Page 1

1	UNITED STATES FEDERAL TRADE COMMISSION
2	and
3	UNITED STATES DEPARTMENT OF JUSTICE
4
5
6
7	SHERMAN ACT SECTION 2 JOINT HEARING
8	PREDATORY PRICING
9	THURSDAY, JUNE 22, 2006
10
11
12
13
14	HELD AT:
15	600 PENNSYLVANIA AVENUE, N.W.
16	WASHINGTON, D.C.
17	UNITED STATES FEDERAL TRADE COMMISSION
18	HEADQUARTERS BUILDING, ROOM 432
19	9:30 A.M. TO 4:00 P.M.
20
21
22
23
24	Reported and transcribed by:
25	Susanne Bergling, RMR-CLR

Page 2

1	MODERATORS:
2	ROBERT POTTER
3	Chief, Legal Policy Section
4	Antitrust Division, Department of Justice
5	and
6	PATRICIA SCHULTHEISS
7	Attorney
8	Bureau of Competition, Federal Trade Commission
9
10	PANELISTS:
11	Morning Session:
12	Patrick Bolton
13	Kenneth G. Elzinga
14	A. Douglas Melamed
15	Janusz Ordover
16
17	Afternoon Session:
18	Tim Brennan
19	John Kirkwood
20	Janet L. McDavid
21	Steven C. Salop
22	Frederick R. Warren-Boulton
23
24
25

Page 3

1	C O N T E N T S
2
3	MORNING SESSION (SELLING):
4	Introduction
5	Presentations:
6	Kenneth Elzinga
7	Janusz Ordover
8	Patrick Bolton
9	A. Douglas Melamed
10	Moderated Discussion
11	Lunch Recess
12
13	AFTERNOON SESSION (BUYING):
14	Introduction
15	Presentations:
16	Tim Brennan
17	John Kirkwood
18	Janet L. McDavid
19	Steven C. Salop
20	Frederick R. Warren-Boulton
21	Moderated Discussion
22
23
24
25

Page 4

1	P R O C E E D I N G S
2	- - - - -
3	MR. POTTER: Thank you for coming, everybody.
4	This is the first substantive hearing on predatory
5	pricing from the Section 2 hearings. My name is Bob
6	Potter. I'm the Chief, Legal Policy Section, Antitrust
7	Division, Department of Justice, and I will be the lead
8	moderator for this morning's session. Sitting to my
9	left is Pat Schultheiss, an attorney with the Federal
10	Trade Commission's Bureau of Competition's Office of
11	Policy and Coordination. She will be the co-moderator
12	for this morning and the lead moderator this afternoon
13	on the buy-side predatory pricing.
14	Before we start, just a couple of housekeeping
15	things that I need to say. One, for the courtesy of the
16	audience and the panelists, please turn off any cell
17	phones, Blackberries or other devices that may make
18	noise during the hearing.
19	Second, the restrooms. The men's restroom is
20	out the double doors to the left, on your left. The
21	ladies restroom is out the double doors, past the
22	elevator bank, to the left, and I saw this morning that
23	neither of them had hot water, so, if you want hot
24	water, you're out of luck.
25	MS. SCHULTHEISS: There is no place in the

Page 5

1	building that has it right now.
2	MR. POTTER: Third, and perhaps most important,
3	in the unlikely event that there is an emergency in the
4	building, please calmly and quickly go out the doors to
5	your right and down the stairs. The Federal Trade
6	Commission has a policy of meeting in the Sculpture
7	Garden, which is on Constitution Avenue. If you don't
8	know where it is, just follow the line of people leaving
9	the building, and I am sure you will get there.
10	This morning, we are very grateful for having a
11	very distinguished panel to talk with us about predatory
12	pricing and Section 2. Our panelists are Ken Elzinga,
13	Professor Ken Elzinga of the University of Virginia;
14	Professor Janusz Ordover of New York University;
15	Professor Patrick Bolton of Columbia University; and
16	Doug Melamed of the law firm Wilmer Hale and former
17	Deputy Assistant Attorney General of the Antitrust
18	Division and Acting Assistant Attorney General of the
19	Antitrust Division.
20	The format for this morning is each of the
21	panelists will give a 10 to 15-minute presentation, then
22	we will have a short break, and then we will have sort
23	of a moderated round table discussion for the rest of
24	the time.
25	We want to thank the panelists. I'll introduce

Page 6

1	them each before their speech as opposed to giving
2	everybody's introduction right now, and for the first
3	instance, I will tell you that although I'll give you a
4	short description, a much longer and better description
5	is contained in the biographical information that we
6	have.
7	Our first speaker this morning is Professor Ken
8	Elzinga of the University of Virginia. Professor
9	Elzinga is the Robert C. Taylor Professor of Economics
10	at UVA. He has a long and distinguished teaching career
11	at UVA, having been a faculty member there, although I'm
12	sure it doesn't look like it, for over 40 years.
13	Even more importantly for today's purposes,
14	Professor Elzinga is a creative and prolific academic
15	writer, having authored more than 70 economic articles,
16	a number of which have focused on predatory pricing.
17	In addition, perhaps even more importantly,
18	Professor Elzinga has been an expert witness in some of
19	the most important predatory pricing cases in the
20	history of antitrust, including Brooke Group,
21	Matsushita, and most recently, Spirit Airlines.
22	With that, please join me in welcoming Professor
23	Elzinga.
24	DR. ELZINGA: Thank you, Bob. I am going to
25	speak from the table here if that's all right, and I

Page 7

1	have got 15 minutes, max, to talk about predatory
2	pricing. That's a big topic. So, hold on to your
3	seats.
4	As was mentioned, I was the economic expert for
5	the defendants in the last two Supreme Court cases on
6	predation, the first one being Matsushita -- that really
7	dates me for some people in this crowd -- and then
8	Brooke Group or what I still call Liggett v. Brown &
9	Williamson, and then also, as was mentioned, I was
10	involved more recently in a predatory pricing case,
11	Spirit Airlines v. Northwest. I did an economic
12	analysis for Spirit, a so-called low-cost carrier. This
13	case had a happy landing for Professor Ordover at the
14	district court level, it had a happy landing for me at
15	the circuit court level, and the final destination of
16	this case is still unknown, but I hope to make a few
17	remarks about it later.
18	When I first started speaking about this
19	subject, before a number of you in this room were even
20	born, there was not much economic analysis embedded in a
21	predatory pricing case. You basically answered two
22	questions. Were prices declining in the market -- not
23	necessarily below cost, mind you, just going down -- and
24	did the defendant generate documents with pugilistic or
25	militaristic metaphors? "We are going to cut off their

Page 8

1	air supply. We are going to squish them like a bug."
2	If I had to pick two events, I am just doing a
3	brief intellectual history here, if I had to pick two
4	events that changed all this, it would be the Court's
5	opinion in Matsushita with its famous line that
6	predatory pricing schemes are rarely tried and even more
7	rarely successful. That statement was based on the
8	Court's exegesis of research about predatory pricing in
9	the economics literature. Almost all of this research
10	suggested that predation would be a strategy that would
11	be difficult to pull off.
12	And the second event was the publication of an
13	article by Don Turner -- the first Assistant AG to
14	enlist an academic economist in the front office, that
15	should always be pointed out -- and Phil Areeda in the
16	Harvard Law Review. It's the most often cited article
17	in antitrust scholarship, led to the Areeda-Turner Test.
18	Now, for this audience, I don't need to review
19	that article or that test, but let me mention for the
20	record how powerful was the hidden economic logic in
21	this famous test by using an iconic product from
22	Matsushita, a 19-inch black and white portable TV set, a
23	consumer electronic products my students today cannot
24	even imagine.
25	Let's say -- and these numbers are not way

