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

1

1UNITED STATES FEDERAL TRADE COMMISSION

2and

3UNITED STATES DEPARTMENT OF JUSTICE

4

5

6

7SHERMAN ACT SECTION 2 JOINT HEARING

8UNDERSTANDING SINGLE-FIRM BEHAVIOR:

9EXCLUSIVE DEALING SESSION

10WEDNESDAY, NOVEMBER 15, 2006

11

12

13

14

15HELD AT:

16UNITED STATES FEDERAL TRADE COMMISSION

17601 NEW JERSEY AVENUE, N.W.

18WASHINGTON, D.C.

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

20

21

22

23

24Reported and transcribed by:

25Susanne Bergling, RMR-CLR

2

1 MODERATORS:

2DAN O'BRIEN

3Chief, Economic Regulatory Section

4Antitrust Division, Department of Justice

5and

6MICHAEL G. VITA

7Assistant Director

8Bureau of Economics, Federal Trade Commission

9

10PANELISTS:

11

12Morning Session:

13Jonathan M. Jacobson

14Howard P. Marvel

15Richard M. Steuer

16Mary W. Sullivan

17Joshua D. Wright

18

19Afternoon Session:

20Stephen Calkins

21Joseph Farrell

22Benjamin Klein

23Abbott (Tad) Lipsky

24

25

3

1C O N T E N T S

2

3MORNING SESSION:

4Introduction

5Presentations:

6   Jonathan M. Jacobson

7   Howard P. Marvel

8   Richard M. Steuer

9   Mary W. Sullivan

10   Joshua D. Wright

11Moderated Discussion

12Lunch Recess

13

14AFTERNOON SESSION:

15Introduction

16Presentations:

17   Stephen Calkins

18   Joseph Farrell

19   Benjamin Klein

20   Abbott (Tad) Lipsky

21Moderated Discussion

22Conclusion

23

24

25

4

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

2- - - - -

3MR. VITA: Good morning, everybody. My name is

4Mike Vita. I am an economist here at the Federal Trade

5Commission. My title is Assistant Director for

6Antitrust in the FTC's Bureau of Economics. My

7co-moderator is Dan O'Brien, Chief of the Economic

8Regulatory Section at the Department of Justice,

9Antitrust Division.

10I am going to be leading the morning session,

11and Dan will be leading the afternoon session, and

12before we get started with the substance of today's

13hearings, I am going to cover a few housekeeping

14matters.

15First, turn off the cell phones. You'll get

16detention if you -- the BlackBerries and any other

17devices that make noises, that's very important.

18Second, for those of you who aren't familiar

19with the setup here at 601 New Jersey, the rest rooms

20are down the hall, past the guard's desk and to the

21left. I think there are signs out there in the lobby to

22guide you.

23Third, a safety tip particularly for visitors.

24In the unlikely event that the building alarms go off,

25which they actually did yesterday, please proceed calmly

5

1and quickly as instructed. Dan and I will keep

2everything calm and orderly. If we must leave the

3building, exit the New Jersey Avenue exit by the guards,

4that's where you probably came in, and follow the stream

5of people running to a gathering point where you can

6await further instructions.

7Finally, we request that you not make any

8comments or ask questions during the session. Thank

9you.

10Okay, today's session concerns exclusive

11dealing, one of the most interesting areas I think of

12all the various topics involving vertical restraints and

13vertical contracts. It has been an active area of

14economic research and an active area of antitrust as

15well. We are honored to have assembled a distinguished

16panel of practitioners and professors who are very

17knowledgeable in the issues we are going to tackle

18today, and there are going to be two sessions, one in

19the morning and then one later in the afternoon.

20I will just briefly introduce the panelists for

21this morning before we get started, and I will give a

22little more detailed introduction as each speaker takes

23his or her turn. I do not know if everybody is in some

24sort of order, but it looks like they are.

25Okay, so immediately to Dan's left is Richard M.

6

1Steuer, who is a partner at Mayer Brown Rowe & Maw, LLP.

2Next to Richard is Mary Sullivan, who is an Assistant

3Professor of Accountancy at George Washington

4University. Next to Mary is Josh Wright, who is

5Assistant Professor of Law at George Mason University

6School of Law. Next to Josh is Howard Marvel, who is a

7Professor of Economics in the Department of Economics at

8Ohio State and also Professor of Law in the Michael

9Moritz College of Law at Ohio State University. And at

10the very end is Jonathan Jacobson, who is a partner at

11Wilson Sonsini Goodrich & Rosati and a Commissioner of

12the Antitrust Modernization Commission.

13So, I think we will just get right into it, and

14let me introduce in detail our first speaker, and in

15those handouts that you got, there is a more detailed

16biographical description of each of the speakers as

17well, and you can also find them on the FTC and

18Department of Justice web sites.

19Our first speaker is Richard Steuer, who is a

20partner at Mayer Brown Rowe & Maw, where he specializes

21in the practice of antitrust law, including litigation,

22mergers and acquisitions, intellectual property

23licensing, franchising and e-commerce. Richard has

24written a book and several articles on antitrust law

25which have appeared in various journals throughout the

7

1country. For three years Richard served as chair of the

2Antitrust Committee of the Association of the Bar of the

3City of New York.

4Richard?

5MR. STEUER: Thanks, Joe.

6In baseball they say you can learn a lot by

7watching, and I have been fortunate over the years to

8have been able to observe a great deal about exclusive

9dealing and in various contexts, both in litigation and

10counseling, and I put what I knew into three articles

11that I have written, and I thought that the best way to

12try to present what I have learned about exclusive

13dealing would be to go through those articles and

14briefly outline what it is that I have learned from

15watching.

16The first one was an article on "Exclusive

17Dealing in Distribution," focusing on how exclusive

18dealing works when you are talking about selling to

19resellers, and this appeared in 1983. I will not take

20very much time on the history, but it is interesting

21that once upon a time, the FTC considered most exclusive

22dealing to be virtually per se unlawful. The Standard

23Stations case in 1949 introduced the rule of

24quantitative substantiality. Then the major case of

25Tampa Electric in 1961 brought in qualitative

8

1substantiality, and then we found a more nuanced rule of

2reason approach with the Beltone case from the FTC in

31982, Jefferson Parish in the Supreme Court in '84, and

4added to that are the nuances of rule of reason analyses

5we get from California Dental.

6Now, what I have found is the level of

7distribution really matters in assessing the impact of

8exclusive dealing. What we are measuring with exclusive

9dealing -- why exclusive dealing is different from other

10restraints -- is that we are looking more at foreclosure

11of competitors than anything else. Exclusive dealing is

12interesting among the vertical restraints. This is the

13one that, although it has almost always been a rule of

14reason offense, plaintiffs win quite often, and what we

15are looking at is something quite different than in

16vertical resale restraints where the restraint is on

17reselling rather than purchasing. Exclusive dealing is

18a restraint on purchasing, not on selling.

19So, the level of distribution could be

20wholesalers. One wholesaler can reach every retailer in

21America, potentially. With retailers, it is different.

22Retailers are chained to a location typically, although

23with the Internet, that is not quite as true anymore,

24and this is a fluid field. Retailers could be in

25chains, but basically they have a universe of consumers

9

1that they reach. Wholesalers are a little bit

2different, because foreclosing wholesalers does not mean

3that you are foreclosed from reaching retailers.

4Foreclosing retailers may or may not mean that you are

5foreclosed from reaching end users. Reaching end users

6is the simplest. To the extent that there is an

7exclusive dealing arrangement tying up 10 percent of end

8users, you have got 10 percent of the market.

9Type of product is important. Shopping products

10are products for which consumers will go from place to

11place to compare prices, to compare features. The fact

12that each dealer only has one brand does not necessarily

13have as much of a foreclosure effect, because consumers

14will not stop at that dealer. They are more likely to

15go and continue shopping, looking at other brands at

16other dealers.

17Convenience products, on the other hand, include

18impulse products, products that a consumer is more

19likely to buy because he or she is at the retailer, and

20that goes to the concept of "can the retailer deliver

21customers?" Is the retailer such that, when you think

22about the nature of the retail operation, a customer

23going to that retailer is going to buy whatever brand

24there is, so that exclusive dealing is going to have a

25more considerable impact.

10

1Another variable that is important to keep in

2mind is alternate channels of distribution -- what is

3sometimes called intertype competition -- and there was

4a rather classic book that Palamountain published in

51955 on that. Today, the variation in intertype

6competition is richer than ever with the rise of the

7Internet and other alternate channels. So, one needs to

8look, when you are dealing with resellers, at what other

9types of means are there, direct sales and so forth, for

10getting the product distributed.

11Another possibility is simply establishing new

12distributors. Is it more efficient, is it more

13competitive, to have competitors with other brands

14establish their own distribution networks than just

15piggyback on the existing distribution network and

16possibly compromising the amount of vigor with which the

17intermediate, the reseller, is pushing each brand? Are

18you better off having one brand at each reseller and

19having them competing against one another?

20Foreclosure is measured in many, many antitrust

21defenses. There is a measure of foreclosure for

22monopolization, for attempted monopolization, under

23Section 3 of the Clayton Act, under Section 1, and I

24recently had an opportunity to study what the different

25tests are, and I will not belabor the point here -- we

11

1do not have time -- but they are all over the lot.

2The interesting thing is "foreclosure" is a term

3that is used throughout the antitrust lexicon, but it

4has a different meaning with each substantive offense,

5and that is important to keep in mind.

6The procompetitive effects when you are going

7through distribution: Combating manufacturer-level free

8riding. This is not the kind of free riding that we

9were talking about in a case like Sylvania where one

10retailer free rides on the efforts of another. This is

11one manufacturer free riding on the efforts of another

12manufacturer, and exclusive dealing, by keeping other

13manufacturers out of a particular wholesaler or

14retailer, prevents that.

15Of course, stimulate distributors. If the

16distributor only has one brand of a product, it is going

17to devote all of its efforts to that brand, but again,

18in measuring how valuable that is, there is a

19distinction between commodities and differentiated

20products. With a differentiated product, there is

21something more for the dealer to explain, typically,

22about the features of the product. With commodities,

23that is probably less so.

24Stimulating suppliers. Exclusive dealing also

25stimulates suppliers to put more time and effort and

12

1money behind their channels of distribution, because

2they know that other brands are not using the same

3retailer or same wholesaler, and they do not have to

4worry about divided loyalties where they are wasting

5their effort.

6Protecting trade secrets is similar. To the

7extent that a manufacturer is providing trade secrets to

8a retailer or a wholesaler on how to sell, if that

9retailer or wholesaler is carrying other brands, it can

10use that kind of information for the benefit of the

11other brands.

12Quality control as well is something that can be

13controlled more directly with exclusive dealing where

14there are not other brands in the house, and that is

15particularly true where retailers or wholesalers are

16doing things with the product, to the product, where, if

17there is some kind of adulteration, it is hard to

18control quality with other brands in there.

19Resale restraints. There is a lot of talk and

20we were talking earlier about whether there is going to

21be a change in the rule on resale price maintenance.

22Some of these same considerations also go into the kind

23of resale restraints we looked at in a case like

24Sylvania, customer restraints, territorial restraints,

25resale price maintenance, but those are all restraints

13

1on selling, not on buying. So, some of these apply, but

2they do not apply in the same way.

3The next thing I looked at ten years later was

4"Discounts That Induce Exclusive Dealing," and this is a

5little bit different again, but yet another nuance. I

6started with single products. In the simplest case,

7there is one product involved, the grand daddy of the

8cases is United Shoe Machinery, 1922, but these cases

9still continue. The latest one, and I am not going to

10dwell on cases, but there is a case this year from the

11Sixth Circuit that the plaintiff won on essentially a

12single product. Big cases out of the U.S. were

13Nutrasweet, which involved one product, and Tetra Pak,

14packaging.

15The important thing to know in these cases is

16whether or not there is an offer you cannot refuse.

17These are discounts to induce exclusive dealing. It is

18not an outright exclusive, but it is basically a deal

19saying if you buy 50 percent of your requirements from

20me, you get one price; if you buy 75 percent, you get

21another price; if you buy 100 percent, you get still

22another price. It does not sound like it is quite as

23much foreclosure as exclusive dealing, and in many

24cases, it is not as much foreclosure, it is perfectly

25fine.

14

1However, sometimes it is essential for the buyer

2to buy some of the product from one brand, and a classic

3case, we talked about learning from observing, there was

4one case that I was involved in where it was almost a

5commodity product. It was a fairly undifferentiated

6product, but it was differentiated in certain quality

7aspects, and because the buyers had to buy a particular

8brand to satisfy their customers, because it was spec'd

9in, there was one company that had 100 percent of the

10manufacturing. When a second company came along and was

11about to turn the key to open their factory, the first

12company came up with a discount schedule, that as long

13as you bought 80 percent from me, you got a much lower

14price. If you only bought 79 percent from me, you got a

15much higher price.

16Well, it turned out that about half of what all

17the customers needed they could not buy from anyone

18else, not because one product was better than the other

19or even very different, but it was spec'd in, they had

20to have it, and so it was an offer they could not

21refuse, because if they bought less than 80 percent,

22they would be paying a lot more for everything that they

23bought. The company that would be trying to break into

24the market would have to replace all of those lost

25discounts on the quantity that they could not have. So,

15

even though it was not really a different product,
1

2analytically, it almost was a different product, because

3there was some quantity that they had to have from the

4other brand.

5A little like bundling. Bundling is almost

6easier to see, because there are different products in

7the bundle. Some of them are products you have got to

8have because they are patented in some cases. Sometimes

9you do not have to have them, and there are ways of

10ameliorating it. I am not going to spend time on

11bundling, because I know you have another program

12devoted to that entirely, and I could spend a whole day

13on bundling.

14The last thing I looked at was, who is

15instigating exclusive dealing, and should it make a

16difference? And particularly, "Customer-Instigated

17Exclusive Dealing." There are mixed motivations on how

18many suppliers you would like to have in the market.

19End users have two different motives. On the one hand,

20they would like to assure that there are plenty of

21suppliers, because they would like to have alternatives,

22and they want to play one supplier off against another

23to get the best price. At the same time, there may be

24cases where if there is a requirements contract -- and a

25requirements contract not only means I will buy

16

1everything from you, but the seller promising I will

2supply everything that you need -- if one buyer can get

3a requirements contract and there are not enough other

4sellers to go around, it could have an impact harming

5competitors of the buyer. So, it is possible that there

6are situations where an end user would have a motive, at

7least in the short term, not to have as many suppliers

8survive.

9Resellers, it is somewhat similar. In the short

10term, if you are an exclusive reseller of a particular

11brand, you would like to see all the other brands

12disappear. They only provide competition to you. In

13the long term, though, if that arrangement is not

14necessarily perpetual, the day may come when you would

15like to have some options with other brands that could

16supply you.

17Now, why would a customer want exclusive

18dealing? The most obvious reason is to induce lower

19prices, to say to a supplier, I am giving all of my

20business to one supplier, and it may be you, but it may

21not be, so sharpen your pencil and give me your best

22price.

23Another reason is to assure a dependable supply,

24and that is the requirements contract. Another is to

25assure quality, in that it is expensive to qualify

17

1suppliers in certain very technical industries, and you

2do not want an unlimited number of them. In some cases,

3assuring uniformity is important. There is a case

4involving auto racing where it was felt to be important

5that everybody have the same tires so that there is a

6level playing field among competitors. And achieving

7logistical efficiencies. In some settings, just having

8fewer suppliers is going to wind up lowering expenses.

9Now, how do you find an appropriate legal

10analysis where it seems that the buyer has instigated

11the exclusive dealing? The supplier's objectives often

12are twofold. One is to foreclose others, and that is

13the one we always look at when we are trying to see an

14impact on competition -- will exclusive dealing

15foreclose other suppliers from having customers or

16having distribution? Another is to achieve

17distributional efficiencies.

18The reseller's objectives are the ones we just

19talked about, pricing, supply, quality, uniformity --

20and there are mixed motives about how strong a reseller

21wants other brands to be.

22The end user's objectives are a little bit

23different. Again, the end user of course wants better

24pricing, may have concerns about delivery, quality,

25uniformity, efficiencies. It is less likely that an end

18

1user who is insisting on giving all of its business to

2one supplier is really in favor of weakening other

3suppliers. There may be those rare cases, but it is

4less likely that that is what you are going to find.

5So, what is the right analysis? When should

6courts second-guess buyers for instigating exclusive

7dealing and replace the buyer's judgment that it wants

8an exclusive with the court's judgment? I think that

9certainly when the buyer has a demonstrable motive to

10eliminate competition at the supplier level so that it

11is helping itself in terms of competition, that is one

12to take a hard look at, but generally, I think it is

13important to trust the buyer's judgment if it is

14instigating exclusive dealing.

15Let me just conclude by saying I hope this quick

16snapshot has highlighted some of the very many

17differences that exist among exclusive dealing

18arrangements. All of us as lawyers and economists are

19always searching for those unifying principles that make

20it easy to do the analysis, but I think what is

21important here is that we not get lazy and overlook that

22some of these variables that we have just been talking

23about really do make a difference to the analysis.

24I will leave it there, and thank you very much.

25(Applause.)

19

1MR. VITA: Thank you, Richard. Insightful and

2on time, perfect.

3Our next speaker is Mary Sullivan, who is an

4Assistant Professor of Accountancy at George Washington

5University. Mary received her Ph.D. from the University

6of Chicago, Department of Economics, and taught

7marketing at Chicago Graduate School of Business from

81987 through 1997. While at Chicago, she conducted

9research on industrial organization and marketing

10issues, such as slotting allowances, brand names and

11trademarks.

12In 1997, Professor Sullivan left academia for

13the U.S. Department of Justice Antitrust Division where

14she worked on a variety of antitrust matters and served

15as Assistant Chief of the Competition Policy Section.

16In 2004, she joined the Accountancy Department

17at George Washington University, and as many of you

18know, Mary's research has been published in numerous

19leading economics journals.

20Mary?

21DR. SULLIVAN: Thank you. I would like to start

22by thanking the DOJ and FTC for inviting me to

23participate in these hearings, and I need to keep track

24of the time very closely, because I have been threatened

25by Dan and Mike that if I go over my time limit, that

20

1they might charge me a slotting allowance, although in

2practice, I have learned that it is very difficult to

3charge one unless you charge it in advance.

4Nonetheless, I will try to stay on track.

5Slotting allowances and payola are two allegedly

6exclusionary practices that receive different regulatory

7treatment. What I am going to do in my talk is address

8whether the different regulatory treatment is warranted.

9Slotting allowances and payola are similar in

10many respects. They are basically the same practice

11used in different settings. Slotting allowances are

12payments made by manufacturers to retailers for stocking

13new products. Payola consists of payments made by

14recording companies to radio stations or DJs for playing

15a particular piece of music. Both practices have

16promotional effect. They serve to increase demand by

17providing exposure to the product or music to consumers.

18In each case, there is a scarce resource that

19needs to be allocated, shelf space in the case of

20slotting allowances and airspace in the case of payola.

21For both types of fees, there are concerns about

22exclusionary effects. If you read news articles or, you

23know, just search the web for these practices, or if you

24have talked to industry participants, you will learn

25that these practices are widely believed to be

21

1exclusionary, and the potential exclusionary effect is a

2major motivating factor in the regulatory scrutiny that

3each of these practices has received.

4Now, oddly, despite their similarities, the

5practices receive different regulatory treatment.

6Slotting allowances are not regulated by the FTC. In

7the FTC's 2001 report on slotting allowances, they said

8that the fees need to be judged on a case-by-case basis

9with attention both to likely competitive harms and to

10likely procompetitive effects. So, they take a basic

11rule of reason approach.

12Alternatively, the FCC does regulate payola.

13According to the FCC regulations, payments are

14prohibited unless an announcement of the endorsement is

15made every time a song is played, and this increases the

16cost of using payola. Now, in addition to the FCC

17regulations, the major recording companies have recently

18settled investigations brought by Elliott Spitzer, as

19many of you are probably aware. I think what is less

20well known about these settlements is that the terms of

21the settlements are more restrictive than the FCC

22regulations, with payola completely banned in most cases

23even if an announcement is made of the endorsement.

24Now, given over the past few years we have

25learned a lot about slotting allowances, both in terms

22

1of the economic theories and in legal challenges, I

2thought it would be an interesting exercise just to go

3through some of the things we have learned to try to get

4some insight as to why payola has received different

5regulatory treatment and whether this makes sense.

6Okay, so we will start with a little bit about

7the theories of exclusion. Can theories of exclusion

8explain slotting allowances and payola? Now, there are

9two general classes of theories that I will talk about.

10There are the popular theories or notions of exclusion,

11and then there are the economic, sort of rigorous

12economic theories of exclusion.

13The popular theory of exclusion, according to

14these theories, the payment of the fees increases the

15cost of introducing a new product or a new song. The

16increased entry cost may exclude manufacturers,

17particularly small ones, and many of the complaints are

18of this nature.

19However, this so-called theory cannot really

20explain exclusion. It is fairly well accepted that

21auctioning scarce resource results in efficient

22allocation, and unless something in the auctioning

23process reduces the number of slots that are available,

24it is very easy to see how this could result in

25exclusion. If a product or song is very promising,

23

1someone will give the product financing in order to

2introduce the product. Therefore, I really don't

3consider this a valid theory of exclusion.

4The other class of theories are the economic

5theories, and the two that I have really looked at for

6the purpose of this talk are Farrell 2001 and Shaffer

72005. Now, without going into much detail at all about

8these theories, all these theories share the feature

9that you need to have a contractual provision for the

10retailer to actually exclude a competitor in return for

11the fees. You must have a situation in which the

12retailer is reducing the number of slots available for

13exclusion to occur and for harm to result from it. So,

14one important conclusion that I take away from these

15theories is that simply paying a slotting allowance is

16not enough to cause exclusion.

17So, the next thing I want to do is take a look

18at the evidence, what do we know about slotting

19allowances and payola, and ask the question whether the

20evidence is consistent with the Farrell/Shaffer type

21theories of exclusion.

22In the case of slotting allowances, the answer

23is sometimes. Occasionally slotting allowances are

24accompanied by a contract to reduce the shelf space

25available to competing manufacturers which could weaken

24

1them and potentially exclude them. According to the

2FTC's 2003 study of slotting allowances, such contracts

3are fairly unusual, but they do occur.

4For payola, the answer is no. There is no

5evidence that exclusionary contracts are being used with

6payola. The evidence that I have seen suggested that

7recording studios are simply trying to use payola in

8return for getting the radio stations to play their

9songs, not that they would not benefit if they could

10exclude a popular song of a competing recording studio.

11I think, you know, if they could exclude a competing

12song, it would allow them to sell more records; however,

13there is simply no evidence at all that that is what is

14happening, and believe me, if you take a look at some of

15the Spitzer settlements, you will see that the evidence

16he collected was quite thorough. What I conclude from

17this is that according to the economic theories of

18exclusion, payola is very unlikely to be exclusionary.

19Now, I also wanted to take a look at some of the

20evidence from the courts to see what the courts say

21about slotting allowances and exclusionary effects.

22This is not really intended to be a comprehensive review

23of the legal cases on slotting allowances. What I did

24do is I looked at two legal challenges to slotting

25allowances that are both important, have been very

25

1influential, and I see cited quite often in other cases.

2In both of these cases, the courts found that the fees

3are a valid means of competing, and here are the two

4cases.

5One of the quotes from the Gruma case is

6particularly 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

9shelf space."

10Now, in this case, the plaintiffs were small

11companies, small tortilla manufacturers who were

12complaining that Gruma, the large manufacturer, was

13buying up all the shelf space and giving it unfavorable

14locations. The Court ruled, well, your tough luck. If

15you want to be in this game, you need to compete for

16shelf space.

17Now, in the Reynolds Tobacco/Philip Morris

18case -- which is often referred to as the retailer

19leaders case, which was the name of the Philip Morris

20program that was being challenged in court -- it was a

21somewhat different situation, because Reynolds, the

22plaintiff in this case, was actually a large company,

23but the conclusion of the Court was the same. In this

24case, the Court concluded that the Philip Morris program

25that involved the payment of slotting allowances

26

1increased industry competition.

2Okay, so if the theory predicts that payola is

3unlikely to be exclusionary and the courts have ruled

4that slotting allowances are an efficient means of

5allocating scarce shelf space, then why -- this leads us

6back to the original question -- why does payola receive

7different regulatory treatment than slotting allowances?

8The answer seems to be that since the air waves are

9owned by the public, there is a belief that radio

10stations should select music on the basis of public

11interest rather than the radio station's commercial

12interest. This view highlights the difference between

13slotting allowances and payola.

14The FTC and the courts see slotting allowances

15as a valid and efficient means of allocating shelf

16space, but the FCC believes payola results in an

17allocation of airspace that is not in the public

18interest apparently because it allows the radio station

19to play music that increases their profits. Now, does

20this make sense?

21Another way of asking that is, will regulating

22payola cause radio stations to select music that is in

23the public interest, whatever that is? The answer is

24no. To see why, it is helpful to understand a little

25bit about how radio stations are going to decide what to

27

1play both with and without payola.

2Now, if payola is banned, radio stations are

3going to earn all of their money from creative --

4selling -- or playing music that appeals to an audience

5that will buy advertisers' products. In other words,

6they are going to earn all of their profits from

7advertising dollars. So, what they are going to do is

8they are going to select music that appeals to people

9who buy the advertisers' products.

10Now, if payola is permitted, radio stations earn

11revenue from both advertising and payola, and this may

12cause the radio stations to change their selection of

13music. They may play more songs that appeal to people

14who buy records and play less songs that appeal to

15people who buy advertised products. It is not obvious

16to me that the selection of music will be more in the

17public interest if payola is banned. In either case,

18the radio stations choose what music to play on the

19basis of what maximizes its profits.

20So, I have several conclusions from this. The

21first conclusion from the analysis, from this exercise,

22is that it seems highly unlikely that payola will

23exclude promising music. This argument of exclusion

24should not be used to support the regulation of payola.

25Second, regulating payola will not help achieve

28

1the goal of serving the "public interest." With or

2without regulations, radio stations will design

3playlists to serve their own commercial interests. This

4is unavoidable.

