Understanding Single-Firm Behavior : Exclusive Dealing Session

1

1	UNITED STATES FEDERAL TRADE COMMISSION
2	and
3	UNITED STATES DEPARTMENT OF JUSTICE
4
5
6
7	SHERMAN ACT SECTION 2 JOINT HEARING
8	UNDERSTANDING SINGLE-FIRM BEHAVIOR:
9	EXCLUSIVE DEALING SESSION
10	WEDNESDAY, NOVEMBER 15, 2006
11
12
13
14
15	HELD AT:
16	UNITED STATES FEDERAL TRADE COMMISSION
17	601 NEW JERSEY AVENUE, N.W.
18	WASHINGTON, D.C.
19	9:30 A.M. TO 4:00 P.M.
20
21
22
23
24	Reported and transcribed by:
25	Susanne Bergling, RMR-CLR

2

1	MODERATORS:
2	DAN O'BRIEN
3	Chief, Economic Regulatory Section
4	Antitrust Division, Department of Justice
5	and
6	MICHAEL G. VITA
7	Assistant Director
8	Bureau of Economics, Federal Trade Commission
9
10	PANELISTS:
11
12	Morning Session:
13	Jonathan M. Jacobson
14	Howard P. Marvel
15	Richard M. Steuer
16	Mary W. Sullivan
17	Joshua D. Wright
18
19	Afternoon Session:
20	Stephen Calkins
21	Joseph Farrell
22	Benjamin Klein
23	Abbott (Tad) Lipsky
24
25

3

1	C O N T E N T S
2
3	MORNING SESSION:
4	Introduction
5	Presentations:
6	Jonathan M. Jacobson
7	Howard P. Marvel
8	Richard M. Steuer
9	Mary W. Sullivan
10	Joshua D. Wright
11	Moderated Discussion
12	Lunch Recess
13
14	AFTERNOON SESSION:
15	Introduction
16	Presentations:
17	Stephen Calkins
18	Joseph Farrell
19	Benjamin Klein
20	Abbott (Tad) Lipsky
21	Moderated Discussion
22	Conclusion
23
24
25

4

1	P R O C E E D I N G S
2	- - - - -
3	MR. VITA: Good morning, everybody. My name is
4	Mike Vita. I am an economist here at the Federal Trade
5	Commission. My title is Assistant Director for
6	Antitrust in the FTC's Bureau of Economics. My
7	co-moderator is Dan O'Brien, Chief of the Economic
8	Regulatory Section at the Department of Justice,
9	Antitrust Division.
10	I am going to be leading the morning session,
11	and Dan will be leading the afternoon session, and
12	before we get started with the substance of today's
13	hearings, I am going to cover a few housekeeping
14	matters.
15	First, turn off the cell phones. You'll get
16	detention if you -- the BlackBerries and any other
17	devices that make noises, that's very important.
18	Second, for those of you who aren't familiar
19	with the setup here at 601 New Jersey, the rest rooms
20	are down the hall, past the guard's desk and to the
21	left. I think there are signs out there in the lobby to
22	guide you.
23	Third, a safety tip particularly for visitors.
24	In the unlikely event that the building alarms go off,
25	which they actually did yesterday, please proceed calmly

5

1	and quickly as instructed. Dan and I will keep
2	everything calm and orderly. If we must leave the
3	building, exit the New Jersey Avenue exit by the guards,
4	that's where you probably came in, and follow the stream
5	of people running to a gathering point where you can
6	await further instructions.
7	Finally, we request that you not make any
8	comments or ask questions during the session. Thank
9	you.
10	Okay, today's session concerns exclusive
11	dealing, one of the most interesting areas I think of
12	all the various topics involving vertical restraints and
13	vertical contracts. It has been an active area of
14	economic research and an active area of antitrust as
15	well. We are honored to have assembled a distinguished
16	panel of practitioners and professors who are very
17	knowledgeable in the issues we are going to tackle
18	today, and there are going to be two sessions, one in
19	the morning and then one later in the afternoon.
20	I will just briefly introduce the panelists for
21	this morning before we get started, and I will give a
22	little more detailed introduction as each speaker takes
23	his or her turn. I do not know if everybody is in some
24	sort of order, but it looks like they are.
25	Okay, so immediately to Dan's left is Richard M.

6

1	Steuer, who is a partner at Mayer Brown Rowe & Maw, LLP.
2	Next to Richard is Mary Sullivan, who is an Assistant
3	Professor of Accountancy at George Washington
4	University. Next to Mary is Josh Wright, who is
5	Assistant Professor of Law at George Mason University
6	School of Law. Next to Josh is Howard Marvel, who is a
7	Professor of Economics in the Department of Economics at
8	Ohio State and also Professor of Law in the Michael
9	Moritz College of Law at Ohio State University. And at
10	the very end is Jonathan Jacobson, who is a partner at
11	Wilson Sonsini Goodrich & Rosati and a Commissioner of
12	the Antitrust Modernization Commission.
13	So, I think we will just get right into it, and
14	let me introduce in detail our first speaker, and in
15	those handouts that you got, there is a more detailed
16	biographical description of each of the speakers as
17	well, and you can also find them on the FTC and
18	Department of Justice web sites.
19	Our first speaker is Richard Steuer, who is a
20	partner at Mayer Brown Rowe & Maw, where he specializes
21	in the practice of antitrust law, including litigation,
22	mergers and acquisitions, intellectual property
23	licensing, franchising and e-commerce. Richard has
24	written a book and several articles on antitrust law
25	which have appeared in various journals throughout the

7

1	country. For three years Richard served as chair of the
2	Antitrust Committee of the Association of the Bar of the
3	City of New York.
4	Richard?
5	MR. STEUER: Thanks, Joe.
6	In baseball they say you can learn a lot by
7	watching, and I have been fortunate over the years to
8	have been able to observe a great deal about exclusive
9	dealing and in various contexts, both in litigation and
10	counseling, and I put what I knew into three articles
11	that I have written, and I thought that the best way to
12	try to present what I have learned about exclusive
13	dealing would be to go through those articles and
14	briefly outline what it is that I have learned from
15	watching.
16	The first one was an article on "Exclusive
17	Dealing in Distribution," focusing on how exclusive
18	dealing works when you are talking about selling to
19	resellers, and this appeared in 1983. I will not take
20	very much time on the history, but it is interesting
21	that once upon a time, the FTC considered most exclusive
22	dealing to be virtually per se unlawful. The Standard
23	Stations case in 1949 introduced the rule of
24	quantitative substantiality. Then the major case of
25	Tampa Electric in 1961 brought in qualitative

8

1	substantiality, and then we found a more nuanced rule of
2	reason approach with the Beltone case from the FTC in
3	1982, Jefferson Parish in the Supreme Court in '84, and
4	added to that are the nuances of rule of reason analyses
5	we get from California Dental.
6	Now, what I have found is the level of
7	distribution really matters in assessing the impact of
8	exclusive dealing. What we are measuring with exclusive
9	dealing -- why exclusive dealing is different from other
10	restraints -- is that we are looking more at foreclosure
11	of competitors than anything else. Exclusive dealing is
12	interesting among the vertical restraints. This is the
13	one that, although it has almost always been a rule of
14	reason offense, plaintiffs win quite often, and what we
15	are looking at is something quite different than in
16	vertical resale restraints where the restraint is on
17	reselling rather than purchasing. Exclusive dealing is
18	a restraint on purchasing, not on selling.
19	So, the level of distribution could be
20	wholesalers. One wholesaler can reach every retailer in
21	America, potentially. With retailers, it is different.
22	Retailers are chained to a location typically, although
23	with the Internet, that is not quite as true anymore,
24	and this is a fluid field. Retailers could be in
25	chains, but basically they have a universe of consumers

9

1	that they reach. Wholesalers are a little bit
2	different, because foreclosing wholesalers does not mean
3	that you are foreclosed from reaching retailers.
4	Foreclosing retailers may or may not mean that you are
5	foreclosed from reaching end users. Reaching end users
6	is the simplest. To the extent that there is an
7	exclusive dealing arrangement tying up 10 percent of end
8	users, you have got 10 percent of the market.
9	Type of product is important. Shopping products
10	are products for which consumers will go from place to
11	place to compare prices, to compare features. The fact
12	that each dealer only has one brand does not necessarily
13	have as much of a foreclosure effect, because consumers
14	will not stop at that dealer. They are more likely to
15	go and continue shopping, looking at other brands at
16	other dealers.
17	Convenience products, on the other hand, include
18	impulse products, products that a consumer is more
19	likely to buy because he or she is at the retailer, and
20	that goes to the concept of "can the retailer deliver
21	customers?" Is the retailer such that, when you think
22	about the nature of the retail operation, a customer
23	going to that retailer is going to buy whatever brand
24	there is, so that exclusive dealing is going to have a
25	more considerable impact.

10

1	Another variable that is important to keep in
2	mind is alternate channels of distribution -- what is
3	sometimes called intertype competition -- and there was
4	a rather classic book that Palamountain published in
5	1955 on that. Today, the variation in intertype
6	competition is richer than ever with the rise of the
7	Internet and other alternate channels. So, one needs to
8	look, when you are dealing with resellers, at what other
9	types of means are there, direct sales and so forth, for
10	getting the product distributed.
11	Another possibility is simply establishing new
12	distributors. Is it more efficient, is it more
13	competitive, to have competitors with other brands
14	establish their own distribution networks than just
15	piggyback on the existing distribution network and
16	possibly compromising the amount of vigor with which the
17	intermediate, the reseller, is pushing each brand? Are
18	you better off having one brand at each reseller and
19	having them competing against one another?
20	Foreclosure is measured in many, many antitrust
21	defenses. There is a measure of foreclosure for
22	monopolization, for attempted monopolization, under
23	Section 3 of the Clayton Act, under Section 1, and I
24	recently had an opportunity to study what the different
25	tests are, and I will not belabor the point here -- we

11

1	do not have time -- but they are all over the lot.
2	The interesting thing is "foreclosure" is a term
3	that is used throughout the antitrust lexicon, but it
4	has a different meaning with each substantive offense,
5	and that is important to keep in mind.
6	The procompetitive effects when you are going
7	through distribution: Combating manufacturer-level free
8	riding. This is not the kind of free riding that we
9	were talking about in a case like Sylvania where one
10	retailer free rides on the efforts of another. This is
11	one manufacturer free riding on the efforts of another
12	manufacturer, and exclusive dealing, by keeping other
13	manufacturers out of a particular wholesaler or
14	retailer, prevents that.
15	Of course, stimulate distributors. If the
16	distributor only has one brand of a product, it is going
17	to devote all of its efforts to that brand, but again,
18	in measuring how valuable that is, there is a
19	distinction between commodities and differentiated
20	products. With a differentiated product, there is
21	something more for the dealer to explain, typically,
22	about the features of the product. With commodities,
23	that is probably less so.
24	Stimulating suppliers. Exclusive dealing also
25	stimulates suppliers to put more time and effort and

12

1	money behind their channels of distribution, because
2	they know that other brands are not using the same
3	retailer or same wholesaler, and they do not have to
4	worry about divided loyalties where they are wasting
5	their effort.
6	Protecting trade secrets is similar. To the
7	extent that a manufacturer is providing trade secrets to
8	a retailer or a wholesaler on how to sell, if that
9	retailer or wholesaler is carrying other brands, it can
10	use that kind of information for the benefit of the
11	other brands.
12	Quality control as well is something that can be
13	controlled more directly with exclusive dealing where
14	there are not other brands in the house, and that is
15	particularly true where retailers or wholesalers are
16	doing things with the product, to the product, where, if
17	there is some kind of adulteration, it is hard to
18	control quality with other brands in there.
19	Resale restraints. There is a lot of talk and
20	we were talking earlier about whether there is going to
21	be a change in the rule on resale price maintenance.
22	Some of these same considerations also go into the kind
23	of resale restraints we looked at in a case like
24	Sylvania, customer restraints, territorial restraints,
25	resale price maintenance, but those are all restraints

13

1	on selling, not on buying. So, some of these apply, but
2	they do not apply in the same way.
3	The next thing I looked at ten years later was
4	"Discounts That Induce Exclusive Dealing," and this is a
5	little bit different again, but yet another nuance. I
6	started with single products. In the simplest case,
7	there is one product involved, the grand daddy of the
8	cases is United Shoe Machinery, 1922, but these cases
9	still continue. The latest one, and I am not going to
10	dwell on cases, but there is a case this year from the
11	Sixth Circuit that the plaintiff won on essentially a
12	single product. Big cases out of the U.S. were
13	Nutrasweet, which involved one product, and Tetra Pak,
14	packaging.
15	The important thing to know in these cases is
16	whether or not there is an offer you cannot refuse.
17	These are discounts to induce exclusive dealing. It is
18	not an outright exclusive, but it is basically a deal
19	saying if you buy 50 percent of your requirements from
20	me, you get one price; if you buy 75 percent, you get
21	another price; if you buy 100 percent, you get still
22	another price. It does not sound like it is quite as
23	much foreclosure as exclusive dealing, and in many
24	cases, it is not as much foreclosure, it is perfectly
25	fine.

14

1	However, sometimes it is essential for the buyer
2	to buy some of the product from one brand, and a classic
3	case, we talked about learning from observing, there was
4	one case that I was involved in where it was almost a
5	commodity product. It was a fairly undifferentiated
6	product, but it was differentiated in certain quality
7	aspects, and because the buyers had to buy a particular
8	brand to satisfy their customers, because it was spec'd
9	in, there was one company that had 100 percent of the
10	manufacturing. When a second company came along and was
11	about to turn the key to open their factory, the first
12	company came up with a discount schedule, that as long
13	as you bought 80 percent from me, you got a much lower
14	price. If you only bought 79 percent from me, you got a
15	much higher price.
16	Well, it turned out that about half of what all
17	the customers needed they could not buy from anyone
18	else, not because one product was better than the other
19	or even very different, but it was spec'd in, they had
20	to have it, and so it was an offer they could not
21	refuse, because if they bought less than 80 percent,
22	they would be paying a lot more for everything that they
23	bought. The company that would be trying to break into
24	the market would have to replace all of those lost
25	discounts on the quantity that they could not have. So,

15

even though it was not really a different product,

1
2	analytically, it almost was a different product, because
3	there was some quantity that they had to have from the
4	other brand.
5	A little like bundling. Bundling is almost
6	easier to see, because there are different products in
7	the bundle. Some of them are products you have got to
8	have because they are patented in some cases. Sometimes
9	you do not have to have them, and there are ways of
10	ameliorating it. I am not going to spend time on
11	bundling, because I know you have another program
12	devoted to that entirely, and I could spend a whole day
13	on bundling.
14	The last thing I looked at was, who is
15	instigating exclusive dealing, and should it make a
16	difference? And particularly, "Customer-Instigated
17	Exclusive Dealing." There are mixed motivations on how
18	many suppliers you would like to have in the market.
19	End users have two different motives. On the one hand,
20	they would like to assure that there are plenty of
21	suppliers, because they would like to have alternatives,
22	and they want to play one supplier off against another
23	to get the best price. At the same time, there may be
24	cases where if there is a requirements contract -- and a
25	requirements contract not only means I will buy

16

1	everything from you, but the seller promising I will
2	supply everything that you need -- if one buyer can get
3	a requirements contract and there are not enough other
4	sellers to go around, it could have an impact harming
5	competitors of the buyer. So, it is possible that there
6	are situations where an end user would have a motive, at
7	least in the short term, not to have as many suppliers
8	survive.
9	Resellers, it is somewhat similar. In the short
10	term, if you are an exclusive reseller of a particular
11	brand, you would like to see all the other brands
12	disappear. They only provide competition to you. In
13	the long term, though, if that arrangement is not
14	necessarily perpetual, the day may come when you would
15	like to have some options with other brands that could
16	supply you.
17	Now, why would a customer want exclusive
18	dealing? The most obvious reason is to induce lower
19	prices, to say to a supplier, I am giving all of my
20	business to one supplier, and it may be you, but it may
21	not be, so sharpen your pencil and give me your best
22	price.
23	Another reason is to assure a dependable supply,
24	and that is the requirements contract. Another is to
25	assure quality, in that it is expensive to qualify

17

1	suppliers in certain very technical industries, and you
2	do not want an unlimited number of them. In some cases,
3	assuring uniformity is important. There is a case
4	involving auto racing where it was felt to be important
5	that everybody have the same tires so that there is a
6	level playing field among competitors. And achieving
7	logistical efficiencies. In some settings, just having
8	fewer suppliers is going to wind up lowering expenses.
9	Now, how do you find an appropriate legal
10	analysis where it seems that the buyer has instigated
11	the exclusive dealing? The supplier's objectives often
12	are twofold. One is to foreclose others, and that is
13	the one we always look at when we are trying to see an
14	impact on competition -- will exclusive dealing
15	foreclose other suppliers from having customers or
16	having distribution? Another is to achieve
17	distributional efficiencies.
18	The reseller's objectives are the ones we just
19	talked about, pricing, supply, quality, uniformity --
20	and there are mixed motives about how strong a reseller
21	wants other brands to be.
22	The end user's objectives are a little bit
23	different. Again, the end user of course wants better
24	pricing, may have concerns about delivery, quality,
25	uniformity, efficiencies. It is less likely that an end

18

1	user who is insisting on giving all of its business to
2	one supplier is really in favor of weakening other
3	suppliers. There may be those rare cases, but it is
4	less likely that that is what you are going to find.
5	So, what is the right analysis? When should
6	courts second-guess buyers for instigating exclusive
7	dealing and replace the buyer's judgment that it wants
8	an exclusive with the court's judgment? I think that
9	certainly when the buyer has a demonstrable motive to
10	eliminate competition at the supplier level so that it
11	is helping itself in terms of competition, that is one
12	to take a hard look at, but generally, I think it is
13	important to trust the buyer's judgment if it is
14	instigating exclusive dealing.
15	Let me just conclude by saying I hope this quick
16	snapshot has highlighted some of the very many
17	differences that exist among exclusive dealing
18	arrangements. All of us as lawyers and economists are
19	always searching for those unifying principles that make
20	it easy to do the analysis, but I think what is
21	important here is that we not get lazy and overlook that
22	some of these variables that we have just been talking
23	about really do make a difference to the analysis.
24	I will leave it there, and thank you very much.
25	(Applause.)

19

1	MR. VITA: Thank you, Richard. Insightful and
2	on time, perfect.
3	Our next speaker is Mary Sullivan, who is an
4	Assistant Professor of Accountancy at George Washington
5	University. Mary received her Ph.D. from the University
6	of Chicago, Department of Economics, and taught
7	marketing at Chicago Graduate School of Business from
8	1987 through 1997. While at Chicago, she conducted
9	research on industrial organization and marketing
10	issues, such as slotting allowances, brand names and
11	trademarks.
12	In 1997, Professor Sullivan left academia for
13	the U.S. Department of Justice Antitrust Division where
14	she worked on a variety of antitrust matters and served
15	as Assistant Chief of the Competition Policy Section.
16	In 2004, she joined the Accountancy Department
17	at George Washington University, and as many of you
18	know, Mary's research has been published in numerous
19	leading economics journals.
20	Mary?
21	DR. SULLIVAN: Thank you. I would like to start
22	by thanking the DOJ and FTC for inviting me to
23	participate in these hearings, and I need to keep track
24	of the time very closely, because I have been threatened
25	by Dan and Mike that if I go over my time limit, that

20

1	they might charge me a slotting allowance, although in
2	practice, I have learned that it is very difficult to
3	charge one unless you charge it in advance.
4	Nonetheless, I will try to stay on track.
5	Slotting allowances and payola are two allegedly
6	exclusionary practices that receive different regulatory
7	treatment. What I am going to do in my talk is address
8	whether the different regulatory treatment is warranted.
9	Slotting allowances and payola are similar in
10	many respects. They are basically the same practice
11	used in different settings. Slotting allowances are
12	payments made by manufacturers to retailers for stocking
13	new products. Payola consists of payments made by
14	recording companies to radio stations or DJs for playing
15	a particular piece of music. Both practices have
16	promotional effect. They serve to increase demand by
17	providing exposure to the product or music to consumers.
18	In each case, there is a scarce resource that
19	needs to be allocated, shelf space in the case of
20	slotting allowances and airspace in the case of payola.
21	For both types of fees, there are concerns about
22	exclusionary effects. If you read news articles or, you
23	know, just search the web for these practices, or if you
24	have talked to industry participants, you will learn
25	that these practices are widely believed to be

21

1	exclusionary, and the potential exclusionary effect is a
2	major motivating factor in the regulatory scrutiny that
3	each of these practices has received.
4	Now, oddly, despite their similarities, the
5	practices receive different regulatory treatment.
6	Slotting allowances are not regulated by the FTC. In
7	the FTC's 2001 report on slotting allowances, they said
8	that the fees need to be judged on a case-by-case basis
9	with attention both to likely competitive harms and to
10	likely procompetitive effects. So, they take a basic
11	rule of reason approach.
12	Alternatively, the FCC does regulate payola.
13	According to the FCC regulations, payments are
14	prohibited unless an announcement of the endorsement is
15	made every time a song is played, and this increases the
16	cost of using payola. Now, in addition to the FCC
17	regulations, the major recording companies have recently
18	settled investigations brought by Elliott Spitzer, as
19	many of you are probably aware. I think what is less
20	well known about these settlements is that the terms of
21	the settlements are more restrictive than the FCC
22	regulations, with payola completely banned in most cases
23	even if an announcement is made of the endorsement.
24	Now, given over the past few years we have
25	learned a lot about slotting allowances, both in terms

22

1	of the economic theories and in legal challenges, I
2	thought it would be an interesting exercise just to go
3	through some of the things we have learned to try to get
4	some insight as to why payola has received different
5	regulatory treatment and whether this makes sense.
6	Okay, so we will start with a little bit about
7	the theories of exclusion. Can theories of exclusion
8	explain slotting allowances and payola? Now, there are
9	two general classes of theories that I will talk about.
10	There are the popular theories or notions of exclusion,
11	and then there are the economic, sort of rigorous
12	economic theories of exclusion.
13	The popular theory of exclusion, according to
14	these theories, the payment of the fees increases the
15	cost of introducing a new product or a new song. The
16	increased entry cost may exclude manufacturers,
17	particularly small ones, and many of the complaints are
18	of this nature.
19	However, this so-called theory cannot really
20	explain exclusion. It is fairly well accepted that
21	auctioning scarce resource results in efficient
22	allocation, and unless something in the auctioning
23	process reduces the number of slots that are available,
24	it is very easy to see how this could result in
25	exclusion. If a product or song is very promising,

23

1	someone will give the product financing in order to
2	introduce the product. Therefore, I really don't
3	consider this a valid theory of exclusion.
4	The other class of theories are the economic
5	theories, and the two that I have really looked at for
6	the purpose of this talk are Farrell 2001 and Shaffer
7	2005. Now, without going into much detail at all about
8	these theories, all these theories share the feature
9	that you need to have a contractual provision for the
10	retailer to actually exclude a competitor in return for
11	the fees. You must have a situation in which the
12	retailer is reducing the number of slots available for
13	exclusion to occur and for harm to result from it. So,
14	one important conclusion that I take away from these
15	theories is that simply paying a slotting allowance is
16	not enough to cause exclusion.
17	So, the next thing I want to do is take a look
18	at the evidence, what do we know about slotting
19	allowances and payola, and ask the question whether the
20	evidence is consistent with the Farrell/Shaffer type
21	theories of exclusion.
22	In the case of slotting allowances, the answer
23	is sometimes. Occasionally slotting allowances are
24	accompanied by a contract to reduce the shelf space
25	available to competing manufacturers which could weaken

24

1	them and potentially exclude them. According to the
2	FTC's 2003 study of slotting allowances, such contracts
3	are fairly unusual, but they do occur.
4	For payola, the answer is no. There is no
5	evidence that exclusionary contracts are being used with
6	payola. The evidence that I have seen suggested that
7	recording studios are simply trying to use payola in
8	return for getting the radio stations to play their
9	songs, not that they would not benefit if they could
10	exclude a popular song of a competing recording studio.
11	I think, you know, if they could exclude a competing
12	song, it would allow them to sell more records; however,
13	there is simply no evidence at all that that is what is
14	happening, and believe me, if you take a look at some of
15	the Spitzer settlements, you will see that the evidence
16	he collected was quite thorough. What I conclude from
17	this is that according to the economic theories of
18	exclusion, payola is very unlikely to be exclusionary.
19	Now, I also wanted to take a look at some of the
20	evidence from the courts to see what the courts say
21	about slotting allowances and exclusionary effects.
22	This is not really intended to be a comprehensive review
23	of the legal cases on slotting allowances. What I did
24	do is I looked at two legal challenges to slotting
25	allowances that are both important, have been very

25

1	influential, and I see cited quite often in other cases.
2	In both of these cases, the courts found that the fees
3	are a valid means of competing, and here are the two
4	cases.
5	One of the quotes from the Gruma case is
6	particularly revealing. In this case, the Court said,
7	"Some of the plaintiffs' losses are due to a
8	'self-inflicted' wound -- they chose not to compete for
9	shelf space."
10	Now, in this case, the plaintiffs were small
11	companies, small tortilla manufacturers who were
12	complaining that Gruma, the large manufacturer, was
13	buying up all the shelf space and giving it unfavorable
14	locations. The Court ruled, well, your tough luck. If
15	you want to be in this game, you need to compete for
16	shelf space.
17	Now, in the Reynolds Tobacco/Philip Morris
18	case -- which is often referred to as the retailer
19	leaders case, which was the name of the Philip Morris
20	program that was being challenged in court -- it was a
21	somewhat different situation, because Reynolds, the
22	plaintiff in this case, was actually a large company,
23	but the conclusion of the Court was the same. In this
24	case, the Court concluded that the Philip Morris program
25	that involved the payment of slotting allowances

26

1	increased industry competition.
2	Okay, so if the theory predicts that payola is
3	unlikely to be exclusionary and the courts have ruled
4	that slotting allowances are an efficient means of
5	allocating scarce shelf space, then why -- this leads us
6	back to the original question -- why does payola receive
7	different regulatory treatment than slotting allowances?
8	The answer seems to be that since the air waves are
9	owned by the public, there is a belief that radio
10	stations should select music on the basis of public
11	interest rather than the radio station's commercial
12	interest. This view highlights the difference between
13	slotting allowances and payola.
14	The FTC and the courts see slotting allowances
15	as a valid and efficient means of allocating shelf
16	space, but the FCC believes payola results in an
17	allocation of airspace that is not in the public
18	interest apparently because it allows the radio station
19	to play music that increases their profits. Now, does
20	this make sense?
21	Another way of asking that is, will regulating
22	payola cause radio stations to select music that is in
23	the public interest, whatever that is? The answer is
24	no. To see why, it is helpful to understand a little
25	bit about how radio stations are going to decide what to

27

1	play both with and without payola.
2	Now, if payola is banned, radio stations are
3	going to earn all of their money from creative --
4	selling -- or playing music that appeals to an audience
5	that will buy advertisers' products. In other words,
6	they are going to earn all of their profits from
7	advertising dollars. So, what they are going to do is
8	they are going to select music that appeals to people
9	who buy the advertisers' products.
10	Now, if payola is permitted, radio stations earn
11	revenue from both advertising and payola, and this may
12	cause the radio stations to change their selection of
13	music. They may play more songs that appeal to people
14	who buy records and play less songs that appeal to
15	people who buy advertised products. It is not obvious
16	to me that the selection of music will be more in the
17	public interest if payola is banned. In either case,
18	the radio stations choose what music to play on the
19	basis of what maximizes its profits.
20	So, I have several conclusions from this. The
21	first conclusion from the analysis, from this exercise,
22	is that it seems highly unlikely that payola will
23	exclude promising music. This argument of exclusion
24	should not be used to support the regulation of payola.
25	Second, regulating payola will not help achieve

28

1	the goal of serving the "public interest." With or
2	without regulations, radio stations will design
3	playlists to serve their own commercial interests. This
4	is unavoidable.
5	Third, prohibiting explicit payment for radio
6	airspace will not make competition for airspace
7	disappear. There is a scarce resource, and there is
8	going to be competition for it. The competition will
9	take a different form. To the extent that recording
10	studios can find loopholes in the regulation, then there
11	will be little effect on the regulation on what is
12	played.
13	So, my own personal conclusion from this is that
14	the regulation of payola it seems to me does not serve
15	the public interest, appears to be wasteful, and leads
16	to needless enforcement costs.
17	Thank you.
18	(Applause.)
19	MR. VITA: Thank you, Mary.
20	DR. SULLIVAN: No slotting allowance?
21	MR. VITA: You are off the hook, for now.
22	DR. SULLIVAN: Okay.
23	MR. VITA: Okay, our next speaker is Joshua
24	Wright, who is an Assistant Professor of Law at George
25	Mason University School of Law, where he teaches in the

29

1	areas of antitrust, contracts, and law and economics.
2	Professor Wright's research focuses on the law and
3	economics of the competitive process for product
4	distribution, including slotting allowances, category
5	management, exclusive dealing and other contractual
6	arrangements. He has published in numerous journals.
7	Professor Wright received his Ph.D. in economics
8	from UCLA, Department of Economics, and he also received
9	his JD from the UCLA School of Law, where he was a
10	managing editor of the UCLA Law Review.
11	Joshua?
12	MR. WRIGHT: Thank you.
13	Okay, so I am going to sort of hop on the back
14	of some of Mary's comments on slotting and do a little
15	less background talking about what they are, since that
16	has already been covered. My comments here, just as a
17	preface to get out of the way, are based on two papers
18	that are up on the FTC web site, which has all of the
19	slides and papers from the other panelists, both
20	co-authored with Ben Klein, who I think will be here in
21	the afternoon.
22	So, a tiny bit more detail on -- I am going to
23	use a slightly different definition of slotting
24	arrangements than Mary used and define the contracts as
25	per unit time payments made by manufacturers to

30

1	retailers for shelf space. There is a couple of
2	differences here. One is that sometimes, and indeed, in
3	the FTC report that has been referenced, you will find a
4	distinction between per unit tying payments and
5	discounts for slotting contracts, and it is an important
6	difference and one that I am going to end up not talking
7	much about here, but there is a discussion in the paper
8	I just referenced on the economics of slotting
9	contracts, on when we might expect the efficient form of
10	a distribution contract to be a per unit tying payment
11	or a discount. That said, I am going to ignore the
12	issue for the next 19 minutes.
13	What else we know about slotting is that they
14	cover both new products and established products. So,
15	they cover -- you know, Coca-Cola pays slotting
16	allowances, products where we do not have any sort of
17	risk imposed on the retailer by giving shelf space to
18	some unproven product. We see slotting allowances on
19	those products as well.
20	What else we know is that they increased, there
21	was a spike in the prevalence and the magnitude of
22	payments somewhere between 1981 and 1984, and over the
23	last 20 years, that trend of increasing and over the
24	products covered and the magnitude of payments has
25	continued.