Page 9

1	off -- that this set was sold by Toshiba, one of the
2	defendants, to Sears for $95, and the average total cost
3	was $100, but the average variable cost was $90. So, we
4	have ATC equals 100, P equals 95, AVC equals 90. Almost
5	everyone at the time believed Toshiba was selling below
6	cost. After all, how could Toshiba survive with that
7	type of price-cost relationship? And it took an
8	instinct for economic reasoning or a recollection of a
9	price theory course to realize that such a price was
10	above the shut-down point, it was cash flow positive,
11	and that Toshiba was better off making the sale to Sears
12	than not making that sale, and the Areeda-Turner article
13	convinced a lot of people, including a lot of people in
14	this building and a building nearby, of something that
15	economists have known since Alfred Marshall, and that
16	is, in economics, what happens at the margin really does
17	matter.
18	What was missing from Areeda-Turner was a way of
19	thinking about the period of recoupment. They set the
20	stage for a more sophisticated -- I did not say highly
21	sophisticated -- but a more sophisticated economic
22	analysis that the Court adopts in Brooke Group. The
23	Court in Brooke Group recognized that even if a firm
24	charged a price below cost, whatever was the cost
25	benchmark, if the firm couldn't recover its losses, it

Page 10

1	was difficult to make a case for antitrust enforcement,
2	because the aspiring predator would simply shoot itself
3	in the foot if there was no recoupment, and this
4	economic logic behind plausible recoupment entailed two
5	analytical constructs.
6	The first one is real clear in Brooke Group and
7	the second one is not transparent. The first is the
8	recognition that predation is like a capital
9	expenditure. In Brooke Group, the Court cites a paper
10	by David Mills and me entitled "Investment in
11	Predation." Economists have always recognized that a
12	dollar invested today requires more than a dollar in
13	future products because of the time value of money, and
14	Brooke Group understood that and applied that logic to
15	predatory pricing, that losses from predation need to be
16	recouped and not just on dollar-for-dollar basis.
17	The second point follows from the first: Unless
18	entry and exit conditions are symmetrical, the
19	recoupment returns for the aspiring monopolist must be
20	enjoyed for a longer time period than the time frame in
21	which the aspiring monopolist shouldered the cost of the
22	predation strategy, and I could just do a footnote here
23	on Matsushita and how much the world has changed.
24	The plaintiffs in Matsushita thought they were
25	making a good case for their side by arguing that the

Page 11

1	Japanese charged prices below cost for years and years
2	and years, over a decade, not recognizing that the
3	longer is that time period, the more difficult it would
4	be -- indeed, I think mathematically impossible given
5	the power of compound interest -- to ever make up the
6	gains.
7	For those of you who are attorneys, and that
8	would be most of you in this room, I'll tell you what I
9	find to be a fascinating war story from Matsushita. I
10	did some back-of-the-envelope calculations as to what a
11	19-inch black and white TV set would have to sell for
12	under the plaintiff's argument that predation had gone
13	on for 15 years, that is, these sets had been sold below
14	cost for 15 years. What would a 19-inch black and white
15	TV set have to sell for? And I found it would be like
16	$800 into infinity.
17	Now, I don't know if this is one of the things
18	that economists talk about when we are not in the
19	presence of antitrust lawyers. The antitrust lawyers
20	thought, don't ever make that argument on the stand,
21	because the plaintiffs will say, well, even the
22	professor on the other side says the television sets
23	will sell for $800 a year into infinity because of this
24	case. And I said, no, that can't be. They can't sell
25	for that much. They sell for $100 now. They are not

Page 12

1	going to go up to 800, trust me.
2	But my point is, the predator wants the period
3	of recoupment to be long, not the period of predation to
4	be long. The financial rewards that a successful
5	predator is going to enjoy is the present value of the
6	sum of each period's future return once the target has
7	conceded the battle.
8	Now, remember, a business firm has some hurdle
9	rate or internal rate of return before it signs onto any
10	investment project. Signing on for a predatory pricing
11	strategy to an economist is no different. The higher is
12	the hurdle rate, the bigger and longer the monthly
13	returns have to be during the period of recoupment.
14	And Grant, if you could show my first slide,
15	please.
16	In my experience, if one plays with the math
17	that I have at the top, which shows the monthly
18	sacrifice and the hurdle rate and the time period versus
19	the monthly return, it's hard to look at past episodes
20	of predation and come up with examples where recoupment
21	is mathematically possible. To my mind, when I try to
22	teach my students just the basic economics of the
23	elementary price theory level class, the important
24	asymmetry for predation is the one in the little box at
25	the bottom, if you can see it on the slide, slow entry

Page 13

1	but quick exits by target firms.
2	Putting the math aside, putting even the
3	diagrams aside, if there is slow entry but quick exits
4	by target firms, then there's a possibility that
5	predation can be successful. There's got to be, in
6	other words, an economic asymmetry between exit and
7	entry conditions in the market, and think about what
8	that means. In most markets where entry is easy, exit
9	is easy. So, predation simply won't work in those
10	markets. And in like fashion, in markets where entry is
11	difficult, that helps an aspiring predator, exit will be
12	slow, and that is bad news for an aspiring predator.
13	So, what the successful predator needs is a market
14	setting where exit is quick, but entry or supply
15	expansion is slow.
16	Now, in the Spirit-Northwest case, one of the
17	factors persuading me that predatory pricing was
18	plausible or rational for Northwest was because the exit
19	of Spirit, that was the target LCC, the target low-cost
20	carrier, took place quickly, but re-entry and supply
21	expansion was difficult. Spirit Airlines pulled
22	capacity out of Detroit quickly when Northwest cut its
23	fares in the two markets that Spirit served, but Spirit
24	could not enter and expand rapidly during Northwest's
25	recoupment period, because Spirit faced an entry barrier

Page 14

1	in the form of access to gates at the Detroit Airport.
2	Now, I went into the Spirit Airline case as
3	someone from Missouri or Chicago, maybe either metaphor
4	fits, but I ended up concluding that Spirit was a victim
5	of predatory pricing by Northwest, and I'll just say as
6	an aside, this is a case in which Fred Kahn should have
7	testified and not myself. Professor Kahn knows more
8	about the economics of airlines than most any group of
9	economists combined, but he was unable to participate,
10	though he was convinced that predation took place, as I
11	slowly -- kicking and screaming -- came to conclude.
12	The pricing trends in the Spirit case are a
13	textbook example of what predatory pricing would look
14	like. If I could have the first slide, this shows the
15	prices in the Philadelphia area -- I think the first
16	one -- yes, in Philadelphia. There were two city pairs,
17	Detroit, Boston and Philadelphia, and you will see that
18	Northwest prices in both of these are high. Spirit
19	enters; Northwest prices fall dramatically. Spirit
20	exits; Northwest prices jump up. If you show the other
21	slide, you will see basically the same scenario.
22	Now, these price trends -- I want to stress
23	this -- they are merely suggestive. They are not
24	dispositive of predatory pricing. Once a pricing
25	scenario like this is observed, then there follows the