5Third, prohibiting explicit payment for radio

6airspace will not make competition for airspace

7disappear. There is a scarce resource, and there is

8going to be competition for it. The competition will

9take a different form. To the extent that recording

10studios can find loopholes in the regulation, then there

11will be little effect on the regulation on what is

12played.

13So, my own personal conclusion from this is that

14the regulation of payola it seems to me does not serve

15the public interest, appears to be wasteful, and leads

16to needless enforcement costs.

17Thank you.

18(Applause.)

19 MR. VITA: Thank you, Mary.

20DR. SULLIVAN: No slotting allowance?

21MR. VITA: You are off the hook, for now.

22DR. SULLIVAN: Okay.

23MR. VITA: Okay, our next speaker is Joshua

24Wright, who is an Assistant Professor of Law at George

25Mason University School of Law, where he teaches in the

29

1areas of antitrust, contracts, and law and economics.

2Professor Wright's research focuses on the law and

3economics of the competitive process for product

4distribution, including slotting allowances, category

5management, exclusive dealing and other contractual

6arrangements. He has published in numerous journals.

7Professor Wright received his Ph.D. in economics

8from UCLA, Department of Economics, and he also received

9his JD from the UCLA School of Law, where he was a

10managing editor of the UCLA Law Review.

11Joshua?

12MR. WRIGHT: Thank you.

13Okay, so I am going to sort of hop on the back

14of some of Mary's comments on slotting and do a little

15less background talking about what they are, since that

16has already been covered. My comments here, just as a

17preface to get out of the way, are based on two papers

18that are up on the FTC web site, which has all of the

19slides and papers from the other panelists, both

20co-authored with Ben Klein, who I think will be here in

21the afternoon.

22So, a tiny bit more detail on -- I am going to

23use a slightly different definition of slotting

24arrangements than Mary used and define the contracts as

25per unit time payments made by manufacturers to

30

1retailers for shelf space. There is a couple of

2differences here. One is that sometimes, and indeed, in

3the FTC report that has been referenced, you will find a

4distinction between per unit tying payments and

5discounts for slotting contracts, and it is an important

6difference and one that I am going to end up not talking

7much about here, but there is a discussion in the paper

8I just referenced on the economics of slotting

9contracts, on when we might expect the efficient form of

10a distribution contract to be a per unit tying payment

11or a discount. That said, I am going to ignore the

12issue for the next 19 minutes.

13What else we know about slotting is that they

14cover both new products and established products. So,

15they cover -- you know, Coca-Cola pays slotting

16allowances, products where we do not have any sort of

17risk imposed on the retailer by giving shelf space to

18some unproven product. We see slotting allowances on

19those products as well.

20What else we know is that they increased, there

21was a spike in the prevalence and the magnitude of

22payments somewhere between 1981 and 1984, and over the

23last 20 years, that trend of increasing and over the

24products covered and the magnitude of payments has

25continued.

31

1So, the anticompetitive theories of slotting,

2first, before I try to explain a procompetitive

3rationale for shelf space contracts. We see slotting

4contracts used by manufacturers with small market

5shares. We see -- in general, the FTC report finds that

6the normative time for these agreements are between six

7months and a year. We see them on products where there

8are not significant economies of scale in manufacture,

9one of the conditions that drives the anticompetitive

10theories in the literature. And also, the

11anticompetitive theories have a difficult time

12explaining the jump in the use of the contracts in the

13middle of the 1980s.

14In terms of the procompetitive story for

15slotting allowances, there are really two important

16economic questions with respect to slotting fees, and

17the first is why you see a separate contract at all,

18right? The first economic intuition one might have is

19why don't we see, like the setting of retail prices in a

20competitive retail market, supermarkets, et cetera, why

21don't we see manufacturers just set the wholesale price

22and allow the retailer to set the level of shelf space

23that is supplied for different products like we let them

24set the price? So, why do we see this separate contract

25for the shelf space?

32

1And the second is, and more related to the panel

2discussion today, is we see sometimes that these

3contracts include exclusivity provisions, unlike the

4payola contracts. We see provisions that say, give me

570 percent of the shelf space, give me a space to sales,

6give me the full exclusive, do not put anyone else on

7the shelf space. So, we see this additional variation

8in the contracts that we are going to need to explain.

9So, I will turn to that second. There are other

10interesting questions, again, the form of the payment

11and these things, which for the moment I am going to

12skip so I can focus on exclusivity.

13So, the answer provided by Ben Klein and myself

14in the paper I alluded to earlier, the intuitive answer

15is what you see on the screen, and it is that slotting

16contracts solve this pervasive incentive incompatibility

17problem where the retailer does not want to supply the

18joint profit maximizing level of promotional shelf space

19under the conditions where the supply and the shelf

20space does not induce consumer switching. So, we have

21cases like McCormick and we have 90 percent of the shelf

22space allocated for spices. Well, supplying additional

23promotional shelf space to spices does not induce a

24greater number of consumers to say I will not shop at

25this retail outlet because they have given 90 percent of

33

1the shelf space to spices, and they have two brands, and

2so I am going to leave. So, we expect to see this

3incentive incompatibility problem solved with a separate

4contract under these conditions.

5Now, I am going to go through a little bit of

6the analysis with a simple model with a little bit of

7math, but here is the intuitive answer. So, the

8fundamental point here is that for many products, and

9differentiated products, we have manufacturers with a

10large profit margin. So, the manufacturers, the

11wholesale price over the marginal cost, this P sub W

12minus the marginal cost of manufacture, is large

13relative to the retailer's incremental profit, whether

14it sells Coke, Pepsi or any brand of soda, okay?

15For a number of products, this is generally the

16case. So, the retailer, when it is making its decision

17on the optimal level of shelf space, promotional shelf

18space to supply to the manufacturer's products, say

19Coca-Cola, does not take into account that these

20promotional sales induced by giving, say, the eye-level

21shelf space, or if you are in the children's cereal

22aisle, the children's eye level shelf space, these

23incremental profits are large for the manufacturer and

24not taken into account by the retailer.

25Now, we can make the same argument with respect

34

1to price competition, but there is a key difference as

2to why we see manufacturers in the retail setting, at

3least, allowing the manufacturers to set the retail

4price, and competition between retailers is sufficient

5to get an optimal jointly profit-maximizing price set

6but not the jointly profit-maximizing level of shelf

7space. So, why do we get prices right and shelf space

8wrong ends up being the question.

9So, unlike the shelf space case, when we are

10talking about price competition, you see here we have

11got on the right-hand side is this large manufacturer's

12margin, that P sub W minus the marginal cost of the

13manufacturers. It is large. It is maybe 10-20 times

14larger than the retailer's margin for a good chunk of

15products. But we have this offsetting effect induced by

16customer switching. So, the intuition here is that

17while the manufacturer's margin is much larger, we have

18got this switching effect, so the quantity response

19faced by the retailer when it changes the price has

20these two different components.

21One, when it reduces the price or increases the

22price of Coca-Cola, there are interbrand effects, so

23sales move from Coke to Pepsi, but there also are

24inter-retailer competitive effects, right? So,

25consumers may end up switching stores when we are

35

1talking about price decisions or at least are more

2likely to do so than when we talk about moving Coke from

3the bottom level to the eye-level shelf space, right?

4So, the key point and argument here is that

5because promotional shelf space does not involve large

6inter-retailer shelf space effects, we do not see

7consumers switching on a number of grocery products. My

8co-author on the paper and dissertation adviser likes to

9use the example of dog collars in the store, right? So,

10there is some exclusive space granted for dog collars,

11and people pay and they compete for this space, but

12nobody switches the stores because there is one dog

13collar versus two, okay?

14And because we have this idea that there are

15these small inter-retailer effects, it is the case that

16we have this incentive incompatibility problem, right,

17and instead of this inequality, if we had the jointly

18profit-maximizing level, we would see at least this

19relationship be approximately equal. The big difference

20is this elasticity from the retailer's perspective of

21the shelf space effect, right?

22And so this is all to illustrate the point that

23where we see these small inter-retailer effects, again,

24this incentive incompatibility problem is pervasive, and

25this is especially so in the supermarket context. Now,

36

1there are some limits on this idea. We do not see --

2the distinction here is not just because of price and

3nonprice competition, okay? There are elements of

4nonprice competition where there are inter-retailer

5effects because all consumers value the service.

6So, the supermarket provides a free parking lot.

7You can go and you park and you do not pay for it, you

8know, when you go in to park. Everyone generally values

9that there is a parking lot, maybe there is lighting

10there so you don't get mugged when you go to the parking

11lot, and everybody values this, and this means, because

12consumers value some nonprice services, then they will

13induce some switching, that for those services, the

14incentive incompatibility problem is solved. The

15retailer will supply those because consumers are all

16willing to pay.

17So, where we see this, the very idea of

18promotional shelf space is to give some sort of

19effective, targeted discount to the marginal consumers

20who are sensitive to allocations in the shelf space,

21right? They are sensitive to what is in the eye-level

22shelf space, and there is a substantial marketing

23literature which demonstrates sometimes some really

24surprising results about how large the effects can be in

25terms of changes in sales when we play around with the

37

1shelf space allocation.

2So, in these fairly general circumstances, the

3disparity in margins and the small inter-retailer

4switching effects from the supply of promotional shelf

5space, the manufacturer wants more shelf space than the

6retailer is willing to supply, and so we need to have

7some separate contract where the manufacturer pays the

8retailer for the supply of the shelf space in order to

9solve this incentive incompatibility problem.

10So, now we have got a situation where Coke is

11paying for the eye-level shelf space to the retailer,

12and it pays them $10,000 per unit time for the month for

13some contracted-for level of shelf space. Now, this

14does not mean that the whole process is over, right?

15So, the manufacturer pays the retailer with this money,

16and the retailer has some incentive to not perform.

17It can provide less than the contracted-for

18level of space. It can otherwise violate the implicit

19contractual understanding between the manufacturer and

20the retailer to sell the space twice, in other words,

21the simple way to think about it. So, it is taking the

22money and not performing under the terms of the deal.

23This is where we get to the function of full or limited

24exclusives in shelf space contracts.

25Now, we see that in the slotting context, at

38

1least a full or a partial exclusive seems to be -- at

2least appears to be thus far -- a necessary condition

3for liability. So, we have some form of exclusive -- we

4have -- well, there is no liability, but Gruma, Conwood,

5McCormick, so we have these cases where the contracts do

6not just buy the shelf space. They specify a

7percentage. They specify a full exclusive. They

8specify limits on the placement of rival products.

9So, there are a number of procompetitive

10rationales for exclusivity terms in these contracts, and

11Mr. Steuer went over many of them, and so I am not going

12to belabor them here, but the key, following from this

13sort of shelf space contracting model, is that an

14exclusive can help facilitate performance of the

15contract, right? The retailer pockets this money and

16can have some short-term incentives to not perform.

17So, a couple of things that exclusivity can do,

18it can efficiently define exactly what the manufacturer

19is purchasing. Purchasing all of the shelf space,

20detecting cheating becomes easy. The other thing it

21does is it allows the retailer to say, you are bidding

22for all or 70 percent or some large fraction of the

23promotional shelf space, and this intensifies the

24bidding process between the manufacturers for the shelf

25space, and this is a good thing in terms of the

39

1antitrust analysis, a good thing for consumers, because

2these shelf space payments are passed on to consumers,

3and that is whether they are discounts or per unit time

4payments.

5Quickly, so I can end here, category management

6contracts are just a form of limited exclusive, where

7what we are doing instead of saying you get 50 percent

8of the space is the retailer delegates the function to

9the manufacturer to allocate the shelf space, and we see

10this in circumstances where consumers' demand for a

11particular brand is high. So, the implicit contract is,

12you get to feature your product, Coca-Cola, and you can

13allocate the shelf space, but if consumers come to me

14and say I have a high demand for Pepsi and you're

15putting it on the bottom or you have run out or you did

16not put it on the shelf, then I know and I terminate the

17agreement, okay?

18Just to finish up, Conwood seems to get this all

19wrong. So, Conwood, despite the sort of atmospheric

20facts and the tortious behavior and lots of bad stuff

21going on, there is some bothersome language in the

22opinion about imposing a standard on category managers

23that is tougher than the standard on monopolists using

24full exclusives, and so the key idea is that exclusive

25dealing can make economic sense in these circumstances

40

1and that we need to make sure that the plaintiffs are

2demonstrating an anticompetitive effect before we engage

3in any sort of balancing under the rule of reason

4analysis.

5I think I went over, sorry.

6MR. VITA: Not too bad.

7(Applause.)

8MR. VITA: Thanks, Josh.

9Okay, our next speaker is Howard Marvel who is a

10Professor of Economics in the Department of Economics at

11Ohio State, and he is also Professor of Law in the

12Moritz College of Law at Ohio State. Howard's work on

13vertical restraints is very well known. He has written

14on a variety of different topics, including resale price

15maintenance and exclusive dealing, and I know those

16papers have appeared in some leading economics journals.

17Howard also has advised the Japanese

18International Trade Ministry, had a post in

19telecommunications, the Federal Trade Commission and the

20National Association of Attorneys General law on

21vertical restraints issues. In addition, he has served

22as an expert in vertical restraint matters for a number

23of firms.

24Howard?

25DR. MARVEL: Okay, I have seen a lot of you

41

1before. I am happy that you have invited me to come

2talk to you outside of the Third Circuit, and the topic

3for today is exclusive dealing.

4It is obvious that exclusive dealing is a very

5common thing that we see every time, when you go to a

6MacDonald's, you do not find a Burger King hamburger,

7and Haagen Dazs has had the exclusive dealing in their

8distribution contracts, car dealers typically have it,

9there is exclusive dealing in beer distribution. It is

10all over the place, and ordinarily we do not think

11anything about it. You know, any business format

12franchise is basically franchise or else, and it is most

13commonly observed for our market leaders, the big guys.

14Anheuser-Busch has it in the Chicago area, it is

15under study, and you don't see that elsewhere. Haagen

16Dazs had contracts with distributors with Steve's, which

17at the time was a premium ice cream. I do not know if

18it is still around. Anybody from Boston? Steve's did

19not have that. The big guys have more reason to

20foreclose, of course, but they have also more to free

21ride upon.

22So, for a long time we had a rule that Richard

23talked about, how tough it was to engage in exclusive

24dealing. The rule seemed to be that if you had market

25dominance or a big share somehow, somehow, and you

42

1practiced exclusion, if you had exclusion in your title

2of whatever the practice was, you were toast. So, it

3was essentially a per se violation.

4Now, exclusion there does not mean foreclosure.

5It just means exclusion from a portion of the market,

6and that is very different than keeping the firm totally

7out of the market. Foreclosure is a different story.

8Now, several of the -- I think John is going to

9talk about the Chicago view and why it is limited, so

10let's run through what the Chicago view of vertical

11restraints is. It is that vertical restraints create

12property rights. So, you have a problem that you want

13to get somebody to do something, but you are afraid that

14at the end of the day they will not do it because the

15fruits of their actions will end up being frittered away

16as other people take advantage of them, okay?

17So, the idea behind vertical restraint is that

18it creates a property right for somebody or other, so

19exclusive territories, for example, create a property

20right for customers that a particular distributor or

21dealer generates, okay? So, I go out to get a customer,

22how do I guarantee if I am the seller who wants that

23customer generated, how do I guarantee the customer gets

24generated? I protect the rights to that customer for

25the guy who actually did the work?

43

1Resale price maintenance is very similar. There

2is a property right for the services that the

3distributor provides, and Josh talked about how this

4sort of works in slotting as well, like exclusive

5dealing, that creates a property right for customers

6that the supplier's actions pull in, and I think that if

7you think about the -- almost all of the things that

8Richard included in his discussion from the 1983 paper,

9they all have that characteristic, that the supplier is

10doing something to pull in customers and those customers

11are being protected through exclusive dealing by -- from

12some sort of bait and switch approach.

13 Now, the problem with exclusive dealing and what

14makes it more serious and more of a worry than

15territories and RPM is that in territories and RPM, the

16supplier is creating a property right for somebody else.

17It says, you do this, and you get to keep the fruits, so

18I would police that. And I am an outsider, and I want

19to have the distribution system to be as effective as I

20possibly can make it be, but with exclusive dealing, the

21property right is for the creator and the monitor of the

22right.

23I give myself the right, and then I protect that

24right, and we have a problem that can emerge there if

25the right is somehow something that you really don't

44

1want the guy to have and be able to protect, and that is

2really what is at the heart of Aspen Ski, because in

3Aspen Skiing, Aspen Skiing and Aspen Highlands

4cooperated to develop the Aspen market as a destination

5for skiers, and then at the end of the day, Aspen Skiing

6said, well, gee, they passed a law here in Aspen where

7you have got to have a three-week rental instead of just

8a one-week minimum rental or a longer rental term, and

9so you essentially locked customers in. You didn't have

10to compete for customers so much, because they said,

11well, we will walk away with rents, and you can see that

12elsewhere.

13If you have a patent holder who has accessories

14for his product, the patent is about to expire, the guy

15may decide to engage in exclusive dealing to try and

16freeze out the accessory guys that he's cooperated with

17to build that product, and believe it or not, I was an

18expert witness in a matter in which I thought exclusive

19dealing was used improperly in this way, so it's not

20clear that these are anticompetitive so much as fraud or

21contracting problems, but they are problems.

22Okay, so the basic exclusive dealing story is

23simply that the manufacturer invests in a product or a

24reputation that brings in customers, if the manufacturer

25confers upon its customers -- its customers onto dealers

45

1who are cloaked in its reputation. So, if I become a

2dealer for a particular manufacturer, then customers

3say, hey, that dealer is essentially certified as

4knowing what he's talking about, so the customer walks

5into the dealer, induced to do so by the manufacturer's

6efforts, and then the dealer says, by the way, I have

7got a better deal for you.

8Now, a requirement for this to work is that the

9customer cost, the cost of generating the customers has

10to be included in the charge for the product. So, if

11you can charge for leads separately, no sweat, okay?

12You just charge for the leads, you do the promotion, the

13customers walk in, and if the dealer who's paid for

14those customers wants to switch them to some other

15product, hey, that's fine, okay, but there are a lot of

16circumstances in which you only charge for the customer

17when they actually buy something, so it is rolled into

18the product price, and this is, again, the way it works

19with royalties in business format franchises, right,

20because MacDonald's brings customers in, but they only

21receive a charge, a payment, for those customers when

22the royalty is generated, okay?

23So, the dealer can avoid this particular charge

24through a bait and switch scheme in which he says, okay,

25you are a customer for firm X, firm X brought you in,

46

1that is what you came looking for, but firm Y has got a

2product that is cheaper, because it does not involve any

3promotion, it is simply a free rider, so why don't you

4switch to that one, and you can trust me, because I am

5firm X's dealer, okay?

6So, what is the evidence for this -- how this

7works, okay? Is there any evidence to suggest that this

8works? Well, you know, "can you hear me now" doesn't

9necessarily need to be Verizon's slogan, it also should

10be a slogan for the hearing aids manufacturers who were

11engaged in exclusive dealing, and they were going out

12and getting a lot of customers to come in, into their

13dealers, and the customer comes in saying I saw an ad

14for Beltone hearing aids or whatever, can you fit me

15with a hearing aid? And the dealer at that point can

16say, yeah, I am a Beltone expert, and by the way, I've

17got a better deal on another hearing aid.

18Now, the interesting evidence on this is that

19the FTC decided to take four of the five hearing aid

20manufacturers who used exclusive dealing, take them out

21and shoot them, because the idea was if you agree not to

22use exclusive dealing, we'll let you off the hook, and

23at the end of about a year or so, the bodies of the

24companies had agreed not to engage in exclusive dealing

25washed up on the shore. They were out of the business.

47

1So, that's a problem in these cases, the

2counterfactual, what would happen if the practice were

3forced to be given up, is very hard to prove until it is

4too late. When you see the corpses, then you know you

5screwed it up.

6The manufacturers in the hearing aids case did

7not recognize the role of exclusive dealing themselves,

8and so they walked away from it. Beltone didn't, but

9the other manufacturers of hearing aids did, and they

10ended up dead in short order, okay?

11Now, after the Chicago explanation came out,

12then we got a game theory counter-revolution, okay? A

13famous paper by Aghion and Bolton sort of launched the

14"why don't we get together, write a contract and screw

15the next guy to come along" approach to contracting,

16which is, I think, a fair way to say what their model

17is. It says, I am in the market now, I am the only guy

18in the market, you're my dealer, there might be somebody

19who comes along later and is better than me. Why don't

20we figure out a way to split the rents from that guy's

21advantage, okay? And the way we will do that is we will

22write a contract between ourselves that has a penalty

23clause, okay, and the penalty clause is such that --

24five minutes, it says. Okay, I'll never get there,

25okay? I am a professor, you know, I am not one of these

48

1lawyer guys. I just talk and talk. That's the way it

2works, but I'll be done.

3Okay, so the Aghion-Bolton idea is that there is

4a contract that is written before the entrant shows up,

5and then we run off with the entrant's rents because of

6the existence of this contracting penalty clause, okay?

7The requirement for that to work is you have got to have

8a contract, right? That is what you have got to have

9before this works, because if the entrant does show up,

10then the dealers run to the entrant if he is better,

11okay?

12There is a second set of theories that are

13contract-based, and you think of the names Segal and

14Whinston, Ramweyer, Rasmussen and Wiley, and these are

15train leaving the station contracts. The train is

16leaving the station, I am the only guy in the market,

17you better sign up with me or else, and then you have

18got to stay with me if I am no longer the only guy in

19the market, okay? So, these both require contracts.

20All of these theories require contracts. No contract,

21no problem, okay? And that is the characteristic of the

22game theory counter-revolution.

23So, is Chicago out the window? Oh, they are,

24because Professor -- or Mr. Jacobson -- what is the

25appropriate -- Mr. -- Mr. Jacobson --

49

1MR. JACOBSON: Hey you, hey you is fine.

2DR. MARVEL: Hey you? Okay, he says, but

3Chicago writers -- post-Chicago writers long ago

4debunked the Chicago School, and it is now common ground

5that in many contexts exclusive dealing can be deployed

6in a way that is both profitable for the dealer and that

7allows the defendant to reap gains from the arrangement

8that far exceed the associated costs. Guess what? I

9agree, okay? True. Absolutely.

10Now, we will wait for the first one of these to

11come along, but it is possible, in principle, for this

12to happen. I do not have the slightest disagreement

13with that.

14Now, a couple of examples of this sort of thing,

15the first from your vintage Chicago School nut case, we

16appreciate the potential reply that it is impossible to

17say that a given practice "never" could injure

18customers. A creative economist -- there are creative

19economists -- could imagine unusual combinations that

20would cause injury in the rare situation, but antitrust

21law applies rules of per se legality to practices that

22almost never injure customers, and who might that be?

23Yes, Chicago.

24Okay, but then we also have this statement the

25literature on anticompetitive exclusive dealing, so

50

1actually what we are talking about today, has focused on

2producing "possibility results" in simple settings to

3counter Chicago School arguments. It is possible that

4something can go wrong, says Mike, okay? Now, he is not

5a Chicago guy, okay, and he is right. He has written

6some of the possibilities, but the possibilities take

7contracts, okay?

8Problems are possible, and the problems involve

9foreclosure. If you get foreclosure, that does not mean

10foreclosing a particular set of dealers. It means

11foreclosing the market. If you get that, that is a

12problem. The benefits are going to be really hard to

13prove from exclusive dealing up front. Again, like I

14said, until you see the bodies wash up on the beach.

15The default rule in these cases is going to

16determine the outcome, okay? If the default is that

17exclusion could be bad, what will happen is that

18exclusion will be found to be bad despite the absence of

19factors suggesting the presence that we might have one

20of the bad theories of exclusion, the proof of concept

21or possibility theories, present. So, if we get the

22default rule wrong, what will happen is that we always

23find that possibility means exclusion, becomes the

24default rule, and we are back to where we started.

25Exclusion plus dominance will equal violation. That is

51

1where we were before. One minute.

2Beltone, forget them, okay?

3So, what should we do about all this in the last

4minute? The first possibility is that all of the

5possibility results that I know of, and even this guy

6Joe Farrell back there who just walked in seems to know

7of, are contract-related, okay? So, why don't we start

8by requiring a contract? No contract, no problem, okay?

9Then, we ought to require some notion that there

10might be something wrong in this market in the sense

11that there be a showing of foreclosure, and success

12should not be defined as foreclosure. If I do better

13than you do, I get a big share of the market, so what?

14And if my dealers then get that share, so what? Success

15should never be considered the equivalent of

16foreclosure.

17But if you get to that point where you have

18found that there is a contract and there is a showing

19that foreclosure is a real problem in this industry in

20the sense that there is not another way to get to

21market, then, and only then, after you have gone past

22those two standards, should you go ahead and run your

23trade-off analysis, and I am reasonably convinced that

24that trade-off will often, if not always, that you will

25find it very difficult to prove that the efficiency

52

1benefits that you are claiming are really present.

2With that, we will be done, okay?

3(Applause.)

4MR. VITA: Our final speaker before we take a

5short break is Jonathan Jacobson, who is a partner at

6Wilson Sonsini Goodrich & Rosati, where he practices

7antitrust law and has taken a lead role in many

8significant antitrust matters over his 30-year career.

9Among other cases, Jonathan was lead counsel for

10Coca-Cola in Pepsico v. Coca-Cola, a leading Section 2

11monopolization case.

12Jonathan was appointed by Congress in 2002 to

13serve on the Antitrust Modernization Commission, which

14is dedicated to studying the nation's antitrust laws and

15considering several changes. He also is the editorial

16chair of the ABA's Antitrust Law Developments and has

17chaired a number of ABA antitrust section committees.

18He has written and edited numerous articles and books on

19antitrust, and his most recent paper co-authored with

20Scott Scherr is entitled, "'No Economic Sense' Makes No

21Sense For Exclusive Dealing."

22 John?

23MR. JACOBSON: Thank you.

24I also want to express particular thanks for

25seating me on the far left wing on this panel. I think

53

1that is entirely appropriate, although I would comment

2that in exclusive dealing cases, I have never

3represented a plaintiff. I would like to, but it has

4always been defense representation so far.

