate when regulated

6monopolists do the excluding.

7Although I have criticized "raising rivals'

8costs," mostly for its emphasis on "rivals," Steve

9deserves enormous credit for pointing out long ago that

10predatory sacrifice and recoupment is unnecessary to

11point out the tactics that raise those costs. My

12difference is that I would focus primarily on the

13complement market.

14Ironically, the test also forgets that once upon

15a time, profit sacrifice implied previously unobserved

16efficiency, not anticompetitive harm. We learned that

17exclusive territories, exclusive dealing, tying, and

18even resale price maintenance must generate efficiencies

19because they reduce demand, making even monopolists

20worse off otherwise. That realization gradually

21reformed most vertical restraint law. Assuming now that

22a profit sacrifice must be anticompetitive forgets

23antitrust history and invites us to repeat mistakes that

24have not been fully undone after nearly a century.

25On equally efficient competitors, I point out

Page 124

1what should be obvious: Inefficient competitors hold

2down price. Complement market monopolization leading to

3their exclusion can raise price and harm consumers.

4Having gone after two sacred cows, I may as well

5finish off the herd: The Grinnell prior possession of

6monopoly test can also impede meritorious exclusion

7cases. It distracts attention away from the complement

8market, focusing instead on the characteristics of who

9monopolized it. Prior dominance could even be a

10defense, but once complement market monopolization is

11shown, it should be up to a defendant to claim that it

12has no consequence because of monopoly elsewhere in the

13production chain.

14Moreover, this test is counterproductive.

15Proving the cost, demand, and entry barriers necessary

16to establish prior dominance undercuts the argument that

17the alleged exclusionary practice makes a difference.

18Using Richard Posner's phrase, the monopoly should be

19"fragile" at worst. An exclusion case will be strongest

20if the sector would be competitive, but for the practice

21under scrutiny.

22Ask whether we would apply these standards to

23mergers. Should all mergers be legal unless one could

24show they would be unprofitable but for anticompetitive

25harm? Should any merger, including to monopoly, be

Page 125

1legal if a more efficient firm buys and eliminates a

2less efficient competitor? Of course not. Even prior

3dominance may make the incremental effect of a merger

4less troubling. If these tests would gut merger law,

5and if exclusion cases are akin to acquisitions in the

6complement market, they do not belong on this side of

7Section 2.

8Consider share-based rather than "all or

9nothing" remedies.

10Analogy to mergers opens the door to more

11creative remedies. Generally, either a practice is

12okay, or it is not and should be stopped. We should

13instead take a share-based approach. Exclusive dealing

14contracts, bundles, or other alleged monopolizing

15practices might have efficiency benefits. The problem

16is not the practices per se, but their scale -- that

17they pre-empt so much of the complement market to raise

18its price significantly. Rather, defendants should be

19allowed to retain the practice, but only over a

20nondominant share of the complement market, 35 percent,

2150 percent or some appropriate number. If the practice

22is actually efficient, it will be kept. If it serves

23only to exclude, this remedy would lead to its

24discontinuance.

25Last, focus on the creation of monopolies, not

Page 126

1their maintenance.

2About two years ago, I gave a talk at the FTC on

3these ideas, entitled "Saving Section 2." As I began,

4an economist there asked, "Why should anyone want to

5save Section 2?" My answer may not have satisfied him,

6but in short, it is that it can and should be saved.

7Were all Section 2, single-firm conduct cases about

8protecting a monopolist's rivals by drawing vague or

9impossible lines between competing just enough and too

10much, I might have agreed with the questioner. However,

11exclusion cases are not about maintaining monopolies but

12creating new ones. In focusing on complement market

13monopolization, such cases can and should be no more

14controversial than collusion and merger cases are today.

15Thank you again for the privilege of allowing me

16to share these observations. I hope I can clarify them

17through responses to any questions you have here and as

18they arise in the future. Thanks very much.

19(Applause.)

20MS. SCHULTHEISS: Our next speaker will be Steve

21Salop, who is a Professor of Economics and Law at the

22Georgetown University Law Center where Steve teaches

23courses in antitrust law and economics and economic

24reasoning for lawyers. Dr. Salop also has a consulting

25practice at CRA International involving a variety of

Page 127

1antitrust issues.

2Before joining the Georgetown faculty, Steve

3held positions at the Federal Trade Commission, and a

4while back, the Civil Aeronautics Board and the Federal

5Reserve Board. Professor Salop has written numerous

6articles in various areas of antitrust economics and

7law, many of which take a, quote unquote, "post-Chicago"

8approach. Professor Salop recently published two

9articles in the Antitrust Law Journal that concern

10exclusionary behavior and monopoly power.

11Of particular importance for today's hearing is

12his article, "Anticompetitive Overbuying By Power

13Buyers." It contrasts predatory versus raising rivals'

14costs, overbuying behavior, and I will let him go

15further into that.

16And in addition, Professor Salop has also

17consulted for Weyerhaeuser in its appeal and district

18court decision. Is that correct?

19DR. SALOP: Yes, that's correct.

20MS. SCHULTHEISS: So, with that, I will hand it

21over to you.

22DR. SALOP: Okay, thank you.

23I want to talk about these two types of

24overbuying, predatory overbuying, what Jack called

25predatory bidding, and raising rivals' costs overbuying,

Page 128

1but to just set it up, I think a key place to begin is a

2notion that Tim Brennan touched on, the fact that in

3Section 2, there are really two distinct paradigms, and

4I believe that the way individual people think about

5Section 2 has a lot to do with which paradigm they have

6in mind, you know, which one animates them, and so I

7want to stress the difference between these two

8paradigms.

9One paradigm is the predatory pricing paradigm,

10seller-side predatory pricing, and the other is the

11raising rivals' costs or non-price predation paradigm.

12Now, in my view, and a key element in what I am going to

13talk about today, and, indeed, much of my work, is that

14conduct that fits into the raising rivals' cost paradigm

15raises much greater concerns than conduct that fits

16within or that people characterize as the predatory

17pricing paradigm.

18Now, we all know the claims about why predatory

19pricing is seldom attempted and rarely succeeds. In the

20short run, the predator loses more money than the

21victim. Secondly, it only works if the victim exits.

22Otherwise, there is no -- they will never be able to

23recoup. And third, consumers benefit from the lower

24prices, and the harm to consumers is mere speculative

25impact in the future. So, for all those reasons, it is

Page 129

1argued that predatory pricing is unlikely to be tried,

2it is unlikely to succeed, and it is unlikely to harm

3consumers, and therefore, we should have a really light

4hand in predatory pricing.

5Taken as a paradigm, conduct that raises rivals'

6costs raises much greater antitrust concerns; hence,

7let's say it is more likely to succeed, more likely to

8harm consumers. Why? Well, first, there is no need to

9induce competitors to exit. If you raise competitors'

10costs, variable costs, they will tend to raise price,

11and the excluding firm will gain even if the competitors

12do not exit. You would rather compete against a

13high-cost competitor than a low-cost competitor.

14Secondly, there is no necessity for short-run

15profit sacrifice, chronological profit sacrifice, of the

16sort there is in predatory pricing. If the rivals'

17costs are increased, they will raise price immediately.

18The predator, the excluding firm, will gain immediately.

19So, there is no issue that the predator has to lose

20money for a while and then only gain later.

21Similarly, there is no short-run consumer

22benefit, and this one I think is very important. In

23predatory pricing, the consumers inevitably benefit from

24that lower price in the short run. In raising rivals'

25costs, there is no inherent consumer benefit. You raise

Page 130

1rivals' costs, they raise price or contract, and the

2defendant raises price, so consumers are harmed

3immediately. There is no such thing as naked predatory

4pricing, you know, just all bad, the way there is with

5naked price-fixing, but there is naked raising rivals'

6costs. One could conceive of that burning down the

7factory, so on and so forth, and such conduct that

8actually shows up in certain cases.

9So, for all those reasons, I think that you have

10these two paradigms that are distinct, and I think most

11of the time you should be thinking in terms of the

12raising rivals' costs paradigm. I think it is a better

13paradigm for Section 2. I think that the predatory

14pricing is the exceptional paradigm, not the norm.

15Well, now let's apply this to anticompetitive

16overbuying. Now, there are -- by overbuying, I mean

17conduct where the defendant goes into the input market

18and bids up the price of the input. Usually if you bid

19up the price of the input, you are generally almost

20surely going to buy more than you would have otherwise.

21So, it is often called overbuying. Indeed, in the

22literature, it was initially referred to as overbuying

23cases, and I guess sort of the classic case that

24economists studied initially was the bauxite aspect of

25the Alcoa case, where Alcoa was alleged to have

Page 131

1overbought bauxite in order to raise the costs to its

2aluminum rivals.

3Well, there are two distinct overbuying

4allegations that correspond analytically to the two

5exclusion paradigms. There is predatory overbuying,

6Jack talked about, and then there is raising rivals'

7costs overbuying, and the difference between these two

8paradigms is the goal of predatory bidding or predatory

9overbuying is to gain monopsony power to the input

10market, as Jack pointed out. The goal of raising

11rivals' costs overbuying is to raise your rivals' costs

12and then gain market power in the downstream output

13market.

14Okay, so if you think about Alcoa, they could

15have overbought bauxite for -- one reason would be to

16ultimately knock out the other purchasers of bauxite so

17it could then be a bauxite monopsonist, and the

18alternative would be that they did it in order to raise

19the price of bauxite to its rivals so that they could

20ultimately monopolize, raise the price, of aluminum. Of

21course, in a given case, you could have both, but at the

22same time, in a case, you could have one or the other.

23Now, interestingly, in the Weyerhaeuser case and

24the Ross-Simmons case, both allegations were made in the

25complaint. It was alleged that Weyerhaeuser, by its

Page 132

1conduct, would ultimately gain monopsony power in the

2timber market, in the purchase of timber, and secondly,

3it was argued that Ross-Simmons would gain market power

4in a downstream alder wood, alder hardwood market, and

5Ross-Simmons did not carry its burden on the raising

6rivals' costs piece, and so the part that has gone up is

7just the predatory overbuying piece.

8Okay, well, how do I think we should evaluate

9these two types of conduct? I want to separate them.

10So, first, the predatory overbuying, as I said, it's

11market power in the upstream market that's the goal, and

12in general, I have in mind a four-step legal standard,

13very close, very, very close, to what Jack Kirkwood

14called for, and I will talk about it and then stress the

15differences.

16So, the four steps would be, you have got to

17show buyer power, monopsony power, and artificially

18inflated input purchasing, and it's really the latter,

19you have to show that they bought more and that the

20price went up. You have to show exit or permanent

21capacity reduction of the input market competitors, that

22should be, and then there has got to be some kind of

23recoupment through buyer-side monopsony power in the

24input market, and finally, and this is very important,

25you have to show net consumer harm.

Page 133

1Now, usually in these cases there will be a

2short-run consumer benefit during the predatory period,

3because when the defendant buys more of the input, it

4will produce more output, and so the price of output

5will go down, and then during a recoupment period, it

6goes in the other direction, so consumers benefit in the

7short run, harmed in the long run, and for that reason,

8in order to show -- in order to gain -- show liability,

9the plaintiff would have to show consumer harm on

10balance.