Page 15

1	mind-numbing exercise of comparing revenues with some
2	measure of variable costs, and this is a difficult task
3	in the best of circumstances. It is by no means simple
4	in the airline passenger industry. In the Spirit case,
5	this was a battle between Professor Ordover, Janusz, for
6	Northwest, and Dr. Dan Kaplan was the economist for
7	Spirit. There was also a recoupment analysis done by my
8	colleague David Mills.
9	Briefly, from my perspective, going back to the
10	little box on the bottom of my first slide, one key to
11	the success for Northwest was simply how quickly Spirit
12	exited and the duration of the recoupment period, and
13	that's consistent with the first slide that I presented.
14	I was going to show one more slide, but in the
15	interests of time, I am going to pass on that.
16	Let me conclude this way: Antitrust always has
17	surprises. That is one of the reasons I have enjoyed
18	being an antitrust economist all these years. Let me
19	close by mentioning the surprise for me in the Spirit
20	case.
21	At the last minute, Spirit's attorneys suggested
22	that a price below average total cost but above average
23	variable cost could be predatory, and the Circuit Court,
24	at the tail end of its opinion, seems to suggest that at
25	least in the market circumstances of this case,

Page 16

1	Northwest's conduct may have been predatory even if its
2	fare structure exceeds, as the Circuit Court put it, and
3	I'm quoting here, "an appropriate measure of average
4	variable costs."
5	Now, Spirit's attorneys were pleased with this
6	little present, I am sure. I can restrain my enthusiasm
7	for the way the Circuit Court closed out its opinion.
8	This might take us into a more European view of
9	predation under Article 82, where prices greater than
10	average variable costs might be construed as predatory
11	and where, as I understand that in Europe, there is a
12	continued interest in intent documents and there is no
13	recoupment requirement, again, as I understand it.
14	Like most economists, I can restrain my
15	enthusiasm for the misuse of intent documents. I hold
16	the opposite view here of what had been the conventional
17	view in antitrust. To me, pugilistic and militaristic
18	metaphors are a welcome signal, not of predation, but of
19	competition in a market that doesn't have a stodgy "live
20	and let live" oligopoly setting, and where you see those
21	documents, to me, the prima facie case is that
22	consumers, albeit not rivals, but consumers are the
23	beneficiaries of head-to-head competition and not
24	predation.
25	MR. POTTER: Thank you.

Page 17

1	DR. ELZINGA: Sure, thank you.
2	(Applause.)
3	MR. POTTER: Our second speaker today is
4	Professor Janusz Ordover. Professor Ordover is a
5	Professor of Economics and a former Director of the
6	Masters in Economics Program at New York University,
7	also Director of Competition Policy Associates in
8	Washington, D.C. I first met him when he was the Deputy
9	Assistant Attorney General for Economics in the
10	Antitrust Division.
11	While at the Antitrust Division, Janusz was a
12	member of the White House deregulation task force. He
13	guided economic analysis of antitrust enforcement and
14	acted as a major liaison between the Justice Department
15	and various regulatory agencies.
16	Professor Ordover has written extensively about
17	predatory pricing and has a great deal of experience as
18	an expert witness in predatory pricing cases. He was an
19	expert for the defendant in the Division's American
20	Airlines case, and he is, as Professor Elzinga said, an
21	expert in the Spirit versus Northwest case on
22	Northwest's behalf.
23	Professor Ordover, welcome.
24	(Applause.)
25	DR. ORDOVER: Well, while we're getting set up,

Page 18

1	thank you very much. It's always a pleasure to
2	participate in these kinds of hearings.
3	Predation, of course, marks a lot of my
4	antitrust life. The first time I unveiled my thinking
5	on the subject of predation was about 1980 at the FTC
6	hearings on predatory conduct, and at that time, I think
7	I was attacked -- Professor Willig and I were attacked
8	by Frank Easterbrook, David Scheffman and Mike Scherer,
9	so essentially from left to right, everybody thought we
10	were completely foolish, and Mike Scherer said it was
11	the worst antitrust paper ever written, unlike the
12	Areeda-Turner paper, obviously, has its own different
13	reputation.
14	And then, just a few years ago, it was my
15	misfortune to fly into Ken Elzinga, who has never seen
16	predation other than the case that I was involved in.
17	Something is wrong here. So, I don't know what's wrong,
18	but I guess maybe I will switch careers in my waning
19	years.
20	In any case, what I wanted to do today is to
21	quickly run through some of the ideas that I have been
22	toying with in the antitrust predation field for some
23	past 20-some odd years and perhaps follow up on some of
24	the comments that Ken made, although I will not try to
25	relitigate Spirit versus Northwest. This will have to

Page 19

1	await Northwest's exit out of bankruptcy. So, unless
2	they get into bankrupt predating, but you never know.
3	So, now we are in a holding pattern until somebody
4	coughs up some money and we can actually go back and
5	litigate the antitrust part of the airline life.
6	In any case, what I wanted to do was just go
7	through a few slides focusing essentially on some of the
8	issues that have been discussed over the years, and that
9	is how to analyze challenged conduct from
10	monopolization, particularly paying some attention to
11	predatory behavior.
12	I was going to simply jettison this whole talk
13	by simply saying one should have no price predation
14	cases, but I thought that would be too quick an exit, so
15	I have to torture you for a bit longer to convince you
16	maybe that we should think about it as a possible
17	solution to our woes in this antitrust patch, without,
18	at the same time, suggesting that we should throw out
19	all kinds of scrutiny of firms' conduct, which consists
20	of much more sophisticated pricing from other aspects of
21	what they do, behavior, what I would often call
22	competitive response package, which is I think a term I
23	coined for my testimony in U.S. V. American Airlines,
24	where actually American Airlines' behavior was not just
25	simply pricing but involved a lot of other things that

Page 20

1	the Government alleged were designed to, in fact, retain
2	or maintain or defend American Airlines' position at its
3	hub, Dallas-Fort Worth.
4	So, I'm always thinking about competitive
5	response packages as strategies designed either to exit
6	the rival or to prevent the rival from coming in or
7	possibly designed to contain the rival, and I think it's
8	the last category of strategies which I believe is of
9	great interest and perhaps should be given a little bit
10	more time than we often do.
11	But in any case, the question becomes, how
12	should the decision-maker delineate conduct that does
13	not harm competition by harming scarce rivals from
14	standard, day-to-day market interactions? And
15	economists have been pulling their hairs out since
16	Areeda-Turner, 1975 paper, so we are now in 31 years, 30
17	years of thinking about it, and there is no solution as
18	evidenced by the articles in the latest Antitrust Law
19	Journal, where everybody is still fighting over
20	important things but without actually coming to any
21	particular conclusion.
22	I have been associated over the years with
23	something called the sacrifice test, but I always
24	thought of sacrifice test actually as a version of the
25	welfare test. In other words, what attracted me and