31

1	So, the anticompetitive theories of slotting,
2	first, before I try to explain a procompetitive
3	rationale for shelf space contracts. We see slotting
4	contracts used by manufacturers with small market
5	shares. We see -- in general, the FTC report finds that
6	the normative time for these agreements are between six
7	months and a year. We see them on products where there
8	are not significant economies of scale in manufacture,
9	one of the conditions that drives the anticompetitive
10	theories in the literature. And also, the
11	anticompetitive theories have a difficult time
12	explaining the jump in the use of the contracts in the
13	middle of the 1980s.
14	In terms of the procompetitive story for
15	slotting allowances, there are really two important
16	economic questions with respect to slotting fees, and
17	the first is why you see a separate contract at all,
18	right? The first economic intuition one might have is
19	why don't we see, like the setting of retail prices in a
20	competitive retail market, supermarkets, et cetera, why
21	don't we see manufacturers just set the wholesale price
22	and allow the retailer to set the level of shelf space
23	that is supplied for different products like we let them
24	set the price? So, why do we see this separate contract
25	for the shelf space?

32

1	And the second is, and more related to the panel
2	discussion today, is we see sometimes that these
3	contracts include exclusivity provisions, unlike the
4	payola contracts. We see provisions that say, give me
5	70 percent of the shelf space, give me a space to sales,
6	give me the full exclusive, do not put anyone else on
7	the shelf space. So, we see this additional variation
8	in the contracts that we are going to need to explain.
9	So, I will turn to that second. There are other
10	interesting questions, again, the form of the payment
11	and these things, which for the moment I am going to
12	skip so I can focus on exclusivity.
13	So, the answer provided by Ben Klein and myself
14	in the paper I alluded to earlier, the intuitive answer
15	is what you see on the screen, and it is that slotting
16	contracts solve this pervasive incentive incompatibility
17	problem where the retailer does not want to supply the
18	joint profit maximizing level of promotional shelf space
19	under the conditions where the supply and the shelf
20	space does not induce consumer switching. So, we have
21	cases like McCormick and we have 90 percent of the shelf
22	space allocated for spices. Well, supplying additional
23	promotional shelf space to spices does not induce a
24	greater number of consumers to say I will not shop at
25	this retail outlet because they have given 90 percent of

33

1	the shelf space to spices, and they have two brands, and
2	so I am going to leave. So, we expect to see this
3	incentive incompatibility problem solved with a separate
4	contract under these conditions.
5	Now, I am going to go through a little bit of
6	the analysis with a simple model with a little bit of
7	math, but here is the intuitive answer. So, the
8	fundamental point here is that for many products, and
9	differentiated products, we have manufacturers with a
10	large profit margin. So, the manufacturers, the
11	wholesale price over the marginal cost, this P sub W
12	minus the marginal cost of manufacture, is large
13	relative to the retailer's incremental profit, whether
14	it sells Coke, Pepsi or any brand of soda, okay?
15	For a number of products, this is generally the
16	case. So, the retailer, when it is making its decision
17	on the optimal level of shelf space, promotional shelf
18	space to supply to the manufacturer's products, say
19	Coca-Cola, does not take into account that these
20	promotional sales induced by giving, say, the eye-level
21	shelf space, or if you are in the children's cereal
22	aisle, the children's eye level shelf space, these
23	incremental profits are large for the manufacturer and
24	not taken into account by the retailer.
25	Now, we can make the same argument with respect

34

1	to price competition, but there is a key difference as
2	to why we see manufacturers in the retail setting, at
3	least, allowing the manufacturers to set the retail
4	price, and competition between retailers is sufficient
5	to get an optimal jointly profit-maximizing price set
6	but not the jointly profit-maximizing level of shelf
7	space. So, why do we get prices right and shelf space
8	wrong ends up being the question.
9	So, unlike the shelf space case, when we are
10	talking about price competition, you see here we have
11	got on the right-hand side is this large manufacturer's
12	margin, that P sub W minus the marginal cost of the
13	manufacturers. It is large. It is maybe 10-20 times
14	larger than the retailer's margin for a good chunk of
15	products. But we have this offsetting effect induced by
16	customer switching. So, the intuition here is that
17	while the manufacturer's margin is much larger, we have
18	got this switching effect, so the quantity response
19	faced by the retailer when it changes the price has
20	these two different components.
21	One, when it reduces the price or increases the
22	price of Coca-Cola, there are interbrand effects, so
23	sales move from Coke to Pepsi, but there also are
24	inter-retailer competitive effects, right? So,
25	consumers may end up switching stores when we are

35

1	talking about price decisions or at least are more
2	likely to do so than when we talk about moving Coke from
3	the bottom level to the eye-level shelf space, right?
4	So, the key point and argument here is that
5	because promotional shelf space does not involve large
6	inter-retailer shelf space effects, we do not see
7	consumers switching on a number of grocery products. My
8	co-author on the paper and dissertation adviser likes to
9	use the example of dog collars in the store, right? So,
10	there is some exclusive space granted for dog collars,
11	and people pay and they compete for this space, but
12	nobody switches the stores because there is one dog
13	collar versus two, okay?
14	And because we have this idea that there are
15	these small inter-retailer effects, it is the case that
16	we have this incentive incompatibility problem, right,
17	and instead of this inequality, if we had the jointly
18	profit-maximizing level, we would see at least this
19	relationship be approximately equal. The big difference
20	is this elasticity from the retailer's perspective of
21	the shelf space effect, right?
22	And so this is all to illustrate the point that
23	where we see these small inter-retailer effects, again,
24	this incentive incompatibility problem is pervasive, and
25	this is especially so in the supermarket context. Now,

36

1	there are some limits on this idea. We do not see --
2	the distinction here is not just because of price and
3	nonprice competition, okay? There are elements of
4	nonprice competition where there are inter-retailer
5	effects because all consumers value the service.
6	So, the supermarket provides a free parking lot.
7	You can go and you park and you do not pay for it, you
8	know, when you go in to park. Everyone generally values
9	that there is a parking lot, maybe there is lighting
10	there so you don't get mugged when you go to the parking
11	lot, and everybody values this, and this means, because
12	consumers value some nonprice services, then they will
13	induce some switching, that for those services, the
14	incentive incompatibility problem is solved. The
15	retailer will supply those because consumers are all
16	willing to pay.
17	So, where we see this, the very idea of
18	promotional shelf space is to give some sort of
19	effective, targeted discount to the marginal consumers
20	who are sensitive to allocations in the shelf space,
21	right? They are sensitive to what is in the eye-level
22	shelf space, and there is a substantial marketing
23	literature which demonstrates sometimes some really
24	surprising results about how large the effects can be in
25	terms of changes in sales when we play around with the

37

1	shelf space allocation.
2	So, in these fairly general circumstances, the
3	disparity in margins and the small inter-retailer
4	switching effects from the supply of promotional shelf
5	space, the manufacturer wants more shelf space than the
6	retailer is willing to supply, and so we need to have
7	some separate contract where the manufacturer pays the
8	retailer for the supply of the shelf space in order to
9	solve this incentive incompatibility problem.
10	So, now we have got a situation where Coke is
11	paying for the eye-level shelf space to the retailer,
12	and it pays them $10,000 per unit time for the month for
13	some contracted-for level of shelf space. Now, this
14	does not mean that the whole process is over, right?
15	So, the manufacturer pays the retailer with this money,
16	and the retailer has some incentive to not perform.
17	It can provide less than the contracted-for
18	level of space. It can otherwise violate the implicit
19	contractual understanding between the manufacturer and
20	the retailer to sell the space twice, in other words,
21	the simple way to think about it. So, it is taking the
22	money and not performing under the terms of the deal.
23	This is where we get to the function of full or limited
24	exclusives in shelf space contracts.
25	Now, we see that in the slotting context, at

38

1	least a full or a partial exclusive seems to be -- at
2	least appears to be thus far -- a necessary condition
3	for liability. So, we have some form of exclusive -- we
4	have -- well, there is no liability, but Gruma, Conwood,
5	McCormick, so we have these cases where the contracts do
6	not just buy the shelf space. They specify a
7	percentage. They specify a full exclusive. They
8	specify limits on the placement of rival products.
9	So, there are a number of procompetitive
10	rationales for exclusivity terms in these contracts, and
11	Mr. Steuer went over many of them, and so I am not going
12	to belabor them here, but the key, following from this
13	sort of shelf space contracting model, is that an
14	exclusive can help facilitate performance of the
15	contract, right? The retailer pockets this money and
16	can have some short-term incentives to not perform.
17	So, a couple of things that exclusivity can do,
18	it can efficiently define exactly what the manufacturer
19	is purchasing. Purchasing all of the shelf space,
20	detecting cheating becomes easy. The other thing it
21	does is it allows the retailer to say, you are bidding
22	for all or 70 percent or some large fraction of the
23	promotional shelf space, and this intensifies the
24	bidding process between the manufacturers for the shelf
25	space, and this is a good thing in terms of the

39

1	antitrust analysis, a good thing for consumers, because
2	these shelf space payments are passed on to consumers,
3	and that is whether they are discounts or per unit time
4	payments.
5	Quickly, so I can end here, category management
6	contracts are just a form of limited exclusive, where
7	what we are doing instead of saying you get 50 percent
8	of the space is the retailer delegates the function to
9	the manufacturer to allocate the shelf space, and we see
10	this in circumstances where consumers' demand for a
11	particular brand is high. So, the implicit contract is,
12	you get to feature your product, Coca-Cola, and you can
13	allocate the shelf space, but if consumers come to me
14	and say I have a high demand for Pepsi and you're
15	putting it on the bottom or you have run out or you did
16	not put it on the shelf, then I know and I terminate the
17	agreement, okay?
18	Just to finish up, Conwood seems to get this all
19	wrong. So, Conwood, despite the sort of atmospheric
20	facts and the tortious behavior and lots of bad stuff
21	going on, there is some bothersome language in the
22	opinion about imposing a standard on category managers
23	that is tougher than the standard on monopolists using
24	full exclusives, and so the key idea is that exclusive
25	dealing can make economic sense in these circumstances

40

1	and that we need to make sure that the plaintiffs are
2	demonstrating an anticompetitive effect before we engage
3	in any sort of balancing under the rule of reason
4	analysis.
5	I think I went over, sorry.
6	MR. VITA: Not too bad.
7	(Applause.)
8	MR. VITA: Thanks, Josh.
9	Okay, our next speaker is Howard Marvel who is a
10	Professor of Economics in the Department of Economics at
11	Ohio State, and he is also Professor of Law in the
12	Moritz College of Law at Ohio State. Howard's work on
13	vertical restraints is very well known. He has written
14	on a variety of different topics, including resale price
15	maintenance and exclusive dealing, and I know those
16	papers have appeared in some leading economics journals.
17	Howard also has advised the Japanese
18	International Trade Ministry, had a post in
19	telecommunications, the Federal Trade Commission and the
20	National Association of Attorneys General law on
21	vertical restraints issues. In addition, he has served
22	as an expert in vertical restraint matters for a number
23	of firms.
24	Howard?
25	DR. MARVEL: Okay, I have seen a lot of you

41

1	before. I am happy that you have invited me to come
2	talk to you outside of the Third Circuit, and the topic
3	for today is exclusive dealing.
4	It is obvious that exclusive dealing is a very
5	common thing that we see every time, when you go to a
6	MacDonald's, you do not find a Burger King hamburger,
7	and Haagen Dazs has had the exclusive dealing in their
8	distribution contracts, car dealers typically have it,
9	there is exclusive dealing in beer distribution. It is
10	all over the place, and ordinarily we do not think
11	anything about it. You know, any business format
12	franchise is basically franchise or else, and it is most
13	commonly observed for our market leaders, the big guys.
14	Anheuser-Busch has it in the Chicago area, it is
15	under study, and you don't see that elsewhere. Haagen
16	Dazs had contracts with distributors with Steve's, which
17	at the time was a premium ice cream. I do not know if
18	it is still around. Anybody from Boston? Steve's did
19	not have that. The big guys have more reason to
20	foreclose, of course, but they have also more to free
21	ride upon.
22	So, for a long time we had a rule that Richard
23	talked about, how tough it was to engage in exclusive
24	dealing. The rule seemed to be that if you had market
25	dominance or a big share somehow, somehow, and you

42

1	practiced exclusion, if you had exclusion in your title
2	of whatever the practice was, you were toast. So, it
3	was essentially a per se violation.
4	Now, exclusion there does not mean foreclosure.
5	It just means exclusion from a portion of the market,
6	and that is very different than keeping the firm totally
7	out of the market. Foreclosure is a different story.
8	Now, several of the -- I think John is going to
9	talk about the Chicago view and why it is limited, so
10	let's run through what the Chicago view of vertical
11	restraints is. It is that vertical restraints create
12	property rights. So, you have a problem that you want
13	to get somebody to do something, but you are afraid that
14	at the end of the day they will not do it because the
15	fruits of their actions will end up being frittered away
16	as other people take advantage of them, okay?
17	So, the idea behind vertical restraint is that
18	it creates a property right for somebody or other, so
19	exclusive territories, for example, create a property
20	right for customers that a particular distributor or
21	dealer generates, okay? So, I go out to get a customer,
22	how do I guarantee if I am the seller who wants that
23	customer generated, how do I guarantee the customer gets
24	generated? I protect the rights to that customer for
25	the guy who actually did the work?

43

1	Resale price maintenance is very similar. There
2	is a property right for the services that the
3	distributor provides, and Josh talked about how this
4	sort of works in slotting as well, like exclusive
5	dealing, that creates a property right for customers
6	that the supplier's actions pull in, and I think that if
7	you think about the -- almost all of the things that
8	Richard included in his discussion from the 1983 paper,
9	they all have that characteristic, that the supplier is
10	doing something to pull in customers and those customers
11	are being protected through exclusive dealing by -- from
12	some sort of bait and switch approach.
13	Now, the problem with exclusive dealing and what
14	makes it more serious and more of a worry than
15	territories and RPM is that in territories and RPM, the
16	supplier is creating a property right for somebody else.
17	It says, you do this, and you get to keep the fruits, so
18	I would police that. And I am an outsider, and I want
19	to have the distribution system to be as effective as I
20	possibly can make it be, but with exclusive dealing, the
21	property right is for the creator and the monitor of the
22	right.
23	I give myself the right, and then I protect that
24	right, and we have a problem that can emerge there if
25	the right is somehow something that you really don't

44

1	want the guy to have and be able to protect, and that is
2	really what is at the heart of Aspen Ski, because in
3	Aspen Skiing, Aspen Skiing and Aspen Highlands
4	cooperated to develop the Aspen market as a destination
5	for skiers, and then at the end of the day, Aspen Skiing
6	said, well, gee, they passed a law here in Aspen where
7	you have got to have a three-week rental instead of just
8	a one-week minimum rental or a longer rental term, and
9	so you essentially locked customers in. You didn't have
10	to compete for customers so much, because they said,
11	well, we will walk away with rents, and you can see that
12	elsewhere.
13	If you have a patent holder who has accessories
14	for his product, the patent is about to expire, the guy
15	may decide to engage in exclusive dealing to try and
16	freeze out the accessory guys that he's cooperated with
17	to build that product, and believe it or not, I was an
18	expert witness in a matter in which I thought exclusive
19	dealing was used improperly in this way, so it's not
20	clear that these are anticompetitive so much as fraud or
21	contracting problems, but they are problems.
22	Okay, so the basic exclusive dealing story is
23	simply that the manufacturer invests in a product or a
24	reputation that brings in customers, if the manufacturer
25	confers upon its customers -- its customers onto dealers

45

1	who are cloaked in its reputation. So, if I become a
2	dealer for a particular manufacturer, then customers
3	say, hey, that dealer is essentially certified as
4	knowing what he's talking about, so the customer walks
5	into the dealer, induced to do so by the manufacturer's
6	efforts, and then the dealer says, by the way, I have
7	got a better deal for you.
8	Now, a requirement for this to work is that the
9	customer cost, the cost of generating the customers has
10	to be included in the charge for the product. So, if
11	you can charge for leads separately, no sweat, okay?
12	You just charge for the leads, you do the promotion, the
13	customers walk in, and if the dealer who's paid for
14	those customers wants to switch them to some other
15	product, hey, that's fine, okay, but there are a lot of
16	circumstances in which you only charge for the customer
17	when they actually buy something, so it is rolled into
18	the product price, and this is, again, the way it works
19	with royalties in business format franchises, right,
20	because MacDonald's brings customers in, but they only
21	receive a charge, a payment, for those customers when
22	the royalty is generated, okay?
23	So, the dealer can avoid this particular charge
24	through a bait and switch scheme in which he says, okay,
25	you are a customer for firm X, firm X brought you in,

46

1	that is what you came looking for, but firm Y has got a
2	product that is cheaper, because it does not involve any
3	promotion, it is simply a free rider, so why don't you
4	switch to that one, and you can trust me, because I am
5	firm X's dealer, okay?
6	So, what is the evidence for this -- how this
7	works, okay? Is there any evidence to suggest that this
8	works? Well, you know, "can you hear me now" doesn't
9	necessarily need to be Verizon's slogan, it also should
10	be a slogan for the hearing aids manufacturers who were
11	engaged in exclusive dealing, and they were going out
12	and getting a lot of customers to come in, into their
13	dealers, and the customer comes in saying I saw an ad
14	for Beltone hearing aids or whatever, can you fit me
15	with a hearing aid? And the dealer at that point can
16	say, yeah, I am a Beltone expert, and by the way, I've
17	got a better deal on another hearing aid.
18	Now, the interesting evidence on this is that
19	the FTC decided to take four of the five hearing aid
20	manufacturers who used exclusive dealing, take them out
21	and shoot them, because the idea was if you agree not to
22	use exclusive dealing, we'll let you off the hook, and
23	at the end of about a year or so, the bodies of the
24	companies had agreed not to engage in exclusive dealing
25	washed up on the shore. They were out of the business.

47

1	So, that's a problem in these cases, the
2	counterfactual, what would happen if the practice were
3	forced to be given up, is very hard to prove until it is
4	too late. When you see the corpses, then you know you
5	screwed it up.
6	The manufacturers in the hearing aids case did
7	not recognize the role of exclusive dealing themselves,
8	and so they walked away from it. Beltone didn't, but
9	the other manufacturers of hearing aids did, and they
10	ended up dead in short order, okay?
11	Now, after the Chicago explanation came out,
12	then we got a game theory counter-revolution, okay? A
13	famous paper by Aghion and Bolton sort of launched the
14	"why don't we get together, write a contract and screw
15	the next guy to come along" approach to contracting,
16	which is, I think, a fair way to say what their model
17	is. It says, I am in the market now, I am the only guy
18	in the market, you're my dealer, there might be somebody
19	who comes along later and is better than me. Why don't
20	we figure out a way to split the rents from that guy's
21	advantage, okay? And the way we will do that is we will
22	write a contract between ourselves that has a penalty
23	clause, okay, and the penalty clause is such that --
24	five minutes, it says. Okay, I'll never get there,
25	okay? I am a professor, you know, I am not one of these

48

1	lawyer guys. I just talk and talk. That's the way it
2	works, but I'll be done.
3	Okay, so the Aghion-Bolton idea is that there is
4	a contract that is written before the entrant shows up,
5	and then we run off with the entrant's rents because of
6	the existence of this contracting penalty clause, okay?
7	The requirement for that to work is you have got to have
8	a contract, right? That is what you have got to have
9	before this works, because if the entrant does show up,
10	then the dealers run to the entrant if he is better,
11	okay?
12	There is a second set of theories that are
13	contract-based, and you think of the names Segal and
14	Whinston, Ramweyer, Rasmussen and Wiley, and these are
15	train leaving the station contracts. The train is
16	leaving the station, I am the only guy in the market,
17	you better sign up with me or else, and then you have
18	got to stay with me if I am no longer the only guy in
19	the market, okay? So, these both require contracts.
20	All of these theories require contracts. No contract,
21	no problem, okay? And that is the characteristic of the
22	game theory counter-revolution.
23	So, is Chicago out the window? Oh, they are,
24	because Professor -- or Mr. Jacobson -- what is the
25	appropriate -- Mr. -- Mr. Jacobson --

49

1	MR. JACOBSON: Hey you, hey you is fine.
2	DR. MARVEL: Hey you? Okay, he says, but
3	Chicago writers -- post-Chicago writers long ago
4	debunked the Chicago School, and it is now common ground
5	that in many contexts exclusive dealing can be deployed
6	in a way that is both profitable for the dealer and that
7	allows the defendant to reap gains from the arrangement
8	that far exceed the associated costs. Guess what? I
9	agree, okay? True. Absolutely.
10	Now, we will wait for the first one of these to
11	come along, but it is possible, in principle, for this
12	to happen. I do not have the slightest disagreement
13	with that.
14	Now, a couple of examples of this sort of thing,
15	the first from your vintage Chicago School nut case, we
16	appreciate the potential reply that it is impossible to
17	say that a given practice "never" could injure
18	customers. A creative economist -- there are creative
19	economists -- could imagine unusual combinations that
20	would cause injury in the rare situation, but antitrust
21	law applies rules of per se legality to practices that
22	almost never injure customers, and who might that be?
23	Yes, Chicago.
24	Okay, but then we also have this statement the
25	literature on anticompetitive exclusive dealing, so

50

1	actually what we are talking about today, has focused on
2	producing "possibility results" in simple settings to
3	counter Chicago School arguments. It is possible that
4	something can go wrong, says Mike, okay? Now, he is not
5	a Chicago guy, okay, and he is right. He has written
6	some of the possibilities, but the possibilities take
7	contracts, okay?
8	Problems are possible, and the problems involve
9	foreclosure. If you get foreclosure, that does not mean
10	foreclosing a particular set of dealers. It means
11	foreclosing the market. If you get that, that is a
12	problem. The benefits are going to be really hard to
13	prove from exclusive dealing up front. Again, like I
14	said, until you see the bodies wash up on the beach.
15	The default rule in these cases is going to
16	determine the outcome, okay? If the default is that
17	exclusion could be bad, what will happen is that
18	exclusion will be found to be bad despite the absence of
19	factors suggesting the presence that we might have one
20	of the bad theories of exclusion, the proof of concept
21	or possibility theories, present. So, if we get the
22	default rule wrong, what will happen is that we always
23	find that possibility means exclusion, becomes the
24	default rule, and we are back to where we started.
25	Exclusion plus dominance will equal violation. That is

51

1	where we were before. One minute.
2	Beltone, forget them, okay?
3	So, what should we do about all this in the last
4	minute? The first possibility is that all of the
5	possibility results that I know of, and even this guy
6	Joe Farrell back there who just walked in seems to know
7	of, are contract-related, okay? So, why don't we start
8	by requiring a contract? No contract, no problem, okay?
9	Then, we ought to require some notion that there
10	might be something wrong in this market in the sense
11	that there be a showing of foreclosure, and success
12	should not be defined as foreclosure. If I do better
13	than you do, I get a big share of the market, so what?
14	And if my dealers then get that share, so what? Success
15	should never be considered the equivalent of
16	foreclosure.
17	But if you get to that point where you have
18	found that there is a contract and there is a showing
19	that foreclosure is a real problem in this industry in
20	the sense that there is not another way to get to
21	market, then, and only then, after you have gone past
22	those two standards, should you go ahead and run your
23	trade-off analysis, and I am reasonably convinced that
24	that trade-off will often, if not always, that you will
25	find it very difficult to prove that the efficiency

52

1	benefits that you are claiming are really present.
2	With that, we will be done, okay?
3	(Applause.)
4	MR. VITA: Our final speaker before we take a
5	short break is Jonathan Jacobson, who is a partner at
6	Wilson Sonsini Goodrich & Rosati, where he practices
7	antitrust law and has taken a lead role in many
8	significant antitrust matters over his 30-year career.
9	Among other cases, Jonathan was lead counsel for
10	Coca-Cola in Pepsico v. Coca-Cola, a leading Section 2
11	monopolization case.
12	Jonathan was appointed by Congress in 2002 to
13	serve on the Antitrust Modernization Commission, which
14	is dedicated to studying the nation's antitrust laws and
15	considering several changes. He also is the editorial
16	chair of the ABA's Antitrust Law Developments and has
17	chaired a number of ABA antitrust section committees.
18	He has written and edited numerous articles and books on
19	antitrust, and his most recent paper co-authored with
20	Scott Scherr is entitled, "'No Economic Sense' Makes No
21	Sense For Exclusive Dealing."
22	John?
23	MR. JACOBSON: Thank you.
24	I also want to express particular thanks for
25	seating me on the far left wing on this panel. I think