11Okay, now, let me go through the steps in a

12little more detail. First of all, note, Jack Kirkwood

13said, well, I'm not sure you need consumer harm, maybe

14it's enough to have supplier harm, so that is one

15difference between our standards to date. Jack actually

16in his article had this last step, consumer harm. So,

17he has broadened his position today.

18Now, the first step, the issue here is the

19question is, is the increased purchasing artificial or

20is it competitively driven? Now, you know, there are a

21lot of good reasons why a firm may increase its input

22purchases in this year versus last year. For example,

23maybe the demand for its output went up. It needs more

24inputs to produce more output. Maybe it is not so much

25its demand went up, maybe it decided to change its

Page 134

1business strategy. Maybe it decided to decide to grow

2its market share rather than go for a high price in the

3output market. Or maybe it got a new production process

4that is more efficient, and that leads it to want more

5inputs in order to expand. Fourth, maybe something

6happened in the input market. It used to have monopsony

7power, and it finds it has lost that monopsony power or

8has less of it, and so it wants to stop acting like a

9monopsonist.

10All four of those reasons would lead it to

11increase its demand for inputs. As its demand for

12inputs goes up, the price of the inputs would tend to

13rise, and its purchases would tend to rise. So, for

14these reasons, I mean, what would be most suspicious

15would be if the defendant bought extra inputs and then

16did not use them, just warehoused them, all right,

17because that would suggest it was not buying more in

18order to produce more output, but rather, did it in

19order to raise price and drive its rivals out of

20business. So, warehousing would be an issue.

21Of course, the fact that it has inventories does

22not necessarily mean it is warehousing. It could have

23been an error. It could have just bought more thinking

24it was going to need it and then it just did not need

25it. So, one has to be careful there.

Page 135

1Because of the concern, you know, that there are

2all these legitimate reasons why you might want to

3purchase more output, and the fact that there is this

4inherent short-run consumer benefit, I am very worried

5that there could be false positives, and for that

6reason, I am willing to put on the Brooke Group style

7test of output priced below cost, where it's really sort

8of that marginal revenue product, or Rick went on to the

9value of marginal product, being less than the input

10price. Because here, like with predatory pricing, there

11is inherent consumer benefit in the short run that may

12or may not be offset by consumer harm in the long run,

13and because there is that balance, it is a lot like

14predatory pricing, and so the rule might be close to the

15predatory pricing rule.

16I also think that in the end, this is why the

17Supreme Court will opt for a Brooke Group kind of test,

18because it is so close to the reasoning in Brooke Group

19that they are going to probably find it irresistible to

20change the rule.

21Now, it is possible that there is no consumer

22benefit in the short run. If the demand for the output

23was perfectly elastic, demand for the input was

24perfectly elastic, in terms of price taker, then, when

25it increases its purchases, it will not reduce the price

Page 136

1that consumers pay for the outputs. There would be no

2consumer benefit. But on the other hand, and under

3those same circumstances, there would not be any

4consumer harm. So, under those circumstances, the

5plaintiff would lose anyway. So, I think the Brooke

6Group, adding the price-cost test in Brooke Group, will

7not cause any damage in this situation.

8Now, when it comes to raising rivals' costs

9overbuying, I feel a more interventional stance is

10necessary. Again, I have got a four-step legal

11standard. I do not have the Brooke Group piece, and, of

12course, the analysis is somewhat different, because the

13goal here is to gain market power in the downstream

14market, not to gain market power in the upstream market.

15So, you would still ask whether there were good reasons

16for the firm to increase its demand for the input, but I

17would not go so far as the Brooke Group test.

18Here, you not only have warehousing could be a

19concern, but also naked purchasing. Now, what I mean by

20naked purchasing is suppose the firm does not even use

21the input. It might buy up some of the input for the

22sole purpose of making it more expensive for its rivals.

23For example, in the Alcoa case, it was alleged that

24Alcoa bought exclusive contracts to electricity from

25utilities where it never had a plan. They just bought

Page 137

1the exclusive so that other aluminum companies could not

2buy electricity in those regions. So, that would be an

3example of naked buying, there where an exclusive wasn't

4overbuying, it was simply buying the exclusive.

5You would never have naked overbuying for

6predatory overbuying, because you can't get a

7monopsony -- why would you want a monopsony? Why would

8you want to knock rivals out and get a monopsony over

9some input that you don't even use? So, you know, it

10just does not compute.

11So, assuming that you can show inflated input

12purchasing, then you go through the standard type of

13raising rivals' costs analysis to show consumer harm.

14You have to show that rivals' costs were raised

15materially, but that's not enough. That's just harm to

16competitors. We know in Section 2 that is not enough.

17You also have to show harm to competition, as you have

18to show downstream market power, some power over price

19downstream. And then, as before, you need to show net

20consumer harm, because the overpurchasing could have

21efficiency benefits. Maybe the firm is producing more

22output and, you know, competing harder, that is a

23benefit. So, you want to have that consumer harm

24standard. As I said, it is more interventionist.

25Now, by consumer harm, I really mean it. I mean

Page 138

1true consumer harm. I do not mean supplier welfare, and

2I do not mean what Robert Bork called the consumer

3welfare standard, which was really a total welfare

4standard. I think Bork was either deceiving people,

5which is what Herb Hovenkamp has said, or maybe Bork was

6just confused, and the reason I think it is possible he

7was confused is that if you care about the total welfare

8standard, then competitor injury is enough to carry the

9day, because competitors are part of total welfare.

10Indeed, it is easy to construct mainstream

11simple examples in which conduct could harm competitors,

12consumers could benefit with lower prices, but yet total

13welfare could fall, and yet under the total welfare

14standard, total welfare would go down, and so if you had

15in mind a total welfare standard, that conduct would be

16illegal. Now, I just cannot believe that Bork would

17want to make such conduct illegal. So, I have chosen to

18believe, out of respect, that he was confused rather

19than attempting to create a consumer protection problem

20or a court protection problem.

21There are other reasons why I think the consumer

22welfare standard is better. It is consistent with

23precedent. It is what the agencies use. It is simple

24to evaluate. Actually, the analysis I have done and

25several other economists have done, it has shown that

Page 139

1actually, even if you care about total welfare, for

2several reasons, the consumer welfare standard actually

3could lead to higher total welfare, and I will go into

4that if that comes up later on.

5Finally, I think the consumer welfare standard

6supports innovation better than does the total welfare

7standard. Again, I will go into that if there is time

8later on.

9So, thank you very much.

10(Applause.)

11MS. SCHULTHEISS: Okay, our next speaker is

12Dr. Rick Warren-Boulton, a principal of MiCRA,

13Microeconomic Consulting and Research Associates, an

14economics consulting firm and research firm specializing

15in antitrust litigation and regulatory matters. Before

16joining MiCRA -

17DR. WARREN-BOULTON: We like to call that

18"MiCRA."

19MS. SCHULTHEISS: I like to call it "MiCRA."

20Rick Warren-Boulton served in a number of positions

21involving economics and antitrust law. He was an

22Associate Professor of Economics at Washington

23University in St. Louis. He was the Chief Economist for

24the Antitrust Division at the Department of Justice. He

25was also a resident scholar at the American Enterprise

Page 140

1Institute.

2Rick has authored numerous publications -- you

3laugh -

4MR. KIRKWOOD: Are you still welcome there?

5DR. WARREN-BOULTON: Yes, for lunch.

6MS. SCHULTHEISS: Old friends, huh, Rick?

7He has authored numerous publications, primarily

8in the application of industrial organization economics

9to antitrust and regulation. He's served as an expert

10witness in numerous cases, including the Department of

11Justice, U.S. versus AT&T, and for the FTC in our FTC

12versus Staples and Office Depot merger. He also was an

13expert with the Department of Justice in the United

14States versus Microsoft.

15In addition, Dr. Warren-Boulton was recently a

16consultant in support of the Ross-Simmons side in its

17effort to convince the Supreme Court to reject

18Weyerhaeuser's petition for certiorari. Rick is yet

19another one of our panelists who has some connection to

20the Weyerhaeuser case.

21Rick, with that...

22DR. WARREN-BOULTON: Thank you. You know you

23are getting old when you start talking about when you

24were an expert witness in the AT&T case.

25MS. SCHULTHEISS: And now they have all

Page 141

1re-merged.

2DR. WARREN-BOULTON: -- in 1981, and in terms of

3my connection with this case, obviously I was -- I tried

4to help Ross-Simmons in arguing to the DOJ and the FTC

5that they should not file what they did, so it was a

6complete failure, but I will just keep trying anyway,

7because I might as well.

8You know, the issue that I would like to talk

9about here today is really a question as to whether or

10not we need a test, or at the extreme, even a safe

11harbor, and I think we need to distinguish between tests

12and a safe harbor for cases of strictly what I would

13call predatory overbidding to achieve monopsony power,

14and I think what we are doing is, in the order of the

15speakers here, we are going from general to specific.

16You know, I think that the much more broader and

17frankly more interesting question is exclusion, and I

18think Tim is right when he says if you come away from

19this with anything, it is, for heaven's sake, do not get

20predatory pricing or predatory overbidding mixed up with

21exclusion, and then I think Steve would say, do not get

22it mixed up with raising rivals' costs, too, and I

23agree, very much so, that those are two very important

24questions.

25I do not think that the analysis, nor the

Page 142

1implications for policy, are the same for the three of

2them, and I think what I am going to talk about today

3is, in fact, the least interesting and least important

4of the three, but that is what I thought my topic was,

5and so that is what I want to talk about.

6You know, I want to begin by distinguishing

7between an economic definition of predation and a legal

8test for predation, because coming up with a definition

9for predation is extraordinarily easy. Any of us can

10define predation. I was thinking about it before. I

11think it is the reverse of pornography. We all know it.

12We don't know it when we see it, but we know how to

13define it.

14You know, testing predation or coming up with a

15test for predation is very difficult, because a test for

16predation involves a balancing of the costs and

17probability of false positives against the costs and

18probability of a false negative, and so finding the

19right tests or the right safe harbors for each

20distinguishable situation is fundamentally an empirical

21question, and it's a different question for every

22situation that you can distinguish.

23What I think for us at the moment to think about

24it is is a couple of implications of that for coming up

25with tests for predatory overbidding to achieve

Page 143

1monopsony power. The first is, there is absolutely no

2need to have the same rule for predatory pricing to

3achieve monopoly power as predatory overbidding to

4achieve monopsony power. You don't need to have the

5same rule unless you can't tell the two situations

6apart. But even a court, to be honest, it seems to me,

7is unlikely to confuse predatory pricing with predatory

8overbuying.