Page 21

1	Professor Willig to thinking about the so-called profit
2	sacrifice approach to delineating procompetitive versus
3	anticompetitive conduct, or at least neutral from
4	anticompetitive conduct, was the notion that at least in
5	some well-defined range of circumstances, these two
6	tests ought to give a pretty close set of answers.
7	In other words, that one was not -- that is, the
8	sacrifice test -- was not somehow biased, setting aside
9	the difficulties of implementation, but it somehow was
10	not biased one way or the other against deterring what
11	would be anticompetitive conduct or what would not be
12	anticompetitive conduct, relating to too much conduct
13	that, in fact, would be harmful. We have been able to
14	show in a variety of circumstances -- in fact, these two
15	tests coincide for the very simple reason that a pursuit
16	of profit, which is the engine of market economies, in
17	fact, is a kind of behavior that generally or frequently
18	does, in fact, conduce to welfare maximization. Seeking
19	profit is a good thing, it is not a bad thing, and
20	therefore, it is not surprising that if you write down
21	your economic model correctly, or at least correctly for
22	the purposes at hand here, that in many circumstances,
23	these two tests will give you the same kind of answer.
24	So, there might be, however, a range of
25	circumstances in which these two tests fall apart by

Page 22

1	virtue of the fact that the basic condition under which
2	they do coincide is potentially difficult to meet, and
3	that condition is incomprehensively stated as the third
4	bullet on this slide, but the basic idea here is that if
5	the incumbent firm can effectively, without creating
6	additional distortions, extract profits by its pricing
7	strategies and other strategies, then any strategy that
8	actually lowers the profits relative to that extraction
9	ought to signal, at least as a first step to the
10	analysis, ought to signal that a firm may have some
11	other aims in pursuing that strategy, something that I
12	think Bernheim and Whinston have now been calling over
13	the years as trying to create market power in what's
14	called a noncoincident market, okay?
15	So, the action takes place in market A, assuming
16	we have it well defined, but the goal essentially turns
17	out to be gaining incremental power or preventing
18	erosion of power in some other market, which Schullman
19	called a noncoincident market, let's say, which could
20	be, in Areeda-Turner world, it could be the same market
21	but in the future day, okay? So, what's the meaning of
22	noncoincident market is actually a little loose, but
23	that's the term that at least Berheim and Whinston in
24	their fine unpublished monograph on exclusionary
25	behavior utilized as a view for analyzing this kind of

Page 23

1	situation.
2	So, it could be a setting in which the whole
3	thing works beautifully. An example is an inferior
4	source of supply, this is the second thing which I think
5	is quite ubiquitous, in which the incumbent firm is
6	faced with competition from another firm or a firm that
7	constrains its ability to exercise pricing freedom,
8	which provides an inferior product, and therefore,
9	enables the incumbent firm to earn supra-competitive
10	profits, at least profits higher than some rents, but
11	getting rid of that firm would, in fact, lift the
12	ceiling and therefore would enable the firm to raise the
13	price even higher.
14	The problem turns out to be that maybe exiting
15	that firm may be just very difficult; however, a
16	circumstance that we have analyzed, Willig and I, under
17	the rubric of systems competition, informs a view of the
18	circumstance in which actually disabling a component
19	that the other firm needs in order to be a full-fledged
20	parcel, bundle and bundle competitor with the incumbent
21	firm will, indeed, lift the ceiling and therefore enable
22	the firm to exercise incremental market power. So, the
23	idea that we have pursued, and the idea which I think is
24	actually fruitful, is that in many circumstances, the
25	goal of the competitive response package is not

Page 24

1	necessarily to kill or to weaken or to disable the
2	person that or the firm against whom this conduct is
3	being perpetrated, but rather, to try to lessen or
4	weaken some other kind of restraint which cooperates
5	complimentarily with the firm whose market presence is
6	being weakened.
7	I think if you look at these Microsoft cases,
8	some of which were discussed along the same lines, this
9	is a fruitful way of thinking about it, but you can
10	immediately see that the economics of the situation is
11	much more difficult than the one instance in the
12	Areeda-Turner case, which is drop the price below some
13	level of cost, you go perhaps profit-negative, assuming
14	you know how to calculate profits, you know how to
15	calculate revenues maybe, you know how to calculate
16	costs maybe, and you can compare the two and see what
17	happens, you are losing money, and as a result of which,
18	it is anticompetitive.
19	But in the situation like this, you don't have
20	to be losing money on anything unless you try to look at
21	the situation in a somewhat different way, which is
22	where the efficient component pricing rule tells you how
23	to look at that situation that I have just described.
24	The efficient component pricing rule for those of you
25	who are not regulatory freaks like myself is a rule that

Page 25

1	tells you what the price of a scarce bottleneck should
2	be if it does not involve any kind of profit sacrifice,
3	okay? So, ECPR is a way of thinking about pricing
4	access, pricing access to the component that is needed
5	in order for the firm to be a viable component system or
6	system competitor.
7	Another example along the same lines, which
8	again focuses on a complicated pricing strategy, not
9	simply dropping price below some measure of cost, was
10	discussed in Ortho v. Abbott. Actually, I worked for
11	Ortho in that case, and there the situation was, again,
12	packaged pricing of a very interesting kind, in part
13	interesting because the buyer insisted on firms offering
14	not only unbundled pricing, but also bundled pricing
15	with a different number of components put in. The buyer
16	needed to buy five tests. There was a regulatory
17	presence out there that required that every blood
18	screening used five tests at that time, I think now it's
19	six, and Abbott was the one that could offer five of
20	them, Ortho could only offer three.
21	Then the question was, could Ortho compete
22	against Abbott if it did not get the access to the
23	remaining two, either because the buyer could create the
24	bundle or because Ortho could buy the necessary input
25	and then resell the bundle? Again, in this case, it

Page 26

1	turns out that there is some discussion that potentially
2	Abbott was pricing the incremental two tests at levels
3	that were unprofitable, that violated some version of
4	what we called a second ago the efficient component
5	pricing rule.
6	What was very complicated in that case was, A,
7	that Ortho did not give me any cost data. So, I
8	couldn't say anything, whether it was true or not, but I
9	did derive the test on a napkin, so other than the
10	Laffer Curve that also was derived on a napkin, this is
11	probably the second most famous napkin in the history of
12	economics. But in any case, the point I'm making is
13	that in this case at least we had a way of dealing with
14	an issue, but we had no reason to explain why this was
15	going on, and I think that's a very important aspect of
16	any predation case, which is that the plaintiff makes a
17	clear connection between the conduct that is at issue
18	and the anticompetitive impact that is being challenged
19	as leading to this anticompetitive outcome down the
20	road.
21	Virgin versus BA, another complicated case that
22	pitted Bernheim against Schmalensee, actually a
23	beautiful battle -- I think it was Schmalensee -- of
24	battle in IO, in which, again, there was no simple
25	pricing strategy, but rather, a complex pricing strategy