53

1	that is entirely appropriate, although I would comment
2	that in exclusive dealing cases, I have never
3	represented a plaintiff. I would like to, but it has
4	always been defense representation so far.
5	So, let's talk about exclusionary conduct and
6	exclusive dealing in particular. There are lots of
7	different exclusionary conduct devices, and these
8	hearings will cover most of them. I actually think
9	ripping your competitor's racks off the shelves is
10	pretty exclusionary, so maybe we can talk about that in
11	the dialogue, but that is one example of exclusionary
12	conduct. The other is price cutting, which is, you
13	know, rarely, rarely, rarely harmful and yields, you
14	know, major significant consumer benefits.
15	Exclusive dealing is in the middle, and it
16	presents a real challenge, because what makes exclusive
17	dealing potentially harmful is the very same mechanism
18	that makes the arrangement efficient and may lead to
19	lower prices for consumers.
20	So, what are the consumer benefits? I think
21	Richard went through them and I will just go through
22	briefly, but basically the distributor, if we are
23	focusing on distribution, which is the typical case, the
24	distributor focuses his or her attention on the
25	supplier's product and becomes a more effective

54

1	distributor, and from the supplier's perspective, the
2	supplier has an incentive to provide the distributor
3	with information and displays and all sorts of that
4	stuff without concern of free riding by competing
5	suppliers.
6	So, these benefits are very important, but they
7	are possible only because the arrangement is exclusive,
8	denying rivals access to the distributor's capabilities.
9	This same exclusivity can have the effect -- and it is
10	not an ephemeral possibility, it can happen, although it
11	is not necessarily the default rule, but it is a real
12	world phenomenon -- that the exclusive can deny the
13	rivals access to customers or supplies and have the
14	effect of driving their costs up and rendering them less
15	effective competitors, less effective constraints on the
16	defendant's market power. And the result of that can
17	be -- and this is the case we need to be alert to -- to
18	allow the supplier to increase prices to consumers as a
19	result of the weakening of that competitive constraint.
20	So, the question is, how do we evaluate
21	exclusive dealing and quasi-exclusive dealing
22	arrangements in light of these simultaneous benefits and
23	harms?
24	Now, today I think, you know, I have not been
25	here for these hearings, I have read a lot of the

55

1	summaries and some of the testimony, but I suspect that
2	there is agreement on really four issues in terms of an
3	overall approach to exclusionary conduct. One, we do
4	want to prohibit behavior that leads to the creation or
5	expansion of significant market power. We want to be
6	careful, and I think that is a principal focus of these
7	hearings, to avoid deterring procompetitive conduct. We
8	want to have rules that businesses can understand and
9	apply so that they know what they are doing is legal or
10	illegal. And we want to provide the courts with
11	sufficiently clear rules so that they can tell in the
12	context of a lawsuit what is illegal and what is not.
13	So, for exclusive dealing, we have applied these
14	goals. I think you can go back to Tampa Electric and
15	say we have had a rule of reason since then, but I will
16	respect Richard's qualification of that and take the
17	rule of reason back to Beltone, which is clearly the
18	first sort of modern formulation of the rule of reason
19	in exclusive dealing cases. And where we are coming to
20	now, I have another paper where I comment that the focus
21	on foreclosure is unfortunate, and my basic point of
22	view on this, and I think where the law is going to come
23	to if it has not come to already, is that in an
24	exclusive dealing case, what the plaintiff must show to
25	prevail is that the net effect of the conduct, including

56

1	the efficiencies, is to raise prices or otherwise harm
2	consumers. And I think, you know, if you look at the
3	major exclusive dealing cases over the last ten years,
4	the results largely -- not entirely -- but are largely
5	consistent with that kind of paradigm.
6	So, the recent debate was spurred in part, I
7	think, by the thinking of folks like Judge Easterbrook,
8	who gave a talk a few years ago saying that we should
9	abandon Section 2 enforcement entirely, but that has led
10	a lot of conservative thinkers and some more mainstream
11	and liberal thinkers, like Steve Salop, to try to
12	determine whether there is a universal test for
13	examining exclusive conduct, and at some level we have
14	been searching for the universal rule ever since Learned
15	Hand's decision in the Alcoa case.
16	I would commend to all of your attention an
17	excellent article in the Antitrust Law Journal a few
18	months ago by Marc Popofsky, that having a
19	one-size-fits-all approach that can be applied equally
20	to practices as diverse as predatory pricing, refusals
21	to deal, ripping your competitors' products off the
22	shelves, has proven to be elusive. And I do not think
23	we have gotten there yet, and I question whether we ever
24	will.
25	The main area of disagreement is the extent that

57

1	we need extraordinary screens to ensure that
2	procompetitive conduct is not deterred. The sort of
3	screens that I would add that we do not see in most
4	areas of the law other than antitrust. Antitrust, at
5	least in the last few years, has been very sensitive to
6	avoid deterring procompetitive conduct at the cost, many
7	recognize, of allowing the occasional illegal behavior
8	to go through.
9	All right, so -- by the way, thank you for not
10	allowing questions from the audience, because Greg
11	Werden is here -- and it is with quite a bit of
12	trepidation, although he and I have had a few
13	discussions on this subject, that I challenge the no
14	economic sense test or Doug Melamed's version, the
15	profit sacrifice test. This issue has gained -- and
16	appropriately so -- a lot of attention, and under at
17	least one articulation of the no economic sense test, a
18	practice is not exclusionary for purposes of Section 2
19	unless it would make no economic sense for the defendant
20	but for the tendency to eliminate or lessen competition.
21	And in varying degrees, some of the advocates of this
22	test urge that it be applied to all single-firm and
23	vertical conduct.
24	If you look at the certiorari brief filed by the
25	Justice Department in the Trinko case and the briefs

58

1	filed in the Court of Appeals in the Dentsply and
2	American Airlines cases, the Justice Department has
3	argued variations on this test as a rule of law. It has
4	not been adopted by any of those courts, but it has been
5	argued with some vigor by the Department of Justice.
6	One of the issues I have with the no economic
7	sense test is that it is fundamentally the Areeda Turner
8	predatory pricing pricing test in new garb. Areeda
9	Turner made a major advance in the law in 1975 when they
10	urged that predatory pricing not be condemned unless it
11	is below cost with a likelihood of recouping the lost
12	profits through the market conditions that will result
13	from the predatory pricing scheme. And their test was
14	acknowledged and stated by them to be an extraordinary
15	test reserved exclusively at that time for price
16	cutting, because price cutting is so rarely harmful and
17	so extraordinarily important to our economy that we want
18	to have a test that really makes sure that errors are
19	purely on the side of allowing the defendant to win
20	rather than the plaintiff to prevail.
21	Now, there have been efforts starting with the
22	article that Janusz Ordover and Bobby Willig put out a
23	few years after that to apply this sort of analysis more
24	regularly to other forms of exclusionary conduct, but in
25	general, we have been asking ourselves the question

59

1	since the no economic sense literature came out, is this
2	purposefully extraordinary test -- and it was designed
3	as an extraordinary test -- is it appropriate to apply
4	it to other types of exclusionary conduct?
5	In my view, as applied to exclusive dealing, the
6	no economic sense test really does make no economic
7	sense, and I say that because exclusive dealing
8	arrangements make economic sense precisely because they
9	lessen competition by rivals for the affected business.
10	So asking that question tells us nothing about whether
11	the arrangement is procompetitive or anticompetitive.
12	Exclusives are usually associated, even in
13	extreme cases like Dentsply, I think you can say that
14	exclusives are usually associated with real efficiencies
15	and sometimes cost very little to implement. So, unless
16	you apply the economic sense test with the rigor that a
17	Greg Werden would, and if you apply it in the real
18	world, it is very easy to come out with the
19	determination that the exclusive makes economic sense
20	for the defendant.
21	But the way in which those efficiencies are
22	achieved, as I said before, is through this mechanism of
23	exclusion. So, the judicial audience, the business
24	audience out there, is wondering, how can I do this?
25	This arrangement makes no economic sense to me unless I

60

1	can exclude my rivals, but that seems to be the test for
2	illegality, so what do I do? And I think the answer to
3	that is you apply a different test.
4	So, exclusive dealing is also interesting and
5	different, as Steve Salop points out, because at least
6	under some scenarios there need be no period in which
7	profits are sacrificed during the course of the
8	exclusive dealing arrangement. You can have
9	simultaneous exclusion and recoupment.
10	All right, recent case, not a federal case,
11	although I will tell you we did our best to get the
12	Justice Department and Federal Trade Commission to file
13	a brief and they politely declined, but the Court came
14	out correctly I think anyway, although it was a 5-4
15	decision, and if you really want to read something
16	interesting, read the dissent in the case. It is a
17	decision that came out less than a month ago out of the
18	Texas Supreme Court, and it involved exclusive
19	promotional agreements with retailers, not exclusive
20	dealing arrangements, but exclusive promotional
21	agreements.
22	In some of the agreements, Coke -- in all of the
23	agreements, Coke had to get a reduced price. In some of
24	the agreements, it provided that the low price had to be
25	the lowest in the store on that particular package. The

61

1	exclusives required the most prominent displays in the
2	stores and also exclusive ads.
3	In return for this, Coke provided very
4	significant lump sum promotional payments and deeply
5	discounted wholesale prices. So, the result was to
6	reduce the retailer's costs, both marginal costs and
7	total costs. Coke had 70 to 80 percent of the market if
8	you accepted the market definition in the case. The
9	result of this was lower prices for Coca-Cola products,
10	and it was not seriously disputed that the level of
11	promotional activity resulted in overall lower prices in
12	the marketplace for carbonated soft drinks as a whole.
13	Now, the exclusivity in that case, the
14	agreements, made economic sense only because the
15	exclusives made more -- made things more difficult for
16	rivals, and the easy example is to ask why would Coke
17	pay thousands of dollars to a supermarket for a
18	promotion? Let's say the promotion is two-liter and
19	you expect that the reduced price would be something
20	like 99 cents. If the consumer is going to walk in the
21	store and the first thing she is going to see is a Pepsi
22	display of two liters at 89 cents, that promotion really
23	is not worth very much for Coke. Why would Coke spend
24	the money for that promotion? Why wouldn't it just
25	figure out some other way to sell soft drinks?

62

1	The problem, as the dissent points out, is that
2	this kind of exclusivity could fail an incautious
3	application of the no economic sense test, but
4	appropriately, the majority upheld the agreements under
5	the rule of reason because there was no showing that
6	they led to increased prices in the market as a whole.
7	Now, I will very briefly talk about Microsoft,
8	and I am not going to go through the whole slide, but
9	the basic concept here is a lot of what Microsoft was
10	doing was virtually costless. Leaving Internet Explorer
11	out of add/remove programs was virtually costless, and
12	if you apply the no economic sense test to Microsoft,
13	you can easily get a situation where the Court would say
14	that this conduct makes economic sense and is,
15	therefore, upheld. I think the Court went through an
16	elaborate recitation of the rule of reason, and I think
17	we have a good precedent there.
18	I had promised not to go over time, and I see
19	that I already have. What I do want to point out is
20	that the focus that we care about in antitrust generally
21	and in exclusive dealing cases as one piece of that
22	overall puzzle is does this behavior injure consumers?
23	Does it raise prices? Does it otherwise injure
24	consumers and the benefit of the bargain that they are
25	going to receive?

63

1	The no economic sense test asks that we bypass
2	that question. My point is simply, let's look at that
3	question directly. Let's try to get to that analysis
4	directly. The shortcut, which if applied incorrectly
5	can lead to very questionable results, is not a
6	necessary route. It does not protect competitive
7	conduct any more than a careful application of the rule
8	of reason would. So, let's just ask the question that
9	we really want the answer to and guide our analysis on
10	that basis.
11	Thank you.
12	(Applause.)
13	MR. VITA: Thank you, John.
14	I think we will take a short break right now.
15	Why don't we come back at -- ten past? -- yeah, ten
16	minutes past, and we will reconvene.
17	(A brief recess was taken.)
18	MR. VITA: All right, let's get started.
19	I think the first thing we will do here is take
20	a few minutes and just open it up to the panel to allow
21	them to react to some of the things that they might have
22	heard and pose questions to the other panelists. So,
23	Jonathan, you came by before and said you had an issue
24	you wanted to raise. I'll let you have the honor of
25	going first.

64

1	MR. JACOBSON: Well, thank you. I previewed
2	this with Howard, because I think the no contract --
3	MR. VITA: Jonathan, speak into the mike.
4	MR. JACOBSON: I think the "no contract, no
5	problem" scheme is a problem, so to speak, and what I
6	would ask Howard is, isn't it a fair observation that
7	you worry more about exclusive dealing the larger the
8	market share of the defendant, and don't you run into
9	cases where the defendant's share is so high -- it is
10	not really the share, but the market power of the
11	defendant -- where the defendant's market power is such
12	that they can enforce exclusives on the offer you can't
13	refuse or the all-or-nothing offer that Richard was
14	referring to with a lot of the detriments that can be
15	associated with exclusive dealing with little or none of
16	the benefits?
17	And again, you know, Microsoft is not a bad
18	example. Those were not contracts at least of any
19	duration in that case. Microsoft basically told Dell
20	and Compaq and Hewlett Packard, you know, here it is,
21	deal with it, and, you know, it was not a really good
22	option for them to go to UNIX, and Apple was not
23	available. So, let me put that one back to you.
24	DR. MARVEL: Well, I guess what I would say is
25	that looking at the economic analysis of exclusive

65

1	dealing and at the places where the game theoretic
2	models have found problems, they are all cases in which
3	there is not an option today and I sign up everybody
4	today and I lock them in, okay? And since that is
5	virtually always the case in all these models, if you
6	find another example of a circumstance in which you say
7	there is a real economic loss that results from this, I
8	would like to see an economic analysis of why there was
9	an economic loss there. So, I wait for some economist,
10	the clever economists that Easterbrook was talking
11	about, to come up with the explanation.
12	I think I probably could for Microsoft as to why
13	Microsoft's behavior might be a problem, but that is not
14	similar to the ones that we have already talked about,
15	okay? So, in -- I hate to do this with Gail here -- but
16	in Dentsply, one of the things that was interesting
17	about that case was that the Justice Department seemed
18	to recognize early on that they needed to provide a de
19	facto contract analysis as to why there was lock-in,
20	okay? So, they said, okay, it is because of inventory
21	investments. I bought so many inventories from these
22	guys, from Dentsply, that if I walk away from them, I am
23	stuck with the inventories, and the alternative
24	explanation in that case said, hey, you really want
25	those inventories to tide you over while you are trying

66

1	to convert customers, right?
2	And so, in fact, in that case, the lock-in
3	turned out not to be lock-in, because Dentsply was happy
4	to buy back those inventories, and the guy that walked
5	away from Dentsply to sign up with rivals found that he
6	sure wanted a hell of a lot more Dentsply teeth than he
7	was going to get. So, there was no lock-in there in
8	that case at all.
9	And again, it is possible to imagine
10	circumstances in which a manufacturer exerts or creates
11	a property right for itself to take advantage of
12	somebody who has sort of cooperated with it to develop a
13	new product, and then the manufacturer says, hey, why
14	don't I seize that new product on my own and define this
15	property right and take that right away from the other
16	guy?
17	That is a problem, but that is almost as much of
18	a fraud or a contract problem as it is an antitrust
19	problem. It becomes an antitrust problem only if you
20	get to the point where it says people are standing on
21	the sidelines unwilling to invest because they are
22	subject to this misappropriation of their up-front
23	investments.
24	So, I can imagine circumstances under which that
25	might work, but I am not sure that you need to attack

67

1	them in this sort of standard exclusive dealing context.
2	MR. JACOBSON: I don't want to hog the mike, and
3	I know Dentsply, we would get a very different view of
4	the facts from people like Gail and Mark Bodde (ph), but
5	what about Lorraine Journal? No contracts, you know --
6	DR. MARVEL: Well, you brought that one up to
7	me, and unfortunately, not being a lawyer -- and I am
8	not a lawyer, even though I am a professor of law -- I
9	am going to have to duck on that one, because I do not
10	know the facts.
11	MR. STEUER: Well, maybe if I can jump in --
12	MR. VITA: Let me remind people, just pull the
13	mikes up close to your face so they actually function.
14	MR. STEUER: It may be that lawyers and
15	economists do not always define "contract" exactly the
16	same way, and lawyers get hung up with the whole Colgate
17	doctrine. In the case that I alluded to before, for
18	example, a monopolist had 100 percent share of the
19	market and came up with a discount schedule that
20	basically made it advantageous for customers who needed
21	to have some of its product to buy all of that kind of
22	product from it so that when a new competitor opened its
23	factory, it was facing the daunting challenge of having
24	to replace all of the discounts that would be lost by
25	potential customers giving up any of it. There was no

68

1	contract.
2	It was similar to a Colgate relationship that
3	way. It was simply a unilateral policy, "Here is my
4	price schedule if you do what I want you to do," and yet
5	it seemed to have all of the foreclosure effect that a
6	bilateral contract would. So, to some extent, maybe we
7	are talking past each other a little bit in terms of the
8	terminology and what is a contract and what is not.
9	DR. MARVEL: Well, maybe so, but one of the
10	things that you brought up, Richard, in your discussion
11	was this NicSand case, right? And one of the things
12	that has really impressed me about the cleverness of the
13	post-Chicago world is how really imaginative they are at
14	coming up with sort of contract-based explanations for
15	why you could have problems, but, of course, the Chicago
16	side does that, too, and you look at Lepage's and
17	NicSand, and those are matters in which the Justice
18	Department says we don't know yet what we should be
19	doing, so let's wait a while before we have the Supreme
20	Court step into that, or at least that is what happened
21	in Lepage's.
22	But, in fact, we are starting to figure out that
23	those things involve -- I mean, maybe Lepage's was
24	collateral damage, because there was a real problem with
25	getting your entire line carried if you are going to a

69

1	discounter, like a WalMart or a K-Mart. So, it is very
2	possible that in a case like that, what you are really
3	trying to do is induce the discounter that you are
4	dealing with -- and this is particularly true for
5	discounters -- to carry a much broader portion of the
6	line than they would otherwise carry, and that is going
7	to increase consumer welfare even though it is going to
8	increase prices or it is going to increase economic
9	welfare.
10	So, I mean, you can get into these circumstances
11	where you say, I don't understand yet why the
12	manufacturer is doing this, so it must be foreclosure,
13	but if you stand back for a while, maybe somebody will
14	come along and say, hey, some of these bundling schemes
15	have the efficiency effects that are pretty significant,
16	and I think that cases like those may just be
17	circumstances in which you are dealing with a guy who is
18	going to carry a very narrow portion of your line, and
19	you do not like that, so you pay him to carry a broader
20	portion, and if somebody -- and you say, well, I am
21	offering you this really good deal to carry the broader
22	portion of the line, and maybe if that excludes somebody
23	else, well, yeah, that could very well do that, but that
24	is not the only effect of it, and so it is a really --
25	these are really tough questions.

70

1	MR. STEUER: Well, Lepage's had a "have to have
2	it" kind of product in the bundle. NicSand is almost
3	more interesting, because it was real competition for
4	the contract, and I am not sure we have seen the last of
5	that case.
6	MR. JACOBSON: Well, it was a 12(b), so...
7	MR. VITA: Anybody else? Josh, Mary, anything
8	you would like to pose to the other speakers before
9	we --
10	MR. WRIGHT: I have one.
11	MR. VITA: Yeah, go ahead.
12	MR. WRIGHT: I maybe was being too sensitive to
13	one of the comments, so I heard it directed at me, but
14	Jonathan had mentioned that he --
15	MR. JACOBSON: Ripping competitors' racks off
16	shelves? Yeah.
17	MR. WRIGHT: So, I think you either
18	mischaracterized what I said, but since I didn't say
19	anything about the shelves, then maybe that's not it,
20	but to be clear, what the paper is about and what we are
21	arguing about in the paper is the economic analysis of
22	category management contracts, giving a procompetitive
23	explanation for why, under some conditions, the retailer
24	may want to delegate to the manufacturer the
25	responsibility of the shelf space allocation decisions.

71

1	That has nothing to do with the decision in Conwood.
2	What the point is about the decision in Conwood
3	is -- and I agree, and I am happy to say, court reporter
4	and everything, that I agree that ripping shelf space --
5	ripping displays down is bad, it is exclusionary. It
6	would be bad --
7	MR. JACOBSON: Makes no economic sense?
8	MR. WRIGHT: -- it would be bad if -- also if
9	the United States Tobacco employees sat out in the
10	parking lot with bats and said don't come in and bring
11	in product. All these things would be bad, but the
12	point is about whether or not there is anticompetitive
13	effect and whether or not there are any foreclosure
14	effects and whether or not the conduct was sufficient or
15	likely to generate anticompetitive effects.
16	I know I am to the right of you on the panel, so
17	I will use someone else. Professor Hovenkamp, in
18	Antitrust Enterprise, using the testimony in the record,
19	estimates the distribution cost increase as something
20	like 33 cents per store per month, and there is some
21	other evidence we talk about in the paper, but the idea
22	is that there is this other question about whether or
23	not there is a likelihood of anticompetitive effect and
24	that even in the case of really nasty, nasty, bad, wrong
25	conduct, we should be asking the question.

72

1	MR. VITA: Mary, do you have anything?
2	DR. SULLIVAN: Ah, no.
3	MR. VITA: Okay, Brandon, why don't we move
4	along then, and what we would like to do is put some
5	propositions up and get some reactions from the panel,
6	and I am going to go ahead -- I am going to read these,
7	they have to be read into the record, so let me just go
8	ahead and read the first one here, and this is a
9	quotation from Justice O'Connor's concurring opinion in
10	Jefferson Parish Hospital District Number 2 versus Hyde,
11	1984, and the statement is, "Exclusive-dealing
12	arrangements are analyzed under the rule of reason."
13	Let me just pose probably a simple question to
14	the panel, and this is more to the lawyers, I think.
15	Does this statement from Justice O'Connor's concurrence
16	in that case accurately summarize the law regarding
17	exclusive dealing? Richard and Joshua, Jonathan?
18	MR. STEUER: I think it does. I think that the
19	rule of reason is still a work in progress since Cal
20	Dental, and we will see what the content is in judging
21	these, but there really are three elements I think that
22	go into it with exclusive dealing. One is the nature of
23	the product and relationship, all the things that I
24	talked about. The second is, of course, the percentage
25	of the market once you have defined it that's

73

1	"foreclosed," and the third element is the duration, the
2	time period. So, I think those are the big moving parts
3	in a rule of reason analysis, and the nuances await the
4	development of the case law.
5	MR. JACOBSON: Yeah, I agree with that. I was
6	actually surprised, because this is also on the first of
7	the questions that you sent out to us yesterday, that
8	this would be perceived as controversial. I mean, the
9	law is fairly clear about this, certainly under Section
10	1, and I think Microsoft and Dentsply, properly read,
11	import this analysis into Section 2. The greater the
12	market power of the defendant, the lower the degree of
13	impairment of rivals you are generally going to require
14	before you see a price effect, but I do not think this
15	is a controversial proposition. So, I wonder what is
16	motivating the inquiry.
17	MR. O'BRIEN: We didn't necessarily think it was
18	controversial, but in this area where we are trying to
19	build some kind of consensus in terms of what we all
20	agree on, we thought we would start simple.
21	MR. JACOBSON: Well, I "concense" this.
22	MR. VITA: Josh, are you on board, too?
23	MR. WRIGHT: I third the motion.
24	MR. VITA: Let me follow up on that, then, and
25	ask again, and anybody can step in here, does anybody

74

1	think there are exclusivity arrangements that should be
2	per se illegal? And similarly, does anyone think there
3	are exclusivity arrangements that are always or nearly
4	always procompetitive and are thus appropriate
5	candidates for a safe harbor? Just if anybody has any
6	thoughts on that, you can step in.
7	MR. JACOBSON: Yeah, but dissent in the Harmar
8	case, four Justices saying that exclusive dealing
9	arrangements with multiple retailers are illegal because
10	Klors as originally understood is correct, but I do not
11	think anyone else believes that, and I think it would be
12	really wrong-headed to circumvent, you know, 30 years
13	now of rule of reason foray after Sylvania, to go back
14	to a per se rule on exclusivity here.
15	I think there are going to be safe harbors, but
16	they are basically going to be low market share safe
17	harbors and in a properly defined market, and the open
18	question in those cases is going to be, well, what if
19	the whole market is tied up with exclusives as in
20	Standard Stations? Do we really look just at the
21	defendant's share of the market as a screen? I think
22	the answer is yes, but I think it is a difficult
23	question.
24	MR. VITA: Anybody else?
25	MR. WRIGHT: Sure.

75

1	MR. VITA: Josh?
2	MR. WRIGHT: The first question I think was are
3	there any that should be per se illegal, no. And the
4	second question is with respect to safe harbors, and I
5	think in addition to the point about safe harbors for
6	exclusives that do not foreclose some significant share
7	of distribution, sort of foreclose trivial shares of
8	distribution, then that is an appropriate place for a
9	safe harbor.
10	And I know there is at least -- I mean, there is
11	not a consensus on this point about the duration of the
12	contracts, but I believe it is certainly the case that
13	short-term arrangements, like the ones we see in
14	slotting, six months in duration, may also be, though I
15	recognize this is subject to probably more debate, may
16	also be appropriate for safe harbors.
17	MR. STEUER: Some courts have misapplied the
18	term "exclusive dealing" to both exclusive selling and
19	exclusive buying. There is almost a safe harbor for
20	exclusive selling other than those rare arrangements
21	where one dealer has the exclusive for every brand there
22	is, and there have been a couple of cases like that.
23	In terms of real exclusive dealing, exclusive
24	buying, there is almost a safe harbor of a third coming
25	out of Jefferson Parish, talking about 30 percent.

76

1	There are some other contexts where 20 percent is surely
2	a safe harbor. I think that Jon is absolutely right,
3	that the tough issue is, well, if somebody has an
4	exclusive for 33 percent, but then there are two others
5	who have 33 percent and 33 percent, and so there is 100
6	percent exclusivity, that becomes more difficult, but
7	Jefferson Parish for practical purposes has introduced a
8	quasi-safe harbor of about a third.
9	MR. VITA: Okay, Howard?
10	DR. MARVEL: That is an awfully small harbor,
11	but on top of that I wanted to ask you about the
12	exclusive -- the exclusive -- which side did you put it
13	on, seller is --
14	MR. STEUER: Exclusive selling and sometimes it
15	is called an exclusive distributorship, "You will be my
16	only dealer in the State of Maryland" or something like
17	that.
18	DR. MARVEL: Yeah, but then turning that around,
19	how do you regard an agreement extracted by a dealer
20	like Toys 'R Us from seller -- a seller where he says,
21	you know, don't sell to my rival the same product that
22	you are selling to me. Is that okay?
23	MR. STEUER: It can be. Again, if it extracts
24	that from every manufacturer, that becomes increasingly
25	a problem. If Toys 'R Us were to enter into an

77

1	agreement with one manufacturer for one product and says
2	"I want to be the exclusive seller of this product," it
3	is rather limited what the impact is. In fact, I think
4	the decree that was finally negotiated specifically
5	provides for some limited exclusivity like that.
6	But if one chain were to become powerful enough
7	to sign up as the exclusive seller of all the toys for
8	all the major manufacturers, obviously everybody else is
9	frozen out, and I think there actually have been a
10	couple of examples like that.
11	DR. MARVEL: So, in Toys 'R Us, what happened,
12	if I recall, was that the Seventh Circuit of all people
13	said that the Toys 'R Us arrangement was not okay, and
14	that is because Toys 'R Us did have this sort of
15	monopoly position in the toy business, and it was
16	unassailable -- because of their unassailable position,
17	they really needed to protect the other poor souls like
18	Sam's Club from the depredations of Toys 'R Us. So --
19	is that right?
20	MR. JACOBSON: Well, another way to --
21	MR. STEUER: Well, Sam's Club or consumers. I
22	mean, the classic example, there was a wholesaler on an
23	island, I think St. Thomas, that was the sole
24	distributor for, it turned out, every single brand of
25	liquor, so that it basically created a bottleneck and

78

1	had monopoly at the distribution level, and to the
2	extent any of these examples approach that almost
3	textbook model, then you have a situation where
4	consumers really do not have other options at which to
5	shop for those particular products.
6	DR. MARVEL: So, is it an advantage to consumers
7	when Toys 'R Us contemplates getting out of the toy
8	business?
9	MR. JACOBSON: Because of WalMart? Look, there
10	were a lot of things going on in the case. One of them
11	was that the facts supported a finding of a horizontal
12	arrangement that was facilitated by Toys 'R Us, and I
13	think that is what concerned Judge Wood most --
14	DR. MARVEL: Right, absolutely.
15	MR. JACOBSON: -- in terms of the significance,
16	but looking at it purely on a vertical basis, at the
17	time there was a credible theory that it was raising
18	prices. Even though Toys 'R Us had a 20 percent market
19	share nationally, there were pockets of the country
20	where the share was in the high 40s, low 50s, and where
21	they were a must-have retailer for Mattel and Hasbro and
22	those other toy stores, and the result of this was that
23	the real, you know, the real discounters were cut off by
24	it, and you could make an arguable case that consumers
25	were paying higher prices as a result.