9You know, when you have predatory pricing, you

10look at the final product price, and it goes down, and

11then it goes up, and when you look at predatory

12overbidding, the initial price goes up, and then it goes

13down. So, it seems to me that trying to order that you

14need consistency between monopsony and monopoly

15situations when what you are looking at is completely

16different behavior seems to me to be unnecessary and

17sort of pointless. It may not be the hobgoblin of

18little minds, but there is absolutely no reason to say

19that just because something is over here in the monopoly

20world, that we should apply the same rule in a monopsony

21world.

22The second is, I think it is very easy to

23distinguish between predatory overbidding and what I

24think is going to be Janet's concern, which is between

25predatory overbidding and either the exercise of

Page 144

1monopsony power or, more importantly, the exercise of

2bargaining power by power buyers. When you have

3predatory overbuying, what happens is your input price

4goes up, and then it goes down. That is what we are

5looking at today. When you are looking at bargaining

6power, in my experience, input prices go down and then

7they stay down. You know, power buyers like Wal-Mart

8pay lower prices, you know, from the get-go. Wal-Mart,

9to my knowledge, has never been accused of paying too

10high a price, even temporarily.

11Once more, of course, is you can distinguish the

12output effect, the classic monopsony problem, is you buy

13too little. You know, Wal-Mart buys a lot. So, it

14seems to me, you know, there should not be much concern

15here about false positives when we are facing

16allegations of predatory overbidding, and I hope that

17Janet will disagree strenuously, and I am sure she will,

18because otherwise, it will be no fun at all, but the

19result of this, I think just as a sort of preamble, is I

20find no urgent need for a safe harbor to deal with

21allegations of predatory overbidding to gain monopsony

22power. I do not think we should expect a flood of cases

23that might be false positives, that if we do not

24immediately nip this in the bud by allowing a

25Brooke Group safe harbor for predatory overbidding, and

Page 145

1I am certain Janet will disagree with that one.

2The second point to me is there is no reason why

3we can't proceed inductively, decide Ross-Simmons and

4any of the other rare cases that arise on the merits,

5and then try to generalize them. There are two

6procedures here. Lawyers, in my experience, the idea is

7that you have a whole bunch of cases, you try them on

8the merits, and then you ask, "Gosh, is there some

9common principle that's going on here?" It is very much

10of an inductive reasoning process, and it works very

11well.

12I think the economist approach is highly

13complementary. We go the reverse. We think

14deductively. We think what are the first principles,

15and we try to deduce, deductively, what the right

16principles are, and, you know, with any luck, it is like

17two people tunneling from opposite sides of the

18mountain, you know, the lawyers going inductively and

19the economists going deductively, and if all goes well,

20we meet in the middle, and if we don't, we just keep

21going until we get the other side, and we have two

22tunnels.

23But, you know, this strikes me as a situation in

24which I think the lawyers -- I mean, look at how many

25years we have been trying deductively to figure out what

Page 146

1the right principle to use is on predatory pricing for,

2you know, on the output side. We can't figure out the

3damned thing. I mean, you know, look how far apart

4those two gentlemen are, opposite sides of the table for

5God's sake. They are never going to agree. So, I think

6that the deductive approach really -- I mean, I think we

7have plenty of time to get a few more Ross-Simmons cases

8and then figure out how to go on.

9But let us suppose -- you know, at this point, I

10think I could sit down and say there is no need for me

11to proceed further, but unfortunately, my panelists

12won't let me do that, and anyway, I think I have another

13six minutes. So, let's ask the question. Let's suppose

14that you did do the unnecessary thing and you did ask,

15well, let's ask, what kind of tests should we sort of

16come up with for predatory overbidding to achieve

17monopsony power? And the question I think you want to

18ask is, what are the questions we want to ask?

19And I think you really want to ask things like,

20is monopsony different from monopoly? Are bidding

21markets -- because remember, this was specifically a

22bidding case -- are bidding markets different from, you

23know, other markets? And I think that the first answer

24is -- on both of them is yes. If you look at the

25difference between monopsony and monopoly, I have a

Page 147

1paper, which I am not posting on the web site until I

2get a chance to sort of, you know, get all the

3information, getting digs at Janet, so I will post it

4later, but, you know, two of the main differences

5between monopoly and monopsony it seems to me is

6monopsony is much rarer, and it is easier to identify.

7Monopsony is rarer for several reasons.

8Basically supply inelasticity is much rarer than demand

9elasticity. We all know this. Why is it? Well,

10producers can substitute easier than consumers in

11general, but most important, you know, in consumption,

12diminishing returns, you know, the demand curve slopes

13downwards, I mean, we all know that.

14On the other hand, in production, the norm is

15constant returns to scale, and so what we tend to get

16is, you know, a situation in which -- I just -- you

17know, I don't like -- I don't have a blackboard or a

18white board, but if you were drawing a demand and supply

19curve, you would feel perfectly normal drawing a

20horizontal supply curve and a downward-sloping demand

21curve, something which would look like this, but most of

22you would think that what's somewhat less usual is to

23have a completely horizontal demand curve, supply -- you

24know, upward-sloping supply curve. I mean,

25fundamentally, demand curves slope downwards. Supply

Page 148

1curves usually are horizontal. So, the reason why we

2get that is I think because there's a real difference.

3There is diminishing returns of consumption and

4relatively constant returns on production.

5The second thing is I think monopsony is much

6easier to identify, monopsony power, for several

7reasons. One is that supply inelasticity is really only

8observed in a few situations, and when I stop and I

9think about what they are, they are most importantly, in

10my experience, when a product is a by-product of some

11other product or when it is an exhaustible natural

12resource. Both of those are very unusual, and, you

13know, both of them actually apply in this particular -

14in the Ross-Simmons case. Alder turns out to be a

15by-product or what they call a come-along product in the

16industry, and it is an exhaustible product. So, you can

17use the screen essentially of a by-product as a -- to

18limit the scope of any decision.

19By the way, there is a second reason, which I

20always find kind of fun, which is why observing

21monopsony is so much rarer than observing monopoly.

22Monopsony is inefficient, and so nearly all monopsony

23situations or potential monopsony situations are solved

24by vertical integration, okay? If you are an aluminum

25plant and you are right next to an electricity dam and

Page 149

1you are the only buyer of electricity, you have

2monopsony power towards that supplier, say, of

3electricity, and exercising monopsony power can be

4profitable. It is even more profitable, though, if you

5buy the dam and get rid of the inefficiency.

6So, vertical integration is profitable because

7it gets rid of the inefficiency of monopsony, okay? But

8vertical integration backwards is much, much easier than

9vertical integration downwards and to the consumer, and

10the nice test for this is if you ask where do we

11actually see monopsony being exercised, in most cases,

12it's situations where vertical integration is not

13possible, and the classic example of that, at least

14since the Civil War, is labor. You cannot buy people,

15okay? And that is why the other day, you know, if you

16were reading the New York Times, you would find that it

17said a class action case, you know, on monopsony, who is

18it? It is by a group of nurses who are being -- the

19assertion is that they are being monopsonized. Now, if

20the hospital could just buy the nurses, there would not

21be a monopsony problem, okay? So, most monopsony cases

22are in labor, because vertical integration does not

23solve the problem.

24All right, the second main difference that I

25think we see between everything else we are talking

Page 150

1about is that these are, of the cases that we are

2talking about here, Ross-Simmons, is it is a bidding

3market, and the question is, just like monopoly is

4different from monopsony, bidding markets are different

5from markets in which there simply is single price. The

6classic problem with predatory pricing has always been,

7and I think it was Steve who introduced this, that it is

8inherently implausible, because the cost to the predator

9with a high market share is so much greater than the

10harm to the victim.

11If you have a 90 percent market share and you

12engage in predatory pricing, you bear 90 percent of the

13costs, and the guy with the 10 percent share bears 10

14percent of the costs, and that is so stupid, you know,

15that it is fundamentally implausible. The guy does not

16believe you are going to keep shooting yourself in the

17foot over and over and over again. So, it is sort of

18hard.

19But, you know, when you can price-discriminate,

20if the potential competitor can price-discriminate, if

21prices are individually negotiated, then the cost to the

22predator is no longer linked to the market share. There

23is no longer any connection between the relative cost of

24the predator and the predatee, you know, and then market

25shares, and so you have a fundamentally different

Page 151

1probability, you know, of a false positive here.

2So, you know, my conclusion, looking at this, is

3that if I am looking at the specific problem that we are

4dealing with here, which is predatory overbidding to

5achieve monopsony power, that if you have a combination

6of three things, you have a very good probability of a

7real case being there and a very low probability of

8having a false positive. Those three things are a very,

9very low supply elasticity of the input, which as I say

10is almost always just when it is either a by-product or

11a natural resource; secondly, the ability of the`

12predator to price-discriminate, particularly in bidding

13markets; and third, very strong barriers to re-entry.

14If you have those three things, then I think you

15have basically passed sort of the structural criteria

16that you need to pass. You don't pass -- you know, I

17think you have to pass all of those three, but if you

18pass those three, then I think you have a very good

19argument here that the predatory overbidding to gain

20monopsony power is, in fact, a realistic effort.

21Recoupment, what I am basically saying, is very

22likely, and then you do not need to get into the kind of

23cost tests that are in Brooke Group that I think most of

24us have looked at and said, the problem with these cost

25tests is that they generate a very -- particularly if

Page 152

1you allow them as safe harbors, they allow a high rate

2of false negatives. If you look at Ross-Simmons, it

3fits all of those, alder is a by-product, prices are

4individually negotiated, and the equipment that the

5sawmills was using is actually specific to alder, okay?

6If Ross-Simmons' equipment could have been used either

7for alder or soft wood, then we would not be here. Then

8basically, you know, Ross-Simmons could have been a

9hit-and-run entrant. They could have come in, they

10could have gone on, but the specificity of the equipment

11means basically that there is a very high barrier to

12re-entry.

13If you look at those three criteria, I think it

14merits what the Ninth Circuit sort of was saying. It

15was saying that predatory overbidding for a resource

16input in a highly inelastic supply, combined with a high

17barrier to entry by the downstream firms, you know, that

18is an extraordinarily narrow set of events, and if you

19look at that and you combine that with a

20no-economic-sense test, to understand the strategy and

21say does it pass the economic sense, I do not think

22there is any need for going into, you know, cost-based

23tests. I think what we could basically do is accept the

24Ninth Circuit decision, construe Ross-Simmons narrowly,

25and we can all wait for the next 20 or 30 years to see

Page 153

1if another case ever turns up.

2You know, finally, in terms of the two gentlemen

3over here, you know, I think we owe an enormous debt of

4gratitude to what used to be called the Chicago School,

5but I guess in this case more the University of Virginia

6school, for 40 years of reducing false positives, and I

7think that is yeoman's work at the academic level. I

8would like to believe that the Reagan Administration and

9Bill Baxter and the people who came in with him were at

10least sort of implementing that, but I think one thing

11we should realize is that getting rid of false negatives

12is a complement to reducing false positives.

13We all know on deterrence that optimal penalties

14are a decreasing function of the probability of a false

15positive, and for those of you who do not quite

16immediately know what that means, what it means

17basically is now that we have DNA testing, the case for

18capital punishment is much stronger. That is the

19easiest way to think of it.