5So, let's talk about exclusionary conduct and

6exclusive dealing in particular. There are lots of

7different exclusionary conduct devices, and these

8hearings will cover most of them. I actually think

9ripping your competitor's racks off the shelves is

10pretty exclusionary, so maybe we can talk about that in

11the dialogue, but that is one example of exclusionary

12conduct. The other is price cutting, which is, you

13know, rarely, rarely, rarely harmful and yields, you

14know, major significant consumer benefits.

15Exclusive dealing is in the middle, and it

16presents a real challenge, because what makes exclusive

17dealing potentially harmful is the very same mechanism

18that makes the arrangement efficient and may lead to

19lower prices for consumers.

20So, what are the consumer benefits? I think

21Richard went through them and I will just go through

22briefly, but basically the distributor, if we are

23focusing on distribution, which is the typical case, the

24distributor focuses his or her attention on the

25supplier's product and becomes a more effective

54

1distributor, and from the supplier's perspective, the

2supplier has an incentive to provide the distributor

3with information and displays and all sorts of that

4stuff without concern of free riding by competing

5suppliers.

6So, these benefits are very important, but they

7are possible only because the arrangement is exclusive,

8denying rivals access to the distributor's capabilities.

9This same exclusivity can have the effect -- and it is

10not an ephemeral possibility, it can happen, although it

11is not necessarily the default rule, but it is a real

12world phenomenon -- that the exclusive can deny the

13rivals access to customers or supplies and have the

14effect of driving their costs up and rendering them less

15effective competitors, less effective constraints on the

16defendant's market power. And the result of that can

17be -- and this is the case we need to be alert to -- to

18allow the supplier to increase prices to consumers as a

19result of the weakening of that competitive constraint.

20So, the question is, how do we evaluate

21exclusive dealing and quasi-exclusive dealing

22arrangements in light of these simultaneous benefits and

23harms?

24Now, today I think, you know, I have not been

25here for these hearings, I have read a lot of the

55

1summaries and some of the testimony, but I suspect that

2there is agreement on really four issues in terms of an

3overall approach to exclusionary conduct. One, we do

4want to prohibit behavior that leads to the creation or

5expansion of significant market power. We want to be

6careful, and I think that is a principal focus of these

7hearings, to avoid deterring procompetitive conduct. We

8want to have rules that businesses can understand and

9apply so that they know what they are doing is legal or

10illegal. And we want to provide the courts with

11sufficiently clear rules so that they can tell in the

12context of a lawsuit what is illegal and what is not.

13So, for exclusive dealing, we have applied these

14goals. I think you can go back to Tampa Electric and

15say we have had a rule of reason since then, but I will

16respect Richard's qualification of that and take the

17rule of reason back to Beltone, which is clearly the

18first sort of modern formulation of the rule of reason

19in exclusive dealing cases. And where we are coming to

20now, I have another paper where I comment that the focus

21on foreclosure is unfortunate, and my basic point of

22view on this, and I think where the law is going to come

23to if it has not come to already, is that in an

24exclusive dealing case, what the plaintiff must show to

25prevail is that the net effect of the conduct, including

56

1the efficiencies, is to raise prices or otherwise harm

2consumers. And I think, you know, if you look at the

3major exclusive dealing cases over the last ten years,

4the results largely -- not entirely -- but are largely

5consistent with that kind of paradigm.

6So, the recent debate was spurred in part, I

7think, by the thinking of folks like Judge Easterbrook,

8who gave a talk a few years ago saying that we should

9abandon Section 2 enforcement entirely, but that has led

10a lot of conservative thinkers and some more mainstream

11and liberal thinkers, like Steve Salop, to try to

12determine whether there is a universal test for

13examining exclusive conduct, and at some level we have

14been searching for the universal rule ever since Learned

15Hand's decision in the Alcoa case.

16I would commend to all of your attention an

17excellent article in the Antitrust Law Journal a few

18months ago by Marc Popofsky, that having a

19one-size-fits-all approach that can be applied equally

20to practices as diverse as predatory pricing, refusals

21to deal, ripping your competitors' products off the

22shelves, has proven to be elusive. And I do not think

23we have gotten there yet, and I question whether we ever

24will.

25The main area of disagreement is the extent that

57

1we need extraordinary screens to ensure that

2procompetitive conduct is not deterred. The sort of

3screens that I would add that we do not see in most

4areas of the law other than antitrust. Antitrust, at

5least in the last few years, has been very sensitive to

6avoid deterring procompetitive conduct at the cost, many

7recognize, of allowing the occasional illegal behavior

8to go through.

9All right, so -- by the way, thank you for not

10allowing questions from the audience, because Greg

11Werden is here -- and it is with quite a bit of

12trepidation, although he and I have had a few

13discussions on this subject, that I challenge the no

14economic sense test or Doug Melamed's version, the

15profit sacrifice test. This issue has gained -- and

16appropriately so -- a lot of attention, and under at

17least one articulation of the no economic sense test, a

18practice is not exclusionary for purposes of Section 2

19unless it would make no economic sense for the defendant

20but for the tendency to eliminate or lessen competition.

21And in varying degrees, some of the advocates of this

22test urge that it be applied to all single-firm and

23vertical conduct.

24If you look at the certiorari brief filed by the

25Justice Department in the Trinko case and the briefs

58

1filed in the Court of Appeals in the Dentsply and

2American Airlines cases, the Justice Department has

3argued variations on this test as a rule of law. It has

4not been adopted by any of those courts, but it has been

5argued with some vigor by the Department of Justice.

6One of the issues I have with the no economic

7sense test is that it is fundamentally the Areeda Turner

8predatory pricing pricing test in new garb. Areeda

9Turner made a major advance in the law in 1975 when they

10urged that predatory pricing not be condemned unless it

11is below cost with a likelihood of recouping the lost

12profits through the market conditions that will result

13from the predatory pricing scheme. And their test was

14acknowledged and stated by them to be an extraordinary

15test reserved exclusively at that time for price

16cutting, because price cutting is so rarely harmful and

17so extraordinarily important to our economy that we want

18to have a test that really makes sure that errors are

19purely on the side of allowing the defendant to win

20rather than the plaintiff to prevail.

21Now, there have been efforts starting with the

22article that Janusz Ordover and Bobby Willig put out a

23few years after that to apply this sort of analysis more

24regularly to other forms of exclusionary conduct, but in

25general, we have been asking ourselves the question

59

1since the no economic sense literature came out, is this

2purposefully extraordinary test -- and it was designed

3as an extraordinary test -- is it appropriate to apply

4it to other types of exclusionary conduct?

5In my view, as applied to exclusive dealing, the

6no economic sense test really does make no economic

7sense, and I say that because exclusive dealing

8arrangements make economic sense precisely because they

9lessen competition by rivals for the affected business.

10So asking that question tells us nothing about whether

11the arrangement is procompetitive or anticompetitive.

12Exclusives are usually associated, even in

13extreme cases like Dentsply, I think you can say that

14exclusives are usually associated with real efficiencies

15and sometimes cost very little to implement. So, unless

16you apply the economic sense test with the rigor that a

17Greg Werden would, and if you apply it in the real

18world, it is very easy to come out with the

19determination that the exclusive makes economic sense

20for the defendant.

21But the way in which those efficiencies are

22achieved, as I said before, is through this mechanism of

23exclusion. So, the judicial audience, the business

24audience out there, is wondering, how can I do this?

25This arrangement makes no economic sense to me unless I

60

1can exclude my rivals, but that seems to be the test for

2illegality, so what do I do? And I think the answer to

3that is you apply a different test.

4So, exclusive dealing is also interesting and

5different, as Steve Salop points out, because at least

6under some scenarios there need be no period in which

7profits are sacrificed during the course of the

8exclusive dealing arrangement. You can have

9simultaneous exclusion and recoupment.

10All right, recent case, not a federal case,

11although I will tell you we did our best to get the

12Justice Department and Federal Trade Commission to file

13a brief and they politely declined, but the Court came

14out correctly I think anyway, although it was a 5-4

15decision, and if you really want to read something

16interesting, read the dissent in the case. It is a

17decision that came out less than a month ago out of the

18Texas Supreme Court, and it involved exclusive

19promotional agreements with retailers, not exclusive

20dealing arrangements, but exclusive promotional

21agreements.

22In some of the agreements, Coke -- in all of the

23agreements, Coke had to get a reduced price. In some of

24the agreements, it provided that the low price had to be

25the lowest in the store on that particular package. The

61

1exclusives required the most prominent displays in the

2stores and also exclusive ads.

3In return for this, Coke provided very

4significant lump sum promotional payments and deeply

5discounted wholesale prices. So, the result was to

6reduce the retailer's costs, both marginal costs and

7total costs. Coke had 70 to 80 percent of the market if

8you accepted the market definition in the case. The

9result of this was lower prices for Coca-Cola products,

10and it was not seriously disputed that the level of

11promotional activity resulted in overall lower prices in

12the marketplace for carbonated soft drinks as a whole.

13Now, the exclusivity in that case, the

14agreements, made economic sense only because the

15exclusives made more -- made things more difficult for

16rivals, and the easy example is to ask why would Coke

17pay thousands of dollars to a supermarket for a

18promotion? Let's say the promotion is two-liter and

19you expect that the reduced price would be something

20like 99 cents. If the consumer is going to walk in the

21store and the first thing she is going to see is a Pepsi

22display of two liters at 89 cents, that promotion really

23is not worth very much for Coke. Why would Coke spend

24the money for that promotion? Why wouldn't it just

25figure out some other way to sell soft drinks?

62

1The problem, as the dissent points out, is that

2this kind of exclusivity could fail an incautious

3application of the no economic sense test, but

4appropriately, the majority upheld the agreements under

5the rule of reason because there was no showing that

6they led to increased prices in the market as a whole.

7 Now, I will very briefly talk about Microsoft,

8and I am not going to go through the whole slide, but

9the basic concept here is a lot of what Microsoft was

10doing was virtually costless. Leaving Internet Explorer

11out of add/remove programs was virtually costless, and

12if you apply the no economic sense test to Microsoft,

13you can easily get a situation where the Court would say

14that this conduct makes economic sense and is,

15therefore, upheld. I think the Court went through an

16elaborate recitation of the rule of reason, and I think

17we have a good precedent there.

18I had promised not to go over time, and I see

19that I already have. What I do want to point out is

20that the focus that we care about in antitrust generally

21and in exclusive dealing cases as one piece of that

22overall puzzle is does this behavior injure consumers?

23Does it raise prices? Does it otherwise injure

24consumers and the benefit of the bargain that they are

25going to receive?

63

1The no economic sense test asks that we bypass

2that question. My point is simply, let's look at that

3question directly. Let's try to get to that analysis

4directly. The shortcut, which if applied incorrectly

5can lead to very questionable results, is not a

6necessary route. It does not protect competitive

7conduct any more than a careful application of the rule

8of reason would. So, let's just ask the question that

9we really want the answer to and guide our analysis on

10that basis.

11Thank you.

12(Applause.)

13MR. VITA: Thank you, John.

14I think we will take a short break right now.

15Why don't we come back at -- ten past? -- yeah, ten

16minutes past, and we will reconvene.

17(A brief recess was taken.)

18MR. VITA: All right, let's get started.

19I think the first thing we will do here is take

20a few minutes and just open it up to the panel to allow

21them to react to some of the things that they might have

22heard and pose questions to the other panelists. So,

23Jonathan, you came by before and said you had an issue

24you wanted to raise. I'll let you have the honor of

25going first.

64

1MR. JACOBSON: Well, thank you. I previewed

2this with Howard, because I think the no contract --

3MR. VITA: Jonathan, speak into the mike.

4MR. JACOBSON: I think the "no contract, no

5problem" scheme is a problem, so to speak, and what I

6would ask Howard is, isn't it a fair observation that

7you worry more about exclusive dealing the larger the

8market share of the defendant, and don't you run into

9cases where the defendant's share is so high -- it is

10not really the share, but the market power of the

11defendant -- where the defendant's market power is such

12that they can enforce exclusives on the offer you can't

13refuse or the all-or-nothing offer that Richard was

14referring to with a lot of the detriments that can be

15associated with exclusive dealing with little or none of

16the benefits?

17And again, you know, Microsoft is not a bad

18example. Those were not contracts at least of any

19duration in that case. Microsoft basically told Dell

20and Compaq and Hewlett Packard, you know, here it is,

21deal with it, and, you know, it was not a really good

22option for them to go to UNIX, and Apple was not

23available. So, let me put that one back to you.

24DR. MARVEL: Well, I guess what I would say is

25that looking at the economic analysis of exclusive

65

1dealing and at the places where the game theoretic

2models have found problems, they are all cases in which

3there is not an option today and I sign up everybody

4today and I lock them in, okay? And since that is

5virtually always the case in all these models, if you

6find another example of a circumstance in which you say

7there is a real economic loss that results from this, I

8would like to see an economic analysis of why there was

9an economic loss there. So, I wait for some economist,

10the clever economists that Easterbrook was talking

11about, to come up with the explanation.

12I think I probably could for Microsoft as to why

13Microsoft's behavior might be a problem, but that is not

14similar to the ones that we have already talked about,

15okay? So, in -- I hate to do this with Gail here -- but

16in Dentsply, one of the things that was interesting

17about that case was that the Justice Department seemed

18to recognize early on that they needed to provide a de

19facto contract analysis as to why there was lock-in,

20okay? So, they said, okay, it is because of inventory

21investments. I bought so many inventories from these

22guys, from Dentsply, that if I walk away from them, I am

23stuck with the inventories, and the alternative

24explanation in that case said, hey, you really want

25those inventories to tide you over while you are trying

66

1to convert customers, right?

2And so, in fact, in that case, the lock-in

3turned out not to be lock-in, because Dentsply was happy

4to buy back those inventories, and the guy that walked

5away from Dentsply to sign up with rivals found that he

6sure wanted a hell of a lot more Dentsply teeth than he

7was going to get. So, there was no lock-in there in

8that case at all.

9And again, it is possible to imagine

10circumstances in which a manufacturer exerts or creates

11a property right for itself to take advantage of

12somebody who has sort of cooperated with it to develop a

13new product, and then the manufacturer says, hey, why

14don't I seize that new product on my own and define this

15property right and take that right away from the other

16guy?

17That is a problem, but that is almost as much of

18a fraud or a contract problem as it is an antitrust

19problem. It becomes an antitrust problem only if you

20get to the point where it says people are standing on

21the sidelines unwilling to invest because they are

22subject to this misappropriation of their up-front

23investments.

24So, I can imagine circumstances under which that

25might work, but I am not sure that you need to attack

67

1them in this sort of standard exclusive dealing context.

2MR. JACOBSON: I don't want to hog the mike, and

3I know Dentsply, we would get a very different view of

4the facts from people like Gail and Mark Bodde (ph), but

5what about Lorraine Journal? No contracts, you know --

6DR. MARVEL: Well, you brought that one up to

7me, and unfortunately, not being a lawyer -- and I am

8not a lawyer, even though I am a professor of law -- I

9am going to have to duck on that one, because I do not

10know the facts.

11MR. STEUER: Well, maybe if I can jump in --

12MR. VITA: Let me remind people, just pull the

13mikes up close to your face so they actually function.

14MR. STEUER: It may be that lawyers and

15economists do not always define "contract" exactly the

16same way, and lawyers get hung up with the whole Colgate

17 doctrine. In the case that I alluded to before, for

18example, a monopolist had 100 percent share of the

19market and came up with a discount schedule that

20basically made it advantageous for customers who needed

21to have some of its product to buy all of that kind of

22product from it so that when a new competitor opened its

23factory, it was facing the daunting challenge of having

24to replace all of the discounts that would be lost by

25potential customers giving up any of it. There was no

68

1contract.

2It was similar to a Colgate relationship that

3way. It was simply a unilateral policy, "Here is my

4price schedule if you do what I want you to do," and yet

5it seemed to have all of the foreclosure effect that a

6bilateral contract would. So, to some extent, maybe we

7are talking past each other a little bit in terms of the

8terminology and what is a contract and what is not.

9DR. MARVEL: Well, maybe so, but one of the

10things that you brought up, Richard, in your discussion

11was this NicSand case, right? And one of the things

12that has really impressed me about the cleverness of the

13post-Chicago world is how really imaginative they are at

14coming up with sort of contract-based explanations for

15why you could have problems, but, of course, the Chicago

16side does that, too, and you look at Lepage's and

17NicSand, and those are matters in which the Justice

18Department says we don't know yet what we should be

19doing, so let's wait a while before we have the Supreme

20Court step into that, or at least that is what happened

21in Lepage's.

22But, in fact, we are starting to figure out that

23those things involve -- I mean, maybe Lepage's was

24collateral damage, because there was a real problem with

25getting your entire line carried if you are going to a

69

1discounter, like a WalMart or a K-Mart. So, it is very

2possible that in a case like that, what you are really

3trying to do is induce the discounter that you are

4dealing with -- and this is particularly true for

5discounters -- to carry a much broader portion of the

6line than they would otherwise carry, and that is going

7to increase consumer welfare even though it is going to

8increase prices or it is going to increase economic

9welfare.

10So, I mean, you can get into these circumstances

11where you say, I don't understand yet why the

12manufacturer is doing this, so it must be foreclosure,

13but if you stand back for a while, maybe somebody will

14come along and say, hey, some of these bundling schemes

15have the efficiency effects that are pretty significant,

16and I think that cases like those may just be

17circumstances in which you are dealing with a guy who is

18going to carry a very narrow portion of your line, and

19you do not like that, so you pay him to carry a broader

20portion, and if somebody -- and you say, well, I am

21offering you this really good deal to carry the broader

22portion of the line, and maybe if that excludes somebody

23else, well, yeah, that could very well do that, but that

24is not the only effect of it, and so it is a really --

25these are really tough questions.

70

1MR. STEUER: Well, Lepage's had a "have to have

2it" kind of product in the bundle. NicSand is almost

3more interesting, because it was real competition for

4the contract, and I am not sure we have seen the last of

5that case.

6MR. JACOBSON: Well, it was a 12(b), so...

7MR. VITA: Anybody else? Josh, Mary, anything

8you would like to pose to the other speakers before

9we --

10MR. WRIGHT: I have one.

11MR. VITA: Yeah, go ahead.

12MR. WRIGHT: I maybe was being too sensitive to

13one of the comments, so I heard it directed at me, but

14Jonathan had mentioned that he --

15MR. JACOBSON: Ripping competitors' racks off

16shelves? Yeah.

17MR. WRIGHT: So, I think you either

18mischaracterized what I said, but since I didn't say

19anything about the shelves, then maybe that's not it,

20but to be clear, what the paper is about and what we are

21arguing about in the paper is the economic analysis of

22category management contracts, giving a procompetitive

23explanation for why, under some conditions, the retailer

24may want to delegate to the manufacturer the

25responsibility of the shelf space allocation decisions.

71

1That has nothing to do with the decision in Conwood.

2What the point is about the decision in Conwood

3is -- and I agree, and I am happy to say, court reporter

4and everything, that I agree that ripping shelf space --

5ripping displays down is bad, it is exclusionary. It

6would be bad --

7MR. JACOBSON: Makes no economic sense?

8MR. WRIGHT: -- it would be bad if -- also if

9the United States Tobacco employees sat out in the

10parking lot with bats and said don't come in and bring

11in product. All these things would be bad, but the

12point is about whether or not there is anticompetitive

13effect and whether or not there are any foreclosure

14effects and whether or not the conduct was sufficient or

15likely to generate anticompetitive effects.

16I know I am to the right of you on the panel, so

17I will use someone else. Professor Hovenkamp, in

18Antitrust Enterprise, using the testimony in the record,

19estimates the distribution cost increase as something

20like 33 cents per store per month, and there is some

21other evidence we talk about in the paper, but the idea

22is that there is this other question about whether or

23not there is a likelihood of anticompetitive effect and

24that even in the case of really nasty, nasty, bad, wrong

25conduct, we should be asking the question.

72

1MR. VITA: Mary, do you have anything?

2DR. SULLIVAN: Ah, no.

3MR. VITA: Okay, Brandon, why don't we move

4along then, and what we would like to do is put some

5propositions up and get some reactions from the panel,

6and I am going to go ahead -- I am going to read these,

7they have to be read into the record, so let me just go

8ahead and read the first one here, and this is a

9quotation from Justice O'Connor's concurring opinion in

10Jefferson Parish Hospital District Number 2 versus Hyde,

111984, and the statement is, "Exclusive-dealing

12arrangements are analyzed under the rule of reason."

13Let me just pose probably a simple question to

14the panel, and this is more to the lawyers, I think.

15Does this statement from Justice O'Connor's concurrence

16in that case accurately summarize the law regarding

17exclusive dealing? Richard and Joshua, Jonathan?

18MR. STEUER: I think it does. I think that the

19rule of reason is still a work in progress since Cal

20Dental, and we will see what the content is in judging

21these, but there really are three elements I think that

22go into it with exclusive dealing. One is the nature of

23the product and relationship, all the things that I

24talked about. The second is, of course, the percentage

25of the market once you have defined it that's

73

1"foreclosed," and the third element is the duration, the

2time period. So, I think those are the big moving parts

3in a rule of reason analysis, and the nuances await the

4development of the case law.

5MR. JACOBSON: Yeah, I agree with that. I was

6actually surprised, because this is also on the first of

7the questions that you sent out to us yesterday, that

8this would be perceived as controversial. I mean, the

9law is fairly clear about this, certainly under Section

101, and I think Microsoft and Dentsply, properly read,

11import this analysis into Section 2. The greater the

12market power of the defendant, the lower the degree of

13impairment of rivals you are generally going to require

14before you see a price effect, but I do not think this

15is a controversial proposition. So, I wonder what is

16 motivating the inquiry.

17MR. O'BRIEN: We didn't necessarily think it was

18controversial, but in this area where we are trying to

19build some kind of consensus in terms of what we all

20agree on, we thought we would start simple.

21MR. JACOBSON: Well, I "concense" this.

22MR. VITA: Josh, are you on board, too?

23MR. WRIGHT: I third the motion.

24MR. VITA: Let me follow up on that, then, and

25ask again, and anybody can step in here, does anybody

74

1think there are exclusivity arrangements that should be

2per se illegal? And similarly, does anyone think there

3are exclusivity arrangements that are always or nearly

4always procompetitive and are thus appropriate

5candidates for a safe harbor? Just if anybody has any

6thoughts on that, you can step in.

7MR. JACOBSON: Yeah, but dissent in the Harmar

8case, four Justices saying that exclusive dealing

9arrangements with multiple retailers are illegal because

10Klors as originally understood is correct, but I do not

11think anyone else believes that, and I think it would be

12really wrong-headed to circumvent, you know, 30 years

13now of rule of reason foray after Sylvania, to go back

14to a per se rule on exclusivity here.

15 I think there are going to be safe harbors, but

16they are basically going to be low market share safe

17harbors and in a properly defined market, and the open

18question in those cases is going to be, well, what if

19the whole market is tied up with exclusives as in

20Standard Stations? Do we really look just at the

21defendant's share of the market as a screen? I think

22the answer is yes, but I think it is a difficult

23question.

24MR. VITA: Anybody else?

25MR. WRIGHT: Sure.

75

1MR. VITA: Josh?

2MR. WRIGHT: The first question I think was are

3there any that should be per se illegal, no. And the

4second question is with respect to safe harbors, and I

5think in addition to the point about safe harbors for

6exclusives that do not foreclose some significant share

7of distribution, sort of foreclose trivial shares of

8distribution, then that is an appropriate place for a

9safe harbor.

10And I know there is at least -- I mean, there is

11not a consensus on this point about the duration of the

12contracts, but I believe it is certainly the case that

13short-term arrangements, like the ones we see in

14slotting, six months in duration, may also be, though I

15recognize this is subject to probably more debate, may

16also be appropriate for safe harbors.

17MR. STEUER: Some courts have misapplied the

18term "exclusive dealing" to both exclusive selling and

19exclusive buying. There is almost a safe harbor for

20exclusive selling other than those rare arrangements

21where one dealer has the exclusive for every brand there

22is, and there have been a couple of cases like that.

23In terms of real exclusive dealing, exclusive

24buying, there is almost a safe harbor of a third coming

25out of Jefferson Parish, talking about 30 percent.

76

1There are some other contexts where 20 percent is surely

2a safe harbor. I think that Jon is absolutely right,

3that the tough issue is, well, if somebody has an

4exclusive for 33 percent, but then there are two others

5who have 33 percent and 33 percent, and so there is 100

6percent exclusivity, that becomes more difficult, but

7Jefferson Parish for practical purposes has introduced a

8quasi-safe harbor of about a third.

9MR. VITA: Okay, Howard?

10DR. MARVEL: That is an awfully small harbor,

11but on top of that I wanted to ask you about the

12exclusive -- the exclusive -- which side did you put it

13on, seller is --

14MR. STEUER: Exclusive selling and sometimes it

15is called an exclusive distributorship, "You will be my

16only dealer in the State of Maryland" or something like

17that.

18DR. MARVEL: Yeah, but then turning that around,

19how do you regard an agreement extracted by a dealer

20like Toys 'R Us from seller -- a seller where he says,

21you know, don't sell to my rival the same product that

22you are selling to me. Is that okay?

23MR. STEUER: It can be. Again, if it extracts

24that from every manufacturer, that becomes increasingly

25a problem. If Toys 'R Us were to enter into an

77

1agreement with one manufacturer for one product and says

2"I want to be the exclusive seller of this product," it

3is rather limited what the impact is. In fact, I think

4the decree that was finally negotiated specifically

5provides for some limited exclusivity like that.

6But if one chain were to become powerful enough

7to sign up as the exclusive seller of all the toys for

8all the major manufacturers, obviously everybody else is

9frozen out, and I think there actually have been a

10couple of examples like that.

11DR. MARVEL: So, in Toys 'R Us, what happened,

12if I recall, was that the Seventh Circuit of all people

13said that the Toys 'R Us arrangement was not okay, and

14that is because Toys 'R Us did have this sort of

15monopoly position in the toy business, and it was

16unassailable -- because of their unassailable position,

17they really needed to protect the other poor souls like

18Sam's Club from the depredations of Toys 'R Us. So --

19is that right?