79

1	So, it was not -- it is not a crazy case. I
2	think it is a tough case, but I do not think it was a
3	crazy case.
4	DR. MARVEL: Well, I brought it up because it is
5	a tough case, but it is not a crazy case that what they
6	were doing was actually in the interest of consumers.
7	In fact, to have reasonably broad distribution of the
8	lines of the toy manufacturers -- and, of course, we
9	have also seen that not only has Toys 'R Us gone
10	belly-up and KB Toys and FAO Schwartz, but also the toy
11	manufacturers are rapidly fading into the sunset. Maybe
12	that is because only one Tickle Me Elmo was -- one
13	variety of Elmo was sold every Christmas at Sam's Club,
14	maybe not, but it does not appear that that industry is
15	a model of good health, and it may possibly be that that
16	is because a vertical restraint that was contributing to
17	not the monopoly behavior, but the good health of the
18	industry, was expunged.
19	MR. JACOBSON: Well, it may also be that our
20	analysis of monopsony power/buyer power is in its
21	infancy and that we really do not understand the
22	ramifications of WalMart, and I think that is the larger
23	issue, and I do not think anyone has a good answer to
24	that.
25	DR. MARVEL: I think that is right, because if

80

1	you look at Conwood, for example, and what Josh was
2	talking about, the Conwood case seems to me to have
3	turned in part upon the, shall we say, hyjinks of the
4	UST representatives who were trashing the Conwood
5	racks --
6	MR. JACOBSON: Right.
7	DR. MARVEL: -- but what it really turned on was
8	what was going on at WalMart, and that was a different
9	tale entirely. They wouldn't dare trash the racks at
10	WalMart, and so it kind of conflated those two things.
11	I mean, I have come up with a number of sort of
12	hair-raising anticompetitive activities that firms used
13	to engage in, and it is easy to come up with these
14	things, but that one is tough, because you start
15	conflating these things, and then you get a decision
16	that is made more on emotion than on what the economics
17	of it are.
18	MR. VITA: Let's go to the next slide, Brandon,
19	and let me just again read this, but this discussion
20	that Howard and Jonathan have been having I think sort
21	of leads into this next proposition and some of the
22	questions surrounding it. Let me just read it.
23	This is a quotation from Posner's Antitrust Law,
24	Second Edition, 2001, and in that book, Posner says, "I
25	propose the following standard for judging practices

81

1	claimed to be exclusionary: In every case in which such
2	a practice is alleged, the plaintiff must prove first
3	that the defendant has monopoly power...all the
4	plausible cases of exclusionary practices involve
5	defendants that have monopoly power."
6	And so let me pose two questions, two related
7	questions, you know, should monopoly power be a
8	requirement for challenging an exclusive dealing
9	arrangement under Section 1 of the Sherman Act and
10	Section 3 of the Clayton Act, and related to that is,
11	can exclusive dealing involving a non-monopolist result
12	in substantial lessening of competition?
13	And I think you two were already starting to
14	discuss that. Let me see if anybody else wants to have
15	any thoughts on that. Richard, Mary, Josh?
16	MR. STEUER: Well, clearly I think one of the
17	toughest areas is that space between 33 percent and 50
18	percent, because when you get above -- where you are in
19	the realm of Section 2 cases, the legalities change. I
20	know this means nothing to economists, but it certainly
21	does in terms of where you can get into court and
22	whether you can stay there.
23	The Microsoft case is an interesting example,
24	because there, in terms of browsers -- and I don't want
25	to dwell on this one case -- but certainly the share at

82

1	the time the case was brought was very low, and that may
2	explain why there was talk about monopoly power in
3	operating systems, but if you look at it purely as a
4	Section 3 type case and not searching for monopoly
5	power, but even at a low market share, was there a
6	danger -- an anticompetitive effect from the types of
7	exclusivity that was being entered into? Purely on the
8	numbers, you would say, no, the share is much too low,
9	and come back when it gets higher, but we all know where
10	that ended up.
11	MR. VITA: Well, let me ask this, and this may
12	be a question more for the economists, although the
13	lawyers are free to jump in, too.
14	Can we articulate or identify necessary
15	conditions in the downstream market that -- conditions
16	that are necessary for the exclusive dealing arrangement
17	to have an anticompetitive effect? Are there certain
18	things that have to be there before we have any ability
19	to infer anticompetitive consequences from an exclusive
20	dealing arrangement?
21	Josh, got any thoughts on that?
22	MR. WRIGHT: Sure. One -- I mean, let me make
23	sure I understand -- I understand the question.
24	MR. VITA: Yeah.
25	MR. WRIGHT: So, when you say competitive

83

1	conditions in the downstream -- you know, the downstream
2	market, so I am envisioning a manufacturer with
3	exclusive deals to a retailer --
4	MR. VITA: Think about that, that's a good
5	scenario.
6	MR. WRIGHT: That's what I would think of as an
7	example.
8	MR. VITA: Yeah.
9	MR. WRIGHT: I mean, substantial foreclosure
10	on -- I mean, the sort of well-known conditions from the
11	literature are that substantial foreclosure of the rival
12	so he can't achieve minimum efficient scale is a
13	necessary condition of most of these models, if not all
14	of these models, and so I think that that is -- you
15	know, in the legal analysis, we can have certainly, you
16	know, in the economics literature is a necessary but not
17	sufficient condition, and, you know, we know in the
18	cases, there are cases that end up on both sides. We
19	have a large foreclosure share but no liability because
20	of short duration or entry conditions or some such, and
21	so I think it is appropriate to use foreclosure as a
22	necessary but not sufficient condition.
23	MR. VITA: What about things like scaled
24	economies in the downstream -- when you talked about
25	scale economies, you were thinking about upstream, but

84

1	what about downstream?
2	MR. WRIGHT: So, in downstream, you can have --
3	there are cases where if you have large economies of
4	scale in distribution, you get -- you can have these
5	exclusionary effects as well.
6	MR. VITA: I mean, if there weren't substantial
7	scale economies downstream, or maybe some other factors
8	as well, do you think it would be possible in the kind
9	of long run or medium run for exclusive dealing
10	arrangements to have an anticompetitive effect? I mean,
11	why wouldn't -- you know, because if you don't have
12	substantial scaled economies and/or sunk costs at the
13	retailing level, why can't the -- supposedly the
14	foreclosed manufacturer get around the --
15	MR. WRIGHT: Right, so if you have -- at the
16	retail level you have -- I am going to frame this a
17	slightly different way, but if you have -- even if you
18	have the manufacturing scale economies but the retail
19	level you have free entry condition, then you are going
20	to have retailers who will re-align the supply
21	contracts, new entrants into the retailers who will
22	re-align the supply contracts, and so you need it at
23	some level, and the theory is you can do it with
24	economies of scale at the manufacturer level, but if you
25	have free entry at the retail level, I think that is

85

1	another problem for the exclusionary dealings.
2	MR. VITA: Jonathan, you looked like you might
3	have had something to add there.
4	MR. JACOBSON: No, I actually agree with that,
5	but it led into one of my sort of favorite topics in the
6	space, which is let's not talk about foreclosure,
7	because if we look at the percentage of distribution or
8	retail outlets foreclosed without examining entry, for
9	example, we may get a large number that's meaningless,
10	and that is why I think we are a lot better off if we
11	get rid of the word "foreclosure" and think about the
12	impairment of the rival, because that is the mechanism
13	that is going to lead to the consumer harm, not the
14	foreclosure, as such.
15	Foreclosure is a part of the analysis, but I
16	think it is only part of the analysis. You have to look
17	at the broader picture. Clearly there have to be
18	impediments to entry downstream.
19	And incidentally, I would agree with Posner's
20	book depending on the definition of "monopoly power."
21	You know, I think if you change it to market power, I
22	think, you know, a lot of people would subscribe to it.
23	I certainly would.
24	DR. SULLIVAN: Yes, I have one comment to make
25	on the -- following up on Josh's comment about free

86

1	entry in the retailing level. I agree that if there is
2	free entry in retailing, this is problematic for
3	theories of exclusion, because the excluded manufacturer
4	can more easily go to one of the new entrant retailers
5	to obtain distribution, but on the other extreme, if you
6	have, say, a monopolistic retailer, then I think that
7	the exclusive dealing arrangements, it is very hard to
8	prove that they would be harmful just because of the one
9	monopoly rent problem. So, I think you need to -- there
10	may be more potential for harm from exclusion in the
11	more intermediate market structures.
12	MR. VITA: Okay. Brandon, let's move on to the
13	next slide.
14	Here's another -- this is yet another quotation
15	from Justice O'Connor in Jefferson Parish Hospital
16	District Number 2 versus Hyde, and the proposition here
17	is, "Exclusive-dealing arrangement 'may be substantially
18	procompetitive by ensuring stable markets and
19	encouraging long-term, mutually advantageous business
20	relationships.'"
21	Let me put a couple of questions out. You know,
22	what are the -- empirically, what kinds of efficiencies
23	do the panelists perceive to be most likely to be most
24	significant in one of these exclusivity arrangements?
25	And think about this, you know, are there efficiencies

87

1	that are sometimes discussed maybe in the academic
2	literature in connection with exclusivity arrangements,
3	but in all likelihood, really aren't likely to exist or
4	likely to be very important empirically in real cases?
5	So, let me put that out there. Anybody --
6	DR. SULLIVAN: Yes, I will take that one just in
7	the sort of specialized area of slotting allowances. In
8	the academic literature, people make a big deal out
9	of -- one of the efficiencies of slotting allowances is
10	that it signals the product quality to retailers of
11	manufacturers' new products in cases where product
12	default is uncertain, and based on a lot of the
13	empirical studies that have been done by people in
14	marketing, that is simply not one of the efficiencies
15	that pops up, and I think the reason is there are quite
16	a few tools that manufacturers use to introduce their
17	products in addition to slotting allowances, and that
18	just -- so, I would feel comfortable ruling that out as
19	an efficiency, although there are plenty of other
20	efficiencies involved in slotting allowances.
21	MR. VITA: Howard?
22	DR. MARVEL: One of the cases that Richard
23	mentioned is the first nuanced case of exclusive dealing
24	I think was Beltone, and I think it is fair to say that
25	if there had not been some very un-nuanced evidence in

88

1	that case, that Beltone would have gone down in flames,
2	because by the time Beltone came up before the
3	Commission, its four principal rivals in that particular
4	channel that it was involved in had all met their
5	demise, and so Beltone was left as the monopolist --
6	thank you very much, FTC -- and at that point, they
7	didn't really have a good explanation for why they were
8	engaging in the exclusive dealing that they were
9	engaging in, but -- and so I don't see how they really
10	could have prevailed in that case unless there was this
11	evidence that was pretty clear that the companies that
12	had to give up the exclusive dealing practice had gone
13	belly-up.
14	So, in some ways John's paper talks about how
15	there probably is not a case that you can find where you
16	cannot determine that there are some advantages, but the
17	real difficult problem is to figure out how important
18	they are, and that is an incredibly difficult trade-off.
19	It is very hard to measure these things.
20	MR. VITA: Let me ask a follow-up on that point.
21	What significance, if any, should be given to observing
22	a challenged exclusive dealing arrangement in a similar
23	but somewhat more competitive market? So, you know,
24	that is sometimes an argument you make or you hear,
25	that, well, you know, this particular arrangement must

89

1	have some competitive benefits, because we see it over
2	here in these other markets that are structurally
3	competitive and where there is no plausible
4	anticompetitive theory of harm. How much -- how
5	powerful are those arguments and what weight should they
6	be given?
7	MR. JACOBSON: I think it is a much more
8	powerful argument if a small company is doing it than if
9	a large company is doing it in the same market. I think
10	looking at comparable markets and saying exclusive
11	dealing works efficiencies there, therefore they must in
12	this other market, really depends on how similar the
13	markets are. I would not make that leap without, you
14	know, a good deal of comparability evidence.
15	MR. VITA: Josh?
16	MR. WRIGHT: A related point, I mean, the nature
17	of the exclusive deal to facilitate some sort of
18	contract or performance, in the slotting example, again,
19	where the contract is over some sort of form of
20	promotion, and you see this a lot in exclusive dealing
21	cases where the underlying relationship between the
22	manufacturer and retailer relies on some sort of
23	promotional effort of the retailer and, in fact, is
24	contracted for, but the nature of performance in these
25	different markets varies a great deal, whether we are

90

1	talking about putting a product on an eye-level shelf
2	space or giving a product demonstration or some other
3	form of promotion.
4	So, the contracted-for conduct varies so much
5	market to market, I think the best you can make out of
6	seeing exclusive in a more competitive but different
7	market is sort of one of a cautious inference that we
8	generally know that exclusives can be procompetitive,
9	which I think there is not much disagreement on anyway.
10	MR. VITA: Okay.
11	MR. JACOBSON: I have a question for Mary. If
12	we renamed it payola, from payola to music leaders or
13	retail music program, do you think we would get a
14	different result?
15	DR. SULLIVAN: No. I think the people at FCC
16	and Elliott Spitzer would figure it out in a second.
17	DR. MARVEL: Why don't we call grocery store
18	slotting allowances payola?
19	DR. SULLIVAN: Well, I think we could, and one
20	thing you could do --
21	MR. JACOBSON: Because we would like to win the
22	cases.
23	DR. SULLIVAN: -- if the FCC regulated slotting
24	allowances, they would require the cashier at the
25	checkout counters to tell the customer each time he or

91

1	she was buying a product for which a slotting allowance
2	had been paid, then say, do you still want to buy it?
3	MR. WRIGHT: Well, as funny as that is,
4	California had proposed at one point -- I think it is
5	still kicking around in committee --
6	MR. JACOBSON: No, it was killed.
7	MR. WRIGHT: It was killed now?
8	MR. JACOBSON: Yeah.
9	MR. WRIGHT: Senate Bill 582, which would have
10	made -- it would have been illegal for -- essentially a
11	retailer would have to tell Pepsi exactly what Coke was
12	paying in terms of its promotional allowances, in terms
13	of the slotting fees, and if you conceive of these
14	things, these payments, as I do, as part of the
15	competitive process, I mean, this is a statute that is
16	a -- it is, you know, a legislatively enforced
17	collusion, right? And so it is silly, but, you know,
18	not silly enough to write down in a bill.
19	DR. MARVEL: Was it going to be the California
20	Raisin and Coca-Cola board? Is that --
21	MR. JACOBSON: It was proposed by a coalition of
22	the same people who represented the plaintiffs in the
23	Gruma case and the Harmar case. I mean, it was serious,
24	and it did get some traction, but it got killed fairly
25	early on in committee.

92

1	MR. VITA: Okay, let's move on then. The next
2	proposition is from Dennis Carlton from his article in
3	the Antitrust Law Journal, "A General Analysis of
4	Exclusionary Conduct and Refusal to Deal -- Why Aspen
5	and Kodak Are Misguided," and Carlton's proposition is
6	as follows:
7	"In the presence of scale economies, exclusive
8	dealing can be a way of depriving Firm 2 (or its
9	distributors) of the necessary scale to achieve
10	efficiencies, even though, absent the exclusivity, Firm
11	1 and Firm 2 would both be large enough to achieve
12	efficiency."
13	So, two related questions for the panel. One,
14	do you agree with Dennis' statement, and secondly, other
15	than its potential to deprive competitors of scale and
16	the resulting effect on prices, are there any other
17	theories of harm from an exclusivity arrangement that
18	should be the subject of antitrust concern?
19	DR. SULLIVAN: I will try the second question.
20	There is a theory -- and this is one I referred to in my
21	presentation -- by Greg Shaffer, a 2005 theory, and he
22	had a theory of exclusion in which scale economies did
23	not play a role, but what was going on is the retailing
24	segment was very, very competitive, and essentially
25	retailers, without exclusion of a manufacturer, would

93

1	earn almost no profits because their segment was so
2	competitive, and they could easily be coerced into going
3	along with an exclusivity deal that would exclude one of
4	the manufacturers because it simply would increase the
5	industry profits, and he developed conditions under
6	which this was true. One might argue that that would be
7	fairly unusual, but it -- you know, it is there.
8	MR. VITA: Anybody else? Dan, did you want to
9	add something?
10	MR. O'BRIEN: I would just like to ask, Mary,
11	following up, in that kind of a theory, if a
12	manufacturer could secretly get to a -- get with a
13	retailer, okay, assuming that everybody else was being
14	coerced into this exclusive with the manufacturer, and
15	negotiate something on the sly, wouldn't they be able to
16	undercut what, you know, the monopoly price that was
17	presumably being set by the other guys?
18	DR. SULLIVAN: I think so, and I think there was
19	something in particular about the nature of the game
20	that Greg set up that allowed him to get this outcome,
21	so I agree that might be -- it might not be that
22	problematic in reality.
23	MR. STEUER: There are a lot of assumptions in
24	here obviously. It makes a huge difference whether the
25	exclusivity is with end users and for how long. If this

94

1	is simply competition for the contract, clearly if one
2	manufacturer can get exclusive arrangements with the
3	bulk of the end users and freeze out the other, that is
4	going to have a profound impact, but if the second
5	manufacturer can survive long enough to go and bid the
6	next time and try to get the contract back, that is very
7	different. We do have some situations in defense, for
8	instance, where there is only going to be one winner of
9	these contracts. They are always exclusive, and yet you
10	do have some back and forth bidding as long as both
11	companies can survive. Here, I presume the assumption
12	is, with economies of scale, that there is a danger that
13	one of the companies disappears off the face of the
14	economic map.
15	MR. JACOBSON: I think this identifies a case --
16	there are certainly exceptions, as Richard points out,
17	but I think this identifies the exclusive dealing case
18	that you ought to worry about, you know, if these
19	conditions are holding, this is the case you ought to
20	worry about. There may be other cases you ought to
21	worry about. There may be cases where this is not a
22	problem because it is competition for the contract, but
23	in terms of our analysis, this is where I think we
24	should focus most of our resources.
25	I would add that this is an excellent article,

95

1	although Aspen I do not think was misguided, although
2	that is debatable, and Kodak was clearly correctly
3	decided.
4	MR. VITA: Whoa.
5	MR. O'BRIEN: Okay, we will try -- if we have
6	got time -- we have got time for one more, I think. Oh,
7	one more on this? On this proposition or one more
8	proposition?
9	MR. VITA: One more proposition, I think.
10	MR. O'BRIEN: I think we can go a little longer.
11	MR. VITA: All right, Brandon?
12	Okay, this next proposition is from Herbert
13	Hovenkamp, Antitrust Enterprise, 2005, and I will read
14	it.
15	"Exclusive dealing is a rule of reason offense,
16	requiring a plaintiff to show that the defendant has
17	significant market power, that the exclusivity agreement
18	serves to deny market access to one or more significant
19	rivals, and that market output to consumers is lower (or
20	prices higher) as a result."
21	A couple of questions for the panel. As to
22	significant market power or some indicator of
23	significant market power, is there or should there be a
24	safe harbor? And does anybody have an -- you know, it
25	says here in my notes that Jonathan in his writing

96

1	suggests courts apply a 40 percent market share safe
2	harbor, and if that -- you know, is that actually true,
3	and does anybody have an alternative minimum requirement
4	that they would prefer?
5	So, let me put those two out, those two
6	propositions out there and see what the panel thinks.
7	MR. JACOBSON: Well, I generally agree with what
8	I said.
9	MR. VITA: Glad to hear that.8
10	MR. JACOBSON: I think this is a pretty good
11	quote. I think "market access" needs a little bit of
12	definition, because I do not think you need -- this was
13	one of the other questions that we had talked about
14	before the program -- I do not think you need total
15	foreclosure. Again, I think the test needs to be the
16	degree of impairment of rivals. So, as long as denying
17	market access is read in that context, I think this is a
18	pretty good analysis.
19	I think 40 percent is a pretty good rough
20	screen. I think Richard's correct to point out that
21	Jefferson Parish is a 30 percent number, but it does not
22	say anything about a screen here or there, but if you
23	look at the subsequent cases, you are not going to find
24	any where the defendants have liability with less than
25	40 percent unless you consider Toys 'R Us an exclusive

97

1	dealing case, and there, you know, there were
2	extenuating circumstances given the horizontality of the
3	agreement.
4	MR. STEUER: And the term in here "significant
5	rivals" is significant, because it really raises the
6	question, who should have a cause of action here? At
7	some point, if there is ample competition in a market
8	and there is exclusive dealing going around, there may
9	be some marginal players who claim that they are being
10	excluded, and those can be emotionally appealing cases
11	in terms of jury appeal, and yet in terms of what the
12	actual effect is on the market, it may be very marginal
13	indeed, and there are not very clear tests right now as
14	to who should be able to bring a claim.
15	MR. O'BRIEN: If I could follow up with that,
16	John, earlier you had said that one of the areas in
17	which there was an agreement, you listed four points,
18	one of which was we want to prevent the enhanced -- you
19	know, practices that enhance market power. I am
20	wondering if you would agree with the last part of this
21	proposition, which is that plaintiffs have to show to
22	successfully bring a case that market output goes down
23	and/or prices go up.
24	MR. JACOBSON: Well, I think what he means is
25	that market output is likely to go down, and if you show

98

1	there is a significant enhancement or creation of market
2	power, I think you have done that. So, I do not think
3	this is inconsistent with that proposition.
4	MR. VITA: Okay, let's move on then.
5	This next proposition is from United States
6	versus Microsoft, the D.C. Circuit en banc decision.
7	The quotation is as follows:
8	"If the monopolist's procompetitive
9	justification stands unrebutted, then the plaintiff must
10	demonstrate that the anticompetitive harm of the conduct
11	outweighs the procompetitive benefit."
12	A couple of questions, and again, this may be a
13	little more for the economists, but anybody can step in.
14	First of all, does economics supply tools that
15	would assist courts in making this kind of assessment,
16	and do courts have the ability to apply these kinds of
17	tests?
18	Let me stop right there and see what the
19	reaction is from the economists on the panel.
20	DR. MARVEL: How about no?
21	MR. VITA: Say again?
22	DR. MARVEL: Do the courts have the tools? No.
23	MR. VITA: Actually, the proposition was, can we
24	as economists supply tools that courts could use? I
25	mean, what kind of analysis, if any, can we provide that

99

1	will allow noneconomists to make the kind of
2	determination that the Court called for in this case?
3	DR. MARVEL: I think that you really need to be
4	very careful about if you show anticompetitive harm, it
5	is pretty clear that you have got anticompetitive harm,
6	then I guess once you have gotten to that point, unless
7	convinced that the procompetitive benefits you are
8	trying to demonstrate will be easily enough measured and
9	ready available in such a way as to make it possible for
10	the courts to do the trade-off, I just think they are
11	awfully hard to prove what they are.
12	So, if you can really show that somebody is
13	locked out by the nature of the arrangement -- and that
14	means from the market, that does not mean from the
15	channel that the manufacturer in question controls, but
16	from the market as a whole -- then it is going to be
17	hard to do this trade-off, but if you have got the
18	anticompetitive harm and people are absolutely convinced
19	that it is there, then I think that that might be
20	enough.
21	MR. VITA: Yes, Josh?
22	MR. WRIGHT: Well, I think in this particular
23	quote, we have to -- there may be differences with
24	respect to what economists can do before and after -- in
25	the first and second clauses, right? The economist

100

1	might have tools to supply with respect to understanding
2	a monopolist's procompetitive justifications. Something
3	we can do is understand why we might see exclusives,
4	understand why conduct might be procompetitive, and the
5	conditions under which those explanations are likely.
6	That is something we can do and should be doing.
7	It is a lot tougher, the challenge of doing the
8	balancing is much tougher, and I guess the part that is
9	not in this quote is that the first step of requiring
10	the plaintiff to show the likelihood of some
11	anticompetitive effect is also an area where economists
12	can contribute by explaining the conditions for
13	anticompetitive effects are either satisfied or they are
14	not.
15	MR. O'BRIEN: Do you want to follow up, John?
16	MR. JACOBSON: I mean, this is what my article
17	is all about, so I do not want to leave this one
18	untouched.
19	A, most cases do not reach the level where you
20	need balancing. The number of cases where you really
21	need to balance it are few and far between. Usually the
22	case will fail because a prima facie case of
23	anticompetitive effect will not be shown. If that is
24	shown and the defendant shows a significant
25	justification, usually the plaintiff gives up at that

101

1	point. So, it is a very rare case that requires
2	balancing.
3	But if balancing is required, I think we need to
4	do it, and to say -- to throw up our hands and say it is
5	too complicated is just completely the wrong answer. We
6	do it every day. This building is filled with people
7	doing that in merger cases. It is done at the Justice
8	Department in merger cases all the time. This is
9	exactly what we do. So, to say that we are not going to
10	do this, it is too complicated, we might as well just
11	get rid of antitrust, because this is the guts of what
12	hard antitrust cases are all about, and we not only want
13	to do this, but we have to do it. This is one issue I
14	feel very strongly about.
15	MR. O'BRIEN: So, I wanted to follow up with
16	Howard, and, John, you may want to chime in on this,
17	too. You are concerned that if we can establish that
18	there may be an anticompetitive effect, that it is often
19	very hard for defendants to come in and argue, well, no,
20	in fact, there are efficiencies and that they offset the
21	anticompetitive effect, and I --
22	DR. MARVEL: No, what I am saying is that if you
23	can really show anticompetitive harm and --
24	MR. O'BRIEN: That may or may not be offset by
25	efficiencies, okay, so that is what I am saying. It may

102

1	or may not be offset, and what I took you to be saying
2	was that --
3	DR. MARVEL: That would make it really tough
4	for -- once you have a compelling demonstration of
5	anticompetitive harm -- and that is compelling for me,
6	not for you --
7	MR. O'BRIEN: Right.
8	DR. MARVEL: -- then I am not so sure that -- it
9	reminds me of the original merger guidelines when they
10	did not allow efficiencies as a defense, and I do not
11	think that that was absolutely nuts. So, if there is a
12	strong demonstration of anticompetitive harm -- and that
13	is not just locking up a channel, that is locking up the
14	market -- then I am not sure how much balancing I want
15	to do at that point.
16	MR. O'BRIEN: I see.
17	MR. JACOBSON: It is a rare case, Dan, it is a
18	rare case where you need to do this, but there can be,
19	at least in theory -- I will tell you, I have never seen
20	one -- but there can be one, at least in theory, where
21	the effect of the exclusives is to create a market
22	structure such that the defendant can raise prices to
23	some extent.
24	However, there may be sufficient dealer focus as
25	one traditional efficiency or other effects that overall