20So, what has happened is we have an accumulated

21experience, we have many cases of alleged predation, we

22have economists working on it, and I think we have

23greatly reduced the level of false positives, and I

24think the problem is, of course, is once you have

25reduced the probability of false positives

Page 154

1significantly, then it becomes efficient and desirable

2to turn to the question of can we reduce the number of

3false negatives. So, I think what we are seeing over

4here, and I know Professor Elzinga is going to shoot me

5for this, but I would say Professor Elzinga has made

6Professor Bolton possible. I leave the two of you to

7duke it out.

8(Applause.)

9MS. SCHULTHEISS: Our final speaker is Janet

10McDavid. Janet McDavid is a partner in the Washington,

11D.C. office of the law firm Hogan & Hartson, where she

12focuses on antitrust and trade regulation, litigation

13and counseling. She is widely recognized as a leading

14authority in antitrust law, has been included in many

15guides to top antitrust lawyers, and an author of many

16books and articles on antitrust law.

17Ms. McDavid was previously the Chair of the

18Section of Antitrust Law of the American Bar Association

19and also a chair of its committee on Section 2 of the

20Sherman Act. As a member of the Antitrust Council to

21the U.S. Chamber of Commerce, Ms. McDavid also brings to

22us a business perspective.

23In addition, of particular relevance, again, to

24today's hearing, Ms. McDavid was a co-author of the

25brief for the Business Round Table and National

Page 155

1Association of Manufacturers in support of

2Weyerhaeuser's petition for certiorari, and Ms. McDavid

3was also the lead author of a recent article in the

4National Law Journal entitled "Predatory Purchasing?"

5And at this, I give Janet the job of cleanup and

6responding to some of the comments that have been made

7thus far.

8MS. McDAVID: I don't think my views will be

9quite as extreme as Rick may have suggested. My

10perspective here is that of a practicing lawyer who has

11to try to advise clients on where the line is between

12conduct that might be unlawful under the antitrust laws

13and conduct in which they can engage with relatively low

14risk, and Rick has suggested that greater clarity in

15this area really is not necessary, we should just allow

16the law to develop for a while and see whether

17inductively or deductively we can establish some rules,

18because the cases are rare.

19I am not seeing dozens of these cases in my

20practice, but I am increasingly seeing a lot of

21monopsony questions coming up from my clients, and so I

22do not think these cases are likely to be as rare as

23perhaps has been hypothesized. I am hoping that these

24hearings or the Weyerhaeuser case in the Supreme Court,

25whichever comes out first, can provide an answer with

Page 156

1respect to the predatory purchasing conundrum that can

2reasonably be applied by businesspeople and their

3lawyers.

4Remember, the purchasing department of even a

5Fortune 100 company does not include these guys. They

6would not know a downward sloping demand curve if it hit

7them in the face. And so the notion that that is a

8standard that can be applied by a businessperson in a

9purchasing department is just unrealistic.

10It is actually quite funny, I think, that Rick

11referred to Justice Stewart's definition of pornography,

12because that is also the analogy that I had here. I

13think the -

14DR. WARREN-BOULTON: Like minds.

15MS. McDAVID: -- Ninth Circuit standard in

16Weyerhaeuser provides much less guidance than Justice

17Stewart's standard with respect to pornography. There

18is no basis here for advising a client, and this is a

19serious mistake, because ultimately, businesspeople have

20to understand the rules. Justice Breyer, when he was

21still on the First Circuit, said that antitrust rules

22must be clear enough for lawyers to explain them to

23clients and be administratively workable, and that in

24formulating antitrust liability standards, the courts

25must consider what advice the lawyer is going to give.

Page 157

1Chairman Majoras made essentially the same point

2when she opened these hearings on Tuesday. The process

3of distinguishing between the permissible and the

4impermissible must be relatively consistent and

5transparent so firms can incorporate it into their

6decision-making.

7So, where does that leave us with the Ninth

8Circuit? We have a standard of liability that allowed a

9jury, using 20/20 hindsight, to determine whether

10Weyerhaeuser paid a higher price than necessary to

11prevent Ross-Simmons from obtaining its inputs at a fair

12price and that Weyerhaeuser may have purchased more logs

13than necessary. How is a business person supposed to

14apply that standard? It is entirely vague, open-ended,

15and subjective. It gives us no way to draw the lines,

16and a jury exercising its 20/20 hindsight can come out

17in a completely different place than a perfectly

18rational businessperson did.

19I think it is significant that Weyerhaeuser

20never lost any money on any of this. The division was

21operating profitably the whole time. I reread the Ninth

22Circuit decision this morning, and the Court talks about

23the need for recoupment in order to render the strategy

24profitable, but it does not actually impose that

25requirement as part of the test. So, the Court gave lip

Page 158

1service to recoupment without ever requiring it as part

2of the standard.

3So, one of the effects of all of this is to prop

4up less efficient firms, such as Ross-Simmons, that are

5unable to operate profitably at the same input prices

6that a more efficient rival can afford to pay. Now,

7firms compete for inputs just as they compete for sales,

8especially if the inputs are scarce, and there is

9language in the Ninth Circuit's decision that suggests

10that the inputs were becoming more scarce. They

11certainly were not more plentiful. So, we have a

12circumstance in which firms are going to be deterred

13from aggressive buying by the threat of liability for

14treble damages as a consequence of a standard that they

15cannot understand and cannot apply.

16How is a firm supposed to know if a jury will

17later determine whether it bought more than it needed,

18whether it paid a price that was too high, whether it

19paid a price that was not fair? If it buys too much,

20there are lots of reasons that that could have happened,

21as Steve was explaining earlier. It might have decided

22to stockpile inventory to preclude future shortages or

23to hedge against a future price increase. That happens.

24We had it happen, for example, during the

25Hurricane Katrina circumstance. Oil companies had

Page 159

1inventories. They were able to sell it off during a

2time of shortage. They might have overestimated their

3needs. Businesses make mistakes. They might have

4assumed that demand was going to increase, and it did

5not. Again, businesses make mistakes. They may have

6planned that sales were going to grow and guessed wrong,

7or they may have chosen to deal with the predictable

8supplier who has a reliable source and paid a slight

9premium to do so.

10Any one of these could, in hindsight, become the

11basis for liability by a jury that is exercising a

12standard based on fairness. Under no other circumstance

13in the antitrust laws, except perhaps the regrettable

14Robinson-Patman Act, do we consider a fairness standard.

15It is simply not administrable. So, people like me will

16have a hard time telling our clients what they should

17do, and the businesspeople on the ground will have a

18very hard time knowing what they should do and what they

19can do safely and fairly.

20Now, I will recommend to my clients, as I do in

21all circumstances, that they document the reasons they

22are doing things if they face a risk. I always tell

23them that I would prefer to create our own legislative

24history, that in the event the inevitable happens, I

25would like there to be evidence in their files that

Page 160

1explains, without us having to go back and explain it

2again, why they did what they did and that there was a

3rational business reason for what they are doing. I

4will also ask them, can you make money if you purchase

5at X price? Now, the unfortunate thing is, while I

6think that is a perfectly rational question to be

7asking, it is not relevant to the analysis in the Ninth

8Circuit, whether or not they were going to be able to

9make money if they paid X price for their inputs.

10I do not understand why the Brooke Group

11standards should not be at least relevant to the

12analysis here. These are standards that business people

13actually can understand. Very few of my clients have

14ever known what their average variable cost is, but they

15do know, if we talk about it in kind of gross terms, if

16you pay X or if you charge X, will you make money or

17will you be operating at a loss?

18They also can understand the notion of whether

19or not it will be profitable in the longer term. These

20are standards businesspeople can understand, and the

21courts have got to give them some kind of guidance here,

22and if the courts fail to do so, I hope that these

23hearings will, because without it, the business people

24are going to be left largely rudderless.

25MS. SCHULTHEISS: Thank you, Janet.

Page 161

1(Applause.)

2MS. SCHULTHEISS: I would like to thank all of

3our panelists, and we are going to take a brief

4ten-minute break, after which we will reconvene and have

5a directed discussion with some questions. Thank you.

6(A brief recess was taken.)

7MS. SCHULTHEISS: Okay, let's get started with

8the second part of our panel, which is the discussion

9period, and before I start asking some specific

10questions of the panelists, I would like to see if I can

11get some consensus from the panelists on some of the

12terms we are using and whether we are understanding them

13correctly.

14As I am hearing the presentations, I hear that

15you are talking about two different things. You are

16talking about predatory conduct that affects the input

17suppliers and the price of the input supplies, and

18another area is the predation or predatory conduct that

19is directed at affecting your competitors in the output

20market. Is that correct?

21Let's just start with Jack and let's just go

22through the panelists and see what you think. Is that a

23correct statement?

24MR. KIRKWOOD: I am not sure I would put it

25exactly that way. I think that Steve's distinction is a

Page 162

1good one, that if you as the dominant firm could raise

2the input costs of a rival, and the question is whether

3your ultimate goal is monopsony power or, instead,

4downstream market power, and the first I have called

5predatory bidding, as did the Ninth Circuit, Steve has

6called it predatory overbuying. The second category, it

7could come from predatory bidding, as I explained, but

8the second category is more typically described as

9raising rivals' costs, or as Tim would say, exclusion,

10and maybe Tim could address the question of to what

11extent is exclusion different from raising rivals'

12costs.

13MS. SCHULTHEISS: Tim?

14DR. BRENNAN: Well, I agree with what Jack said.

15Again, I agree with the distinction as he put it or, you

16know, that Steve made. I would not call both predatory,

17because as soon as you do that, then you start using

18predatory tests in both contexts, and I think for

19reasons that I have argued and Steve has suggested as

20well and others, that I think those are inappropriate.

21The main reason I make the distinction, I mean,

22I suppose there is a bit of a long story here, but just

23to keep it short, the two are -- well, first, when you

24say that the harm from something involves hurting

25rivals, then I think you invite people to go back into

Page 163

1predation land, and I would rather call it something

2else, and the reason I would call it the something else,

3when I am calling it complement market monopolization,

4is that in order to exclude people, in order to raise

5their costs, whatever it is, you have to create and

6exercise market power over something that they need, and

7because antitrust authorities have ways of thinking

8about that sort of thing, that is what we do with

9mergers, that is what we do with cartels, you could in

10some sense take an enormous part of -- as Steve, I

11think, aptly put it -- the great lion's share of Section

122 cases and get them out from under the controversy that

13attaches to them because of the view that what Section 2

14is about is in some sense about competing so hard that

15people get hurt too much.

16MS. SCHULTHEISS: Steve, would you like to

17address that?

18DR. SALOP: Yes, sure, I agree with Jack that I

19got it right.

20DR. BRENNAN: I figured it was relatively safe.

21DR. SALOP: I agree with the first two sentences

22of what Tim said. I think when you use the term

23"predatory," it is a loaded term. I never really liked

24the term "nonprice predation" for that reason. That is

25why I like calling it exclusion or raising rivals'

Page 164

1costs.