20MR. JACOBSON: Well, another way to --

21MR. STEUER: Well, Sam's Club or consumers. I

22mean, the classic example, there was a wholesaler on an

23island, I think St. Thomas, that was the sole

24distributor for, it turned out, every single brand of

25liquor, so that it basically created a bottleneck and

78

1had monopoly at the distribution level, and to the

2extent any of these examples approach that almost

3textbook model, then you have a situation where

4consumers really do not have other options at which to

5shop for those particular products.

6DR. MARVEL: So, is it an advantage to consumers

7when Toys 'R Us contemplates getting out of the toy

8business?

9MR. JACOBSON: Because of WalMart? Look, there

10were a lot of things going on in the case. One of them

11was that the facts supported a finding of a horizontal

12arrangement that was facilitated by Toys 'R Us, and I

13think that is what concerned Judge Wood most --

14DR. MARVEL: Right, absolutely.

15MR. JACOBSON: -- in terms of the significance,

16but looking at it purely on a vertical basis, at the

17time there was a credible theory that it was raising

18prices. Even though Toys 'R Us had a 20 percent market

19share nationally, there were pockets of the country

20where the share was in the high 40s, low 50s, and where

21they were a must-have retailer for Mattel and Hasbro and

22those other toy stores, and the result of this was that

23the real, you know, the real discounters were cut off by

24it, and you could make an arguable case that consumers

25were paying higher prices as a result.

79

1So, it was not -- it is not a crazy case. I

2think it is a tough case, but I do not think it was a

3crazy case.

4DR. MARVEL: Well, I brought it up because it is

5a tough case, but it is not a crazy case that what they

6were doing was actually in the interest of consumers.

7In fact, to have reasonably broad distribution of the

8lines of the toy manufacturers -- and, of course, we

9have also seen that not only has Toys 'R Us gone

10belly-up and KB Toys and FAO Schwartz, but also the toy

11manufacturers are rapidly fading into the sunset. Maybe

12that is because only one Tickle Me Elmo was -- one

13variety of Elmo was sold every Christmas at Sam's Club,

14maybe not, but it does not appear that that industry is

15a model of good health, and it may possibly be that that

16is because a vertical restraint that was contributing to

17not the monopoly behavior, but the good health of the

18industry, was expunged.

19MR. JACOBSON: Well, it may also be that our

20analysis of monopsony power/buyer power is in its

21infancy and that we really do not understand the

22ramifications of WalMart, and I think that is the larger

23issue, and I do not think anyone has a good answer to

24that.

25DR. MARVEL: I think that is right, because if

80

1you look at Conwood, for example, and what Josh was

2talking about, the Conwood case seems to me to have

3turned in part upon the, shall we say, hyjinks of the

4UST representatives who were trashing the Conwood

5racks --

6MR. JACOBSON: Right.

7DR. MARVEL: -- but what it really turned on was

8what was going on at WalMart, and that was a different

9tale entirely. They wouldn't dare trash the racks at

10WalMart, and so it kind of conflated those two things.

11I mean, I have come up with a number of sort of

12hair-raising anticompetitive activities that firms used

13to engage in, and it is easy to come up with these

14things, but that one is tough, because you start

15conflating these things, and then you get a decision

16that is made more on emotion than on what the economics

17of it are.

18MR. VITA: Let's go to the next slide, Brandon,

19and let me just again read this, but this discussion

20that Howard and Jonathan have been having I think sort

21of leads into this next proposition and some of the

22questions surrounding it. Let me just read it.

23This is a quotation from Posner's Antitrust Law,

24Second Edition, 2001, and in that book, Posner says, "I

25propose the following standard for judging practices

81

1claimed to be exclusionary: In every case in which such

2a practice is alleged, the plaintiff must prove first

3that the defendant has monopoly power...all the

4plausible cases of exclusionary practices involve

5defendants that have monopoly power."

6And so let me pose two questions, two related

7questions, you know, should monopoly power be a

8requirement for challenging an exclusive dealing

9arrangement under Section 1 of the Sherman Act and

10Section 3 of the Clayton Act, and related to that is,

11can exclusive dealing involving a non-monopolist result

12in substantial lessening of competition?

13And I think you two were already starting to

14discuss that. Let me see if anybody else wants to have

15any thoughts on that. Richard, Mary, Josh?

16MR. STEUER: Well, clearly I think one of the

17toughest areas is that space between 33 percent and 50

18percent, because when you get above -- where you are in

19the realm of Section 2 cases, the legalities change. I

20know this means nothing to economists, but it certainly

21does in terms of where you can get into court and

22whether you can stay there.

23The Microsoft case is an interesting example,

24because there, in terms of browsers -- and I don't want

25to dwell on this one case -- but certainly the share at

82

1the time the case was brought was very low, and that may

2explain why there was talk about monopoly power in

3operating systems, but if you look at it purely as a

4Section 3 type case and not searching for monopoly

5power, but even at a low market share, was there a

6danger -- an anticompetitive effect from the types of

7exclusivity that was being entered into? Purely on the

8numbers, you would say, no, the share is much too low,

9and come back when it gets higher, but we all know where

10that ended up.

11MR. VITA: Well, let me ask this, and this may

12be a question more for the economists, although the

13lawyers are free to jump in, too.

14Can we articulate or identify necessary

15conditions in the downstream market that -- conditions

16that are necessary for the exclusive dealing arrangement

17to have an anticompetitive effect? Are there certain

18things that have to be there before we have any ability

19to infer anticompetitive consequences from an exclusive

20dealing arrangement?

21Josh, got any thoughts on that?

22MR. WRIGHT: Sure. One -- I mean, let me make

23sure I understand -- I understand the question.

24MR. VITA: Yeah.

25MR. WRIGHT: So, when you say competitive

83

1conditions in the downstream -- you know, the downstream

2market, so I am envisioning a manufacturer with

3exclusive deals to a retailer --

4MR. VITA: Think about that, that's a good

5scenario.

6MR. WRIGHT: That's what I would think of as an

7example.

8MR. VITA: Yeah.

9MR. WRIGHT: I mean, substantial foreclosure

10on -- I mean, the sort of well-known conditions from the

11literature are that substantial foreclosure of the rival

12so he can't achieve minimum efficient scale is a

13necessary condition of most of these models, if not all

14of these models, and so I think that that is -- you

15know, in the legal analysis, we can have certainly, you

16know, in the economics literature is a necessary but not

17sufficient condition, and, you know, we know in the

18cases, there are cases that end up on both sides. We

19have a large foreclosure share but no liability because

20of short duration or entry conditions or some such, and

21so I think it is appropriate to use foreclosure as a

22necessary but not sufficient condition.

23MR. VITA: What about things like scaled

24economies in the downstream -- when you talked about

25scale economies, you were thinking about upstream, but

84

1what about downstream?

2MR. WRIGHT: So, in downstream, you can have --

3there are cases where if you have large economies of

4scale in distribution, you get -- you can have these

5exclusionary effects as well.

6MR. VITA: I mean, if there weren't substantial

7scale economies downstream, or maybe some other factors

8as well, do you think it would be possible in the kind

9of long run or medium run for exclusive dealing

10arrangements to have an anticompetitive effect? I mean,

11why wouldn't -- you know, because if you don't have

12substantial scaled economies and/or sunk costs at the

13retailing level, why can't the -- supposedly the

14foreclosed manufacturer get around the --

15MR. WRIGHT: Right, so if you have -- at the

16retail level you have -- I am going to frame this a

17slightly different way, but if you have -- even if you

18have the manufacturing scale economies but the retail

19level you have free entry condition, then you are going

20to have retailers who will re-align the supply

21contracts, new entrants into the retailers who will

22re-align the supply contracts, and so you need it at

23some level, and the theory is you can do it with

24economies of scale at the manufacturer level, but if you

25have free entry at the retail level, I think that is

85

1another problem for the exclusionary dealings.

2MR. VITA: Jonathan, you looked like you might

3have had something to add there.

4MR. JACOBSON: No, I actually agree with that,

5but it led into one of my sort of favorite topics in the

6space, which is let's not talk about foreclosure,

7because if we look at the percentage of distribution or

8retail outlets foreclosed without examining entry, for

9example, we may get a large number that's meaningless,

10and that is why I think we are a lot better off if we

11get rid of the word "foreclosure" and think about the

12impairment of the rival, because that is the mechanism

13that is going to lead to the consumer harm, not the

14foreclosure, as such.

15Foreclosure is a part of the analysis, but I

16think it is only part of the analysis. You have to look

17at the broader picture. Clearly there have to be

18impediments to entry downstream.

19And incidentally, I would agree with Posner's

20book depending on the definition of "monopoly power."

21You know, I think if you change it to market power, I

22think, you know, a lot of people would subscribe to it.

23I certainly would.

24DR. SULLIVAN: Yes, I have one comment to make

25on the -- following up on Josh's comment about free

86

1entry in the retailing level. I agree that if there is

2free entry in retailing, this is problematic for

3theories of exclusion, because the excluded manufacturer

4can more easily go to one of the new entrant retailers

5to obtain distribution, but on the other extreme, if you

6have, say, a monopolistic retailer, then I think that

7the exclusive dealing arrangements, it is very hard to

8prove that they would be harmful just because of the one

9monopoly rent problem. So, I think you need to -- there

10may be more potential for harm from exclusion in the

11more intermediate market structures.

12MR. VITA: Okay. Brandon, let's move on to the

13next slide.

14Here's another -- this is yet another quotation

15from Justice O'Connor in Jefferson Parish Hospital

16District Number 2 versus Hyde, and the proposition here

17is, "Exclusive-dealing arrangement 'may be substantially

18procompetitive by ensuring stable markets and

19encouraging long-term, mutually advantageous business

20relationships.'"

21Let me put a couple of questions out. You know,

22what are the -- empirically, what kinds of efficiencies

23do the panelists perceive to be most likely to be most

24significant in one of these exclusivity arrangements?

25And think about this, you know, are there efficiencies

87

1that are sometimes discussed maybe in the academic

2literature in connection with exclusivity arrangements,

3but in all likelihood, really aren't likely to exist or

4likely to be very important empirically in real cases?

5So, let me put that out there. Anybody --

6DR. SULLIVAN: Yes, I will take that one just in

7the sort of specialized area of slotting allowances. In

8the academic literature, people make a big deal out

9of -- one of the efficiencies of slotting allowances is

10that it signals the product quality to retailers of

11manufacturers' new products in cases where product

12default is uncertain, and based on a lot of the

13empirical studies that have been done by people in

14marketing, that is simply not one of the efficiencies

15that pops up, and I think the reason is there are quite

16a few tools that manufacturers use to introduce their

17products in addition to slotting allowances, and that

18just -- so, I would feel comfortable ruling that out as

19an efficiency, although there are plenty of other

20efficiencies involved in slotting allowances.

21MR. VITA: Howard?

22DR. MARVEL: One of the cases that Richard

23mentioned is the first nuanced case of exclusive dealing

24I think was Beltone, and I think it is fair to say that

25if there had not been some very un-nuanced evidence in

88

1that case, that Beltone would have gone down in flames,

2because by the time Beltone came up before the

3Commission, its four principal rivals in that particular

4channel that it was involved in had all met their

5demise, and so Beltone was left as the monopolist --

6thank you very much, FTC -- and at that point, they

7didn't really have a good explanation for why they were

8engaging in the exclusive dealing that they were

9engaging in, but -- and so I don't see how they really

10could have prevailed in that case unless there was this

11evidence that was pretty clear that the companies that

12had to give up the exclusive dealing practice had gone

13belly-up.

14So, in some ways John's paper talks about how

15there probably is not a case that you can find where you

16cannot determine that there are some advantages, but the

17real difficult problem is to figure out how important

18they are, and that is an incredibly difficult trade-off.

19It is very hard to measure these things.

20MR. VITA: Let me ask a follow-up on that point.

21What significance, if any, should be given to observing

22a challenged exclusive dealing arrangement in a similar

23but somewhat more competitive market? So, you know,

24that is sometimes an argument you make or you hear,

25that, well, you know, this particular arrangement must

89

1have some competitive benefits, because we see it over

2here in these other markets that are structurally

3competitive and where there is no plausible

4anticompetitive theory of harm. How much -- how

5powerful are those arguments and what weight should they

6be given?

7MR. JACOBSON: I think it is a much more

8powerful argument if a small company is doing it than if

9a large company is doing it in the same market. I think

10looking at comparable markets and saying exclusive

11dealing works efficiencies there, therefore they must in

12this other market, really depends on how similar the

13markets are. I would not make that leap without, you

14know, a good deal of comparability evidence.

15MR. VITA: Josh?

16MR. WRIGHT: A related point, I mean, the nature

17of the exclusive deal to facilitate some sort of

18contract or performance, in the slotting example, again,

19where the contract is over some sort of form of

20promotion, and you see this a lot in exclusive dealing

21cases where the underlying relationship between the

22manufacturer and retailer relies on some sort of

23promotional effort of the retailer and, in fact, is

24contracted for, but the nature of performance in these

25different markets varies a great deal, whether we are

90

1talking about putting a product on an eye-level shelf

2space or giving a product demonstration or some other

3form of promotion.

4So, the contracted-for conduct varies so much

5market to market, I think the best you can make out of

6seeing exclusive in a more competitive but different

7market is sort of one of a cautious inference that we

8generally know that exclusives can be procompetitive,

9which I think there is not much disagreement on anyway.

10MR. VITA: Okay.

11MR. JACOBSON: I have a question for Mary. If

12we renamed it payola, from payola to music leaders or

13retail music program, do you think we would get a

14different result?

15DR. SULLIVAN: No. I think the people at FCC

16and Elliott Spitzer would figure it out in a second.

17DR. MARVEL: Why don't we call grocery store

18slotting allowances payola?

19DR. SULLIVAN: Well, I think we could, and one

20thing you could do --

21MR. JACOBSON: Because we would like to win the

22cases.

23DR. SULLIVAN: -- if the FCC regulated slotting

24allowances, they would require the cashier at the

25checkout counters to tell the customer each time he or

91

1she was buying a product for which a slotting allowance

2had been paid, then say, do you still want to buy it?

3MR. WRIGHT: Well, as funny as that is,

4California had proposed at one point -- I think it is

5still kicking around in committee --

6MR. JACOBSON: No, it was killed.

7MR. WRIGHT: It was killed now?

8MR. JACOBSON: Yeah.

9MR. WRIGHT: Senate Bill 582, which would have

10made -- it would have been illegal for -- essentially a

11retailer would have to tell Pepsi exactly what Coke was

12paying in terms of its promotional allowances, in terms

13of the slotting fees, and if you conceive of these

14things, these payments, as I do, as part of the

15competitive process, I mean, this is a statute that is

16a -- it is, you know, a legislatively enforced

17collusion, right? And so it is silly, but, you know,

18not silly enough to write down in a bill.

19DR. MARVEL: Was it going to be the California

20Raisin and Coca-Cola board? Is that --

21MR. JACOBSON: It was proposed by a coalition of

22the same people who represented the plaintiffs in the

23Gruma case and the Harmar case. I mean, it was serious,

24and it did get some traction, but it got killed fairly

25early on in committee.

92

1MR. VITA: Okay, let's move on then. The next

2proposition is from Dennis Carlton from his article in

3the Antitrust Law Journal, "A General Analysis of

4Exclusionary Conduct and Refusal to Deal -- Why Aspen

5and Kodak Are Misguided," and Carlton's proposition is

6as follows:

7 "In the presence of scale economies, exclusive

8dealing can be a way of depriving Firm 2 (or its

9distributors) of the necessary scale to achieve

10efficiencies, even though, absent the exclusivity, Firm

111 and Firm 2 would both be large enough to achieve

12efficiency."

13So, two related questions for the panel. One,

14do you agree with Dennis' statement, and secondly, other

15than its potential to deprive competitors of scale and

16the resulting effect on prices, are there any other

17theories of harm from an exclusivity arrangement that

18should be the subject of antitrust concern?

19DR. SULLIVAN: I will try the second question.

20There is a theory -- and this is one I referred to in my

21presentation -- by Greg Shaffer, a 2005 theory, and he

22had a theory of exclusion in which scale economies did

23not play a role, but what was going on is the retailing

24segment was very, very competitive, and essentially

25retailers, without exclusion of a manufacturer, would

93

1earn almost no profits because their segment was so

2competitive, and they could easily be coerced into going

3along with an exclusivity deal that would exclude one of

4the manufacturers because it simply would increase the

5industry profits, and he developed conditions under

6which this was true. One might argue that that would be

7fairly unusual, but it -- you know, it is there.

8MR. VITA: Anybody else? Dan, did you want to

9add something?

10MR. O'BRIEN: I would just like to ask, Mary,

11following up, in that kind of a theory, if a

12manufacturer could secretly get to a -- get with a

13retailer, okay, assuming that everybody else was being

14coerced into this exclusive with the manufacturer, and

15negotiate something on the sly, wouldn't they be able to

16undercut what, you know, the monopoly price that was

17presumably being set by the other guys?

18DR. SULLIVAN: I think so, and I think there was

19something in particular about the nature of the game

20that Greg set up that allowed him to get this outcome,

21so I agree that might be -- it might not be that

22problematic in reality.

23MR. STEUER: There are a lot of assumptions in

24here obviously. It makes a huge difference whether the

25exclusivity is with end users and for how long. If this

94

1is simply competition for the contract, clearly if one

2manufacturer can get exclusive arrangements with the

3bulk of the end users and freeze out the other, that is

4going to have a profound impact, but if the second

5manufacturer can survive long enough to go and bid the

6next time and try to get the contract back, that is very

7different. We do have some situations in defense, for

8instance, where there is only going to be one winner of

9these contracts. They are always exclusive, and yet you

10do have some back and forth bidding as long as both

11companies can survive. Here, I presume the assumption

12is, with economies of scale, that there is a danger that

13one of the companies disappears off the face of the

14economic map.

15MR. JACOBSON: I think this identifies a case --

16there are certainly exceptions, as Richard points out,

17but I think this identifies the exclusive dealing case

18that you ought to worry about, you know, if these

19conditions are holding, this is the case you ought to

20worry about. There may be other cases you ought to

21worry about. There may be cases where this is not a

22problem because it is competition for the contract, but

23in terms of our analysis, this is where I think we

24should focus most of our resources.

25I would add that this is an excellent article,

95

1although Aspen I do not think was misguided, although

2that is debatable, and Kodak was clearly correctly

3decided.

4MR. VITA: Whoa.

5MR. O'BRIEN: Okay, we will try -- if we have

6got time -- we have got time for one more, I think. Oh,

7one more on this? On this proposition or one more

8proposition?

9MR. VITA: One more proposition, I think.

10MR. O'BRIEN: I think we can go a little longer.

11MR. VITA: All right, Brandon?

12Okay, this next proposition is from Herbert

13Hovenkamp, Antitrust Enterprise, 2005, and I will read

14it.

15"Exclusive dealing is a rule of reason offense,

16requiring a plaintiff to show that the defendant has

17significant market power, that the exclusivity agreement

18serves to deny market access to one or more significant

19rivals, and that market output to consumers is lower (or

20prices higher) as a result."

21A couple of questions for the panel. As to

22significant market power or some indicator of

23significant market power, is there or should there be a

24safe harbor? And does anybody have an -- you know, it

25says here in my notes that Jonathan in his writing

96

1suggests courts apply a 40 percent market share safe

2harbor, and if that -- you know, is that actually true,

3and does anybody have an alternative minimum requirement

4that they would prefer?

5So, let me put those two out, those two

6propositions out there and see what the panel thinks.

7MR. JACOBSON: Well, I generally agree with what

8I said.

9MR. VITA: Glad to hear that.8

10MR. JACOBSON: I think this is a pretty good

11quote. I think "market access" needs a little bit of

12definition, because I do not think you need -- this was

13one of the other questions that we had talked about

14before the program -- I do not think you need total

15foreclosure. Again, I think the test needs to be the

16degree of impairment of rivals. So, as long as denying

17market access is read in that context, I think this is a

18pretty good analysis.

19I think 40 percent is a pretty good rough

20screen. I think Richard's correct to point out that

21Jefferson Parish is a 30 percent number, but it does not

22say anything about a screen here or there, but if you

23look at the subsequent cases, you are not going to find

24any where the defendants have liability with less than

2540 percent unless you consider Toys 'R Us an exclusive

97

1dealing case, and there, you know, there were

2extenuating circumstances given the horizontality of the

3agreement.

4MR. STEUER: And the term in here "significant

5rivals" is significant, because it really raises the

6question, who should have a cause of action here? At

7some point, if there is ample competition in a market

8and there is exclusive dealing going around, there may

9be some marginal players who claim that they are being

10excluded, and those can be emotionally appealing cases

11in terms of jury appeal, and yet in terms of what the

12actual effect is on the market, it may be very marginal

13indeed, and there are not very clear tests right now as

14to who should be able to bring a claim.

15MR. O'BRIEN: If I could follow up with that,

16John, earlier you had said that one of the areas in

17which there was an agreement, you listed four points,

18one of which was we want to prevent the enhanced -- you

19know, practices that enhance market power. I am

20wondering if you would agree with the last part of this

21proposition, which is that plaintiffs have to show to

22successfully bring a case that market output goes down

23and/or prices go up.

24 MR. JACOBSON: Well, I think what he means is

25that market output is likely to go down, and if you show

98

1there is a significant enhancement or creation of market

2power, I think you have done that. So, I do not think

3this is inconsistent with that proposition.

4MR. VITA: Okay, let's move on then.

5This next proposition is from United States

6versus Microsoft, the D.C. Circuit en banc decision.

7The quotation is as follows:

8 "If the monopolist's procompetitive

9justification stands unrebutted, then the plaintiff must

10demonstrate that the anticompetitive harm of the conduct

11outweighs the procompetitive benefit."

12A couple of questions, and again, this may be a

13little more for the economists, but anybody can step in.

14First of all, does economics supply tools that

15would assist courts in making this kind of assessment,

16and do courts have the ability to apply these kinds of

17tests?

18Let me stop right there and see what the

19reaction is from the economists on the panel.

20DR. MARVEL: How about no?

21MR. VITA: Say again?

22DR. MARVEL: Do the courts have the tools? No.

23MR. VITA: Actually, the proposition was, can we

24as economists supply tools that courts could use? I

25mean, what kind of analysis, if any, can we provide that

99

1will allow noneconomists to make the kind of

2determination that the Court called for in this case?

3DR. MARVEL: I think that you really need to be

4very careful about if you show anticompetitive harm, it

5is pretty clear that you have got anticompetitive harm,

6then I guess once you have gotten to that point, unless

7convinced that the procompetitive benefits you are

8trying to demonstrate will be easily enough measured and

9ready available in such a way as to make it possible for

10the courts to do the trade-off, I just think they are

11awfully hard to prove what they are.

12So, if you can really show that somebody is

13locked out by the nature of the arrangement -- and that

14means from the market, that does not mean from the

15channel that the manufacturer in question controls, but

16from the market as a whole -- then it is going to be

17hard to do this trade-off, but if you have got the

18anticompetitive harm and people are absolutely convinced

19that it is there, then I think that that might be

20enough.

21MR. VITA: Yes, Josh?

22MR. WRIGHT: Well, I think in this particular

23quote, we have to -- there may be differences with

24respect to what economists can do before and after -- in

25the first and second clauses, right? The economist

100

1might have tools to supply with respect to understanding

2a monopolist's procompetitive justifications. Something

3we can do is understand why we might see exclusives,

4understand why conduct might be procompetitive, and the

5conditions under which those explanations are likely.

6That is something we can do and should be doing.

7It is a lot tougher, the challenge of doing the

8balancing is much tougher, and I guess the part that is

9not in this quote is that the first step of requiring

10the plaintiff to show the likelihood of some

11anticompetitive effect is also an area where economists

12can contribute by explaining the conditions for

13anticompetitive effects are either satisfied or they are

14not.

15MR. O'BRIEN: Do you want to follow up, John?

16MR. JACOBSON: I mean, this is what my article

17is all about, so I do not want to leave this one

18untouched.

19A, most cases do not reach the level where you

20need balancing. The number of cases where you really

21need to balance it are few and far between. Usually the

22case will fail because a prima facie case of

23anticompetitive effect will not be shown. If that is

24shown and the defendant shows a significant

25justification, usually the plaintiff gives up at that

101

1point. So, it is a very rare case that requires

2balancing.

3But if balancing is required, I think we need to

4do it, and to say -- to throw up our hands and say it is

5too complicated is just completely the wrong answer. We

6do it every day. This building is filled with people

7doing that in merger cases. It is done at the Justice

8Department in merger cases all the time. This is

9exactly what we do. So, to say that we are not going to

10do this, it is too complicated, we might as well just

11get rid of antitrust, because this is the guts of what

12hard antitrust cases are all about, and we not only want

13to do this, but we have to do it. This is one issue I

14feel very strongly about.

15MR. O'BRIEN: So, I wanted to follow up with

16Howard, and, John, you may want to chime in on this,

17too. You are concerned that if we can establish that

18there may be an anticompetitive effect, that it is often

19very hard for defendants to come in and argue, well, no,

20in fact, there are efficiencies and that they offset the

21anticompetitive effect, and I --

22DR. MARVEL: No, what I am saying is that if you

23can really show anticompetitive harm and --

24MR. O'BRIEN: That may or may not be offset by

25efficiencies, okay, so that is what I am saying. It may

102

1or may not be offset, and what I took you to be saying

2was that --

3DR. MARVEL: That would make it really tough

4for -- once you have a compelling demonstration of

5anticompetitive harm -- and that is compelling for me,

6not for you --

7MR. O'BRIEN: Right.

8DR. MARVEL: -- then I am not so sure that -- it

9reminds me of the original merger guidelines when they

10did not allow efficiencies as a defense, and I do not

11think that that was absolutely nuts. So, if there is a

12strong demonstration of anticompetitive harm -- and that

13is not just locking up a channel, that is locking up the

14market -- then I am not sure how much balancing I want

15to do at that point.

16MR. O'BRIEN: I see.

17MR. JACOBSON: It is a rare case, Dan, it is a

18rare case where you need to do this, but there can be,

19at least in theory -- I will tell you, I have never seen

20one -- but there can be one, at least in theory, where

21the effect of the exclusives is to create a market

22structure such that the defendant can raise prices to

23some extent.