103

1	output of the product is increased. Think about your
2	resale price maintenance cases, the same -- it is the
3	same type of analysis, and if you can show -- first of
4	all, the burden is on the plaintiff, not the defendant,
5	but if the defendant can put in evidence to say that
6	notwithstanding the price increase, we are going to have
7	a significant overall market output effect that is going
8	to be procompetitive, I think you have got to entertain
9	that defense, and then I think you have got to see
10	whether that is true at the end of the day. Is the net
11	effect going to be to increase output or not?
12	MR. O'BRIEN: I guess I -- I am sorry.
13	DR. MARVEL: I think maybe if I can go, John's
14	point, I think part of the disagreement with -- the
15	implicit disagreement here is in my determination of
16	what constitutes an anticompetitive effect, because I
17	certainly would not agree to that parenthetical remark
18	that Hovenkamp had that said that prices are higher than
19	they would have been if the restraint was taken away.
20	Well, you cannot do that, because all of these
21	explanations talk about setting up a property right that
22	allow you to get a return on your investment which could
23	very well take the form of, you know, if you shift up
24	the demand curve, you are going to get a higher price
25	and greater output. If you get more output, end of

104

1	story. If it is a higher price, that does not really
2	tell you much of anything, and so that is I think part
3	of what we are -- we may be agreeing, somehow have a
4	different setup.
5	MR. O'BRIEN: So, following up on that, Howard,
6	I am curious how you feel about something like the no
7	economic sense test as a way to, you know, ask is there
8	a plausible efficiency rationale and, you know, maybe
9	short-circuit this balancing some.
10	DR. MARVEL: Sorry, but I -- it hurts me, but I
11	would have to agree with John on that one. I do not
12	like the test.
13	MR. O'BRIEN: Okay. Why don't you like the
14	test?
15	DR. MARVEL: I think your explanation is that
16	there is always economic sense in these practices, and I
17	think that that is right, that there will always be some
18	plausible argument that could be made. Unless we are
19	talking about gunning down your rivals or some such,
20	anything short of that, you are probably going to be
21	able to come up with some plausible argument on behalf
22	of that.
23	MR. JACOBSON: One convert, not a bad morning.
24	MR. VITA: Well, with that, then, we will bring
25	the morning session to a close. I would like to thank

105

1	the panelists. This was a really great discussion, and
2	I think everybody got a lot out of it. So, thanks very
3	much.
4	(Applause.)
5	(Whereupon, at 12:19 p.m., a lunch recess was
6	taken.)
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

106

1	AFTERNOON SESSION
2	(1:31 p.m.)
3	MR. O'BRIEN: Okay, let's get started. Well,
4	welcome to the second exclusive dealing panel of the day
5	in what is part of our ongoing series of public hearings
6	on single-firm conduct. My name is Dan O'Brien. I am
7	the Chief of the Economic Regulatory Section at the
8	Antitrust Division, and I will be moderating this
9	session along with Mike Vita, who is the Assistant
10	Director in the Economics Group, the Bureau of Economics
11	at the Federal Trade Commission.
12	The Department of Justice and the FTC are
13	jointly sponsoring these hearings to help advance the
14	development of the law concerning the treatment of
15	unilateral conduct under the antitrust laws.
16	Transcripts and other materials from the prior sessions
17	are available on the DOJ and FTC web sites, and I just
18	wanted to advertise that upcoming panels include a panel
19	on bundled loyalty discounts on November 29th, obviously
20	a practice that is somewhat related to exclusive
21	dealing, which is the topic for today, and then there is
22	a panel on misleading and deceptive conduct on December
23	6th.
24	So, today's session concerns the law and
25	economics of exclusive dealing. It was 40 years ago in

107

1	the Brown Shoe case that the Supreme Court made a very
2	strong statement against exclusive dealing, asserting
3	that it conflicts with the central policy against
4	contracts that take away the freedom of purchasers to
5	buy in an open market.
6	Since that time, the treatment of exclusive
7	dealing by the courts has changed fairly dramatically
8	over time, and the economics of exclusive dealing has
9	progressed, identifying both procompetitive and
10	anticompetitive aspects of the practice depending on a
11	range of circumstances.
12	We have a very distinguished group of panelists
13	here this afternoon to talk about these developments and
14	the current state of affairs from both the legal and
15	economic perspectives. My goals from today's panel are,
16	first, to highlight some areas hopefully where there is
17	some consensus on the effects of exclusive dealing and
18	how to treat it, but also maybe identify questions that
19	remain unsettled so we can have some consensus about the
20	questions that need to be addressed as we move forward.
21	So, before introducing the panelists, I just
22	wanted to thank my colleagues at the FTC and at the
23	Antitrust Division, particularly June Lee and the
24	economics staff at the Antitrust Division and Joe
25	Matelis in Legal Policy. The two of them together did a

108

1	lot of the work in putting together this panel.
2	The organization of the panel is going to be as
3	follows: We have four panelists. They will give
4	presentations of approximately 15 minutes. Then we will
5	take a short break. Then the panelists will have a few
6	minutes to respond to the other presentations if they so
7	desire, and then we will have a moderated discussion,
8	and we can go until around 4:00 p.m.
9	So, the order of the panelists, in case people
10	are wondering, will be Steve Calkins first, Tad Lipsky
11	second, Joe Farrell and then Ben Klein. So, let me
12	introduce Stephen Calkins. He is our first speaker.
13	Stephen Calkins is Professor of Law and Director
14	of Graduate Studies at Wayne State University Law School
15	where he teaches courses and seminars on antitrust,
16	trade regulation, consumer law and torts.
17	From 1995 to 1997, Steve served as General
18	Counsel of the Federal Trade Commission. Steve lectures
19	widely throughout the U.S. and abroad, most recently in
20	Europe and New Zealand. He has authored many
21	publications on competition and consumer law and policy,
22	and he has served on the editorial boards of well-known
23	journals in antitrust.
24	Stephen?
25	DR. CALKINS: Thank you. Thank you for the

109

1	introduction. What was not said is that I am actually
2	the most novice of all the people who are speaking here
3	today. I mean, you go over everybody else, and they
4	have been an expert witness in one or more of the
5	leading cases, they have litigated one or more of the
6	leading cases. Richard Steuer, in the previous session,
7	got up and proceeded to point out that he had published
8	three articles specifically on exclusive dealing. I
9	have never been an expert witness on exclusive dealing,
10	I have never litigated, I have never done an article
11	about exclusive dealing, as such, you know, we are
12	talking about somebody who is just not in the same
13	ballpark. So, with great humility, let me just tell you
14	that I am trying to sort out my own thinking and to
15	learn from all these geniuses.
16	To do that, we need to start somewhere, and so I
17	found one interesting case that I thought I would begin
18	just looking at a little bit, and here is a court
19	opinion that talks about how exclusive dealing "may well
20	be of economic advantage to buyers as well as to
21	sellers, and thus, indirectly of advantage to the
22	consuming public," and these advantages may often
23	explain why there are exclusive dealing contracts, and
24	if you wanted to go and understand whether they were
25	harmful or beneficial, you would look at a series of

110

1	tests.
2	You would look at "evidence that competition has
3	flourished, despite use of the contracts," or you would
4	look at the conformity of the length of their terms to
5	the reasonable requirements of the field of commerce, or
6	you would look at the status of the defendant as a
7	struggling newcomer or an established competitor or the
8	defendant's degree of market control, and you would go
9	through all this sort of stuff, but the opinion goes on
10	and says that to do this would just be extremely
11	difficult and to sort everything out would be an immense
12	challenge and, using words very similar to sort of the
13	basic sort of Areeda Hovenkamp mantra, we need to have
14	tests that are administerable by courts, we need to have
15	rules that can be enforced without wasting a lot of
16	societal resources on hopelessly complex litigation that
17	can't lead to any predictable outcomes, and so for
18	reasons of administrative efficiency, exclusive dealing
19	contracts should almost all be illegal, because this was
20	the original Standard Oil/Standard Stations case with
21	those thoughtful observations about the procompetitive
22	benefits of exclusive dealing, but the conundrum, the
23	difficulties, of litigating.
24	So, when I sat down and took a look to start my
25	sort of thinking about this and went back in time, I

111

1	said, golly, what an interesting beginning place, and I
2	then decided to pull out key dates in exclusive dealing
3	history, and we began with the classic Supreme Court
4	cases, which have been reviewed a little bit in the
5	morning session, and I will not mention them except that
6	Standard Oil you know, Brown Shoe was just referenced,
7	the classical Supreme Court cases were certainly
8	important moments in exclusive dealing history.
9	That led us to the key year of 1977 when all of
10	antitrust, as we know it, changed with Continental TV,
11	and then along came Robert Bork and the antitrust
12	paradox -- actually, along came all of the Chicago
13	School -- but Bork in particular is associated with
14	exclusive dealing, because he said so emphatically that,
15	by golly, there is only one monopoly profit. Exclusive
16	dealing cannot increase a monopolist's monopoly profit,
17	and so, therefore, "if Standard finds it worthwhile to
18	purchase exclusivity, the reason is not the barring of
19	entry but some more sensible goal such as obtaining the
20	special selling effort of the outlet," emphatically
21	saying that one cannot increase the profit of the
22	monopolist, and so there must be a procompetitive
23	justification, and those Supreme Court cases were just
24	dead wrong, a really clarion call for a different way of
25	looking at exclusive dealing.

112

1	As mentioned in the previous session, that call
2	was picked up first in the courts or the adjudicative
3	bodies in the Beltone Electronics opinion, where the
4	Court specifically relies on Bork and the antitrust
5	paradox to take a different approach to exclusive
6	dealing, the Federal Trade Commission, leading the way
7	to a new day of exclusive dealing decision-making, even
8	if we learned in the last session at the cost of having
9	sacrificed four of the five competitors, but
10	nonetheless, having led the way, that was followed
11	shortly thereafter by Jefferson Parish. Of course, it
12	is always cute, we refer to the Jefferson Parish
13	exclusive dealing holding, and it wasn't a holding at
14	all. It was part of the concurrence of Justice
15	O'Connor, but we all think of it as the holding from
16	Jefferson Parish where she emphatically said exclusive
17	dealing is judged more permissively than tying, it is
18	rule of reason, and "exclusive dealing is unreasonable
19	restraint on trade only when a significant fraction of
20	buyers or sellers are frozen out of a market by the
21	exclusive deal."
22	And since then, if you look at things that have
23	happened and you sort of parade through the exclusive
24	dealing cases that we know, which I throw up on the
25	screen in front of you or I throw up more of the

113

1	exclusive dealing cases that we know or I throw up more
2	of the exclusive dealing cases that we know, the one
3	great unifying principle is, of course, that the
4	defendant always wins. There are a few exceptions, but
5	overwhelmingly, the judicial treatment of exclusive
6	dealing ever since Beltone Electronics came down has
7	been that defendants win these cases, and you can find
8	support in the case law for all sorts of pro-defendant
9	propositions, with exclusive dealing being strongly
10	presumed to be legal if there is a market share of less
11	than 40 percent, if the restraint is of less than a
12	year, the contract is of less than a year, if the
13	contract is easily cancellable, if we do not have a
14	complete and total foreclosure, see the words in
15	Jefferson Parish, if there are no entry barriers, and
16	on, there are probably other ones as well, a whole
17	series of different principles, standards under which
18	defendants have won these cases, and that's a whole lot
19	of the exclusive dealing story, and then there is the
20	"but" part of the whole thing that makes our life
21	slightly interesting here.
22	There are three things to mention. The first,
23	the post-Chicago literature, I have reason to suspect,
24	although I did not look at his slides, that Joe Farrell
25	will reference a little of this, and it can be done in

114

1	all sorts of wonderful mathematical sophistication. I
2	think of the lesson as a common sense story of
3	collective action.
4	There was recently a case that Tad knows dearly,
5	the Coca-Cola case just decided by the Texas Supreme
6	Court. I do not know anything about the facts of that
7	case, and I have no opinion on the case. I do not know
8	what happened down there, but one of the things that
9	allegedly happened was that Coca-Cola paid retailers not
10	to allow 7-Up in its stores, and if you think about that
11	for a minute, you know, it sort of sets out the
12	collective action story very crisply. Why would a
13	retailer agree not to carry 7-Up when it knows that if
14	in the long run there is no 7-Up, that is probably bad
15	for retailers? And the answer is, of course, that if a
16	payment goes to a single retailer, that single retailer
17	can collect the payment knowing that its excluding of
18	7-Up is not really going to make a difference in the
19	long run, and you do not have all the retailers getting
20	together and agreeing that they will resist Coca-Cola,
21	because that would be illegal under the antitrust laws,
22	and so each separate retailer looking at its individual
23	self-interest can quite reasonably say, I will agree not
24	to allow 7-Up in my store, even though in the long run,
25	that is against the collective interests of all of them,

115

1	and it is because of that kind of a collective action
2	problem that exclusive dealing can sometimes harm
3	competition in the long run because one can have an
4	exclusive dealing arrangement that helps someone today,
5	with all the benefit going to that one entity, in the
6	long run, there is harm, but the harm is shared widely,
7	and so, therefore, you have a mismatch between the
8	benefit of the harm, a collective action problem, and
9	therefore, mischief can be worked.
10	Two cases have come along that have sort of set
11	out the -- sort of the other ways of thinking about
12	exclusive dealing, being Microsoft and Dentsply. People
13	in this room know those cases far more than I do, but
14	just mentioning a couple of points quickly, Microsoft
15	is -- you can find several different points in the
16	Microsoft opinion on exclusive dealing. This is one
17	where the District Court had said that there must be
18	complete and total exclusion before there is a
19	violation, and the Court of Appeals wrote that "even
20	assuming the holding is correct," and went on to say
21	there could still be a violation, thereby suggesting
22	that that holding may not be correct.
23	It went on and said there could be a violation
24	because there is a different standard under Section 2
25	than under Section 1, and even if something might be

116

1	lawful under Section 1, it could be unlawful when
2	engaged in by a monopolist. The Court asked rather
3	tough questions about the justifications for the
4	practices going on there, specifically saying that with
5	respect to one practice, where 14 of the 15 top Internet
6	access providers had contracts to work only with
7	Microsoft, the justification was to keep them focused on
8	Microsoft's product, "which is to say it wants to
9	preserve its power in the operating system market, that
10	is not an unlawful end, but neither is it a
11	procompetitive justification," thereby raising nice
12	questions about the difference between a benefit to the
13	seller and a benefit that qualifies as a procompetitive
14	justification.
15	Also of interest to the Microsoft case is we had
16	a very economically sophisticated court unable to resist
17	quoting some language indicating subjective intent.
18	"Kill the cross-platform Java by growing the polluted
19	Java market," so on and so forth, finding some comfort
20	in the words that business people had used to describe
21	what they were doing, and then finally being troubled,
22	even though we did not have total exclusion. So, we
23	have a whole series of interesting points that come out
24	of the Microsoft case.
25	In the Dentsply case, what did we have in

117

1	Dentsply? You had something where you had an at-will
2	contract, and yet the Court of Appeals said that was not
3	reason for the defendant to prevail, because
4	realistically, wholesalers are not going to give up $22
5	million in sales in order to pick up $200,000, and so an
6	at-will contract does not really give a new entrant
7	realistic access to the market. So, also, there was
8	talk about monopoly maintenance as a separate kind of
9	problem, and once again, we had reference to subjective
10	intent evidence.
11	So, where am I at that point in terms of, as I
12	end, little lessons that I draw from my sort of going
13	over things, and they are very tentative, because I
14	really have not thought these things through all the
15	way. I am learning, okay, but tentative things that I
16	might throw out as propositions.
17	One, it should be possible for a short-term
18	contract or contract that is cancellable still to be
19	found to be unlawful. It should be possible for there
20	to be illegality without total exclusion. Section 2
21	standards should be tougher than Section 1 standards.
22	It does not make sense to take all of the teaching of
23	Section 1 exclusive dealing cases and then import them
24	unthinkingly into the world of Section 2. If you have a
25	firm with a 75-80 percent market share and entry

118

1	barriers and lots of power, it ought to be tougher than
2	on a smaller, less powerful firm.
3	I hesitantly think that it is -- this will not
4	be popular with some of my panelists -- sometimes it is
5	interesting and possibly informative, if done very
6	carefully, to look at subject intent evidence to help
7	you sort through these difficult things. Clearly it
8	makes sense to scrutinize the procompetitive
9	justifications that are being offered up in a case that
10	otherwise looks troubling. The classic procompetitive
11	story is that the manufacturer has expended resources to
12	bring a consumer into the store who will then be bait
13	and switched off to another product. Well, you know, do
14	the facts fit that story or not? In Dentsply, the Court
15	thought they did not fit that story but went on to try
16	to really sort of sort through what is the
17	justification. It should not be enough just to say it
18	is a nonprice vertical restraint.
19	I personally would not think that one should
20	require a plaintiff to prove that prices have increased.
21	I mean, think again about your classic exclusive dealing
22	situation would be something where we are trying to
23	cause problems in the future. Go back to my Coke paying
24	to have 7-Up not around. The reason to do that is so
25	that things will be better for Coca-Cola in year two or

119

1	three or four or five, and one can have a lessening of
2	competition without prices today being affected. The
3	hard question here is the long-run competitive effects,
4	though, can't be a complete defense to say that current
5	prices have not gone up.
6	So, also we would say that the legal standard
7	really does matter in these cases. Going back to
8	previous sessions that you have had, you heard a lot
9	about the no economic sense test in the last session.
10	Another standard that can make a big difference in
11	exclusive dealing cases is whether you choose to adopt
12	the Posner "Exclude an equally efficient firm" test.
13	Were you to adopt that, which I would not favor, that
14	would make it much harder for a plaintiff to win an
15	exclusive dealing case.
16	And finally, in closing, pretty much on time, it
17	is interesting as you survey the landscape that there is
18	a whole lot of theory, not a great deal of empirical
19	evidence, and so I hope that this program, if nothing
20	else, inspires some people to go out there and get their
21	hands dirty and bring forth more empirical evidence.
22	Thanks very much.
23	(Applause.)
24	MR. O'BRIEN: Okay, our next speaker is Tad
25	Lipsky. Tad is a partner at Latham & Watkins and a

120

1	former Deputy Assistant Attorney General at DOJ. Tad's
2	30-year legal career has been devoted mainly to
3	antitrust, and it spans virtually every facet of
4	competition law.
5	From 1981 to 1983, Tad served as Deputy
6	Assistant Attorney General at DOJ under William Baxter.
7	Following government service, Tad developed a broad U.S.
8	and international antitrust practice, successfully
9	managing a variety of important antitrust matters.
10	As chief antitrust lawyer for the Coca-Cola
11	Company from 1992 to 2002, Mr. Lipsky conducted and
12	supervised competition matters before courts and
13	antitrust authorities in the U.S., Canada, the EU, EU
14	Member States, and dozens of other jurisdictions. He is
15	a frequent author and speaker on antitrust topics.
16	Tad?
17	MR. LIPSKY: Thank you very much. Until a few
18	moments ago, I had forgotten how stupid it was to follow
19	Steve Calkins to the podium, because he knows more about
20	whatever he speaks about than anybody else and expressed
21	his interesting views so trenchantly and with such great
22	humor that that is a very tough standard, but I will do
23	my little bit and see if we can find something to agree
24	on. I think we can find a few things to disagree on,
25	and we will see where it goes.

121

1	Exclusive dealing is a very elastic label. It
2	applies to a lot of different kinds of things. We have
3	already heard mention of the fact that tying, certain
4	kinds of bundling and price discounting can have effects
5	very similar to exclusive dealing, and therefore, when
6	you talk about exclusive dealing, you also need to be
7	considering a bunch of its very, very close relatives,
8	and so we are talking about implicitly, at least, a very
9	broad category of business conduct and competitive
10	phenomena.
11	Now, on the plus side, for our policy evaluation
12	of exclusive dealing, it has never been a per se
13	offense, which is a very good thing. It is a little
14	like saying, well, in Eastern Europe, they have a little
15	better luck re-adopting capitalism, because they were
16	capitalists within living memory, whereas in the old
17	Soviet Union, in the heart of Mother Russia, that was
18	not the case, and so there is no great body of learning,
19	there is no familiarity in the culture, and similarly,
20	with exclusive dealing, although it is true that back in
21	the Standard Stations days and when we were dealing with
22	the International Salt comment, that under Section 3 of
23	Clayton, you could condemn exclusive dealing either if
24	the defendant had market power or if there was not an
25	insubstantial amount of foreclosure, that is coming

122

1	within an eyelash of saying it is per se, but we never
2	quite got there.
3	There was always a little bit of procompetitive
4	culture left in exclusive dealing, and so -- as a matter
5	of fact, even in the dark ages, between the decision in
6	Schwinn, all vertical agreements are illegal per se,
7	until the release from bondage in 1977 with Sylvania
8	taking the nonprice verticals out of that category, I am
9	not aware of any decision going whole hog and saying,
10	well, that because of Schwinn, now we have to say that
11	exclusive dealing is per se. Even in those dark days,
12	we never had a rule for exclusive dealing that said
13	basically shoot on sight.
14	So, now, having escaped per se condemnation, I
15	think it was easier for exclusive dealing cases to kind
16	of re-absorb the economic learning, to talk about
17	procompetitive justifications, to insist upon genuine
18	proof that the process of competition had been
19	obstructed before liability could be imposed. We went
20	from Standard Stations, we went to Tampa Electric, which
21	basically said, well, even quantitative foreclosure does
22	not really give us the story that we want to hear when
23	we are talking rule of reason. And so, in effect, this
24	evolution is kind of a testament to just how thoroughly
25	the microeconomic analytic approach has been absorbed in

123

1	the antitrust enforcement industry, the enforcement
2	agencies, the courts, counselors, what have you, and
3	this is all very much to the good. This is as it should
4	be.
5	But one result of this emergence into the more
6	full-blown consideration of justifications and actual
7	competitive effects is that the role of market power and
8	monopoly power have been pushed to the fore, and for
9	most kinds of exclusive dealing claims, you need to have
10	market power or monopoly power at one level in order to
11	have any kind of a plausible theory of restraint, and so
12	now it has become a topic that is addressed more under
13	the Section 2 standards than under Sherman 1 or Clayton
14	3, and that is fine. So, that focuses, to the extent
15	that these issues come up under the Section 2 rubric,
16	that focuses you on monopoly power, because it is a
17	required element of proof in every Section 2 case, or in
18	an attempt case, of course, the reasonable likelihood of
19	monopoly power being attained -- and it also means
20	that -- it really brings us down to I think the main
21	discussion, the main subject of discussion, which is the
22	definition of monopolizing conduct, and, of course, that
23	is a much broader area, and let's see what light we can
24	shed on the exclusive dealing aspect.
25	Well, one of my colleagues, Steve Calkins, has

124

1	already alluded to the fact that if you look at
2	exclusive dealing cases, there are not many in which
3	plaintiffs win, and it is interesting that some of those
4	cases are really not Section 1 or Clayton 3 cases
5	anymore, they are Section 2 cases, oddly enough, in
6	which the decision-maker for one reason or another
7	failed to condemn exclusive dealing under Sherman 1 or
8	Clayton 3, but only under Section 2, and that would
9	include U.S. v. Microsoft, Lepage's v. 3M, sort of in
10	the margins of exclusive dealing, one of those forms of
11	bundling, and then we have heard about U.S. v. Dentsply.
12	Now, within the broader debate about legal
13	standards for monopolizing conduct, exclusive dealing I
14	think is more or less kind of a classic example. What
15	do we have to go on when somebody is challenged for
16	their conduct under Section 2? Well, we have Grinnell,
17	we have Aspen, exclusion on the basis of something other
18	than efficiency; we have Image Technical Services, not
19	the part that everybody has had seminars about and
20	talked about for years and years and years, and Salop
21	said this and somebody else said that and it is
22	post-Chicago -- no, it is pre-post-Chicago -- okay, it
23	is post-modernist Chicago, but the point is there is a
24	second part of Kodak versus Image Technical, which say
25	what you will about the tying part, the first part of

125

1	the Supreme Court opinion, there is that second part
2	that makes some extremely broad characterizations of
3	what it takes to -- broad and vague characterizations --
4	of what it takes to prove monopolization. That part of
5	the opinion was so good that when Image Technical got to
6	go back and have its trial, it did not even bother with
7	all the hard post-Chicago stuff in the first part. It
8	just relied on that great language in the second part of
9	the opinion. So, it is really a question of
10	deconstructing and coming up with a monopolistic conduct
11	standard that can be applied sensibly to the generality
12	of these cases.
13	Now, I will put all my cards right on the table
14	and say I am not one of those who says there is
15	salvation to be had in taking the vague language of
16	Grinnell and the vague language of Aspen and the vague
17	language of the second Section 2 part of Image Technical
18	versus Kodak and trying to put some kind of a
19	microeconomic overlay on it, whether it is no economic
20	sense, profit sacrifice, exclusion of equally efficient
21	competitor. I think all of those things can come in
22	very handy. I mean, if you see a monopolist doing
23	something that causes it losses, you are entitled to
24	inquire, is it an eleemosynary motive, was it a mistake,
25	or was the monopolist taking money and paying for

126

1	something, and was it a competitive restraint? So, I do
2	not want to suggest that those concepts are useless, but
3	I think they are not going to get us the distance to a
4	standard under Section 2 for judging exclusive dealing.
5	As a matter of fact, I am prepared to say that
6	as a general matter, any standard that is simply stated
7	and purported to apply to the generality of exclusive
8	dealing cases cannot possibly give you specific enough
9	guidance to decide any particular case. This is just
10	one of those situations where we are kind of stuck with
11	the dilemma that Steve referred to in the initial part
12	of his remarks when he was quoting from Standard Station
13	saying, well, you know, we would love to consider all
14	these justifications, but, you know, it would not be an
15	administerable rule of law, there is nothing you could
16	do with it. Therefore, we are going to have a per se
17	rule based on quantitative substantiality.
18	It is a little bit like my favorite footnote in
19	Topco. Remember United States v. Topco, which was a
20	horizontal case, and it was a bunch of independent
21	grocers who had banded together and had arranged to have
22	their own private label line of products to offer in the
23	grocery store so they could compete with A&P Ann Page
24	and Safeway's whatever, and the District Court had said,
25	well, this is very procompetitive as a rule of reason,

127

1	case dismissed, and the Supreme Court said, oh, no, oh,
2	no, when you are talking about a horizontal restraint --
3	and it was a territorial restraint in that particular
4	case -- what the Supreme Court said is you don't
5	consider all that stuff, it is per se, and then they
6	dropped a footnote that said, well, look, if Congress
7	would like to adopt a rule of reason for this kind of
8	restraint and send the courts off into the wilds of
9	economic theory -- that's the exact phrase they use in
10	that footnote in Topco -- Congress can go to that, but
11	we are not going to, per se illegal, next case. So, we
12	have got a similar situation here.
13	Exclusive dealing could be good, could be bad,
14	depends on a lot of different factors, very hard to
15	formulate a different -- a reformulation of a general
16	standard that is going to apply in all circumstances,
17	and so I have very little faith in any such
18	reformulation. I think we are just stuck, you know,
19	courts do what they do. You have got a difficult area
20	where it is hard to make a judgment. Actually, as I
21	think as I am going to talk about toward the end of my
22	remarks, which will be soon, what I am basically saying
23	is if the courts find it difficult to take such an
24	amorphous standard and apply it to this practice, what
25	we have to have is better courts.