2Now, nonprice predation sometimes involves

3prices, like here, prices are involved, and when you

4call it predation, that tends to plug into Brooke Group,

5and as I said, I think there are two vastly different

6paradigms.

7MS. SCHULTHEISS: So, in terms of what we are

8speaking about here today, Rick focused on the narrow

9with the predatory overbidding or overbuying that is in

10the Weyerhaeuser case, but then there is a separate area

11of conduct that involves some type of exclusionary

12conduct, what you call raising rivals' costs, what Tim

13might call exclusion, and looking at the complementary

14market.

15DR. SALOP: I think what all three of us were

16taking issue with in your question was the word

17"predatory." Had you just said "conduct that gives

18market power in the input market" and "conduct that

19gives market power in the output market," we all would

20have -

21MS. SCHULTHEISS: Do you all agree with that

22one?

23DR. BRENNAN: More or less.

24MS. SCHULTHEISS: What about you, Rick?

25DR. WARREN-BOULTON: First of all, yes, I think

Page 165

1they are enormously different. I think that the first

2is monopsony and the second is monopoly, and I think it

3is incumbent upon a plaintiff to say which one -- which

4world he's in. I agree that the false positive/false

5negative trade-off is enormously different between the

6two. I don't see how people can really -- or people

7should not confuse them. I don't think they are

8particularly relevant. I think all three of those

9situations should be handled, you know, quite

10separately, because they have different structure

11requirements.

12The final question, would I distinguish them in

13terms of predation/nonpredation, no, I would distinguish

14them I think in terms of, you know, simple pricing

15versus, you know, more complex exclusion, you know,

16price or nonprice. I mean, most of the really

17interesting problems here really are exclusion done in

18interesting and complicated and strategic ways and how

19do we handle that. But again, I think at the very

20beginning, it is really incumbent on people to say which

21box they are in in any case.

22MS. SCHULTHEISS: Janet?

23MS. McDAVID: I think that Steve and Jack got it

24right from my perspective.

25MS. SCHULTHEISS: Okay. In terms of the

Page 166

1antitrust agencies, would you agree, then, that we

2should be more concerned with the latter category, that

3being, if you want to call it, either the exclusionary

4conduct or raising rivals' costs, versus the pure

5predatory buying towards, you know, monopsony over the

6input market? And I would wonder -- I want to find out

7if you agree with that.

8MR. KIRKWOOD: I would be inclined to agree with

9it, yes. I have not done the kind of research that

10Steve has done into the frequency of raising rivals'

11costs problems, but from what I have read of what he and

12others have written, it seems to be a more common

13problem. One of the key elements of the analysis of

14predatory bidding we have done is that it seems to be

15quite rare. So, yes, in terms of where you would target

16your enforcement resources, sure, surely.

17MS. SCHULTHEISS: Tim?

18DR. BRENNAN: Yes, mostly for the reasons Steve

19said before, that the predation parts of things, whether

20it is buying or selling, are going to be rare for the

21reasons he outlined and I do not need to repeat, and so

22for that reason my expectation would be to worry about

23things where people do not have to sacrifice a lot and

24take a lot of risks in order to get some speculative

25benefit down the road.

Page 167

1MS. SCHULTHEISS: Steve, I take it you agree.

2DR. SALOP: Yes, I agree with what they said,

3but they didn't answer your question, which is, well,

4what the agency should not be doing, and frankly -

5MS. SCHULTHEISS: I do not know if it is

6necessarily what we should or should not, but our

7primary focus. Should we be worried about one more than

8the other?

9DR. SALOP: The way you phrased it is something

10that I think is all too apparent in this administration,

11which is you are spending an awful lot of time deciding

12what you should not do, and you are not doing a heck of

13a lot, and so I think that in the exclusionary conduct

14area, the agencies ought to get involved and start

15looking for exclusionary conduct cases rather than

16protecting monopolists.

17MS. SCHULTHEISS: Okay.

18DR. WARREN-BOULTON: I think I am probably

19agreeing with everybody. First of all, I agree that -

20I mean, the whole point is that, you know, predatory

21overbidding to get monopsony power, as I said, is

22incredibly rare. So, the answer to the first question

23is yes, and oddly enough, I would agree with Steve that

24I think exclusion is a real problem, and so if you are

25going to spend more money on something, I would have you

Page 168

1spend it on exclusion.

2MS. SCHULTHEISS: Janet?

3MS. McDAVID: I would urge the agencies actually

4to play a role in the Section 2 enforcement area

5generally, because too often these cases arise in the

6context of disputes between business rivals, each of

7whom brings baggage, whereas the agencies are trying to

8do the right thing, and a case that an agency brings has

9been vetted carefully as opposed to just being the

10pissing contest between somebody who may have been

11forced out of business and thinks that, of course, it

12was unfair and something must have been wrong, it

13couldn't have been an inefficient firm, it couldn't have

14been incompetent. The agencies, I think, bring an

15important balancing rule to the enforcement in the

16Section 2 area generally and in this area as well.

17MS. SCHULTHEISS: Let me ask you another

18question, Janet, because you were talking about in terms

19of counseling clients and what have you. Would you

20agree, then, with Rick and most of the other panelists,

21I think, that the overbidding or overbuying that we see

22in the Weyerhaeuser case is an unusual situation?

23MS. McDAVID: I have not seen a lot of this. I

24think it probably is an unusual situation, but the

25plaintiff's bar is extremely entrepreneurial, and I do

Page 169

1not think we should assume that it will continue to be

2an unusual circumstance.

3MS. SCHULTHEISS: So, you -

4MS. McDAVID: And virtually every major firm in

5this country sues in the Ninth Circuit. So, unless the

6law gets clarified, I think we will see these cases

7because they will get syndicated.

8MS. SCHULTHEISS: Steve, did you want to say

9something in response to that?

10DR. SALOP: Yes, I just wanted to -- I just

11jotted down some overbuying cases before, and so I know

12we see sort of the classic, you know, Ross-Simmons cases

13being rare, but one, as I said before, Ross-Simmons

14alleged raising rivals' costs as well as -- as well

15as -

16MS. SCHULTHEISS: Right.

17DR. SALOP: -- predatory bidding, and in terms

18of the history of antitrust, you know, among sort of the

19cases I teach and run across, we have got not just the

20timber cases, we have got the tobacco case was an

21overbuying case, the beef case, the Monfort case that -

22I mean, the beef case was a conspiracy in overbuying,

23the Monfort case was about overbuying, went to the

24Supreme Court, Socony-Vacuum was about overbuying, Alcoa

25had a piece that was involved in overbuying, and in

Page 170

1terms of what is going on right now, stuff the agency

2has in its shop, all the stuff about shelf space is all

3about overbuying, too, you know, could be thought of as

4an overbuying issue. There could be overbuying of

5patents. Alcoa, you know, allegedly tied up some

6patents, so it made it harder for people in -- not sort

7of classic predatory bidding, but, you know, you could

8certainly have overbuying of patents, too.

9So, I do not think it is a -- I think this idea

10that it is just exhaustible resources that are

11by-products, I know that covers the old molybdenum case,

12one of my favorite cases when I was at the FTC, but I

13think it is rather broader than that.

14DR. WARREN-BOULTON: Can I just make a -- I

15mean, I have not gone back and looked at the cases, but

16I would question how many cases are there -

17MS. SCHULTHEISS: Can you speak into the

18microphone?

19DR. WARREN-BOULTON: I'm sorry, how many cases

20are there in overbidding to gain monopsony power. Steve

21cites a tobacco case. I do not think that was to gain

22monopsony power for tobacco growers. I thought that was

23to exclude other, you know, cigarette producers. The

24beef case, I do not think they ever managed to -- it was

25ever shown that they ever managed to raise the price of

Page 171

1anything. The others I cannot remember, but, I mean, I

2guess the question is, are there any cases other than

3Reid, which was I don't know how long ago, and

4Ross-Simmons that are predatory overbidding to achieve

5monopsony power? I mean, I do not want to say sui

6generis, because I do not know how to pronounce it

7correctly, but it has got to be something like that.

8Sui?

9MS. McDAVID: Sui.

10DR. WARREN-BOULTON: Sue me, sue you.

11MS. SCHULTHEISS: All right, Jack would like to

12say something.

13MR. KIRKWOOD: Yes, Steve is certainly right

14that there have been a number of overbuying allegations,

15but if you look narrowly at how many of these cases are

16predatory bidding cases, American Tobacco, as Rick and I

17discussed last night, that seems fairly clearly a

18raising rivals' costs case as opposed to a predatory

19bidding case, because the kind of tobacco the major

20producers bid up was a tobacco they did not use at all

21and continued not to use it. So, they were not trying

22to get monopsony power in it, because they were not

23using it as an input.

24And in beef, though there was a predatory

25bidding allegation, as Rick suggests, they were not able

Page 172

1to raise the price at all. And in Socony, that was

2behavior designed to facilitate a cartel. Monfort had a

3predatory bidding concern, but it was a concern raised

4as part of an effort to enjoin a merger. So, there was

5not any actual evidence of predatory bidding, and the

6Court analyzed it as a predatory pricing case.

7MS. SCHULTHEISS: Steve, and then I want to get

8Janet's reaction.

9DR. SALOP: Of course, the Supreme Court

10misanalyzed Monfort, but putting that aside, I mean, I

11agree, a lot of these are not predatory overbuying

12cases, they are other things, but the Ross-Simmons jury

13instruction does not require you to show anything other

14than they paid more than necessary or bought more than

15necessary or paid more than a fair price. So, all of

16them would be swept in if we go inductively and just

17allow the Ross-Simmons jury instruction to stand and

18screw things up for another 15 years until Ken has to

19get involved and clean up the area.

20MS. McDAVID: That is precisely my concern, is

21that if this is the law in the Ninth Circuit, virtually

22anything can be brought as one of these cases, and they

23probably will not be tried to judgment in the end,

24because the defendants cannot take the risk of trying

25those cases to judgment with the treble damages burden.

Page 173

1MS. SCHULTHEISS: I am going to let Tim have the

2last word on this particular point, and then I would

3like to move on.

4DR. BRENNAN: I think that the reasons that Rick

5in particular elaborated on why these cases would be

6rare in terms of the circumstances of the market,

7exhaustible resources, the bidding market and stuff, I

8think those are all important to keep in mind. I do not

9think, though -- one should be careful about making the

10distinction, one, between buying and selling, because

11that can be just an arbitrary function on the nature of

12the market.

13Let me give you an example, which is pipelines.

14I don't know as much -- I don't know as much now about

15how the pipeline sector works as I used to, but

16pipelines are kind of funny in the following sense: The

17way oil pipelines worked is that they sold oil pipeline

18delivery services to oil wells basically. The way the

19gas pipeline industry worked was that they bought gas

20from gas wells and then resold it at the end of the

21pipeline. If the oil pipeline was exercising market

22power against the oil wells, it would involve monopoly,

23that they were raising the price of pipeline services.

24If they were -- if the gas pipeline was exercising

25market power against the gas wells, it would be

Page 174

1monopsony. They would be driving down the price they

2paid for the natural gas.