24However, there may be sufficient dealer focus as

25one traditional efficiency or other effects that overall

103

1output of the product is increased. Think about your

2resale price maintenance cases, the same -- it is the

3same type of analysis, and if you can show -- first of

4all, the burden is on the plaintiff, not the defendant,

5but if the defendant can put in evidence to say that

6notwithstanding the price increase, we are going to have

7a significant overall market output effect that is going

8to be procompetitive, I think you have got to entertain

9that defense, and then I think you have got to see

10whether that is true at the end of the day. Is the net

11effect going to be to increase output or not?

12MR. O'BRIEN: I guess I -- I am sorry.

13DR. MARVEL: I think maybe if I can go, John's

14point, I think part of the disagreement with -- the

15implicit disagreement here is in my determination of

16what constitutes an anticompetitive effect, because I

17certainly would not agree to that parenthetical remark

18that Hovenkamp had that said that prices are higher than

19they would have been if the restraint was taken away.

20Well, you cannot do that, because all of these

21explanations talk about setting up a property right that

22allow you to get a return on your investment which could

23very well take the form of, you know, if you shift up

24the demand curve, you are going to get a higher price

25and greater output. If you get more output, end of

104

1story. If it is a higher price, that does not really

2tell you much of anything, and so that is I think part

3of what we are -- we may be agreeing, somehow have a

4different setup.

5MR. O'BRIEN: So, following up on that, Howard,

6I am curious how you feel about something like the no

7economic sense test as a way to, you know, ask is there

8a plausible efficiency rationale and, you know, maybe

9short-circuit this balancing some.

10DR. MARVEL: Sorry, but I -- it hurts me, but I

11would have to agree with John on that one. I do not

12like the test.

13MR. O'BRIEN: Okay. Why don't you like the

14test?

15DR. MARVEL: I think your explanation is that

16there is always economic sense in these practices, and I

17think that that is right, that there will always be some

18plausible argument that could be made. Unless we are

19talking about gunning down your rivals or some such,

20anything short of that, you are probably going to be

21able to come up with some plausible argument on behalf

22of that.

23MR. JACOBSON: One convert, not a bad morning.

24MR. VITA: Well, with that, then, we will bring

25the morning session to a close. I would like to thank

105

1the panelists. This was a really great discussion, and

2I think everybody got a lot out of it. So, thanks very

3much.

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

1AFTERNOON SESSION

2(1:31 p.m.)

3MR. O'BRIEN: Okay, let's get started. Well,

4welcome to the second exclusive dealing panel of the day

5in what is part of our ongoing series of public hearings

6on single-firm conduct. My name is Dan O'Brien. I am

7the Chief of the Economic Regulatory Section at the

8Antitrust Division, and I will be moderating this

9session along with Mike Vita, who is the Assistant

10Director in the Economics Group, the Bureau of Economics

11at the Federal Trade Commission.

12The Department of Justice and the FTC are

13jointly sponsoring these hearings to help advance the

14development of the law concerning the treatment of

15unilateral conduct under the antitrust laws.

16Transcripts and other materials from the prior sessions

17are available on the DOJ and FTC web sites, and I just

18wanted to advertise that upcoming panels include a panel

19on bundled loyalty discounts on November 29th, obviously

20a practice that is somewhat related to exclusive

21dealing, which is the topic for today, and then there is

22a panel on misleading and deceptive conduct on December

23 6th.

24So, today's session concerns the law and

25economics of exclusive dealing. It was 40 years ago in

107

1the Brown Shoe case that the Supreme Court made a very

2strong statement against exclusive dealing, asserting

3that it conflicts with the central policy against

4contracts that take away the freedom of purchasers to

5buy in an open market.

6Since that time, the treatment of exclusive

7dealing by the courts has changed fairly dramatically

8over time, and the economics of exclusive dealing has

9progressed, identifying both procompetitive and

10anticompetitive aspects of the practice depending on a

11range of circumstances.

12We have a very distinguished group of panelists

13here this afternoon to talk about these developments and

14the current state of affairs from both the legal and

15economic perspectives. My goals from today's panel are,

16first, to highlight some areas hopefully where there is

17some consensus on the effects of exclusive dealing and

18how to treat it, but also maybe identify questions that

19remain unsettled so we can have some consensus about the

20questions that need to be addressed as we move forward.

21So, before introducing the panelists, I just

22wanted to thank my colleagues at the FTC and at the

23Antitrust Division, particularly June Lee and the

24economics staff at the Antitrust Division and Joe

25Matelis in Legal Policy. The two of them together did a

108

1lot of the work in putting together this panel.

2The organization of the panel is going to be as

3follows: We have four panelists. They will give

4presentations of approximately 15 minutes. Then we will

5take a short break. Then the panelists will have a few

6minutes to respond to the other presentations if they so

7desire, and then we will have a moderated discussion,

8and we can go until around 4:00 p.m.

9So, the order of the panelists, in case people

10are wondering, will be Steve Calkins first, Tad Lipsky

11second, Joe Farrell and then Ben Klein. So, let me

12introduce Stephen Calkins. He is our first speaker.

13Stephen Calkins is Professor of Law and Director

14of Graduate Studies at Wayne State University Law School

15where he teaches courses and seminars on antitrust,

16trade regulation, consumer law and torts.

17From 1995 to 1997, Steve served as General

18Counsel of the Federal Trade Commission. Steve lectures

19widely throughout the U.S. and abroad, most recently in

20Europe and New Zealand. He has authored many

21publications on competition and consumer law and policy,

22and he has served on the editorial boards of well-known

23journals in antitrust.

24Stephen?

25DR. CALKINS: Thank you. Thank you for the

109

1introduction. What was not said is that I am actually

2the most novice of all the people who are speaking here

3today. I mean, you go over everybody else, and they

4have been an expert witness in one or more of the

5leading cases, they have litigated one or more of the

6leading cases. Richard Steuer, in the previous session,

7got up and proceeded to point out that he had published

8three articles specifically on exclusive dealing. I

9have never been an expert witness on exclusive dealing,

10I have never litigated, I have never done an article

11about exclusive dealing, as such, you know, we are

12talking about somebody who is just not in the same

13ballpark. So, with great humility, let me just tell you

14that I am trying to sort out my own thinking and to

15learn from all these geniuses.

16To do that, we need to start somewhere, and so I

17found one interesting case that I thought I would begin

18just looking at a little bit, and here is a court

19opinion that talks about how exclusive dealing "may well

20be of economic advantage to buyers as well as to

21sellers, and thus, indirectly of advantage to the

22consuming public," and these advantages may often

23explain why there are exclusive dealing contracts, and

24if you wanted to go and understand whether they were

25harmful or beneficial, you would look at a series of

110

1tests.

2You would look at "evidence that competition has

3flourished, despite use of the contracts," or you would

4look at the conformity of the length of their terms to

5the reasonable requirements of the field of commerce, or

6you would look at the status of the defendant as a

7struggling newcomer or an established competitor or the

8defendant's degree of market control, and you would go

9through all this sort of stuff, but the opinion goes on

10and says that to do this would just be extremely

11difficult and to sort everything out would be an immense

12challenge and, using words very similar to sort of the

13basic sort of Areeda Hovenkamp mantra, we need to have

14tests that are administerable by courts, we need to have

15rules that can be enforced without wasting a lot of

16societal resources on hopelessly complex litigation that

17can't lead to any predictable outcomes, and so for

18reasons of administrative efficiency, exclusive dealing

19contracts should almost all be illegal, because this was

20the original Standard Oil/Standard Stations case with

21those thoughtful observations about the procompetitive

22benefits of exclusive dealing, but the conundrum, the

23difficulties, of litigating.

24So, when I sat down and took a look to start my

25sort of thinking about this and went back in time, I

111

1said, golly, what an interesting beginning place, and I

2then decided to pull out key dates in exclusive dealing

3history, and we began with the classic Supreme Court

4cases, which have been reviewed a little bit in the

5morning session, and I will not mention them except that

6Standard Oil you know, Brown Shoe was just referenced,

7the classical Supreme Court cases were certainly

8important moments in exclusive dealing history.

9That led us to the key year of 1977 when all of

10antitrust, as we know it, changed with Continental TV,

11and then along came Robert Bork and the antitrust

12paradox -- actually, along came all of the Chicago

13School -- but Bork in particular is associated with

14exclusive dealing, because he said so emphatically that,

15by golly, there is only one monopoly profit. Exclusive

16dealing cannot increase a monopolist's monopoly profit,

17and so, therefore, "if Standard finds it worthwhile to

18purchase exclusivity, the reason is not the barring of

19entry but some more sensible goal such as obtaining the

20special selling effort of the outlet," emphatically

21saying that one cannot increase the profit of the

22monopolist, and so there must be a procompetitive

23justification, and those Supreme Court cases were just

24dead wrong, a really clarion call for a different way of

25looking at exclusive dealing.

112

1As mentioned in the previous session, that call

2was picked up first in the courts or the adjudicative

3bodies in the Beltone Electronics opinion, where the

4Court specifically relies on Bork and the antitrust

5paradox to take a different approach to exclusive

6dealing, the Federal Trade Commission, leading the way

7to a new day of exclusive dealing decision-making, even

8if we learned in the last session at the cost of having

9sacrificed four of the five competitors, but

10nonetheless, having led the way, that was followed

11shortly thereafter by Jefferson Parish. Of course, it

12is always cute, we refer to the Jefferson Parish

13exclusive dealing holding, and it wasn't a holding at

14all. It was part of the concurrence of Justice

15O'Connor, but we all think of it as the holding from

16Jefferson Parish where she emphatically said exclusive

17dealing is judged more permissively than tying, it is

18rule of reason, and "exclusive dealing is unreasonable

19restraint on trade only when a significant fraction of

20buyers or sellers are frozen out of a market by the

21exclusive deal."

22And since then, if you look at things that have

23happened and you sort of parade through the exclusive

24dealing cases that we know, which I throw up on the

25screen in front of you or I throw up more of the

113

1exclusive dealing cases that we know or I throw up more

2of the exclusive dealing cases that we know, the one

3great unifying principle is, of course, that the

4defendant always wins. There are a few exceptions, but

5overwhelmingly, the judicial treatment of exclusive

6dealing ever since Beltone Electronics came down has

7been that defendants win these cases, and you can find

8support in the case law for all sorts of pro-defendant

9propositions, with exclusive dealing being strongly

10presumed to be legal if there is a market share of less

11than 40 percent, if the restraint is of less than a

12year, the contract is of less than a year, if the

13contract is easily cancellable, if we do not have a

14complete and total foreclosure, see the words in

15Jefferson Parish, if there are no entry barriers, and

16on, there are probably other ones as well, a whole

17series of different principles, standards under which

18defendants have won these cases, and that's a whole lot

19of the exclusive dealing story, and then there is the

20"but" part of the whole thing that makes our life

21slightly interesting here.

22There are three things to mention. The first,

23the post-Chicago literature, I have reason to suspect,

24although I did not look at his slides, that Joe Farrell

25will reference a little of this, and it can be done in

114

1all sorts of wonderful mathematical sophistication. I

2think of the lesson as a common sense story of

3collective action.

4There was recently a case that Tad knows dearly,

5the Coca-Cola case just decided by the Texas Supreme

6Court. I do not know anything about the facts of that

7case, and I have no opinion on the case. I do not know

8what happened down there, but one of the things that

9allegedly happened was that Coca-Cola paid retailers not

10to allow 7-Up in its stores, and if you think about that

11for a minute, you know, it sort of sets out the

12collective action story very crisply. Why would a

13retailer agree not to carry 7-Up when it knows that if

14in the long run there is no 7-Up, that is probably bad

15for retailers? And the answer is, of course, that if a

16payment goes to a single retailer, that single retailer

17can collect the payment knowing that its excluding of

187-Up is not really going to make a difference in the

19long run, and you do not have all the retailers getting

20together and agreeing that they will resist Coca-Cola,

21because that would be illegal under the antitrust laws,

22and so each separate retailer looking at its individual

23self-interest can quite reasonably say, I will agree not

24to allow 7-Up in my store, even though in the long run,

25that is against the collective interests of all of them,

115

1and it is because of that kind of a collective action

2problem that exclusive dealing can sometimes harm

3competition in the long run because one can have an

4exclusive dealing arrangement that helps someone today,

5with all the benefit going to that one entity, in the

6long run, there is harm, but the harm is shared widely,

7and so, therefore, you have a mismatch between the

8benefit of the harm, a collective action problem, and

9therefore, mischief can be worked.

10Two cases have come along that have sort of set

11out the -- sort of the other ways of thinking about

12exclusive dealing, being Microsoft and Dentsply. People

13in this room know those cases far more than I do, but

14just mentioning a couple of points quickly, Microsoft

15is -- you can find several different points in the

16Microsoft opinion on exclusive dealing. This is one

17where the District Court had said that there must be

18complete and total exclusion before there is a

19violation, and the Court of Appeals wrote that "even

20assuming the holding is correct," and went on to say

21there could still be a violation, thereby suggesting

22that that holding may not be correct.

23It went on and said there could be a violation

24because there is a different standard under Section 2

25than under Section 1, and even if something might be

116

1lawful under Section 1, it could be unlawful when

2engaged in by a monopolist. The Court asked rather

3tough questions about the justifications for the

4practices going on there, specifically saying that with

5respect to one practice, where 14 of the 15 top Internet

6access providers had contracts to work only with

7Microsoft, the justification was to keep them focused on

8Microsoft's product, "which is to say it wants to

9preserve its power in the operating system market, that

10is not an unlawful end, but neither is it a

11procompetitive justification," thereby raising nice

12questions about the difference between a benefit to the

13seller and a benefit that qualifies as a procompetitive

14justification.

15Also of interest to the Microsoft case is we had

16a very economically sophisticated court unable to resist

17quoting some language indicating subjective intent.

18"Kill the cross-platform Java by growing the polluted

19Java market," so on and so forth, finding some comfort

20in the words that business people had used to describe

21what they were doing, and then finally being troubled,

22even though we did not have total exclusion. So, we

23have a whole series of interesting points that come out

24of the Microsoft case.

25In the Dentsply case, what did we have in

117

1Dentsply? You had something where you had an at-will

2contract, and yet the Court of Appeals said that was not

3reason for the defendant to prevail, because

4realistically, wholesalers are not going to give up $22

5million in sales in order to pick up $200,000, and so an

6at-will contract does not really give a new entrant

7realistic access to the market. So, also, there was

8talk about monopoly maintenance as a separate kind of

9problem, and once again, we had reference to subjective

10intent evidence.

11So, where am I at that point in terms of, as I

12end, little lessons that I draw from my sort of going

13over things, and they are very tentative, because I

14really have not thought these things through all the

15way. I am learning, okay, but tentative things that I

16might throw out as propositions.

17One, it should be possible for a short-term

18contract or contract that is cancellable still to be

19found to be unlawful. It should be possible for there

20to be illegality without total exclusion. Section 2

21standards should be tougher than Section 1 standards.

22It does not make sense to take all of the teaching of

23Section 1 exclusive dealing cases and then import them

24unthinkingly into the world of Section 2. If you have a

25firm with a 75-80 percent market share and entry

118

1barriers and lots of power, it ought to be tougher than

2on a smaller, less powerful firm.

3I hesitantly think that it is -- this will not

4be popular with some of my panelists -- sometimes it is

5interesting and possibly informative, if done very

6carefully, to look at subject intent evidence to help

7you sort through these difficult things. Clearly it

8makes sense to scrutinize the procompetitive

9justifications that are being offered up in a case that

10otherwise looks troubling. The classic procompetitive

11story is that the manufacturer has expended resources to

12bring a consumer into the store who will then be bait

13and switched off to another product. Well, you know, do

14the facts fit that story or not? In Dentsply, the Court

15thought they did not fit that story but went on to try

16to really sort of sort through what is the

17justification. It should not be enough just to say it

18is a nonprice vertical restraint.

19I personally would not think that one should

20require a plaintiff to prove that prices have increased.

21I mean, think again about your classic exclusive dealing

22situation would be something where we are trying to

23cause problems in the future. Go back to my Coke paying

24to have 7-Up not around. The reason to do that is so

25that things will be better for Coca-Cola in year two or

119

1three or four or five, and one can have a lessening of

2competition without prices today being affected. The

3hard question here is the long-run competitive effects,

4though, can't be a complete defense to say that current

5prices have not gone up.

6So, also we would say that the legal standard

7really does matter in these cases. Going back to

8previous sessions that you have had, you heard a lot

9about the no economic sense test in the last session.

10Another standard that can make a big difference in

11exclusive dealing cases is whether you choose to adopt

12the Posner "Exclude an equally efficient firm" test.

13Were you to adopt that, which I would not favor, that

14would make it much harder for a plaintiff to win an

15exclusive dealing case.

16And finally, in closing, pretty much on time, it

17is interesting as you survey the landscape that there is

18a whole lot of theory, not a great deal of empirical

19evidence, and so I hope that this program, if nothing

20else, inspires some people to go out there and get their

21hands dirty and bring forth more empirical evidence.

22Thanks very much.

23(Applause.)

24MR. O'BRIEN: Okay, our next speaker is Tad

25Lipsky. Tad is a partner at Latham & Watkins and a

120

1former Deputy Assistant Attorney General at DOJ. Tad's

230-year legal career has been devoted mainly to

3antitrust, and it spans virtually every facet of

4competition law.

5From 1981 to 1983, Tad served as Deputy

6Assistant Attorney General at DOJ under William Baxter.

7Following government service, Tad developed a broad U.S.

8and international antitrust practice, successfully

9managing a variety of important antitrust matters.

10As chief antitrust lawyer for the Coca-Cola

11Company from 1992 to 2002, Mr. Lipsky conducted and

12supervised competition matters before courts and

13antitrust authorities in the U.S., Canada, the EU, EU

14Member States, and dozens of other jurisdictions. He is

15a frequent author and speaker on antitrust topics.

16Tad?

17MR. LIPSKY: Thank you very much. Until a few

18moments ago, I had forgotten how stupid it was to follow

19Steve Calkins to the podium, because he knows more about

20whatever he speaks about than anybody else and expressed

21his interesting views so trenchantly and with such great

22humor that that is a very tough standard, but I will do

23my little bit and see if we can find something to agree

24on. I think we can find a few things to disagree on,

25and we will see where it goes.

121

1Exclusive dealing is a very elastic label. It

2applies to a lot of different kinds of things. We have

3already heard mention of the fact that tying, certain

4kinds of bundling and price discounting can have effects

5very similar to exclusive dealing, and therefore, when

6you talk about exclusive dealing, you also need to be

7considering a bunch of its very, very close relatives,

8and so we are talking about implicitly, at least, a very

9broad category of business conduct and competitive

10phenomena.

11Now, on the plus side, for our policy evaluation

12of exclusive dealing, it has never been a per se

13offense, which is a very good thing. It is a little

14like saying, well, in Eastern Europe, they have a little

15better luck re-adopting capitalism, because they were

16capitalists within living memory, whereas in the old

17Soviet Union, in the heart of Mother Russia, that was

18not the case, and so there is no great body of learning,

19there is no familiarity in the culture, and similarly,

20with exclusive dealing, although it is true that back in

21the Standard Stations days and when we were dealing with

22the International Salt comment, that under Section 3 of

23Clayton, you could condemn exclusive dealing either if

24the defendant had market power or if there was not an

25insubstantial amount of foreclosure, that is coming

122

1within an eyelash of saying it is per se, but we never

2quite got there.

3There was always a little bit of procompetitive

4culture left in exclusive dealing, and so -- as a matter

5of fact, even in the dark ages, between the decision in

6Schwinn, all vertical agreements are illegal per se,

7until the release from bondage in 1977 with Sylvania

8taking the nonprice verticals out of that category, I am

9not aware of any decision going whole hog and saying,

10well, that because of Schwinn, now we have to say that

11exclusive dealing is per se. Even in those dark days,

12we never had a rule for exclusive dealing that said

13basically shoot on sight.

14So, now, having escaped per se condemnation, I

15think it was easier for exclusive dealing cases to kind

16of re-absorb the economic learning, to talk about

17procompetitive justifications, to insist upon genuine

18proof that the process of competition had been

19obstructed before liability could be imposed. We went

20from Standard Stations, we went to Tampa Electric, which

21basically said, well, even quantitative foreclosure does

22not really give us the story that we want to hear when

23we are talking rule of reason. And so, in effect, this

24evolution is kind of a testament to just how thoroughly

25the microeconomic analytic approach has been absorbed in

123

1the antitrust enforcement industry, the enforcement

2agencies, the courts, counselors, what have you, and

3this is all very much to the good. This is as it should

4be.

5But one result of this emergence into the more

6full-blown consideration of justifications and actual

7competitive effects is that the role of market power and

8monopoly power have been pushed to the fore, and for

9most kinds of exclusive dealing claims, you need to have

10market power or monopoly power at one level in order to

11have any kind of a plausible theory of restraint, and so

12now it has become a topic that is addressed more under

13the Section 2 standards than under Sherman 1 or Clayton

143, and that is fine. So, that focuses, to the extent

15that these issues come up under the Section 2 rubric,

16that focuses you on monopoly power, because it is a

17required element of proof in every Section 2 case, or in

18an attempt case, of course, the reasonable likelihood of

19monopoly power being attained -- and it also means

20that -- it really brings us down to I think the main

21discussion, the main subject of discussion, which is the

22definition of monopolizing conduct, and, of course, that

23is a much broader area, and let's see what light we can

24shed on the exclusive dealing aspect.

25Well, one of my colleagues, Steve Calkins, has

124

1already alluded to the fact that if you look at

2exclusive dealing cases, there are not many in which

3plaintiffs win, and it is interesting that some of those

4cases are really not Section 1 or Clayton 3 cases

5anymore, they are Section 2 cases, oddly enough, in

6which the decision-maker for one reason or another

7failed to condemn exclusive dealing under Sherman 1 or

8Clayton 3, but only under Section 2, and that would

9include U.S. v. Microsoft, Lepage's v. 3M, sort of in

10the margins of exclusive dealing, one of those forms of

11bundling, and then we have heard about U.S. v. Dentsply.

12Now, within the broader debate about legal

13standards for monopolizing conduct, exclusive dealing I

14think is more or less kind of a classic example. What

15do we have to go on when somebody is challenged for

16their conduct under Section 2? Well, we have Grinnell,

17we have Aspen, exclusion on the basis of something other

18than efficiency; we have Image Technical Services, not

19the part that everybody has had seminars about and

20talked about for years and years and years, and Salop

21said this and somebody else said that and it is

22post-Chicago -- no, it is pre-post-Chicago -- okay, it

23is post-modernist Chicago, but the point is there is a

24second part of Kodak versus Image Technical, which say

25what you will about the tying part, the first part of

125

1the Supreme Court opinion, there is that second part

2that makes some extremely broad characterizations of

3what it takes to -- broad and vague characterizations --

4of what it takes to prove monopolization. That part of

5the opinion was so good that when Image Technical got to

6go back and have its trial, it did not even bother with

7all the hard post-Chicago stuff in the first part. It

8just relied on that great language in the second part of

9the opinion. So, it is really a question of

10deconstructing and coming up with a monopolistic conduct

11standard that can be applied sensibly to the generality

12of these cases.

13Now, I will put all my cards right on the table

14and say I am not one of those who says there is

15salvation to be had in taking the vague language of

16Grinnell and the vague language of Aspen and the vague

17language of the second Section 2 part of Image Technical

18versus Kodak and trying to put some kind of a

19microeconomic overlay on it, whether it is no economic

20sense, profit sacrifice, exclusion of equally efficient

21competitor. I think all of those things can come in

22very handy. I mean, if you see a monopolist doing

23something that causes it losses, you are entitled to

24inquire, is it an eleemosynary motive, was it a mistake,

25or was the monopolist taking money and paying for

126

1something, and was it a competitive restraint? So, I do

2not want to suggest that those concepts are useless, but

3I think they are not going to get us the distance to a

4standard under Section 2 for judging exclusive dealing.

5As a matter of fact, I am prepared to say that

6as a general matter, any standard that is simply stated

7and purported to apply to the generality of exclusive

8dealing cases cannot possibly give you specific enough

9guidance to decide any particular case. This is just

10one of those situations where we are kind of stuck with

11the dilemma that Steve referred to in the initial part

12of his remarks when he was quoting from Standard Station

13saying, well, you know, we would love to consider all

14these justifications, but, you know, it would not be an

15administerable rule of law, there is nothing you could

16do with it. Therefore, we are going to have a per se

17rule based on quantitative substantiality.

18It is a little bit like my favorite footnote in

19Topco. Remember United States v. Topco, which was a

20horizontal case, and it was a bunch of independent

21grocers who had banded together and had arranged to have

22their own private label line of products to offer in the

23grocery store so they could compete with A&P Ann Page

24and Safeway's whatever, and the District Court had said,

25well, this is very procompetitive as a rule of reason,

127

1case dismissed, and the Supreme Court said, oh, no, oh,

2no, when you are talking about a horizontal restraint --

3and it was a territorial restraint in that particular

4case -- what the Supreme Court said is you don't

5consider all that stuff, it is per se, and then they

6dropped a footnote that said, well, look, if Congress

7would like to adopt a rule of reason for this kind of

8restraint and send the courts off into the wilds of

9economic theory -- that's the exact phrase they use in

10that footnote in Topco -- Congress can go to that, but

11we are not going to, per se illegal, next case. So, we

12have got a similar situation here.

13Exclusive dealing could be good, could be bad,

14depends on a lot of different factors, very hard to

15formulate a different -- a reformulation of a general

16standard that is going to apply in all circumstances,

17and so I have very little faith in any such

18reformulation. I think we are just stuck, you know,

19courts do what they do. You have got a difficult area

20where it is hard to make a judgment. Actually, as I

21think as I am going to talk about toward the end of my

22remarks, which will be soon, what I am basically saying

23is if the courts find it difficult to take such an

24amorphous standard and apply it to this practice, what

25we have to have is better courts.

128

1Now, we have mentioned that defendants almost

2always win. So what? So what? I have no great faith

3in the numerology of one loss statistics. The real

4question is whether anticompetitive conduct gets struck

5down in these cases and procompetitive conduct is

6exonerated, and by that standard, as I read the same

7cases that Steve has obviously read -- and he has

8probably spent a lot more time reading them than I have

9and has read a lot more cases as well -- but I find it

10very difficult to say that something is seriously awry.