Page 27

1	that Bernheim showed leads to an equilibrium in which
2	there's relatively cheap exclusion but in which no price
3	is technically below marginal cost, simply understood,
4	yet as we know, all of these tests that we have in front
5	of us do have some flavor of comparing something to
6	cost, and again, what Bernheim tried to demonstrate in
7	that case, that a simple comparison of price to
8	something like marginal cost may be a flawed way to go
9	if you put that pricing in a strategy that British
10	Airways allegedly developed in a broader context.
11	Quantity-forcing contracts, I think we will skip
12	that, only because we have to, A, rush, and B, we will
13	talk about it in the fall, so I am going to skip that
14	unless it comes up in questions.
15	Just because I don't believe that true price
16	predation is an antitrust offense that is of great
17	interest, it does not mean that we as economists and you
18	as enforcers do not have plenty to focus on. I believe
19	that business strategies, these competitive response
20	packages, that have a strong commitment value, are
21	actually a more relevant focus than just simply pure
22	price predation, which creates all types of problems as
23	these papers in ALJ demonstrate.
24	Commitment to discount, which is Virgin versus
25	BA, commitment to product design, commitment to defend

Page 28

1	lucrative markets, which I call the "new era" tying
2	models, network economies, commitments to effectively
3	raising rival's cost of competing, are the types of
4	strategies that we are now slowly beginning to
5	understand with the help of very fancy economic models
6	and beautiful game theory.
7	The question that I think we will have to leave
8	for Patrick to help us answer is whether or not we can
9	actually fashion workable tests that will take into
10	account these kinds of complications that economists
11	have been focusing on.
12	Thank you.
13	(Applause.)
14	MR. POTTER: Thank you, Janusz.
15	Our next speaker is Professor Patrick Bolton.
16	Professor Bolton is the David Zalaznick Professor of
17	Business. He began as Assistant Professor at U-Cal
18	Berkeley, then moved to Harvard. Then he was the John
19	H. Scully Professor of Finance and Economics at
20	Princeton University.
21	Professor Bolton's research and areas of
22	interest are in contract theory and contracting issues
23	in corporate finance and industrial organization. One
24	of his particular areas of research is the impact of
25	strategic economic game models on predatory pricing

Page 29

1	theory.
2	Professor Bolton, welcome.
3	DR. BOLTON: Thank you, Bob. It's a pleasure
4	and an honor to be on this panel.
5	Unlike Professors Elzinga and Ordover, I have no
6	experience as an expert, haven't had that pleasure, and
7	if you want, I'm a new entrant. We will see whether
8	this will elicit predatory response from the economists.
9	So, my interest, as Robert just alluded to, my
10	interest in this topic came from reading the original
11	McGee article, which claimed that predation couldn't be
12	a rational economic strategy, and, you know, I read this
13	article again and again, and I just was not convinced,
14	and this led me later on to write a theory piece with
15	David Scharfstein where we outlined how predation could
16	be a rational strategy if it took the form of financial
17	predation, and I will say a little bit more about that
18	in my presentation.
19	And then later, I had the good fortune of
20	meeting with Joe Brodley, who introduced me to the new
21	developments in policy under Brooke Group and
22	highlighted some of the problems with the new policy and
23	also some of the new opportunities and challenges, and
24	that then led to our, in my view, very fruitful
25	collaboration on our article, which I will make the

Page 30

1	centerpiece of my brief presentation today.
2	So, I thought I would start by saying first, you
3	know, where are potential areas of agreement among
4	economists and legal scholars and where there are still
5	areas of disagreement. I would argue that this is
6	relatively easy, that we are all in agreement on the
7	general definition on predatory pricing. Namely, it's a
8	price reduction that is only profitable because of the
9	added market power the predator gains from eliminating,
10	disciplining or inhibiting the competitive conduct, and
11	to summarize what both Professors Elzinga and Ordover
12	said earlier, you can distinguish two phases in any
13	predatory pricing episode, a sacrifice phase and a
14	recoupment phase. As Professor Elzinga wrote elsewhere,
15	you can think of predation as an investment in market
16	power. So, I would say that there is general agreement
17	on this characterization.
18	Where there is more disagreement is on policy,
19	and, well, there had been long disagreements on basic
20	economic premise, whether predation is an economically
21	rational strategy and how prevalent predatory pricing
22	episodes are. My sense is that this is an area of
23	convergence, at least on the first bullet point. I
24	think nowadays it is more and more widely accepted that
25	predation can be an economically rational strategy.

Page 31

1	On the second bullet point, I think there are
2	still some areas of disagreement, but I would argue that
3	over time, things have moved in the direction of
4	thinking of predatory pricing as being more prevalent
5	than we thought before and also more likely to succeed
6	than we thought before, in part because our initial
7	beliefs were built on writing, McGee's writing,
8	suggesting that it couldn't be rational, and those
9	writings, I would argue, are now obsolete.
10	There are, however, still very sharp
11	disagreements on the legal standard. Some people argue
12	that we should have simple rules. Others have argued
13	that we should always err on the side of
14	under-deterrence to reduce the risk of false positives,
15	and the policy under Brooke Group is characterized as
16	both being simple and under-deterring. I would argue
17	against this.
18	Now, let me skip the description of Brooke
19	Group, because I imagine most of you are familiar with
20	it, so it involves both a cost test and a recoupment
21	test, and let me emphasize potential problems first with
22	the new policy, and namely, when we look at the facts on
23	what happened post-Brooke, what we find is that since
24	Brooke, plaintiffs have not prevailed in a single case,
25	and almost all cases have been decided by summary

Page 32

1	judgment, and it is only very recently that we are
2	seeing some action on predatory pricing, particularly in
3	the case of Spirit versus Northwest.
4	So, what are the problems with the present
5	policy? Well, first of all, and I think we will discuss
6	this later on in the question time, I would argue that
7	the basic problem with the present policy is that the
8	cost test is highly unreliable. Professor Elzinga
9	earlier qualified proving a cost test as a mind-numbing
10	exercise. I would fully agree with that. I would say
11	that when you go into the details of trying to prove a
12	cost test, you will lose track of the economics of the
13	problem and of the case, and in particular, a very
14	narrow interpretation of the cost test, price being
15	below average variable cost, is a very poor proxy for
16	measuring profit sacrifice, which is what we are trying
17	to go after.
18	Another problem with current policy, we have
19	never gone to a point where we had to ask about a
20	possible efficiency defense on the part of the
21	defendant. There has never been any talk of applying
22	the same rigorous recoupment criteria that the plaintiff
23	has to fulfill on the defendant in proving an efficiency
24	defense. I would argue we should go in that direction.
25	But just to emphasize, I think that the major

Page 33

1	problem with present policy is its failure to focus on
2	the main issues, and those are what is the predatory
3	strategy, what strategy drives alleged predation, first
4	of all, and second, what are the possible dynamic
5	efficiencies and how do you balance procompetitive and
6	predatory effects? And this is where our article takes
7	off and proposes an alternative approach, which I would
8	summarize as taking away some weight off the cost test
9	and emphasizing instead intent, bringing back intent,
10	but intent as structured by an economic analysis, and so
11	this is what in my short time I want to briefly go into.
12	So, specifically, we are thinking that any
13	approach based on intent should be based on strategic
14	analysis of predatory pricing, and in our article, we
15	emphasize at least two well-proven strategies, which are
16	financial market predation and reputation effect
17	predation. We also discuss test market predation. Of
18	course, as Professor Ordover highlighted, predation can
19	take many different and complex forms, and one should
20	not necessarily reduce one's self to just those few
21	strategies, and one should allow for any
22	well-articulated and rational strategy that might be
23	used. I might comment on that later on.
24	Anyway, so what we argue in our paper is that
25	this approach has two advantages. One is that it can