3The triangles, the welfare losses, the dead

4weight stuff, all that stuff would be exactly the same

5in both cases, at least qualitatively it would be the

6same, but one's monopoly and one's monopsony, and so

7it's purely in some sense a -- I don't know, I won't say

8it's arbitrary or artificial, but it is the way those

9markets just happen to work. So, one needs to be

10careful about making a distinction in that way, that one

11can turn monopoly into monopsony in some sectors without

12a great deal of difficulty.

13DR. WARREN-BOULTON: Yes, I -

14MS. SCHULTHEISS: Rick is going to have to get

15the last word I see.

16DR. WARREN-BOULTON: -- I very much agree. You

17can be regarded as buying a service or selling, and it

18can be based on form. I think one implication that I

19think is kind of interesting is that in terms of a

20welfare loss, you know, the idea that somehow welfare

21losses are only important if they happen downstream

22rather than upstream strikes me as kind of bizarre.

23MS. SCHULTHEISS: Let's get to that in a second,

24because I want to get to that issue.

25DR. WARREN-BOULTON: But that is one of the

Page 175

1reasons why, because we can characterize it as upstream

2or downstream sort of at will.

3MS. SCHULTHEISS: Okay. Does everybody agree

4that the jury instruction in Weyerhaeuser was just

5wrong?

6MS. McDAVID: Yes.

7MS. SCHULTHEISS: Is there anybody that

8disagrees with that?

9DR. WARREN-BOULTON: You know, I am an

10economist, I am not an attorney. If I was giving a set

11of jury instructions to a group of economists, it would

12be quite different from those instructions.

13MS. McDAVID: How about if they were grocers and

14guys who pump gas, because that is who your jury is made

15up of.

16DR. WARREN-BOULTON: So if I had a set of

17instructions that said something like if you were a

18group of economists and I wanted to have a recoupment,

19how would I say that in a way that a group of grocers

20would understand? You know, maybe this is the best way

21the Court could think of to explain it to a group of

22grocers. I'm not sure it would have been a better

23result if we would have explained it in sort of

24mathematical formulas.

25MS. McDAVID: It probably would have been worse.

Page 176

1DR. WARREN-BOULTON: The question is, what does

2the audience grasp?

3DR. SALOP: You think saying they bought more

4than necessary is a good proxy for recoupment? Is that

5what you just said?

6DR. WARREN-BOULTON: No, paying more than

7necessary, it seems to me to be to a -- I don't know, an

8ordinary person's idea of was there a profit sacrifice

9here, all right, which is at least one element.

10MS. SCHULTHEISS: But in terms of precedent, how

11can that possibly -- you are just agreeing with the

12statement I just said, then. You would not have that

13kind of an objection to the standard that was set out in

14the jury instruction.

15DR. WARREN-BOULTON: You know, I am saying

16anybody here, I think anybody here could probably today

17come up with a better set of jury instructions than

18that. How bad it is, I do not know.

19MS. SCHULTHEISS: I do want -

20DR. SALOP: Do you think a reasonable jury could

21reach the right conclusion if they tried to apply that

22jury instruction?

23DR. WARREN-BOULTON: You know, actually, I have

24been on a jury, and I have to tell you that my personal

25experience is that what a jury decides has little, if

Page 177

1nothing, to do with the jury instructions. I was on a

2cocaine bust jury in the District of Columbia, and we

3convicted him of carrying a machete. We did that

4because two people in the room said under no conditions

5would I convict a male in the District of Columbia

6carrying cocaine. We decided that he should go to jail

7for three months, and we cast around to try to think of

8what kind of thing would get him to jail for three

9months. So, we convicted him of carrying a machete,

10so...

11MS. SCHULTHEISS: All right, I get the point.

12There have been some different standards put

13out, and I would like to get some input, because it

14appears -- I mean, we have talked about consumer

15welfare, we have talked about total welfare, no economic

16sense, and they have come into play in different ways,

17and I would like to start with you, Janet.

18Which of the standards do you think makes most

19sense in this type of a case, in the case where you are

20dealing with either overbidding or some kind of

21overbuying situation?

22MS. McDAVID: Well, I am taken with Steve's

23standard, which is effectively a rule of reason analysis

24applied to the sort of circumstances we have here and

25which incorporates as part of the analysis the Brooke

Page 178

1Group test, because I think it gives us some grounding.

2MS. SCHULTHEISS: In dealing with the Brooke

3Group test, though, then, do you have to have power then

4in the output market in order to get that recoupment?

5MS. McDAVID: You have to be able to make it up

6somewhere, and exactly where you are going to make it up

7might vary based on the particular circumstance and as

8to whether it's a raising rivals' cost case or a

9predation sort of case.

10MS. SCHULTHEISS: But if you are recouping

11solely on the input side, on the supply-side, then you

12really do not necessarily have the consumer welfare

13harm.

14MS. McDAVID: You may not.

15MS. SCHULTHEISS: And in that case, would you

16find that there would not be a violation or there should

17not be an antitrust violation found?

18MS. McDAVID: I am actually not terribly

19troubled by also applying the antitrust laws and

20allowing the victim to be the supplier, just as we do in

21the cases Gail used to investigate when she headed the

22health care shop where we looked at large insurance

23companies and whether they were going to acquire

24sufficient power over doctors and hospitals. I think it

25is a reasonable inquiry.

Page 179

1MS. SCHULTHEISS: So, the consumer welfare test

2should not necessarily be the end all and be all, okay.

3MS. McDAVID: Not the exclusive. It should be

4the major concern, but it would not exclude entirely

5concern to suppliers.

6MS. SCHULTHEISS: Okay, Rick?

7DR. WARREN-BOULTON: Yes, I just categorically

8disagree strenuously. I mean, I just do not see any

9difference between a dollar in producer surplus that

10goes to some guy who is growing timber and a dollar to a

11guy who is buying a unit of lumber, nor do the Merger

12Guidelines. I mean, I think the economics literature is

13absolutely unambiguous on this. The welfare losses from

14monopsony are, you know, the same as the welfare losses

15from monopoly, consumer surplus.

16I mean, I think Roger Noll wrote a very nice

17article right before everybody else's article that

18basically walked through that, you know, and as Tim

19points out, is that, you know, trying to find some

20distinction as to whether or not it is upstream versus

21downstream is completely arbitrary. It certainly cannot

22be a distinction that says, "Gosh, if it is a loss by

23producer, it does not count."

24In that case, what you say is all input

25monopolies are legal? I mean, you do not care if it is

Page 180

1an input monopoly because, you know, the thing is bought

2by a firm who does not necessarily pass it on? I mean,

3the idea of restricting this to consumer surplus to me

4is bizarre. I think basically you have got an

5externality test, basically says I have got the bad guy

6who is doing something, and the question is, is he

7hurting the rest of society? And the rest of society

8are the people he sells to and the people he buys from,

9and those people stand, you know, on the same footing.

10So, you know, I think what we used to call a

11consumer welfare test when we were trying to read the

12names back in the good old days is the sum of consumer

13surplus and producer surplus. I think what most of us

14say is you ignore the profits of the monopolist, but

15that is it. Everybody else is in.

16MS. SCHULTHEISS: Steve, is that what you mean

17when you say consumer welfare or consumer harm?

18DR. SALOP: No, I thought my slides were very

19clear on that, and I have got actually a paper that I

20submitted to the Antitrust Modernization Commission on

21this issue. I think by consumer welfare I mean true

22consumer welfare. I think that people who want to say

23suppliers, losses to supplier welfare should be enough,

24should consider whether they think harm to competitors

25should be enough to carry an antitrust violation, and

Page 181

1the Supreme Court has made it very clear that harm to

2competitors is not enough, and I think the same thing

3should be true with harm to suppliers.

4The tricky part of this -- and I am sorry, this

5is kind of a long answer -- the tricky part is that it

6is quite clear to make an agreement -

7MS. SCHULTHEISS: Could you speak more into the

8mic? I'm sorry.

9DR. SALOP: I'm sorry.

10It's quite clear that naked agreements among

11competitors, buyer-side competitors, to fix prices that

12they make to inputs is illegal and should be per se

13illegal, and -

14MS. SCHULTHEISS: Well, that is horizontal,

15though.

16DR. SALOP: But that is where antitrust starts.

17MS. SCHULTHEISS: Right.

18DR. SALOP: So, I think it is worth thinking

19about that case, because a lot of people that I have

20talked to when I say it should be about consumer harm,

21not about supplier harm, they pretty quickly think about

22the buyer cartel cases, and the way I would distinguish

23is even in the buyer cartel side in the following way:

24Suppose you have an agreement among competitors

25to jointly set -- I will not use the loaded term

Page 182

1"fix" -- to jointly set the price that they pay for

2inputs. I think if that is naked, it is quite clear

3that is per se illegal and should be per se illegal, but

4to the extent that they have a justification, a

5procompetitive, i.e., pro-consumer welfare justification

6for that, then it ought to -- and I think it does -- go

7into the rule of reason, just like in VMI.

8If a group of sellers jointly sets a price, it

9goes into a rule of reason if they have a procompetitive

10justification for their actions, and then when you get

11into the rule of reason, I think it makes sense that

12consumer welfare rules, that it is not enough -- once

13they show that consumer benefit, then the burden would

14go to the plaintiff to prove that consumers are harmed.

15It is not enough to show suppliers are harmed, and it is

16not, I do not think, a balancing between the losses to

17the input suppliers versus the effect on consumers and

18versus the gains to the buyers that engage in the joint

19price setting.

20So, you know, I think antitrust is a consumer

21welfare prescription, it is about consumer welfare, and

22we should stick with that. We should stick with that

23here, particularly in a situation where -- you know,

24with this overbuying, where in a standard case,

25consumers gain in the short run from it, just as they

Page 183

1gain in the short run during the predatory period from

2predatory pricing.

3MS. SCHULTHEISS: Okay, Tim?

4DR. BRENNAN: A few very quick observations.

5First, I just happened to go to the AMC deliberations on

6mergers last week, and the only thing -- almost the only

7thing they argued about was the welfare standard, and

8the only arguments in favor of the consumer welfare

9standard that were given were essentially critical

10rhetoric, that the basis of public support for antitrust

11is if people believe it is about consumers rather than

12about the economy as a whole, but there was not a

13substantive argument offered in its favor. So, I do go

14with that.

15As far as the harm to competitors being unduly

16counted goes, I suspect that in these cases, that is

17balanced out by profits to the perpetrator or gains to

18the consumers or someplace, that that is all going to be

19just a transfer, and you still end up with the assorted

20dead weight losses versus efficiencies that we are

21familiar with.

22As far as what sort of tests to use, more

23specifically, on what I would call the exclusion cases,

24I basically view those as essentially horizontal, taking

25up an ever larger share of this complement market

Page 184

1through exclusive dealing contracts or whatever it might

2be and that we have horizontal tools for looking at

3that. So, that is what I would use there.

4On the predatory buying or predation cases

5generally, I do not know enough to know at what point

6one hits that balance between type I and type II error,

7but that to me is what it is about, I think rather than

8attempting to get each case exactly right, and I am just

9going to leave it at that.