128

1	Now, we have mentioned that defendants almost
2	always win. So what? So what? I have no great faith
3	in the numerology of one loss statistics. The real
4	question is whether anticompetitive conduct gets struck
5	down in these cases and procompetitive conduct is
6	exonerated, and by that standard, as I read the same
7	cases that Steve has obviously read -- and he has
8	probably spent a lot more time reading them than I have
9	and has read a lot more cases as well -- but I find it
10	very difficult to say that something is seriously awry.
11	I have cases where I would disagree with what is
12	going on, but there are two cases in the -- well, I have
13	talked about Microsoft, U.S. v. Microsoft, Lepage's v.
14	3M, U.S. v. Dentsply. I have listed -- have I listed in
15	my -- well, anyway, three cases I could name where the
16	defendants won, three recent important cases where the
17	defendants won, Pepsico versus Coca-Cola, this is the
18	New York case affirmed by the Second Circuit where
19	basically the Second Circuit said you do not get a trial
20	on the proposition that the reason quick-service
21	restaurants do not buy Pepsi-Cola is that Pepsi-Cola
22	cannot figure out a way to deliver the syrup to the
23	restaurants. Whatever reason there is for the relevant
24	market shares in quick-service restaurants for
25	carbonated soft drinks, it is not that Pepsi-Cola could

129

1	not figure out a way to get its product delivered.
2	Omega Environmental versus Gilbarco, I do not
3	know any more than what you, the average case reader,
4	knows. I had no involvement with that case. Then we
5	have Harmar Bottling, which is, again, a case that I do
6	know something about. I am not sure the facts bear the
7	characterization that Steve was giving it. I do not
8	want to get into a cat fight with him over that, but I
9	will just say that I think the result in that case was
10	correct, and so of the cases I know, of the cases I have
11	read about and tried to understand, I do not think you
12	can say that defendants are winning in cases where they
13	should not win.
14	So, you know, we need to figure out a way to
15	assess exclusive dealing efficiently, and basically, as
16	I say, my message is there is some exclusive dealing
17	that is good, some exclusive dealing that is bad.
18	Harmar took about 14 years to tell one from the other,
19	and my main message is that there has got to be a way of
20	getting to an efficient resolution of these cases much
21	more quickly. As a matter of fact, I would consider
22	whether -- I might regret this if it became a sound
23	bite, but if there is a sound bite I would give you,
24	let's have the antitrust enforcement mechanism, let's
25	adopt as a policy objective, that in the area of

130

1	exclusive dealing, we want to reduce the duration and
2	the expense of deciding whether exclusive dealing in a
3	particular case is good or bad. Let's reduce the
4	duration and expense by an order of magnitude so that a
5	Harmar, which took 14 years to litigate, takes, say, 14
6	months to litigate.
7	Now, in this column, I have very high praise for
8	the Ann Bingaman suit against Microsoft which resulted
9	in the 1994 consent decree. I know that there was some
10	investigation prior to the time that the DOJ got the
11	file in that case, but I remember being incredibly
12	impressed for two reasons with that effort. Number
13	one -- well, other than feeling that the result was
14	right. It was a consent decree, but I think it did the
15	right thing.
16	Number one, it was about exactly one year
17	between the time that the Department of Justice got the
18	file in that case and the date that the decree was
19	entered, and number two, it was a very specific,
20	targeted form of relief. It was a doable form of
21	relief. So, if you can do an exclusive dealing case
22	that quickly and come up with a result that concrete in
23	a year, it forgives almost any other defect that you can
24	find in that case, because on that time scale, you can
25	correct for your mistakes. You can, you know, do in

131

1	year two what you failed to do in year one, or vice
2	versa. So, litigation efficiency is an extremely
3	important consideration, and we ought to figure out ways
4	for a great increase in litigation efficiency.
5	One minute, that is exactly what I need.
6	So, here are some ideas for enhancing the
7	efficiency of this process, and I think a lot of the
8	tools are already at hand. Daubert, it has already been
9	used in an exclusive dealing context. Let's have more
10	of it. Let's make sure that expert testimony is forced
11	to go through and survive a plausibility test, the
12	Daubert standard. Let's make sure that the plausibility
13	formulation in Matsushita and Brooke Group, even though
14	that is relative to predatory pricing, a plausibility
15	test should also be applied to other types of antitrust
16	claims, including exclusive dealing, help filter out
17	losing claims early, and focus remaining claims on all
18	phases for the remainder of the litigation, so you are
19	not carrying forward speculative theories and going
20	through the wasteful discovery and legal motions and so
21	forth that that involves.
22	Second, expand the use of neutral expert or
23	expert panels, and I want to emphasize here, it is not
24	just in a strict Rule 706 sense, in other words, an
25	expert witness providing economic testimony to a judge

132

1	in a matter in litigation, like Fred Kahn's testimony in
2	the New York versus -- the Nabisco Brands case. That
3	was a very effective use of a 706 expert, but we need
4	ways to bring specialized knowledge about antitrust
5	cases, discovery, theories, the nature of the market, we
6	need to put those resources at the service of the courts
7	that are having these exclusive dealing litigation
8	things litigated before them.
9	And the last one I won't go through due to the
10	shortness of time, but the Manual for Complex Litigation
11	does contain a few things about antitrust, but perhaps
12	of the ideas that we could expand, the sort of helpful
13	guidance, the identification of issues, the suggestion
14	of efficiency-enhancing methods of resolving complex
15	litigation, expand it specifically in the area of
16	monopolization and exclusive dealing for the use of the
17	courts.
18	So, just to sum up, I do not think that our
19	exclusive dealing jurisprudence is in crisis. I kind of
20	like where the law is. Some exclusive dealing is good,
21	some exclusive dealing is bad, it is not per se legal,
22	it is not per se illegal, but if we could reduce the
23	time it takes to tell the difference between good
24	exclusive dealing and bad exclusive dealing by an order
25	of magnitude, I think that would be a very worthy goal

133

1	for the antitrust policy.
2	(Applause.)
3	MR. O'BRIEN: Thank you, Tad.
4	Okay, our next speaker, shifting gears to a
5	couple of economists, is Joe Farrell. He is Professor
6	of Economics at the University of California, Berkeley,
7	and he is a Fellow of the Econometric Society, former
8	editor of the Journal of Industrial Economics and former
9	President of the Industrial Organization Society.
10	Currently he's the senior consultant for Charles River
11	Associates.
12	Joe's published widely articles on a broad range
13	of topics in industrial organization and microeconomics,
14	including exclusive dealing. He has substantial policy
15	experience as well, having served as Chief Economist at
16	the Federal Communications Commission from '96 to '97
17	and Deputy Assistant Attorney General for Economics at
18	the Antitrust Division from 2000 to 2001.
19	Joe?
20	DR. FARRELL: Well, I am an economist. I am
21	going to talk about economics for a few minutes, and
22	then I am going to talk about the law. I feel all right
23	about this because I hear a lot of lawyers talking about
24	economics.
25	So, economics for the most part in antitrust

134

1	analysis has focused on the question, what should we do
2	if we knew really quite a lot about the case, okay? And
3	in the area of exclusive dealing, I think a bland and
4	very fair summary of economics in this area is both
5	efficiency and anticompetitive effects and explanations
6	of exclusive dealing are very possible, and on both
7	sides of that, the analysis is really quite subtle, and
8	I am going to spend a few minutes on this. In terms of
9	the efficiency explanations, I am going to focus on the
10	investment incentive theory, which I think Ben Klein is
11	also going to talk about a form of. In terms of
12	anticompetitive effects, I am going to talk about what I
13	think is the leading example, though not the only
14	example, of an economic structure to understand
15	anticompetitive effects of exclusive dealing.
16	So, in terms of the investment incentives, you
17	will often hear it said that exclusive dealing is
18	efficient if you have to motivate relationship-specific
19	investment or some such phrase as that, okay? As far as
20	I know, the state of the art in the economics literature
21	on these arguments is the article by Elias Segal and
22	Michael Whinston in the Rand Journal, 2000. They start
23	out by showing that in what appears to be quite a
24	general model, relationship-specific investments, that
25	is, investments that have no value outside the

135

1	relationship, are not -- repeat, not -- an efficiency
2	rationale for exclusivity.
3	They then continue to show that investments that
4	are not in that strict sense relationship-specific, that
5	have a spillover to deals between the customer and the
6	potential entrant, might or might not be an efficiency
7	rationale for exclusivity. It depends on quite a number
8	of things. It depends on who is doing the investment.
9	Is it the buyer or the seller? It depends on how it
10	spills over. Is it a complement or a substitute with
11	the efficiency of potential deals between the buyer and
12	an entrant? It depends on the bargaining structure
13	between the buyer and the seller. It depends on what is
14	the nature of any investment by us absent the exclusive
15	dealing. And that is all within their model. If you
16	step outside that model, it also depends on whether
17	their model sort of applies or sort of does not apply.
18	So, I am going to leave you for the moment with
19	the thought, how is a court likely to be able to
20	disentangle all this in addressing an asserted
21	efficiency rationale along the lines of investment
22	incentives?
23	Now, what about the other side of the courtroom,
24	divide and conquer exclusion, Rasmussen and Ramseyer and
25	Wiley, 1991, corrected, beefed up and radically improved

136

1	by Segal and Whinston in the American Economic Review,
2	2000, show that exclusion can profitably and harmfully
3	work against end users; however, although I think that
4	is very well understood and accepted, the fact is their
5	models involve buyers who are end users.
6	In most cases that I am aware of, exclusive
7	dealing is not a deal struck with end users. It is a
8	deal struck with retailers or distributors or someone
9	else intermediate in the value chain between the
10	manufacturer and the end users. That makes a lot of
11	difference.
12	So, interestingly, a year or two ago, there
13	appeared to be economics literature, two broadly
14	parallel articles, papers, one by Fumagalli and Motta,
15	which I believe has been published or is about to be
16	published in the American Economic Review, and one by
17	John Simpson and Abraham Wickelgren, and within the last
18	24 hours, I have learned about other articles by Yong
19	and Shaffer that may be somewhat along the same lines,
20	and both of these articles address the question, how
21	does the RRWSW theory of anticompetitive exclusive
22	dealing change when you recognize that the buyers in the
23	model, in practice, should be replaced by buyers who are
24	not end users?
25	Well, there are two forces, okay? One force is

137

1	that intermediate buyers, nonfinal buyers, actually do
2	not care that much if the price goes up or stays high,
3	provided it goes up or stays high to all of them,
4	because then it gets passed through downstream, okay?
5	How much that is true depends on the details of the
6	market structure and so on, but that tends to be true.
7	That lowers their resistance to things that maintain
8	monopoly upstream relative to what it would be if they
9	were end users. So, that you would expect would make
10	anticompetitive exclusive dealing easier.
11	Another force, however, is that if you have a
12	nonfinal buyer who holds out and does not sign the
13	exclusive deal, then an entrant can come to him and say,
14	"Aha, I will give you a lower price than all your tied
15	up rivals will be getting. You can expand. You and I
16	can meet my scale requirements, and you will make a
17	bundle of money." So, that dynamic potentially makes it
18	harder to have anticompetitive exclusive dealing.
19	Well, Fumagalli and Motta found conclusively
20	that it went one way, and Simpson and Wickelgren found
21	conclusively that it went the other way, and which way
22	Yong and Shaffer come out, I do not know yet. Which of
23	them is right and when? Well, I attempted to diagnose
24	this in my Antitrust Bulletin article last year. My
25	attempted diagnosis is that it depends on whether in

138

1	that last situation where you had one hold-out buyer,
2	the incumbent is then able to or does adjust the price
3	that it charges the tied buyers. So, I believe
4	Fumagalli and Motta assumed that it does not, and
5	Simpson and Wickelgren assumed that it does, or maybe it
6	is the other way around, okay?
7	When I put this tentative diagnosis to one of
8	the four economists -- and I will not say which one --
9	the response I got was, "Ah, that is interesting, I am
10	not sure." That is telling, I think, because it says
11	that it is kind of unlikely that a court is going to do
12	a very good job of disentangling all of these difficult
13	concepts. Now, the optimistic view is this is just the
14	beginning of the economic exploration of this topic, and
15	come the year 2010, we will understand it well and in a
16	way that is good enough for us to brief courts on it,
17	and maybe that will happen, okay, but I take from this
18	two things.
19	One is economics is making progress, that is
20	great, I hope to participate, but the other is, it is
21	pretty subtle and it will probably stay pretty subtle,
22	if not get more subtle.
23	All right, so we are doing antitrust under
24	uncertainty. We are not in the world where we can say
25	exactly what is going on and work out the welfare

139

1	consequences, okay? Let's take that as an assumption
2	for now.
3	Well, traditionally at this point economists
4	plunge into Bayesian mode and talk about type one errors
5	and type two errors and so on. Underlying what I am
6	going to say, there certainly is a Bayesian framework,
7	okay, but I am not going to talk explicitly in Bayesian
8	terms. I am going to talk in jurisprudential terms,
9	because my lawyer colleagues on this panel have been
10	talking economics, so I want to get back at them.
11	So, I am going to talk about the role of
12	presumptions and burdens of proof, and I am going to
13	talk about two presumptions that should be extremely
14	important in antitrust policy and about what I
15	personally think -- although I cannot prove -- is a very
16	worrying trend that has been taking place in the
17	relative strength of these two presumptions.
18	So, what are these two presumptions? Number
19	one, in economic policy generally, in market economies,
20	we have a laissez-faire presumption. The Government
21	should not intervene in stuff unless it is reasonably
22	sure that intervention will help. I think that is a
23	pretty good idea.
24	Number two, in antitrust particularly, we should
25	protect competition unless we are reasonably sure that

140

1	some alternative is better, okay? So, I think at a very
2	grand, 40,000-foot level, you can view a lot of what
3	goes on in antitrust jurisprudence as being a tug of war
4	or back and forth between these two presumptions.
5	Now, I put competition in quotes on this slide
6	for a reason, and that reason is when you look at it too
7	closely, things get a little out of focus, and you do
8	not exactly know what that word means, okay? And that
9	has led us, I believe, over the course of the decades
10	towards the tempting solution of redefining the word
11	"competition" to mean what is good. So, here is a test
12	of that, okay?
13	What happens when you hear someone refer to the
14	possibility that a merger to monopoly would reduce
15	marginal costs so much that it would be good for
16	efficiency and consumers? Well, if that were true,
17	let's say you knew it was true, it would be a good
18	thing. Would it be procompetitive? I think a lot of
19	people would say yes, because it is a good thing, but
20	that is ridiculous. It is not procompetitive. It is
21	pro-consumer, it is pro-efficiency. It is not
22	procompetitive.
23	So, if we are going to use words in their real
24	meaning rather than redefining them so that the
25	definition does our policy analysis for us, we have got

141

1	to be a little careful about doing stuff like that.
2	Now, of course, the antitrust law protects
3	"competition," so tautologically, redefining the word
4	would be a good idea, it would lead us to do good
5	policy, if we always knew what was going on, okay? So,
6	given that the law protected competition, it would be a
7	very smart move on the part of benevolent antitrust
8	enforcers and courts and so on to redefine the word
9	"competition" so that the law then protects whatever is
10	good, okay?
11	However, there is a problem with doing this. A,
12	we do not always know what is going on exactly, and B --
13	B only applies given A -- attempting to have a
14	presumption in favor of protecting competition makes no
15	sense if you define competition to mean what is good,
16	okay, because if you knew that something was good, you
17	would want to do it, and that is not a presumption in
18	favor of protecting competition. So, for there to be
19	any meaning to the presumption in favor of competition,
20	it has to be a presumption in favor of something that
21	has not yet been proved to be good, okay?
22	So, this I think casts an interesting light on
23	the slide that I heard this morning -- and I was not
24	taking notes on who said it -- but somebody said
25	something along the following lines, or if I misheard

142

1	it, it has certainly been said within the last week --
2	that because there are perfectly plausible efficiency
3	justifications for exclusive dealing, plaintiffs should
4	be required to prove that there is an anticompetitive
5	effect, okay? That, of course, would be obviously right
6	if we could always prove what is true, but if we cannot
7	always prove what is true, it is not obviously right.
8	It might still be right, but it is not obviously right,
9	okay?
10	So, in order to explore this, let me, with
11	tongue in cheek, put the shoe on the other foot, okay,
12	and let's suppose that we applied the same redefinition
13	to the laissez-faire presumption, okay? So, we have
14	this presumption that says the Government should not
15	intervene unless it is pretty sure that intervention is
16	a good thing, okay? So, now let's suppose that we
17	defined laissez-faire as the good outcome, and we
18	defined intervention as the bad outcome.
19	Now, if the Government wants to come along and
20	insist that you paint your bedroom walls blue, not
21	white, you can't say that is intervention, because you
22	have not proved that it is a bad thing, okay? Well,
23	that is obviously pretty stupid.
24	So, I come out of this thinking it would be a
25	good idea for us to make sure that words go on meaning

143

1	what they mean, and it is very dangerous -- it has had
2	some good consequences, but it is nevertheless very
3	dangerous -- to redefine words to make them do your
4	policy analysis for you.
5	So, antitrust intellectual history, to the
6	extent that I understand it -- in less than one
7	minute -- in the bad old days, anything that could be
8	presented as a reduction of competition was illegal.
9	That was bad, because quite often, things that can be
10	presented as a reduction of competition are actually
11	good. The good new days, we have got to analyze the
12	effects of things that seem to be capable of being
13	presented as a reduction in competition, because you
14	would not want to ban those things if they are actually
15	good, okay?
16	What I am worried about is the possibility that
17	we are drifting into the not so good new days where it
18	is difficult to prevent things that are in some sense
19	reductions of competition unless you can actually prove
20	that those things are bad. Now, of course, you would
21	not want policy to prevent those things unless they are
22	bad, but that is very different from unless you can
23	prove that they are bad.
24	Now, the final bullet on this slide, which is
25	quite important, I talked about these ideas very briefly

144

1	with some people in Europe over the summer, and they
2	were aghast. Why were they aghast? Because they said
3	we have spent years trying to move away from a
4	descriptive basis of liability towards an effects-based,
5	economics-based concept of liability, and now, you are
6	coming from over there and trying to undo that. Well, I
7	take that seriously, so I am not going gung ho on a
8	policy proposal here, but it does seem to me that if too
9	much burden of proof is being imposed, that is a
10	problem.
11	Let me finish with this slide, dark matter, do
12	the physics, okay? It is a good idea to intervene only
13	if intervention benefits efficiency or consumers. It is
14	maybe not such a good idea to intervene only if you can
15	specifically prove that and how it would do so, okay?
16	There are multiple benefits of competition in
17	most circumstances. Often, there are concrete,
18	predictable, provable price effects, okay? Merger
19	simulation has been a very powerful tool in exploring
20	that. There is also the much vaguer and harder to pin
21	down possibility that having a bunch of different firms
22	doing different things and independent of each other can
23	lead us to benefits that are much harder to prove or
24	even define or even point to ex ante, okay? I call this
25	the dark matter of competition policy, because as in

145

1	astrophysics, if it exists, it is quite likely to be a
2	lot bigger and more important than the stuff that you
3	can see. So, watch out when you are imposing burdens of
4	proof.
5	(Applause.)
6	MR. O'BRIEN: Thank you, Joe.
7	Our final speaker is Ben Klein. He is Professor
8	Emeritus of Economics at UCLA and a director at LECG.
9	Ben is an internationally recognized expert on antitrust
10	economics. He was a Professor of Economics at UCLA for
11	34 years where he published numerous articles on a range
12	of topics, including antitrust, contracts and
13	intellectual property.
14	He currently serves on the board of editors of
15	five academic journals. Over the past 25 years, he has
16	consulted extensively on antitrust issues and has made
17	numerous presentations to state, federal and foreign
18	regulatory agencies and courts.
19	Ben?
20	DR. KLEIN: Thank you, Dan.
21	I am going to be talking mostly about economics,
22	and although what I am going to say is subtle, you
23	should not reach the conclusion from what Joe said that
24	because the arguments are subtle that, therefore,
25	anything goes. Just find the economist that is going to

146

1	make the argument you want to hear. I think there is
2	truth out there. This is moving us along on coming up
3	with what is the economic foundation for some commonly
4	used procompetitive justifications.
5	This is a paper that I am working on with Andres
6	Lerner. The paper is posted on the web site, and I
7	think it is important to go through these procompetitive
8	justifications in terms of the economics, because the
9	danger I see is the exact opposite one. I think that we
10	are moving in the direction that if you find a practice
11	that does not have efficiencies, it is becoming a
12	sufficient condition, if it is something that is being
13	used by a firm, a large firm, it is a sufficient
14	condition for antitrust liability, because the very
15	nature of an exclusive dealing contract is
16	"exclusionary," and then when you get to the balancing,
17	you have nothing on one side of the scale.
18	Although the paper discusses many exclusive
19	dealing cases, we concentrate on Dentsply, and that is
20	what I am going to concentrate on today, and it is
21	because the court in that case used economics to reject
22	two very common procompetitive justifications, both free
23	riding and this undivided dealer loyalty justification,
24	and the principles that I am going to be giving you here
25	can be applied to a number of very different claimed

147

1	justifications, and we do it in the paper.
2	So, in terms of Dentsply, as I said, Dentsply
3	illustrates that actually the economic foundations for
4	procompetitive justifications are actually pretty
5	narrow, and the Court rejected Dentsply's claim, in
6	particular, that exclusive dealing was used to prevent
7	dealer free riding on manufacturer-supplied promotional
8	investments. This is the classic Howard Marvel
9	rationale, where the manufacturer makes investments in a
10	dealer, you know, like they build out a dealership or
11	engage in dealer training, and then the dealer uses
12	those manufacturer investments to sell a rival product,
13	and that is the classic free riding argument. The Court
14	rejected that, and the Court rejected the undivided
15	loyalty argument, that somehow you give somebody an
16	exclusive so they will more actively promote Dentsply's
17	product.
18	The Court rejected the free riding rationale
19	basically because the Court found it was contrary to the
20	facts, that number one, Dentsply did not make any
21	investments in the dealers that they could then free
22	ride on by using them to sell rival products. There was
23	no evidence, essentially no evidence in the case, that
24	the Dentsply dealers were actually switching buyers to
25	rival products. And finally, that there was testimony

148

1	by Dentsply executives that if there was not an
2	exclusive, they actually would have invested more -- you
3	see, the usual economic argument is the purpose of
4	exclusive dealing is to encourage the manufacturer to
5	make investments, and one way it is encouraged to make
6	investments is to prevent free riding that it knows that
7	these investments are going to be used to sell its
8	product, and the Court said, you know, the Dentsply
9	executives actually testified that if we did not have
10	exclusive dealing, we would have had to make more
11	promotional investments.
12	In terms of the other argument, the Dentsply
13	Court rejected the undivided loyalty argument, and here
14	it was not really just the facts. It was basically the
15	theory that this theory about enhancing dealer services
16	cannot be a justification for exclusive dealing, because
17	in general, competition between dealers is going to lead
18	them to supply the desired quantity of promotional
19	services, as the Court said, the dealers have the
20	incentive in competing with other dealers to make sure
21	that they supply the right kind of services.
22	See, basically the problem that Dentsply ran
23	into is although this undivided loyalty argument has
24	been accepted by a number of courts, Judge Robinson in
25	this case knew a lot of economics, and in particular,

149

1	she knew Howard Marvel's argument and had read the
2	article, and Howard Marvel was the expert that Dentsply
3	had hired for this, and she said, no, even in your
4	expert's article, he says that you can generally leave
5	it up to competition to put dealers to supply the right
6	services. It is only when you have this problem, this
7	inter-dealer free riding problem described in Sylvania,
8	you know, and that is a problem where the customer goes
9	to one full-service dealer and gets some kind of dealer
10	services and then goes to another dealer and buys the
11	product, you have that inter-dealer free riding problem,
12	and in that circumstance, maybe competition among
13	dealers will not give you the right quantity of dealer
14	services, but that is a problem that would not be
15	corrected with exclusive dealing, because even if you
16	had exclusive dealing and you had this kind of problem,
17	the exclusive dealer would say, no, get the services
18	from somebody else and then come and buy the
19	manufacturer's product from me.
20	So, as I said, although this rationale has been
21	accepted by a number of courts, Judge Robinson said, you
22	know, basically you can leave it up to competition, and
23	this undivided loyalty makes absolutely no economic
24	sense.
25	In contrast to basically the established

150

1	economics, I think the expanded economic framework that
2	I am going to present here shows that these arguments
3	make sense, that free riding is much more general than
4	you would think, and the dealer undivided loyalty makes
5	sense, and it is based upon two common sense business
6	propositions.
7	Number one, that manufacturers often want their
8	dealers, even dealers that are competing with one
9	another, to supply more promotion than the dealers would
10	independently provide on their own, and number two, that
11	exclusive dealing by creating this undivided dealer
12	loyalty actually increases the dealer's incentives to
13	supply these desired services and to more actively
14	promote the manufacturer's products.
15	So, in terms of the logic of what I am going to
16	do, first I am going to discuss the first proposition,
17	and hopefully people have been here in the morning and
18	heard Josh Wright, who has done this already, but
19	basically the first proposition is manufacturers, in
20	general, cannot leave it entirely up to the dealer
21	competition to get the quantity and the type of services
22	they want supplied, and the logic of the argument is
23	there is another step where, therefore, they have to
24	contract with their dealers, either explicitly or
25	implicitly, to solve this problem and to make sure that

151

1	they adequately promote their product. That leads to
2	these free riding problems, which I will discuss are
3	much broader than the classic Marvel free riding
4	problems. And then finally, that exclusive dealing is
5	commonly an element in those contractual arrangements
6	that gets the individual dealers' incentives then
7	aligned more with the manufacturer's incentives.
8	So, let me do the first proposition first, that
9	manufacturers often want their dealers to supply more
10	promotion than the dealers would independently decide to
11	provide, and the basic reason for this is that the
12	dealers do not take account of the manufacturer
13	profitability on incremental sales, that the dealer does
14	something that increases the manufacturer's sales, and
15	the dealer gets only a part of that incremental profit,
16	in many cases only a very small part of the incremental
17	profit.
18	Now, in general, this is not a problem for
19	dealer price and nonprice competition that has
20	significant inter-dealer quantity effects. So, in
21	general, when a dealer provides a desirable service like
22	free parking or lowers its price a little bit and makes
23	a little bit more sales, even though they might have a
24	small margin in terms of the total profit being earned
25	by the manufacturer and retailer together when they make

152

1	that extra sale, because they are getting consumers to
2	switch from other dealers, because there is large
3	inter-dealer effects, you get an equilibrium where you
4	get the desired quantity of the services provided, but
5	with promotional activity, the primary effect is not
6	really inter-dealer, but it is primarily inter-brand,
7	that you just make an extra sale for the manufacturer,
8	and there are no significant inter-dealer quantity
9	effects. Then you have this problem where the dealer,
10	by not taking account of the incremental profit, is
11	going to supply less than the desired promotional
12	services of pushing the manufacturer's product. In
13	addition, dealers cannot charge consumers directly for
14	those services, because the promotion is, in effect, a
15	price discount.
16	So -- I am going to have to go faster --
17	manufacturers solve this problem -- although I am going
18	to be talking about violating these contracts, I can
19	always violate, you know, this one --
20	MR. O'BRIEN: There is no red string we can pull
21	--
22	DR. KLEIN: No, there is no self-enforcement
23	problem here, although -- anyway, I am wasting my time.
24	Manufacturers solve this problem of insufficient
25	dealer promotion by contracting with and compensating

153

1	dealers for providing increased promotion, and the
2	contract may be explicit or it -- you know, in plenty of
3	the cases, like in Standard Fashion, they explicitly
4	said you have to have a certain amount of display space,
5	you have to have a "lady attendant" there full-time,
6	they used a few words like that. Most of the times it
7	is really understood that you are going to make your
8	best efforts, and they compensate dealers in these
9	things by giving them a valuable distributorship in the
10	sense that if they get terminated because they are not
11	pushing the product adequately, they are going to lose
12	this future rent stream, and the threat of termination
13	is what gets them to perform as desired.
14	However, because dealers are contracting to
15	supply more promotion than they would otherwise, you
16	know, do in their own independent interests, there is an
17	inherent problem in that they have an incentive to
18	violate the contract and free ride on the manufacturer's
19	compensation arrangement, basically because you are
20	getting a valuable dealership, like in Beltone, they
21	gave them an exclusive territory. In Standard Fashion,
22	they had minimum resale price maintenance. Whatever it
23	is, you have something valuable, but you are getting it
24	on all your sales, and you therefore have an incentive
25	just to do that pushing at the end and save the cost if

154

1	you are a dealer, and still you are getting most of the
2	compensation.
3	In terms of this contract, dealers may violate
4	the contract and free ride in three distinct ways, and
5	the first way is the standard case where the dealers use
6	the manufacturer-supplied investments to sell rival
7	products, and that is part of the contractual
8	arrangement. Look, we will give you these complementary
9	assets to help you push our product, and that is one
10	that you know about, but there is two other free riding
11	problems.
12	Second is the dealers may just use the
13	manufacturer paid for promotion to sell rival products,
14	that they are being compensated with this valuable
15	dealership, and on the margin, they are just going to
16	switch, and the profit incentive is really the same as
17	one, but you do not have to find these manufacturer
18	assets there.
19	And the third one is the dealers may just
20	under-supply the manufacturer's paid-for promotion, as I
21	said, because on the margin, they are getting paid on
22	all these inframarginal sales, and on the margin, it
23	really does not pay for them to spend all this money on
24	pushing the products on the margin if it was not for
25	this contract.