Page 34

1	reduce the risk of false positives, and second, that it
2	puts the spotlight back on what we are really trying to
3	determine, which is discriminate between procompetitive
4	and anticompetitive effects, and there we can use intent
5	as our guide, evidence of intent as a guide to possible
6	defense, and what I mean by intent is not what Professor
7	Elzinga has referred to as militaristic and pugilistic
8	language, but evidence of a deliberate effort to exclude
9	and evidence of pursuit of a predatory strategy.
10	So, in our article, we outline five legal
11	elements to a predatory pricing test. Let me enumerate
12	them first, and then I will go into some of them in more
13	detail. The first element, which is straightforward, is
14	there should be a facilitating market structure. The
15	second element is the scheme of predation and supporting
16	evidence. Third, probable recoupment. Fourth, price
17	below cost. And those four elements would constitute a
18	prima facie case of predatory pricing.
19	I have put the fourth element in brackets here
20	to emphasize the fact that we try to de-emphasize the
21	cost test, and we would agree with the appeals court
22	opinion in the Spirit Airlines case that predatory
23	pricing which is above some measure of average variable
24	cost but below average total cost, that kind of pricing
25	could be predatory. Then, however, we add, if you

Page 35

1	de-emphasize the cost test, we want to add as a safe
2	harbor the -- allow for an efficiency defense.
3	So, how do we prove those elements? Well, some
4	of them are straightforward, and I will not go into -­
5	so, facilitating market structure is any evidence of
6	market power. The scheme of predation and supporting
7	evidence, I want to give you an example of how you go
8	about doing this. So, I will in particular take out of
9	our article the example we have on financial market
10	predation, and so under this element, what is important
11	is to establish that the conditions to implement a given
12	strategy are present and to provide direct or
13	circumstantial evidence showing that this strategy is
14	being implemented.
15	Recoupment, again, this is relatively
16	straightforward. You would want to show evidence of
17	exclusion and disciplining of rivals, and we stressed
18	the idea that second, that you should emphasize probable
19	recoupment instead of actual recoupment, because what
20	matters is whether at the time when this strategy was
21	being chosen, whether at that time, at the time of the
22	information the incumbent had at that time, whether it
23	made sense to implement such a strategy, and we know, as
24	in our own investments in finance, we know that at the
25	time when we make a decision of investment, we make an

Page 36

1	analysis using this kind of cash flows that suggests
2	that we have a positive net present value investment,
3	but that is no guarantee that when we actually undertake
4	the investment, it will end up being profitable. So, we
5	would emphasize probable recoupment, and in particular,
6	put a lot of weight on market structure that makes
7	recoupment likely in the future.
8	Let me also emphasize here the "or related" in
9	brackets, and this is a point that Professor Ordover
10	emphasized, that recoupment shouldn't just be seen in
11	the narrow market where predation takes place. It could
12	be obtained in a related -- I forget the term you
13	used -­
14	DR. ORDOVER: Noncoincident.
15	DR. BOLTON: -- noncoincident market.
16	On price below cost, I do not have much to add
17	to what I have said already except that in the paper we
18	emphasize that a better measure than average variable
19	cost would be average avoidable cost, and a better
20	measure for long-run average cost would be long-run
21	average incremental cost. I do not want to go further
22	into this, because making fine distinctions about these
23	definitions could end up being a mind-numbing exercise,
24	and it just highlights the difficulty with applying the
25	cost test.

Page 37

1	So, what I would like to emphasize, though, is
2	that we would argue that failure to meet the cost test,
3	in particular, failure to establish pricing below
4	average variable cost, should not be grounds for a
5	dismissal on summary judgment and that, in fact, the way
6	to go would be to balance the cost test with an
7	efficiency defense. So, I would argue that if you are
8	able to show that there was pricing below average total
9	cost but above average variable cost but that there was
10	absolutely no efficiency defense, plausible efficiency
11	defense provided, that that would then make a strong
12	case for predatory pricing.
13	So, the efficiency defense, we spent a lot of
14	time in the paper on that, because one of the weaknesses
15	of the policy under Brooke Group and the Areeda-Turner
16	Test is that it really neglects looking at efficiencies,
17	and so we would argue that an efficiency defense does
18	provide safe harbor in itself for price competition that
19	benefits consumers, and we distinguish between defensive
20	defenses and market-expanding defenses and provide in
21	the paper an approach to proving those defenses. So,
22	defensive defenses, we mean by unilateral best response
23	mainly and minimizing losses from unexpected market
24	developments, and as for market-expanding defenses, we
25	really mean here promotional pricing, learning-by-doing,

Page 38

1	and network externalities.
2	So, let me move on perhaps in the few minutes
3	that I have left with an illustrative example. How do
4	you prove financial market predation in a particular
5	case? So, very briefly, the theory here, you know, what
6	is financial market predation, why does it work?
7	Well, the reason why it works is because in
8	corporate finance, there are imperfections -- and there
9	is enormous literature on this -- there are
10	imperfections in capital markets due to agency problems
11	in lending, and as I have argued and have written in my
12	paper with David Scharfstein, a predator can take
13	advantage of those imperfections and drive out an
14	entrant by basically drying up financing.
15	So, how do you go about proving financial market
16	predation? So, we distinguish five essential
17	preconditions. One, the prey's dependent on outside
18	financing. The prey's outside funding depends on its
19	cash flow. Three, predation will reduce the prey's cash
20	flow sufficient to threaten its continued viability.
21	All these are fairly straightforward. Four, the
22	predator knows of the prey's dependence on outside
23	funding or can be assumed to know based on easily
24	accessible facts or rational conjecture. And five, the
25	predator can finance predation internally or has

Page 39

1	substantially better access to external credit than the
2	prey.
3	I think in the Spirit Airlines case, I quickly
4	looked at it, most of these elements you would be able
5	to establish.
6	So, the example we have in the paper is about
7	entry into the cable TV market in Sacramento. This is a
8	case that predates Brooke, and here are the facts. So,
9	this is an entrant with outside financing amounting to
10	$6 million, entered in a small district in the
11	Sacramento area, the Arden District, serving 5000 homes,
12	and the entrant's intention was, of course, to reach a
13	bigger market share and expand gradually in the
14	Sacramento area. The incumbent Sacramento TV company
15	responded to this entrant with drastic price-cutting,
16	and after eight months, the entrant exited. So, how
17	would we prove a scheme of financial predation here?
18	Well, first of all, the dependence on outside
19	funding, what do we know? What are the facts here?
20	Well, first of all, the prey obtained funds through a
21	loan, and the entrant's owners were unwilling to commit
22	more capital than they had initially. Secondly, outside
23	financing depends on cash flow. Well, the incumbent
24	targeted its price reductions on the entrant's customers
25	and potential customers, and that obviously had the