10MS. SCHULTHEISS: I am going to let Steve make a

11quick response and then Jack.

12DR. SALOP: I am not surprised that the AMC was

13confused during the hearings on the welfare standard.

14Rick Rule testified that Bork used the Williamson

15Diagram, he called it consumer welfare, as Rick said,

16that was the old days when we were trying to confuse

17people. The Supreme Court cited the Bork book.

18Therefore, what the Supreme Court meant by consumer

19welfare is total welfare, and that is the law, you know,

20so I can certainly see why the AMC would get confused.

21With respect to what the law ought to be, you

22know, I think it is a complicated argument, but to Tim I

23would say the following: It is not just a transfer.

24Suppose you have entry into an industry by a relatively

25high-cost entrant, enters the market, prices begin to

Page 185

1come down, benefiting consumers, and then the

2monopolist, which has got much lower costs, kills the

3entrant, prices go back up, suppose demand is relatively

4inelastic? Well, in that situation, the killing of the

5entrant would raise total welfare, and I thought you

6said in your statement that standards like that were no

7good, but regardless of what you said, it is quite clear

8that that is conduct that I think ought to be illegal.

9MS. SCHULTHEISS: I want to get to Jack now.

10MR. KIRKWOOD: Sure, sure. We are looking at

11various welfare standards now, which is inevitably a

12somewhat complex topic. One choice is between total

13welfare and what Rick called third-party welfare, so let

14me just talk about total welfare versus consumer

15welfare.

16That is ultimately a value choice. You can

17think of the famous Williamsonian case where the merger

18lowers cost but raises price. My value judgment is that

19antitrust ought to stop that. There are arguments pro

20and con, but the legislative history seems to reflect

21that judgment, and every court that has ever faced that

22issue has come out the same way.

23On the more difficult and more judgmental choice

24of supplier welfare versus consumer welfare, I agree

25with much of what Steve said and originally wrote my

Page 186

1article using a consumer welfare test, even in monopsony

2cases, because generally there is a link between the

3adverse impact on suppliers and the adverse impact on

4consumers, but people raised two issues with me.

5One, what about the case law? The case law

6generally favors supplier welfare in a monopsony or

7cartel case, and two, what about those instances, of

8which Steve has described, where there is an adverse

9impact on suppliers, suppliers are exploited, just like

10in that merger, prices to them are lower, but no impact

11whatsoever on consumers? Does that allow the practice?

12And it does not seem it should.

13DR. WARREN-BOULTON: Can I just make an example,

14and I am responding really to Steve, to his statement

15that somehow once we start talking about producer

16welfare, we have to take into account the welfare of

17competitors, and I do not think that is true. Let me

18give you an example.

19Suppose that I am a Kansas wheat farmer, and

20what I do is I burn down all my neighbor's fields in the

21county. Now, do we have a harm here? Yes. The

22question is, is it an antitrust harm?

23Now, you know, it is unlikely to me that I am

24going to burn down all my neighbor's wheat farms because

25I think that as a result, you know, there will be a

Page 187

1shortage of wheat and I will be able to raise the price

2of wheat. So, I do not have any consumer harm at all.

3Let's suppose that there is no input problem at all.

4Then I typically have harm. I do not have an antitrust

5harm. I still presumably go to jail for arson, right?

6So, I have harm, but it is not an antitrust harm.

7Antitrust harm gets triggered when I have a

8market impact, and it could be one of two things. It

9could be that I really burn down enough wheat to

10actually raise the Chicago price of wheat, unlikely, but

11if somebody came to me and said, you know, what the guy

12did is he burned down all his neighbor's wheat farms,

13and why did he do it? He said because there is a local

14labor market for workers, and if I burn down and put out

15of business all my local -- you know, who are with me in

16the local labor market, I will be able to reduce the

17prices I have to pay my workers.

18Now, that is monopsony. Now, do I think that is

19an antitrust violation? Yes, I would say that that is

20an antitrust violation. I do not care whether it is

21raising the price of wheat or reducing the wage rate of

22farmers, nor do I think, in particular, that somehow

23that one, you know, on some ethical standard, they are

24any different. In fact, wheat bread consumers are

25probably richer than farm laborers. So, there is some

Page 188

1distinction. I do not think you have to say that

2somehow by bringing in producer surplus, we are somehow

3worrying about competitors.

4DR. SALOP: Well, you said something very

5different. You said you would be willing to find an

6antitrust violation where there is no consumer harm.

7That is different from saying that you are adopting the

8total welfare standard as the overarching standard to

9govern antitrust cases, because if you were adopting the

10total welfare standard to govern antitrust cases, then

11the simple business tort of burning down your neighbor's

12fields would lower total welfare, and it could support

13an antitrust violation.

14Certainly if you and your neighbor burned down

15everybody else's fields, so you would not get into all

16this complexity -

17MS. SCHULTHEISS: Let's deal with a single firm,

18which is what we are really focusing on.

19DR. SALOP: Actually, we are not. We are

20focusing on what the welfare standard should be.

21MS. SCHULTHEISS: Right, but in connection with

22Section 2.

23DR. SALOP: Well, why would you have a different

24welfare standard for Section 1 and Section 2? Why would

25you want to make -- why would you want to gerrymander

Page 189

1antitrust and make it incoherent? You know, Tim pointed

2out this notion of the pipeline, you know, if you have

3different rules governing how -- if the rules are very

4different according to whether the pipeline does a

5tolling agreement where they sell services or whether

6they buy the gas and resell it at the other end, you are

7going to have incoherent antitrust. If you are going to

8have dramatically different rules for tying, exclusive

9dealing, vertical mergers, where one's per se illegal,

10one's rule of reason, and one's virtually per se legal,

11well, since lawyers can characterize conduct pretty

12easily as any one of those three boxes, you are going to

13end up with very incoherent antitrust, which we did

14until the eighties. So, I think saying this is Section

152, not Section 1, that is a recipe for disaster.

16MS. SCHULTHEISS: Janet, did you want to respond

17to that?

18MS. McDAVID: No.

19MS. SCHULTHEISS: Okay. Rick?

20DR. WARREN-BOULTON: Well, the broad question

21is, gee, do we have to have the same rules in every

22situation? My answer is no. If you can distinguish

23between situations, then you can have the same rules -

24you can have different rules. There is no -

25MS. SCHULTHEISS: But would the test be the

Page 190

1same?

2DR. WARREN-BOULTON: Sure. I mean, I think the

3test for naked price-fixing is different than the

4test -- you know, for the safe harbor for naked

5price-fixing is very small. I think the safe harbor

6for, you know, predatory pricing should be quite large.

7MS. SCHULTHEISS: No, I mean in terms of the

8consumer welfare versus total welfare, that test should

9be the same across all of the violations or different?

10Steve is arguing I think that it should be the same

11welfare test regardless of the violation.

12DR. WARREN-BOULTON: Oh, I see, yes. You can

13argue what is the best welfare, but somehow suppliers

14are upstream or downstream, welfare does not change

15depending on -

16MS. SCHULTHEISS: So, you agree with Steve,

17then, that whatever welfare test -- whatever welfare

18test is chosen should apply across all of the various

19violations you're looking at?

20DR. WARREN-BOULTON: As long as he chooses my

21welfare test, then -

22MS. SCHULTHEISS: Janet, would you agree with

23that? Yes, no?

24MS. McDAVID: I think we should probably move

25on.

Page 191

1MS. SCHULTHEISS: Well, I think the test is

2important, though, and I think whether you apply the

3same test or not is an important issue.

4DR. WARREN-BOULTON: It has a huge effect on the

5case in question.

6MS. SCHULTHEISS: And I really would want to

7know whether you think we should be using a consumer

8welfare test or a total welfare test regardless of the

9type of violation you are looking at.

10MS. McDAVID: Well, I think I started by saying

11much as Rick did, that the suppliers are entitled to a

12competitive market just as buyers are, and so I would

13not exclude a remedy for suppliers by using a test that

14focused entirely on consumers.

15MS. SCHULTHEISS: Exclusively.

16DR. SALOP: Suppose two firms get together and

17exchange a technology, a labor-saving technology, that

18enables them to produce the same amount of output with

19less labor, and as a result of that agreement, the firms

20demand less labor and the wage rate goes down or some

21other input, if you will, and suppliers -- here the

22suppliers of labor are harmed. Do they have standing to

23bring an antitrust case against that agreement?

24DR. BRENNAN: No, because total welfare went up.

25MS. SCHULTHEISS: Jack?

Page 192

1DR. WARREN-BOULTON: Total welfare -- I mean -

2MS. SCHULTHEISS: Let's do this in order. Jack

3looked like he wanted to respond. Let me let Jack

4respond first.

5MR. KIRKWOOD: Yes, that was a point I didn't

6comment on before, but I do agree with Steve there. If

7you do have a case in which there is supplier harm but

8consumer benefit, then I would go with the consumer

9welfare standard. I think that does make antitrust more

10coherent.

11DR. BRENNAN: Total welfare went up in that

12standard, so I would stick with it.

13DR. SALOP: Yes. Now, how do you know that

14total welfare went up? Did you do that calculation?

15DR. BRENNAN: It is a -- I guess it is a

16presumption for me, kind of 101, yes.

17MS. SCHULTHEISS: I mean, I am going to give

18Rick and Janet one more chance, and then we will get off

19this.

20DR. WARREN-BOULTON: How about we get a group of

21consumers that get together a group of monopsonized

22consumers, is this somehow a good thing? I don't think

23so.

24DR. SALOP: That is the cartel case decided by

25Judge Breyer at the time, and he said it was okay.

Page 193

1DR. WARREN-BOULTON: Well -

2MS. McDAVID: I think in this -- in the

3circumstance that Steve has posited, we have got an

4integrated joint venture that will be evaluated under

5the rule of reason, there is an efficiency and a

6business rationale, and under the rule of reason, you

7probably do not conclude it is illegal.

8MS. SCHULTHEISS: Well, I want to get to

9something else that is completely off of this topic for

10a few minutes, because I want to make sure we have time

11for it, and that is related to relief and remedies.

12How do you deal with relief and remedies and in

13particular in the overbidding situation? Should a court

14enjoin the defendant's pricing? How do you deal with

15that issue in the context of this type of conduct?

16MS. McDAVID: Well, in the private party

17litigation, it is going to be damages.

18MS. SCHULTHEISS: I mean in a government case.

19MS. McDAVID: In a government case, probably -

20well, then you need to have a standard, and a standard

21of fairness does not allow a remedy.

22MS. SCHULTHEISS: But even if you have the

23Brooke Group standard, what is the Government's remedy

24in that situation on the predatory bidding, you know,

25raise your prices or lower your prices? I mean -

Page 194

1MS. McDAVID: It is pretty unlikely to me that

2this is going to be a government case, although I think

3there is a role for the Government to play in Section 2

4enforcement, an important role. These cases are going

5to come up in disputes among rivals.