11I have cases where I would disagree with what is

12going on, but there are two cases in the -- well, I have

13talked about Microsoft, U.S. v. Microsoft, Lepage's v.

143M, U.S. v. Dentsply. I have listed -- have I listed in

15my -- well, anyway, three cases I could name where the

16defendants won, three recent important cases where the

17defendants won, Pepsico versus Coca-Cola, this is the

18New York case affirmed by the Second Circuit where

19basically the Second Circuit said you do not get a trial

20on the proposition that the reason quick-service

21restaurants do not buy Pepsi-Cola is that Pepsi-Cola

22cannot figure out a way to deliver the syrup to the

23restaurants. Whatever reason there is for the relevant

24market shares in quick-service restaurants for

25carbonated soft drinks, it is not that Pepsi-Cola could

129

1not figure out a way to get its product delivered.

2Omega Environmental versus Gilbarco, I do not

3know any more than what you, the average case reader,

4knows. I had no involvement with that case. Then we

5have Harmar Bottling, which is, again, a case that I do

6know something about. I am not sure the facts bear the

7characterization that Steve was giving it. I do not

8want to get into a cat fight with him over that, but I

9will just say that I think the result in that case was

10correct, and so of the cases I know, of the cases I have

11read about and tried to understand, I do not think you

12can say that defendants are winning in cases where they

13should not win.

14So, you know, we need to figure out a way to

15assess exclusive dealing efficiently, and basically, as

16I say, my message is there is some exclusive dealing

17that is good, some exclusive dealing that is bad.

18Harmar took about 14 years to tell one from the other,

19and my main message is that there has got to be a way of

20getting to an efficient resolution of these cases much

21more quickly. As a matter of fact, I would consider

22whether -- I might regret this if it became a sound

23bite, but if there is a sound bite I would give you,

24let's have the antitrust enforcement mechanism, let's

25adopt as a policy objective, that in the area of

130

1exclusive dealing, we want to reduce the duration and

2the expense of deciding whether exclusive dealing in a

3particular case is good or bad. Let's reduce the

4duration and expense by an order of magnitude so that a

5Harmar, which took 14 years to litigate, takes, say, 14

6months to litigate.

7Now, in this column, I have very high praise for

8the Ann Bingaman suit against Microsoft which resulted

9in the 1994 consent decree. I know that there was some

10investigation prior to the time that the DOJ got the

11file in that case, but I remember being incredibly

12impressed for two reasons with that effort. Number

13one -- well, other than feeling that the result was

14right. It was a consent decree, but I think it did the

15right thing.

16Number one, it was about exactly one year

17between the time that the Department of Justice got the

18file in that case and the date that the decree was

19entered, and number two, it was a very specific,

20targeted form of relief. It was a doable form of

21relief. So, if you can do an exclusive dealing case

22that quickly and come up with a result that concrete in

23a year, it forgives almost any other defect that you can

24find in that case, because on that time scale, you can

25correct for your mistakes. You can, you know, do in

131

1year two what you failed to do in year one, or vice

2versa. So, litigation efficiency is an extremely

3important consideration, and we ought to figure out ways

4for a great increase in litigation efficiency.

5One minute, that is exactly what I need.

6So, here are some ideas for enhancing the

7efficiency of this process, and I think a lot of the

8tools are already at hand. Daubert, it has already been

9used in an exclusive dealing context. Let's have more

10of it. Let's make sure that expert testimony is forced

11to go through and survive a plausibility test, the

12Daubert standard. Let's make sure that the plausibility

13formulation in Matsushita and Brooke Group, even though

14that is relative to predatory pricing, a plausibility

15test should also be applied to other types of antitrust

16claims, including exclusive dealing, help filter out

17losing claims early, and focus remaining claims on all

18phases for the remainder of the litigation, so you are

19not carrying forward speculative theories and going

20through the wasteful discovery and legal motions and so

21forth that that involves.

22Second, expand the use of neutral expert or

23expert panels, and I want to emphasize here, it is not

24just in a strict Rule 706 sense, in other words, an

25expert witness providing economic testimony to a judge

132

1in a matter in litigation, like Fred Kahn's testimony in

2the New York versus -- the Nabisco Brands case. That

3was a very effective use of a 706 expert, but we need

4ways to bring specialized knowledge about antitrust

5cases, discovery, theories, the nature of the market, we

6need to put those resources at the service of the courts

7that are having these exclusive dealing litigation

8things litigated before them.

9And the last one I won't go through due to the

10shortness of time, but the Manual for Complex Litigation

11does contain a few things about antitrust, but perhaps

12of the ideas that we could expand, the sort of helpful

13guidance, the identification of issues, the suggestion

14of efficiency-enhancing methods of resolving complex

15litigation, expand it specifically in the area of

16monopolization and exclusive dealing for the use of the

17courts.

18So, just to sum up, I do not think that our

19exclusive dealing jurisprudence is in crisis. I kind of

20like where the law is. Some exclusive dealing is good,

21some exclusive dealing is bad, it is not per se legal,

22it is not per se illegal, but if we could reduce the

23time it takes to tell the difference between good

24exclusive dealing and bad exclusive dealing by an order

25of magnitude, I think that would be a very worthy goal

133

1for the antitrust policy.

2(Applause.)

3MR. O'BRIEN: Thank you, Tad.

4Okay, our next speaker, shifting gears to a

5couple of economists, is Joe Farrell. He is Professor

6of Economics at the University of California, Berkeley,

7and he is a Fellow of the Econometric Society, former

8editor of the Journal of Industrial Economics and former

9President of the Industrial Organization Society.

10Currently he's the senior consultant for Charles River

11Associates.

12Joe's published widely articles on a broad range

13of topics in industrial organization and microeconomics,

14including exclusive dealing. He has substantial policy

15experience as well, having served as Chief Economist at

16the Federal Communications Commission from '96 to '97

17and Deputy Assistant Attorney General for Economics at

18the Antitrust Division from 2000 to 2001.

19Joe?

20DR. FARRELL: Well, I am an economist. I am

21going to talk about economics for a few minutes, and

22then I am going to talk about the law. I feel all right

23about this because I hear a lot of lawyers talking about

24economics.

25So, economics for the most part in antitrust

134

1analysis has focused on the question, what should we do

2if we knew really quite a lot about the case, okay? And

3in the area of exclusive dealing, I think a bland and

4very fair summary of economics in this area is both

5efficiency and anticompetitive effects and explanations

6of exclusive dealing are very possible, and on both

7sides of that, the analysis is really quite subtle, and

8I am going to spend a few minutes on this. In terms of

9the efficiency explanations, I am going to focus on the

10investment incentive theory, which I think Ben Klein is

11also going to talk about a form of. In terms of

12anticompetitive effects, I am going to talk about what I

13think is the leading example, though not the only

14example, of an economic structure to understand

15anticompetitive effects of exclusive dealing.

16So, in terms of the investment incentives, you

17will often hear it said that exclusive dealing is

18efficient if you have to motivate relationship-specific

19investment or some such phrase as that, okay? As far as

20I know, the state of the art in the economics literature

21on these arguments is the article by Elias Segal and

22Michael Whinston in the Rand Journal, 2000. They start

23out by showing that in what appears to be quite a

24general model, relationship-specific investments, that

25is, investments that have no value outside the

135

1relationship, are not -- repeat, not -- an efficiency

2rationale for exclusivity.

3They then continue to show that investments that

4are not in that strict sense relationship-specific, that

5have a spillover to deals between the customer and the

6potential entrant, might or might not be an efficiency

7rationale for exclusivity. It depends on quite a number

8of things. It depends on who is doing the investment.

9Is it the buyer or the seller? It depends on how it

10spills over. Is it a complement or a substitute with

11the efficiency of potential deals between the buyer and

12an entrant? It depends on the bargaining structure

13between the buyer and the seller. It depends on what is

14the nature of any investment by us absent the exclusive

15dealing. And that is all within their model. If you

16step outside that model, it also depends on whether

17their model sort of applies or sort of does not apply.

18So, I am going to leave you for the moment with

19the thought, how is a court likely to be able to

20disentangle all this in addressing an asserted

21efficiency rationale along the lines of investment

22incentives?

23Now, what about the other side of the courtroom,

24divide and conquer exclusion, Rasmussen and Ramseyer and

25Wiley, 1991, corrected, beefed up and radically improved

136

1by Segal and Whinston in the American Economic Review,

22000, show that exclusion can profitably and harmfully

3work against end users; however, although I think that

4is very well understood and accepted, the fact is their

5models involve buyers who are end users.

6In most cases that I am aware of, exclusive

7dealing is not a deal struck with end users. It is a

8deal struck with retailers or distributors or someone

9else intermediate in the value chain between the

10manufacturer and the end users. That makes a lot of

11difference.

12So, interestingly, a year or two ago, there

13appeared to be economics literature, two broadly

14parallel articles, papers, one by Fumagalli and Motta,

15which I believe has been published or is about to be

16published in the American Economic Review, and one by

17John Simpson and Abraham Wickelgren, and within the last

1824 hours, I have learned about other articles by Yong

19and Shaffer that may be somewhat along the same lines,

20and both of these articles address the question, how

21does the RRWSW theory of anticompetitive exclusive

22dealing change when you recognize that the buyers in the

23model, in practice, should be replaced by buyers who are

24not end users?

25Well, there are two forces, okay? One force is

137

1that intermediate buyers, nonfinal buyers, actually do

2not care that much if the price goes up or stays high,

3provided it goes up or stays high to all of them,

4because then it gets passed through downstream, okay?

5How much that is true depends on the details of the

6market structure and so on, but that tends to be true.

7That lowers their resistance to things that maintain

8monopoly upstream relative to what it would be if they

9were end users. So, that you would expect would make

10anticompetitive exclusive dealing easier.

11Another force, however, is that if you have a

12nonfinal buyer who holds out and does not sign the

13exclusive deal, then an entrant can come to him and say,

14"Aha, I will give you a lower price than all your tied

15up rivals will be getting. You can expand. You and I

16can meet my scale requirements, and you will make a

17bundle of money." So, that dynamic potentially makes it

18harder to have anticompetitive exclusive dealing.

19Well, Fumagalli and Motta found conclusively

20that it went one way, and Simpson and Wickelgren found

21conclusively that it went the other way, and which way

22Yong and Shaffer come out, I do not know yet. Which of

23them is right and when? Well, I attempted to diagnose

24this in my Antitrust Bulletin article last year. My

25attempted diagnosis is that it depends on whether in

138

1that last situation where you had one hold-out buyer,

2the incumbent is then able to or does adjust the price

3that it charges the tied buyers. So, I believe

4Fumagalli and Motta assumed that it does not, and

5Simpson and Wickelgren assumed that it does, or maybe it

6is the other way around, okay?

7When I put this tentative diagnosis to one of

8the four economists -- and I will not say which one --

9the response I got was, "Ah, that is interesting, I am

10not sure." That is telling, I think, because it says

11that it is kind of unlikely that a court is going to do

12a very good job of disentangling all of these difficult

13concepts. Now, the optimistic view is this is just the

14beginning of the economic exploration of this topic, and

15come the year 2010, we will understand it well and in a

16way that is good enough for us to brief courts on it,

17and maybe that will happen, okay, but I take from this

18two things.

19One is economics is making progress, that is

20great, I hope to participate, but the other is, it is

21pretty subtle and it will probably stay pretty subtle,

22if not get more subtle.

23All right, so we are doing antitrust under

24uncertainty. We are not in the world where we can say

25exactly what is going on and work out the welfare

139

1consequences, okay? Let's take that as an assumption

2for now.

3Well, traditionally at this point economists

4plunge into Bayesian mode and talk about type one errors

5and type two errors and so on. Underlying what I am

6going to say, there certainly is a Bayesian framework,

7okay, but I am not going to talk explicitly in Bayesian

8terms. I am going to talk in jurisprudential terms,

9because my lawyer colleagues on this panel have been

10talking economics, so I want to get back at them.

11So, I am going to talk about the role of

12presumptions and burdens of proof, and I am going to

13talk about two presumptions that should be extremely

14important in antitrust policy and about what I

15personally think -- although I cannot prove -- is a very

16worrying trend that has been taking place in the

17relative strength of these two presumptions.

18So, what are these two presumptions? Number

19one, in economic policy generally, in market economies,

20we have a laissez-faire presumption. The Government

21should not intervene in stuff unless it is reasonably

22sure that intervention will help. I think that is a

23pretty good idea.

24Number two, in antitrust particularly, we should

25protect competition unless we are reasonably sure that

140

1some alternative is better, okay? So, I think at a very

2grand, 40,000-foot level, you can view a lot of what

3goes on in antitrust jurisprudence as being a tug of war

4or back and forth between these two presumptions.

5Now, I put competition in quotes on this slide

6for a reason, and that reason is when you look at it too

7closely, things get a little out of focus, and you do

8not exactly know what that word means, okay? And that

9has led us, I believe, over the course of the decades

10towards the tempting solution of redefining the word

11"competition" to mean what is good. So, here is a test

12of that, okay?

13What happens when you hear someone refer to the

14possibility that a merger to monopoly would reduce

15marginal costs so much that it would be good for

16efficiency and consumers? Well, if that were true,

17let's say you knew it was true, it would be a good

18thing. Would it be procompetitive? I think a lot of

19people would say yes, because it is a good thing, but

20that is ridiculous. It is not procompetitive. It is

21pro-consumer, it is pro-efficiency. It is not

22procompetitive.

23So, if we are going to use words in their real

24meaning rather than redefining them so that the

25definition does our policy analysis for us, we have got

141

1to be a little careful about doing stuff like that.

2Now, of course, the antitrust law protects

3"competition," so tautologically, redefining the word

4would be a good idea, it would lead us to do good

5policy, if we always knew what was going on, okay? So,

6given that the law protected competition, it would be a

7very smart move on the part of benevolent antitrust

8enforcers and courts and so on to redefine the word

9"competition" so that the law then protects whatever is

10good, okay?

11However, there is a problem with doing this. A,

12we do not always know what is going on exactly, and B --

13B only applies given A -- attempting to have a

14presumption in favor of protecting competition makes no

15sense if you define competition to mean what is good,

16okay, because if you knew that something was good, you

17would want to do it, and that is not a presumption in

18favor of protecting competition. So, for there to be

19any meaning to the presumption in favor of competition,

20it has to be a presumption in favor of something that

21has not yet been proved to be good, okay?

22So, this I think casts an interesting light on

23the slide that I heard this morning -- and I was not

24taking notes on who said it -- but somebody said

25something along the following lines, or if I misheard

142

1it, it has certainly been said within the last week --

2that because there are perfectly plausible efficiency

3justifications for exclusive dealing, plaintiffs should

4be required to prove that there is an anticompetitive

5effect, okay? That, of course, would be obviously right

6if we could always prove what is true, but if we cannot

7always prove what is true, it is not obviously right.

8It might still be right, but it is not obviously right,

9okay?

10So, in order to explore this, let me, with

11tongue in cheek, put the shoe on the other foot, okay,

12and let's suppose that we applied the same redefinition

13to the laissez-faire presumption, okay? So, we have

14this presumption that says the Government should not

15intervene unless it is pretty sure that intervention is

16a good thing, okay? So, now let's suppose that we

17defined laissez-faire as the good outcome, and we

18defined intervention as the bad outcome.

19Now, if the Government wants to come along and

20insist that you paint your bedroom walls blue, not

21white, you can't say that is intervention, because you

22have not proved that it is a bad thing, okay? Well,

23that is obviously pretty stupid.

24So, I come out of this thinking it would be a

25good idea for us to make sure that words go on meaning

143

1what they mean, and it is very dangerous -- it has had

2some good consequences, but it is nevertheless very

3dangerous -- to redefine words to make them do your

4policy analysis for you.

5So, antitrust intellectual history, to the

6extent that I understand it -- in less than one

7minute -- in the bad old days, anything that could be

8presented as a reduction of competition was illegal.

9That was bad, because quite often, things that can be

10presented as a reduction of competition are actually

11good. The good new days, we have got to analyze the

12effects of things that seem to be capable of being

13presented as a reduction in competition, because you

14would not want to ban those things if they are actually

15good, okay?

16What I am worried about is the possibility that

17we are drifting into the not so good new days where it

18is difficult to prevent things that are in some sense

19reductions of competition unless you can actually prove

20that those things are bad. Now, of course, you would

21not want policy to prevent those things unless they are

22bad, but that is very different from unless you can

23prove that they are bad.

24Now, the final bullet on this slide, which is

25quite important, I talked about these ideas very briefly

144

1with some people in Europe over the summer, and they

2were aghast. Why were they aghast? Because they said

3we have spent years trying to move away from a

4descriptive basis of liability towards an effects-based,

5economics-based concept of liability, and now, you are

6coming from over there and trying to undo that. Well, I

7take that seriously, so I am not going gung ho on a

8policy proposal here, but it does seem to me that if too

9much burden of proof is being imposed, that is a

10problem.

11Let me finish with this slide, dark matter, do

12the physics, okay? It is a good idea to intervene only

13if intervention benefits efficiency or consumers. It is

14maybe not such a good idea to intervene only if you can

15specifically prove that and how it would do so, okay?

16There are multiple benefits of competition in

17most circumstances. Often, there are concrete,

18predictable, provable price effects, okay? Merger

19simulation has been a very powerful tool in exploring

20that. There is also the much vaguer and harder to pin

21down possibility that having a bunch of different firms

22doing different things and independent of each other can

23lead us to benefits that are much harder to prove or

24even define or even point to ex ante, okay? I call this

25the dark matter of competition policy, because as in

145

1astrophysics, if it exists, it is quite likely to be a

2lot bigger and more important than the stuff that you

3can see. So, watch out when you are imposing burdens of

4proof.

5(Applause.)

6MR. O'BRIEN: Thank you, Joe.

7Our final speaker is Ben Klein. He is Professor

8Emeritus of Economics at UCLA and a director at LECG.

9Ben is an internationally recognized expert on antitrust

10economics. He was a Professor of Economics at UCLA for

1134 years where he published numerous articles on a range

12of topics, including antitrust, contracts and

13intellectual property.

14He currently serves on the board of editors of

15five academic journals. Over the past 25 years, he has

16consulted extensively on antitrust issues and has made

17numerous presentations to state, federal and foreign

18regulatory agencies and courts.

19Ben?

20DR. KLEIN: Thank you, Dan.

21I am going to be talking mostly about economics,

22and although what I am going to say is subtle, you

23should not reach the conclusion from what Joe said that

24because the arguments are subtle that, therefore,

25anything goes. Just find the economist that is going to

146

1make the argument you want to hear. I think there is

2truth out there. This is moving us along on coming up

3with what is the economic foundation for some commonly

4used procompetitive justifications.

5This is a paper that I am working on with Andres

6Lerner. The paper is posted on the web site, and I

7think it is important to go through these procompetitive

8justifications in terms of the economics, because the

9danger I see is the exact opposite one. I think that we

10are moving in the direction that if you find a practice

11that does not have efficiencies, it is becoming a

12sufficient condition, if it is something that is being

13used by a firm, a large firm, it is a sufficient

14condition for antitrust liability, because the very

15nature of an exclusive dealing contract is

16"exclusionary," and then when you get to the balancing,

17you have nothing on one side of the scale.

18Although the paper discusses many exclusive

19dealing cases, we concentrate on Dentsply, and that is

20what I am going to concentrate on today, and it is

21because the court in that case used economics to reject

22two very common procompetitive justifications, both free

23riding and this undivided dealer loyalty justification,

24and the principles that I am going to be giving you here

25can be applied to a number of very different claimed

147

1justifications, and we do it in the paper.

2So, in terms of Dentsply, as I said, Dentsply

3illustrates that actually the economic foundations for

4procompetitive justifications are actually pretty

5narrow, and the Court rejected Dentsply's claim, in

6particular, that exclusive dealing was used to prevent

7dealer free riding on manufacturer-supplied promotional

8investments. This is the classic Howard Marvel

9rationale, where the manufacturer makes investments in a

10dealer, you know, like they build out a dealership or

11engage in dealer training, and then the dealer uses

12those manufacturer investments to sell a rival product,

13and that is the classic free riding argument. The Court

14rejected that, and the Court rejected the undivided

15loyalty argument, that somehow you give somebody an

16exclusive so they will more actively promote Dentsply's

17product.

18The Court rejected the free riding rationale

19basically because the Court found it was contrary to the

20facts, that number one, Dentsply did not make any

21investments in the dealers that they could then free

22ride on by using them to sell rival products. There was

23no evidence, essentially no evidence in the case, that

24the Dentsply dealers were actually switching buyers to

25rival products. And finally, that there was testimony

148

1by Dentsply executives that if there was not an

2exclusive, they actually would have invested more -- you

3see, the usual economic argument is the purpose of

4exclusive dealing is to encourage the manufacturer to

5make investments, and one way it is encouraged to make

6investments is to prevent free riding that it knows that

7these investments are going to be used to sell its

8product, and the Court said, you know, the Dentsply

9executives actually testified that if we did not have

10exclusive dealing, we would have had to make more

11promotional investments.

12In terms of the other argument, the Dentsply

13Court rejected the undivided loyalty argument, and here

14it was not really just the facts. It was basically the

15theory that this theory about enhancing dealer services

16cannot be a justification for exclusive dealing, because

17in general, competition between dealers is going to lead

18them to supply the desired quantity of promotional

19services, as the Court said, the dealers have the

20incentive in competing with other dealers to make sure

21that they supply the right kind of services.

22See, basically the problem that Dentsply ran

23into is although this undivided loyalty argument has

24been accepted by a number of courts, Judge Robinson in

25this case knew a lot of economics, and in particular,

149

1she knew Howard Marvel's argument and had read the

2article, and Howard Marvel was the expert that Dentsply

3had hired for this, and she said, no, even in your

4expert's article, he says that you can generally leave

5it up to competition to put dealers to supply the right

6services. It is only when you have this problem, this

7inter-dealer free riding problem described in Sylvania,

8you know, and that is a problem where the customer goes

9to one full-service dealer and gets some kind of dealer

10services and then goes to another dealer and buys the

11product, you have that inter-dealer free riding problem,

12and in that circumstance, maybe competition among

13dealers will not give you the right quantity of dealer

14services, but that is a problem that would not be

15corrected with exclusive dealing, because even if you

16had exclusive dealing and you had this kind of problem,

17the exclusive dealer would say, no, get the services

18from somebody else and then come and buy the

19manufacturer's product from me.

20So, as I said, although this rationale has been

21accepted by a number of courts, Judge Robinson said, you

22know, basically you can leave it up to competition, and

23this undivided loyalty makes absolutely no economic

24sense.

25In contrast to basically the established

150

1economics, I think the expanded economic framework that

2I am going to present here shows that these arguments

3make sense, that free riding is much more general than

4you would think, and the dealer undivided loyalty makes

5sense, and it is based upon two common sense business

6propositions.

7Number one, that manufacturers often want their

8dealers, even dealers that are competing with one

9another, to supply more promotion than the dealers would

10independently provide on their own, and number two, that

11exclusive dealing by creating this undivided dealer

12loyalty actually increases the dealer's incentives to

13supply these desired services and to more actively

14promote the manufacturer's products.

15So, in terms of the logic of what I am going to

16do, first I am going to discuss the first proposition,

17and hopefully people have been here in the morning and

18heard Josh Wright, who has done this already, but

19basically the first proposition is manufacturers, in

20general, cannot leave it entirely up to the dealer

21competition to get the quantity and the type of services

22they want supplied, and the logic of the argument is

23there is another step where, therefore, they have to

24contract with their dealers, either explicitly or

25implicitly, to solve this problem and to make sure that

151

1they adequately promote their product. That leads to

2these free riding problems, which I will discuss are

3much broader than the classic Marvel free riding

4problems. And then finally, that exclusive dealing is

5commonly an element in those contractual arrangements

6that gets the individual dealers' incentives then

7aligned more with the manufacturer's incentives.

8So, let me do the first proposition first, that

9manufacturers often want their dealers to supply more

10promotion than the dealers would independently decide to

11provide, and the basic reason for this is that the

12dealers do not take account of the manufacturer

13profitability on incremental sales, that the dealer does

14something that increases the manufacturer's sales, and

15the dealer gets only a part of that incremental profit,

16in many cases only a very small part of the incremental

17profit.

18Now, in general, this is not a problem for

19dealer price and nonprice competition that has

20significant inter-dealer quantity effects. So, in

21general, when a dealer provides a desirable service like

22free parking or lowers its price a little bit and makes

23a little bit more sales, even though they might have a

24small margin in terms of the total profit being earned

25by the manufacturer and retailer together when they make

152

1that extra sale, because they are getting consumers to

2switch from other dealers, because there is large

3inter-dealer effects, you get an equilibrium where you

4get the desired quantity of the services provided, but

5with promotional activity, the primary effect is not

6really inter-dealer, but it is primarily inter-brand,

7that you just make an extra sale for the manufacturer,

8and there are no significant inter-dealer quantity

9effects. Then you have this problem where the dealer,

10by not taking account of the incremental profit, is

11going to supply less than the desired promotional

12services of pushing the manufacturer's product. In

13addition, dealers cannot charge consumers directly for

14those services, because the promotion is, in effect, a

15price discount.

16So -- I am going to have to go faster --

17manufacturers solve this problem -- although I am going

18to be talking about violating these contracts, I can

19always violate, you know, this one --

20MR. O'BRIEN: There is no red string we can pull

21--

22DR. KLEIN: No, there is no self-enforcement

23problem here, although -- anyway, I am wasting my time.

24Manufacturers solve this problem of insufficient

25dealer promotion by contracting with and compensating

153

1dealers for providing increased promotion, and the

2contract may be explicit or it -- you know, in plenty of

3the cases, like in Standard Fashion, they explicitly

4said you have to have a certain amount of display space,

5you have to have a "lady attendant" there full-time,

6they used a few words like that. Most of the times it

7is really understood that you are going to make your

8best efforts, and they compensate dealers in these

9things by giving them a valuable distributorship in the

10sense that if they get terminated because they are not

11pushing the product adequately, they are going to lose

12this future rent stream, and the threat of termination

13is what gets them to perform as desired.