Page 40

1	effect of reducing cash flow. Predation will reduce
2	cash flow and threaten viability. Again, that is easy
3	to establish in this case.
4	The predator knows of the prey's dependence on
5	outside funding. Well, here it turns out there is
6	evidence, intent evidence, memorandum from the
7	incumbent's files that speaks of sending a message to
8	the entrant's bankers. Well, that's relatively easy to
9	establish here. And then finally, the predator has
10	better access to credit than the prey. Again, that is
11	an easy proof in this particular case.
12	So, let me -- sorry for having stepped over my
13	time -- so, let me just quickly conclude with
14	highlighting one potential concern with our approach,
15	and that is something that Posner mentions in his second
16	edition of his antitrust book, and he argues that one
17	concern one might have with evidence of intent is that
18	it's really "a function of luck and of defendant's legal
19	sophistication." So, we would argue that this concern
20	is reduced if the plaintiff is also required to prove,
21	as we articulate in our article, all the other elements,
22	and if what you are required to establish is the
23	implementability of a rational predatory strategy.
24	So, let me end with that.
25	MR. POTTER: Thank you very much.

Page 41

1	(Applause.)
2	MR. POTTER: Our final speaker today is also the
3	only lawyer on the panel, although Doug is very used to
4	dealing with economists, so I am sure it will not be a
5	problem for him to follow them.
6	Doug is a partner at Wilmer Hale, and he is the
7	co-chair of the firm's Antitrust and Competition
8	Department. He has significant experience in a number
9	of government investigations, both government and
10	private litigation, substantial antitrust counseling,
11	and some of that counseling in investigatory work, in
12	litigation work, has involved predatory pricing.
13	From 1996 to 2001, Mr. Melamed served as the
14	Principal Deputy Assistant Attorney General in the
15	Antitrust Division and then as the Acting Assistant
16	Attorney General in the Antitrust Division. He's a
17	prolific writer, a frequent speaker, always has
18	interesting viewpoints that are well thought out. His
19	most recent -­
20	MR. MELAMED: Don't raise the bar, please.
21	MR. POTTER: -- his most recent article, which
22	appears in the summer 2006 Antitrust Law Journal
23	provides a thought-provoking commentary on whether there
24	are unifying principles under Section 2.
25	Mr. Melamed, welcome.

Page 42

1	MR. MELAMED: Thank you.
2	Well, I am a lawyer, and much though I enjoy
3	listening to economists and talking to them, I am going
4	to be talking as a lawyer now and giving a lawyer's
5	perspective on some aspects of the predatory pricing
6	issue.
7	Let me start by saying, I think Brooke Group was
8	correctly decided, an important decision, it brought
9	needed rigor and order to predatory pricing law, but I
10	am concerned about what has happened to it in the life
11	of the law. There is a kind of -- I do not know if this
12	is the right word -- a kind of rarefaction of Brooke
13	Group that I think has done some mischief, and let me
14	tell you what I mean.
15	As everyone knows, Brooke Group has proven to be
16	a defendant friendly standard. As Professor Bolton
17	noted, no plaintiff has won a predatory pricing case
18	post-Brooke Group. Not surprisingly, therefore, when
19	price is an element of the allegedly unlawful strategy,
20	the defendant argues that the standard to be applied by
21	the Court should be Brooke Group, and, of course, they
22	are entitled to do that, because if that's the law, they
23	ought to make that argument, and certainly I have done
24	that myself.
25	But if it is not a straightforward price-cutting

Page 43

1	case, if it is a little complicated, the plaintiff says,
2	"No, no, no, this was different, bundled discounts,
3	aggressive buying, low prices conditioned on exclusivity
4	or other preferential treatment and so on." So, you
5	have a legal dispute. Does Brooke Group apply? Is this
6	the right category, predatory pricing, in which Brooke
7	Group applies, or does the conduct at issue belong in a
8	different category?
9	And there is a kind of a notion that there is an
10	apparent precision of Brooke Group, the price-cost test
11	and the recoupment test, that is uniquely valuable but
12	uniquely applicable to predatory pricing, and one
13	consequence of this is that when the Court decides in
14	this kind of stovepipe analysis that the conduct before
15	it really should not be considered predatory pricing,
16	too often, courts seem to find themselves in a kind of
17	"deer in the headlights, what do I do now" posture, and
18	the result is incoherent decisions like LePage's or
19	courts affirming nonsensical and meaningless jury
20	instructions like Weyerhaeuser and basically a casting
21	about in the way that Professor Elhauge had spoke of
22	Section 2 as a kind of incoherent mess.
23	I think this stovepipe or essentialist way of
24	looking at predatory pricing has created these kinds of
25	dichotomies as categorization, and it has inhibited the

Page 44

1	development of a more robust antitrust jurisprudence,
2	one that can help courts make reasoned decisions about
3	conduct that they do not think falls into a precise,
4	well-established category, whether it be exclusive
5	dealing or predatory pricing or whatever.
6	Put differently, instead of inducing from Brooke
7	Group principles of broader application in the kind of
8	common law tradition which antitrust has in other
9	contexts involved, the process seems to have separated
10	predatory pricing from other forms of exclusionary
11	conduct, and it's done so because there has been in what
12	I call this rarefaction a number of propositions about
13	predatory pricing that are taken for granted or thought
14	to be true or thought to be unique to predatory pricing,
15	and I want to express some skepticism about that. There
16	is a lot of these propositions I have in mind, four or
17	maybe three depending on my time, and I want to express
18	skepticism either that they are true or that they are
19	unique to predatory pricing or perhaps both.
20	So, proposition one, to apply the price-cost
21	test, we need to select some term of art from the
22	economists as our measure of cost, average variable cost
23	or something like that. Now, this is a big topic. I
24	will make just a couple observations.
25	Almost everyone seems to agree that some kind of

Page 45

1	incremental cost is the right measure, because we want
2	to know whether the allegedly predatory sales cost so
3	much that either the defendant must have intended some
4	predatory scheme or, at the very least, that the cost of
5	the sales exceeded the amount consumers were willing to
6	pay for them and therefore resulted in a welfare loss.
7	Areeda and Turner say, "Well, marginal cost is
8	the right test, but it's hard to prove, so let's use
9	average variable cost as a proxy," and now we have this
10	debate for 30 years, "Well, average variable cost really
11	isn't a good proxy, we should use average long-run
12	incremental cost or average total cost, may depend on
13	the circumstances," and you all probably read the
14	article, too, the discussion paper which went through
15	this discussion at great length.
16	Why are we even having this conversation? Why
17	are we debating these categories about technical
18	economic jargon that might have made sense in the
19	Areeda-Turner world in 1975, a simple static price
20	series model, and you can draw the ABC curve, the
21	marginal cost curve, and you can talk about these
22	metaphors, what's going on in the real word, but that
23	doesn't make any sense in the real world as I have
24	experienced it as a lawyer.
25	Areeda talked about additional increments of