6MS. SCHULTHEISS: So, you are thinking the

7overbidding context, it is not the type of case that the

8Government should be getting into?

9MS. McDAVID: It is the type of case that I

10doubt the Government will get into. It is the type -

11the Government typically intervenes in cases where there

12is kind of a broader principle to be generated. This is

13typically an intrafirm dispute or interfirm dispute, and

14they rarely get into those circumstances.

15MS. SCHULTHEISS: Well, when we are looking at

16what test to apply, though, you know, for example, you

17know, we did file an amicus brief in this case,

18obviously concerned with the test and the jury

19instructions from the Ninth Circuit, if it had been a

20government case, and I mean this sincerely, what kind of

21relief could the Government get in any kind of an

22overbuying monopsony type case like this? Does anybody

23have any -

24DR. WARREN-BOULTON: I am about to agree with

25Janet. I think there are situations in which we rely

Page 195

1primarily on deterrents -- there are situations in which

2deterrents, ex post penalties, are simply much, much

3less expensive than ex ante penalties. I mean, you

4think of how would you have hypothetically solved

5Weyerhaeuser. Well, I suppose to me the answer is you

6deter basically through either private litigation or a

7fine. How could you prevent this situation? Would you

8have to divest? You would have to sort of break up

9Weyerhaeuser.

10It seems to me that facing an alternative,

11looking at the costs of preventing bad behavior through

12structural means, that is to say, you know, forcing

13Weyerhaeuser to sell off sawmills versus simply having

14an ex post, you know, you will face damages if you, you

15know, behave badly. A behavioral remedy is probably

16better than a structural remedy, and if it is

17behavioral, it seems to me I agree with Janet, it is

18really something -- it is for private litigation.

19Steve?

20MS. SCHULTHEISS: Would you agree with that,

21DR. SALOP: No, I thought -- I think there is an

22answer. First of all, what the Government usually does

23is it gets an injunction. You tell them not to violate

24the standard anymore, and once you know what your

25standard is -

Page 196

1MS. SCHULTHEISS: We have to have a standard,

2yes.

3DR. SALOP: -- which hopefully you will know

4once you bring the case, then you tell them, don't

5violate the standard.

6Secondly, I am not sure why Rick was so negative

7with respect to structural relief. If what Weyerhaeuser

8did was it knocked its rivals out of business and

9thereby got a monopsony, then the way to jump-start

10buy-side competition is to make Weyerhaeuser divest some

11of its mills to re-establish that competition, or

12perhaps if Weyerhaeuser simply has one big mill, you

13would make Weyerhaeuser subsidize the entry of the small

14sawmills that it knocked out of business. So, there is

15a potential structural remedy there, or, of course,

16maybe this would be one of the places where the FTC

17should seek disgorgement.

18MS. McDAVID: You see, that is precisely the

19remedy for private damages. That's what private damages

20are designed to achieve, is what Steve has just

21described.

22DR. WARREN-BOULTON: I agree.

23DR. BRENNAN: I agree with Steve, and I think

24Ken said it earlier today, about if you have a

25structural remedy, that that would be a better thing to

Page 197

1do. I think the instance was slots, you know, making

2that more competitive or something for airports.

3The one thing that came up in the morning

4discussion and also here that I honestly do not

5understand is that somehow that seeking damages is

6different and that this question only comes up with the

7Government seeking injunctive relief, because if you

8have a world where damages are collectible, then people

9out there have to know what to do to avoid having to pay

10damages, and the threat of damages in predation cases

11is, I think, going to involve some kind of, in effect,

12price regulation that says, you know, if you are bidding

13now, if you set your price -- if you pay -- and that's

14in essence the whole problem with the jury instruction,

15right, that if you pay this much, it's okay, but if you

16go beyond that, then you are paying too much, and then

17you are liable to damages, and it is the threat of

18damages that basically you are going to force people to

19say, how high a price can I pay and I can't go above

20that.

21So, there is something I think inevitable about

22this kind of law apart from the Government's specific

23remedies that says that some prices are okay and some

24prices aren't okay.

25DR. SALOP: But that is not a lot different than

Page 198

1saying you can have exclusive dealing arrangements with

2six supermarkets but not eight. You talked earlier

3about the share remedy, so why is one less administrable

4than the other?

5DR. BRENNAN: I mean, I think that is a good

6question. I mean, it's -- you know, we are more

7comfortable with merger law than we are with price

8regulation, and maybe we shouldn't be.

9MS. SCHULTHEISS: Jack?

10MR. KIRKWOOD: Yes, Pat, your question is

11excellent, because it forces all of us who are thinking

12about what standards should be ideal to think about them

13in the concrete situation where a court might enter an

14injunction that incorporates the standard, and so if I

15think a particular standard is appropriate, that would

16incline me to believe that an injunction that wrote it

17down would be appropriate, too.

18I am a little reluctant, for the reason Ken

19suggested, antitrust normally does not get into direct

20price regulation, and you would think in the kind of

21structured rule of reason that Steve and I have been

22advocating, though with some differences, that there

23would be more flexibility in the way a court would

24interpret such a standard in an actual case than the way

25a court might interpret an order that was written down

Page 199

1where someone was seeking contempt sanctions. So, I

2would be a little reluctant to enter an order here and

3would want to pay attention to the possibility of

4structural relief.

5Weyerhaeuser had six different mills in the

6area, so it is not inconceivable, and it acquired those

7mills rather than building them internally.

8MS. SCHULTHEISS: So, you would envision that it

9would be possible for the Government to seek relief

10other than just an injunction?

11MR. KIRKWOOD: Yes.

12MS. SCHULTHEISS: Putting aside damages, you

13know, the DOJ going for treble damages?

14MR. KIRKWOOD: I mean, the preference, as Janet

15suggested and Rick, is for -

16MS. SCHULTHEISS: Deterrence.

17MR. KIRKWOOD: -- private action.

18MS. SCHULTHEISS: Rick?

19DR. WARREN-BOULTON: One little problem, and I

20think it was echoed by somebody this morning, who said

21that, you know, the harm to competitors may correlate

22but is not a very good measure of the harm to either

23consumers or to producers. In the Ross-Simmons case,

24the person who is suing for damages is Ross-Simmons,

25another sawmill. I mean, the whole theory of this is

Page 200

1the person who should be suing for damages is the timber

2owner. I mean, they are the guys who supposedly have

3been harmed by this, who are being monopsonized.

4MS. McDAVID: They weren't sure, they were

5getting a little extra.

6DR. WARREN-BOULTON: The question is, why do we

7have the competitor suing rather than the timberer, and

8the answer is because the competitor gets hurt first,

9right? Now, we could, of course, simply have said

10forget it, let's wait, right? But having the competitor

11sued is a sort of an ex ante, prophylactic way to

12prevent presumably the monopsony harm to the person you

13are really worried about, you know, who probably never

14turned up in this litigation, which was the people who

15were actually sawing the timber. They are the guys who

16were supposedly monopsonized.

17MS. SCHULTHEISS: I am going to ask Ken and

18Patrick, since they have spent the afternoon with us and

19have clearly given a lot of thought to these issues, and

20we talked about it extensively this morning on the

21sell-side, whether you have any response to the remedies

22on this side of it, of the issue.

23DR. ELZINGA: I just have a couple reactions.

24First of all, I must confess -

25MS. SCHULTHEISS: Is there a mic near you?

Page 201

1MS. McDAVID: First of all, you notice they are

2sitting closer together.

3MR. KIRKWOOD: And further away from us.

4DR. ELZINGA: But you will notice it is I who

5moved, and I did that as a symbolic gesture in response

6to what Rick said, and also I moved to my left, which is

7uncharacteristic for me, as my students would know.

8Two things: First of all, this has been

9extremely helpful to me. I didn't realize this whole

10topic on the buy-side was such a big issue, and that is

11a reflection perhaps of my being out of a particular

12loop, but this was a great way to learn about it.

13I also have rarely seen a group of antitrust

14experts basically in as much agreement as this group has

15had. I mean, you can talk about, well, is it consumer

16welfare or total welfare, and everybody can get into

17snipping about that, but I thought there was remarkable

18consensus among the panelists on the topic. So, thanks.

19Patrick, you are going to have to move down

20here.

21MS. SCHULTHEISS: Patrick?

22DR. WARREN-BOULTON: And now for something

23completely different.

24DR. BOLTON: Ken pretty much stole my thunder.

25I am very much in agreement. Just on the last point

Page 202

1about damages, it occurred to me, some of the issues

2were brought up similar to the case, the

3Sotheby's-Christie's -

4MS. SCHULTHEISS: The auction case?

5DR. BOLTON: -- cartel, yes, where there was an

6issue who was harmed and how do you -- was it the -- was

7it the buyers of art or was it the sellers and who

8should be the recipient of the damages, and in that

9case, I forget how it was decided in the end, but there

10was reason to believe that the wrong party was

11collecting the damages. Maybe someone else will -

12DR. WARREN-BOULTON: I agree.

13MS. SCHULTHEISS: Does anybody know?

14MS. McDAVID: I don't remember who got them.

15DR. WARREN-BOULTON: Yes.

16MS. SCHULTHEISS: But it was a damages issue.

17DR. WARREN-BOULTON: Well, the question was, is

18if you read the auction -- it is pretty much like the

19Social Security question, you know, if you pose -

20Social Security taxes, who pays it, is it labor or is it

21the firm? If you ask anybody on the street, they say,

22oh, I pay half of it and my employer pays half of it.

23If you ask an economist, and he will say 95 percent is

24paid for by the workers, but the incidence of a tax

25bears very little resemblance to the accounting and

Page 203

1collection of that tax, and I think in the same way, it

2is extraordinarily difficult to find out who actually

3pays for antitrust violations by simply looking at the

4accounting incidence.

5DR. SALOP: Yes, I -

6MS. SCHULTHEISS: We have two minutes.

7DR. SALOP: Okay, so I will just make a law

8professor remark. There is Illinois Brick and there is

9Hanover Shoe, and that says the direct purchasers get to

10sue the direct purchasers or sellers of the art. They

11get the money. There is no pass-on defense by the

12cartel, and the -- you know, the Supreme Court had a

13reason for that, and so the -- I would think the proper

14people to get the money under the current law are the

15sellers of the art, not the buyers.

16MS. SCHULTHEISS: Okay. Well, I think with

17that -

18DR. ELZINGA: I worked for the judge on that

19case. Most of it went to the lawyers.

20MS. SCHULTHEISS: Well, on that note -

21MS. McDAVID: The entrepreneurial plaintiff's

22bar.

23DR. WARREN-BOULTON: Well, at least some of it

24went to you.

25MS. SCHULTHEISS: On that note, we will wrap

Page 204

1this session up, and I ask you to join me in giving a

2hand to our panelists. Thank you very much.

3(Applause.)

4(Whereupon, at 3:58 p.m., the hearing was

5concluded.)

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

Page 205

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, PREDATORY PRICING

4DATE: JUNE 22, 2006

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

12DATED: 7/9/06

13

14

15

16SUSANNE 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

25DIANE QUADE