155

1	Dealer free riding need not involve manufacturer
2	investments or dealer switching. That is the
3	implication of this. So, for example, in free riding
4	one, which is the one you all know about, that one
5	involves manufacturer investments and dealer switching.
6	That is what the Court in Dentsply said, there is no
7	free rider problem here. But free riding, too, the
8	dealers are just using the paid promotion to sell the
9	rival products, and that one can occur without any
10	manufacturer investments whatsoever. They are just free
11	riding on the compensation arrangement.
12	Free riding number three, where dealers are
13	undersupplying what the manufacturers are paying for,
14	that one occurs without any manufacturer investments or
15	without any dealer switching, okay, and exclusive
16	dealing may be used to mitigate all thee forms of free
17	riding, and it prevents free riding types one and two by
18	just preventing the switching of sales to rival
19	products, and it prevents free riding number three by
20	creating this undivided dealer loyalty by promoting the
21	incentive of the dealers to promote the manufacturer's
22	product more intensively that aligns the incentives.
23	So, how does exclusive dealing, that third type,
24	how does the exclusive dealing increase the dealer's
25	incentive to promote? And remember, we are operating in

156

1	the context, you know, why did the Dentsply Court reject
2	this as making absolutely no economic sense? And that
3	is because there is all this competition between
4	dealers, and that is all that is necessary to get the
5	services provided unless there is a Sylvania type
6	problem, and the example that we go through in the paper
7	is this.
8	Consider this case where a customer is thinking
9	about buying a car and is leaning towards the purchase
10	of a Honda, and he goes into a Toyota dealership to
11	check out the Toyota, but really, you know, it is -- but
12	just to make sure, let me just check out the Toyota.
13	So, that is the hypothetical example.
14	Then I have this -- look at that. So -- and
15	under that -- there is a Honda and there is a Toyota,
16	and Mdh is the profit margin that the dealer earns if it
17	sells a Honda, and Mdt is the profit margin for the
18	dealer if it sells a Toyota, and the Toyota dealer is
19	deciding, what about this, even though they are leaning
20	towards the Honda?
21	Well, a nonexclusive dealer will not make its
22	best efforts to sell the Toyota if it has both cars
23	there, and basically -- now, do not get scared -- but
24	the dealer is going to choose a level of Toyota
25	promotional service as S that maximizes its

157

1	profitability. So, it chooses S, that maximizes the
2	profitability, which is the difference between the
3	margin on the Toyota minus the margin on the Honda,
4	times the probability that they will make the Toyota
5	sale if he starts telling them how great the Toyota is,
6	whether they will buy the Toyota, and that probability,
7	p(S) is a positive function of how much S the person
8	chooses, minus the cost of supplying S, and obviously
9	there is a positive marginal cost. The more S, the
10	higher those costs.
11	And since it costs the dealer less to sell the
12	Honda in this nonexclusive context, the dealer can earn
13	a higher net profit margin on selling the Honda, the
14	dealer goes to the one where the marginal cost of
15	providing additional services is equal to the
16	probability -- increased probability, as you supply --
17	of making the Toyota sale as you're supplying more
18	services, times the difference in the margin between the
19	Toyota and the Honda, and you can assume that they could
20	sell the Honda -- the customer comes in and sells -- and
21	wants to buy the Honda, and the salesperson can at a
22	zero cost sell the Honda and say, you know, you are
23	right, Honda is much better. Come here, I will get you
24	a good price.
25	So, under these circumstances -- for example, if

158

1	the dealer's margin on the two cars were the same, so
2	that Mdt and Mdh were the same number, that difference
3	would be zero, and clearly the dealer would supply
4	absolutely no services in trying to sell the Toyota. It
5	would be cheaper for the dealer to just write up the
6	sale for the Honda. But by selling the Honda rather
7	than promoting the Toyota, the dealer is free riding.
8	He is engaging in that third type of free riding that we
9	were talking about. The dealer is not switching. The
10	dealer is not actively promoting the rival Honda brand
11	as an alternative to Toyota, you know, for customers who
12	come in and want Toyota, as occurs in free riding one
13	and two. Instead, the dealer is violating the implicit
14	dealer contract for the Toyota by failing to actively
15	promote the Toyota automobiles.
16	Alternatively, if it was an exclusive, you know
17	that undivided loyalty is going to lead dealers to
18	expand their promotional efforts, and it is just going
19	to go to the point where the marginal costs of
20	additional efforts in pushing the Toyota is exactly
21	equal to how much it can make on the Toyota, times the
22	increased probability that the promotion makes it more
23	likely that they will make the sale. So, undivided
24	loyalty is clearly in that case going to lead to that,
25	and that is what you sometimes see courts saying, you

159

1	know, that if you do not make the -- you know, what
2	happens if you are an exclusive Toyota, basically it
3	means if you do not sell the Toyota, you do not make any
4	sale, and so it is common sense -- and, you know, this
5	is the business people who understand this -- it is
6	common sense that undivided loyalty is going to give you
7	an incentive to promote more, and in the paper, it is a
8	function of -- you still do not get to the point where
9	the dealer has the right incentive in terms of
10	maximizing the total profit of the manufacturer and
11	dealer together. That is the last thing on the
12	left-hand side. So, they still have to have these --
13	the manufacturer still has to have these implicit
14	self-enforced contracting and -- to go all the way to
15	the end, but basically the role of exclusive dealing is
16	that it aligns the incentives that are here.
17	So, I am done. The lessons, other than that I
18	put too much down here, okay? Lesson one, the Court's
19	rejection of Dentsply's procompetitive rationale is an
20	example of a common error that I think occurs in cases
21	of trying to fit the facts of a case into a preconceived
22	economic model rather than developing a model to fit the
23	facts of the case, and the preconceived theory, economic
24	theory here, that the Court adopted was basically, you
25	know, interdealer competition will lead dealers to

160

1	supply the type and quantity of promotional services,
2	unless you had that Sylvania type free riding problem,
3	and -- you know, because there are more likely to be
4	valid procompetitive justifications for exclusive
5	dealing, one of the implications I think is that this no
6	economic sense test is less likely to be a useful test
7	for antitrust liability, that there may be efficiency
8	justifications for exclusive -- people talk about the
9	Dentsply case as an easy case because there is nothing
10	on one side of the scale. There is obviously something
11	on one side of the scale is what I am trying to say, but
12	clearly, even though there is an efficiency
13	justification, you may have anticompetitive effects.
14	I think that the facts of that case, there were
15	significant anticompetitive effects, and Jonathan
16	Jacobson makes this point in his excellent latest
17	article in the Antitrust Law Journal. What he doesn't
18	do is he does not answer the Court's finding that there
19	was absolutely no economic basis for Dentsply's
20	undivided loyalty and free riding justification. So, in
21	that case, you would not get the wrong answer if you
22	used the no economic sense test, but the only reason you
23	do not get the wrong answer is because you do not really
24	understand the procompetitive justifications.
25	So, as I said in the beginning, I think the

161

1	greater danger is not that -- you know, the way some
2	people are advocating this no economic sense test as a
3	necessary condition for antitrust liability. I think
4	the danger is that the courts are going to use a no
5	economic sense test as a sufficient condition for
6	antitrust liability when a large firm uses exclusive
7	dealing, and it is not only that I am giving you that
8	there are other valid procompetitive rationales, but I
9	think as economists and as regulators we have to be more
10	humble that just because we have not figured this out
11	yet, there is lots of other procompetitive efficiency
12	justifications, and we cannot assume that the purpose of
13	a restraint is anticompetitive.
14	How much did I violate the contract by?
15	MR. O'BRIEN: Ten minutes.
16	(Applause.)
17	MR. O'BRIEN: Okay, we are going to break until
18	about five past 3:00, okay?
19	(A brief recess was taken.)
20	MR. O'BRIEN: Okay, let's get started.
21	Okay, I would like to start out our sort of
22	post-speech session here by asking folks if they would
23	like to comment on the remarks of others on the panel,
24	and I guess I will ask for a volunteer first rather than
25	being systematic about it. We will find out who has the

162

1	most burning comments to make about what someone else
2	said.
3	Okay, Joe.
4	DR. FARRELL: Well, I have one question for a
5	fellow panelist, which is relatively specific, I think.
6	Ben, in your model, you didn't have time to
7	present all of it, but I would like to ask, have you
8	offline, as it were, closed the loop and shown actual
9	harm to buyer, or is it just that the buyer who was
10	leaning towards buying a Honda ended up buying a Honda
11	and, of course, the Honda -- Honda likes that, the
12	dealer apparently likes that, the customer seems to like
13	that, although the welfare economics of this promotion
14	stuff, of course, are a little subtle. Toyota, of
15	course, does not like it.
16	Where does this go and how does the whole thing
17	play out with and without exclusive dealing as opposed
18	to just Toyota would like S to be higher in the short
19	run?
20	DR. KLEIN: All we do in the paper is present
21	the procompetitive efficiency justification. We do not
22	do the other side of the scale in terms of is there any
23	anticompetitive effect. In some cases, there will and
24	in some cases there will not be an anticompetitive
25	effect, and, you know, and as I suggested in Dentsply,

163

1	even though there may have been a legitimate
2	procompetitive rationale, forget undivided dealer
3	loyalty in that case, that does not mean that
4	arrangement was not, on net, anticompetitive and harmful
5	to consumers ultimately by creating a barrier to entry
6	to competitors.
7	But there are so many cases out there where we
8	know -- I mean, the case I love is this Joyce Beverages
9	case that I am certain Tad knows about, where you
10	have -- you have RC Cola having their distributor only
11	have one cola, the RC Cola, because they want undivided
12	loyalty. Well, in that case, RC Cola has 5 percent of
13	the cola market and a lot smaller share if you define
14	the market more broadly to have all carbonated drinks.
15	So, in that case, we clearly know there is no
16	anticompetitive effect.
17	But basically there was this mystery in the
18	literature, why are they really having an exclusive
19	dealing arrangement there, because there does not seem
20	to be any specific investments, and there does not seem
21	to be this dealer switching, but what they want is when
22	the salesman goes into the supermarket, that they are
23	going to push RC Cola and not any other brand.
24	So, if you want to get the ultimate question,
25	that would depend upon the particular case, and you

164

1	would have to examine that particular case.
2	DR. FARRELL: But you say you only do the
3	procompetitive justification. What do you demand of it
4	in order to call it a procompetitive justification, just
5	that Toyota would like it?
6	DR. KLEIN: Well, look, I do not want to get
7	caught up in a language thing about --
8	DR. FARRELL: Okay, sorry, no.
9	DR. KLEIN: -- you know, we will do linguistic
10	philosophy later. My feeling -- all I mean by it is
11	that somebody doing what you would consider the right
12	thing or the good thing or something and balancing it, I
13	am just looking -- I am just presenting an economic
14	foundation for this legitimate procompetitive
15	justification.
16	I mean, the crazy thing is if you look in the
17	marketing literature, people are talking about this all
18	the time. It is just economists, you know, a little bit
19	of economics can be a very dangerous thing, and it is
20	only the economists that say competition should give you
21	the services, everything is fine. So, if you talk to
22	business people, marketing people, they all know that
23	this makes -- and it makes a lot of common sense.
24	So, in some sense, as I said, Dentsply was
25	unlucky enough to have the judge that knew economics,

165

1	and that is the only reason they got into problems in
2	terms of the procompetitive justification, plus they
3	were unfortunate enough to choose an expert that
4	explicitly wrote in his article that the argument makes
5	absolutely no sense. So, he could not present -- he did
6	not -- Howard did not present the argument at trial, but
7	the company did in terms of answers to interrogatories,
8	and they said, what are you talking about? Your own
9	expert says this makes no economic sense.
10	And then the other interesting thing about it,
11	and this is the connection between anticompetitive and
12	procompetitive justifications is strange, because the
13	Justice Department -- and Gail would know this -- the
14	Justice Department, in trying to demonstrate the
15	anticompetitive effect, spent all this time in their
16	findings of fact to show how important this dealer
17	channel was to promoting the Dentsply products and how
18	rivals would be at a competitive disadvantage because
19	they did not have access to that channel.
20	So, you just look at all the findings of fact,
21	and it not only demonstrates that there was a
22	significant potential anticompetitive effect, but it
23	also demonstrates that there is a significant
24	procompetitive justification for motivating the dealers
25	to do a good job. So, you have that tension, but

166

1	basically -- well, maybe I should not monopolize this
2	thing.
3	MR. O'BRIEN: I mean, I have a follow-up for Ben
4	on this point. I mean, in talking about Dentsply and
5	just more generally your theory, there were two types of
6	free riding beyond the traditional one that you cited.
7	Dealers can free ride, effectively, on their own
8	promotion on behalf of a manufacturer, right, which
9	maybe they are doing in conjunction with the
10	manufacturer or somehow they have arranged a contract to
11	get that done, and the other one was that dealers may
12	violate this implicit contract by just under-investing
13	rather than free riding by steering customers to a
14	rival.
15	I am wondering if you have any specific evidence
16	that you can cite from the Dentsply case, if your
17	knowledge of the cases is deep enough, that one of these
18	two types of free riding that you identify in your paper
19	was actually present.
20	DR. KLEIN: Well, I do not think -- well, I
21	think the Marvel type free riding was not present, and
22	he did try to present an argument, and I think the facts
23	made it very difficult for him, and it is too bad that
24	he is not on the panel, because he would disagree, so I
25	think the first free riding did not exist, and I think

167

1	the second free riding did not exist basically because
2	there was no switching to rival brands.
3	I mean, I think there was one example where
4	there was some disagreement about whether they tried to
5	switch it to someone else. So, I do not think those
6	other two were there, but in terms of the third one, all
7	the evidence you need for that is that promotion is
8	important for the manufacturer to make sales, and as I
9	said, the Justice Department went to great lengths in
10	terms of trying to demonstrate the anticompetitive
11	effect to demonstrate the existence of that, so that is
12	there, and then all you need in addition to that is that
13	incremental sales are profitable for the manufacturer,
14	and those two conditions were clearly met in Dentsply.
15	DR. CALKINS: The great thing about the Howard
16	Marvel theory is that it is one that we lay people can
17	understand, namely, that it is good for everybody for a
18	manufacturer to run ads saying you can have your hearing
19	improved by getting it tested and going to a dealer and
20	getting this new improved kind of technology to use for
21	your hearing aid, and that drives consumers to go to the
22	dealership to try it out, and that is good for consumers
23	because they are finding out information.
24	It is good for the overall industry because
25	total sales of hearing aids will go up, because all

168

1	these consumers are being driven into the dealership and
2	are getting their ears tested, and it is all sorts of
3	wonderful stuff. And then, if, when the consumer gets
4	there, there is the old bait and switch and they are
5	sent off to buy the el cheapo discount brand, well, the
6	bad consequence of that is there will be less of that
7	advertising about the new, improved technology and, you
8	know, science of hearing aids and such, which is
9	something that is good for the whole industry, good for
10	consumers, good for everybody, it will now be lost,
11	because the manufacturer will not spend money on that.
12	So, you can easily tell a simple layperson
13	story, if you let everybody get switched off, you will
14	no longer have those ads being run, and I look forward
15	to reading the article, but merely saying that -- I
16	mean, if you then say that if you have exclusive
17	dealing, it is good for RC Cola because they are going
18	to make more money and have more sales, well, I can have
19	a warm feeling about that just because they have got a 5
20	percent share, and God knows, if you are RC Cola, you
21	are going up against people that you need all the help
22	you can get to go up against them. So, we can see
23	there.
24	You know, if you were to tell the RC Cola story
25	where you had somebody that had an 80 percent share and

169

1	climbing, before I then sat back and said, boy, I am
2	really concerned about maybe intervening and causing
3	harm here, I would like to at least make sure I
4	understood what is the equivalent of the lost nice
5	advertising that is going to happen if you intervene in
6	that type of situation.
7	DR. KLEIN: Well, Steve -- can I answer it,
8	then? I mean, I agree with you on your main point,
9	Steve, that, you know, with RC Cola, that we can be
10	pretty much assured that inter-manufacturer cooperation
11	or competition is going to pass on these benefits to
12	consumers, but if you are talking about -- the analogy
13	is really identical about lost advertising, because it
14	is either lost advertising by the manufacturer or lost
15	advertising by the dealers. It is just in some cases,
16	it is efficient for the manufacturer to do the
17	promotion, and in some cases, it is efficient for the
18	dealer to do the promotion, and if you do not have the
19	exclusive in one case, you do not get the manufacturer
20	advertising, and in this case, you do not get the dealer
21	pushing the product, and if you think that there is a
22	benefit from lost advertising, then it is totally
23	analogous in the two cases.
24	You know, if you start to do a calculation --
25	and you -- even in the standard case with the

170

1	manufacturer, you know, when the consumers are switched
2	to the discount brand, they almost always pay a lower
3	price. It is not -- you know, they are not being
4	deceived, that they think they are getting the higher --
5	DR. CALKINS: I understand. I think my problem,
6	and I will confess, I was sitting here with my back to
7	the screen, but I understood the Marvel advertising of
8	hearing, you know, development is a good thing. When
9	you hold out as your public good having a used car
10	salesman sit there and harass you into switching from
11	this model to that model, and I being a kid from
12	Detroit, noticing that you are using entirely Japanese
13	brand models, I...
14	MR. O'BRIEN: At the risk of allowing Ben more
15	monopolization time here, I just want to push it just a
16	little bit further on the Dentsply, and if others do not
17	have a question right off the bat here, and you said
18	that this third type of free riding, which is that you
19	would under-invest as a retailer --
20	DR. KLEIN: Right.
21	MR. O'BRIEN: -- you think was present, but the
22	evidence that you cited for that was that investment was
23	important, and that does not seem to demonstrate to me
24	that they actually would have necessarily under-invested
25	but for the exclusive dealing arrangement. You know,

171

1	was there another way for them to ensure that
2	investment? I mean, it seems to be a bit of a leap to
3	me.
4	DR. KLEIN: Well, that's --
5	MR. O'BRIEN: You know, and that sort of
6	statement strikes me as, you know, it might not be hard
7	to argue that that efficiency is there in almost every
8	case.
9	DR. KLEIN: No, that is a problem that you have
10	with these cases. That is why I said you cannot
11	adopt --
12	MR. O'BRIEN: A no economic sense test?
13	DR. KLEIN: Yes, you know, because I think the
14	efficiency is -- I would not say universally present,
15	but it is a motivation. I forgot what your other
16	question was, but, you know, it is important -- if it is
17	important to the manufacturer, we just know from the
18	economics that if there is an exclusive, the incentives
19	are going to be aligned, and if they do not make the
20	sale of that product, they are going to not make any
21	profit. So, you do know that they are going to push it
22	more.
23	So, I mean, it just follows logically, but you
24	would need to see what they adopted -- oh, yeah, so
25	you -- there may be a less restrictive way, and then we

172

1	can talk about a less restrictive standard here if that
2	is the question you want to move to, but in cases where
3	it looks like the practice might have also some
4	foreclosure problems and anticompetitive effects -- I
5	hope I am using the right language -- you may impose
6	this burden on the manufacturer to come up with a less
7	restrictive way of doing it, and, you know, maybe they
8	chose this not just because of the efficiency effects
9	but also because of the -- it increased their market
10	power, so...
11	MR. O'BRIEN: Okay.
12	DR. KLEIN: I mean, it makes it very, very
13	difficult in terms of this balancing. The important
14	point is, you know, you are not going to have these easy
15	cases anymore where there is nothing on -- I mean, you
16	will still have easy cases where you do not have the
17	anticompetitive effect on one side of the scale, but you
18	are not going to have these cases, I think, if you
19	accept this where there is nothing on the other side.
20	MR. O'BRIEN: Okay. Do any other panelists have
21	any questions for any of the other panelists?
22	(No response.)
23	MR. O'BRIEN: If not, we have a series of
24	propositions we would like to walk through to see where
25	we might be able to reach consensus, where we have open

173

1	issues, and hopefully this will spawn some additional
2	conversation about both what was said during the session
3	and perhaps some new things.
4	So, let me start with -- where is our -- oh, you
5	have got it, okay. Let's go to the first proposition.
6	Okay, I am going to read it, because we need to read it
7	for the transcript here.
8	Exclusive-dealing arrangements are analyzed
9	under the rule of reason.
10	First, does everybody agree -- and this is
11	really more for the lawyers -- that that is the way the
12	analysis of exclusive dealing goes today?
13	DR. CALKINS: Yeah. I mean, that -- yes -- yes,
14	I'll agree to say that, and B, for a whole lot of the
15	cases, it is consistent with the general idea that under
16	the rule of reason, the defendant always wins.
17	MR. O'BRIEN: Okay. So, nobody disagrees with
18	that point. Well, perhaps the point that was just made,
19	but nobody disagrees with the proposition, correct?
20	Does anybody think that there are exclusivity
21	arrangements that should be per se illegal?
22	(No response.)
23	MR. O'BRIEN: No, I guess that is the answer.
24	DR. KLEIN: Move on.
25	MR. O'BRIEN: No.

174

1	Does anybody think there are exclusivity
2	arrangements that are always or nearly always
3	procompetitive and thus good candidates for a safe
4	harbor?
5	DR. CALKINS: Well, presumably a very small
6	exclusive would be -- would fit anybody's idea of a safe
7	harbor.
8	MR. O'BRIEN: And when you say "small
9	exclusive," you mean a small percentage of the market
10	or --
11	DR. CALKINS: Yeah, it is very -- it is hard to
12	imagine a court or an enforcer being concerned about an
13	exclusive below -- choose your figures. Some might
14	choose 20 percent, some might choose 30 percent, some
15	might choose 40 percent, but I think everybody would
16	agree that below some percent, no agency should worry
17	about it, and no court should find illegality unless,
18	you know, you have some reason to think that that number
19	is just, you know, totally misleading and the real
20	number will be totally different in six months when the
21	contracts kick in or something.
22	MR. O'BRIEN: Okay, fair enough.
23	Anybody else? That one was --
24	DR. KLEIN: I would like to -- I would like to
25	ask Steve a question on this one. You know, your

175

1	opinion about the foreclosure standard somehow being
2	lower when it comes to Section 2 rather than Section 1,
3	I mean, if somebody is a monopolist or is likely to be a
4	monopolist, I could see that it is more likely that they
5	are going to meet the critical share number, but why
6	should that critical number, whether you say it is 40 or
7	whatever, why should somehow it be lower? It sounded
8	like that is what you said from your presentation,
9	should be a tougher standard.
10	DR. CALKINS: If I did, I misspoke slightly.
11	What I meant to say is that -- well, specifically, is
12	that in the Microsoft case, the defendants argued that
13	because this practice is lawful under Section 1, it
14	must, as a necessity, be lawful under Section 2, and I
15	was just saying that I do not think that is correct,
16	that, you know, take your extreme of a dominant firm
17	that everybody would agree is a monopolist on the one
18	hand, and on the other hand, your RC Cola kind of a
19	thing. I am not saying whether or not, you know,
20	exactly where one would say there is a difference, but I
21	would think that one should be much more likely to be
22	concerned about something being done by a dominant firm
23	that is --
24	DR. KLEIN: Right, obviously, but why should
25	there be a different standard under Section 2 than under

176

1	Section 1? I mean, I think we are in trouble here
2	basically because Justice did not appeal the Section 1
3	no liability in Microsoft and Dentsply, and if you read
4	the Court, you know that the appeals court would have
5	overturned both of those things, but -- you know, and
6	then I think we would be in a lot better shape, but the
7	idea that somehow we should have a different standard
8	and principle when you are doing the first step of a
9	Section 2 -- I agree, if somebody is a dominant firm,
10	they are much more likely to have anticompetitive
11	foreclosure under Section 1 and under Section 2, but why
12	should there be a lower hurdle showing the
13	anticompetitive effect under Section 2?
14	DR. CALKINS: Well, part of this goes -- I mean,
15	in all of this, it is trying to make a judgment about
16	how likely a particular practice is to be harmful to
17	competition, and I was just saying that -- well,
18	specifically, is that there are a whole series of sort
19	of ways that firms with fairly modest market shares have
20	been able to persuade courts to get rid of exclusive
21	dealing cases, but where you have a dominant firm, I am
22	not saying that there is a magic difference. I am just
23	saying that, as you recognized, you would think longer
24	and harder about something being done by a dominant firm
25	that is a clear monopoly than by some firm that is a

177

1	trivial firm, and so just because you are told that
2	something would be lawful under -- you find some Section
3	1 case out there where some foreclosure level was a
4	motion for summary judgment for a defendant, that does
5	not mean that in every case with the most extreme
6	monopolist you would grant summary judgment without
7	thinking long and hard about it.
8	MR. O'BRIEN: Okay, let's move on to the next
9	proposition.
10	Okay, I think this one will be easy, too. The
11	proposition is from Posner's Antitrust Law.
12	I propose the following standard for judging
13	practices claimed to be exclusionary: "In every case in
14	which such a practice is alleged, the plaintiff must
15	prove first that the defendant has monopoly power. All
16	the plausible cases of exclusionary practices involve
17	defendants that have monopoly power."
18	First, does everybody agree with that?
19	MR. LIPSKY: Uh-oh.
20	MR. O'BRIEN: Can exclusive dealing involving a
21	non-monopolist result in a substantial lessening of
22	competition?
23	DR. KLEIN: Yes.
24	DR. FARRELL: All statements containing the word
25	"all" are false except for this one and perhaps a

178

1	handful of others. I think there is a real problem with
2	a subtle, complex and imperfectly understood topic
3	having courts, judges, make grand and sweeping
4	pronouncements. The law, as I understand it, in a
5	precedent-based system tries hard not to change over
6	time, and our understanding tends to change over time,
7	and that creates a lot of trouble. So, it is not like I
8	am out here saying, oh, and the following large category
9	of cases, firms without monopoly power or without market
10	power or something, can do a lot of harm with exclusive
11	dealing. There have been some theories developed under
12	which that can happen.
13	I think the consensus currently is that that is
14	not such a big worry, but we do not really know yet, and
15	freezing stuff in place by grand pronouncements that say
16	"all," I am not sure it is such a great idea.
17	DR. CALKINS: The larger consequence, if that is
18	the law, is that any time a -- well, any time a
19	plaintiff has failed to hire one of these fancy
20	economists and satisfactorily define a market in which
21	the defendant has a well-defined market share of more
22	than 75 or 80 percent, there is a very good chance that
23	a Court would grant a motion for summary judgment or a
24	motion to dismiss, because when you have rules like
25	that, lots of courts operationalize it by saying, okay,