14However, because dealers are contracting to

15supply more promotion than they would otherwise, you

16know, do in their own independent interests, there is an

17inherent problem in that they have an incentive to

18violate the contract and free ride on the manufacturer's

19compensation arrangement, basically because you are

20getting a valuable dealership, like in Beltone, they

21gave them an exclusive territory. In Standard Fashion,

22they had minimum resale price maintenance. Whatever it

23is, you have something valuable, but you are getting it

24on all your sales, and you therefore have an incentive

25just to do that pushing at the end and save the cost if

154

1you are a dealer, and still you are getting most of the

2compensation.

3In terms of this contract, dealers may violate

4the contract and free ride in three distinct ways, and

5the first way is the standard case where the dealers use

6the manufacturer-supplied investments to sell rival

7products, and that is part of the contractual

8arrangement. Look, we will give you these complementary

9assets to help you push our product, and that is one

10that you know about, but there is two other free riding

11problems.

12Second is the dealers may just use the

13manufacturer paid for promotion to sell rival products,

14that they are being compensated with this valuable

15dealership, and on the margin, they are just going to

16switch, and the profit incentive is really the same as

17one, but you do not have to find these manufacturer

18assets there.

19And the third one is the dealers may just

20under-supply the manufacturer's paid-for promotion, as I

21said, because on the margin, they are getting paid on

22all these inframarginal sales, and on the margin, it

23really does not pay for them to spend all this money on

24pushing the products on the margin if it was not for

25this contract.

155

1Dealer free riding need not involve manufacturer

2investments or dealer switching. That is the

3implication of this. So, for example, in free riding

4one, which is the one you all know about, that one

5involves manufacturer investments and dealer switching.

6That is what the Court in Dentsply said, there is no

7free rider problem here. But free riding, too, the

8dealers are just using the paid promotion to sell the

9rival products, and that one can occur without any

10manufacturer investments whatsoever. They are just free

11riding on the compensation arrangement.

12Free riding number three, where dealers are

13undersupplying what the manufacturers are paying for,

14that one occurs without any manufacturer investments or

15without any dealer switching, okay, and exclusive

16dealing may be used to mitigate all thee forms of free

17riding, and it prevents free riding types one and two by

18just preventing the switching of sales to rival

19products, and it prevents free riding number three by

20creating this undivided dealer loyalty by promoting the

21incentive of the dealers to promote the manufacturer's

22product more intensively that aligns the incentives.

23So, how does exclusive dealing, that third type,

24how does the exclusive dealing increase the dealer's

25incentive to promote? And remember, we are operating in

156

1the context, you know, why did the Dentsply Court reject

2this as making absolutely no economic sense? And that

3is because there is all this competition between

4dealers, and that is all that is necessary to get the

5services provided unless there is a Sylvania type

6problem, and the example that we go through in the paper

7is this.

8Consider this case where a customer is thinking

9about buying a car and is leaning towards the purchase

10of a Honda, and he goes into a Toyota dealership to

11check out the Toyota, but really, you know, it is -- but

12just to make sure, let me just check out the Toyota.

13So, that is the hypothetical example.

14Then I have this -- look at that. So -- and

15under that -- there is a Honda and there is a Toyota,

16and Mdh is the profit margin that the dealer earns if it

17sells a Honda, and Mdt is the profit margin for the

18dealer if it sells a Toyota, and the Toyota dealer is

19deciding, what about this, even though they are leaning

20towards the Honda?

21Well, a nonexclusive dealer will not make its

22best efforts to sell the Toyota if it has both cars

23there, and basically -- now, do not get scared -- but

24the dealer is going to choose a level of Toyota

25promotional service as S that maximizes its

157

1profitability. So, it chooses S, that maximizes the

2profitability, which is the difference between the

3margin on the Toyota minus the margin on the Honda,

4times the probability that they will make the Toyota

5sale if he starts telling them how great the Toyota is,

6whether they will buy the Toyota, and that probability,

7p(S) is a positive function of how much S the person

8chooses, minus the cost of supplying S, and obviously

9there is a positive marginal cost. The more S, the

10higher those costs.

11And since it costs the dealer less to sell the

12Honda in this nonexclusive context, the dealer can earn

13a higher net profit margin on selling the Honda, the

14dealer goes to the one where the marginal cost of

15providing additional services is equal to the

16probability -- increased probability, as you supply --

17of making the Toyota sale as you're supplying more

18services, times the difference in the margin between the

19Toyota and the Honda, and you can assume that they could

20sell the Honda -- the customer comes in and sells -- and

21wants to buy the Honda, and the salesperson can at a

22zero cost sell the Honda and say, you know, you are

23right, Honda is much better. Come here, I will get you

24a good price.

25So, under these circumstances -- for example, if

158

1the dealer's margin on the two cars were the same, so

2that Mdt and Mdh were the same number, that difference

3would be zero, and clearly the dealer would supply

4absolutely no services in trying to sell the Toyota. It

5would be cheaper for the dealer to just write up the

6sale for the Honda. But by selling the Honda rather

7than promoting the Toyota, the dealer is free riding.

8He is engaging in that third type of free riding that we

9were talking about. The dealer is not switching. The

10dealer is not actively promoting the rival Honda brand

11as an alternative to Toyota, you know, for customers who

12come in and want Toyota, as occurs in free riding one

13and two. Instead, the dealer is violating the implicit

14dealer contract for the Toyota by failing to actively

15promote the Toyota automobiles.

16Alternatively, if it was an exclusive, you know

17that undivided loyalty is going to lead dealers to

18expand their promotional efforts, and it is just going

19to go to the point where the marginal costs of

20additional efforts in pushing the Toyota is exactly

21equal to how much it can make on the Toyota, times the

22increased probability that the promotion makes it more

23likely that they will make the sale. So, undivided

24loyalty is clearly in that case going to lead to that,

25and that is what you sometimes see courts saying, you

159

1know, that if you do not make the -- you know, what

2happens if you are an exclusive Toyota, basically it

3means if you do not sell the Toyota, you do not make any

4sale, and so it is common sense -- and, you know, this

5is the business people who understand this -- it is

6common sense that undivided loyalty is going to give you

7an incentive to promote more, and in the paper, it is a

8function of -- you still do not get to the point where

9the dealer has the right incentive in terms of

10maximizing the total profit of the manufacturer and

11dealer together. That is the last thing on the

12left-hand side. So, they still have to have these --

13the manufacturer still has to have these implicit

14self-enforced contracting and -- to go all the way to

15the end, but basically the role of exclusive dealing is

16that it aligns the incentives that are here.

17So, I am done. The lessons, other than that I

18put too much down here, okay? Lesson one, the Court's

19rejection of Dentsply's procompetitive rationale is an

20example of a common error that I think occurs in cases

21of trying to fit the facts of a case into a preconceived

22economic model rather than developing a model to fit the

23facts of the case, and the preconceived theory, economic

24theory here, that the Court adopted was basically, you

25know, interdealer competition will lead dealers to

160

1supply the type and quantity of promotional services,

2unless you had that Sylvania type free riding problem,

3and -- you know, because there are more likely to be

4valid procompetitive justifications for exclusive

5dealing, one of the implications I think is that this no

6economic sense test is less likely to be a useful test

7for antitrust liability, that there may be efficiency

8justifications for exclusive -- people talk about the

9Dentsply case as an easy case because there is nothing

10on one side of the scale. There is obviously something

11on one side of the scale is what I am trying to say, but

12clearly, even though there is an efficiency

13justification, you may have anticompetitive effects.

14I think that the facts of that case, there were

15significant anticompetitive effects, and Jonathan

16Jacobson makes this point in his excellent latest

17article in the Antitrust Law Journal. What he doesn't

18do is he does not answer the Court's finding that there

19was absolutely no economic basis for Dentsply's

20undivided loyalty and free riding justification. So, in

21that case, you would not get the wrong answer if you

22used the no economic sense test, but the only reason you

23do not get the wrong answer is because you do not really

24understand the procompetitive justifications.

25So, as I said in the beginning, I think the

161

1greater danger is not that -- you know, the way some

2people are advocating this no economic sense test as a

3necessary condition for antitrust liability. I think

4the danger is that the courts are going to use a no

5economic sense test as a sufficient condition for

6antitrust liability when a large firm uses exclusive

7dealing, and it is not only that I am giving you that

8there are other valid procompetitive rationales, but I

9think as economists and as regulators we have to be more

10humble that just because we have not figured this out

11yet, there is lots of other procompetitive efficiency

12justifications, and we cannot assume that the purpose of

13a restraint is anticompetitive.

14How much did I violate the contract by?

15MR. O'BRIEN: Ten minutes.

16(Applause.)

17MR. O'BRIEN: Okay, we are going to break until

18about five past 3:00, okay?

19(A brief recess was taken.)

20MR. O'BRIEN: Okay, let's get started.

21Okay, I would like to start out our sort of

22post-speech session here by asking folks if they would

23like to comment on the remarks of others on the panel,

24and I guess I will ask for a volunteer first rather than

25being systematic about it. We will find out who has the

162

1most burning comments to make about what someone else

2said.

3Okay, Joe.

4DR. FARRELL: Well, I have one question for a

5fellow panelist, which is relatively specific, I think.

6Ben, in your model, you didn't have time to

7present all of it, but I would like to ask, have you

8offline, as it were, closed the loop and shown actual

9harm to buyer, or is it just that the buyer who was

10leaning towards buying a Honda ended up buying a Honda

11and, of course, the Honda -- Honda likes that, the

12dealer apparently likes that, the customer seems to like

13that, although the welfare economics of this promotion

14stuff, of course, are a little subtle. Toyota, of

15course, does not like it.

16Where does this go and how does the whole thing

17play out with and without exclusive dealing as opposed

18to just Toyota would like S to be higher in the short

19run?

20DR. KLEIN: All we do in the paper is present

21the procompetitive efficiency justification. We do not

22do the other side of the scale in terms of is there any

23anticompetitive effect. In some cases, there will and

24in some cases there will not be an anticompetitive

25effect, and, you know, and as I suggested in Dentsply,

163

1even though there may have been a legitimate

2procompetitive rationale, forget undivided dealer

3loyalty in that case, that does not mean that

4arrangement was not, on net, anticompetitive and harmful

5to consumers ultimately by creating a barrier to entry

6to competitors.

7But there are so many cases out there where we

8know -- I mean, the case I love is this Joyce Beverages

9case that I am certain Tad knows about, where you

10have -- you have RC Cola having their distributor only

11have one cola, the RC Cola, because they want undivided

12loyalty. Well, in that case, RC Cola has 5 percent of

13the cola market and a lot smaller share if you define

14the market more broadly to have all carbonated drinks.

15So, in that case, we clearly know there is no

16anticompetitive effect.

17But basically there was this mystery in the

18literature, why are they really having an exclusive

19dealing arrangement there, because there does not seem

20to be any specific investments, and there does not seem

21to be this dealer switching, but what they want is when

22the salesman goes into the supermarket, that they are

23going to push RC Cola and not any other brand.

24So, if you want to get the ultimate question,

25that would depend upon the particular case, and you

164

1would have to examine that particular case.

2DR. FARRELL: But you say you only do the

3procompetitive justification. What do you demand of it

4in order to call it a procompetitive justification, just

5that Toyota would like it?

6DR. KLEIN: Well, look, I do not want to get

7caught up in a language thing about --

8DR. FARRELL: Okay, sorry, no.

9DR. KLEIN: -- you know, we will do linguistic

10philosophy later. My feeling -- all I mean by it is

11that somebody doing what you would consider the right

12thing or the good thing or something and balancing it, I

13am just looking -- I am just presenting an economic

14foundation for this legitimate procompetitive

15justification.

16I mean, the crazy thing is if you look in the

17marketing literature, people are talking about this all

18the time. It is just economists, you know, a little bit

19of economics can be a very dangerous thing, and it is

20only the economists that say competition should give you

21the services, everything is fine. So, if you talk to

22business people, marketing people, they all know that

23this makes -- and it makes a lot of common sense.

24So, in some sense, as I said, Dentsply was

25unlucky enough to have the judge that knew economics,

165

1and that is the only reason they got into problems in

2terms of the procompetitive justification, plus they

3were unfortunate enough to choose an expert that

4explicitly wrote in his article that the argument makes

5absolutely no sense. So, he could not present -- he did

6not -- Howard did not present the argument at trial, but

7the company did in terms of answers to interrogatories,

8and they said, what are you talking about? Your own

9expert says this makes no economic sense.

10And then the other interesting thing about it,

11and this is the connection between anticompetitive and

12procompetitive justifications is strange, because the

13Justice Department -- and Gail would know this -- the

14Justice Department, in trying to demonstrate the

15anticompetitive effect, spent all this time in their

16findings of fact to show how important this dealer

17channel was to promoting the Dentsply products and how

18rivals would be at a competitive disadvantage because

19they did not have access to that channel.

20So, you just look at all the findings of fact,

21and it not only demonstrates that there was a

22significant potential anticompetitive effect, but it

23also demonstrates that there is a significant

24procompetitive justification for motivating the dealers

25to do a good job. So, you have that tension, but

166

1basically -- well, maybe I should not monopolize this

2thing.

3MR. O'BRIEN: I mean, I have a follow-up for Ben

4on this point. I mean, in talking about Dentsply and

5just more generally your theory, there were two types of

6free riding beyond the traditional one that you cited.

7Dealers can free ride, effectively, on their own

8promotion on behalf of a manufacturer, right, which

9maybe they are doing in conjunction with the

10manufacturer or somehow they have arranged a contract to

11get that done, and the other one was that dealers may

12violate this implicit contract by just under-investing

13rather than free riding by steering customers to a

14rival.

15I am wondering if you have any specific evidence

16that you can cite from the Dentsply case, if your

17knowledge of the cases is deep enough, that one of these

18two types of free riding that you identify in your paper

19was actually present.

20DR. KLEIN: Well, I do not think -- well, I

21think the Marvel type free riding was not present, and

22he did try to present an argument, and I think the facts

23made it very difficult for him, and it is too bad that

24he is not on the panel, because he would disagree, so I

25think the first free riding did not exist, and I think

167

1the second free riding did not exist basically because

2there was no switching to rival brands.

3I mean, I think there was one example where

4there was some disagreement about whether they tried to

5switch it to someone else. So, I do not think those

6other two were there, but in terms of the third one, all

7the evidence you need for that is that promotion is

8important for the manufacturer to make sales, and as I

9said, the Justice Department went to great lengths in

10terms of trying to demonstrate the anticompetitive

11effect to demonstrate the existence of that, so that is

12there, and then all you need in addition to that is that

13incremental sales are profitable for the manufacturer,

14and those two conditions were clearly met in Dentsply.

15DR. CALKINS: The great thing about the Howard

16Marvel theory is that it is one that we lay people can

17understand, namely, that it is good for everybody for a

18manufacturer to run ads saying you can have your hearing

19improved by getting it tested and going to a dealer and

20getting this new improved kind of technology to use for

21your hearing aid, and that drives consumers to go to the

22dealership to try it out, and that is good for consumers

23because they are finding out information.

24It is good for the overall industry because

25total sales of hearing aids will go up, because all

168

1these consumers are being driven into the dealership and

2are getting their ears tested, and it is all sorts of

3wonderful stuff. And then, if, when the consumer gets

4there, there is the old bait and switch and they are

5sent off to buy the el cheapo discount brand, well, the

6bad consequence of that is there will be less of that

7advertising about the new, improved technology and, you

8know, science of hearing aids and such, which is

9something that is good for the whole industry, good for

10consumers, good for everybody, it will now be lost,

11because the manufacturer will not spend money on that.

12So, you can easily tell a simple layperson

13story, if you let everybody get switched off, you will

14no longer have those ads being run, and I look forward

15to reading the article, but merely saying that -- I

16mean, if you then say that if you have exclusive

17dealing, it is good for RC Cola because they are going

18to make more money and have more sales, well, I can have

19a warm feeling about that just because they have got a 5

20percent share, and God knows, if you are RC Cola, you

21are going up against people that you need all the help

22you can get to go up against them. So, we can see

23there.

24You know, if you were to tell the RC Cola story

25where you had somebody that had an 80 percent share and

169

1climbing, before I then sat back and said, boy, I am

2really concerned about maybe intervening and causing

3harm here, I would like to at least make sure I

4understood what is the equivalent of the lost nice

5advertising that is going to happen if you intervene in

6that type of situation.

7DR. KLEIN: Well, Steve -- can I answer it,

8then? I mean, I agree with you on your main point,

9Steve, that, you know, with RC Cola, that we can be

10pretty much assured that inter-manufacturer cooperation

11or competition is going to pass on these benefits to

12consumers, but if you are talking about -- the analogy

13is really identical about lost advertising, because it

14is either lost advertising by the manufacturer or lost

15advertising by the dealers. It is just in some cases,

16it is efficient for the manufacturer to do the

17promotion, and in some cases, it is efficient for the

18dealer to do the promotion, and if you do not have the

19exclusive in one case, you do not get the manufacturer

20advertising, and in this case, you do not get the dealer

21pushing the product, and if you think that there is a

22benefit from lost advertising, then it is totally

23analogous in the two cases.

24You know, if you start to do a calculation --

25and you -- even in the standard case with the

170

1manufacturer, you know, when the consumers are switched

2to the discount brand, they almost always pay a lower

3price. It is not -- you know, they are not being

4deceived, that they think they are getting the higher --

5DR. CALKINS: I understand. I think my problem,

6and I will confess, I was sitting here with my back to

7the screen, but I understood the Marvel advertising of

8hearing, you know, development is a good thing. When

9you hold out as your public good having a used car

10salesman sit there and harass you into switching from

11this model to that model, and I being a kid from

12Detroit, noticing that you are using entirely Japanese

13brand models, I...

14MR. O'BRIEN: At the risk of allowing Ben more

15monopolization time here, I just want to push it just a

16little bit further on the Dentsply, and if others do not

17have a question right off the bat here, and you said

18that this third type of free riding, which is that you

19would under-invest as a retailer --

20DR. KLEIN: Right.

21MR. O'BRIEN: -- you think was present, but the

22evidence that you cited for that was that investment was

23important, and that does not seem to demonstrate to me

24that they actually would have necessarily under-invested

25but for the exclusive dealing arrangement. You know,

171

1was there another way for them to ensure that

2investment? I mean, it seems to be a bit of a leap to

3me.

4DR. KLEIN: Well, that's --

5MR. O'BRIEN: You know, and that sort of

6statement strikes me as, you know, it might not be hard

7to argue that that efficiency is there in almost every

8case.

9DR. KLEIN: No, that is a problem that you have

10with these cases. That is why I said you cannot

11adopt --

12MR. O'BRIEN: A no economic sense test?

13DR. KLEIN: Yes, you know, because I think the

14efficiency is -- I would not say universally present,

15but it is a motivation. I forgot what your other

16question was, but, you know, it is important -- if it is

17important to the manufacturer, we just know from the

18economics that if there is an exclusive, the incentives

19are going to be aligned, and if they do not make the

20sale of that product, they are going to not make any

21profit. So, you do know that they are going to push it

22more.

23So, I mean, it just follows logically, but you

24would need to see what they adopted -- oh, yeah, so

25you -- there may be a less restrictive way, and then we

172

1can talk about a less restrictive standard here if that

2is the question you want to move to, but in cases where

3it looks like the practice might have also some

4foreclosure problems and anticompetitive effects -- I

5hope I am using the right language -- you may impose

6this burden on the manufacturer to come up with a less

7restrictive way of doing it, and, you know, maybe they

8chose this not just because of the efficiency effects

9but also because of the -- it increased their market

10power, so...

11MR. O'BRIEN: Okay.

12DR. KLEIN: I mean, it makes it very, very

13difficult in terms of this balancing. The important

14point is, you know, you are not going to have these easy

15cases anymore where there is nothing on -- I mean, you

16will still have easy cases where you do not have the

17anticompetitive effect on one side of the scale, but you

18are not going to have these cases, I think, if you

19accept this where there is nothing on the other side.

20MR. O'BRIEN: Okay. Do any other panelists have

21any questions for any of the other panelists?

22(No response.)

23MR. O'BRIEN: If not, we have a series of

24propositions we would like to walk through to see where

25we might be able to reach consensus, where we have open

173

1issues, and hopefully this will spawn some additional

2conversation about both what was said during the session

3and perhaps some new things.

4So, let me start with -- where is our -- oh, you

5have got it, okay. Let's go to the first proposition.

6Okay, I am going to read it, because we need to read it

7for the transcript here.

8Exclusive-dealing arrangements are analyzed

9under the rule of reason.

10First, does everybody agree -- and this is

11really more for the lawyers -- that that is the way the

12analysis of exclusive dealing goes today?

13DR. CALKINS: Yeah. I mean, that -- yes -- yes,

14I'll agree to say that, and B, for a whole lot of the

15cases, it is consistent with the general idea that under

16the rule of reason, the defendant always wins.

17MR. O'BRIEN: Okay. So, nobody disagrees with

18that point. Well, perhaps the point that was just made,

19but nobody disagrees with the proposition, correct?

20Does anybody think that there are exclusivity

21arrangements that should be per se illegal?

22(No response.)

23MR. O'BRIEN: No, I guess that is the answer.

24DR. KLEIN: Move on.

25MR. O'BRIEN: No.

174

1Does anybody think there are exclusivity

2arrangements that are always or nearly always

3procompetitive and thus good candidates for a safe

4harbor?

5DR. CALKINS: Well, presumably a very small

6exclusive would be -- would fit anybody's idea of a safe

7harbor.

8MR. O'BRIEN: And when you say "small

9exclusive," you mean a small percentage of the market

10 or --

11DR. CALKINS: Yeah, it is very -- it is hard to

12imagine a court or an enforcer being concerned about an

13exclusive below -- choose your figures. Some might

14choose 20 percent, some might choose 30 percent, some

15might choose 40 percent, but I think everybody would

16agree that below some percent, no agency should worry

17about it, and no court should find illegality unless,

18you know, you have some reason to think that that number

19is just, you know, totally misleading and the real

20number will be totally different in six months when the

21contracts kick in or something.

22MR. O'BRIEN: Okay, fair enough.

23Anybody else? That one was --

24DR. KLEIN: I would like to -- I would like to

25ask Steve a question on this one. You know, your

175

1opinion about the foreclosure standard somehow being

2lower when it comes to Section 2 rather than Section 1,

3I mean, if somebody is a monopolist or is likely to be a

4monopolist, I could see that it is more likely that they

5are going to meet the critical share number, but why

6should that critical number, whether you say it is 40 or

7whatever, why should somehow it be lower? It sounded

8like that is what you said from your presentation,

9should be a tougher standard.

10DR. CALKINS: If I did, I misspoke slightly.

11What I meant to say is that -- well, specifically, is

12that in the Microsoft case, the defendants argued that

13because this practice is lawful under Section 1, it

14must, as a necessity, be lawful under Section 2, and I

15was just saying that I do not think that is correct,

16that, you know, take your extreme of a dominant firm

17that everybody would agree is a monopolist on the one

18hand, and on the other hand, your RC Cola kind of a

19thing. I am not saying whether or not, you know,

20exactly where one would say there is a difference, but I

21would think that one should be much more likely to be

22concerned about something being done by a dominant firm

23that is --

24DR. KLEIN: Right, obviously, but why should

25there be a different standard under Section 2 than under

176

1Section 1? I mean, I think we are in trouble here

2basically because Justice did not appeal the Section 1

3no liability in Microsoft and Dentsply, and if you read

4the Court, you know that the appeals court would have

5overturned both of those things, but -- you know, and

6then I think we would be in a lot better shape, but the

7idea that somehow we should have a different standard

8and principle when you are doing the first step of a

9Section 2 -- I agree, if somebody is a dominant firm,

10they are much more likely to have anticompetitive

11foreclosure under Section 1 and under Section 2, but why

12should there be a lower hurdle showing the

13anticompetitive effect under Section 2?

14DR. CALKINS: Well, part of this goes -- I mean,

15in all of this, it is trying to make a judgment about

16how likely a particular practice is to be harmful to

17competition, and I was just saying that -- well,

18specifically, is that there are a whole series of sort

19of ways that firms with fairly modest market shares have

20been able to persuade courts to get rid of exclusive

21dealing cases, but where you have a dominant firm, I am

22not saying that there is a magic difference. I am just

23saying that, as you recognized, you would think longer

24and harder about something being done by a dominant firm

25that is a clear monopoly than by some firm that is a

177

1trivial firm, and so just because you are told that

2something would be lawful under -- you find some Section

31 case out there where some foreclosure level was a

4motion for summary judgment for a defendant, that does

5not mean that in every case with the most extreme

6monopolist you would grant summary judgment without

7thinking long and hard about it.

8MR. O'BRIEN: Okay, let's move on to the next

9proposition.

10Okay, I think this one will be easy, too. The

11proposition is from Posner's Antitrust Law.

12I propose the following standard for judging

13practices claimed to be exclusionary: "In every case in

14which such a practice is alleged, the plaintiff must

15prove first that the defendant has monopoly power. All

16the plausible cases of exclusionary practices involve

17defendants that have monopoly power."

18First, does everybody agree with that?

19MR. LIPSKY: Uh-oh.

20MR. O'BRIEN: Can exclusive dealing involving a

21non-monopolist result in a substantial lessening of

22competition?

23DR. KLEIN: Yes.

24DR. FARRELL: All statements containing the word

25"all" are false except for this one and perhaps a

178

1handful of others. I think there is a real problem with

2a subtle, complex and imperfectly understood topic

3having courts, judges, make grand and sweeping

4pronouncements. The law, as I understand it, in a

5precedent-based system tries hard not to change over

6time, and our understanding tends to change over time,

7and that creates a lot of trouble. So, it is not like I

8am out here saying, oh, and the following large category

9of cases, firms without monopoly power or without market

10power or something, can do a lot of harm with exclusive

11dealing. There have been some theories developed under

12which that can happen.

13I think the consensus currently is that that is

14not such a big worry, but we do not really know yet, and

15freezing stuff in place by grand pronouncements that say

16"all," I am not sure it is such a great idea.