Page 46

1	output. I have rarely had a client say to me, "I'm
2	thinking of pushing more widgets off my production line.
3	How low can I go in price?" That's not how the problem
4	comes up in the real world, and if it looks like that,
5	there's a lot more going on.
6	The kind of predatory pricing problems I've
7	counseled clients on in recent years are things like
8	this: Price offerings to early adopters in a de facto
9	standards war; prices in two-sided markets; decisions to
10	assign a plant or an airplane to one market or one
11	segment rather than another. In these situations, I
12	think these terms of art that economists have, they are
13	very valuable in their models and their heuristic
14	exercises, don't have much value, and even if they have
15	value to the economists, they don't have much value to
16	the lawyer and the client.
17	What I find is valuable is saying to the client,
18	when I'm talking about costs, "What are the costs you
19	are incurring to engage in the strategy at issue that
20	you wouldn't otherwise have incurred?" Clients
21	understand that question, and it's not always a trivial
22	question, but I think it's one they can answer. So, I
23	think avoidable costs -- and I don't mean that as some
24	technical term, I mean simply as the but for costs of
25	the allegedly unlawful conduct -- is the cost measure,

Page 47

1	okay?
2	Proposition two, price-cutting is beneficial to
3	consumers, so we should therefore have a standard that
4	errs in favor of avoiding false positives. Then Judge
5	Breyer, in the wonderful "bird in the hand" metaphor, I
6	think most famously perhaps articulated that.
7	Here is my concern: Sure, price-cutting is good
8	for consumers, no question about that. So are all sorts
9	of other things that companies do for consumers. In
10	fact, as I understand, from what the economists tell us,
11	that innovation does a lot more for welfare than
12	improving allocative efficiency by some price cuts and
13	supra-competitive down toward competitive levels. So,
14	why don't we -- and innovation, by the way, could be
15	inventing the PC or it could be coming up with an
16	improved method of distribution because of tying
17	arrangements or because of exclusive dealing. It could
18	be anything that improves the value of the product to
19	consumers.
20	In fact, cost of sale reductions could be
21	beneficial certainly to a total welfare sense and
22	ultimately to consumers as well. So, why do we single
23	out price-cutting, which I don't think has any unique
24	benefits to consumers?
25	Now, there is one thing about price-cutting that

Page 48

1	is different, and that is it's unambiguously in the
2	interests of consumers. A product improvement, you
3	know, the car with the air conditioner might look like
4	it's better, but maybe consumers would rather have
5	better mileage. So, there is some ambiguity about
6	whether other forms of conduct benefit consumers, but
7	why do we have a legal superstructure built on the
8	premise that pricing is unique?
9	At some point, if we do that simply because it's
10	easier to identify the consumer benefit, don't we begin
11	to look like the economists searching for the keys under
12	the light post? At the very least, when the defendant
13	is able to show that his conduct is benefiting
14	consumers, why treat predatory pricing any differently?
15	Proposition three, the recoupment requirement is
16	central to and a great contribution to predatory pricing
17	law. Let me be clear. I strongly believe there should
18	be something like a recoupment requirement at least in
19	the sense of a market power screen; that is to say, a
20	plaintiff ought to have to prove that the allegedly
21	predatory scheme will pay off for the defendant by
22	creating additional market power or preserving market
23	power that will guard against -- kind of belt and
24	suspenders -- a mistake in the application of the
25	price-cost test, and it will preserve antitrust

Page 49

1	violations for those cases where there is competitive
2	harm, and we won't worry about the others.
3	I think, in fact, there should be such a screen
4	in all cases of exclusionary conduct. The problem is, I
5	think in many quarters, including some of my
6	predecessors this morning, the recoupment test is
7	understood to mean that the plaintiff should prove,
8	should quantify, the defendant's investment in the
9	predatory strategy and then quantify his
10	supracompetitive returns during the recoupment period,
11	discount them by risk and uncertainty and time, and
12	conclude that the recoupment exceeds the investment.
13	Now, I think evidence of that sort, on that
14	issue, whether introduced by the plaintiff or the
15	defendant, should be relevant in a predatory pricing
16	case, because it certainly illuminates the likelihood
17	that what is going on here is some exclusionary conduct,
18	but I am very skeptical of the notion that that should
19	be an element of the offense. It clearly complicates
20	the proceedings, increases costs. It may be an
21	impossible burden for the plaintiff in a multi-market
22	reputation effect recoupment story.
23	If taken literally, you would have to go to a
24	profit-maximizing standard to figure out the defendant's
25	investment in the predatory strategy, because you

Page 50

1	wouldn't be asking simply what did it cost him below
2	cost, you would be asking how much in profits did he
3	sacrifice. It's not necessary in order to identify
4	anticompetitive conduct, because if we think we got the
5	price-cost test right and the guy is selling below cost,
6	you can actually, it seems to me, infer that he expects
7	to recoup. It's not needed, because the market power
8	screen will identify the cases of competitive harm. And
9	finally -- and this is a point that I don't know that
10	it's original to me, but I haven't seen it before -- I
11	think it is an illusion that we're measuring something
12	about the welfare effects of the conduct when we use a
13	recoupment screen.
14	The welfare question in predatory pricing is
15	whether the welfare gains, consumer or total, during the
16	rivalry period, the competitive period, are greater than
17	or less than the welfare costs, consumer or total,
18	during the recoupment period, but the recoupment test
19	doesn't measure either of those. The recoupment test
20	measures producer surplus in the competitive period
21	versus producer surplus in the recoupment period, and it
22	doesn't take a whole lot of imagination to think of
23	situations where the results could be different, where
24	you could have, for example, recoupment but no welfare
25	loss from an allegedly predatory strategy.

Page 51

1	So, proposition four, in applying the Brooke
2	Group price-cost test, the price of the product at issue
3	is the appropriate price to compare to cost. That in my
4	view is only partially correct. Obviously you look at
5	the price of or the revenues generated by the additional
6	sales attributable to the predatory conduct. You don't
7	look at the price of, of course, the inframarginal
8	units, the units that would have been sold anyhow,
9	because those units didn't exclude the rival or at least
10	they didn't exclude them by reason of the
11	anticompetitive conduct.
12	But that's not all there is to it. Suppose
13	we're in a two-sided market. Suppose you're cutting
14	price on circulation of the newspaper in order to
15	generate more readers and therefore more advertising
16	revenues. Surely you want to take into account the
17	incremental advertising revenues.
18	Suppose you have complimentary revenues. You
19	know, the Government didn't accuse Microsoft of
20	predatory pricing because the browser was free when
21	bundled with the operating system. Because it was a
22	plausible story that it increased revenues, we didn't -­
23	increased revenues for the operating system. What about
24	revenues lost from inframarginal sales; that is to say,
25	the sales that the defendant would have made anyhow even

Page 52

1	if he had not engaged in the predatory scheme, but he
2	would have made them at a higher price if he hadn't cut
3	prices?
4	To me, another way of putting that question is,
5	are we concerned with the incremental revenues or the
6	revenues from incremental sales? The law chooses wisely
7	in my view the latter. It ignores the loss of
8	inframarginal revenues, I think -- I know Professor
9	Bolton may disagree with this -- I think the law wisely
10	ignores that, because if you want to go into those lost
11	inframarginal revenues, you have to have a profit
12	maximization test, you know, what would have been the
13	profit-maximizing outcome of the strategy, and that is
14	in most cases going to be virtually impossible it seems
1