179

1	any market share below 70 percent, I grant a motion for
2	summary judgment and do not explore anything else about
3	what is going on, and that in my judgment is too
4	sweeping a broom to use. That was a bad way to phrase
5	that, wasn't it?
6	MR. O'BRIEN: Okay, Tad?
7	MR. LIPSKY: I think I can agree with the last
8	sentence there, that all the plausible cases -- I am a
9	little confused, though, whether this statement in
10	context, was it limited to exclusive dealing or is it
11	meant to be applied more broadly to other types of
12	exclusionary practices? I guess that there -- you know,
13	I am trying to recall. Wasn't there a -- there were
14	some Commission consent decrees in cases involving water
15	pumps for fire trucks. It was a multiple defendant
16	situation where there was actually a fairly plausible
17	theory of cartelizing, and I do not think you could have
18	found, at least not with any logical consistency, that
19	both of the competitors were monopolists.
20	So, I guess that is a limiting case, but I would
21	be closer to agreeing with this if you were talking
22	about cases other than those in which a cartelizing
23	theory for challenging the exclusive dealing was the
24	theory of liability.
25	Am I right about this FTC decree? Does anybody

180

1	remember that?
2	DR. CALKINS: There was a pump case. There
3	was -- there was a case like that.
4	MR. LIPSKY: Okay, so it is actually -- it is
5	probably real, presumptively real.
6	MR. VITA: It is called Waters Hale (ph).
7	MR. LIPSKY: Excellent, okay, thank you.
8	DR. KLEIN: If Posner had restated it in terms
9	of market power instead of monopoly power --
10	MR. LIPSKY: That would be fine.
11	DR. KLEIN: -- I assume we could all agree,
12	right?
13	MR. LIPSKY: Yes, that would be fine.
14	MR. O'BRIEN: So, this statement is about
15	monopoly power or market power on the part of the
16	defendant. I am wondering if any of you think that
17	conditions relating to market power or market structure
18	in the downstream market have an effect on the extent to
19	which exclusive dealing can be anticompetitive. That
20	was not stated well, but what should we make of the
21	downstream market structure in terms of the likelihood
22	that exclusive dealing can have an anticompetitive
23	effect?
24	DR. FARRELL: Well, I mean, I talked briefly
25	earlier about the developing economics of understanding

181

1	the role of downstream competition in that and, you
2	know, fairly plausible seeming analyses have come out
3	with very different answers so far, so watch this space,
4	and that perhaps should be a pretty strong warning
5	against making strong statements at this point.
6	MR. O'BRIEN: Would you be willing to say that
7	some kind of barrier to entry in the downstream market
8	is necessary for anticompetitive exclusive dealing?
9	DR. FARRELL: Well, I think -- see, you are
10	talking about a lot of abstract nouns here, and I am
11	sorry, I cannot put on a southern U.S. accent, but I
12	would like to.
13	DR. WERTHER: Can you do any U.S. accent?
14	MR. LIPSKY: I thought that was a Berkeley
15	accent.
16	DR. CALKINS: You have got such a lovely accent.
17	DR. FARRELL: I think Strunken White might have
18	said if you are getting confused, try to decrease the
19	abstract nouns and increase the active verbs, and I
20	think that is a pretty good proscription for thinking
21	straight. So, let's try that.
22	Instead of talking about market power and market
23	share and dominance and exclusive dealing and so on,
24	let's ask the following question: If I come up with a
25	better way of doing things than the incumbent is doing

182

1	or I am less greedy than the incumbent and I am willing
2	to give consumers a better deal, am I stymied in my
3	attempt to do so by these deals that people have struck?
4	That is the core question, and a lot of the time, the
5	answer will be no, I am not stymied if there are small
6	shares of this or that. Sometimes I will be.
7	So, for example, if you look at the Microsoft
8	case, Microsoft had no need to completely keep NetScape
9	out and wasn't trying to keep NetScape out and charge a
10	lot of money for Internet Explorer. They just had to
11	make sure that NetScape did not become sufficiently
12	widely distributed that people would start writing to it
13	and say, yeah, I -- that is a rather different case from
14	the one we would generically think of. You have to be
15	careful and I think should be pretty reluctant to kind
16	of lay down these firm rules.
17	Now, having said that, I also am very aware
18	that, you know, attempts to do full-blown rule of reason
19	analysis are also dangerous, right, given the subtlety
20	of what is going on and given the capabilities and
21	noncapabilities of courts.
22	I am a big fan in theory of -- I have never been
23	up close when it has happened -- of court-appointed
24	experts. I think that could probably improve the
25	process a lot quite generally, but especially when you

183

1	are dealing with subtle and difficult issues.
2	DR. CALKINS: Clearly everybody would say that
3	it matters how easy our new entrant can gain access to
4	the customers to whom it is trying to sell, and if it is
5	very easy to do that, then exclusive dealing will not
6	present any problems. As you phrased the question, you
7	used the magic word "entry barriers," and as you know,
8	that has lots of different definitions, and choose your
9	right definition and defendants will almost always
10	prevail; choose different definitions, and they might
11	not.
12	It also raises the question as to whether you
13	are looking at a total exclusion standard or at simply
14	making it much more expensive, time-consuming and risky
15	in order to gain access, and so you have staked out a
16	position or the quote here has staked out a position
17	which might mean things that I would not be comfortable
18	with.
19	MR. O'BRIEN: Right. So, just one follow-up to
20	that, I guess this is directed to Joe, the Fumagalli and
21	Moto models and the Simpson and Wickelgren model, in the
22	simplest cases, you have homogeneous producers
23	downstream with no economies of scale or very small
24	economies of scale, and it strikes me in the context of
25	those models that it would be very easy for a firm to

184

1	enter at both levels and disentangle any anticompetitive
2	effect that is being contemplated. I am wondering if
3	you have thought about that, or maybe I am wrong.
4	DR. FARRELL: You know, it has been a while
5	since I read the models, so I do not remember
6	technically whether what you say is right. Clearly if
7	you really have homogeneous products and fixed costs and
8	sunk costs are very small, then you would think -- and
9	you would want to know why not if somebody was claiming
10	not -- that a firm could enter at both levels.
11	On the other hand, there certainly are
12	industries where at any given time the industry may
13	behave quite competitively involving the pass-through
14	dynamics that we were talking about, and yet there are
15	big sunk costs lying behind it, and that may be the more
16	relevant case for that kind of analysis.
17	MR. O'BRIEN: Anyone else? Okay, next slide.
18	Okay, I think this is an uncontroversial slide
19	as well. We will see. Maybe the questions will be more
20	interesting.
21	"Exclusive-dealing arrangements --" this is a
22	quote from Jefferson Parish. "Exclusive-dealing
23	arrangements 'may be substantially procompetitive by
24	ensuring stable markets and encouraging long-term
25	mutually advantageous business relationships.'"

185

1	Yes, Joe?
2	DR. FARRELL: I hate to be a curmudgeon, but
3	stable markets are not exactly what antitrust aims for.
4	Actually, maybe we should try to encourage unstable
5	markets where the status quo could be disrupted at any
6	moment by some pesky firm that maybe has not shown up
7	before, or maybe has, and is willing to take a lower
8	margin or has a better way of doing things.
9	Now, I am not saying that the basic point here,
10	that exclusive dealing arrangements "may be good" is
11	wrong, but I do not like that language.
12	MR. O'BRIEN: Okay. Well, you pick the --
13	DR. CALKINS: And while you are complaining, you
14	could complain about the mutually advantageous business
15	relationship, because that could be good for consumers,
16	and if it is just dividing up a surplus between two
17	businesses, it could be bad for consumers.
18	DR. KLEIN: Yeah, I --
19	MR. O'BRIEN: Ben Klein, do you have a view on
20	that?
21	DR. KLEIN: Well, who knows what Justice
22	O'Connor is referring to, but if she means by
23	encouraging long-term mutually advantageous business
24	that it encourages people to make specific investments
25	in the relationship, relationship-specific investments,

186

1	then I think she is correct and that she should not go
2	through it now, but that is one of the problems I had
3	with Joe's presentation, is that the Segal and Whinston
4	criticism of that rationale for exclusive dealing is
5	just wrong, and it is logically correct, but there is
6	assumptions being made in that that are very, very
7	unrealistic, and in particular, they are just -- well, I
8	better not go into it.
9	But, you know, so if that is what she is saying,
10	I would agree with her very much, but it is so vague,
11	right, but if she is just saying there is -- it
12	sometimes may be good...
13	MR. O'BRIEN: What efficiencies, which I assume
14	are the second object of this sentence, there are
15	numerous efficiencies that have been discussed about
16	exclusive dealing that we might classify into that
17	second phrase. What are the most significant and most
18	likely in an exclusive dealing arrangement?
19	And similarly, what efficiencies have been
20	asserted most often do you think are least likely to
21	actually exist?
22	DR. CALKINS: Oh, the best is the classic Marvel
23	free riding, manufacturers spending money, bringing in
24	the customer, then there's the old bait and switch to
25	the other product. That would be the classic and the

187

1	best.
2	MR. O'BRIEN: So, what efficiencies are often
3	asserted in exclusive dealing cases that you think may
4	not actually exist very often? Anybody?
5	DR. KLEIN: I hope nobody says this focused
6	dealer effort, but I guess one of the things I should
7	say is the justification that Microsoft offered, the
8	procompetitive justification for the exclusive dealing
9	arrangement with the Internet access providers, sounded
10	like a focus -- the way you presented it, it sounded
11	like a focused dealer incentive, but what they wanted
12	was -- the argument they presented was something to the
13	effect that they wanted the developers to focus on the
14	Windows APIs, which meant they wanted to have a monopoly
15	in Windows so that when developers were developing their
16	programs, they would only develop Windows programs,
17	which is a very different argument than, you know, you
18	want -- they did not want the Internet access providers
19	to promote their product. That is not what they were
20	doing.
21	They were talking about a different type of
22	focus there, but that argument I think the Court
23	correctly rejected as making no sense other than you
24	want a monopoly. You want to maintain your monopoly.
25	MR. O'BRIEN: What significance, if any, should

188

1	be given to observing an exclusive dealing arrangement
2	in a similar competitive market when you are analyzing a
3	case where there is exclusive dealing, maybe in a market
4	that exhibits some more market power in some ways than
5	the other, but otherwise has similarities?
6	DR. FARRELL: Well, at a technical level, there
7	certainly have been analyses that show that in some
8	circumstances, exclusive dealing engaged in by, let's
9	say, all members of an oligopolistic manufacturing
10	sector, whether downstream industry, can soften
11	competition and be in that sense anticompetitive, even
12	conditional on, you know, a flourishing oligopoly
13	structure, and let's face facts, we are never dealing
14	with perfectly competitive industries when we are
15	talking about these cases, so oligopoly is what you mean
16	by the word "competitive" here.
17	There are other analyses that suggest that
18	exclusive dealing can actually sharpen competition. I
19	think it is fair to say that that literature is both
20	unsettled and in a state of nonferment, the nonferment
21	because nobody seems very excited about it. People are
22	really more interested in the monopoly-preserving
23	possibilities I think than the oligopoly-softening
24	possibilities, and that may be a legitimate choice of
25	emphasis, where to put our intellectual resources, or it

189

1	may just be, you know, what happens to be fun for
2	assistant professors to do these days.
3	DR. KLEIN: I think we have to be very careful
4	when we start talking about oligopoly-softening, and I
5	guess Joe would say I have this bias, this laissez-faire
6	bias, but I can imagine unilateral behavior -- you know,
7	a gasoline company decides they are going to locate
8	their station not next to another station but a couple
9	of blocks away, because if they locate it next to the
10	station, it is going to be more intensive competition.
11	People are going to be able to compare the prices.
12	We do not want to go in and micro-regulate the
13	competitive process. You know, you hire an economist,
14	and let's assume they draw the welfare triangles, and
15	they say consumers are better off if that person puts
16	the station next to the other station, and even though
17	it has -- let's assume it has the effect of sharpening
18	competition if we do that, we do not want to regulate
19	that behavior, at least I do not want to, even though
20	the calculation would come out that way.
21	So, I think it is dangerous to start talking
22	about oligopoly-softening of competition in general, and
23	basically I guess I have a prior that we are just going
24	to mess things up and we should just leave it up to the
25	competitive process, unless there is a -- you know, you

190

1	have this first step where you need some major
2	anticompetitive effect in terms of foreclosure.
3	So, I guess my comment was not totally
4	irrelevant, because we are talking about Section 2
5	unilateral behavior, even though it has nothing to do
6	with exclusive dealing.
7	DR. CALKINS: Trying to psycho-analyze your
8	question, I think you were -- I am guessing that you
9	were referring to the argument you sometimes see made
10	that, look, over here in this market, which we all
11	stipulate is competitive, this practice is occurring,
12	and so, therefore, it must follow as the night follows
13	the day that when that same practice is being engaged in
14	by this complete and total monopolist, it deserves
15	summary judgment very promptly on that ground alone,
16	and --
17	MR. O'BRIEN: That is a good psycho-analysis.
18	Yes, that is what I was hoping somebody would address.
19	DR. CALKINS: And I myself do not buy into that
20	theory in the little that I have done thinking about it,
21	but my thinking is still at a preliminary stage.
22	MR. VITA: Well, it is not so much -- maybe,
23	Dan, a competitive market versus a noncompetitive
24	market, but the individual -- the size of the firm or
25	the mark -- the firm's specific market power. Like the

191

1	RC Cola example somebody alluded to before, RC has some
2	exclusive relationship with its bottlers or something, I
3	think it was, and you look at RC Cola, and they are a
4	small fry. I mean, they do not matter anywhere. So,
5	you look at that and you say, well, obviously they are
6	doing that. They cannot possibly have any kind of
7	foreclosure mode or some monopolization motive. It has
8	to be some sort of value creation that induces them to
9	do that.
10	Is it fair to say that when you do -- then you
11	look at Coke, for example, maybe doing the same kind of
12	thing, some other firm with substantial market share or
13	market power possibly? At least it says you have got to
14	consider the efficiency story. You can't rule it out.
15	There is a possibility that there is value creation,
16	that there is something inefficient about it, but not
17	necessarily -- the fact that RC does it doesn't
18	vindicate Coke's usage, that debate is not over, but
19	that does say to you -- you know, we have got to take
20	that seriously.
21	DR. FARRELL: Yeah, I think you said it right,
22	you know, unless there is something about that industry
23	or market that I do not know, you can presumably infer
24	from RC's use of these exclusives that there is
25	something other than monopoly preservation going on, but

192

1	that does not mean that there is not monopoly
2	preservation going on.
3	DR. KLEIN: Exactly.
4	DR. FARRELL: It does not mean there is either.
5	MR. LIPSKY: Thanks.
6	MR. O'BRIEN: Okay, next slide.
7	Okay, so anticompetitive effects, this is a --
8	this is actually a quote from Dennis Carlton's paper on
9	the Aspen and Kodak case.
10	"In the presence of scale economies, exclusive
11	dealing can be a way of depriving Firm 2 (or its
12	distributors) of the necessary scale to achieve
13	efficiencies, even though, absent the exclusivity, Firm
14	1 and Firm 2 would both be large enough to achieve
15	efficiency."
16	So, this is the standard scale economy argument
17	about excluding your rivals so that it cannot reach
18	efficient scale, and I guess my question is, does the
19	panel see that as the primary anticompetitive theory of
20	exclusive dealing that we ought to be focused on?
21	MR. LIPSKY: Well, I will take a stab at that.
22	Certainly, you know, in a static sense, it is hard to
23	argue with this proposition, and I think this is
24	consistent with the notion that there are stories
25	associated with exclusive dealing where you are trying

193

1	to compel two-stage entry basically, and I think some of
2	those are good stories. Probably we would not agree on
3	which ones were good stories.
4	I heard John Jacobson the other day talking
5	about Pullman, and I disagreed with him on that one, and
6	then Motion Picture, and I disagreed with him on that
7	one, but it does not -- and also United Shoe Machinery,
8	and I disagree with him on that one, but I think we
9	could find -- I think we could find a two-stage entry
10	story that held together, and so I would say I agree
11	with this.
12	But I would also interject -- and I have said in
13	other contexts -- there is kind of an endemic temptation
14	or tendency in the system, in the investigation and the
15	litigation system, to underestimate supply flexibility.
16	I mean, you know, supply flexibility is not -- or new
17	entry is not always an answer, and so I would hate for
18	my remarks to be misconstrued. There are industries in
19	which the barriers to entry are such that if you have a
20	two-stage story, it is a serious problem, but I think
21	there is a tendency to look at what is right in front of
22	you, to, you know, fail to predict the rise of the
23	Internet or the mobile phone, you know, falling in price
24	by 75 percent over five years or, you know, some other
25	alarming and unpredicted new technology or new

194

1	development, and because the dynamic aspect is so
2	important, I think this is a theme that needs to be
3	hammered again and again.
4	So, what I guess I am saying, yes, I agree with
5	this, but it is narrow -- I would like to make my
6	agreement as narrow as humanly possible.
7	MR. O'BRIEN: Anybody else?
8	DR. KLEIN: Tad, you sounded like an expert
9	witness there.
10	DR. CALKINS: I was hoping that Tad could tell
11	me how to get a mobile phone bill that is 75 percent
12	lower.
13	MR. O'BRIEN: So, Joe, based on your remarks, I
14	guess I would ask, do you think this is the primary
15	story of competitive harm that we should be focused on
16	in analyzing exclusive dealing, or should some of the
17	other theories that you mentioned, I guess in particular
18	Simpson/Wickelgren, maybe some of these two-stage models
19	of oligopoly where exclusive dealing can play a role,
20	are those things we should be concerned about, or is
21	this number one and number two?
22	DR. FARRELL: Well, I disagree with the
23	question. I think the primary focus should be based on
24	what is going on in the market at hand, and we should
25	adjust the tools to fit the facts and not prejudge what

195

1	the theory is going to be.
2	Having said that, I think I said in my earlier
3	remarks that I believe this Rasmussen, Ramseyer and
4	Wiley or Segal/Whinston theory, which is being referred
5	to here, is the one that people talk about most. I tend
6	to suspect that it is the main one. I would add -- I
7	mean, you have to interpret efficiencies carefully, so,
8	for example, scale to fully reward innovation, is that
9	achieving efficiencies?
10	But broadly speaking, I think this is what most
11	economists think of most of the time when they think
12	about anticompetitive exclusive dealing, and I think
13	that may well be right, but I think we should be open to
14	whatever the facts of a particular case say.
15	MR. O'BRIEN: All right. Anybody else?
16	(No response.)
17	MR. O'BRIEN: Let's go to 7.
18	Okay, this is from the Microsoft case, and the
19	quotation is:
20	"If the monopolist's procompetitive
21	justification stands unrebutted, then the plaintiff must
22	demonstrate that the anticompetitive harm of the conduct
23	outweighs the procompetitive benefit."
24	I guess my question is -- well, first, does that
25	make sense to you, and secondly -- this is maybe more

196

1	for the economists, although equally for the lawyers --
2	does economics supply tools to do this?
3	DR. KLEIN: Try Joe.
4	DR. FARRELL: Well, let's see. I mean, clearly
5	in order to plunge into enforcement, we would not want
6	to go ahead if the anticompetitive harm of the conduct
7	is outweighed by the procompetitive benefit. Using the
8	term "procompetitive benefit" in -- I am not sure
9	whether it is the same way or not as Ben uses it, but I
10	am using it to mean actual benefits to efficiency and
11	consumers, not just kind of non-anticompetitive
12	rationales.
13	This, of course, is part of a bigger decision
14	tree that the Microsoft Court laid out. In thinking
15	through a burden-shifting process like that, you have to
16	think about a number of things, and I do not know how
17	much the Court thought through these things. I am
18	pretty sure I know how much they knew the necessary data
19	required to do it exactly right, which is not a
20	criticism, because nobody has that data either.
21	You have to think both about whether in most
22	cases this is true or that is true, but also about if
23	this is true, is it going to be easy to prove, or is it
24	quite likely to be true but be hard to prove? And that
25	really gets back to what I hope was the main theme that

197

1	came out of my talk earlier, that in my opinion, there
2	are often benefits of open, free-wheeling competition
3	that are very difficult to pin down and almost
4	impossible to prove, and I think that needs to be kept
5	in mind when we lay down these decision trees.
6	Did the Microsoft Court keep that in mind? To
7	some extent. Did it do it the right amount? I have no
8	idea, and I doubt that they really know either.
9	DR. CALKINS: If the question is should one
10	think about the competitive harm that is likely, should
11	one think about the procompetitive benefit, the answer
12	to that is entirely yes.
13	On the other hand, can you read this statement
14	to say that if there is any tiny procompetitive benefit,
15	perhaps using anybody's definition of "procompetitive,"
16	does that mean that the defendant always wins unless the
17	plaintiff is able, with great specificity, to precisely
18	quantify the anticompetitive harm, precisely quantify
19	the anticompetitive benefit, and then precisely
20	calculate that one is more than the other?
21	Well, it may well be that if that is what one
22	means, then what one is saying is that any time there is
23	any benefit that can be characterized as procompetitive,
24	the defendant will always win, and so if that is where
25	you ended up, that might not be a good place, but that

198

1	does not mean that you should not think about the
2	procompetitive benefit.
3	DR. KLEIN: Go ahead, Tad.
4	MR. LIPSKY: No, go ahead.
5	DR. KLEIN: No --
6	MR. O'BRIEN: Go ahead, Tad.
7	MR. LIPSKY: Well, I was just going to say that
8	we always have to consider the fact, you know, there was
9	a day not so long ago when you could expect a follow-on
10	litigation from cartel cases that were litigated and won
11	by the Department of Justice. You would get a guilty
12	plea in a price-fixing case, and then we transitioned --
13	I am not sure exactly what the history is or how we got
14	here, but then we got to the point where there was a
15	story in a newspaper saying that there was a
16	price-fixing investigation, boom, 80 private class
17	action -- purported class action treble damage suits
18	against everybody in the industry, and then we got to
19	the state where there -- you get the same thing even in
20	these conduct type cases, which are not cartel cases,
21	and there are follow-on class actions for Dentsply, and
22	there were follow-on class actions for this, that and
23	the other outside of the price-fixing area, and that
24	combined with, you know, indirect purchaser statutes and
25	all kinds of things that happen in antitrust litigation

199

1	generally I think creates the fear that there are some
2	legitimate procompetitive practices that the perpetrator
3	cannot afford a defense, and I think that is a very
4	troublesome phenomenon.
5	I guess the thought is provoked by Joe's comment
6	that there are -- you know, there is sort of a -- maybe
7	we should indulge a presumption that when things are
8	loosened up a little, and there are fewer strong ties,
9	you know, partial vertical relationships, maybe that is
10	the way we want markets to function, but I think the
11	system in general works pretty well if we require -- you
12	know, we always have the ultimate burden of proof on the
13	plaintiff, so that if the defendant can come up with a
14	sensible justification, a justification that can be
15	persuasive with the fact-finder, then yes, the right
16	standard is, if the defendant has something good to say
17	for his practice, let's adopt a rule that the plaintiff
18	does not win unless the plaintiff persuades that the
19	negative effect on competition outweighs the
20	procompetitive effect.
21	And true enough, part of what I was saying
22	earlier is, yes, it is the wiles of economic theory. It
23	is the unadministerable, you know, battle between the
24	economic experts and all the other facts in the case,
25	but what is the alternative? The alternative is

200

1	Standard Stations, or worse, and we know that is wrong,
2	so that is why I would like -- I keep trying to bring
3	into the conversation this institutional element.
4	Let's not -- once we decide it is a balance,
5	let's not just throw confetti in the air. Let's try to
6	focus on what the applied micro tells us about what
7	rationales deserve to be explored and what facts could
8	rule various theories of efficiency or theories of
9	restraint in or out. Let's organize that process so we
10	do not just have a U.S. versus IBM every time there is,
11	you know, a 13-year slog or a 14-year slog like Harmar
12	versus Coca-Cola every time we have a difficult
13	exclusive dealing issue.
14	DR. CALKINS: I really misunderstood you, Tad.
15	I thought when you said you wanted to go to the 18-month
16	model, you wanted to go back to the days of Standard
17	Stations, and I just --
18	DR. KLEIN: Per se. But to answer your question
19	about whether we have the tools to do this, I guess
20	economists have the tools -- I was on a panel with Steve
21	Salop where I said I -- even if I were the judge, I
22	wouldn't know exactly how to do it, and he said, you
23	know, that is all economists know how to do, you know,
24	want to take away your doctorate or something, but when
25	you -- obviously you have to go to balancing.

201

1	I mean, I am pretty cynical about this, because
2	I do not know -- I do not think the courts have done
3	this, and I do not know what to tell them to do. I
4	mean, I think they go backwards, and they figure out --
5	you know, they do some kind of implicit balancing, and
6	then they say -- they make it easy and they say it was
7	not an anticompetitive effect or there is no
8	procompetitive efficiency rationale, and I do not know
9	what exactly we should have them do, other than we know
10	we want them to hire more economists, right?
11	But it is a -- I think that is the ultimate
12	question, because you do have to do the balancing, and I
13	do -- I mean, it is a legal question, but I do think the
14	burden should be placed on the plaintiff at that point,
15	because I have this prior bias about the competitive
16	process. So, I agree with the legal rule, but then what
17	exactly are you doing -- and it should -- it should not
18	be a close thing, because that is my -- and I think that
19	is the way the law is or it should be, that it should
20	not be a very close thing that we are balancing, and it
21	should not be something -- you know, there should be
22	this first step that you have to show a very clear
23	anticompetitive effect before you go forward in any way,
24	and that is going to get rid of most of the cases.
25	Steve will say that is why the defendants win

202

1	all the time, but they do not always win, because you
2	have the Dentsplies and you have the Microsoft, and I
3	think that is enough to get efficiency in the economy.
4	DR. FARRELL: There is this article by Priest
5	and Klein -- I do not know if that is you --
6	DR. KLEIN: Yes, that is me.
7	DR. FARRELL: -- saying that whatever the rules
8	are, the litigated cases are going to be close ones.
9	So, I do not think we can have a rule that litigated
10	cases are not allowed to be close.
11	MR. O'BRIEN: Okay, well, we have run past our
12	time, and I think it is Ben's fault, by about four
13	minutes. So, thank you very much everybody.
14	(Applause.)
15	(Whereupon, at 4:04 p.m., the hearing was
16	concluded.)
17
18
19
20
21
22
23
24
25

203

1

C E R T I F I C A T I O N O F R E P O R T E R

2

DOCKET/FILE NUMBER: P062106

3

CASE TITLE: SECTION 2 HEARING

4

DATE: NOVEMBER 15, 2006

5

6

I HEREBY CERTIFY that the transcript contained

7

herein is a full and accurate transcript of the notes

8

taken by me at the hearing on the above cause before the

9

FEDERAL TRADE COMMISSION to the best of my knowledge and

10

belief.

11

12

DATED: 12/4/2006

13

14

15

16

SUSANNE BERGLING, RMR-CLR

17

18

C E R T I F I C A T I O N O F P R O O F R E A D E R

19

20

I HEREBY CERTIFY that I proofread the transcript

21

for accuracy in spelling, hyphenation, punctuation and

22

format.

23

24

25

DIANE QUADE

Updated January 2, 2024