17DR. CALKINS: The larger consequence, if that is

18the law, is that any time a -- well, any time a

19plaintiff has failed to hire one of these fancy

20economists and satisfactorily define a market in which

21the defendant has a well-defined market share of more

22than 75 or 80 percent, there is a very good chance that

23a Court would grant a motion for summary judgment or a

24motion to dismiss, because when you have rules like

25that, lots of courts operationalize it by saying, okay,

179

1any market share below 70 percent, I grant a motion for

2summary judgment and do not explore anything else about

3what is going on, and that in my judgment is too

4sweeping a broom to use. That was a bad way to phrase

5that, wasn't it?

6MR. O'BRIEN: Okay, Tad?

7MR. LIPSKY: I think I can agree with the last

8sentence there, that all the plausible cases -- I am a

9little confused, though, whether this statement in

10context, was it limited to exclusive dealing or is it

11meant to be applied more broadly to other types of

12exclusionary practices? I guess that there -- you know,

13I am trying to recall. Wasn't there a -- there were

14some Commission consent decrees in cases involving water

15pumps for fire trucks. It was a multiple defendant

16situation where there was actually a fairly plausible

17theory of cartelizing, and I do not think you could have

18found, at least not with any logical consistency, that

19both of the competitors were monopolists.

20So, I guess that is a limiting case, but I would

21be closer to agreeing with this if you were talking

22about cases other than those in which a cartelizing

23theory for challenging the exclusive dealing was the

24theory of liability.

25Am I right about this FTC decree? Does anybody

180

1remember that?

2DR. CALKINS: There was a pump case. There

3was -- there was a case like that.

4MR. LIPSKY: Okay, so it is actually -- it is

5probably real, presumptively real.

6MR. VITA: It is called Waters Hale (ph).

7MR. LIPSKY: Excellent, okay, thank you.

8DR. KLEIN: If Posner had restated it in terms

9of market power instead of monopoly power --

10MR. LIPSKY: That would be fine.

11DR. KLEIN: -- I assume we could all agree,

12right?

13MR. LIPSKY: Yes, that would be fine.

14MR. O'BRIEN: So, this statement is about

15monopoly power or market power on the part of the

16defendant. I am wondering if any of you think that

17conditions relating to market power or market structure

18in the downstream market have an effect on the extent to

19which exclusive dealing can be anticompetitive. That

20was not stated well, but what should we make of the

21downstream market structure in terms of the likelihood

22that exclusive dealing can have an anticompetitive

23effect?

24DR. FARRELL: Well, I mean, I talked briefly

25earlier about the developing economics of understanding

181

1the role of downstream competition in that and, you

2know, fairly plausible seeming analyses have come out

3with very different answers so far, so watch this space,

4and that perhaps should be a pretty strong warning

5against making strong statements at this point.

6MR. O'BRIEN: Would you be willing to say that

7some kind of barrier to entry in the downstream market

8is necessary for anticompetitive exclusive dealing?

9DR. FARRELL: Well, I think -- see, you are

10talking about a lot of abstract nouns here, and I am

11sorry, I cannot put on a southern U.S. accent, but I

12would like to.

13DR. WERTHER: Can you do any U.S. accent?

14MR. LIPSKY: I thought that was a Berkeley

15accent.

16DR. CALKINS: You have got such a lovely accent.

17DR. FARRELL: I think Strunken White might have

18said if you are getting confused, try to decrease the

19abstract nouns and increase the active verbs, and I

20think that is a pretty good proscription for thinking

21straight. So, let's try that.

22Instead of talking about market power and market

23share and dominance and exclusive dealing and so on,

24let's ask the following question: If I come up with a

25better way of doing things than the incumbent is doing

182

1or I am less greedy than the incumbent and I am willing

2to give consumers a better deal, am I stymied in my

3attempt to do so by these deals that people have struck?

4That is the core question, and a lot of the time, the

5answer will be no, I am not stymied if there are small

6shares of this or that. Sometimes I will be.

7So, for example, if you look at the Microsoft

8case, Microsoft had no need to completely keep NetScape

9out and wasn't trying to keep NetScape out and charge a

10lot of money for Internet Explorer. They just had to

11make sure that NetScape did not become sufficiently

12widely distributed that people would start writing to it

13and say, yeah, I -- that is a rather different case from

14the one we would generically think of. You have to be

15careful and I think should be pretty reluctant to kind

16of lay down these firm rules.

17Now, having said that, I also am very aware

18that, you know, attempts to do full-blown rule of reason

19analysis are also dangerous, right, given the subtlety

20of what is going on and given the capabilities and

21noncapabilities of courts.

22I am a big fan in theory of -- I have never been

23up close when it has happened -- of court-appointed

24experts. I think that could probably improve the

25process a lot quite generally, but especially when you

183

1are dealing with subtle and difficult issues.

2DR. CALKINS: Clearly everybody would say that

3it matters how easy our new entrant can gain access to

4the customers to whom it is trying to sell, and if it is

5very easy to do that, then exclusive dealing will not

6present any problems. As you phrased the question, you

7used the magic word "entry barriers," and as you know,

8that has lots of different definitions, and choose your

9right definition and defendants will almost always

10prevail; choose different definitions, and they might

11not.

12It also raises the question as to whether you

13are looking at a total exclusion standard or at simply

14making it much more expensive, time-consuming and risky

15in order to gain access, and so you have staked out a

16position or the quote here has staked out a position

17which might mean things that I would not be comfortable

18with.

19MR. O'BRIEN: Right. So, just one follow-up to

20that, I guess this is directed to Joe, the Fumagalli and

21Moto models and the Simpson and Wickelgren model, in the

22simplest cases, you have homogeneous producers

23downstream with no economies of scale or very small

24economies of scale, and it strikes me in the context of

25those models that it would be very easy for a firm to

184

1enter at both levels and disentangle any anticompetitive

2effect that is being contemplated. I am wondering if

3you have thought about that, or maybe I am wrong.

4DR. FARRELL: You know, it has been a while

5since I read the models, so I do not remember

6technically whether what you say is right. Clearly if

7you really have homogeneous products and fixed costs and

8sunk costs are very small, then you would think -- and

9you would want to know why not if somebody was claiming

10not -- that a firm could enter at both levels.

11On the other hand, there certainly are

12industries where at any given time the industry may

13behave quite competitively involving the pass-through

14dynamics that we were talking about, and yet there are

15big sunk costs lying behind it, and that may be the more

16relevant case for that kind of analysis.

17MR. O'BRIEN: Anyone else? Okay, next slide.

18Okay, I think this is an uncontroversial slide

19as well. We will see. Maybe the questions will be more

20interesting.

21"Exclusive-dealing arrangements --" this is a

22quote from Jefferson Parish. "Exclusive-dealing

23arrangements 'may be substantially procompetitive by

24ensuring stable markets and encouraging long-term

25mutually advantageous business relationships.'"

185

1Yes, Joe?

2DR. FARRELL: I hate to be a curmudgeon, but

3stable markets are not exactly what antitrust aims for.

4Actually, maybe we should try to encourage unstable

5markets where the status quo could be disrupted at any

6moment by some pesky firm that maybe has not shown up

7before, or maybe has, and is willing to take a lower

8margin or has a better way of doing things.

9Now, I am not saying that the basic point here,

10that exclusive dealing arrangements "may be good" is

11wrong, but I do not like that language.

12MR. O'BRIEN: Okay. Well, you pick the --

13DR. CALKINS: And while you are complaining, you

14could complain about the mutually advantageous business

15relationship, because that could be good for consumers,

16and if it is just dividing up a surplus between two

17businesses, it could be bad for consumers.

18DR. KLEIN: Yeah, I --

19MR. O'BRIEN: Ben Klein, do you have a view on

20that?

21DR. KLEIN: Well, who knows what Justice

22O'Connor is referring to, but if she means by

23encouraging long-term mutually advantageous business

24that it encourages people to make specific investments

25in the relationship, relationship-specific investments,

186

1then I think she is correct and that she should not go

2through it now, but that is one of the problems I had

3with Joe's presentation, is that the Segal and Whinston

4criticism of that rationale for exclusive dealing is

5just wrong, and it is logically correct, but there is

6assumptions being made in that that are very, very

7unrealistic, and in particular, they are just -- well, I

8better not go into it.

9But, you know, so if that is what she is saying,

10I would agree with her very much, but it is so vague,

11right, but if she is just saying there is -- it

12sometimes may be good...

13MR. O'BRIEN: What efficiencies, which I assume

14are the second object of this sentence, there are

15numerous efficiencies that have been discussed about

16exclusive dealing that we might classify into that

17second phrase. What are the most significant and most

18likely in an exclusive dealing arrangement?

19And similarly, what efficiencies have been

20asserted most often do you think are least likely to

21actually exist?

22DR. CALKINS: Oh, the best is the classic Marvel

23free riding, manufacturers spending money, bringing in

24the customer, then there's the old bait and switch to

25the other product. That would be the classic and the

187

1best.

2MR. O'BRIEN: So, what efficiencies are often

3asserted in exclusive dealing cases that you think may

4not actually exist very often? Anybody?

5DR. KLEIN: I hope nobody says this focused

6dealer effort, but I guess one of the things I should

7say is the justification that Microsoft offered, the

8procompetitive justification for the exclusive dealing

9arrangement with the Internet access providers, sounded

10like a focus -- the way you presented it, it sounded

11like a focused dealer incentive, but what they wanted

12was -- the argument they presented was something to the

13effect that they wanted the developers to focus on the

14Windows APIs, which meant they wanted to have a monopoly

15in Windows so that when developers were developing their

16programs, they would only develop Windows programs,

17which is a very different argument than, you know, you

18want -- they did not want the Internet access providers

19to promote their product. That is not what they were

20doing.

21They were talking about a different type of

22focus there, but that argument I think the Court

23correctly rejected as making no sense other than you

24want a monopoly. You want to maintain your monopoly.

25MR. O'BRIEN: What significance, if any, should

188

1be given to observing an exclusive dealing arrangement

2in a similar competitive market when you are analyzing a

3case where there is exclusive dealing, maybe in a market

4that exhibits some more market power in some ways than

5the other, but otherwise has similarities?

6DR. FARRELL: Well, at a technical level, there

7certainly have been analyses that show that in some

8circumstances, exclusive dealing engaged in by, let's

9say, all members of an oligopolistic manufacturing

10sector, whether downstream industry, can soften

11competition and be in that sense anticompetitive, even

12conditional on, you know, a flourishing oligopoly

13structure, and let's face facts, we are never dealing

14with perfectly competitive industries when we are

15talking about these cases, so oligopoly is what you mean

16by the word "competitive" here.

17There are other analyses that suggest that

18exclusive dealing can actually sharpen competition. I

19think it is fair to say that that literature is both

20unsettled and in a state of nonferment, the nonferment

21because nobody seems very excited about it. People are

22really more interested in the monopoly-preserving

23possibilities I think than the oligopoly-softening

24possibilities, and that may be a legitimate choice of

25emphasis, where to put our intellectual resources, or it

189

1may just be, you know, what happens to be fun for

2assistant professors to do these days.

3DR. KLEIN: I think we have to be very careful

4when we start talking about oligopoly-softening, and I

5guess Joe would say I have this bias, this laissez-faire

6bias, but I can imagine unilateral behavior -- you know,

7a gasoline company decides they are going to locate

8their station not next to another station but a couple

9of blocks away, because if they locate it next to the

10station, it is going to be more intensive competition.

11People are going to be able to compare the prices.

12We do not want to go in and micro-regulate the

13competitive process. You know, you hire an economist,

14and let's assume they draw the welfare triangles, and

15they say consumers are better off if that person puts

16the station next to the other station, and even though

17it has -- let's assume it has the effect of sharpening

18competition if we do that, we do not want to regulate

19that behavior, at least I do not want to, even though

20the calculation would come out that way.

21So, I think it is dangerous to start talking

22about oligopoly-softening of competition in general, and

23basically I guess I have a prior that we are just going

24to mess things up and we should just leave it up to the

25competitive process, unless there is a -- you know, you

190

1have this first step where you need some major

2anticompetitive effect in terms of foreclosure.

3So, I guess my comment was not totally

4irrelevant, because we are talking about Section 2

5unilateral behavior, even though it has nothing to do

6with exclusive dealing.

7DR. CALKINS: Trying to psycho-analyze your

8question, I think you were -- I am guessing that you

9were referring to the argument you sometimes see made

10that, look, over here in this market, which we all

11stipulate is competitive, this practice is occurring,

12and so, therefore, it must follow as the night follows

13the day that when that same practice is being engaged in

14by this complete and total monopolist, it deserves

15summary judgment very promptly on that ground alone,

16and --

17MR. O'BRIEN: That is a good psycho-analysis.

18Yes, that is what I was hoping somebody would address.

19DR. CALKINS: And I myself do not buy into that

20theory in the little that I have done thinking about it,

21but my thinking is still at a preliminary stage.

22MR. VITA: Well, it is not so much -- maybe,

23Dan, a competitive market versus a noncompetitive

24market, but the individual -- the size of the firm or

25the mark -- the firm's specific market power. Like the

191

1RC Cola example somebody alluded to before, RC has some

2exclusive relationship with its bottlers or something, I

3think it was, and you look at RC Cola, and they are a

4small fry. I mean, they do not matter anywhere. So,

5you look at that and you say, well, obviously they are

6doing that. They cannot possibly have any kind of

7foreclosure mode or some monopolization motive. It has

8to be some sort of value creation that induces them to

9do that.

10Is it fair to say that when you do -- then you

11look at Coke, for example, maybe doing the same kind of

12thing, some other firm with substantial market share or

13market power possibly? At least it says you have got to

14consider the efficiency story. You can't rule it out.

15There is a possibility that there is value creation,

16that there is something inefficient about it, but not

17necessarily -- the fact that RC does it doesn't

18vindicate Coke's usage, that debate is not over, but

19that does say to you -- you know, we have got to take

20that seriously.

21DR. FARRELL: Yeah, I think you said it right,

22you know, unless there is something about that industry

23or market that I do not know, you can presumably infer

24from RC's use of these exclusives that there is

25something other than monopoly preservation going on, but

192

1that does not mean that there is not monopoly

2preservation going on.

3DR. KLEIN: Exactly.

4DR. FARRELL: It does not mean there is either.

5MR. LIPSKY: Thanks.

6MR. O'BRIEN: Okay, next slide.

7Okay, so anticompetitive effects, this is a --

8this is actually a quote from Dennis Carlton's paper on

9the Aspen and Kodak case.

10"In the presence of scale economies, exclusive

11dealing can be a way of depriving Firm 2 (or its

12distributors) of the necessary scale to achieve

13efficiencies, even though, absent the exclusivity, Firm

141 and Firm 2 would both be large enough to achieve

15efficiency."

16So, this is the standard scale economy argument

17about excluding your rivals so that it cannot reach

18efficient scale, and I guess my question is, does the

19panel see that as the primary anticompetitive theory of

20exclusive dealing that we ought to be focused on?

21MR. LIPSKY: Well, I will take a stab at that.

22Certainly, you know, in a static sense, it is hard to

23argue with this proposition, and I think this is

24consistent with the notion that there are stories

25associated with exclusive dealing where you are trying

193

1to compel two-stage entry basically, and I think some of

2those are good stories. Probably we would not agree on

3which ones were good stories.

4I heard John Jacobson the other day talking

5about Pullman, and I disagreed with him on that one, and

6then Motion Picture, and I disagreed with him on that

7one, but it does not -- and also United Shoe Machinery,

8and I disagree with him on that one, but I think we

9could find -- I think we could find a two-stage entry

10story that held together, and so I would say I agree

11with this.

12But I would also interject -- and I have said in

13other contexts -- there is kind of an endemic temptation

14or tendency in the system, in the investigation and the

15litigation system, to underestimate supply flexibility.

16I mean, you know, supply flexibility is not -- or new

17entry is not always an answer, and so I would hate for

18my remarks to be misconstrued. There are industries in

19which the barriers to entry are such that if you have a

20two-stage story, it is a serious problem, but I think

21there is a tendency to look at what is right in front of

22you, to, you know, fail to predict the rise of the

23Internet or the mobile phone, you know, falling in price

24by 75 percent over five years or, you know, some other

25alarming and unpredicted new technology or new

194

1development, and because the dynamic aspect is so

2important, I think this is a theme that needs to be

3hammered again and again.

4So, what I guess I am saying, yes, I agree with

5this, but it is narrow -- I would like to make my

6agreement as narrow as humanly possible.

7MR. O'BRIEN: Anybody else?

8DR. KLEIN: Tad, you sounded like an expert

9witness there.

10DR. CALKINS: I was hoping that Tad could tell

11me how to get a mobile phone bill that is 75 percent

12lower.

13MR. O'BRIEN: So, Joe, based on your remarks, I

14guess I would ask, do you think this is the primary

15story of competitive harm that we should be focused on

16in analyzing exclusive dealing, or should some of the

17other theories that you mentioned, I guess in particular

18Simpson/Wickelgren, maybe some of these two-stage models

19of oligopoly where exclusive dealing can play a role,

20are those things we should be concerned about, or is

21this number one and number two?

22DR. FARRELL: Well, I disagree with the

23question. I think the primary focus should be based on

24what is going on in the market at hand, and we should

25adjust the tools to fit the facts and not prejudge what

195

1the theory is going to be.

2Having said that, I think I said in my earlier

3remarks that I believe this Rasmussen, Ramseyer and

4Wiley or Segal/Whinston theory, which is being referred

5to here, is the one that people talk about most. I tend

6to suspect that it is the main one. I would add -- I

7mean, you have to interpret efficiencies carefully, so,

8for example, scale to fully reward innovation, is that

9achieving efficiencies?

10But broadly speaking, I think this is what most

11economists think of most of the time when they think

12about anticompetitive exclusive dealing, and I think

13that may well be right, but I think we should be open to

14whatever the facts of a particular case say.

15MR. O'BRIEN: All right. Anybody else?

16(No response.)

17MR. O'BRIEN: Let's go to 7.

18Okay, this is from the Microsoft case, and the

19quotation is:

20"If the monopolist's procompetitive

21justification stands unrebutted, then the plaintiff must

22demonstrate that the anticompetitive harm of the conduct

23outweighs the procompetitive benefit."

24I guess my question is -- well, first, does that

25make sense to you, and secondly -- this is maybe more

196

1for the economists, although equally for the lawyers --

2does economics supply tools to do this?

3 DR. KLEIN: Try Joe.

4DR. FARRELL: Well, let's see. I mean, clearly

5in order to plunge into enforcement, we would not want

6to go ahead if the anticompetitive harm of the conduct

7is outweighed by the procompetitive benefit. Using the

8term "procompetitive benefit" in -- I am not sure

9whether it is the same way or not as Ben uses it, but I

10am using it to mean actual benefits to efficiency and

11consumers, not just kind of non-anticompetitive

12rationales.

13This, of course, is part of a bigger decision

14tree that the Microsoft Court laid out. In thinking

15through a burden-shifting process like that, you have to

16think about a number of things, and I do not know how

17much the Court thought through these things. I am

18pretty sure I know how much they knew the necessary data

19required to do it exactly right, which is not a

20criticism, because nobody has that data either.

21You have to think both about whether in most

22cases this is true or that is true, but also about if

23this is true, is it going to be easy to prove, or is it

24quite likely to be true but be hard to prove? And that

25really gets back to what I hope was the main theme that

197

1came out of my talk earlier, that in my opinion, there

2are often benefits of open, free-wheeling competition

3that are very difficult to pin down and almost

4impossible to prove, and I think that needs to be kept

5in mind when we lay down these decision trees.

6Did the Microsoft Court keep that in mind? To

7some extent. Did it do it the right amount? I have no

8idea, and I doubt that they really know either.

9DR. CALKINS: If the question is should one

10think about the competitive harm that is likely, should

11one think about the procompetitive benefit, the answer

12to that is entirely yes.

13On the other hand, can you read this statement

14to say that if there is any tiny procompetitive benefit,

15perhaps using anybody's definition of "procompetitive,"

16does that mean that the defendant always wins unless the

17plaintiff is able, with great specificity, to precisely

18quantify the anticompetitive harm, precisely quantify

19the anticompetitive benefit, and then precisely

20calculate that one is more than the other?

21Well, it may well be that if that is what one

22means, then what one is saying is that any time there is

23any benefit that can be characterized as procompetitive,

24the defendant will always win, and so if that is where

25you ended up, that might not be a good place, but that

198

1does not mean that you should not think about the

2procompetitive benefit.

3DR. KLEIN: Go ahead, Tad.

4MR. LIPSKY: No, go ahead.

5DR. KLEIN: No --

6MR. O'BRIEN: Go ahead, Tad.

7MR. LIPSKY: Well, I was just going to say that

8we always have to consider the fact, you know, there was

9a day not so long ago when you could expect a follow-on

10litigation from cartel cases that were litigated and won

11by the Department of Justice. You would get a guilty

12plea in a price-fixing case, and then we transitioned --

13I am not sure exactly what the history is or how we got

14here, but then we got to the point where there was a

15story in a newspaper saying that there was a

16price-fixing investigation, boom, 80 private class

17action -- purported class action treble damage suits

18against everybody in the industry, and then we got to

19the state where there -- you get the same thing even in

20these conduct type cases, which are not cartel cases,

21and there are follow-on class actions for Dentsply, and

22there were follow-on class actions for this, that and

23the other outside of the price-fixing area, and that

24combined with, you know, indirect purchaser statutes and

25all kinds of things that happen in antitrust litigation

199

1generally I think creates the fear that there are some

2legitimate procompetitive practices that the perpetrator

3cannot afford a defense, and I think that is a very

4troublesome phenomenon.

5I guess the thought is provoked by Joe's comment

6that there are -- you know, there is sort of a -- maybe

7we should indulge a presumption that when things are

8loosened up a little, and there are fewer strong ties,

9you know, partial vertical relationships, maybe that is

10the way we want markets to function, but I think the

11system in general works pretty well if we require -- you

12know, we always have the ultimate burden of proof on the

13plaintiff, so that if the defendant can come up with a

14sensible justification, a justification that can be

15persuasive with the fact-finder, then yes, the right

16standard is, if the defendant has something good to say

17for his practice, let's adopt a rule that the plaintiff

18does not win unless the plaintiff persuades that the

19negative effect on competition outweighs the

20procompetitive effect.

21And true enough, part of what I was saying

22earlier is, yes, it is the wiles of economic theory. It

23is the unadministerable, you know, battle between the

24economic experts and all the other facts in the case,

25but what is the alternative? The alternative is

200

1Standard Stations, or worse, and we know that is wrong,

2so that is why I would like -- I keep trying to bring

3into the conversation this institutional element.

4Let's not -- once we decide it is a balance,

5let's not just throw confetti in the air. Let's try to

6focus on what the applied micro tells us about what

7rationales deserve to be explored and what facts could

8rule various theories of efficiency or theories of

9restraint in or out. Let's organize that process so we

10do not just have a U.S. versus IBM every time there is,

11you know, a 13-year slog or a 14-year slog like Harmar

12versus Coca-Cola every time we have a difficult

13exclusive dealing issue.

14DR. CALKINS: I really misunderstood you, Tad.

15I thought when you said you wanted to go to the 18-month

16model, you wanted to go back to the days of Standard

17Stations, and I just --

18DR. KLEIN: Per se. But to answer your question

19about whether we have the tools to do this, I guess

20economists have the tools -- I was on a panel with Steve

21Salop where I said I -- even if I were the judge, I

22wouldn't know exactly how to do it, and he said, you

23know, that is all economists know how to do, you know,

24want to take away your doctorate or something, but when

25you -- obviously you have to go to balancing.

201

1I mean, I am pretty cynical about this, because

2I do not know -- I do not think the courts have done

3this, and I do not know what to tell them to do. I

4mean, I think they go backwards, and they figure out --

5you know, they do some kind of implicit balancing, and

6then they say -- they make it easy and they say it was

7not an anticompetitive effect or there is no

8procompetitive efficiency rationale, and I do not know

9what exactly we should have them do, other than we know

10we want them to hire more economists, right?

11But it is a -- I think that is the ultimate

12question, because you do have to do the balancing, and I

13do -- I mean, it is a legal question, but I do think the

14burden should be placed on the plaintiff at that point,

15because I have this prior bias about the competitive

16process. So, I agree with the legal rule, but then what

17exactly are you doing -- and it should -- it should not

18be a close thing, because that is my -- and I think that

19is the way the law is or it should be, that it should

20not be a very close thing that we are balancing, and it

21should not be something -- you know, there should be

22this first step that you have to show a very clear

23anticompetitive effect before you go forward in any way,

24and that is going to get rid of most of the cases.

25Steve will say that is why the defendants win

202

1all the time, but they do not always win, because you

2have the Dentsplies and you have the Microsoft, and I

3think that is enough to get efficiency in the economy.

4DR. FARRELL: There is this article by Priest

5and Klein -- I do not know if that is you --

6DR. KLEIN: Yes, that is me.

7DR. FARRELL: -- saying that whatever the rules

8are, the litigated cases are going to be close ones.

9So, I do not think we can have a rule that litigated

10cases are not allowed to be close.

11MR. O'BRIEN: Okay, well, we have run past our

12time, and I think it is Ben's fault, by about four

13minutes. So, thank you very much everybody.

14(Applause.)

15(Whereupon, at 4:04 p.m., the hearing was

16concluded.)

17

18

19

20

21

22

23

24

25

203

1C E R T I F I C A T I O N O F R E P O R T E R

2DOCKET/FILE NUMBER: P062106

3CASE TITLE: SECTION 2 HEARING

4DATE: NOVEMBER 15, 2006

5

6I HEREBY CERTIFY that the transcript contained

7herein is a full and accurate transcript of the notes

8taken by me at the hearing on the above cause before the

9FEDERAL TRADE COMMISSION to the best of my knowledge and

10belief.

11

12
 DATED: 12/4/2006


13

14

15

16
 SUSANNE BERGLING, RMR-CLR


17

18C E R T I F I C A T I O N O F P R O O F R E A D E R

19

20I HEREBY CERTIFY that I proofread the transcript

21for accuracy in spelling, hyphenation, punctuation and

22format.

23

24

25
 DIANE QUADE